Promoting Green Urban Development in Africa: Enhancing the relationship between urbanization, environmental assets and ecosystem services PART II: EVALUATING THE POTENTIAL RETURNS TO INVESTING IN GREEN URBAN DEVELOPMENT IN DURBAN Promoting Green Urban Development in Africa: Enhancing the relationship between urbanization, environmental assets and ecosystem services PART II: EVALUATING THE POTENTIAL RETURNS TO INVESTING IN GREEN URBAN DEVELOPMENT IN DURBAN This page intentionally blank. Authors Jane Turpie, Gwyneth Letley, Robynne Chyrstal, Stefan Corbella and Derek Stretch Prepared for AECOM on behalf of The World Bank by Anchor Environmental with support from The Nature Conservancy Prepared by Anchor Environmental Consultants 8 Steenberg House, Silverwood Close, Tokai 7945 www.anchorenvironmental.co.za 2017 COPYRIGHT PREFACE AND ACKNOWLEDGMENTS © 2017 International Bank for Reconstruction and Development / The World Bank This study forms one of the case studies of a larger study on Green Urban Development commissioned by the World Bank and co-funded by The Nature Conservancy. Anchor Environmental Consultants (Anchor) was subcontracted 1818 H Street NW by AECOM to undertake case studies in three cities: Kampala, Uganda; Dar es Salaam, Tanzania; and Durban, Washington DC 20433 South Africa. Each city was consulted as to the focus of the case study. In the case of Durban, the city requested a Telephone: 202-473-1000 study to evaluate Durban’s natural capital and its role in Green Urban Development (GUD). The study is made up Internet: www.worldbank.org of two parts. The first part consists of providing an updated, spatial estimate of the value of natural capital in the eThekwini Municipal Area and the second part involves undertaking a scenario analysis to evaluate the potential This work is a product of the staff of The World Bank with external contributions. The findings, interpretations, and returns to investing in GUD with a focus on the role of natural systems. This study builds on the preparation of an conclusions expressed in this work do not necessarily reflect the views of The World Bank, its Board of Executive Environmental Profile for eThekwini Municipality by AECOM. Directors, or the governments they represent. The study was led by Dr Jane Turpie of Anchor Environmental Consultants. Dr Liz Day of Freshwater Consulting Group The World Bank does not guarantee the accuracy of the data included in this work. The boundaries, colors, and Gwyn Letley of Anchor Environmental undertook the ecological and green urban design aspects of the study and denominations, and other information shown on any map in this work do not imply any judgment on the part of The associated costings. Dr Robynne Chrystal of CCS consulting undertook the hydrological modelling work under the World Bank concerning the legal status of any territory or the endorsement or acceptance of such boundaries. guidance of Prof Derek Stretch, and Dr Stef Corbella of CCS Consulting prepared the infrastructure costing model. April 2017 We are grateful to the eThekwini municipal staff for their interest and support of this project, in particular to eThekwini Municipality to the eThekwini Environmental Planning and Climate Protection Department for providing Rights and Permissions relevant GIS data and associated explanations for the Durban Metropolitan Open Space System. The material in this work is subject to copyright. Because The World Bank encourages dissemination of its Thanks to Roland White and Chyi-Yun Huang of the World Bank and Diane Dale, Brian Goldberg and John Bachmann knowledge, this work may be reproduced, in whole or in part, for noncommercial purposes as long as full attribution of AECOM, Dr Timm Kroeger of TNC, Jeff Wielgus and Mike Toman of the World Bank for the inputs into the study to this work is given. design and comments on an earlier draft. Any queries on rights and licenses, including subsidiary rights, should be addressed to the Publishing and Knowledge Division, The World Bank Group, 1818 H Street NW, Washington, DC 20433, USA; fax: 202-522-2625; e-mail: pubrights@worldbank.org. EXECUTIVE SUMMARY Introduction livelihood and property value losses which in turn affect Urbanisation is taking place at an unprecedented municipal finances and GDP outputs. rate throughout the world, often outpacing plans and Green urban development is an approach that aims the capacity of city managers. As a result, natural to minimize the impacts of urbanization on the open space areas in cities are being degraded and environment, and tackles the core problems of pollution diminished, and problems such as flooding, air and and waste, the consumption of natural resources, the water pollution are getting worse. The environmental loss of urban open space and the degradation and problems associated with increased hardened surfaces loss of biodiversity, as well as mitigation of the urban and the loss of natural areas and ecosystem services contribution to climate change. In addition to a range are particularly acute in developing country cities, of policy interventions, this involves investing in natural where a lack of regulation and resources has led to poor capital as well as use of green structural engineering planning, the expansion of informal settlements in high and conventional grey infrastructure. Green urbam risk, marginal areas, and the inability to adequately development includes (1) sanitation services and manage the quantity and quality of surface water flows. regulations to minise pollution, (2) applying “green While conventional storm water conveyance measures engineering” approaches to urban problems such as contribute to reducing flooding impacts, they have not stormwater management, (3) controlling consumption been able to keep ahead of the problem and have also and carbon emissions, (4) protecting natural assets and contributed to pollution and degradation of downstream (5) maintaining parks, street trees and gardens. aquatic systems. The aim of this study was to explore, using a case study However, great strides have been made in the design and scenario-based approach, the potential costs and of more sustainable engineering mechanisms to deal This page intentionally blank. benefits of undertaking a green urban development with urban flooding and water quality problems, and approach to address some of the main environmental the management and planning of cities is increasingly issues described above, and to explore the potential focusing on a more holistic approach that includes the tradeoffs between different types of interventions, conservation of natural areas as part of a green urban with an emphasis on assessing the desirable balance development (GUD) strategy. A GUD strategy does not between engineered interventions and the conservation only focus on surface water issues but also involves the of natural open space areas. The study focuses on three maintenance of natural open space areas for recreation elements of green urban development, all of which which is essential for human health and wellbeing. impact on ecosystems and biodiversity: sewage and solid One of the challenges of green urban development will waste management, active stormwater management be to find the right balance between natural, semi- and the conservation of natural systems and riparian natural, innovative and conventional built infrastructure. corridors. Understanding the costs and benefits associated with the different types of measures is important and The study involved modelling current flooding and requires careful consideration of their potential benefits water quality in the Umhlatuzana - Umbilo catchment, and cost effectiveness in managing urban environmental and determining the change in water quality and flood problems. hydrographs under a series of hypothetical scenarios in which the past development of the area had involved Durban, located within the eThekwini Municipality on different combinations and extents of green urban the east coast of South Africa, is rich in biodiversity, but development measures including better sanitation, faces a number of environmental and developmental stormwater management and conservation measures. challenges. While green open space areas make up some The economic implications of these changes were 33% of the total area within the municipality, less than assessed in terms of implications for aquatic ecosystem one third of this falls within the urban edge and only health as well as the infrastructure costs, and losses in about 10% is formally protected and just under 7% is property, tourism and fishery benefits that would have actively managed. Rapid urbanisation and the continued been avoided under these alternative scenarios. The expansion of informal settlements contributes to the relative costs and benefits of different scenarios were degradation and loss of natural systems and biodiversity. then evaluated using a cost benefit approach. Key environmental issues include more frequent and intense flooding events, solid waste pollution, elevated flows and nutrients as a result of increases in wastewater Study area: the Umhlatuzana-Umbilo Catchment outflows, erosion, poor air and water quality, overexploitation of natural resources, and the spread of The Umhlatuzana-Umbilo catchment is situated in the alien invasive species. All of these issues contribute to centre of the eThekwini Municipal Area (EMA) and rising infrastructure and human health costs as well as covers an area of approximately 272 km2. It incorporates EVALUATING THE POTENTIAL RETURNS TO INVESTING IN GREEN URBAN DEVELOPMENT IN DURBAN PAGE i Durban’s city centre, it’s harbour (Durban Bay) and rated as Good in the middle reaches of the Umbilo River important industrial areas and has a significant formal and Poor in the lower reaches. Invertebrate and fish residential population. The catchment is undulating communities in this river are rated as Poor. Water quality with steep river valleys in the upper and middle conditions in the upper catchment of the Umhlatuzana catchment, while the lower catchment is relatively River, in particular, are considered acceptable – a result flat. The catchment has a natural mean annual runoff of the lower density of settlements and the presence (MAR) of 43.25 million m3 and the two main rivers are of largely unmodified riparian buffers in this part of the the Umhlatuzana River and the Umbilo River which flow catchment. into Durban Bay. The EMA has a subtropical climate with humid wet summers and mild dry winters. The mean annual precipitation is just over 1000 mm and most of Design and potential extent of selected types of the rain falls between September and March. interventions Sewage and solid waste management Much of the catchment has been developed and is dominated by urban settlement, commercial and Within the catchment study area there are a number industrial land use, with some agriculture in the upper of informal settlements without adequate sanitation. ES Figure 1 Different types of measures used in stormwater management (Source: this study) parts. There are a number of informal settlements Without adequate sanitation raw sewage, litter in the catchment and these tend to be located along and other pollutants end up in rivers and streams. steep river banks and in river floodplains. There are Under green urban development, these and any new water supply (in the case of rainwater harvesting) and release water, so that they are ready for the next high around 6000 ha of natural vegetation in the catchment, settlements would be provided with urine diversion the provision of sports and recreational opportunities. rainfall event, such as detention basins and extended dominated by woodland and forest. Most of the intact dehydration toilets and adequate services as a minimum The latter is particularly the case for the vegetated detention or infiltration trenches. It was necessary to forest is found along the steep river valleys in the requirement. options which have greater aesthetic appeal. Green consider localised measures that capture and/or slowly upper and middle catchment. The North Park Nature measures include both engineering solutions and the infiltrate runoff at source as well as regional measures Reserve, Kenneth Stainbank Nature Reserve and Bluff Active stormwater management (“green engineering”) protection or restoration of natural systems in riparian that capture runoff off-site to achieve meaningful overall Nature Reserve provide some protection to these Urban drainage management has changed significantly and catchment areas. Within flood-prone areas, scale of impact and thereby also creating a treatment natural systems. Durban Bay is one of South Africa’s over the last few decades, from a conventional conservation of natural green infrastructure such as train process. Selected source control measures included larger estuaries which in spite of a high degree of ‘rapid disposal’ approach to a more integrated and riparian buffers, functional floodplain areas and large infiltration trenches, subsurface soakaways, permeable transformation is still of conservation importance. At the sustainable ‘design with nature’ approach. There has conservation areas can potentially enhance the value paving and green roofs. Regional measures included head of the estuary a small 15 ha pocket of mangroves been a proliferation of related approaches going under of development setbacks and conveyance measures. detention basins and treatment wetlands. are protected as part of the Bayhead Natural Heritage terms such as Integrated Urban Water Management Non-structural measures can also be considered “green” Site. (IUWM), Water Sensitive Urban Design (WSUD), urban but do not involve physical construction, rather using Conservation areas and riparian corridors stormwater Best Management Practices (BMPs), policies and laws, public awareness raising, training and Among the most feasible options identified were the Water pollution and flooding are two of the main Sustainable Urban Drainage Systems (SUDS) and Low education to enhance urban areas and reduce risks and protection, restoration and/or enhancement of natural environmental issues associated with this catchment. Impact Development (LID). These describe a number impacts of anthropogenic activities. systems. The maximum possible extent of each of these Pollution comes from a number of different sources of measures to address flooding and/or water quality options was determined on the basis of GIS information including non-point source pollution such as problems. These tend to be categorised into passive This study sought to find a suitable set of “green” on land use, soils and slope. pesticides, fertilisers and industrial and residential and active structural measures. Passive measures aim to measures that could be implemented in combination runoff, stormwater outflows, point source pollution convey water and protect areas from flooding whereas to address flooding and water quality problems in Currently there are approximately 6000 ha of natural from industrial discharge points (factories) and urban active measures aim to modify the hydrograph and the study area, while also contributing to a green vegetation in the catchment as per the D’MOSS plan. infrastructure (WWTW) resulting in high nutrient address water quality by retarding water movement and urban development path for the city. A review was Taking into account future planning as per municipal concentrations, point source pollution from discharges increasing infiltration or storage. The active measures, carried out on stormwater management options, scheme development zonation plans, the amount from informal settlements, and the presence of alien which seek to reduce the effects of urbanisation on the their efficacy in terms of various criteria, their cost- of natural vegetation in the catchment is reduced to invasive vegetation within the riparian zone resulting in quantity and quality of catchment runoff, can be further effectiveness and the necessary or suitable conditions 2800 ha, or 11% of the total catchment area. It was erosion and sedimentation. Flooding in the catchment categorised into source, local and regional controls, as for their implementation in the Umbilo-Umhlatuzana determined that a total of 7000 ha would be required to is exacerbated by the high levels of litter and dead summarised in Figure I. catchment. Based on this review, GIS land cover data meet conservation targets, accounting for approximately vegetation which block culverts and drains causing and criteria based on slope, depth to groundwater 28% of the catchment area. rivers to overtop and burst their banks. An enormous While conveyance measures tend to be highly effective and soil drainage characteristics, areas of possible amount of plastic litter washes into the rivers and out for reducing flood exposure/risk, they achieve little application were identified and a set of measures that Riparian buffer zones along waterways intercept to sea during high rainfall events. Pollution and flooding water quality improvement, vary in terms of cost- had both a high feasibility of implementation and that sediments, nutrients, pesticides and litter in diffuse has a major impact on the biodiversity and ecological effectiveness and generally produce relatively small would be complementary in terms of their effects on the or sheet surface runoff thereby reducing the amount functioning of the river systems. co-benefits. Indeed, they are more likely to lead to environment in the study area were selected. of pollutants entering rivers and streams. They may externalities such as damage to aquatic ecosystems. also provide habitat and wildlife corridors and can Water quality in the middle and lower reaches of the “Green” (environmentally-friendly) measures that seek Due to unsuitable soil, high ground water levels or be important for reducing erosion, slowing down Umhlatuzana and Umbilo Rivers was assessed by the to ameliorate the impacts of urban development on steepness of certain areas in the catchment, a number floodwaters by increasing roughness (and thus National River Health Programme as Poor, due mainly quantity and quality of flows also vary in their cost- of active structural measures were excluded, including reducing downstream peaks) and providing river bank to urban impacts. Habitat integrity, aquatic macro effectiveness and may have to be applied in combination bioretention areas, grassed swales, filter strips and sand stabilisation. In this study, riparian buffers extended 15 invertebrates and fish are rated as Fair in the middle and/or at scale for effective flood protection, but are filters. Rainwater harvesting tanks were excluded due m on either side for smaller rivers and streams, 30 m on and upper reaches of the Umhlatuzana River but are important for water quality. They also present much to cost as well as the limited flood attenuation benefits. either side of major rivers and on one side of canalised considered Poor in the lower reaches. Habitat integrity is greater opportunities for delivering co-benefits, such as A better option would be to use measures that slowly rivers (1900 ha). PAGE ii EVALUATING THE POTENTIAL RETURNS TO INVESTING IN GREEN URBAN DEVELOPMENT IN DURBAN EVALUATING THE POTENTIAL RETURNS TO INVESTING IN GREEN URBAN DEVELOPMENT IN DURBAN PAGE iii Scenario setup Scenarios 4 to 8 were set up to test the effects of each subcatchment. The modelled runoff flows were Each scenario consisted of a combination of the different In Scenario 2 (GUD without sanitation), all the different amounts of natural areas (conservation coupled with EMC values to estimate the concentration interventions described above, applied to different green engineering and conservation measures were areas and river buffers) on flooding and water quality. and total load of TSS, TIN and TP. Modelled suspended extents in the catchment. Since GUD measures are implemented but the sanitation backlog was not Scenarios 9 to 12 were set up to test the effects sediment loads were multiplied by a factor of 1.25 in unlikely to have much positive impact in the absence addressed. This was to test whether the GUD measures of different combinations of green engineering order to account for bed load. of adequate sanitation, it was decided to include full designed to address water quality would still be effective interventions. Scenarios 13 to 15 were set up to explore the effects of implementing both green engineering and By comparing the modelled sediment and nutrient sanitation (as required by existing legislation) in most if sanitation were not properly addressed. conservation measures to different extents. outputs for each scenario versus the baseline, it was of the scenarios. A total of 15 scenarios were designed Scenario 3 (“Clean Baseline”, BAU + SWM) was the same possible to estimate the difference made by GUD and analysed (Table I). All scenarios assumed that the as Scenario 1 except that the sanitation backlog and interventions to the sediment loads and nutrient catchment was fully developed (as per municipal scheme litter problems were addressed (i.e. informal settlements Scenario modelling concentrations and loads transported to Durban Bay. zonation plans, effective 2014), except that scenarios were serviced and any growth in sewage output was The water quality data were used as inputs into a River with medium or high levels of conservation meant that Several models were used during the scenario analysis balanced by recycling). A comparison of Scenario 3 to Ecosystem Health assessment tool to evaluate river the development was more compact. All scenarios had to determine the potential costs and benefits associated Scenario 1 allowed a test of the effect of sanitation alone system changes, into an Estuarine Ecosystem Services the same number of households and the same amount with using a green urban development approach to on water quality. Scenario 2 was compared to Scenario model to estimate changes in the value of selected of commercial and industrial activity. addressing environmental issues. Inputs into the models 15, and showed that in this particular catchment the services, and into an assessment to determine harbour included the extent and quality of natural areas, the Scenario 1 (BAU) had full development as planned, measures being tested would work with or without dredging costs avoided. extent, design and performance of a range of GUD but with the same level of backlog of sanitation and sanitation, given the relatively small area of unserviced engineering solutions for flood attenuation and water Changes in the quantity of green open space areas solid waste services as at present. The total area of informal settlements. quality amelioration, and the amount and design were used to determine changes in carbon storage. The informal settlements remained the same, i.e. 1% of total capacity of conventional conveyance and waste water The remaining scenarios all included full sanitation and carbon storage value for the EMA was used to determine catchment area. No green engineering measures were treatment infrastructure. litter prevention programmes, so “+SWM” is implied a per hectare cost for D’MOSS which was then used implemented and the amount of natural open space in Scenarios 3-15. Scenario 3 effectively provided a to determine a carbon storage costs for the minimum, was reduced to the planned extent of 2800 ha. This A hydrologic (hydrology + hydraulic) model was set “Clean Baseline” against which to evaluate the relative medium and maximum extents of conservation areas was termed the “Baseline.” Note that the baseline is up for the Umhlatuzana-Umbilo catchment using the net benefits of different engineering and conservation under each scenario. not the same as the present-day situation, which was PC-SWMM software. This model was set up to run used to develop and calibrate the models, and against measures applied to different extents. Scenarios design flood events in order to determine the influence Changes in the quantity and quality of green open space which water quality estimates are also compared out of 4-15 were compared with Scenario 3, under the prior of green urban development interventions on flood areas were inputs into a Tourism Value model and a interest. assumption that adequate sanitation is both imperative hydrographs at strategic points relating to the location Hedonic Pricing Model, which estimated differences and a prerequisite to GUD. of existing flood conveyance infrastructure. The flood in tourism and property values, respectively. The hydrographs generated under baseline conditions (fully- comprehensiveness of these models varied according to ES Table I List of all scenarios and the interventions applied to different extents under each scenario developed catchment as planned, business as usual) priorities, availability of potentially suitable modelling Conservation / were compared with hydrographs generated under platforms and models, and data availability. The Estuary Sanitation and waste Green engineering (GE) / Non-structural sw each of the different GUD scenarios. This provides an management (SWM) Active stormwater management management Ecosystem Services Model, Tourism Value Model and indication of the impacts of GUD interventions and the Hedonic Pricing Model built on models developed for Source Detention Treatment Riparian Conservation difference can be construed as an estimation of the Litter Controls Basins wetlands buffers areas the municipality-wide eThekwini ecosystem services flood attenuation benefit obtained from implementing valuation study (Turpie et al. 2017). A cost benefit Scenarios Sanitation removal S D W R Cons these interventions in the catchment. The additional analysis (CBA) was set up to assess the overall impacts 1. Baseline: flood volumes without these interventions that would BAU (sanitation - - - - - - ● and benefits associated with the GUD approach. Sanitation occur under different return period flood events, would backlog backlog) require larger drains, culverts, and other conveyance 2. GUD without infrastructure, depending on the size of the event these sanitation - ● ●● ● ● ● ●●● Impacts on sediments and water quality constructed flood management assets are designed to 3. Clean baseline: deal with. Thus a second model was developed in order Both green engineering and natural interventions ● ● - - - - ● significantly reduced sediment loads into the harbour. BAU + SWM to estimate the capital costs of the structures required 4. Cons2 ● ● - - - - ●● under the baseline versus GUD scenarios. The difference, Conservation areas and riparian buffers (Scenarios together with associated differences in discounted 4 – 8) had a significant impact on reducing TSS loads Conservation 5. Cons3 ● ● - - - - ●●● measures annual maintenance costs, was the total life-cycle flood into the harbour. An increase in conservation area 6. R ● ● - - - ● ● from the minimum to medium and or maximum extent infrastructure cost saving from the GUD measures. 7. R + Cons2 ● ● - - - ● ●● translated into 18% and 32% reductions in TSS loads, 8. R + Cons3 ● ● - - - ● ●●● A sediment and nutrient model, also set up in PC- respectively. The addition of river buffers led to a further 9. S2 ● ● ●● - - - ● SWMM, produced water quality outputs in the form of 31-36% reduction in the annual TSS loading, and with engineering Total Suspended Solids (TSS) and nutrient concentrations the maximum extent of natural interventions (Scenario 10. D ● ● - ● - - ● Green and loads at specific points of interest in the catchment. 8) annual TSS loads entering the harbour were reduced 11. S+D ● ● ● ● - - ● by 63%. Source controls and detention basins each had The pollutant washoff from a given landuse during 12. S2+D ● ● ●● ● - - ● periods of wet weather was characterized in the model a relatively low impact on TSS loads into the harbour. 13. S+D+R ● ● ● ● - ● ● by using a user-defined Event Mean Concentration However their effect was noticeable at certain points Combined 14. S2+D+W+R+Cons2 ● ● ●● ● ● ● ●● (EMC). Model subcatchment parameters were derived higher in the catchment where they had significant by area-weighting the various land use parcels within impacts on reducing TSS loads by up to 90%. Avoided 15. S2+D+W+R+Cons3 ● ● ●● ● ● ● ●●● PAGE iv EVALUATING THE POTENTIAL RETURNS TO INVESTING IN GREEN URBAN DEVELOPMENT IN DURBAN EVALUATING THE POTENTIAL RETURNS TO INVESTING IN GREEN URBAN DEVELOPMENT IN DURBAN PAGE v harbour dredging costs ranged from R0.46 million per year and conservation areas also played a major role in Gains in ecosystem values a prerequisite to GUD, and to specifically focus on the for Scenario 4 to R1.75 million per year for Scenario 15. determining overall river habitat integrity, through the The decrease of nutrient and sediment loads into Durban potential added benefits of the less conventional GUD provision of longitudinal corridors for faunal movement Bay led to gains in estuarine and marine fishery values measures. Scenarios 4 – 8 and Scenario 10 had positive Recycling of wastewater and improved sanitation had (important in an increasingly urban environment). net present values (NPV). NPVs increased with increasing ranging from 9% for Scenario 4 to 55% for Scenario 15, the greatest impact on reducing nutrient loads into the conservation area from riparian buffers alone (Scenario worth R0.6 - R3.5 million per annum. harbour. This led to a 30% reduction in total inorganic 6) to riparian buffers in conjunction with the full area nitrogen (TIN) loads and a 35% reduction in total Impacts on flooding and infrastructure requirements The effect of a more compact but greener development of natural open space required to meet conservation phosphorous (TP) loads entering the harbour each year. approach on property value was estimated to be targets (Scenario 8), with the latter having a NPV of R5.9 Taken alone, green engineering measures were Neither natural areas nor green engineering measures significant, with property premiums associated with billion (using a discount rate of 6% over 20 years). This more efficient at reducing peak flows than natural (other than treatment wetlands) had much additional natural vegetation in a good condition increasing from was because maintaining natural areas had relatively low interventions, although the maxium extent of effect on nutrient loads entering the habour, although R887 million for the minimum extent to R1.8 billion for costs and high benefits. However, the estimates derived conservation and riparian areas was highly effective. significant localised effects were seen in the catchment. the medium conservation extent and R3 billion under from this study suggest that scenarios that include the Riparian buffers were the least effective intervention. The addition of treatment wetlands (Scenarios 14 and 15) the maximum conservation extent. implementation of source control measures (Scenarios T However, the more compact development scenarios resulted in significant further reductions in TIN and TP, 9, 11-15) would have a negative NPV. This was due to with bigger conservation areas had a significant impact with Scenario 15 reducing TIN by 63% and TP by 53 %. The annual nature-based tourism value associated with the very high costs of these measures, particularly of on peak flows with an 8% reduction in the impervious surface area in the catchment translating to a 15% natural open space areas in the U60F catchment was soakaways in residential areas. Scenario 10 (detention decrease in peak flows during a 2- and 5- year return estimated to be R205 million under the minimum extent, basins only), had a positive NPV, of R180 million, due to Impacts on river health R439 million for the medium extent and R512 million the relatively low costs of implementing this measure. period floods. When the ratio of conservation to Changes in river condition or health associated with each developed area was increased in the upper catchment, under the maximum conservation area scenarios. This This scenario had a higher NPV than Scenario 6; the of the different scenarios were modelled in terms of a significant reduction in peak flows was seen. While translates into a net present value of R2.35, R5.04 and result of higher cost infrastructure savings. instream water quality, which was interpreted using the there was a significant improvement from minimum to R5.87 billion for the minimum, medium and maximum South African national guidelines for determining river extents of conservation areas, respectively. We did The results suggested that the compact development medium conservation, the difference between medium health on a scale from A to F. The analysis focussed on not estimate the effect of different scenarios on the options with larger proportions of green open space and maximum conservation scenarios was relatively total suspended solids (TSS), total phosphorous (TP) and very high tourism value associated with beaches. The were far more effective than using engineering small because much of the added conservation area was total inorganic nitrogen (TIN). management of the U60F catchment is not expected measures alone. The open space areas not only deliver in the lower catchment. to have a major impact on these values, but the ecosystem services relating to the primary stormwater The present-day concentrations of water quality On average the source control measures reduced peak management of several other eThekwini catchments management objectives but also directly provide variables in the river systems generally follow expected flows by 10%. Soakaways in residential areas contributed would be expected to have this added effect. amenity value that is translated into property and patterns for urban environments, with the highest the most to this reduction, which was expected given tourism values. It is acknowledged, nevertheless, that concentrations of TIN and TP occurring in the dry their large scale of implementation compared to the Increased areas of land under conservation result the latter benefits were more difficult to estimate than season (winter), when dilution is lowest, and TSS other source controls. Detention basins were found to in increased increased avoidance of climate change other benefits such as the engineering cost savings, tending to be lowest in the dry season, when flows be more effective at reducing peak flows during small damages. Based on the extent of D’MOSS under each and therefore have some degree of uncertainty. The are too low to mobilise sediments. Phosphorus-based to medium return period floods than higher return of the conservation extents, it was estimated that most uncertain estimates were those of tourism value, enrichment in the catchment is significant, with all but period floods. During small to medium return period the damage costs avoided by retaining carbon stocks for which we have used average values in the absence the upper reaches of the Umhlatuzana River being in an floods, peak flows at the bottom of the catchment were within the Umhlatuzane-Umbilo catchment would be of reliable estimates of marginal changes in value unacceptably poor state (Category E or F on a scale of A reduced by 9-35%. The reduction was higher at points in approximately R3.2 million, R2.7 million and R1.3 million associated with changes in the”green-ness” of Durban. to F). . TIN concentrations were also bad, althoughnone the upper catchment. per annum for the maximum, medium and minimum Nevertheless, our estimates could be considered as of the sites were in a worse category than Category E. extents of conservation areas considered. This translates conservative; firstly, the demand for conservation areas If all interventions were applied together, peak flows to a net present value of R37 million, R31 million and R15 is likely to increase as the city grows and incomes rise. A BAU approach will result in worsening of water were reduced by 45 – 50% for the smaller return million for carbon storage, respectively. Secondly, the cost-benefit analysis does not take into quality compared to present. Recycling and sanitation periods (0.5- and 1-year), 25 – 30% for the medium account a range of other potential benefits of the GUD (Scenario 3) resulted in significant improvement in river return periods (2- and 5-year) and by 15 – 20% for In addition to the above, retaining natural areas in interventions such as air quality and the existence value condition, particularly in reaches previously affected the high return periods (10- and 20-year). However, the catchment also provides biodiversity benefits. of biodiversity. by runoff from poorly serviced informal settlements. combining conservation areas and green engineering Larger conservation areas retain more viable species While riparian buffers and conservation areas (scenarios interventions has a small additional effect relative to populations and together with riparian corridors To provide some perspective, the initial capital 4-8) yielded improvements in terms of TP levels, green other scenarios, which suggests that green engineering provide ecological connectivity which is critical for requirement associated with each scenario (Table III) was engineering measures alone or in combination wtih and conservation interventions (in combination with the movement of organisms and the resultant flows compared to the eThekwini Municipality capital budget conservation measures (Scenarios 9 – 15) achieved the compact development) are largely substitutable in terms of ecosystem services. Ecological connectivity is of R6.73 billion for 2016/17 (eThekiwini Municipality most significant impacts, generally accounting for an of their effects on flooding. particularly important in urban systems where the 2015). Scenarios 4 – 8 and scenario 10 have a funding improvement in water quality condition by at least one location and distribution of green open space is critical requirement that is less than 1% of the municipal capital PES category. The dramatic improvement was assumed The different interventions reduced capital cost for delivery of ecosystem services. budget. Scenarios 11 and 12 would require 13.8% and in part to result from the fact that the measures are requirements for flood conveyance by R19 million 11.4% of the capital budget respectively and scenario 15 implemented across the catchment. Treatment wetlands (riparian buffers) to R226 million (all interventions). would require 96.1% of the budget. The requirements exerted significant effects on water quality in the . The relatively small saving of 0.5 – 6% compared to Cost benefit analysis for scenarios 9, 13, and 14 are all higher than the reaches where they were implemented. the percentage change in flood sizes is to some extent Costs and benefits were examined relative to the proposed capital budget for the municipality, however, an artefact of the present tendency for overdesign of scenario of full development with adequate sanitation it is expected that capital requirements for large-scale Interventions such as riparian buffers appeared to have conveyance infrastructure (due to solid waste, etc.). (Scenario 3; Table III), under the assumption that implementation would occur over several years (or little effect on instream nutrient concentrations but adequate waste management is both imperative and decades) as an area develops. had significant implications for TSS. Riparian buffers PAGE vi EVALUATING THE POTENTIAL RETURNS TO INVESTING IN GREEN URBAN DEVELOPMENT IN DURBAN EVALUATING THE POTENTIAL RETURNS TO INVESTING IN GREEN URBAN DEVELOPMENT IN DURBAN PAGE vii ES Table II Total present value of costs and benefits and NPV (R millions) for all scenarios (2015 Rands, 6% discount rate, 20 years), initial capital estimating the impacts of these kinds of interventions costly at today’s prices, even when accounting for requirement and the capital requirement as a percentage of the annual capital budget for each scenario. at scale. Most previous studies have analysed these economies of scale. WOn the other hand, the simpler Total present Total present Initial capital problems at a micro-catchment scale, making it difficult green engineering measures, such as detention basins value of costs value of benefits NPV requirement % of EM capital to assess the economic implications of a change in for reducing peak flows and treatment wetlands for Scenario (R millions) (R millions) (R millions) (R millions) budget 2016/17 policy. Thus, while these are first-cut estimates which improving water quality, were shown to be highly cost- 4. Cons2 323 3 718 3 394 2 0.0% warrant further refinement, they provide a useful effective, viable interventions. 5. Cons3 246 5 832 5 586 3 0.0% step towards the informing policies to guide the city’s 6. R 16 67 50 1 0.0% sustainable development path. The retention of signifant natural areas within the catchment, which may require more compact 7. R + Cons2 204 3 754 3 550 4 0.1% The results showed that in a catchment with little development, was not only found to be highly effective 8. R + Cons3 262 5 992 5 730 5 0.1% unserviced informal settlement (1% of the area) recycling at reducing sedimentation and flooding problems, 9. S2 12 344 293 -12 052 8 960 133.1% the equivalent of a third of sewage outputs along with but has the added benefit of yielding high amenity 10. D 27 207 179 22 0.3% complete sanitation services has the most significant value realised as property and tourism value as well 11. S+D 1 546 231 - 1 315 928 13.8% impact on nutrient loads entering the rivers and harbour, as intangible and unknown values associated with 12. S2+D 12 372 355 - 12 016 8 982 133.5% and that treatment wetlands make a significant further maintaining biodiversity and ecological connectivity. impact. Sediment loads can be effectively dealt with Although riparian buffers have limited influence on 13. S+D+R 1 284 277 - 1 008 765 11.4% using either natural interventions (particularly river water quality and flooding in urban environments, the 14. S2+D+W+R+Cons2 9 888 3 955 - 5 933 7 035 104.5% buffers) or green engineering interventions (source value in maintaining biodiversity is also very high, as 15. S2+D+W+R+Cons3 9 181 6 148 - 3 033 6 464 96.0% controls and detention basins), and these appear to they are e critical for providing connectivity between be substitutable to a large degree when the former is terrestrial systems, rivers and estuaries. brought about through more compact development. The CBA results were subjected to sensitivity analysis Conclusions Simillarly, both natural and green engineering Because conservation with compact development incurs by varying assumptions of certain GUD intervention Under a business as usual scenario, the continued interventions were highly effective at reducing flood very low costs in comparison to other interventions, unit costs and benefits, and discount rate. Alternative growth of urban areas in Africa will result in a further peaks, and were also substitutable to a large degree in the net benefits of this strategy far outweighed any discount rates of 3% and 9% were applied to the NPV deterioration of the natural environment and living terms of this function. This suggests that some sensible other. Compact development coupled with the other calculations. The alternative discount rates examined conditions, a loss of values associated with green open combination could be applied, with green engineering interventions creates the greenest city, in terms of did not have a significant impact on NPV calculations, space areas, and increased costs in reducing risks to interventions being strategically located within the water quality and biodiversity conservation goals, and with the NPV remaining negative at discount rates of people that result from environmental problems. The catchment for maximum overall effect. is an economically justifiable strategy in terms of overall 3% and 9% for Scenarios 9 and 11-15 and positive for notion of Green Urban Development is therefore highly costs and benefits.  Maintaining large natural areas and Scenarios 4-8 and Scenario 10. Again, this highlights attractive, as it allows citites to grow in a way that However, not all green engineering interventions riparian buffers should therefore be a primary strategy, the significant influence of soakaway implementation maintains their resilience and maintains standards of are equally viable. Our estimates suggested that the along with the strategic positioning of cost-effective costs on the overall NPV for these scenarios. Soakaways living and quality of life. However, few studies have large-scale application of low-impact stormwater green engineering measures. represent 80% of the cost for Scenario 15 and 89% of investigated what following a more sustainable green management measures (i.e. source controls) is extremely the cost for Scenario 14 but contribute approximately urban development path will actually cost, and whether R124 million in benefits, with R116 million of this being these costs can be justified. Moreover, what Green stormwater infrastructure cost savings. It was found that Urban Development should look like is also not well if soakaways were not implemented but all other source defined, in terms of the degree to which is includes the controls were still included then the NPV for Scenario 15 conservation of river buffers and other natural areas, would be approximately R4.8 billion and for Scenario 14 the mimicking of natural processes through innovative would be R2.6 billion. Or if the unit cost of soakaways engineering design or the protection of downstream was reduced by 8% as a result of economies of scale, areas through conventional engineering measures. competition and improved technologies then the NPV In eThekwini Municipality, these issues have to be for Scenario 15 would be positive at R198 million. considered as plans for the growth of Durban are laid Reducing the extent of soakaway implementation to 62% out. In this study, we tested the idea of green urban of residential areas would also result in a positive NPV development by backcasting a range of scenarios for a for Scenario 15 of approximately R65 million. However, well-developed catchment. the NPV associated with other source control scenarios would remain negative. This study attempted to analyse a highly complex problem in a fairly large catchment area, and as such, The CBA results suggest that source control measures a great deal more work will still be needed in order are costly, particularly the implementation of soakaways to narrow the potential error margin. Our study in residential areas. Detention basins and treatment found a lack of empirical studies to inform modelling wetlands are the most affordable engineering measures, assumptions, which suggests that much benefit could with positive net benefits. Retaining natural areas in be obtained from the implementation and monitoring the catchment has the highest net benefits, due to the of pilot programmes. Furthermore, there are few well- significant co-benefits associated with amenity values. developed modelling platforms that are capable of PAGE viii EVALUATING THE POTENTIAL RETURNS TO INVESTING IN GREEN URBAN DEVELOPMENT IN DURBAN EVALUATING THE POTENTIAL RETURNS TO INVESTING IN GREEN URBAN DEVELOPMENT IN DURBAN PAGE ix ACRONYMS AND ABBREVIATIONS TABLE OF CONTENTS BAU Business as usual LID Low Impact Development BCDA Black Communities Development Act LULC Land Use Land Cover EXECUTIVE SUMMARY I BMP Best Management Practice MAR Mean Annual Runoff ACRONYMS AND ABBREVIATIONS X BWSP Basic Water and Sanitation Programme NTU Nephelometric Turbidity Units CBA Cost Benefit Analysis PES Present Ecological State 1. INTRODUCTION 1 CSCM Catchment, Stormwater and Coastal RB Riparian Buffer 1.1 Background ...........................................................................................................................................................................1 Management RP Return Period DB Detention Basin 1.2 Durban’s environmental issues............................................................................................................................................2 SC Source control DEM Digital Elevation Model 1.3 Green urban development....................................................................................................................................................3 SDF Spatial Development Plan DWA Department of Water Affairs 1.4 Study objectives ...................................................................................................................................................................4 SRTM Shuttle Radar Topography Mission D’MOSS Durban Metropolitan Open Space System 1.5 Study approach ....................................................................................................................................................................4 SUDS Sustainable Urban Development Systems EM eThekwini Municipality TEEB The Economics of Ecosystems and Biodiversity II. THE UMHLATUZANA-UMBILO CATCHMENT AREA 7 EMA eThekwini Municipal Area TIN Total Inorganic Nitrogen EMC Event Mean Concentration 2.1 Location and extent...............................................................................................................................................................7 TN Total Nitrogen EPCPD Environmental Planning and Climate 2.2 Geography and climate.........................................................................................................................................................8 TOC Total Organic Carbon Protection Department 2.3 Historical and current land cover.........................................................................................................................................9 TP Total Phosphorous GDP Gross Domestic Product 2.4 Pollution and flooding...........................................................................................................................................................9 TSS Total Suspended Solids GIS Geographic Information System 2.5 Water quality....................................................................................................................................................................... 12 TW Treatment Wetland GUD Green Urban Development WDT Watershed Delineation Tool 2.6 River condition.................................................................................................................................................................... 12 IDP Integrated Development Plan WMA Water Management Area InVEST Integrated Valuation of Ecosystem Services III. DESIGN AND POTENTIAL EXTENT OF SELECTED GREEN URBAN DEVELOPMENT and Tradeoffs WRC Water Resource Commission INTERVENTIONS FOR THE STUDY AREA 13 IUWM Integrated Urban Water Management WSUD Water Sensitive Urban Design 3.1 Sewage and solid waste management ............................................................................................................................ 13 KZN KwaZulu-Natal WTW Water Treatment Works 3.1.1 � Dealing with sewage................................................................................................................................................ 13 LFTEA Less Formal Township Establishment Act WWTW Waste Water Treatment Works 3.1.2 Managing solid waste.............................................................................................................................................. 13 3.2 Active stormwater management (green engineering)...................................................................................................... 13 3.2.1 Overview of stormwater management.................................................................................................................... 13 3.2.2 Review and feasibility of green engineering measures......................................................................................... 16 3.2.3 Potential extent of selected interventions...............................................................................................................17 3.3 Conservation of natural systems and biodiversity........................................................................................................... 19 3.3.1 Riparian buffers....................................................................................................................................................... 19 3.3.2 Conservation areas and compact development.................................................................................................... 19 IV. SCENARIO SET-UP 23 4.1 Approach............................................................................................................................................................................. 23 4.2 Scenario elements............................................................................................................................................................. 23 4.3 Scenarios ........................................................................................................................................................................... 24 PAGE x EVALUATING THE POTENTIAL RETURNS TO INVESTING IN GREEN URBAN DEVELOPMENT IN DURBAN EVALUATING THE POTENTIAL RETURNS TO INVESTING IN GREEN URBAN DEVELOPMENT IN DURBAN PAGE xi V. SCENARIO MODELLING AND RESULTS 27 A1.3.8 Sand filters............................................................................................................................................................ 72 5.1 Overview.............................................................................................................................................................................. 27 A1.3.9 Bio-retention areas............................................................................................................................................... 73 5.2 Water quality improvements and related benefits........................................................................................................... 28 A1.3.10 Detention basins....................................................................................................................................................74 5.2.1 Impacts on sediment and nutrient concentrations and loads.............................................................................. 28 A1.3.11 Constructed treatment wetlands......................................................................................................................... 75 5.2.2 Impacts on river condition....................................................................................................................................... 32 A1.4 Non-structural interventions...............................................................................................................................................76 5.2.3 Avoided costs due to sediment retention............................................................................................................... 37 A1.4.1 Sweeping and solid waste management..............................................................................................................76 5.2.4 Impacts on estuary condition and fishery values.................................................................................................. 38 A1.4.2 River cleaning and stewardship............................................................................................................................76 5.3 Flood attenuation benefits................................................................................................................................................. 39 A1.4.3 Riparian buffers ................................................................................................................................................... 77 5.3.1 Impacts on flood peaks............................................................................................................................................ 39 A1.4.4 Catchment reforestation...................................................................................................................................... 78 5.3.2 Avoided costs due to flood attenuation...................................................................................................................47 A1.5 Relative performance of different measures................................................................................................................... 79 5.4 Amenity benefits of increased conservation areas.......................................................................................................... 49 A1.5.1 Average cost effectiveness in terms of peak flow and volume reduction......................................................... 80 5.4.1 Amenity value to locals............................................................................................................................................ 49 A1.5.2 Average cost effectiveness in terms of water quality amelioration................................................................... 80 5.4.2 Nature-based tourism value.................................................................................................................................... 50 A1.5.3 Overall effectiveness, cost-effectiveness and potential co-benefits................................................................. 82 5.5 Avoided climate change costs ...........................................................................................................................................51 APPENDIX 2. SCENARIO ASSUMPTIONS 83 VI. COST BENEFIT ANALYSIS 53 A2.1 Sanitation measures.......................................................................................................................................................... 83 6.1 Framework.......................................................................................................................................................................... 53 A2.1.1 Scenario 1-2 (sanitation backlog)........................................................................................................................ 83 6.2 Scenario costs.................................................................................................................................................................... 53 A2.1.2 Scenarios 3-15 (full sanitation) .......................................................................................................................... 84 6.3 Cost-benefit analysis.......................................................................................................................................................... 55 A2.2 Stormwater source controls .............................................................................................................................................. 84 6.4 Sensitivity analysis ............................................................................................................................................................ 58 A2.2.1 Design and extent of the different interventions................................................................................................ 84 A2.2.2 Cost assumptions................................................................................................................................................. 86 VII. CONCLUSIONS AND POLICY RECOMMENDATIONS 59 A2.3 Treatment wetlands ........................................................................................................................................................... 86 VIII. REFERENCES 61 A2.4 PC_SWMM model assumptions ....................................................................................................................................... 87 A2.4.1 Assumptions regarding untreated urban runoff quality..................................................................................... 87 APPENDIX 1: URBAN STORMWATER MANAGEMENT OPTIONS 65 A2.4.2 Assumptions for stormwater management interventions.................................................................................. 88 A1.1 Overview.............................................................................................................................................................................. 65 A1.2 Passive engineering measures to improve conveyance.................................................................................................. 65 APPENDIX 3. FLOOD MODELLING 93 A1.2.1 Drains and swales................................................................................................................................................. 65 A3.1 Model setup........................................................................................................................................................................ 93 A1.2.2 Enlargement of river channel/canalisation/levees/dredging............................................................................ 65 A3.1.1 Baseline information, software and GIS layers................................................................................................... 93 A1.2.3 Hydraulic bypass................................................................................................................................................... 66 A3.1.2 Subcatchment delineation and flow lines........................................................................................................... 94 A1.3 Active engineering measures to retard runoff.................................................................................................................. 66 A3.1.3 Point sources ........................................................................................................................................................ 96 A1.3.1 Permeable pavements ......................................................................................................................................... 66 A3.1.4 Hydraulic parameters .......................................................................................................................................... 96 A1.3.2 Infiltration trenches .............................................................................................................................................. 67 A3.1.5 Soil infiltration ...................................................................................................................................................... 99 A1.3.3 Soakaways (sub-surface infiltration trenches).................................................................................................... 68 A3.1.6 Storm design events........................................................................................................................................... 101 A1.3.4 Green roofs............................................................................................................................................................ 69 A3.1.7 U60F model......................................................................................................................................................... 101 A1.3.5 Rainwater harvesting ........................................................................................................................................... 70 A3.2 Points of interest for scenario analysis........................................................................................................................... 102 A1.3.6 Vegetated swales...................................................................................................................................................71 A3.3 Model calibration ............................................................................................................................................................. 102 A1.3.7 Filter strips ............................................................................................................................................................ 72 A3.3.1 Rainfall selection and application .................................................................................................................... 102 PAGE xii EVALUATING THE POTENTIAL RETURNS TO INVESTING IN GREEN URBAN DEVELOPMENT IN DURBAN EVALUATING THE POTENTIAL RETURNS TO INVESTING IN GREEN URBAN DEVELOPMENT IN DURBAN PAGE xiii A3.3.2 Design rainfall generation for U60F................................................................................................................... 103 LIST OF FIGURES A3.3.3 Available measured data.................................................................................................................................... 103 Figure 1.1 ................1 Schematic representation of Green Urban Development and associated terminology (Source: authors own interpretation). A3.3.4 Calibration of flows and water levels................................................................................................................. 103 Figure 1.2 Schematic diagram of key environmental issues in Durban (divided into “brown issues” and “green issues”), their causes and their A3.4 Assumptions and limitations........................................................................................................................................... 104 consequences (Source: author’s own analysis)...............................................................................................................................................2 Figure 1.3 The main elements of a green urban development policy (Source: authors)...............................................................................................3 A3.4.1 The use of design rainfall................................................................................................................................... 104 Figure 2.1 The location of Umhlatuzana-Umbilo catchment in the EMA........................................................................................................................7 A3.4.2 Groundwater and baseflows............................................................................................................................... 104 Figure 2.2 Elevation map showing the main elevation bands as one moves from the upper catchment to the lower catchment............................. 8 A3.4.3 Stormwater network data................................................................................................................................... 104 Figure 2.3 Average monthly rainfall (mm) for Durban, 1961-2003 (Source: StatsSA 2005)...........................................................................................8 Figure 2.4 Map of the historical vegetation within the Umhlatuzana-Umbilo catchment.............................................................................................9 APPENDIX 4: INFRASTRUCTURE COST ESTIMATE METHOD 105 Figure 2.5 Current landcover in the Umhlatuzana-Umbilo catchment.........................................................................................................................10 Figure 2.6 The Cato Manor informal settlement is situated along the banks of the Mkumbane River.......................................................................10 A4.1 Overview............................................................................................................................................................................ 105 Figure 2.7 Sewage and litter from the informal settlement ends up in the Mkumbane River (Source: Google Earth)..............................................11 A4.2 Identifying existing infrastructure.................................................................................................................................... 105 Figure 2.8 Photos taken of the Umhlatuzana River after a flood event in February 2016 showing the large amounts of plastic and other waste washed into the river...........................................................................................................................................................................11 A4.3 Assigning rainfall return periods..................................................................................................................................... 105 Figure 3.1 Different types of measures used in stormwater management (Source: Turpie et al. 2017). These measures are described in A4.4 Cost estimate of the infrastructure................................................................................................................................. 105 detail in Appendix 1........................................................................................................................................................................................14 A4.4.1 Bridges................................................................................................................................................................. 105 Figure 3.2 ...............17 Soil drainage rates in the Umhlatuzana-Umbilo catchment (Source: South African atlas of agrohydrology and climatology). Figure 3.3 The properties selected as suitable for application of source controls in the catchment (a) green roofs, (b) permeable paving, A4.4.2 Culverts................................................................................................................................................................ 105 (c) infiltration trenches and (d) soakaway pits..............................................................................................................................................18 A4.4.3 Canals.................................................................................................................................................................. 105 Figure 3.4 ......................................................................................20 The potential extent of riparian buffers in the Umhlatuzana-Umbilo catchment. Figure 3.5 The total area of transformed land and natural open space (OS) in the catchment under status quo, minimum conservation, A4.4.4 Pipes.................................................................................................................................................................... 105 medium conservation and maximum conservation.....................................................................................................................................20 A4.5 Flow vs dimension relationship....................................................................................................................................... 106 Figure 3.6 The extent of (a) conservation areas that will remain under existing development plans, (b) current conservation areas and (c) conservation areas that would have met biodiversity targets in the catchment........................................................................................21 A4.5.1 Pipe scaling......................................................................................................................................................... 106 Figure 5.1 The approach used for the scenario analysis and determining the costs and benefits associated with green urban development......27 A4.5.2 Culvert scaling..................................................................................................................................................... 106 Figure 5.2 Annual TSS load (tonnes) for different outfalls into Durban Harbour for all scenarios...............................................................................29 Figure 5.3 The annual TSS loadings at different points along the Umbilo River (left) and the Umhlatuzana River (right). Refer to Figure A6. 1 A4.5.3 Estimating the flow type..................................................................................................................................... 106 for the monitoring station locations in the catchment. . .............................................................................................................................30 A4.6 Cost comparison............................................................................................................................................................... 106 Figure 5.4 Annual TP and TIN loads for different outfalls entering Durban Harbour for all scenarios........................................................................31 A4.7 Additional information ..................................................................................................................................................... 106 Figure 5.5 Effects of different modelled scenarios on concentrations of Total Phosphorus and Total Suspended Solids (TSS), respectively. Site locations as shown in Figure A6.1. .........................................................................................................................................................34 Figure 5.6 Effects of different modelled scenarios on total inorganic nitrogen (TIN) concentrations at different monitoring sites in the APPENDIX 5. SEDIMENT AND NUTRIENT MODELLING 109 study area. Site locations as shown in Figure A6.1........................................................................................................................................35 A5.1 Model setup...................................................................................................................................................................... 109 Figure 5.7 Schematic summary of the linkages from water quality parameters to estuary ecosystem services for a case of deteriorating ..................................................................................................................................................................................................38 water quality.. A5.1.1 Rainfall................................................................................................................................................................. 109 Figure 5.8 SWMM model of the U60F catchment including flow paths (yellow and red lines), detention basins (green squares), WWTWs (larger black dots), water quality monitoring stations (smaller black dots), extent of source controls for commercial and A5.2 Points of interest for scenario analysis........................................................................................................................... 110 industrial areas (grey shaded) and residential areas (purple shaded areas). The Umhlatuzana-Umbilo catchment, the Amanzimyana Stream catchment and all other contributing catchments have been outlined.................................................................41 A5.3 Model calibration ............................................................................................................................................................. 110 Figure 5.9 Summary of the simulated peak flows for all scenarios and all return periods for the Umhlatuzana/Umbilo Canal outfall into the harbour............................................................................................................................................................................................................42 APPENDIX 6: RIVER ECOSYSYEM HEALTH ASSESSMENT 113 Figure 5.10 Summary of the relative change in peak flows compared to the baseline for all scenarios and all return periods for the Umhlatuzana/Umbilo Canal outfall into the harbour. . ................................................................................................................................42 A6.1 Approach for assessing river condition........................................................................................................................... 113 Figure 5.11 Percent impervious surface area and associated relative change in peak flow for the conservation scenarios compared to the A6.2 Model outcomes............................................................................................................................................................... 116 baseline . .........................................................................................................................................................................................................43 Figure 5.12 Simulated peak flows for a 2- and 5-year rainfall return period for the Baseline and Scenarios 4 and 5 at sampling points (a) R_Mkumbaan_01 and (b) R_Zana_35..................................................................................................................................................... 44 Figure 5.13 Hydrographs showing a 2-year rainfall return period flood for Scenarios 9, 11 and 12 at sampling points (a) Umbilo_27, (b) R_ Mkumbaan_01, (c) R_Chats_15 and (d) R_Zana_35. Note that time units are reporting time steps from the model with one time unit representing 5 minutes.................................................................................................................................................................. 44 Figure 5.14 Simulated peak flows for a 2-, 5-, 10- and 20-year rainfall return period for Scenario 10 compared to the Baseline at sampling points (a) R_Zana_35, (b) R_Zana_29, (c) R_Zana_10 and (d) R_Mkumbaan. Note that time units are reporting time steps from the model with one time unit representing 5 minutes........................................................................................................................45 PAGE xiv EVALUATING THE POTENTIAL RETURNS TO INVESTING IN GREEN URBAN DEVELOPMENT IN DURBAN EVALUATING THE POTENTIAL RETURNS TO INVESTING IN GREEN URBAN DEVELOPMENT IN DURBAN PAGE xv Figure 5.15 Aerial screenshot of the Pinetown industrial region for the comparison of scenario 10 (with detention basins) and scenario Figure A3.10 Thiessen polygons determined for the available rain gauges relevant to the U60F catchment..............................................................103 11 (detention basins and source controls). The noticeable difference is that the southern conduit system (red lines) flows via a Figure A5.1 Measure rainfall and simulated flow, depth, P, TIN, TSS concentrations for monitoring station (R_ZANA_10) on the Umhlatuzana detention basin (green box) before joining with the northern stormwater system (yellow lines) on the lower right hand....................46 River...............................................................................................................................................................................................................111 Figure 5.16 Simulated peak flows for a 2-, 5-, 10- and 20-year rainfall return period for Scenarios 10 and 11 at points (a) CJ4_56Umb (red line Figure A5.2 Measured rainfall and simulated flow, depth, P, TIN, TSS concentrations for monitoring station (R_UMBILO_13) on the Umbilo in Figure 5.15 with detention basin and source controls) and (b) CJ5_55Umb (yellow line in Figure 5.15 with only source controls River...............................................................................................................................................................................................................112 upstream). Note that time units are reporting time steps from the model and one time unit represents 5 minutes.............................46 Figure A5.3 Measured water quality from water sampling stations on the Umbilo and Umhlatuzana Rivers at the same locations as the .........................................................49 Figure 5.17 Infrastructure capital cost saving (R million) for scenarios 4 – 15 when compared to the Baseline. simulations....................................................................................................................................................................................................112 Figure 6.1 The total present value (R millions) of costs and benefits for all the scenarios relative to the baseline scenario of full Figure A6.1 Location of water quality sampling sites for which data have been used in the hydrological model......................................................113 development with adequate sanitation (Scenario 3)....................................................................................................................................57 Figure A6.2 Effects of different modelled scenarios on total phosphorus concentrations at different monitoring sites in the study area.............116 .........................................................................................................66 Figure A1.1 Schematic representation of levees at two side of the watercourse.. Figure A6. 2 (cond.) Effects of different modelled scenarios on total phosphorus concentrations at different monitoring sites in the study Figure A1.2 Schematic representation of a hydraulic bypass...........................................................................................................................................66 area..................................................................................................................................................................................................117 Figure A1.3 Permeable paving allows water to soak into the gravel sub-base, temporarily holding the water before it soaks into the ground, Figure A6.3 Effects of different modelled scenarios on total suspended solids (TSS) concentrations at different monitoring sites in the study or passes to an outfall (Source: susdrain, www.susdrain.org)......................................................................................................................67 area................................................................................................................................................................................................................118 Figure A1.4 Soakaways are square or circular excavations either filled with rubble or other aggregate fill that are able to attenuate and treat Figure A6. 3 (cond) Effects of different modelled scenarios on total suspended solids (TSS) concentrations at different monitoring sites in significant amounts of stormwater runoff. They can be grouped and linked together to drain large areas such as highways and the study area..................................................................................................................................................................................119 industrial areas................................................................................................................................................................................................68 Figure A6. 3 (cond) Effects of different modelled scenarios on total suspended solids (TSS) concentrations at different monitoring sites in Figure A1.5 Green roofs achieve runoff treatment and infiltration through the construction of vegetative cover on roofs which increases the study area..................................................................................................................................................................................120 storage, evapotranspiration and attenuation (Source: susdrain, www.susdrain.org)................................................................................69 Figure A6.4 Effects of different modelled scenarios on total inorganic nitrogen (TIN) concentrations at different monitoring sites in the study Figure A1.6 Diagram of a rainwater harvesting system. The first picture shows high stormwater runoff with none of the rain being collected area................................................................................................................................................................................................................121 whereas the second picture shows how rainfall is trapped and collected from the roofs in tanks and the amount of runoff entering streams and rivers is significantly reduced. . .................................................................................................................................70 Figure A6. 4 (cond.) Effects of different modelled scenarios on total inorganic nitrogen (TIN) concentrations at different monitoring sites in the study area............................................................................................................................................................................. 122 Figure A1.7 Swales are shallow grassed or vegetated channels used to collect and/or move water (Source:  susdrain, www.susdrain.org)............71 Figure A6. 4 (cond.) Effects of different modelled scenarios on total inorganic nitrogen (TIN) concentrations at different monitoring sites Figure A1.8 Filter strips are maintained grassed areas of land that are used to manage shallow overland stormwater runoff through several in the study area............................................................................................................................................................................. 123 filtration processes. They are usually located as strips adjacent to development areas, roads and waterways. ....................................72 Figure A1.9 Bio-retention areas are landscaped depressions employed to manage runoff by passing it through several natural processes. Rain gardens are an example of a bio-retention area. (Source: susdrain, www.susdrain.org)..................................................................73 Figure A1.10 (a) Lamination effect due to the flood plain storage and (b) Example of flood plain storage in San Paulo, Brazil (Giugni et al. 2012)....74 LIST OF TABLES Figure A1.11 Constructed treatment wetlands are man-made systems designed to mimic natural wetland systems (Source:  susdrain, www. susdrain.org)....................................................................................................................................................................................................75 Table 3.1 The rules and criteria applied to different active stormwater management measures based on soil drainage, slope and water table Figure A1.12 Riparian buffers are located adjacent to streams and river channels. They can either be made up of grasses and smaller plants characteristics.................................................................................................................................................................................................16 as in picture (a) or they can be densely vegetated with trees and bushes as in picture (b). They provide a buffer between Table 3.2 Annual Average Treatment Performance Capabilities for surface flow wetlands (Source: Kadlec & Knight 1995) assuming wetland adjacent land uses such as agriculture and residential areas and waterways.. ...........................................................................................77 influent is a “typical municipal effluent”. . ....................................................................................................................................................19 Figure A1.13 The effect of revegetation on discharge upstream and downstream of the Murrumbidgee in Australia Table 4.1 Summary of interventions considered in the scenarios. .............................................................................................................................24 ....................................................................................................................................................................78 (Source: Rutherford et al. 2007). Table 4.2 Scenarios used in the analysis. Levels of each intervention are described using symbols for ease of comparison. sw = stormwater. ..25 Figure A1.14 Catchment reforestation will aid in runoff infiltration reducing the overall amount of stormwater reaching rivers and streams. Reforestation will also aid in removing sediments and nutrients. ..............................................................................................................79 Table 5.1 Comparison of different systems for the categorisation of river health/condition data, simplified after DWAF (2008). ..............................................................................................................................................................32 Figure A1.15 Comparison of average cost per unit volume of runoff reduction for various stormwater management options, based on data in the literature...............................................................................................................................................................................................80 Table 5.2 The total annual dredging cost for each scenario, and annual and NPV dredging costs avoided (R millions) for each scenario when compared to the baseline...............................................................................................................................................................................38 Figure A1.6 Comparison of cost per unit mass of pollutant/nutrient reduction for various stormwater management options.................................81 Table 5.3 Estimated gains in estuarine and marine fishery values value due to a reduction in TSS and nutrients for scenarios 4-15 when Figure A2.1 Applying source controls to a subcatchment in PC-SWMM.........................................................................................................................89 compared to Scenario 3 as the baseline........................................................................................................................................................39 Figure A2.2 Profile of a detention basin showing the berm and the outlet pipe............................................................................................................89 Table 5.4 Stormwater infrastructure cost savings (R millions) for scenarios 4-15, when compared to the Baseline...............................................48 ..............................................90 Figure A2.3 Layout of modelled detention basins, shown as green squares, within the U60F quaternary catchment.. Table 5.5 The property premium associated with natural vegetation in a good condition and the nature-based tourism value for minimum, Figure A2.4 Example of in-situ detention basin constructed by the EM in the Hillcreast region of the U60F subcatchment. ...................................90 medium and maximum conservation extents in the catchment for each scenario. ..................................................................................51 Figure A3.1 Landuse Categories and D’MOSS subcategories for all Conservation Areas within the EMA....................................................................94 Table 6.1 Summary of the construction and maintenance costs (R millions) associated with GUD interventions for each scenario, excluding sanitation costs. Time frame of 20 years and a discount rate of 6% used. . ...............................................................................................55 Figure A3.2 Information required to delineate subcatchments: topographical aerial survey, DEM and river centre lines based on river flow paths................................................................................................................................................................................................................95 Table 6.2 Present value of costs and benefits (R millions) for all scenarios (2015 Rands, 6% discount rate, 20 years)............................................56 Figure A3.3 Current available stormwater shapefile (the red lines represent the flow paths, yellow lines are stormwater conduits and blue Table 6.3 Total present value of costs and benefits and NPV (R millions) for all scenarios (2015 Rands, 6% discount rate, 20 years), initial dots represent stormwater junctions)...........................................................................................................................................................95 capital requirement and the capital requirement as a percentage of the annual capital budget for each scenario. . ............................58 Figure A3.4 Newly available stormwater network shapefiles from the EM’s SMS audit................................................................................................96 Table 6.4 The impact of soakaways on NPV for Scenarios 9, 12, 14 and 15 (R millions, 6%, 20 yrs.) ........................................................................58 Figure A3.5 Percent impervious area for two different areas of contrasting landuse. The top value represents the %Imperv using approach Table A1.1 Measured pollutant removal capacities of selected stormwater management options and technologies (Source: Armitage et al. 1 and the bottom value represents the %Imperv using approach 2. ..........................................................................................................98 2013)................................................................................................................................................................................................................81 Figure A3.6 Map of Green-Ampt Parameters developed by UKZN (Source: Sinclair 2015)......................................................................................... 100 Table A1.2 Relative merits (indicated by number of “X”) of different measures for stormwater and flood risk management, based on the ......................................................................................82 literature. Measures considered in this study area are marked with an asterisk.. .........................................................................................................101 Figure A3.7 SCS 24-hour rainfall distributions (not to scale) (Source: SCS 1984). Table A2.1 Sewage output generated in the newly developed areas . .........................................................................................................................83 Figure A3.8 Snapshot of the PCSWMM model for catchment U60F..............................................................................................................................101 Table A2.2 General Effluent Limits...................................................................................................................................................................................84 Figure A3.9 Aerial image of all of the Durban Harbour outfalls (red triangles). The stormwater network is shown by the yellow lines. . ...............102 PAGE xvi EVALUATING THE POTENTIAL RETURNS TO INVESTING IN GREEN URBAN DEVELOPMENT IN DURBAN EVALUATING THE POTENTIAL RETURNS TO INVESTING IN GREEN URBAN DEVELOPMENT IN DURBAN PAGE xvii Table A2.3 The extent of implementation of source controls .......................................................................................................................................85 1. INTRODUCTION Table A2.4 The total area and volume (ha, m3) of each source control intervention for each of the source control scenarios.................................85 Table A2.5 The area, construction and maintenance costs associated with implementing source controls..............................................................86 1.1 Background that help to ameliorate the effects of urbanization on Table A2.6 Source control implementation costs (R millions) .......................................................................................................................................86 flooding and water quality closer to the source of the Urbanisation is taking place at an unprecedented rate Table A2.7 Annual Average Treatment Performance Capabilities for surface flow wetlands (Source: Kadlec & Knight 1995) assuming wetland problem. In addition, management and planning of cities influent is a “typical municipal effluent” (see Table A2. 8). .........................................................................................................................87 throughout the world, often outpacing plans and the is increasingly taking a holistic approach that includes capacity of city managers to provide the necessary Table A2.8 Typical municipal final effluent concentrations on which wetland treatment performance outlined in Table 3.2 is based the use and conservation of semi-natural and natural services. As a result, urban ecosystems are being (after Kadlec & Knight 1995). .........................................................................................................................................................................87 areas within cities as part of a green urban development degraded and lost, and problems such as flooding, air Table A2.9 Event Mean Concentration (EMC) data used in water quality modelling ...................................................................................................87 strategy. This not only contributes to solving surface and water pollution are becoming worse. This has led Table A2.10 Assumed pollutant reduction based on 1: 0.5 year event (Source: Georgia 2001).....................................................................................88 water problems but also maintains areas for recreation to negative impacts on health, income, productivity and which is essential for human health and wellbeing. All Table A2.11 Assumed treated pollutant concentrations of wetland effluent based on removal efficiencies given in Table 3.2..................................91 quality of life, as well as stretching local and national of this aligns well with the concept of “green urban Table A3.1 WWTWs located within the study area. .......................................................................................................................................................96 government finances. These environmental problems development”, the essence of which is development that are particularly acute in developing country cities, Table A3.2 Hydraulic input properties required for each subcatchment.......................................................................................................................97 minimises impacts on and/or enhances the value of the where a lack of resources and regulation has led to poor Table A3.3 Estimates of Manning’s roughness coefficient (N values) for overland flow. A summary from three different sources natural environment. (Source: Rossman 2015).................................................................................................................................................................................99 planning and the expansion of informal settlements, often in high risk areas. Addressing urban environmental issues requires Table A3.4 Values used for the depression storage based on landuse (ASCE, 1992)....................................................................................................99 a combination of engineering, spatial planning, Table A3.5 Soil parameters (Source: Rawls 1983). ........................................................................................................................................................ 100 Because urbanisation leads to the hardening of surfaces, environmental management and other interventions. Table A3.6 Data used for calibration of flows and water levels....................................................................................................................................103 the importation of water to supply urban inhabitants and One of the challenges of achieving green urban the production of wastewater and sewage, managing Table A3.7 Sensitivity of runoff volume and peak flow to surface runoff parameters (EPA, 2015)........................................................................... 104 development will be to shift the focus from reliance the quantity and quality of surface water flows is one Table A4.1 Return periods assigned to each of the infrastructure categories............................................................................................................105 on conventional grey infrastructure and “end of pipe” of the most important challenges of city planners and Table A4.2 Construction rates applied to the bridge culvert category........................................................................................................................107 measures, and find the right balance between ecological engineers. While conventional measures have involved and green or grey engineered infrastructure (Figure Table A4.3 Construction rates applied to the bridge pipe category. ............................................................................................................................107 “end-of-pipe” interventions to convey these problems 1.1) to tackle problems closer to source and maintain Table A4.4 Construction rates applied to the canal and culvert category...................................................................................................................107 away, these measures have often not been able to keep healthier, more vibrant and more resilient cities. This ahead of the problems, and have also contributed to Table A4.5 Construction rates applied to the pipe category........................................................................................................................................107 includes the strategic protection of natural habitats the pollution and degradation of aquatic systems within Table A4.6 The cost of pipes in 2016 delivered to central Durban...............................................................................................................................107 within cities for biodiversity protection and the delivery and downstream of urban areas. This together with the of ecosystem services. However, there is very little Table A5.1 Event Mean Concentration (EMC) data for different landuse types ........................................................................................................ 109 encroachment of developments into the natural habitats understanding of the costs and benefits of creating or Table A5.2 Water quality monitoring stations along the Umhlatuzana and Umbilo rivers ........................................................................................110 within urban areas and at their margins, has led to the protecting these different types of green infrastructure, Table A5.3 A summary of effluent water quality data at the outfalls of the WWTWs in the U60F catchment.........................................................111 loss of biodiversity and ecosystem services. especially in developing countries. It is also important Table A6.1 Comparison of different systems for the categorisation of river health/condition data, after DWAF (2008). National guidelines Great strides have been made in the design of more to understand to what extent it would be more cost for the determination of the ecological reserve with regard to water quality recommend the use of numeric ratings 0-1. ...............114 sustainable engineering mechanisms to deal with urban effective to rely on green rather than grey infrastructure Table A6.2 Threshold values for variables considered in this study, using ranges defined for each River Health Category (see Table 5.15). problems, including innovative water retention measures to solve key environmental problems. The values shown in each row represent the upper threshold value of that category. . .........................................................................114 such as porous pavements, green roofs and bio-swales Table A6.3 Guidelines to inform Present Ecological State ratings for turbidity/clarity (after DWAF 2008)...............................................................115 Table A6.4 Present Ecological State (PES) metrics and explanations after (DWA 2013), and used in this study to infer qualitative change in Figure 1.1 Schematic representation river condition as a result of scenarios involving attenuation of runoff and the provision of riparian corridors and buffers ...............115 of Green Urban Development and associated terminology (Source: authors own interpretation) PAGE xviii EVALUATING THE POTENTIAL RETURNS TO INVESTING IN GREEN URBAN DEVELOPMENT IN DURBAN EVALUATING THE POTENTIAL RETURNS TO INVESTING IN GREEN URBAN DEVELOPMENT IN DURBAN PAGE 1 1.2 Durban’s environmental issues Poor air and water quality is caused by traffic and These impacts will be further impacted by increasing industrial emissions, effluents and polluted runoff poverty and lack of economic growth, continued rapid Durban, located within the eThekwini Municipality on city has to deal with frequent floods. While Durban has especially from informal settlements where there is a urbanisation and the impacts of climate change. Climate the east coast of South Africa, is a subtropical “garden a relatively well-developed drainage system, flooding sanitation backlog. The unplanned expansion of informal change is predicted to change temperature, sea levels, city” which is rich in biodiversity, but faces many of the problems are exacerbated by the increased hard settlements and the lack of sufficient sanitation services rainfall patterns and storm events in a way that will same environmental and developmental challenges surfaces and the solid waste that accumulates in storm to these areas has resulted in the direct discharge both exacerbate water supply problems and flooding, as other African cities1. Rapid urbanisation and the water conduits, particularly plastic bottles. Beachfront of effluents into rivers, dams and wetland areas. and which will have direct and indirect impacts on growth of informal settlements is putting major financial development and sea level rise also contribute to Approximately 72% of informal settlements lack formal ecosystems and biodiversity within the EMA. Therefore pressure on the city, which also needs to encourage increasing risks of flooding in coastal areas. Another sanitation infrastructure, contributing significantly to just dealing with the current set of problems will be an development to keep up with the growing demand for factor contributing to high flows and flooding is the the degradation of freshwater ecosystems across the important step towards developing resilience against jobs. Unless they are valued as part of the development importation of water to supply city inhabitants. In formal study area. The release of untreated or poorly treated these future threats. solution, the tandem growth of informal settlements areas, the resulting waste water makes its way to sewage effluent from waste water treatment works directly into and formal developments will steadily contribute to the treatment works, which then discharge their wastes rivers as a result of lack of capacity or overflow during degradation and loss of natural systems and biodiversity into the drainage systems. Combined with the effects within and around the city. Undeveloped land or green of hardened surfaces, these elevated flows and nutrient storm events contributes to ever increasing pollution 1.3 Green urban development levels in freshwater systems. The groundwater has also open space areas make up 33% of the total area within levels lead to erosion of river banks and have an impact Green urban development is an approach that aims become polluted as a result of poor waste disposal sites, the municipality, however less than one third of this falls on estuarine ecology and functioning. Very little of this to minimize the impacts of urbanization on the pit latrines and septic tanks being used across the city. within the urban edge. Some 26% of the land area in the waste water is recycled. Yet, ironically, the EMA faces environment and enhance environmental values OECD The increased pollution loads in freshwater systems EMA remains undeveloped and non-degraded (EPCPD major water shortages as the supply of potable water (2013). The approach is advocated to increase, rather are not only a human health hazard but also lead to the 2012). However, only about 10% of the open space areas from the surrounding catchments is outstripped by than limit, the development potential of cities. Given eutrophication of rivers and estuaries. within the eThekwini Municipal Area (EMA) is formally growing demand. Currently, water is supplied primarily the degree to which global population and economic protected and just under 7% is actively managed by the uMngeni catchment to the city but it is expected The unplanned expansion of informal settlements and production is increasingly urbanized, It is also vital for (eThekwini Municipality 2014). that water from other catchments will soon be needed planned expansion of formal settlements, agriculture, global welfare. to meet growing demands. Water is also becoming more overexploitation of natural resources, and alien Durban’s key environmental issues and their causes and expensive as local resources are depleted and more We define green urban development as a range of invasive species have led to the loss and degradation consequences are summarised in Figure 1.2. The “brown water is imported into the city from greater distances actions that tackle the core problems of pollution of natural habitats within and around the urban area. issues” relate to environmental quality (air and water) away. and waste, the consumption of natural resources, the This exacerbates the “brown issues” referred to above and flooding. During the summer rainfall months, the loss of urban open space and the degradation and (Figure 1.2). For example, illegal sand mining activities loss of biodiversity, as well as mitigation of the urban along or in some of the river channels intensifies contribution to climate change (Figure 1.3). This reqires the problem of bank erosion, and the loss of natural a combination of indirect and direct interventions vegetation cover leads to sedimentation which clogs that will serve synergistically to develop vibrant, drainage channels. The degradation and loss of natural resilient cities that are both greener in appearance habitats and biodiversity, or “green issues”, also leads and greener in terms of their local, regional and global to the loss of amenity, natural resources and carbon. Figure 1.2 Schematic diagram of key environmental issues in Durban (divided into “brown issues” and “green 1 For more information about the state of the ecosystems and issues”), their causes and their consequences (Source: the main drivers and causes of the environmental issues in the author’s own analysis) eThekwini Municipal Area (EMA) see the accompanying eThekwini Urban Environmental Profile Report (World Bank 2015). Figure 1.3 The main elements of a green urban development policy (Source: authors) PAGE 2 EVALUATING THE POTENTIAL RETURNS TO INVESTING IN GREEN URBAN DEVELOPMENT IN DURBAN EVALUATING THE POTENTIAL RETURNS TO INVESTING IN GREEN URBAN DEVELOPMENT IN DURBAN PAGE 3 environmental impacts, and in which natural systems ƒƒ Securing the protection, restoration or rehabilitation Because of the strong linkages of GUD to catchment effluent management measures and (c) conservation provide meaningful refugia for biodiversity and are of selected natural areas in order to maintain hydrology, the case study was for a selected catchment areas in conjunction with more compact development used to advantage in supplying valuable ecosystem biodiversity and valuable ecosystem services. Natural area, rather than a selected administrative area of (so as to be able to support the same number of people services. In addition to a range of policy interventions, systems within cities contribute to livelihoods through Durban. A back-casting approach was used, for a and industry in the catchment). In order to inform the this involves investing in natural capital as well as the provisioning of natural resources, contribute to catchment that is already developed, rather than choice of stormwater management measures, a review use of green structural engineering and conventional human health and wellbeing, property value and designing alternative development scenarios in one was carried out on their efficacy, costs and the necessary grey infrastructure. “Green” is synonymous with tourism through the provision of aesthetic and of the prospective development nodes in the EMA. or suitable conditions for their implementation. Based “environmentally friendly”, and does not just refer to recreational amenity value, and contribute ecosystem This was to allow the investigation of the potential on this, and available GIS data on land cover, slope and vegetated areas. services such as flood control, sediment retention, air costs and benefits of alternative policy measures while soils of the catchment, the long list of possible measures and water quality amelioration, carbon storage, keeping other factors constant (industries, households was reduced to a set of measures that had both a high We therefore propose that the green urban pollination of crops and provision of nursery areas for inhabitants, etc.). The Umhlatuzana – Umbilo catchment feasibility of implementation in the study area and that development should include the following: marine fisheries. As cities grow the remaining natural was selected because it is one of the most developed would be complementary in terms of their effects on areas within them become increasingly important as catchment areas in the EMA. Nevertheless, this flood risk and water quality amelioration. The potential ƒƒ Tackling the problems of air, water and solid waste refugia for biodiversity. All of these functions are lost catchment still contains significant undeveloped areas extent of their implementation was then estimated and pollution through the provision of solid and liquid however, if they are excessively degraded and that are zoned for future development. Therefore our mapped. Finally, a set of scenarios was devised which waste management services and enforcement of fragmented. Thus cities need to plan and manage a approach had to include some assumptions regarding included the full combination and various subsets of appropriate regulations to control effluents and system of natural open space areas within them and future development for these areas following a these measures. emissions from a wide variety of sources. This is not also take care to minise the damages to aquatic “business as usual” approach. This was done based on only a necessity from a social and human health ecosystems downstream. densities from comparable urban typologies in the area. The implications of the different scenarios were assessed perspective, but is also a prerequisite to the success of based on the costs of the interventions, the cost savings all other green urban development interventions; The study required the development of a hydrological due to reduced flood risk, ecosystem health and the 1.4 Study objectives model of the catchment area to model the effects of avoided losses of ecosystem services. The cost savings ƒƒ Tackling the issue of replacing natural with built the interventions on storm flows and water quality. due to reduced flood risk were modelled based on surfaces in a more environmentally friendly manner. The aim of this study was to explore, using a case study Modelling was carried out using the PC-SWMM changes in the required design specifications for the This includes implementation of sustainable and scenario-based approach, the potential costs and modelling system which was set up at a fine resolution existing type of stormwater conveyance infrastructure stormwater management systems that focus on benefits of undertaking a green urban development for event-based flood modelling and a courser resolution in the study area, as a result of changes in the size neutralising the impacts of hardened surfaces on approach to address some of the main environmental for continuous water quality modelling. The current of the relevant return-period floods. Changes in stormwater flows using attenuation measures such as issues described above, and to explore the potential status quo was modelled and calibrated using existing ecosystem health were estimated based on changes detention ponds, infiltration trenches, porous paving tradeoffs between different types of interventions, with data on flows and water quality. Following this, a new in nutrient loads and concentrations in relation to and green roofs. It also includes energy- and water- emphasis on assessing the desirable balance between baseline was set up for the fully developed catchment. current status quo. Changes in the value of ecosystem efficient planning and building design. engineered interventions and the conservation of All scenarios involved interventions to be compared with services was based on the changes in ecosystem size natural open space areas. The study focuses on three the fully developed baseline, as if the catchment had and quality, and models of the relationship between ƒƒ Tackling water and energy consumption. The former is elements of green urban development, all of which been planned and developed differently from the outset. these and their value which were developed as part primarily to ensure the sustainable supply of water impact on ecosystems and biodiversity: sewage and solid A total of 15 scenarios was run to compare the costs and of the accompanying study of the value of eThekwini’s from surrounding surface and groundwater source waste management, active stormwater management outcomes of different types of interventions. ecosystem services (Turpie et al. 2017). areas as well as the impacts of water use within cities. and the conservation of natural systems and riparian Dealing with this problem also helps with the problem corridors. The interventions under consideration included (a) of waste water management, particularly if waste- meeting sanitation needs, (b) various stormwater and water recycling is used as part of the solution. Tackling The study aimed to provide a proof of concept, based energy consumption has multiple advantages. It on high-level exploratory analysis and simple scenarios addresses the necessity of reducing carbon emissions in order to facilitate and promote dialogue rather than to reduce the risks of climate change both globally and providing a blueprint for action. It is hoped that the locally. Given the high reliance by urban households on study will provide a useful step towards the preparation wood fuel in Africa, it would also address local air of a Strategic Environmental Assessment to guide the pollution, as well as deforestation in the areas beyond city’s sustainable development path. cities, which occurs at great cost to biodiversity and society, and which also exacerbates environmental problems such as flooding. 1.5 Study approach The overall approach was to model current flooding and ƒƒ Investing in greening measures. These include creating water quality in the Umhlatuzana - Umbilo catchment and maintaining recreational green open space areas and to determine the potential change in water such as parks, and investing in the planting of trees quality and flood hydrographs at selected points in the and gardens along city streets. These will not only catchment after implementation of a range of green provide aesthetic enhancement but will contribute to urban development measures including sanitation, the reduction of air pollution (Beckett et al. 1998, Jim stormwater management and conservation measures. & Chen 2008) and mitigate against urban heat island The relative effectiveness and cost-effectiveness of effects (Akbari et al. 2001, EPA 2014). different types of measures were then evaluated. PAGE 4 EVALUATING THE POTENTIAL RETURNS TO INVESTING IN GREEN URBAN DEVELOPMENT IN DURBAN EVALUATING THE POTENTIAL RETURNS TO INVESTING IN GREEN URBAN DEVELOPMENT IN DURBAN PAGE 5 This page intentionally blank. PAGE 6 EVALUATING THE POTENTIAL RETURNS TO INVESTING IN GREEN URBAN DEVELOPMENT IN DURBAN II. THE UMHLATUZANA-UMBILO CATCHMENT AREA 2.1 Location and extent The Umhlatuzana-Umbilo catchment (quaternary a number of commercial and industrial areas which are catchment U60F) is located in the centre of the EMA situated both in the lower and middle catchment. The (Figure 2.1), incorporating the city centre, harbour, and catchment covers an area of approximately 272 km2. Figure 2.1 The location of Umhlatuzana-Umbilo catchment in the EMA EVALUATING THE POTENTIAL RETURNS TO INVESTING IN GREEN URBAN DEVELOPMENT IN DURBAN PAGE 7 2.2 Geography and climate humid with a maximum average daily temperature of 2.3 Historical and current land cover 42% being woodland, 35% forest, 14% grassland, and 28°C (Roberts & O’Donoghue 2013). The mean annual 9% thicket. Most of the intact forest is located along the The catchment extends from sea level to about 750m Historically, the vegetation in the catchment was precipitation is over 1000 millimetres and the rainy steep river valleys in the upper and middle catchment. elevation (Figure 2.2). It has a natural mean annual dominated by coastal belt grassland, sandstone sourveld season falls between September and March (Roberts & The North Park Nature Reserve, Kenneth Stainbank runoff (MAR) of 43.25 million m3 and the two main rivers and coastal belt thornveld (Figure 2.4). Coastal forest O’Donoghue 2013). The change in temperature between Nature Reserve and Bluff Nature Reserve are located are the Umhlatuzana River and the Umbilo River which and scarp forest was mostly found along the river valleys seasons is greatest in the higher altitude areas in the in the catchment providing some protection to these flow into the Durban Bay Harbour (Figure 2.1). and steep gorges, with scarp forest generally restricted west. Rainfall tends to be highest along the coast but natural systems. Durban Bay Harbour is one of South to higher elevations (Figure 2.4). Figure 2.5 shows the Durban has a subtropical climate with humid wet rainfall seasonality is greatest in the west, with most Africa’s larger estuaries which in spite of a high degree extent to which landcover has changed within the summers and mild dry winters (EPCPD 2012). The warm of the rainfall falling within the summer months. The of transformation is still of conservation importance. catchment and the amount of natural open space areas Agulhas Current that flows southwards along the coast predominant winds blow parallel to the coastline in a At the head of the estuary a small 15 ha pocket of that remain, mostly in fragmented patches. Most of the has a moderating influence on the climate, keeping north-easterly and south-westerly direction. mangroves are protected as part of the Bayhead Natural catchment has been transformed and is dominated by winter temperatures mild and summers warm and Heritage Site. formal and informal urban settlement, and commercial and industrial land (Figure 2.5). In the upper catchment there is a relatively large amount of agricultural land (Figure 2.5). 2.4 Pollution and flooding Water pollution and flooding are two of the main There are a number of informal settlements such as environmental issues associated with the Umhlatuzana- Cato Manor in the lower catchment, as well as peri- Umbilo catchment. Pollution comes from a number of urban settlements such as Tshelimnyama in the upper different sources including industrial, residential and catchment. The informal settlement at Cato Manor is agricultural runoff, stormwater outflows, solid waste located along the Mkumbane River, a tributary of the and effluent from various waste water treatment works Umbilo River. The informal structures have been erected (WWTWs), with three major WWTWs (the Hillcrest along the steep river banks and in the flood plain WWTW, Umhlatuzana WWTW and Umbilo WWTW) (Figure 2.6). located on the two main rivers. These pollution sources together have a major impact on the biodiversity Currently there are just over 6000 ha of natural and ecological functioning of the river systems (DWA vegetation in the Umhlatuzana-Umbilo catchment, with 2013) as well as Durban Bay. Within the catchment Figure 2.2 Elevation map showing the main elevation bands as one moves from the upper catchment to the lower catchment 160 140 Mean rainfall (mm) 120 100 80 60 40 20 Figure 2.3 Average monthly rainfall (mm) for Durban, 1961-2003 0 (Source: StatsSA 2005) Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Figure 2.4 Map of the historical vegetation within the Umhlatuzana-Umbilo catchment PAGE 8 EVALUATING THE POTENTIAL RETURNS TO INVESTING IN GREEN URBAN DEVELOPMENT IN DURBAN EVALUATING THE POTENTIAL RETURNS TO INVESTING IN GREEN URBAN DEVELOPMENT IN DURBAN PAGE 9 Figure 2.7 Sewage and litter from the informal settlement ends up in the Mkumbane River (Source: Google Earth) Figure 2.5 Current landcover in the Umhlatuzana-Umbilo catchment Figure 2.6 The Cato Manor informal settlement is situated along the banks of the Mkumbane River study area there are a number of informal settlements burst their banks (Figure 2.8). An enormous amount of without adequate sanitation, such as Cato Manor plastic litter washes into the rivers and out to sea during informal settlement which is situated on the banks of high rainfall events. Some of this ends up on Durban’s the Mkumbane River, a tributary of the Umbilo River. main beaches. Not only does this have impacts on the Without adequate sanitation raw sewage, litter and aesthetic (and by implication the tourism) value of the other pollutants end up in the river (Figure 2.7). beaches themselves but there is also increasing concern about the impacts of plastic litter on marine ecosystems Flooding in the catchment is exacerbated by the high (Bergmann et al. 2015), particularly with regard to levels of litter and dead vegetation which block culverts entanglement and plastic ingestion (e.g. Ryan 1990). and drains causing rivers and streams to overtop and Figure 2.8 Photos taken of the Umhlatuzana River after a flood event in February 2016 showing the large amounts of plastic and other waste washed into the river. PAGE 10 EVALUATING THE POTENTIAL RETURNS TO INVESTING IN GREEN URBAN DEVELOPMENT IN DURBAN EVALUATING THE POTENTIAL RETURNS TO INVESTING IN GREEN URBAN DEVELOPMENT IN DURBAN PAGE 11 2.5 Water quality from waste-water treatment works, sewer overflows or leaks, point source and diffuse runoff from poorly The middle and lower reaches of the Umbilo and serviced informal settlements and runoff from areas Umhlatuzana river catchments are highly impacted such as bus stations, informal markets and general street and water quality in both of these rivers is considered runoff, contaminated by human and other animal waste poor (RHP 2002, DWA 2013). However, water quality (e.g. dogs). conditions in the upper catchment of the Umhlatuzana River, in particular, are considered “acceptable” (DWA 2013), in part because of the low density of settlements and the presence of extensive areas of 2.6 River condition largely unmodified riparian buffers in this part of the South Africa’s National River Health Programme (RHP catchment. 2002) rated the the Umhlatuzana and Umbilo Rivers as associated in a Fair overall condition – this rating In addition, Moodley (2014) analysed heavy metal considers the riparian and instream condition of the concentrations at 17 sites in total in the Umhlatuzana, rivers. It is however derived from several components, Umbilo and Amanzimyana Rivers during the 2011-2012 as follows: wet- and dry seasons, and found the following: ƒƒ In the Umhlatuzana River: ƒƒ The Umhlatuzana River upstream of its confluence with the Umbilo River had copper, aluminium and ͵͵ aquatic macroinvertebrate communitiues are rated nickel at concentrations above DWAF (1996) target as in Fair health; thresholds during the dry season (suggesting effluent outputs); ͵͵ fish communities are also rated in a Fair condition in the upper and middle reaches of the catchment but ƒƒ The Umbilo, Amanzimyana and Umhlatuzana River are considered in a Poor condition in the lower downstream of its confluence with the Umbilo River reaches, where riparian habitat quality has also had elevated concentrations of mercury, vanadium, deteriorated to Poor; lead and chromium at all sites during the dry season (suggesting industrial effluent releases); ͵͵ water quality in the middle and lower reaches of the river is Poor (due mainly to urban impacts). ƒƒ Copper, aluminium and lead were problematic in these sites during the wet season as well, albeit at lower ƒƒ In the Umbilo River: concentrations (presumably as a result of dilution); ͵͵ habitat integrity is rated as in Good condition in the ƒƒ Ammonia-nitrogen was above guideline middle reaches and Poor condition in the lower concentrations at all sites in all systems, in both the reaches; wet and dry season; ͵͵ water quality is sonsidered Poor in the middle and ƒƒ Orthophosphate was elevated above DWAF (1996’s) lower reaches of the River, again due mainly to threshold for hypertrophic conditions in all systems urban impacts; during the wet season, suggesting possible surface runoff influences rather than effluent inflows; and ͵͵ invertebrates and fish populations are rated as in a Poor category, as a result of the heavy pollution ƒƒ Sites downstream of nature reserves on the Umbilo load. and Umhlatuzana Rivers showed a clear reduction in ammonia-nitrogen, orthophosphate and sulphur concentrations in river water. Human health concerns as a result of bacterial or other pathogen contamination centre on both the Umbilo and Umhlatuzana Rivers (DWA 2013), with high Escherichia coli counts recorded in the Umbilo River, at Paradise Valley Nature Reserve and below the Umbilo WWTW, along with high nutrient loading and potentially toxic levels of unionised ammonia (DWA 2013). The Umhlatuzana River shows high E.coli counts at Kenneth Stainbank Nature Reserve (DWA 2013). Sources of bacterial loading are assumed to include periodic pulses of poorly-treated sewage effluent PAGE 12 EVALUATING THE POTENTIAL RETURNS TO INVESTING IN GREEN URBAN DEVELOPMENT IN DURBAN III. DESIGN AND POTENTIAL EXTENT OF SELECTED GREEN UR- BAN DEVELOPMENT INTERVENTIONS FOR THE STUDY AREA 3.1 Sewage and solid waste management the project focuses on raising awareness and generating employment. In Durban, the Sihlanzimvelo Stream Cleaning Project has been very successful in areas of the 3.1.1 Dealing with sewage municipality where a number of rivers were considered Under green urban development, informal settlements critical in terms of health and functioning. It was would be provided with urine diversion dehydration assumed that a program similar to this was implemented toilets (UDDTs) and adequate services as a minimum in the catchment. requirement. Growing formal residential areas also provide a potential problem in that they lead to an A multi-faceted solid waste programme is required to increase in the output of treated sewage effluent. Under eliminate or at least drastically reduce the loads of solid a green urban development policy it is assumed that the waste entering the river systems in the study area. This these potential effects are neutralised by recycling an would require not only localised action, but broader equivalent amount of sewage effluent (see Appendix 2 municipal-level strategies (such as by-laws, incentive for more detail). These innovative technologies are very measures and compulsory recycling and linked servicing) much a part of green urban development thinking, which and even national strategies (legislation and taxation). also addresses water and energy supply and demand. 3.2 Active stormwater management 3.1.2 Managing solid waste (green engineering) Sediments and litter accumulate until they are either manually removed or are transported by the wind and/ 3.2.1 Overview of stormwater management or stormwater runoff into the drainage system. Once in the drainage system, they can contribute to blockages The increases in impermeable surfaces within urban and increased flood risk, as well as providing health risks. areas prevent infiltration and cause higher levels of This is a significant and ongoing problem in the study surface runoff during storm events than would have area (Figure 2.8). happened naturally, creating flooding problems in downstream areas (Armitage et al. 2013). This problem Solid waste problems have escalated in the last half is exacerbated by the fact that urban stormwater runoff century with the advent of plastic packaging and throw- generally contains litter, debris, and sediments which away consumerism and business strategies, and are also lead to blockages of the systems designed to convey inadequately managed in most Afrian cities, including water. They also contain bacteria, heavy metals and Durban. Addressing this problem is not only a critical nutrients, which means that floodwaters can become a component of stormwater management, but is also pollution hazard. All of this can have negative impacts on justifiable in terms of reducing impacts on the amenity property, urban infrastructure and downstream natural value of green open spaces and the devastating impacts habitats, as well as urban inhabitants. With an increase on freshwater and ocean environments and biodiversity. in urbanisation worldwide and the associated impact of increasing stormwater runoff on aquatic ecosystems, the Measures to address this include bans, taxes and management of urban drainage has become a critically refund systems on packaging and community-based important challenge (Fletcher et al. 2015). river cleaning programmes as well as the more traditional measures of street sweeping and municipal Various storm water management measures have waste collection. Community stewardship programs been designed that address flooding, water quality can have multi-sectoral impacts as they generate problems, or both (see Appendix 1 for a detailed employment opportunities, provide awareness, review). These approaches are increasingly being applied safeguard communities and provide city-wide services in development planning in South Africa, and their such as functioning river systems that are clean and inclusion (in any form) has recently become mandatory clear of litter. Sections of rivers are maintained by for new developments in the City of Cape Town and in cooperatives which are responsible for removing alien Durban with the aim that the urbanisation effects of vegetation, rubble and any solid waste blocking the new developments are effectively “neutralised”. In this free flow of water down the stream or river. They study, a review of alternative measures was carried out are also responsible for maintaining the grass and and their potential applicability to the study area was other vegetation along the banks of the waterway. evaluated in order to inform the development of feasible The cooperatives generally consist of members of the scenarios for the analysis. community that are unemployed and vulnerable and EVALUATING THE POTENTIAL RETURNS TO INVESTING IN GREEN URBAN DEVELOPMENT IN DURBAN PAGE 13 Stormwater management measures are classified into measures are inadequate, as is the case in a large Box 3.1 Active structural stormwater management measures (see Appendix 1 for more detail). structural and non-structural measures, with structural proportion of African cities, this leads to water quality Source controls measures being further subdivided into passive and problems that are too severe to be treated by the green Permeable paving Rainwater harvesting active measures (Figure 3.1). engineering measures incorporated in sustainable urban Promotes infiltration Collect & store rainfall drainage systems. This reality is not emphasised in the of stormwater runoff from hardened surfaces ƒƒ Passive structural measures aim to convey water and literature because SUDS have largely been developed into sub-layers or (e.g. roofs) for later use. protect areas from flooding. Examples are levees, in higher income countries and cities. Investment in increasing the channel capacity by clearing of debris or underlying substrata. Water that is collected adequate sanitation systems can therefore be seen as a increasing its cross-section, and constructing hydraulic Suitable for can be used to fundamental imperative for green urban development, bypasses (waterways) to divert high flows. pedestrian & supplement potable and something that has to be addressed first in most African cities. vehicular traffic. water supply. ƒƒ Active structural measures aim to modify the Infiltration trenches Soakaways hydrograph (i.e. reduce flood peak and volume) and The active structures or “green” engineering measures Excavated trenches Underground storage address water quality by retarding water movement, tend to be grouped as either source, local or regional lined with geotextile areas packed with course by increasing infiltration or storage in the catchment controls (Thampapillai & Musgrave 1985, Kundzewicz & filled with rock. aggregate or other porous area (Topa et al. 2014). These can be referred to as 2002, Armitage et al. 2013; Figure 3.1 and Box 3.1, see Water infiltrates into media that gradually “green” (sustainable/environmentally-friendly) Appendix 1 for details): surrounding soils discharges stormwater engineering measures. from the bottoms & into the surrounding soil. ƒƒ Source controls tend to be used to manage stormwater runoff as close to the source as possible, sides of the trench. Similar in operation to ƒƒ Non-structural measures do not involve physical construction but use knowledge, practice or generally within the boundaries of the property and infiltration trenches. agreements to reduce risks and impacts through include measures such as green roofs, soakaways and Green roofs behavioral changes, in particular through policies, permeable paving. Multi-layered system that covers the roof of a building with public awareness raising, training and education vegetative cover. Intercepts & retains precipitation. Provides ƒƒ Local controls are usually used to manage runoff as a (Kundzewicz 2002). These include flood warning second line of defence typically in public areas, along ecological, aesthetic and amenity benefits on top of reducing systems, land use regulations such as development roadways and adjacent to parks such as filter strips volume & attenuating peak flows. setbacks which identify where development can and and swales. cannot occur, or to what elevation structures should Local controls locate their lowest habitable floor to; regulations that ƒƒ Regional controls are used to manage runoff as the Bio-retention areas Filter strips require flood proofing and retrofitting of buildings may last line of defence and are generally large-scale Landscaped Maintained grassed increase the strength against flood actions; elevation interventions constructed on municipal land such as depressions. Rely on areas of land, used to of buildings may avoid completely the inundation. detention ponds and wetlands (Armitage et al. 2013). engineered soils, manage shallow Flood insurance and relocations also belong to this Measures that retard flows generally also contribute enhanced overland stormwater typology of measures. towards improving water quality, and vegetated areas vegetation & runoff through several further contribute to water quality amelioration filtration to remove filtration processes in a Although sanitation might be categorised as a non- where flows are slow enough. The various stormwater pollution & reduce very similar manner to structural measure in the context of stormwater mamagement measures are described in more detail in downstream runoff. buffer strips. management, there is a significant element of Appendix 1. traditional built infrastructure required to address Vegetated swales Sand filters sewage reticulation and treatment. Where sanitation Shallow vegetated A sedimentation channels with flat & chamber linked to sloped sides, an underground designed to store & filtration chamber convey runoff as filled with sand or well as remove other media pollutants. through which stormwater runoff can pass. Regional controls Detention basins Treatment wetlands Temporarily stores runoff for short periods after high Man-made systems that are designed to mimic rainfall events. The natural wetland systems. Marshy areas of shallow captured water either water that are either infiltrates into partially or underlying soil layers completely covered or is drained into the in aquatic vegetation. downstream Effectively remove watercourse at a nutrients & Figure 3.1 Different types of measures used in stormwater management (Source: Turpie et al. 2017). These measures are described in detail in Appendix 1. predetermined rate. sediments. Diagrams sourced from: https://help.xpsolutions.com; http://www.lets-do-diy.com; http://theleafyagenda.wordpress.com; http://faculty- legacy.arch.tamu.edu/bdvorak/research.htm PAGE 14 EVALUATING THE POTENTIAL RETURNS TO INVESTING IN GREEN URBAN DEVELOPMENT IN DURBAN EVALUATING THE POTENTIAL RETURNS TO INVESTING IN GREEN URBAN DEVELOPMENT IN DURBAN PAGE 15 3.2.2 Review and feasibility of green engineering measures The design requirements and limitations of green area, the long list of possible measures was reduced engineering measures described in the literature, to a set of measures that had both a high feasibility such as slope, depth to groundwater and soil drainage of implementation in the study area and that would characteristics, were obtained from the most recent be complementary in terms of their effects on the and comprehensive studies (see Armitage et al. environment. The potential extent of each of the 2013, Morales Torres et al. 2015) as well as expert selected measures was then estimated and mapped. opinion (Table 3.1). Based on GIS data for the study Table 3.1 The rules and criteria applied to different active stormwater management measures based on soil drainage, slope and water table characteristics Intervention Soil drainage Slope (%) Water Table (m) Selected Rainwater harvesting n/a n/a n/a No Infiltration trenches 0.25-0.5 (medium) <4% >5m (low) Yes controls Source Soakaway/subsurface infiltration beds >0.5 (well drained) <15% >5m (low) Yes Permeable/porous pavements >0.5 (well drained) n/a >5m (low) Yes Green roofs n/a n/a n/a Yes Bioretention areas, e.g. "raingardens" >0.5 (well drained) <12% >5m (low) No controls Sand filters >0.5 (well drained) <6% >5m (low) No Local Filter strips >0.5 (well drained) 2-6% >5m (low) No Vegetated (grassed) swales n/a <4% n/a No Regional controls Detention basins n/a <15% >5m (low) Yes Treatment wetlands n/a 0-0.5% (flat) >5m (low) Yes Figure 3.2 Soil drainage rates in the Umhlatuzana-Umbilo catchment (Source: South African atlas of agrohydrology and climatology) Rainwater harvesting tanks were not considered alternative to conventional gutter systems. However, it desirable at a large scale due to cost as well as the became clear that most of the roads in the middle and limited flood benefits, as tanks would fill up early in the upper catchment fall outside of the criteria required for 3.2.3 Potential extent of selected interventions high flows and then releasing the stored water over rainy season and would not provide further storage. effective implementation of grassed swales (<4% slope). a period of time. They are usually designed to fill and A better option would be to use measures that slowly empty within 48 hours of a storm event. Whilst the other It is necessary to consider localised measures that 3.2.3.1 Stormwater source controls release water, so that they are ready for the next high GUD measures capture runoff at the source, detention rainfall event, such as infiltration trenches, soakaways capture and/or slowly infiltrate runoff at source (i.e. Source controls are on-site measures that are designed basins capture runoff off-site from larger public surface and detention basins. on-site buildings, roofs, parking lots etc.) and regionally to reduce or neutralise the impact of hardened areas, such as roads. (i.e. off-site such as roads, walkways and other extensive surfaces on runoff. There are a range of options that A number of active structural measures were not paved areas). In doing so a “treatment train” can be all contribute to this, and the mix of these options on a The most effective location and the size of detention feasible because of unsuitable soil drainage (Figure 3.2), created where runoff from local source areas is captured particular site may depend on physical limitations as well basins was determined based on the difference in runoff high ground water levels, or the steepness of certain on site and runoff from larger areas is captured off site. as the costs involved. between pre-development (natural) and the post- areas in the catchment. Many interventions require well- Furthermore, at point-source pollution areas, such as development (fully-developed) scenarios within the drained soils (i.e. high drainage rates), which are mostly WWTW, constructed wetlands could be designed to For each measure, we estimated the potential extent of catchment. This was estimated using the PC- SWMM in the lower catchment (Figure 3.2), where effectiveness improve water quality. implementation. In some cases, where costs were very model described in Appendix 2. Detention basins were of the intervention would be lower. Certain conveyance high and full implementation unlikely to be feasible, the positioned at the outlet of the subcatchments with the and engineering measures, such as vegetated swales, Thus four types of source (on-site) control measures maximum extent considered in the analysis was less than highest difference in volumes. filter strips and bioretention areas are limited by (green roofs, permeable paving, infiltration trenches the potential extent (see next section). The potential slope and therefore their use on a large scale would andsoakaway pits) and two types of regional control extent of each source control measure is shown in not be possible because of the hilliness of the study measures (detention basins, treatment wetlands) were Figure 3.3 and the assumptions regarding extent, design 3.2.3.3 Treatment wetlands area. Grassed swales, for example, were considered considered as the feasible green engineering measures and costs are discussed in more detail in Appendix Constructed treatment wetlands are designed to as a possible measure due their low construction and for implementation in the study area. Their design and 2. Technical details and assumptions relating to the improve polluted runoff and waste water effluent quality maintenance costs and their ability to improve water potential extent is described further below. implementation of source controls in the PC-SWMM and provide some limited control of runoff volumes. quality, attenuate peak flows and provide a “green” model are also provided in Appendix 2. Wetlands can be effective in terms of removal of low levels of pollutants (i.e. a polishing function) and also 3.2.3.2 Detention basins provide buffering functions, aesthetic value and wildlife habitat. Constructed wetlands are located as close to Detention basins temporarily detain stormwater runoff the source of pollution as possible so as to maximise the from roads and other hardened surfaces, by capturing impact on improving water quality. PAGE 16 EVALUATING THE POTENTIAL RETURNS TO INVESTING IN GREEN URBAN DEVELOPMENT IN DURBAN EVALUATING THE POTENTIAL RETURNS TO INVESTING IN GREEN URBAN DEVELOPMENT IN DURBAN PAGE 17 The treatment wetlands were situated at point source biodiversity targets for the catchment. These tow pollution outlets in the study area; at the three existing interventions are ideally combined, with the riparian WWTW. It was assumed that the runoff entering the corridors providing connectivity between the protected wetlands was being treated to general standards terrestrial systems and other aquatic systems, including and that the wetlands further treat runoff to specific the estuary and ocean. The potential design and extent standards (see Appendix 2). of these interventions is explored in more detail below. In order to estimate the extent of the wetland for costing purposes, the removal rate was used to estimate 3.3.1 Riparian buffers the area required to meet the calculated concentrations Riparian buffer zones along waterways intercept based on the given efficiency (Table 3.2). Therefore the sediments, nutrients, pesticides and litter in diffuse outflow from each WWTW was used in conjunction with or sheet surface runoff thereby reducing the amount nutrient and sediment concentrations to determine daily of pollutants entering rivers and streams. They may loads for nutrients and sediments. The removal rate also provide habitat and wildlife corridors and can and removal efficiency data provided in Table 3.2 was be important for reducing erosion, slowing down then used to calculate the area that would be required floodwaters by increasing roughness (and thus to treat these estimated daily loads. Based on these reducing downstream peaks) and providing river assumptions, the extent of the treatment wetlands bank stabilisation. They also provide space for river ranged from 2 ha at the Hillcrest WWTW to 16 ha at the rehabilitation, reducing the need for future river Umhlatuzana WWTW to 40 ha at the Umbilo WWTW. lining. In combination, riparian buffers overlap with Technical details and assumptions relating to the conservation areas. implementation of treatment wetlands in the PC-SWMM model are given in Appendix 2. In this study, riparian buffers extended 15 m on either Table 3.2 Annual Average Treatment Performance Capabilities side for smaller rivers and streams, 30 m on either side for surface flow wetlands (Source: Kadlec & Knight of major rivers and on one side for rivers that have 1995) assuming wetland influent is a “typical municipal been canalised (Figure 3.4). These widths are broadly effluent”. based on the 2014 Draft National River Buffer Guidelines Removal rate Efficiency (Macfarlane et al. 2014). This produces a total area of Parameter (kg/ha/day) (% per ha/day) about 1900 ha. The buffers were clipped in GIS into BOD 10.0 67 the historical vegetation layer for the catchment to TSS 10.0 67 determine what vegetation would be located within the NH4-N 4.7 62 buffered areas. TIN 6.9 69 TP 0.95 48 3.3.2 Conservation areas and compact development Metals 0.1 50 Currently there are approximately 6000 ha of natural vegetation (forest, woodland, thicket and grassland) in the catchment as per the D’MOSS plan (medium 3.3 Conservation of natural systems and conservation; Figure 3.5). Most of the remaining natural biodiversity open space areas are undevelopable. Taking into account Conservation of natural systems and biodifversity within future planning as per municipal scheme development urban areas includes the protection of viable areas zonation plans (effective 2014), the amount of natural containing representative habitats and biodiversity, vegetation in the catchment is reduced to 2800 ha, or ensuring that the layout of protected natural areas 11% of the total catchment area (minimum conservation; is conducive to allowing ecological processes such as Figure 3.6). the movements between systems, and ensuring that downstream aquatic ecosystems are protected from Based on conservation targets set out in the eThekwini excesccive degradation through changes in the quality or Municipality Systematic Conservation Plan, a quanity of freshwater flows. conservation plan was determined as if starting from the undeveloped catchment (historical vegetation), but The latter is achieved in part through the interventions selecting currently undeveloped areas first. Using targets described above, and is considered one of the major for each vegetation type (forest, woodland, thicket and objectives of this study. In addition, two types of grassland) and the maximum remnant extent of each conservation intervention were considered for the study type of vegetation in the catchment, the additional area: the development of riparian buffer zones that amount and distribution of natural vegetation was Figure 3.3 The properties selected as suitable for application of source controls in the catchment (a) green roofs, (b) permeable paving, (c) estimated. Defining the additional conservation areas infiltration trenches and (d) soakaway pits. protect a band of natural vegetation on either side of rivers and streams, and the retention of a representative was done by following the current D’MOSS plan so as to amount of terrestrial natural open space that meets retain connectivity and to identify nodes that could be PAGE 18 EVALUATING THE POTENTIAL RETURNS TO INVESTING IN GREEN URBAN DEVELOPMENT IN DURBAN EVALUATING THE POTENTIAL RETURNS TO INVESTING IN GREEN URBAN DEVELOPMENT IN DURBAN PAGE 19 further increased. Vegetation was also retained on steep slopes. Overall it was determined that a total of 7000 ha were required to meet conservation targets, accounting for approximately 28% of the catchment area (Figure 3.6). Figure 3.4 The potential extent of riparian buffers in the Umhlatuzana-Umbilo catchment Figure 3.5 The total area of transformed land and natural open space (OS) in the catchment under status quo, minimum conservation, medium Figure 3.6 The extent of (a) conservation areas that will remain under existing development plans, (b) current conservation areas and (c) conservation and maximum conservation conservation areas that would have met biodiversity targets in the catchment PAGE 20 EVALUATING THE POTENTIAL RETURNS TO INVESTING IN GREEN URBAN DEVELOPMENT IN DURBAN EVALUATING THE POTENTIAL RETURNS TO INVESTING IN GREEN URBAN DEVELOPMENT IN DURBAN PAGE 21 The planned reduction in D’MOSS, current D’MOSS and ideal conservation plan were used in this study as alternative levels of conservation – minimum, medium and maximum - in the sceanrios analysed. Figure 3.5 shows the difference between the conservation options in terms of how much natural open space in a good condition is conserved and how much land has been transformed into other land uses, while keeping other factors in the catchment constant (industries, households, inhabitants, etc.). The status quo represents the current situation in the catchment where parts of the D’MOSS have become degraded. Medium conservation represents the current status quo situation had all natural vegetation in the catchment not been degraded. PAGE 22 EVALUATING THE POTENTIAL RETURNS TO INVESTING IN GREEN URBAN DEVELOPMENT IN DURBAN IV. SCENARIO SET-UP 4.1 Approach development path been followed, but one that doesn’t change the amount of households or economic activities This study aimed to evaluate the incorporation of in the catchment. This approach avoided making a large various types of green measures in urban development, amount of detailed (and probably wrong) planning all of which contribute to stormwater management, assumptions as would have been required if scenarios protection of biodiversity and maintenance of ecosystem were applied to greenfields areas earmarked for future services. These measures include novel, green development. The study is not about the costs and engineering solutions that mimic natural processes benefits of retrofitting these measures in the study area. (active stormwater management) to complement the Indeed, in some cases this would be impossible. Rather use of more traditional conveyance infrastructure that the study uses backcasting of an existing catchment removes the flood and waste water beyond the urban to derive useful policy recommendations for the environment, as well as conservation of riparian and greenfields areas. terrestrial areas. These measures will vary in efficiency for different purposes (e.g. retarding stormflows), and While the study area is the most developed catchment in from a stormwater management perspective may or the EMA, it is not yet fully developed as per the existing may not have co-benefits (e.g. recreational amenity). In zonation plans. Thus it was still necessary to make some order to evaluate the relative investment that should be assumptions about the characteristics of the latter areas. made in such alternative measures, this study has used a These were taken as being the same as those of nearby scenario evaluation approach. developed zones of the same type. It is important to note that the sanitation situation In scenarios that included increased conservation area is usually the primary factor driving water quality in relative to full development (medium and maximum African cities. The influx of poor people into cities leads conservation scenarios), it was assumed that a compact to the development of informal settlements and a development approach had been followed that used backlog in the setting up of adequate sanitation systems land more efficiently, incorporating taller buildings which requires considerable investment. If drainage and more compact infrastructure. Passive structural systems receive significant quantities of raw sewage, stormwater management measures (drainage systems, then measures to address water quality other than culverts etc.) are assumed to be in place as at present. conventional conveyance (e.g. out to sea) are less likely to be effective as they might be overwhelmed. Therefore whether the sanitation backlog has been addressed is an 4.2 Scenario elements important assumption that has to be made explicit in the scenarios. Since sanitation and solid waste management The green urban development interventions considered are existing obligations of all cities for health reasons are described in detail in Chapter 3. These were grouped alone, these do not form part of the economic analysis in into seven types, with one of these (source controls) this study. The study focuses on evaluation of different being comprised of four sub-types of interventions. green urban development measures in the presence of The grouping of source controls was done because adequate sanitation and solid waste management. they all have the same basic aim to retard flows during storm events, and could be applied in any combination, The scenario analysis used a backcasting approach in depending on their relative costs. Different levels were evaluating the alternatives. This involved assessing considered for source controls and conservation areas. what the outcomes would have been, had a different EVALUATING THE POTENTIAL RETURNS TO INVESTING IN GREEN URBAN DEVELOPMENT IN DURBAN PAGE 23 Table 4.1 Summary of interventions considered in the scenarios. Table 4.2 Scenarios used in the analysis. Levels of each intervention are described using symbols for ease of comparison. sw = stormwater. Intervention Description Possible extents if included Conservation / Sanitation and waste Green engineering (GE) / Non-structural stormwater Sanitation Sanitation measures implemented to meet existing needs and those Fully serviced management (SWM) Active stormwater management associated with planned development. These include improved compliance management and monitoring at existing WWTW, improved sanitation in informal Source Detention Treatment Riparian Conservation settlement areas and adequate sanitation for the planned development Litter Sanitation Controls Basins wetlands buffers areas areas. While meeting the sanitation backlog and servicing planned removal Scenarios S D W R Cons development areas will generate further sewage outputs, it was assumed that at least an equivalent amount of wastewater would be recycled, 1. Baseline: - - - - - - ● Sanitation resulting in no net increase in WWTW outputs. BAU (sanitation backlog) backlog Litter removal Measures to prevent and/or clean up solid waste (e.g. ban on plastic bags, Fully dealt with 2. GUD without bottles/compulsory refunds, community stewardship arrangements). - ● ●● ● ● ● ●●● sanitation Source controls On-site measures to reduce or neutralise the effects of built areas on runoff. Medium = all non-residential The medium extent included the application of green roofs, permeable properties 3. Clean baseline: ● ● - - - - ● paving and infiltration trenches on all commercial and industrial buildings Maximum = all properties BAU + SWM in the catchment. The maximum extent of implementation also included soakaways on all residential buildings 4. Cons2 ● ● - - - - ●● Conservation 5. Cons3 ● ● - - - - ●●● measures Detention basins Applied at strategic points throughout the study area to reduce or neutralise All sub-catchments the effects of built areas on runoff at a broader scale, and therefore also 6. R ● ● - - - ● ● addressing increased runoff from roads, walkways and other hardened surfaces. 7. R + Cons2 ● ● - - - ● ●● Treatment Artificial wetlands constructed below waste water treatment works for All WWTWs 8. R + Cons3 ● ● - - - ● ●●● wetlands polishing effluent and improving water quality. 9. S2 ● ● ●● - - - ● engineering Riparian buffers 15-30 m of natural vegetation is maintained along streams and rivers in All streams and rivers (1900 ha) 10. D ● ● - ● - - ● Green order to intercept and improve the quality of non-point source surface runoff from urban areas before it enters water courses as well as to provide 11. S+D ● ● ● ● - - ● a conservation corridor. 12. S2+D ● ● ●● ● - - ● Conservation Compact development designed to retain significant conservation areas, to Minimum = 2800 ha as planned 13. S+D+R ● ● ● ● - ● ● Combined areas (facilitated reduce hardened surface area and meet biodiversity conservation targets. Medium = 6000 ha as at present by compact This is fully achieved in the maximum conservation scenario, partly achieved Maximum = 7000 ha meeting 14. S2+D+W+R+Cons2 ● ● ●● ● ● ● ●● development) in the medium conservation scenario (which retains the current extent) and conservation targets minimally in the default scenario in which there is no compact development. 15. S2+D+W+R+Cons3 ● ● ●● ● ● ● ●●● 4.3 Scenarios it was decided to include full sanitation (as required by Scenario 1 (BAU) had full development as planned, but 1 allowed a test of the effect of sanitation alone on existing legislation) in most of the scenarios. A total of with the same level of backlog of sanitation and solid water quality. A range of scenarios was then constructed. Each 15 scenarios were designed and analysed (Table 4.2). waste services as at present. No green engineering scenario consisted of a combination of the different The remaining scenarios all included full sanitation and All scenarios assumed that the catchment was fully measures were implemented and the amount of natural interventions described above, applied to different litter prevention programmes, so “+SWM” is implied developed (as per municipal scheme zonation plans, open space was reduced to the planned extent of 2800 extents in the catchment (Table 4.2). The possible in Scenarios 3-15. Scenario 3 effectively provided a effective 2014), except that scenarios with medium or ha. This was termed the “Baseline”. Note that the number of combinations is very large, so these were “Clean Baseline” against which to evaluate the relative high levels of conservation meant that the development baseline is not the same as the present-day situation, chosen, not randomly, but with specific questions in net benefits of different engineering and conservation was more compact. All scenarios had the same number which was used to develop and calibrate the models, mind. Since GUD measures are unlikely to have much measures applied to different extents. Scenarios of households and the same amount of commercial and and against which water quality estimates are also positive impact in the absence of adequate sanitation, 4-15 were compared with Scenario 3, under the prior industrial activity. compared out of interest. assumption that adequate sanitation is both imperative In Scenario 2 (GUD without sanitation), all the and a prerequisite to GUD. green engineering and conservation measures were implemented but the sanitation backlog was not Scenarios 4 to 8 were set up to test the effects of addressed. This was to test the hypothesis that GUD different amounts of natural areas (conservation measures designed to address water quality are not areas and river buffers) on flooding and water quality. effective if sanitation is not properly addressed. Scenarios 9 to 12 were set up to test the effects of different combinations of green engineering Scenario 3 (BAU + SWM) was the same as Scenario 1 interventions. Scenarios 13 to 15 were set up to explore except that the sanitation backlog and litter problems the effects of implementing both green engineering and were addressed. A comparison of Scenario 3 to Scenario conservation measures to different extents. PAGE 24 EVALUATING THE POTENTIAL RETURNS TO INVESTING IN GREEN URBAN DEVELOPMENT IN DURBAN EVALUATING THE POTENTIAL RETURNS TO INVESTING IN GREEN URBAN DEVELOPMENT IN DURBAN PAGE 25 This page intentionally blank. PAGE 26 EVALUATING THE POTENTIAL RETURNS TO INVESTING IN GREEN URBAN DEVELOPMENT IN DURBAN V. SCENARIO MODELLING AND RESULTS 5.1 Overview A number of models were used during the scenario analysis to determine the potential benefits of different combinations and extents of green urban development interventions (Figure 5.1). Inputs into the models included the extent and quality of natural and semi- natural areas, the extent, design and performance of a range of green urban development engineering solutions for flood attenuation and water quality amelioration, and the amount and design capacity of conventional conveyance and waste water treatment infrastructure. In addition, assumptions were also made regarding the management of solid waste. Figure 5.1 The approach used for the scenario analysis and determining the costs and benefits associated with green urban development. EVALUATING THE POTENTIAL RETURNS TO INVESTING IN GREEN URBAN DEVELOPMENT IN DURBAN PAGE 27 A hydrologic (hydrology + hydraulic) model, carried out models developed for the municipality-wide eThekwini Recycling of wastewater and improved sanitation had the total combined attenuation capacity of source using PC-SWMM, produced flood hydrographs at specific ecosystem services valuation study (Turpie et al. 2017). the greatest impact on reducing nutrient loads into the controls was orders of magnitude higher. Considering points in the catchment and a sediment and nutrient harbour. Simulated TSS loads were highest under the that the reductions of TSS loads were similar for Scenario model, also carried out in PC-SWMM, produced water The possibility of also developing a dispersion model to BAU scenario with a total of just over 13 000 tonnes 9 and 10, it implies that detention basins are more quality outputs in the form of total suspended solids estimate water quality impacts on surrounding beach entering the harbour each year (Figure 5.2). This was effective at reducing TSS loadings than source controls. (TSS) and nutrient concentrations and loads at specific areas was considered but was not included in the study reduced to 11 845 tonnes under the Clean Baseline, a points of interest in the catchment. due to the expected low level of impact in this case. result of improved sanitation in the catchment and the Conservation areas and riparian buffers (Scenarios 4 – It should be noted that this might be highly relevant reduction in WWTW discharges to present-day levels 8) had a significant impact on reducing TSS loads into Information on flood hydrographs was used as an for other catchment areas in the EMA, particularly the due to recycling. Note that this change is significant, but the harbour (Figure 5.3). The increase in conservation input into the Infrastructure Costing Model in order uMngeni catchment. relatively small compared to what it might be in a typical area (medium and maximum extent) corresponds to an to value the reductions in flood risk under different African city catchment with a much higher proportion increase in the proportion of D’MOSS area by about 14 scenarios. The water quality data were inputs into a of unserviced informal settlements. Also, much of the and 17% and a decrease in TSS loads by about 18 and River Ecosystem Health assessment tool to evaluate 5.2 Water quality improvements and related change in this case comes from the reduction in WWTW 32% respectively. The establishment of river buffers river system changes, into an Estuarine Ecosystem benefits discharges. corresponded to a further 7% increase in conservation Services model to estimate changes in the value of those area and a further 31-36% reduction in the annual services, and were also used to estimate reductions in Both source controls and conservation measures improved TSS loading. The implementation of river buffers had dredging costs. 5.2.1 Impacts on sediment and nutrient concentrations and on the Clean Baseline situation. Source controls and a two-fold effect in terms of the modelling, i.e. the loads detention basins, when implemented independently change in landuse alters the hydraulic parameters and Changes in the quantity and quality of green open Annual sediment and nutrient loads were estimated (Scenarios 9 and 10), were not as effective at reducing the inclusion of treatment nodes (first-order decay space areas (and beaches) were inputs into a Tourism by simulating the total suspended sediment (TSS), sediment loads when compared to the conservation equations) reduce TSS loads along the flow path. Value model and a Hedonic Pricing Model, which total phosphorous (TP) and total inorganic nitrogen scenarios, and a combination of these interventions Scenario 8 had the largest impact of the non-structural estimated resultant changes in the tourism and property (TIN) loads using the PC-SWMM model (see Box 5.1). produced the best result (Figure 5.2 and Figure 5.3). interventions, reducing annual TSS loads by 63% (Figure values, respectively. The comprehensiveness of these By comparing the modelled sediment and nutrient 5.2 and Figure 5.3). models varied according to priorities, availability of outputs for each scenario, it was possible to estimate the The total capacity of detention basins within the potentially suitable modelling platforms and models, difference made by GUD interventions to the sediment subcatchment was approximately 450 000 m3, whereas and data availability. Some of the models built on and nutrient loads transported into Durban Bay. Note: A comparison of Scenario 2 and Scenario 15 showed that the application of green engineering and Box 5.1 Summary of sediment and nutrient modelling approach (details in Appendix 5). conservation measures in this catchment are still effective without the additional sanitation measures applied in Scenario 15, though the overall effect isn’t quite as good. Similar results were found for nutrients. This is because Annual sediment and nutrient loads were estimated by simulating the TSS, TP and TIN loads using the PC- the overall level of sanitation in this catchment is already relatively high. In comparison, the same green engineering SWMM model described in detail in Appendix 5. TSS, TP and TIN loads were simulated for one year from and conservation measures would be ineffective in Kampala where the proportion of the catchment under August 2013 to August 2014 (total annual precipitation of 572 mm for the Durban city area, lower than MAP). unserviced informal settlements is extremely high (Turpie et al 2016). Scenario 2 is not considered further and the The pollutant washoff from a given landuse during periods of wet weather was characterized in the model remainder of the report focuses on the relative benefits of different interventions in relation to a Clean Baseline. by using a user defined Event Mean Concentration (EMC). The Event Mean Concentration is a case of Rating Curve Washoff where the exponent is 1.0 and the coefficient represents the washoff pollutant concentration in mg/L. In each case build-up is continuously depleted as washoff proceeds, and washoff ceases when there is no more build-up available. The EMCs for TSS, TP and TIN were derived from literature and applied to the different landuse categories. Model subcatchment parameters were derived by area weighting the various land use parcels within each subcatchment. The modelled runoff flows were coupled with EMC values to estimate the concentration and total load of TSS and nutrients at different points in the catchment. The data can also be applied in PC-SWMM to determine the water quality volume to be catered for in stormwater management devices, such as source controls and detention basins. The percentage contribution of bed load and suspended sediments to the total sediment load is different for each catchment. For this study suspended sediment loads were multiplied by a factor of 1.25 in order to account for bed load. This is the factor generally applied in South Africa (Msadala et al. 2010, after Rooseboom 1992). The Umhlatuzana and Umbilo Rivers and other stormwater outfalls within the U60F catchment discharge directly into Durban Harbour. In the PC-SWMM model, there were 79 harbour outfalls in total. In addition to the outfalls, pollutant loads were simulated at a number of water quality monitoring stations situated throughout the catchment. A total of five monitoring stations on the Umhlatuzana River, three stations on the Umbilo River, and one station on the Mkumbaan River were included as points of interest. Monthly water quality data collected by the eThekwini Water and Sanitation Department were used for calibration and validation of the model. Simulations were run for a one-week period from the 1 – 7 July 2014 at a two second time interval, and the simulated values were within reasonable range of the measured values. The assumptions regarding untreated urban runoff quality, and the effects of stormwater interventions are outlined in Appendix 2. Figure 5.2 Annual TSS load (tonnes) for different outfalls into Durban Harbour for all scenarios PAGE 28 EVALUATING THE POTENTIAL RETURNS TO INVESTING IN GREEN URBAN DEVELOPMENT IN DURBAN EVALUATING THE POTENTIAL RETURNS TO INVESTING IN GREEN URBAN DEVELOPMENT IN DURBAN PAGE 29 Source controls (Scenario 9) and detention basins on TSS loads in the Umhlatuzana River due to a higher The accumulation of TIN and TP in the catchment is therefore most efficient at reducing nutrient loadings (Scenario 10), when implemented on their own, had a reduction in WWTW discharges (BAU compared to the predominantly derived from discharges from WWTWs. into the harbour (Scenarios 14 and 15). Source controls, low impact on TSS loads with decreases of 22% and 19% Baseline). The addition of treatment wetlands reduce For Scenarios 3 to 5, discharges from WWTWs detention basins and river buffers only helped to reduce respectively. Detention basins were only implemented in the TSS load from the WWTWs by about 68%, but accounted for 65-70% of the total annual TP loading the total TP and TIN loading generated from surface the Umhlatuzana and Umbilo Rivers and decreased the only reduce the total TSS load into the harbour by less into the harbour and about 50-60% of the total TIN washoff by a maximum of 15 and 20% respectively. total TSS load in this sub-catchment by 24%. than 10% overall. The results indicate that river buffers loading. The implementation of treatment wetlands was (Scenarios 6 – 8) were also effective at reducing TSS The total annual TSS loading at the different points loads. The most notable outcome of these simulations, along the Umbilo River and Umhlatuzana River are however, is that the preservation of the maximum extent shown in Figure 5.3. The three detention basins at the of D’MOSS areas with river buffers has similar effects as top of the Umbilo catchment were very effective at implementing a full treatment train approach. reducing TSS loads with almost a 90% reduction (Figure 5.3). Detention basins are most effective when water is The annual nutrient loading results have shown allowed to pond and sediments can settle. The efficacy that just by improving sanitation in the catchment of the detention basins diminished towards the lower (BAU compared to the Clean Baseline) there is a 30% ends of the catchment, where fewer were implemented reduction in total nitrogen (TIN) loads and a 35% and/or where flows were higher. The efficiency of the reduction in total phosphorous (TP) loads entering detention basins could be improved if their capacity the harbour each year (Figure 5.4). With sanitation in the lower reaches was increased or if more were backlogs taken care of, neither natural areas nor green implemented. engineering measures implemented on their own had much additional effect (Figure 5.4), although significant Overall, the source controls had a similar effect to localised effects were seen. The addition of treatment the detention basins, reducing TSS loads into the wetlands (Scenarios 14 and 15) resulted in the largest harbour by approximately 22% suggesting that these reduction in TIN and TP, with Scenario 15 reducing TIN measures are largely substitutable. The simulated by 63% and TP by 53%.. results for Scenarios 11 and 12 show that the inclusion of soakaways in residential areas with other source While conservation areas and riparian buffers have controls (i.e. permeable paving, green roofs and shown to be particularly effective at reducing TSS loads infiltration trenches) reduce TSS loads by an additional into the harbour and attenuating peak flows during 8%. Source controls were most effective at removing TSS small to medium flood events, these interventions when used in combination with detention basins. The (Scenarios 4 – 8) were less effective at reducing TIN implementation of a treatment train approach (multiple and TP. Scenario 8 reduces TIN loads by 19% and TP measures) was extremely efficient at reducing TSS loads, loads by 14%. Results suggested that source controls even with higher WWTW inputs. The accumulation of and detention basins, when implemented on their TSS in the catchment is predominantly derived from own (Scenario 9 and 10), are also not very effective surface washoff. The contribution from WWTWs is at reducing nutrient loads. The implementation of a significantly lower (Figure 5.3). The introduction of treatment train was most effective at reducing nutrient sanitation measures have a more significant impact loads into the harbour. Figure 5.3 The annual TSS loadings at different points along the Umbilo River (left) and the Umhlatuzana River (right). Refer to Figure A6. 1 for the monitoring station locations in the catchment. Figure 5.4 Annual TP and TIN loads for different outfalls entering Durban Harbour for all scenarios PAGE 30 EVALUATING THE POTENTIAL RETURNS TO INVESTING IN GREEN URBAN DEVELOPMENT IN DURBAN EVALUATING THE POTENTIAL RETURNS TO INVESTING IN GREEN URBAN DEVELOPMENT IN DURBAN PAGE 31 5.2.2 Impacts on river condition Changes in modelled water quality at the ten monitoring (ZANA_35), where dry season TP concentrations were Changes in river condition associated with each of the quality, and qualitatively, in terms of some of the broad sites in the catchment (Figure A6.1) were interpreted generally much lower than in the wet season. The reason different scenarios were assessed quantitatively (see Box parameters considered in assessments of ecosystem with regard to the river health categories for water for this may be that most of the phosphorus in the river 5.2, Appendix 6), in terms of modelled instream water health (see Section 2.6). quality outlined above (see Box 5.2 and Appendix 6). water in these reaches is attached to sediments and/or is Figure 5.5 and Figure 5.6 compare concentrations of in a particulate form, rather than being orthophosphate Total Phosphorus and Total Suspended Solids (TSS), and derived. The site lies upstream of any waste water Box 5.2 Summary of river ecosystem health assessment approach (details in Appendix 6) Total Inorganic Nitrogen (TIN), respectively, averaged treatment works, and sources of orthophosphate into The quantitative assessments used modelled concentrations of the three parameters included in the across all sites for each modelled assessment date, the system appear to be limited. During the dry season, hydrological and water quality model, namely total phosphorous (TP), total suspended solids (TSS) and Total for each scenario. Scenario 0 (“Present day”) provides mobilisation of sediments was also generally reduced Inorganic Nitrogen (TIN). Modelled hourly concentrations of these parameters were available for the 10 sites actual water quality data derived from the various river and this might explain the low measured dry season total shown in Figure 5.8 for a one-year period. The time series dataset was simplified to mean daily concentrations. monitoring sites. phosphorus concentrations in data from this site. These were presented in terms of different river health or condition categories, as recommended in South Merging of data from all sites provided a basic Phosphorus-based enrichment in the catchment is African (national) draft guidelines (DWAF 2008) and interpreted in Table 5.1. The range and/or threshold values assessment only of changes in water quality as a result significant, with no sites under Present Day conditions for the variables considered in this study were also taken from DWAF (2008), using the ranges defined for each of the different scenarios, but at least allowed broad being better than a Category D with regard to this River Health Catergory (see Appendix 6 for more detail). trends to be considered. By contrast, the data presented nutrient, and most sites (with the exception of ZANA_35, Table 5.1 Comparison of different systems for the categorisation of river health/condition in Appendix 6 provided a more detailed comparison which ranges between Category D and E for phosphorus) data, simplified after DWAF (2008). from key monitoring points in the catchment, modelled being in a Category F for most of the time (see Figure Difference from natural Category Natural to poor category for each of the different scenarios. These figures A6. 2). The implication of this is that for meaningful No change A Natural allowed for spatial differences in the effects of different rehabilitation of water quality to be achieved, scenarios to be determined. This is useful for developing a substantial reduction in instream phosphorus Small change B Good recommendations around prioritising different measures concentrations is necessary. Since “sustainable” aquatic Moderate change C Fair for different river reaches, and where rehabilitation of ecosystems are assumed to lie in a condition of Category Large change D river water quality above a certain minimum condition D or better (Kleynhans et al. 2005), this implies that Serious change E may be required. achievement of PES ratings better than Category E with Poor Extreme change F regard to total phosphorus should be a prerequisite of The Present Day concentrations of water quality future management strategies and the scenarios were Total suspended solid (TSS) data could not be interpreted in this manner. DWAF (2008) does not in fact quantify variables in the river systems generally followed evaluated with this in mind. TSS concentrations for different health rating values, on the basis that these data are not routinely measured by expected patterns for urban environments, with the the Department of Water and Sanitation (DWS). Existing water quality guidelines for TSS are limited. Even when highest concentrations of TIN and TP occurring in the Although concentrations of TIN were higher than for Reference (“Natural”) Condition TSS data are available for a particular system, their value is often restricted. dry season (winter), when dilution as a result of surface phosphorus, the ecological implications of nitrogen This is because of the tight links between sediment transport and flow velocity, and the largescale differences runoff is lowest, and TSS tending to be lowest in the nutrients are less severe, with data for sites in the in sediment transport depending on discharge. In the current situation, therefore, while modelled TSS dry season, when flows are assumed to be too low to catchment (see Figure A6. 4) showing that Present Day concentrations allowed comparison between different scenarios, and some comment on likely removal rates of mobilise high sediment loads (Figure 5.5 and Figure 5.6). TIN concentrations were never in a worse category than other parameters expected to be associated with inorganic sediments (e.g. heavy metals and total phosphorus) Category E, and in a Category B to C in the upper reaches they could not be used to infer absolute erosion and sedimentation rates. Guidelines for the interpretation of When individual sites were assessed (refer to Appendix of the Umhlatuzana River (ZANA_35). links between turbidity and river health (after DWAF 2008) were also drawn on in this regard. 6), the exception to the above was seen to be the least- impacted upper reaches of the Umhlatuzana River The qualitative assessments carried out in this project considered the metrics used in Present Ecological State (PES) assessments of turbidity (for which TSS is considered a surrogate value, see Appendix 6) to infer qualitative change in river condition as a result of scenarios involving attenuation of runoff and the provision of riparian corridors and buffers. Changes in TSS data produced by the model were used to infer (but not quantify) changes in sediment transport and erosion. In addition to limitations in the applicability of TSS data in inferring catchment-scale erosion and sediment processes, the following limitations must also be considered with regard to the approach taken to assessment of the impact of the different scenarios on river condition, and in particular, on river water quality, in this study, namely: ƒƒ The assessments are limited to TP, TSS and TIN and do not take account of other variables such as ammonia- nitrogen, various heavy metal concentrations; bacteria, salinity; ƒƒ The lack of data for other important water quality parameters means that interacting parameters are not considered (e.g. the influences of dissolved oxygen, pH, temperature and (in the case of some heavy metals) water hardness); and ƒƒ The lack of data regarding the organic component of TSS and the proportions of total ammonia and orthophosphate comprising TIN and TP respectively. All threshold values, guidelines, metrics and assumptions used to assess river ecosystem health are outlined in Appendix 6. PAGE 32 EVALUATING THE POTENTIAL RETURNS TO INVESTING IN GREEN URBAN DEVELOPMENT IN DURBAN EVALUATING THE POTENTIAL RETURNS TO INVESTING IN GREEN URBAN DEVELOPMENT IN DURBAN PAGE 33 Figure 5.6 Effects of different modelled scenarios on total inorganic nitrogen (TIN) concentrations at different monitoring sites in the study area. Site locations as shown in Figure A6.1.. Scenario 1 (BAU) outputs indicated that in most cases, concentrations (see modelled TSS in Figure A6. 3). This is a BAU approach would result in worsening of water assumed to be the result of upstream detention of less- quality compared to present (and largely unsustainable impacted water flows (see Appendix 2), and as a result, conditions, below PES Category D), thus providing increased concentrations downstream (Figure A2. 3 shows motivation for implementation of GUD approaches. the locations of modelled instream detention ponds, with Under the GUD scenarios, the modelled data need to ponds located upstream of the Hillcrest WWTW). be assessed with regard to where GUD structures are actually placed, taking into account river character / The addition of sanitation without any other GUD condition in different reaches and the existing known measures (Scenario 3 – Clean Baseline) resulted in sources and types of water quality impairment. some improvement in river water quality compared to Assessment of the upper sites on the Umhlatuzana River, Scenario 1 (BAU), and even compared to the Present for example (ZANA_35) show that implementation of Day. The impact on a site-by-site basis is not dramatic, basic sanitation measures (Scenario 3) as well as the however, and reflects the fact that while the water addition of basic GUD measures would have little impact quality impacts of informal settlements on downstream on dry season water quality and only a slight impact on (estuary) loading may be profound, their actual impacts wet season water quality. This is because most of the at a river level are localised, affecting their receiving proposed measures had little application in this part of rivers and the downstream reaches only. Thus the the catchment, which is least affected by poor sanitation implementation of improved sanitation measures issues. Moreover, TSS concentrations at the downstream would affect only those areas currently affected by Figure 5.5 Effects of different modelled scenarios on concentrations of Total Phosphorus and Total Suspended Solids (TSS), respectively. Site site (ZANA_34), downstream of the Hillcrest WWTW, informal settlements – e.g. the middle reaches of the locations as shown in Figure A6.1. were potentially negatively affected compared to BAU Umhlatuzana River and the lower (Cato Manor) reaches PAGE 34 EVALUATING THE POTENTIAL RETURNS TO INVESTING IN GREEN URBAN DEVELOPMENT IN DURBAN EVALUATING THE POTENTIAL RETURNS TO INVESTING IN GREEN URBAN DEVELOPMENT IN DURBAN PAGE 35 of the Umbilo River (see Figure 2.6). In catchments had little effect on instream nutrient concentrations. important effects of the implementation of GUD 5.2.3 Avoided costs due to sediment retention where informal settlements comprise a significant They did, however, have significant implications for TSS, measures such as removal of heavy metals from urban The avoided sedimentation was taken as the difference proportion of landuse, changes in water quality as highlighting the modelled assumption that these factors runoff that might be applicable to this system were not in annual sediment yield from the catchment between a result of improved sanitation would obviously be play a significant role in reducing erosion of the stream considered in the model. the modelled scenarios. The avoided costs were more pronounced. It must be stressed in this regard, banks and bed (thus decreasing downstream sediment estimated using dredging data provided by Transnet for moreover, that the modelled data indicate only patterns load) as well as actively trapping sediment in diffuse From a strategic perspective, the above findings need Durban Harbour for the period 1 April 2015 – 31 March of major nutrient concentrations. The potentially more runoff and thus reducing sediment concentrations. At to be considered in terms of the main objectives of 2016. Maintenance dredging involves the removal of significant effects of bacterial loading on river water the same time, however, it is noted that interventions inclusion of GUD interventions in a particular catchment sediments from channels, basins and berths within the quality and its fitness for human or other use has not such as detention basins would also play a role in management strategy. From a sustainable practice harbour. Dredging of channels and basins costs R85 per been explored in this study. sediment removal, thus explaining the significant perspective, all aquatic ecosystems should be managed m3 on average and dredging of berths costs R636 per effect of Scenario 10 on all aspects of water quality. Of so as to fall within a Category D or better. The modelled m3. A total of just under 182 000 m3 of sediment was The modelled data shown in Appendix 6, Figure 5.5 and interest, however, is the modelled effect of Scenario 10 water quality data indicate that this would be readily dredged from the harbour over the period 2015/16, Figure 5.6 generally show that while improved sanitation on dry season flows throughout the study area, showing achievable in terms of nitrogen nutrients throughout at an overall average cost of R229 per m3 (Transnet affects phosphorus concentrations to some extent, as do a marked increase in TSS concentrations at this time, the catchment, and generally but not always achievable NPA). However, most of the sediment removed from the implementation of riparian buffers and conservation out of keeping with the assumed role of detention in terms of phosphorus enrichment, provided that the the harbour through maintenance dredging is not areas (scenarios 4-8) it is in fact the implementation of ponds in terms of sediment trapping. Modelling of full suite of GUD measures was implemented. This derived from river inputs but is from existing estuarine source control measures (see scenario 9) as well as the sediment loads in river systems is complex, and this means that sustainable management of the affected sediments that are shifted by the movement of large implementation of typical GUD measures (scenarios pattern is assumed to reflect problems with the model river systems requires more than simple application of ships through the harbour (Transnet NPA). The “silt 10 – 15) that assert by far the most significant impacts rather than real patterns of elevated sedimentation. existing legislation with regard to the quality of effluent canal” at the top end of the estuary is not dredged on on total phosphorus concentrations. Almost throughout To some extent, however, the absence of riparian inputs, in highly developed urban areas. Upstream parts a regular basis and it is estimated that less than 5% of the study area, these measures are responsible for buffers and conservation areas from Scenario 10 means of the catchment (e.g. ZANA_35) would require far less the annual volume of dredged sediment is of fluvial reducing modelled phosphorus concentrations by at that sediment availability in terms of the model was stringent application of GUD measures in order to meet origin (Clive Greyling, Transnet NPA pers. comm.). This least one PES category. Nevertheless, these figures need also elevated at all times in the catchment under this minimum water quality thresholds for Category D PES, corresponds to the modelled current TSS load for the to be interpreted with some caution. To some degree, scenario. reflecting the lower levels of impact from existing or U60F catchment which represents 4% of the annual the dramatic reduction in instream concentrations of future landuse scenarios in this area. This outcome is volume of dredged sediment. The proportion of total phosphorus is the result of the more generic application While interventions such as riparian buffers and not unexpected – increased density of settlement and/ sediment due to the bedload was included in the of source control measures at a catchment level, conservation areas may have little measurable effect or cumulative catchment development generates more estimation of mean annual sediment loads. The bedload compared to the site specific application of measures on other aspects of water quality (e.g. phosphorus or impacts with regard to water quality and quantity, and is generally thought to comprise about 20% (by mass) of such as improved sanitation, of relevance only to areas nitrogen nutrient concentrations) they do play a major by implication should require increased effort to offset the total sediment yield from rivers. It also represents affected by informal settlement. role in determining overall river habitat integrity or such impacts. the coarser fraction of the sediment yield spectrum. condition, through the provision of longitudinal corridors The results suggested that the application of However, where the main objective of GUD measures Msadala et al. (2010), used a factor of 1.25 to cater for for faunal movement (important in an increasingly detention ponds (scenario 10) was by far the most is simply to protect downstream ecosystems (e.g. the bedload and non-uniformity in suspended sediment urban environment) as well as areas in which indigenous significant intervention with regard to bringing about estuary), the design and spatial layout of GUD measures concentrations in order to estimate the mean annual riparian vegetation can be established and allowed to an improvement in instream water quality nutrient can be more strategic, allowing a focus on loading rather sediment load (after Rooseboom et al. 1992). spread into downstream reaches, improving indigenous concentrations, although not with regard to TSS habitat quality and function. While such attributes do than concentration of variables. This should, however, The annual TSS loads generated under different reduction (see Figure A6. 3, Appendix 6). The modelled not add to water quality PES Category, they do add still be considered with regard to South Africa’s own scenarios were compared to Scenario 3 as the baseline. effects of the detention ponds were, however, slightly significantly to overall habitat quality, and should thus be legislation, which requires the maintenance of minimum The annual TSS loads were converted from kg to m3 problematic in that all of the sediment ponds were regarded as playing a more substantial role in catchment sustainable (i.e. Category D) water quality standards for using a sediment density of 1350 kg/m3 (Rooseboom assumed to be located along water courses. In practice, ecosystem function than suggested by the data under all aquatic ecosystems. This said, it is also acknowledged 1992) and multiplied by a factor of 1.25 to cater for this is unlikely to be ecologically sanctioned in significant current consideration. that river condition is a composite of many variables, bedload and non-uniformity in suspended sediment streams, as a result of the ecological implications, and including water quality, and that achievement of concentrations. The mean annual sediment load was the catchment-scale impacts of TSS removal from The modelled data included sites at the downstream minimum standards with regard to alien plant clearing then multiplied by the overall average dredging cost, the system in Scenario 10 are assumed to be over- ends of the catchment, notably sites OF1 and OF2. Total and indigenous plant restoration may be as important with the difference in costs between the scenarios emphasized by the model. Associated with TSS removal phosphorus data were not available for these sites. for restoring sustainable riverine ecosystems as representing the costs avoided for dredging of Durban would, moreover, be the removal of total phosphorus, TSS and TIN data shown in Figure A6. 3 and Figure A6. improvement of water quality. harbour. with Campbell (2001) for example estimating that 4 respectively indicated similar effects of the different some 70% percent of total phosphorus from informal scenarios on river condition at OF2_UMB at the inflow The application of detention basins in the model is one A decrease in TSS loads into the harbour resulted in settlements is bound to TSS. Implementation of into the estuary to those discussed generally above. area where an extended modelling and assessment a decrease in annual maintenance dredging costs approaches such as the use of treatment wetlands However, the results for OF1_UMB at the downstream period would probably be useful. The detention associated with maintaining channel depths in the clearly have the most impact on reaches downstream of end of the small Umhlatuzana system, just upstream basins have been modelled as instream systems – a harbour. The modelled scenario results suggest that the the WWTWs where they are implemented – thus total of OF2_UMB, showed no clear patterns in TSS or TIN practice that would not be ecologically defendable in annual dredging costs avoided range from R0.46 million phosphorus and TIN concentrations are most impacted reduction as a result of implementation of the selected systems with high actual or desired ecological integrity. for Scenario 4 to R1.75 million for Scenario 15 (Table 5.2) by scenarios involving these measures (scenarios 14 scenario interventions, suggesting that landuse in the Moreover, in practice, SUDS approaches commonly as a result of decreased sediment loads into the harbour. and 15) at sites downstream of WWTWs (i.e. ZANA_34; catchment of this small system (a highly industrialised incorporate many small off-channel detention basins, This equates to a net present value for dredging costs ZANA_28 and Umbilo_04). urban area) may not respond to the scenarios as that would not have the implications for flow rates and avoided of R5.3 million to R20.1 million (Table 5.2). developed in this study. For example, the system was associated water quality that appear to characterise the Interventions such as riparian buffers (scenarios 6-8) not included in the development of significant riparian modelled scenarios involving the use of detention ponds, and conservation areas (varying between scenarios) corridors along the larger major rivers. Moreover, and Scenario 10 in particular. PAGE 36 EVALUATING THE POTENTIAL RETURNS TO INVESTING IN GREEN URBAN DEVELOPMENT IN DURBAN EVALUATING THE POTENTIAL RETURNS TO INVESTING IN GREEN URBAN DEVELOPMENT IN DURBAN PAGE 37 Table 5.2 The total annual dredging cost for each scenario, and annual and NPV dredging costs avoided (R millions) for each scenario when collated and summarised for the most recent National result of improved estuary condition were assessed by compared to the baseline Biodiversity Assessment of estuaries (van Niekerk & comparing the effect of water quality improvements, i.e. NPV dredging costs Turpie 2012). The relationships between input loads a reduction in nutrient and TSS loads between scenarios. Total annual dredging % decrease in TSS Annual dredging costs avoided (20 yrs., 6%, R and water quality, and between water quality and The gain in estuarine and marine fishery values as a Scenario cost (R millions) loads avoided (R millions) millions) fish abundance determined in the Durban ESV study result of decreased nutrient and sediment loads into 3. Clean Baseline 2.51 (Turpie et al. 2017) were used to estimate the potential Durban Bay when compared to Scenario 3 are shown in 4. Cons2 2.05 18% 0.46 5.3 improvements in the fishery and nursery value of Table 5.3. The percentage gain in estuarine and marine 5. Cons3 1.70 32% 0.81 9.3 Durban Bay as a result of the GUD interventions. fishery values compared to the baseline ranged from 9% 6. R 1.36 46% 1.15 13.2 for Scenario 4 to 55% for Scenario 15, translating into an Increased fish exports to the marine environment annual gain of R0.6 - R3.5 million and a net present value 7. R + Cons2 1.15 54% 1.36 15.6 and improvements in the estuarine fishery value as a of R6.8 – R40.2 million. 8. R + Cons3 0.92 63% 1.59 18.2 9. S2 1.95 22% 0.56 6.5 Table 5.3 Estimated gains in estuarine and marine fishery values value due to a reduction in TSS and nutrients for scenarios 4-15 when compared to Scenario 3 as the baseline. 10. D 2.04 19% 0.47 5.4 % gain in estuarine & marine 11. S+D 1.43 43% 1.08 12.4 fishery values Annual value gained NPV 12. S2+D 1.22 51% 1.29 14.8 Scenario (compared to the baseline) (R millions) (20 yrs., 6%, R millions) 13. S+D+R 1.15 54% 1.36 15.7 4. San+MCons 9% 0.60 6.83 14. S2+D+W+R+Cons2 0.97 61% 1.54 17.7 5. San+HCons 18% 1.14 13.04 15. S2+D+W+R+Cons3 0.76 70% 1.75 20.1 6. San+Rip 28% 1.79 20.57 7. San+Rip+MCons 36% 2.29 26.27 8. San+Rip+HCons 46% 2.94 33.73 5.2.4 Impacts on estuary condition and fishery values 9. San+SC2 12% 0.74 8.53 A loss of natural system function and integrity in the 10. San+DB 10% 0.61 7.03 The relationships between water quality and fish catchment results in a loss in estuary functioning as abundance have been well studied in KwaZulu-Natal 11. San+SC+DB 26% 1.65 18.97 nutrients and TSS loads increase. This has an impact on (Cyrus et al. 1987, 1988). Expert understanding of these 12. San+SC2+DB 33% 2.12 24.32 estuary condition and changes the values and outputs and other relationships is used in the quantification of 13. San+SC+DB+Rip 36% 2.30 26.35 associated with different ecosystem services – such estuarine responses to changes in in the quantity and 14. AllGUD1 44% 2.79 31.98 as tourism and recreation, nursery value and fishery quality of freshwater inflows to estuaries. A substantial outputs. 15. AllGUD2 55% 3.50 40.19 amount of work has also been carried out by groups of estuarine specialist scientists to describe the present A deterioration of water quality in estuaries as a result status of estuaries throughout South Africa, following of increases in nutrient and suspended sediment loads standardised methods of describing estuarine health would lead to impacts on fish stocks, estuarine fishery developed for the setting of environmental flow values and the export of fish to marine fisheries (Figure requirements as well as for estuary management more 5.7). Water quality improvements, such as reduced generally (Turpie et al. 1999, 2012). These require the 5.3 Flood attenuation benefits nutrient and TSS loads, will have the reverse effect with description and scoring of all the abiotic and biotic increased fish exports to the marine environment as a components of estuaries, including water quality result of improved estuary condition. 5.3.1 Impacts on flood peaks variables and fish communities. These scores have been Urbanisation and its associated increase in impervious software interface (see Box 5.3). The main aim of surface area are synonymous with increases in flood the hydrological modelling for the scenario analysis peaks. Urban stormwater systems are designed to included investigating the effects of the implementation effectively drain a catchment and reduce the impacts of of various types and/or combinations of green urban development on flood flows, thereby altering catchment development measures, such as source controls, hydrology. Traditionally, urban stormwater systems drain detention basins, riparian buffers and conservation frequent (minor) storms and associated runoff from areas on flood hydrographs at selected points within the properties to natural watercourses. During severe storm Umhlatuzana-Umbilo catchment. events, open spaces and road systems are considered acceptable drainage components of major stormwater The PC-SWMM model for the Umhlatuzana – Umbilo systems. catchment, which includes flow paths, the location of detention basins, WWTWs, water quality monitoring Hydrological modelling of the Umhlatuzana-Umbilo stations and the extent of source controls in commercial, catchment was performed in the US-EPA SWMM5 industrial and residential areas is shown in Figure 5.8. Figure 5.7 Schematic summary of the linkages from water quality parameters to estuary ecosystem services for a case of deteriorating water hydrology and hydraulics engine, using the PC-SWMM quality. PAGE 38 EVALUATING THE POTENTIAL RETURNS TO INVESTING IN GREEN URBAN DEVELOPMENT IN DURBAN EVALUATING THE POTENTIAL RETURNS TO INVESTING IN GREEN URBAN DEVELOPMENT IN DURBAN PAGE 39 Box 5.3 Summary of the flood modelling method (details in Appendix 3) SWMM5 is an integrated, physically-based model that was selected for this purpose based on a review of a range of possible modelling options and discussions with eThekwini Municipality’s Catchment, Stormwater and Coastal Management (CSCM) department. CSCM had recently completed the migration of completed HEC-RAS hydrologic models of the EMA to SWMM5. GIS subcatchment data available for this project from the eThekwini Municipality (EM) flood studies were in the order of 1 km2 and larger. The subcatchments were further discretised into smaller subcatchments, in the order of 0.2 km2, within the U60F catchment. A spatial analysis tool was then used to process the flow paths, watershed boundaries, and river centre lines. The outlet points for the model were then identified and selected; i.e. estuaries and stormwater infrastructure. EM flood models, based on geometric HecRAS files, were imported into the PC-SWMM model. The stormwater network shapefile for the catchment was incomplete and contained numerous errors and inconsistencies. These networks were amended where possible. A Stormwater Management System audit (visual inspection and assessment) of all stormwater infrastructure, recently carried out by the eThekwini Municipality, produced a number of new shapefiles that were imported into the model and connected to exisiting stormwater networks and flow paths. was included where availabale. The original stormwater shapefile was merged with the new SMS shapefiles and both were connected to the HECRAS and WDT flow paths. Available current GIS landuse files (e.g. zoning files, D’MOSS) were collated and reviewed. These files were concatenated into one consistent landuse polygon shapefile and the landuse classifications were reviewed and summarised into a common set of landuse conventions. The resulting shapefile was ‘groundtruthed’ using aerial imagery for the U60F catchment. A number of input parameters are required for SWMM5. These include hydraulic parameters, soil infiltration properties, rainfall and water quality parameters. Figure 5.8 SWMM model of the U60F catchment including flow paths (yellow and red lines), detention basins (green squares), WWTWs (larger The determination of the catchment characteristics was estimated using a spatial analyst tool for zonal black dots), water quality monitoring stations (smaller black dots), extent of source controls for commercial and industrial areas (grey shaded) and residential areas (purple shaded areas). The Umhlatuzana-Umbilo catchment, the Amanzimyana Stream catchment and statistics. Raster files were generated to represent the information required for the hydraulic and all other contributing catchments have been outlined. hydrological models, with reference to each subcatchment. The most significant input hydraulic parameter is the percentage of impervious area (Imperv. %). The hydraulic parameters were assigned to each landuse classification based on literature. The largest proportion of rainfall losses over pervious areas generally The Umhlatuzana-Umbilo Canal is the main stormwater effective at reducing peak flows during higher return occur due to soil infiltration. The Green-Ampt method was adopted for this study. This method provides outfall that discharges into Durban Harbour. Simulated periods. a soil memory as opposed to a broad brush coefficient approach. Three user-specified soil parameters peak flows for the Canal, under different return periods, were used; i.e. capillary suction head, saturated hydraulic conductivity, and the maximum available are shown in Figure 5.9 and the relative change from It is important to consider two key points when moisture deficit. Average daily abstractions and return flows/discharges were added as point sources at the the baseline is shown in Figure 5.10. These results interpreting these results: appropriate junctions. suggest that structural GUD measures, in combination with conservation areas, can have a significant impact 1. For low flow events (0.5- and 1-year return period), A user-defined hyetograph was used as the precipitation input into the model. The hyetograph was created on reducing peak flows, especially at lower flood return caution should be applied when interpreting what using the total daily mean-areal precipitation depths derived by Smithers & Schulze (2000) for a 24-hour periods. constitutes a significant change in peak flows as design storm. The temporal distribution was derived using a synthetic SCS Type II distribution for 2-, 5-, 10- numerical instabilities in the model may appear to be and 20-year return periods. As expected, Scenarios 14 and 15 have the most relatively large in this regime (i.e. a 1m3/s variation significant impact on peak flows across the different may be represented by a 10% difference). It is The Umhlatuzana and Umbilo Rivers and other stormwater outfalls within the U60F subcatchment flood return periods (Figure 5.9, Figure 5.10). therefore more appropriate to consider the absolute discharge directly into Durban Harbour. In the SWMM model, there are 79 outfalls in total. The peak flows Conservation areas and riparian buffers (Scenario change rather than the relative change; and and flow volumes were estimated for each scenario for a 2-, 5-, 10- and 20-year return period. When 7 and Scenario 8) have an impact on reducing peak considering the peak flows downstream in isolation, the effectiveness of each GUD intervention can be flows, especially at the medium return periods (2- and 2. Stormwater networks are generally designed to undetectable due to the overall impacts of the catchment and therefore flood hydrographs were produced 5-year) when compared to other scenarios that include manage flows caused by lower rainfall return periods at different points throughout the catchment in order to isolate the effect of GUD interventions at a higher only structural interventions. When the structural (i.e. the EM stormwater management design policy is resolution. Assumptions regarding effects of stormwater management interventions on flows are outlined GUD interventions are applied independently, such as based on a 2- to 5-year return period). Once flows in Appendix 3. in Scenario 9, with only source controls, their impact exceed the capacity of the stormwater network (i.e. is significantly reduced, or in Scenario 10 (detention during high return periods), surcharges occur in the basins) where the impact on low flood return periods stormwater system and water tends to flow along the is significant but declines at high return periods (Figure surface. The model was setup to include continuity in 5.10). This demonstrates the importance of applying the total flows, but if a junction is surcharged, the a treatment train approach which includes a mixture peak flow will be underestimated. Although this does of structural GUD interventions in conjunction with not provide a relative change in peak flows in this conservation measures. Furthermore, the treatment regime, this is what realistically occurs in-situ. train approach (i.e. Scenarios 14 and 15) is the most PAGE 40 EVALUATING THE POTENTIAL RETURNS TO INVESTING IN GREEN URBAN DEVELOPMENT IN DURBAN EVALUATING THE POTENTIAL RETURNS TO INVESTING IN GREEN URBAN DEVELOPMENT IN DURBAN PAGE 41 Impervious surfaces reduce infiltration and increase other source controls implemented in commercial and surface runoff, leading to higher peak flows and industrial areas only (i.e. permeable paving, green roofs reduced runoff response times. The results show that and infiltration trenches). This was partly explained by as the amount of conservation area in the catchment the fact that the soakaways were implemented on a increases and the percentage of impervious surface area larger scale, and so had a larger storage capacity. decreases, the relative change in peak flows increases (Figure 5.11). In fact, an 8% reduction in the impervious The effect of different measures at attenuating peak surface area of the catchment (from the baseline to flows varies spatially throughout the catchment and as Scenario 4) corresponds to about a 15% decrease in peak such there is rationale for isolating the effectiveness flows. However, the distribution of D’MOSS areas within of source controls and detention basins at a higher the catchment also had an influence on peak flows, as resolution within the catchment. When considering the illustrated by the relative change between Scenarios peak flows downstream in isolation, the effectiveness of 4 and 5. Although conservation areas covered a larger each intervention can be undetectable due to the overall extent under Scenario 5, this only corresponded to a impacts of the catchment. 1% increase in impervious area. The location of the conservation areas within the catchment is therefore The impact of different source control measures on important. Examples of this can be seen when peak reducing peak flows was assessed by examining the flows are assessed at different sampling points, rather hydrographs for specific points in the catchment (Figure than as an overall impact at the outfall into the harbour. 5.13). For some points where there was no detention At a sampling station which drains from an area north- basin situated upstream, the peak flows were equal for east of the Umbilo River (R_Mkumbaan_01, refer Figure Scenarios 9 and 12 (see R_Chats_15, Figure 5.13), and 5.8) peak flows increase from Scenario 4 to 5 (Figure only a small difference was seen between Scenario 11 5.12a), due to the changes in the amount and location of and 12. At other points, where detention basins were conservation area under each scenario. However, in the implemented upstream, their impact was noticeable, Figure 5.9 Summary of the simulated peak flows for all scenarios and all return periods for the Umhlatuzana/Umbilo Canal outfall into the especially their ability to reduce initial runoff (as shown harbour. upper catchment above the Hillcrest WWTW where the amount of conservation area is increased from Scenario by the orange line of Scenario 9 in Figure 5.13 for the 4 to Scenario 5, the reduction in peak flow was found to Umbilo and Mkumbaan points). The change in peak be more significant for Scenario 5 (Figure 5.12b). flows for stations situated in the highest parts of the catchment were found to much lower due to fewer While the implementation of river buffers (Scenarios source controls being implemented in these areas. The 6 – 8) generally reduced peak flows, their impact was difference in peak flows observed for Scenario 11 and relatively minor when compared to other measures 12 highlight the effectiveness of soakaways, which were (Figure 5.11). Source control measures implemented implemented in Scenario 12 but not in Scenario 11. alone (Scenario 9) reduced overall peak flows by about 10%. Soakaways in residential areas were more effective at attenuating peak flows when compared with the Figure 5.10 Summary of the relative change in peak flows compared to the baseline for all scenarios and all return periods for the Umhlatuzana/Umbilo Canal outfall into the harbour. Figure 5.11 Percent impervious surface area and associated relative change in peak flow for the conservation scenarios compared to the baseline PAGE 42 EVALUATING THE POTENTIAL RETURNS TO INVESTING IN GREEN URBAN DEVELOPMENT IN DURBAN EVALUATING THE POTENTIAL RETURNS TO INVESTING IN GREEN URBAN DEVELOPMENT IN DURBAN PAGE 43 Figure 5.12 Simulated peak flows for a 2- and 5-year rainfall return period for the Baseline and Scenarios 4 and 5 at sampling points (a) R_ Mkumbaan_01 and (b) R_Zana_35. Detention basins (Scenario 10) were found to be more periods and is reduced during higher return period flows, effective at reducing peak flows during small to medium especially above the 500m3/s regime. This is largely due return periods than during large flood events (Figure to the capacity of the detention basins which is rapdily 5.14). The hydrographs at three different sampling exceeded during higher flood return periods resulting in locations on the Umhlatuzana River and one location on overflow. The purpose of a detention basin is to extend the Mkumbaan River show the difference in peak flows the time of concentration and thus regulate the flow between Scenario 10 and the Baseline (Figure 5.14). The downstream. observed difference is much larger in the smaller return Figure 5.14 Simulated peak flows for a 2-, 5-, 10- and 20-year rainfall return period for Scenario 10 compared to the Baseline at sampling points (a) R_Zana_35, (b) R_Zana_29, (c) R_Zana_10 and (d) R_Mkumbaan. Note that time units are reporting time steps from the model with one time unit representing 5 minutes. Another example, further illustrating the effect of the point chosen on the northern system (conduit source control measures at isolated points in the CJ5_55Umb) does not have a detention basin upstream. catchment, is shown in Figure 5.15 and Figure 5.16. This The simulation plots and results confirm the source area represents the Pinetown industrial region of the controls have no effect at higher return periods, whilst catchment. There are two main stormwater systems the combination of source controls and detention basin that manage the runoff in this area, connecting at the continue to reduce peak flows for all return periods. This lower right corner of the image. This region incorporates is illustrated in Figure 5.16. both source controls and detention basins, however Figure 5.13 Hydrographs showing a 2-year rainfall return period flood for Scenarios 9, 11 and 12 at sampling points (a) Umbilo_27, (b) R_ Mkumbaan_01, (c) R_Chats_15 and (d) R_Zana_35. Note that time units are reporting time steps from the model with one time unit representing 5 minutes. PAGE 44 EVALUATING THE POTENTIAL RETURNS TO INVESTING IN GREEN URBAN DEVELOPMENT IN DURBAN EVALUATING THE POTENTIAL RETURNS TO INVESTING IN GREEN URBAN DEVELOPMENT IN DURBAN PAGE 45 5.3.2 Avoided costs due to flood attenuation The reduced flood volumes that would occur under to estimate the capital costs of the structures required different return period flood events (e.g. 1:10 years), under the present versus the GUD scenarios (see Box 5.4, would have required smaller drains, culverts, etc., Appendix 4). The difference, together with associated depending on the size of the event they are designed to differences in maintenance costs, is the total life-cycle deal with. Thus a second model was developed in order cost avoided. Box 5.4 Summary of the method used to estimate infrastructure cost savings (details in Appendix 4) An analysis was undertaken to determine the difference in the replacement cost of stormwater infrastructure (including both conveyance and storage) between the existing stormwater infrastructure in the EMA and infrastructure that would be designed to different specifications depending on the amount of natural vegetation within the EMA. The adopted methodology for this analysis focused on estimating the cost difference of the infrastructure with and without natural/undeveloped areas in the EMA. The relationship between cost and flow for the status quo was established and used to estimate the cost of infrastructure based on changes caused to the flow as a result of having natural areas within the EMA developed into the average land use, i.e. what happens to the flows if the current D’MOSS is replaced with the average land use type within that catchment? The methodology used to determine stormwater infrastructure engineering costs is outlined as follows: 1. Identify all stormwater infrastructure within the study area 2. Divide the infrastructure into four major categories (bridges, canals, culverts and pipes) 3. Determine the dimensions of all infrastructure in the study area Figure 5.15 Aerial screenshot of the Pinetown industrial region for the comparison of scenario 10 (with detention basins) and scenario 11 (detention basins and source controls). The noticeable difference is that the southern conduit system (red lines) flows via a 4. Estimate the costs of the existing infrastructure based on these dimensions detention basin (green box) before joining with the northern stormwater system (yellow lines) on the lower right hand. 5. Assign rainfall design return periods to each category of infrastructure 6. Calculate the maximum open channel flow from the Manning’s equation (i.e. the threshold flow) 7. Estimate a scaling relationship between infrastructure dimensions and flow using theoretical uniform flows 8. Simulate the design rainfall return periods and estimate the maximum design flows for each scenario (status quo versus average land use) 9. Use the maximum flows, the existing infrastructure dimensions and the threshold flows to scale the existing infrastructure to required infrastructure dimensions under each scenario 10. Use the new infrastructure dimensions to estimate the cost of the scaled infrastructure 11. Compare the cost of the existing infrastructure to the cost of the scenario infrastructure Because much of the flood conveyance infrastructure is overdesigned for various reasons including the problems of blockages by litter, the estimation initially yielded a small cost to increase the size of the structures to deal with the difference in flows. Since this is clearly downward biased, a correction was then applied to adjust for this overdesign and produce a more comparable estimate from which to derive the realistic difference in value. Figure 5.16 Simulated peak flows for a 2-, 5-, 10- and 20-year rainfall return period for Scenarios 10 and 11 at points (a) CJ4_56Umb (red line in Figure 5.15 with detention basin and source controls) and (b) CJ5_55Umb (yellow line in Figure 5.15 with only source controls upstream). Note that time units are reporting time steps from the model and one time unit represents 5 minutes. PAGE 46 EVALUATING THE POTENTIAL RETURNS TO INVESTING IN GREEN URBAN DEVELOPMENT IN DURBAN EVALUATING THE POTENTIAL RETURNS TO INVESTING IN GREEN URBAN DEVELOPMENT IN DURBAN PAGE 47 The savings in the capital cost requirements for The conservation of natural areas (Scenario 8), the full flood conveyance were estimated to range from R19 extent of source controls (Scenario 9 and 12) and the million for scenario 6 to R226 million for scenario combination of all GUD measures with conservation 15 when compared to the baseline (Table 5.4). This measures (Scenario 14 and 15) represent the highest represents a 0.5% - 6% capital cost saving in stormwater infrastructure cost savings (Figure 5.17). The total cost infrastructure. Including an estimated 6% of capital costs saving, or flood attenuation value, associated with the as an annual repair and maintenance cost (eThekwini implementation of Scenario 15 is R382 million and for Municipality 2015), this suggests that the cost savings Scenario 8 is R234 million. Therefore the contribution associated with the scenarios have a net present value of large areas of natural vegetation and riparian buffers ranging from R33 – R382 million (Table 5.4). (Scenario 8), or compact development, have a significant impact on reducing peak flows, however the addition The inclusion of 23 detention basins in the subcatchment of structural GUD measures (i.e. full extent source (Scenario 10) provides a cost saving of R194 million, controls and detention basins; Scenario 13, 14 and 15) whereas the implementation of all source controls would provide additional cost savings of R82 million, (Scenario 9) provides a higher cost saving of R278 R102 million, and R146 million, respectively, in terms of million. Soakaways in residential areas were more stormwater infrastructure cost savings (Table 5.4, Figure efficient at attenuating peak flows when compared 5.17). with implementing source controls in commercial and industrial areas only (Scenario 9 compared with Scenario 11 and 12) and therefore, provide a higher cost saving when included (Table 5.4). Table 5.4 Stormwater infrastructure cost savings (R millions) for scenarios 4-15, when compared to the Baseline. Total capital costs Capital cost saving Maintenance Cost (NPV 20 Total cost saving Scenario (R millions) (R millions) yrs., 6%) (R millions) Figure 5.17 Infrastructure capital cost saving (R million) for scenarios 4 – 15 when compared to the Baseline 3. Clean Baseline 3 920 4. Cons2 3 836 84 58 143 5. Cons3 3 859 61 42 104 6. R 3 901 19 13 33 5.4 Amenity benefits of increased conservation areas 7. R + Cons2 3 832 88 61 149 Amenity values included the tourism value generated by current D’MOSS which was used to develop the hedonic 8. R + Cons3 3 781 139 95 234 expenditure by domestic and foreign visitors to Durban, model. Therefore it was assumed that the areas of and the value of green open space areas to residents of natural vegetation for the medium conservation extent 9. S2 3 755 165 113 278 Durban. These values are mutually exclusive and can be were all in a good condition and from this the model 10. D 3 805 115 79 194 added. results for the properties within the catchment could 11. S+D 3 802 118 81 200 be used to calculate the associated premiums. To 12. S2+D 3 733 187 129 316 estimate the premiums that would be associated with 5.4.1 Amenity value to locals the maximum conservation extent, it was assumed 13. S+D+R 3 781 139 96 234 The hedonic analysis that formed part of the that the extra 1000 ha gained from the medium to 14. S2+D+W+R+Cons2 3 717 203 140 343 accompanying study of the value of eThekwini’s maximum conservation area was managed and in a good 15. S2+D+W+R+Cons3 3 694 226 156 382 ecosystem services (Turpie et al. 2017; Letley & Turpie condition. Based on this assumption, a ratio was used 2016) was used to determine the avoided losses of to increase the amount of natural vegetation in a good property value under different scenarios. The hedonic condition surrounding each of the properties included study found that both the amount and condition of in the model. For the minimum conservation extent, natural open space areas in the EMA have significant which is just less than half of the area of the medium positive impacts on property values. Therefore an conservation extent, it was assumed that the amount of increase in the amount of natural vegetation in a good natural vegetation in a good condition surrounding each condition within the catchment would have a positive property would be half that of the natural vegetation impact on property price premiums. Overall property surrounding properties under the medium conservation values were therefore expected to increase with the extent. medium and maximum conservation interventions and The properties in the catchment were grouped by census decrease with the minimum conservation intervention. sub-place. For sub-places that did not fall completely The hedonic model was used to calculate the property within the boundary of the catchment, the number of price premiums attached to the conservation areas properties within the study area was estimated based in the U60F catchment under different scenarios. The on the percentage area of the sub-place within the medium extent of conservation area represents the catchment. The eThekwini Municipality planning scheme documents were used to estimate the increase in the PAGE 48 EVALUATING THE POTENTIAL RETURNS TO INVESTING IN GREEN URBAN DEVELOPMENT IN DURBAN EVALUATING THE POTENTIAL RETURNS TO INVESTING IN GREEN URBAN DEVELOPMENT IN DURBAN PAGE 49 number of properties in the catchment that would occur 5.4.2 Nature-based tourism value Table 5.5 The property premium associated with natural vegetation in a good condition and the nature-based tourism value for minimum, medium and maximum conservation extents in the catchment for each scenario. under the full development scenario. It was assumed The tourism model used geo-tagged Google Earth that the proposed new development areas in the Extent of conservation Total property Annual nature-based NPV nature-based Panoramio photographs to determine the overall catchment have on average one dwelling unit per 650 m2 areas premium tourism value tourism value leisure tourism value in the EMA and formed part of Scenario (min, med, max) (R million) (R million) (R million, 20 yrs., 6%) based on an average taken from the central and outer- the accompanying study of the value of eThekwini’s west scheme planning reports for general and special 3. Clean Baseline ● 887 205 2 351 ecosystem services (Turpie et al. 2017). This approach residential units. The total area of new development in was used to determine the changes in tourism value 4. Cons2 ●● 1 750 439 5 035 the catchment was then divided by 650 m2 to get the under different scenarios in catchment U60F. Using GIS 5. Cons3 ●●● 3 051 512 5 873 estimated number of new properties in the catchment the average R/ha values for natural open space areas 6. R ● 887 205 2 351 to include in the analysis. This equated to 26 000 new in the U60F catchment were multiplied by the area of 7. R + Cons2 ●● 1 750 439 5 035 properties, roughly 36% of the current total in the natural open space within each polygon to get a total 8. R + Cons3 ●●● 3 051 512 5 873 catchment. This percentage increase was then used to nature-based tourism value for the current D’MOSS estimate the average property premium associated with 9. S2 ● 887 205 2 351 in the catchment, representative of the medium natural open space in a good condition for each sub- conservation extent in this study. The percentage change 10. D ● 887 205 2 351 place. in conservation extent from medium to maximum and 11. S+D ● 887 205 2 351 medium to minmum was then used to infer changes in 12. S2+D ● 887 205 2 351 The effect of natural open space in a good condition on tourism value. It was assumed that the change in tourism 13. S+D+R ● 887 205 2 351 property values was obtained from the estimated model value was consistent across the conservation areas, coefficients, which provide the percentage change 14. S2+D+W+R+Cons2 ●● 1 750 439 5 035 irrespective of location in the catchment. In the absence in property value given a unit change in the value of 15. S2+D+W+R+Cons3 ●●● 3 051 512 5 873 of a dedicated choice-expeirment study, this approach natural open space in a good condition, given all other provides a ball-park estimate of the potential differences things being equal. The aggregate effect of open space in the nature-based tourism value associated with in the catchment for minimum, medium and maximum natural systems under different scenarios. 5.5 Avoided climate change costs conservation extent was then estimated by applying The carbon storage analaysis that formed part of Based on the extent of D’MOSS under each of the the regression results to the entire stock of residential The nature-based annual tourism value associated with the accompanying study of the value of eThekwini’s conservation extents, it was estimated that the houses within each sub-place as explained above. natural open space areas in the U60F catchment was ecosystem services (Turpie et al. 2017) was used to damage costs resulting from a loss of the carbon stocks The sub-place premiums were summed to get a total estimated to be R439 million for the medium extent determine the damages avoided of carbon storage under within the Umhlatuzane-Umbilo catchment would be premium for the catchment for each conservation area conservation area (Table 5.5). This value was based different scenarios. The damage costs to South Africa approximately R3.2 million, R2.7 million and R1.3 million extent. on the number of photographic uploads associated resulting from a loss of the carbon stocks within the EMA per annum for the maximum, medium and minimum with natural areas and the amount of natural open was estimated to be approximately R34.3 million per extents of conservation areas. This translates to a net It should be noted that the latter value estimate is only space in the catchment. This annual value increased annum. This equates to a rough estimate of R457 per ha present value of R37 million, R31 million and R15 million, a partial estimate of the amenity value to locals, in that to R512 million under the maximum conservation area of D’MOSS within the EMA. Using the amount of D’MOSS respectively. it only included the value reflected in premiums paid for and decreased to R205 million under the minimum area found within each conservation area (minimum, properties. It does not include values to locals who do extent. This translates into a net present value of R2.35, medium and maximum) the total carbon storage value not live in proximity to open space areas but who make R5.04 and R5.87 billion for the minimum, medium and for each conservation extent could be estimated. use of them. maximum extents of conservation areas, respectively. Based on the hedonic analysis that formed part of We did not estimate the effect of different scenarios on the accompanying study of the value of eThekwini’s the very high tourism value associated with beaches. ecosystem services (Turpie et al. 2017), the total The management of the U60F catchment is not expected premium associated with natural vegetation in a good to have a major impact on these values, but the condition under the maximum conservation extent management of several other eThekwini catchments was estimated to be R3 billion. It was estimated to be would be expected have this added effect. R1.8 billion for the medium conservation extent, and R887 million for the minimum extent. The extent of conservation areas and the total premium associated with these areas for each scenario are shown in Table 5.5. PAGE 50 EVALUATING THE POTENTIAL RETURNS TO INVESTING IN GREEN URBAN DEVELOPMENT IN DURBAN EVALUATING THE POTENTIAL RETURNS TO INVESTING IN GREEN URBAN DEVELOPMENT IN DURBAN PAGE 51 This page intentionally blank. PAGE 52 EVALUATING THE POTENTIAL RETURNS TO INVESTING IN GREEN URBAN DEVELOPMENT IN DURBAN VI. COST BENEFIT ANALYSIS 6.1 Framework use, climate, household incomes and urbanisation (EAI 2006). It is therefore important to incorporate some A cost-benefit analysis was undertaken to evaluate the form of sensitivity analysis so as to adequately assess the potential economic viability of implementing different robustness of the estimates. green urban development measures to address some of the main environmental issues, such as water quality We applied a time frame of 20 years and discount rate and flooding, in the Umhlatuzana-Umbilo catchment. of 6%, and sensitivity analysis using alternative discount rates of 3% and 9%. Cost-benefit analysis is a conceptual framework and tool used to evaluate the viability and desirability of projects and policies based on costs and benefits accumulating over time (Hanley & Spash 1993, Pearce et al. 2006). 6.2 Scenario costs It involves the adjustment of future benefits and costs Determining the overall implementation costs of to their present value equivalent through the process stromwater management interventions is difficult of discounting at a rate which reflects the potential as there are a number of factors that affect costing. rate of return on alternative investments or the pure These factors are discussed in more detail in rate of time preference. For a project to be considered Box 6.1. economically viable, the net present value (NPV) must be positive. This places greater weight on values occurring closer to the present, which means that the Construction costs of the different green future benefits of restoration projects will be down- engineering and conservation measures were weighted compared with the upfront investment costs, obtained from the literature. The unit costs for and therefore have to be relatively larger (in current the source control measures were adjusted based value terms) in order for a project to generate a positive on the information provided in Box 6.1 taking into net present value. Projects can also be evaluated by account the evidence for significantly reduced unit estimating the internal rate of return (IRR), which is the costs as a result of economies of scale. The source discount rate at which the total net present value of the controls implemented in the catchment cover vast project falls to zero. areas; significantly more than the 500 properties The implicit assumption of the above is that the costs and 15 ha described as large scale in other studies and benefits of a project can be determined with (see Box 6.1). It was therefore assumed that the certainty. In reality however, accurately estimating unit costs for green roofs and soakaways, the two all variables in a cost-benefit analysis becomes a interventions implemented at the largest scale in challenge as a result of the way in which estimates are the catchment, would be reduced by up to 80% and assessed and forecast (EAI 2006). Studies are limited by the unit costs for permeable paving and infiltration availability of data and resources, as well as uncertainty trenches would be reduced by 50%. in the consideration of changes in factors such as land EVALUATING THE POTENTIAL RETURNS TO INVESTING IN GREEN URBAN DEVELOPMENT IN DURBAN PAGE 53 Box 6.1 Factors influencing the cost of SUDS municipaliy an average of R50 000 per ha of land can cover an extensive area and the scenarios that include be used (Richard Boon, pers. comm.). For this analysis these interventions therefore have the highest overall A number of factors influence the cost of implementing SUDS (Armitage et al. 2013). These include: the amount of land under conservation for the medium costs, in particular, those scenarios that include the and maximum extents were costed, with the difference full extent of source control implementation (i.e. ƒƒ Project scale and unit costs; ƒƒ Contractor vs. public works crew; between these and the Baseline representing the Scenario 9, 12, 14 and 15). It is the implementation of opportunity cost of conservation. soakaways in all residential areas that exaggerates these ƒƒ New build vs. retrofit; ƒƒ Flexibility in site selection and site suitability; and costs compared to other scenarios. For example, the ƒƒ Regulatory requirements; ƒƒ Levels of experience with the technologies, by both A summary of the GUD construction costs, maintenance difference in GUD construction costs between Scenario designers and contractors (Lampe et al., 2005). costs, conservation management costs, and the 11 and Scenario 12 is R8 billion, the result of soakaways ƒƒ Public vs. private design and construction; opportunity costs of conservation associated with each being included in Scenario 12 but not in Scenario 11. scenario are shown in Table 6.1. The source controls The most important factors are scale and whether the interventions are implemented as part of a new build or are retrofitted. Larger sites offer the opportunity for economies of scale to be realised (Shaffer et al. 2009, Table 6.1 Summary of the construction and maintenance costs (R millions) associated with GUD interventions for each scenario, excluding sanitation costs. Time frame of 20 years and a discount rate of 6% used. Committee on Climate Change 2012). Economies of scale refers to the reduction of per-unit costs through an increase in production volume. The materials and construction costs associated with SUDS such as green roofs GUD measures Conservation areas & riparian buffers and permeable paving decrease considerably over time as a result of increased take up, economies of scale, and Construction Maintenance cost Management costs Full development Opportunity cost Total cost greater competition between suppliers and installers (Shaffer et al. 2009). A review of available case studies scenarios capital cost (NPV 20 yrs., 6%) (NPV 20 yrs., 6%) (R millions) on the costs of SUDS schemes in the U.K. was undertaken by DEFRA (2009), Committee on Climate Change 3. Clean Baseline - - - - - (2012) and McKibbin (2015). These studies confirmed that the unit costs associated with SUDS decrease with development size as well as with higher density developments. In fact, SUDS capital costs decreased by up 4. Cons2 - - 28 160 188 to 80% per property as well as per ha when the scale of the project increased from less than 100 properties 5. Cons3 - - 36 210 246 to more than 500 properties, and from less than 1 ha to more than 15 ha. Shaffer et al. (2009) reported that 6. R - - 16 - 16 the prices of permeable construction materials had fallen considerably over time as a result of improved 7. R + Cons2 - - 44 160 204 technologies and increased demand for SUDS. A cost-benefit analysis of green roofs in urban areas in Helsinki by Nurmi et al. (2013) found that private benefits associated with small scale (property-level) implementation 8. R + Cons3 - - 52 210 262 are in most cases not high enough to justify the investment in green roofs. However, they found that with 9. S2 8 960 3 385 - - 12 344 a higher rate of implementation (at the city wide scale, 50% of all buildings with flat roofs) the unit costs 10. D 22 5 - - 27 decreased by up to 50% and significant public benefits emerged, such as air quality improvements and amenity 11. S+D 928 618 - - 1 546 values. The same study highlighted the importance of implementing incentive or regulation based supportive policies as well as investment in research and development to allow for lower cost green roof installation to 12. S2+D 8 982 3 390 - - 12 372 encourage large scale implementation. 13. S+D+R 764 504 16 - 1 284 14. S2+D+W+R+Cons2 7 031 2 653 44 160 9 888 In addition to economies of scale, the cost of installing SUDS solutions into an existing development (i.e. retrofitting) involves higher costs compared to one designed as part of a new development. A number of 15. S2+D+W+R+Cons3 6 459 2 460 52 210 9 181 DEFRA reports have found that project savings could have been far greater if the SUDS layout had been considered earlier, or upfront, in the development process. It is assumed that for this study the source control interventions were designed and implemented in the catchment as part of new build developments, i.e. it was assumed that when the buildings and parking areas were developed the source controls were part of the design process and were and not retrofitted, thereby reducing unit costs. It should be noted that estimating the costs for such a large scale project is difficult and as a result the costs 6.3 Cost-benefit analysis provided for the stormwater management interventions are not formal estimates, but are preliminary in Costs and benefits of GUD scenarios were examined estimates derived from this study suggest that scenarios order to obtain a ball-park estimate. relative to the baseline scenario of full development that include the implementation of source control with adequate sanitation (Scenario 3), under the measures (Scenarios 9, 11-15) would have a negative assumption that adequate sanitation is both imperative net present value. This was due to the very high costs of The cost of managing riparian buffers and conservation analysis. Opportunity costs of conservation are the and a prerequisite to GUD. Both the total costs and these measures, particularly of soakaways in residential areas was based on conservation management costs costs associated with foregone opportunities to total benefits varied considerably across the different areas (included in the scenarios with full source control taken from James et al. (1999) and Frazee et al. (2003). convert the land into other profitable uses. In the scenarios (Table 6.2, Figure 6.1). Scenarios 4 – 8 and implementation, “SC2”). The only difference between The average annual conservation management cost was eThekwini Municipality, land acquisition is regarded Scenario 10 have positive net present values, with the Scenarios 11 and 12 was the implementation of estimated to be R755/ha (2015 Rands). Based on a total as an important method for securing environmentally benefits significantly outweighing the costs. Net present soakaways in residential areas, with relatively little effect extent of 1900 ha the total annual management costs important areas and has been used successfully in values increased with increasing conservation area from on benefits (Figure 6.1). associated with riparian buffers was R1.4 million. The protecting biodiversity in the EMA (Boon et al. 2016). riparian buffers alone (Scenario 6) to riparian buffers in annual cost of managing the conservation areas was Over the last decade, approximately 619 ha of priority conjunction with the full extent of conservation required Scenario 10 (detention basins only), had a positive net estimated to be R2.1 million for the minimum extent, biodiversity land has been acquired (Boon et al. 2016). to meet biodiversity targets (Scenario 8), with the latter present value of R180 million, due to the relatively low R4.5 million for the medium extent and R5.3 million for The cost of acquiring land varies and is costed based having a net present value of R5.9 billion. costs of implementing this measure. This scenario had the maximum extent. on the amount of developable land (outside of flood a higher net present value than Scenario 6, the result of zones, not on steep slopes) within the land zoning This was because maintaining natural areas had higher infrastructure cost savings (Table 6.2). The opportunity costs of conservation, in the form parcel. However, based on the past experiences of the relatively low costs and high benefits. However, the of land acquisition costs, were also included in the PAGE 54 EVALUATING THE POTENTIAL RETURNS TO INVESTING IN GREEN URBAN DEVELOPMENT IN DURBAN EVALUATING THE POTENTIAL RETURNS TO INVESTING IN GREEN URBAN DEVELOPMENT IN DURBAN PAGE 55 S2+D+W+ R+Cons3 6 459 52 210 2 460 9 181 22 382 40 20 2 163 3 521 6 148 (3 033)   15. improvement, and beach pollution improvement are assumed large enough to offset the costs of litter control and therefore excluding costs and benefits from all * Litter control costs and benefits not included in CBA figures. The co-benefits of litter control which include biodiversity protection, health benefits, amenity S2+D+W+ R+Cons2 7 031 44 160 2 653 9 888 16 343 32 18 863 2 684 3 955 (5 933) 14. 13. S+D+R 764 16 - 504 1 284 - 234 26 16 - - 277 (1 008) 12. S2+D (12 016) 8 982 - - 3 390 12 372 - 316 24 15 - - 355 11. S+D 928 - - 618 1 546 - 200 19 12 - - 231 (1 315) 10. D 22 - - 5 27.5 - 194 7 5 - - 207 179 (12 052) 8 960 - - 3 385 12 344 - 278 9 6 - - 293 9. S2 Figure 6.1 The total present value (R millions) of costs and benefits for all the scenarios relative to the baseline scenario of full development Present value of costs and benefits (R millions) for all scenarios (2015 Rands, 6% discount rate, 20 years) with adequate sanitation (Scenario 3) Cons3 8. R + - 52 210 - 262 22 234 34 18 2 163 3 521 5 992 5 730 The results suggest that the compact development cost-benefit analysis does not take into account a range options that allow for a higher proportion of green open of other potential benefits of the GUD interventions. space are far more effective than using engineering These include improvements in air quality, as well Cons2 7. R + - 44 160 - 204 16 149 26 16 863 2 684 3 754 3 550 measures alone. This is because the open space areas as intangible benefits such as the existence value of not only deliver ecosystem services relating to the biodiversity. primary stormwater management objectives but also directly provide amenity value that is translated into To provide some perspective, the initial capital - 16 - - 16 - 32.8 20.6 13.2 - - 67 50 6. R property values and tourism values. It is acknowledged, requirement associated with each scenario (Figure 6.1, nevertheless, that the latter benefits were more Table 6.3) was compared to the eThekwini Municipality difficult to estimate than other benefits such as the capital budget of R6.73 billion for 2016/17 (eThekwini 5. Cons3 Municipality 2015). Scenarios 4 – 8 and scenario 10 - 36 210 - 246 22 104 13 9 2 163 3 521 5 832 5 586 engineering cost savings, and therefore have some degree of uncertainty. The most uncertain estimates have a funding requirement that is less than 1% of were those of tourism value, for which we have used the municipal capital budget. Scenarios 11 and 12 average values in the absence of reliable estimates of would require 13.8% and 11.4% of the capital budget 4. Cons2 - 188 160 - 348 16 143 7 5 863 2 684 3 718 3 370 marginal changes in value associated with changes in respectively and scenario 15 would require 96.1% of scenarios does not bias CBA outcomes. the green-ness of Durban. Nevertheless, our estimates the budget. The requirements for scenarios 9, 13, and could be considered as conservative. Firstly, the tourism/ 14 are all higher than the proposed capital budget for recreational benefits associated with conservation areas the municipality, however, it is expected that capital Stormwater infrastructure savings are likely to understate future values of these benefits, requirements for large-scale implementation would Conservation management cost Total present value of benefits which are expected to increase as the city develops occur over several years (or decades) as an area Nature-based tourism value Benefits relative to Baseline Total present value of costs Estuary services (fisheries) further, becomes larger and incomes rise. Secondly, the develops. Costs relative to Baseline Total Maintenance Cost Dredging cost savings Property premiums   Net Present Value Construction cost Opportunity cost Carbon storage Table 6.2 PAGE 56 EVALUATING THE POTENTIAL RETURNS TO INVESTING IN GREEN URBAN DEVELOPMENT IN DURBAN EVALUATING THE POTENTIAL RETURNS TO INVESTING IN GREEN URBAN DEVELOPMENT IN DURBAN PAGE 57 Table 6.3 Total present value of costs and benefits and NPV (R millions) for all scenarios (2015 Rands, 6% discount rate, 20 years), initial capital requirement and the capital requirement as a percentage of the annual capital budget for each scenario. Total present Total present Initial capital value of costs value of benefits NPV requirement % of EM capital Scenario (R millions) (R millions) (R millions) (R millions) budget 2016/17 4. Cons2 323 3 718 3 394 2 0.0% 5. Cons3 246 5 832 5 586 3 0.0% 6. R 16 67 50 1 0.0% 7. R + Cons2 204 3 754 3 550 4 0.1% 8. R + Cons3 262 5 992 5 730 5 0.1% 9. S2 12 344 293 -12 052 8 960 133.1% 10. D 27 207 179 22 0.3% 11. S+D 1 546 231 - 1 315 928 13.8% 12. S2+D 12 372 355 - 12 016 8 982 133.5% 13. S+D+R 1 284 277 - 1 008 765 11.4% 14. S2+D+W+R+Cons2 9 888 3 955 - 5 933 7 035 104.5% 15. S2+D+W+R+Cons3 9 181 6 148 -3 033 6 464 96.0% 6.4 Sensitivity analysis The results were further tested under varying discount if the unit cost of soakaways was reduced by 8% as a rates and assumptions of costs and benefits for one result of economies of scale, competition and improved GUD intervention. Costs and benefits were varied for technologies then the NPV for Scenario 15 would be soakaway implementation as this intervention had the positive at R7 million. However, the NPV associated with largest influence on NPV. Alternative discount rates of other source control scenarios would remain negative 3% and 9% were applied to the NPV calculations. The (Table 6.4). alternative discount rates did not have a significant impact on NPV calculations, with the NPV remaining The table results suggest that the implementation negative at discount rates of 3% and 9% for Scenarios 9 of source controls is only feasible if incorporated in and 11-15 and positive for Scenarios 4-8 and Scenario conjunction with conservation areas, riparian buffers 10. Again, this highlights the significant influence of and regional controls (detention basins and treatment soakaway implementation costs on the overall NPV for wetlands) to maximise benefits associated with these these scenarios. measures. Detention basins alone have a positive NPV, as do riparian buffers. This result is encouraging. The Comparing Scenarios 11 and 12 provides an indication sensitivity analysis has shown that a more conservative into the costs and benefits associated with soakaways, approach to source control implementation is needed, as the only difference between the two is the especially with regards to soakaways in residential areas. implementation of soakaways in residential areas. The difference in benefits between these two scenarios is The main findings from the CBA suggest that R124 million, with R116 million of this being stormwater source control measures are costly, particularly the infrastructure cost savings. Soakaways represent 80% of implementation of soakaways in residential areas. the cost for Scenario 15 and 89% of the cost for Scenario Detention basins and treatment wetlands are the 14 but contribute approximately R124 million in benefits. most affordable engineering measures, with positive This suggests that if soakaways were not implemented net benefits. Retaining natural areas in the catchment but all other source controls were still included then the has the highest net benefits, due to the significant co- NPV for Scenario 15 would be approximately R4.6 billion benefits associated with amenity values. and for Scenario 14 would be R2.4 billion (Table 6.4). Or Table 6.4 The impact of soakaways on NPV for Scenarios 9, 12, 14 and 15 (R millions, 6%, 20 yrs.) 9. 12. 14. 15. Net Present Value (R millions) S2 S2+D S2+D+W+R+Cons2 S2+D+W+R+Cons3 NPV Base Case -12 052 -12 016 -5 933 -3 033 NPV Soakaway unit costs reduced by a further 8% -7 721 -7 686 -2 594 7 NPV Soakaways not included -1 226 -1 190 2 415 4 568 PAGE 58 EVALUATING THE POTENTIAL RETURNS TO INVESTING IN GREEN URBAN DEVELOPMENT IN DURBAN VII. CONCLUSIONS AND POLICY RECOMMENDATIONS Under a business as usual scenario, the continued with using either natural interventions (particularly growth of urban areas in Africa will result in a further river buffers) or green engineering interventions (source deterioration of the natural environment and living controls and detention basins), and these appear to conditions, a loss of values associated with green open be substitutable to a large degree when the former is space areas, and increased costs in reducing risks to brought about through more compact development. people that result from environmental problems. The Simillarly, both natural and green engineering notion of Green Urban Development is therefore highly interventions were highly effective at reducing flood attractive, as it allows citites to grow in a way that peaks, and were also substitutable to a large degree in maintains their resilience and maintains standards of terms of this function. This suggests that some sensible living and quality of life. However, few studies have combination could be applied, with green engineering investigated what following a more sustainable green interventions being strategically located within the urban development path will actually cost, and whether catchment for maximum overall effect. these costs can be justified. Moreover, what Green Urban Development should look like is also not well However, not all green engineering interventions are defined, in terms of the degree to which is includes the equally viable. Our estimates suggested that the conservation of river buffers and other natural areas, large-scale application of low-impact stormwater the mimicking of natural processes through innovative management measures (i.e. source controls) is extremely engineering design or the protection of downstream costly at today’s prices, even when accounting for areas through conventional engineering measures. economies of scale. WOn the other hand, the simpler In eThekwini Municipality, these issues have to be green engineering measures, such as detention basins considered as plans for the growth of Durban are laid for reducing peak flows and treatment wetlands for out. In this study, we tested the idea of green urban improving water quality, were shown to be highly cost- development by backcasting a range of scenarios for a effective, viable interventions. well-developed catchment. The retention of signifant natural areas within This study attempted to analyse a highly complex the catchment, which may require more compact problem in a fairly large catchment area, and as such, development, was not only found to be highly effective a great deal more work will still be needed in order at reducing sedimentation and flooding problems, to narrow the potential error margin. Our study but has the added benefit of yielding high amenity found a lack of empirical studies to inform modelling value realised as property and tourism value as well assumptions, which suggests that much benefit could as intangible and unknown values associated with be obtained from the implementation and monitoring maintaining biodiversity and ecological connectivity. of pilot programmes. Furthermore, there are few well- Although riparian buffers have limited influence on developed modelling platforms that are capable of water quality and flooding in urban environments, the estimating the impacts of these kinds of interventions value in maintaining biodiversity is also very high, as at scale. Most previous studies have analysed these they are e critical for providing connectivity between problems at a micro-catchment scale, making it difficult terrestrial systems, rivers and estuaries. to assess the economic implications of a change in policy. Thus, while these are first-cut estimates which Because conservation with compact development incurs warrant further refinement, they provide a useful very low costs in comparison to other interventions, step towards the informing policies to guide the city’s the net benefits of this strategy far outweighed any sustainable development path. other. Compact development coupled with the other interventions creates the greenest city, in terms of The results showed that in a catchment with little water quality and biodiversity conservation goals, and unserviced informal settlement (1% of the area) recycling is an economically justifiable strategy in terms of overall the equivalent of a third of sewage outputs along with costs and benefits.  Maintaining large natural areas and complete sanitation services has the most significant riparian buffers should therefore be a primary strategy, impact on nutrient loads entering the rivers and along with the strategic positioning of cost-effective harbour, and that treatment wetlands make a significant green engineering measures. further impact. Sediment loads can be effectively dealt EVALUATING THE POTENTIAL RETURNS TO INVESTING IN GREEN URBAN DEVELOPMENT IN DURBAN PAGE 59 This page intentionally blank. PAGE 60 EVALUATING THE POTENTIAL RETURNS TO INVESTING IN GREEN URBAN DEVELOPMENT IN DURBAN VIII. REFERENCES Armitage, N., Vice, M., Fisher-Jeffes, L., Winter, K., Department of Water Affairs and Forestry, South Africa. Spiegel, A., & Dunstan, J. 2013. Alternative Technology 2002. South African National Water Quality Monitoring for Stormwater Management. The South African Programmes Series. National Eutrophication Monitoring Guidelines for Sustainable Drainage Systems. WRC Programme. Implementation Manual. Department of Report No. TT 558/13. Water Affairs and Forestry. Pretoria. ASCE, 1992. Design and Construction of Urban Department of Water Affairs and Forestry, South Africa. Stormwater Management Systems, New York, USA. 2008. Methods for determining the Water Quality component of the Ecological Reserve. Second Draft. Bergmann, M., Gutow, L. and Klages, M. (eds). 2015. Prepared by P-A Scherman of Scherman Consulting. Marine Anthropogenic Litter. Springer International Publishing, Cham. Department of Water Affairs and Forestry. 1996. South African Water Quality Guidelines (second edition), Bester, A.J., Jacobs, H. E., van der Merwe, J. & Fuamba, Volume 7: Aquatic Ecosystems. M. 2010. Unit cost-functions for value estimation of waterborne sewer infrastructure. WISA 2010 Conference Department of Water Affairs and Forestry. 1999. Proceedings. Resource Directed Measures for Protection of Water Resources. Volume 3: River Ecosystems Version 1.0, Campbell, L.A. 2001. A Study on the Fate of Urban/ Pretoria. Resource Directed Measures for Protection of Stormwater Runoff from Alexandra Township in the Water Resources, Pretoria, South Africa. Jukskei River. WRC Report No. 598/3/01. Water Research Commission, Pretoria, South Africa. Department of Water Affairs (DWA), 2013. Classification of Water Resources and Determination of the Cerfonteyn, M. and Day, E. (Liz). 2010. Diep River Water Comprehensive Reserve and Resource Quality Objectives Quality Study. Draft report to the City of Cape Town. in the Mvoti to Umzimkulu Water Management Area: Status quo assessment, IUA delineation and biophysical Construction Industry Research and Information node identification. Prepared by: Rivers for Africa eFlows Association (CIRIA). 2007. The SUDS Manual. Report Consulting (Pty) Ltd. C697. Environmental Planning and Climate Protection CSIR 2008. Sand supply from rivers within the eThekwini Department (EPCPD) 2012. eThekwini Municipality jurisdiction: impllications for coastal sand budgets Systematic Conservation Plan. eThekwini Municipality. and resource economics. Theron, A. K., de Lange, W., Nahman, A. & Hardwick, D. (eds). Report published by Environment Agency (EA) 2015. Cost estimation for SUDS CSIR. Report No: CSIR/NRE/ECO/ER/2008/0096/C – summary of evidence. Report SC080039/R9. Bristol, United Kingdom. De Committee on Climate Change 2012. Costs and Benefits of Sustainable Drainage Systems. Economics eThekwini Municipality 2008. Design Manual: Guidelines of flood risk adaptation measures. Haskoning UK Ltd, And Policy For The Design Of Stormwater Drainage Water. United Kingdom. And Stormwater Management Systems. Engineering Unit, Coastal Stormwater and Catchment Management Department for Environment, Food & Rural Affairs Department. (DEFRA). 2009. Impact Assessment - Local Flood Risk Management and the increased use of Sustainable eThekwini Municipality 2014. State of Biodiversity Drainage Systems. London: Defra. Report 2011/2012. Compiled by the Environmental Planning and Climate Protection Department. De Martino, G., De Paola, F., Fontana, N., Marini, G., & Ranucci, A. 2012. Experimental assessment of level Environmental Planning and Climate Protection pool routing in preliminary design of floodplain storage. Department (EPCPD) 2012. eThekwini Municipality Science of the Total Environment, 416, 142-147. Systematic Conservation Plan. eThekwini Municipality. Department of Environmental Protection. 2010. Fletcher, T., Duncan, H., Lloyd, S, & Poelsma, P. 2003. Riparian Forest Buffer Guidance. Bureau of Watershed Stormwater Flow and Quality and the Effectiveness Management, the Pennsylvania Clean Streams Law. of Non-proprietary Stormwater Treatment Measures. Document 394-5600-00. Draft report for the NSW EPA. April 2003. Cooperative Research Centre for Catchment Hydrology, Melbourne, Victoria. EVALUATING THE POTENTIAL RETURNS TO INVESTING IN GREEN URBAN DEVELOPMENT IN DURBAN PAGE 61 Fletcher, T.D., Shuster, W.F., Hunt, W.F., Ashley, R., Butler, Kleynhans, C., Louw, M., Thirion, C., Rossouw, N. and Morales Torres, A., Perales Momparler, S., Jefferies, Ryan, P.G. 1990. The marine plastic debris problem off D., Arthur, S., Trowsdale, S., Barraud, S., Semadeni- Rowntree, K. 2005. River ecoclassification. Manual for C. & Domenech, I.A. 2015. Report on Stormwater southern Africa: types of debris, their environmental Davies, A., Bertrand-Krajewski, J-L., Mikkelsen, P.S., Ecostatus Determination. Report to the Water Research Management. E2STORMED PROJECT, Improvement effects, and control measures. Proceedings of the Rivard, G., Uhl, M., Dagenais, D. & Viklander, M. 2015. Commission. of energy efficiency in the water cycle by the use Second International Conference on Marine Debris, SUDS, LID, BMPs, WSUD and more – The evolution of innovative storm water management in smart NOAA Technical Memorandum, NMFS, SWFC 154: 85- and application of terminology surrounding urban Kundzewicz, Z. W. 2002. Non-structural flood protection Mediterranean cities, www.e2stormed.eu 102. drainage. Urban Water Journal 12:7. 525-542, DOI: and sustainability. Water International, 27(1), 3-13. 10.1080/1573062X.2014.916314. Msadala, V., Gibson, L., Le Roux, J., Rooseboom, A & Severn Trent Water, 2015. Severn Trent Environmental Lampe, L., Barrett, M., Woods-Ballard, B., Andrews, A., Basson, GR. 2010. Sediment yield prediction for South Protection Scheme (STEPS). United Kingdom Food and Agricultural Organisation of the United Nations Martin, P., & Kellagher, R. 2005. Performance and Whole Africa: 2010 Edition. WRC Report No. 1765/1/10 (FAO). 2011. Forests beneath the grass. Edited by Durst, Life Costs of Best Management Practices and Sustainable Shaffer, P., Wilson, S., Brindle, F. & Baffoe-Bonnie, B. P.B., Sajise, P. & Leslie, R.N. Proceedings of the Regional Urban Drainage Systems. Water Environment The South Natural Resources Conservation Service (NRCS). 2005. 2009. Understanding permeable and impermeable Workshop on Advancing the Application of Assisted African SUDS Guidelines Research Foundation Project Economics and Survival of Hand-Planted Riparian Forest surfaces. Technical report on surfacing options and cost Natural Regeneration for Effective Low-Cost Restoration. 01-CTS-21T, Water Environment Research Foundation. Buffers in West Central Maine. Report ME-TP-2005-1, benefit analysis. Department for Communities and Local Illinois. Government, London, United Kingdom Frazee, S.R., Cowling, R.M. Pressey, R.L., Turpie, J.K. Letley, G. & Turpie, J.K. 2016. The contributions of & Lindenberg, D. Estimating the costs of conserving a natural and man-made green open space to property Natural Water Retention Measures (NWRM) Project. Scotland and Northern Ireland Forum For Environmental biodiversity hotspot: a case-study of the Cape Floristic value in Durban, South Africa: a hedonic pricing study. In 2013. Individual NWRM: Infiltration trenches. http:// Research (SNIFFER) 2007. Retrofitting Sustainable Urban Region, South Africa. Biological Conservation 112: 275– prep. www.nwrm.eu Water Solutions. Project UE3(05)UW5. 290. Liu, Y., Bralts, V.F. & Engel, B.A. 2015. Evaluating the Nurmi, V., Votsis, A., Perrels, A. & Lehvavirta, S. 2013. Smithers, J.C., 2002. The Estimation for Design Georgia State. 2001. Georgia Stormwater Management effectiveness of management practices on hydrology Cost-benefit analysis of green roofs in urban areas: Rainfall for Durban Unicity. Draft Report to eThekwini Manual. Volume 2: Technical Handbook. Report and water quality at watershed scale with a rainfall- case study in Helsinki. Finnish Meteorological Institute. Municipality. School of Bioresources Engineering prepared by AMEC Earth and Environmental, Center runoff model. Science of the Total Environment 511: Report 2013:2. and Environmental Hydrology. University of Natal, for Watershed Protection; Debo and Associates; Jordan 298–308. Pietermaritzburg, South Africa. Jones and Goulding; Atlanta Regional Commission. Opperman, J.J. 2014. A Flood of Benefits: Using Green http://www.georgiastormwater.com. MacFarlane, DM, Bredin, IP, Adams, J., Zungu, MM, Bate, Infrastructure to Reduce Flood Risks. The Nature Smithers, J.C. & Schulze, R.E, 2002. Design Rainfall and G.C., and Dickens, C.W.S. 2014. Preliminary guidelines Conservancy, Arlington, Virginia. Flood Estimation in South Africa. School of Bioresources Giugni, M., De Paola, F., Topa M.E. 2012. Guidelines on for the determination of buffer zones for rivers, wetlands Engineering and Environmental Hydrology, University of engineering design and management of storm water and estuaries. Final Consolidated Water Research Ouyang, Y., Leininger, T.D. & Moran, M. 2013. Impacts of Natal, Pietermaritzburg, South Africa. WRC Project No: systems. CLUVA project deliverable D2.2. Available at Commission Report. WRC Report No TT610/14. Water reforestation upon sediment load and water outflow in K5/1060 http://www.cluva.eu/deliverables/CLUVA_D2.2.pdf. Research Commission. Pretoria. the Lower Yazoo River Watershed, Mississippi. Ecological Accessed 1 March 2016. Engineering 61: 394– 406. Spatial Development Framework (SDF) 2014. eThekwini McKibbin, D. 2015. Sustainable drainage system Municipality. Hull City Council, 2015. Hull City Council Local Flood Risk provisions within the Water and Sewerage Services Bill. Rawls, W.J., Brakensiek, D. L. & Miller, N. 1983. Green- Management Strategy. United Kingdom Research and Information Service Bill Paper. Northern Ampt infiltration parameters from soils data. Journal of Statistics South Africa, 2005. South African Statistics, Ireland Assembly. Report NIAR 279-15. Hydraulic Engineering 109: 62-70 2004/05. Published by Statistics South Africa, Pretoria Integrated Development Plan (IDP), 2014. eThekwini 0001. Municipality Draft Integrated Development Plan: Five McPherson, G., Simpson, J.R., Peper, P.J. & Xiao, Q. 1999. Rieck, C. & von Muench, E. 2011. Technology review year plan 2012/13 to 2016/17. Annual Review 2014/2015. Benefit-Cost analysis of Modesto’s Municipal urban of urine diversion dehydration toilets (UDDTs): Design Stovin, V.R. & Swan, A.D. 2007. Retrofit SUDS – cost forest. Journal of Arboriculture 25(5): 235-248. principles, urine and faeces management. Sustainable estimates and decision support tools. Proceedings of the James, A.N., Green, M.J.B. & Paine, J.R. 1999. A Global sanitation - ecosan program, GIZ. Institution of Civil Engineers, Water Management 160, Review of Protected Area Budgets and Staffing. WCMC – Michie, R. 2010. Cost Estimate to Restore Riparian Forest Issue WM4. World Conservation Press, Cambridge, UK. vi + 46pp. Buffers and Improve Stream Habitat in the Willamette Roberts, D. & O’Donoghue, S. 2013. Urban Basin, Oregon. State of Oregon, Department of environmental challenges and climate change action SWWM, 2009. Stormwater Management Model Jiang, Y., Yuan, Y. & Piza, H. 2015. A review of Environmental Quality, Watershed Management Section. in Durban, South Africa. Environment and Urbanisation Applications Manual. Report: EPA/600/R-03/000. United applicability and effectiveness of low impact Report available at: http://www.deq.state.or.us/wq/ 25(2): 299 – 319. States. development/green infrastructure practices in arid/ tmdls/docs/WillametteRipCost030310.pdf. Accessed 01 semi-arid United States. Environments 2: 221–249; March 2016. Rooseboom, A., Verster, E., Zietsman, H.L. & Lotriet, TEEB, 2010. The Economics of Ecosystems and doi:10.3390/environments2020221 H.H., 1992: The development of the new Sediment Yield Biodiversity Ecological and Economic Foundations. Moodley, K. 2014. An investigation into the pollution Map of South Africa. WRC report no 297/2/92, report for Edited by Pushpam Kumar. Earthscan, London and Joksimovic, D. & Alam, Z. 2014. Cost efficiency of low status of the Durban Harbour river catchments, Water Research Commission by Sigma Beta, Pretoria. Washington. impact development (LID) stormwater management kwaZulu-Natal, South Africa. MSc Thesis for the School practices. Procedia Engineering 89: 734 – 741. of Agricultural, Earth and Environmental Sciences, Rutherford, I., Anderson, B. & Ladson, A. 2007. Managing Thampapillai, D.J., & Musgrave, W.F. 1985. Flood damage University of KwaZulu-Natal the effects of riparian vegetation on flooding. In mitigation: a review of structural and nonstructural Kadlec, R.H. and Knight, R.L. 1996. Treatment wetlands. Lovett, S. & Price, P. (eds); Principles for riparian lands measures and alternative decision frameworks. Water CRC Press LLC. Florida. management. Land & Water Australia, Canberra. Resources Research, 21(4), 411-424. PAGE 62 EVALUATING THE POTENTIAL RETURNS TO INVESTING IN GREEN URBAN DEVELOPMENT IN DURBAN EVALUATING THE POTENTIAL RETURNS TO INVESTING IN GREEN URBAN DEVELOPMENT IN DURBAN PAGE 63 Topa, M. E., Giugni, M., & De Paola, F. 2014. Off- UN-HABITAT. 2008. Constructed Wetlands Manual. stream floodplain storage: Numerical modeling and UN-HABITAT Water for Asian Cities Programme Nepal, experimental analysis. Journal of Irrigation and Drainage Kathmandu. Engineering, 141(1), 04014040. United States Environmental Protection Agency (US EPA). Toronto and Region Conservation Authority (TRCA) 1999. National Best Management Practices for Storm 2013. Assessment of Life Cycle Costs for Low Impact Water, Phase II. United States Environmental Protection Development Stormwater Management Practices. Agency online guideline. Cited in Taylor (2005). Sustainable Technologies Evaluation Program. Van Niekerk, L. & Turpie, J.K. 2012. South African TNC, 2015. Upper Tana-Nairobi Water Fund: A Business National Biodiversity Assessment 2011: Technical Case. Available at: http://www. Report. Volume 3: Estuary Component. CSIR Report nature.org/ourinitiatives/regions/africa/upper-tana- Number CSIR/NRE/ECOS/ER/2011/0045. Council for nairobi-water-fund-business-case.pdf. Accessed 1 March Scientific and Industrial Research, Stellenbosch. 2012. Waidler, D., White, M., Steglich, E. Wang, S. Williams, J., Turpie, JK., Adams, J.B., Joubert, A., Harrison, T.D., Jones, C.A. & Srinivasan, R. 2009. Conservation practice Colloty, B.M., Maree, R.C., Whitfield, A.K., Wooldridge, modeling guide for SWAT and APEX. Draft - August 2009, T.H., Lamberth, S.J., Taljaard, S. & van Niekerk, L. 2002 provided by Spatial Sciences Laboratory, Texas A&M Assessment of the conservation priority status of South University, College Station, Texas African estuaries for use in management and water allocation. Water SA. 28: 191-206. World Bank 2015. Promoting Green Urban Development in African cities eThekwini, South Africa. Urban Turpie, J.K., Day, L., Gelo Kutela, D., Letley, G. & Forsythe, Environmental Profile. Washington. K. 2016. A preliminary investigation of the potential costs and benefits of rehabilitation of the Nakivubo Xiao, Q. & McPherson, G. 2002. Rainfall interception by wetland, Kampala. World Bank Report. Santa Monica’s municipal forest. Urban Ecosystems 6: 291-302. Turpie, J.K., Letley, G., Chrystal, R. & Corbello, S. 2017. The value of Durban’s natural capital and its role in Green Urban Development. Part 1: A spatial valuation of the natural and semi-natural open space areas in eThekwini Municipality. World Bank Report. PAGE 64 EVALUATING THE POTENTIAL RETURNS TO INVESTING IN GREEN URBAN DEVELOPMENT IN DURBAN APPENDIX 1: URBAN STORMWATER MANAGEMENT OPTIONS A1.1 Overview management measures now tend to be designed to address both flooding and water quality problems, with Within urban areas, rain that falls onto impermeable many measures addressing both of these. Both WSUD surfaces such as streets, parking lots, pavements and and SUDS approaches are increasingly being applied roofs is unable to filtrate into the ground as it would in development planning in South Africa, and their normally do in undeveloped areas, leading to higher inclusion is mandatory for new developments in the City levels of surface runoff during storm events than of Cape Town and in Durban. The results from this study would have happened naturally and creating flooding provide the opportunity to understand more about problems in downstream areas (Armitage et al. 2013). the impact that GUD interventions can have on future This problem is exacerbated by the fact that urban development in Durban, adding to the already relevant stormwater runoff generally contains litter, debris, and important policies that are in place. and sediments which lead to blockages of the systems designed to convey water, and they also contain bacteria, heavy metals and nutrients, which means that floodwaters can become a pollution hazard. All A1.2 Passive engineering measures to of this can have negative impacts on property, urban improve conveyance infrastructure and natural habitats, as well as urban These measures are designed to protect areas from inhabitants. With an increase in urbanisation worldwide flooding by avoiding or mitigating the water flow off and the associated impact of increasing stormwater stream over the riverbanks, or accommodating the runoff on aquatic ecosystems, the management of urban flood adjusting the riverbed carrying out channel drainage has become a critically important challenge improvement. Therefore, these kinds of measures try (Fletcher et al. 2015). to constrain the inundation without modification of the hydrograph. Examples are levees, cleaning from debris Urban drainage management has changed significantly or increasing of section of the riverbed, and hydraulic over the last few decades, from a conventional bypass, also known as waterways. They involve physical ‘rapid disposal’ approach to a more integrated and construction to reduce or avoid possible impacts of sustainable ‘design with nature’ approach (Fletcher et hazards, or application of engineering techniques to al. 2015). The early traditional attitudes towards urban achieve hazard-resistance and resilience in structures or drainage management focused on trying to dispose systems. These kind of measures alter the streamflow of stormwater in the fastest way possible with not of rivers and channels, resulting in the reduction of the much consideration for surrounding ecosystems or frequency and severity of floods. For example, reservoirs for downstream water quality impacts. In the 1980s reduce peak flows; levees and flood walls confine flows and 1990s a new focus on urban stormwater runoff within predetermined channels; improvements to and water quality developed around the world, which channels reduce the peak stages; and flood ways help concentrated on a more catchment-wide management divert excess flow. and restoration approach to urban drainage as opposed to the standard end-of-catchment solution (Fletcher et al. 2015). This embodied a more holistic approach A1.2.1 Drains and swales which treats stormwater runoff problems at source and These convey flows from built-up areas via small minimises environmental degradation, while delivering channels, and can generally deal with small floods of 1-2 environmental, economic and social benefits. This year return period. rapid development in the field of urban drainage saw a number of terms being used to define similar concepts (Fletcher et al. 2015). Terms such as “Water Sensitive A1.2.2 Enlargement of river channel/canalisation/levees/ Urban Design” (WSUD), “Low Impact Development ” dredging (LID), “Sustainable Urban Drainage Systems” (SUDS), “Integrated Urban Water Management ” (IUWM) Excavation of a river channel involves either deepening and “Best Management Practices” (BMPs) were first or widening the channel to increase flood control used by professionals in countries such as Australia, capacity. A river can be made to carry larger discharges New Zealand, United States, England and Scotland to by improving the hydraulic condition of the channel describe this new approach to urban drainage and are through measures such as dredging. Similarly, levees all essentially synonymous. “Green infrastructure” is (embankments) can be built to increase the conveyance another term commonly used in this context, and tends capacity of the channel. to refer to any environmentally-friendly stormwater management structures, natural and semi-natural systems used in stormwater management. Storm water EVALUATING THE POTENTIAL RETURNS TO INVESTING IN GREEN URBAN DEVELOPMENT IN DURBAN PAGE 65 Levees are generally built as an embankment (i.e. earthwork). In the urban context, if there is not enough land area to build such earth structures, then they are constructed with reinforced concrete or masonry walls. The levees location is designed according to the inundation analyses; their scope is to prevent the inundation of floodplain. Their height is designed to prevent the inundation associated to a specific return period. Once the height is defined, their design will follow geotechnical rules if they are made with earth, or structural rules if they are concrete walls. To analyse Figure A1.1 Schematic representation of levees at two side of the watercourse. the efficiency of such structures, they are modelled in the hydraulic routine as a modification of the digital Figure A1.3 Permeable paving allows water to soak into the gravel sub-base, temporarily holding the water before it soaks into the ground, or elevation model. passes to an outfall (Source: susdrain, www.susdrain.org) A1.2.3 Hydraulic bypass areas (Armitage et al. 2013). They create temporary Advantages A hydraulic bypass is a new channel built to laminate subsurface storage of stormwater runoff thereby the peak discharge crossing the floodable area in the • Significantly reduce stormwater discharge rates and volumes enhancing the natural capacity of the ground to store urban context. The new channel takes part of the from impervious areas • Reduce peak flows to watercourses reducing the risk of flooding and drain water. Infiltration trenches allow water to discharge and brings it to the final destination through downstream and reduce the effects of pollution in runoff on the infiltrate into the surrounding soils from the bottoms an alternative path. Construction of a hydraulic bypass environment and sides of the trench. They usually have a rectangular is very expensive and requires the identification of the • Flexible and tailored solution that can suit the proposed usage Figure A1.2 Schematic representation of a hydraulic bypass. and design life vertical cross-section and are designed to receive alternative path for the new channel. • Allows for dual use of space, so there is no additional land take. stormwater runoff from adjacent properties and They increase the ‘usable’ area by utilising roadways, driveways transportation links such as asphalt roads and footpaths and parking lots as stormwater drainage areas • Good community acceptability (Armitage et al. 2013). The amount of water that can be A1.3 Active engineering measures to retard and allows stormwater runoff to infiltrate into the • Stormwater runoff that is stored can be used to recharge disposed of by an infiltration trench within a specified runoff substratum, ultimately promoting the recharge of the the groundwater table and also be used for several domestic time is dependent on the infiltration potential of the groundwater table. The stored rainwater can also be purposes surrounding soil, the size of the trench, and the bulk The active structural measures are to modify the reused for a number of purposes such as watering • Lined permeable pavement systems can be utilised where hydrograph by reducing and delaying the maximum peak foundation or soil conditions limit infiltration processes density of the fill material. Stormwater runoff is treated gardens and lawns (Armitage et al. 2013). by physical filtration to remove solids, adsorption onto discharge. Examples include floodplain storage (in-line Limitations or off-line) that stores the flood volume temporarily the material in the trench and biochemical reactions Permeable pavements generally do not remove litter in an adequate upstream capacity, leading to flood • Cannot be used where large sediment loads may be washed or involving micro-organisms in the soil. and other debris from stormwater runoff as it tends to carried onto the surface of the paving attenuation as a result of the discharge being gradually remain on the surface as the water infiltrates. Soluble • The implementation is generally limited to sites with slopes less released (Topa et al. 2014). When the discharge falls pollutants, however, do pass through the permeable than 5% Advantages below the maximum allowable flow, the flood volume is layer and surfaces that have an aggregate sub-base can • Risk of long-term clogging and weed growth if poorly maintained • Increases stormwater infiltration and corresponding groundwater • Not normally suitable for high traffic volumes and speeds greater released back to the river (De Martino et al. 2012). Off- provide good water quality treatment. Permeable paving than about 50 km/hr, or for usage by heavy vehicles and/or high recharge stream floodplain storages are often used since they do can be used in a variety of locations, such as parking lots, point loads • Decrease the frequency and extent of flooding • Effective in removing suspended particulates from stormwater not interfere with the natural drainage pattern between private and public roads, industrial storage and loading • The pollutant removal ability of permeable pavements is lower • Due to their relatively narrow cross section, they can be utilised than most other SuDS options. the stream and the floodplain, and only an outlet areas, bike lanes, walkways and terraces (Armitage et in most urban areas structure is needed to regulate the outflow discharge. al. 2013). The use of this paving is however restricted to • Negligible visual impact as they are generally below ground Permeable pavements are relatively expensive to slopes that are less than 5%, or ideally flat, as the high Limitations construct and can have high maintenance costs. velocity stormwater from steep slopes does not have A1.3.1 Permeable pavements However, they are incredibly efficient at reducing peak • If situated in coarse soil strata, groundwater contamination is a sufficient time to infiltrate before being washed away. flows and reducing runoff volume as well as reducing possibility Permeable pavements refer to pavements that are • Restricted to areas with permeable soils To ensure the long term effectiveness of permeable pollutants. They remove approximately 60-95% of TSS, constructed in such a way that they promote the • Not appropriate on unstable or uneven land, or on steep slopes pavements, regular inspections and maintenance are 70-90% of hydrocarbons, 50-80% of total phosphorous, • Prone to failure if sediment, debris and/or other pollutants are infiltration of stormwater runoff through the surface recommended. Blockage of the fine stone aggregate can 65-80% of total nitrogen and 60-95% of heavy metals able to clog the gravel surface and/or into the sub-layers or underlying substrata (Armitage • backfilled aggregate material sometimes be an issue and requires cleaning or replacing (Armitage et al. 2013). Permeable pavements do not et al. 2013). Permeable paving provides a surface if this does occur. This fine aggregate in the joints and provide any amenity, social or ecological benefits. that is suitable for pedestrian and/or vehicular traffic In the first year of construction maintenance is while allowing rainwater to infiltrate through the slots is known to trap the most pollutants, including important, especially after the first large rainfall event. surface. There are a number of different alternatives heavy metals. While clogging may be a maintenance A1.3.2 Infiltration trenches The trench needs to be assessed for performance and for the surface material, including brick pavers, porous concern, the often enormous infiltration capacity of any sediment and debris build up which can cause concrete, porous asphalt, stone chip, and permeable permeable pavement systems means that considerable Infiltration trenches are excavated trenches which clogging (Armitage et al. 2013). Removal and cleaning of concrete block pavers (Armitage et al. 2013). Permeable clogging can be tolerated (Armitage et al. 2013). are lined with geotextile and filled with rock, or stone may be necessary. paving is usually constructed on top of a coarse gravel other granular materials, and are designed to receive base which creates the temporary storage facilities stormwater runoff from contiguous properties in urban PAGE 66 EVALUATING THE POTENTIAL RETURNS TO INVESTING IN GREEN URBAN DEVELOPMENT IN DURBAN EVALUATING THE POTENTIAL RETURNS TO INVESTING IN GREEN URBAN DEVELOPMENT IN DURBAN PAGE 67 The construction costs associated with infiltration (Armitage et al. 2013). They range in depth from between A1.3.4 Green roofs Advantages trenches are not very high, making them one of 1 – 4 metres but are usually approximately 1.5 metres in Green roofs comprise a multi-layered system that covers the more cost effective options in terms of their depth when serving a single building. the roof of a building with vegetative cover (Armitage • Good removal capability of atmospherically deposited urban pollutants ability to reduce runoff volume and treat pollutants. et al. 2013). The use of vegetative roof covers and roof • Can be designed to closely mimic the pre-development state of Their maintenance costs can be higher than other The basic construction costs include clearing and buildings gardens is an important source control for stormwater interventions, especially in areas that have fine grained removing of topsoil, surface bed preparation, pit • Ecological, aesthetic and amenity benefits runoff as they are designed to intercept and retain • Can be constructed on both new and already existing buildings soils. Infiltration trenches remove approximately 70- excavation, supplying and laying filter fabric or precipitation close to where it falls (i.e. at the source) • Help to insulate and regulate buildings against temperature 80% of TSS, 60-80% of total phosphorous, 25-60% of geotextile, supplying and laying of aggregate fill or extremes reducing the volume of runoff and attenuating peak total nitrogen, and 60-90% of heavy metals (Armitage porous media, supplying and laying of building sand, • Can be applied to high density urban areas flows. Green roofs provide great benefits in densely • May improve air quality et al. 2013). Their amenity and conservation value is supplying and laying of slotted pipes, top soiling of urbanised areas where there tends to be less space for • No additional land take poor, however they are generally constructed under the verged areas, and grassing of surface area. some of the other BMP interventions. Green roofs are ground and so the aesthetic impact is negligible. Limitations usually constructed on flat or gently sloping roof tops no Advantages greater than 30 degrees. The vegetative layer sits upon • More costly than conventional roof-runoff practices due to their a drainage layer which in turn lies upon a water proof added structural, vegetative and professional requirements • Have reasonable design lives of up to 20 years if maintained A1.3.3 Soakaways (sub-surface infiltration trenches) (professionals are required to ensure implementation of the properly and relatively easy to construct membrane to prevent any leakage below (Armitage et al. waterproofing and plant requirements Soakaways usually comprise an underground storage • Significantly decrease stormwater runoff volume, peak flow and 2013). Green roofs that are constructed in this manner • Opportunities for retrofitting may be limited by roof structure rate area that is packed with course aggregate or other • Particularly effective in removing particulate and suspended typically have weights of between 40 – 60 kg per m2. (size, strength etc.) • Not appropriate for steep roofs porous media that gradually discharges stormwater into stormwater runoff pollutants The structural design of the green roof needs to account • Detention of water within green roof storage layer may result in the surrounding soil (Armitage et al. 2013). Soakaways • Reduce downstream erosion and flooding for the additional weight of the green roof component failure to the waterproofing membranes which in turn may cause • Minimal net land take are similar to infiltration trenches in their operation materials and expected water detention volumes leakage or cause roof collapse • Plant varieties may be quite limited. Using indigenous vegetation and are also known as sub-surface infiltration beds or Limitations (Armitage et al. 2013). is best sub-surface infiltration trenches). They usually handle • Usually limited to relatively small connected areas roof runoff from single buildings, such as large industrial • They do not function well when constructed on steep slopes or in Green roofs are particularly effective when constructed Maintenance of green roofs include irrigation during buildings. Multiple soakaways can be linked to each unstable areas on roofs with large surface areas such as commercial establishment of vegetation, inspection for bare other to drain larger areas such as parking lots or • Sub-drain piping systems must be utilised or industrial buildings or large residential blocks. • when soakaways are implemented in very patches, weeds and plants that require replacement. major roadways. The type of aggregate material used • fine silt and clay stratum because of the low Irrigation may be required to keep the roof green during Leaf litter removal may be required for certain systems determines the infiltration characteristics of the device. • infiltration rates particularly dry periods. • Sedimentation within the collection and any possible stresses related to the roof and building Modular geo-cellular structures provide relatively • chambers will cause a gradual reduction in There are three main types of green roofs, namely: structure need to be checked. high stormwater treatment and rates of groundwater • the storage capacity recharge (Armitage et al. 2013). • Ecological and amenity value is poor extensive green roofs, intensive green roofs and simple Green roofs are expensive to construct and are one intensive green roofs (Armitage et al. 2013). Extensive of the least cost-effective options in terms of the cost The size of the soakaway is dependent on the porosity of The amount of water disposed of by soakaways depends green roofs generally incorporate low growing and low per unit reduction of runoff volume or pollutant loads. the aggregate used to fill the excavated pit. The soakaway on the infiltration potential of the surrounding soil, the maintenance vegetation that covers the whole roof Green roofs remove approximately 60-95% of TSS and empties either by percolation of the stormwater directly size of the pit and the bulk density of the fill material. surface. The roof is only accessed for maintenance 60-95% of heavy metals (Armitage et al. 2013). They into the underlying soil or via perforated sub-drains The amount of water retained by a soakaway is based on purposes. Usually indigenous vegetation such as mosses, provide a number of social and aesthetic benefits such installed within the pit. Soakaways are usually designed to the roof area of the building and the peak rainfall event herbs and grasses are used as they are relatively as air quality improvement in urban areas, temperature store the entire volume from a design storm and be able (mm) during the flood season. Soakaways are estimated self-sustaining. Intensive green roofs incorporate control, and amenity value. to infiltrate at least half of this volume within 24 hours to to be retain 70-80% of TSS, 25-60% of total nitrogen, 60- planters and trees and tend to have a high level of create further capacity for the runoff from subsequent 80% of total phosphorous, 60-90% of E.coli and 60-90% accessibility (Armitage et al. 2013). Rainwater harvesting rainfall events (Armitage et al. 2013). A single soakaway of heavy metals. Soakaways are relatively cost-effective interventions are often used as the primary irrigation can serve an area of roughly 1000 m2 but groups of in terms of runoff reduction as well as in terms of their source for intensive green roof flora. These roof systems soakaways can serve areas as large as 100 000 m2 ability to remove suspended solids. require more intensive and frequent maintenance. Simple intensive green roofs are a combination of both extensive and intensive green roofs, having both larger plants as well as low lying ground cover. These roofs generally require high levels of maintenance such as cutting, fertilizing and watering – which requires increased accessibility (Armitage et al. 2013). Figure A1.4 Soakaways are square or circular excavations either filled with rubble or other aggregate fill that are able to attenuate and treat significant amounts of stormwater runoff. They can be grouped and linked together to drain large areas such as highways and Figure A1.5 Green roofs achieve runoff treatment and infiltration through the construction of vegetative cover on roofs which increases industrial areas. storage, evapotranspiration and attenuation (Source: susdrain, www.susdrain.org) PAGE 68 EVALUATING THE POTENTIAL RETURNS TO INVESTING IN GREEN URBAN DEVELOPMENT IN DURBAN EVALUATING THE POTENTIAL RETURNS TO INVESTING IN GREEN URBAN DEVELOPMENT IN DURBAN PAGE 69 A1.3.5 Rainwater harvesting A1.3.6 Vegetated swales Rainwater harvesting systems collect and store rainfall especially if the systems are to be placed in residential Swales are shallow vegetated channels with flat and (Armitage et al. 2013). Swales have the potential to from hardened surfaces for later use. With minimal areas (Armitage et al. 2013). The following water balance sloped sides that are designed to store and convey manage stormwater indefinitely if they are properly treatment the water that is collected could be used equation is often used to calculate the volume of usable runoff as well as remove pollutants. Although swales are maintained. Maintenance activities tend to include to supplement the potable water supply and can be rainfall or the annual collectable rainfall: usually lined with grass, a variety of different types of regular mowing of grassed surfaces, weed control, re- used for a number of activities such as toilet flushing vegetation can be used to suit the specific site (Armitage seeding of bare ground, frequent clearing of litter and and irrigating crops and gardens (Armitage et al. V = R x A x C x FE et al. 2013). Swales serve as an alternative option to debris, and watering during extended dry periods. 2013). Storage of runoff from roofs and other elevated Where: the more typical roadside kerb or gutter and generally impervious surfaces is provided by rainwater tanks, V = volume of usable rainwater (l) have a larger stormwater storage capacity so they help Advantages barrels, cisterns or other storage structures until the R = average rainfall over a period (mm) to reduce runoff volumes and peak stormwater flows water is required (Armitage et al. 2013). The utilisation A = Area contributing to runoff (m2) (Armitage et al. 2013). Their ability to store and convey • Usually less expensive and more aesthetically pleasing than kerbs and their associated concrete- and stone-lined channels of stormwater as a water source not only saves potable C = runoff coefficient (0-1) significant volumes means that they require relatively • Runoff from adjacent impermeable areas is often completely water but it also significantly reduces the stormwater FE = filter efficiency (0-1) large surface areas in order to function effectively. infiltrated in-situ using swales discharge from roofs. Rainwater harvesting systems are • Reduce stormwater runoff volumes and delay runoff peak flows For a standard flat roof the runoff coefficient is 0.4 and • Retain particulate pollutants as close to the source as possible known to be particularly useful during extreme rainfall Swales are commonly used in combination with other • Easy to incorporate into landscaping with low capital costs events as they help to protect receiving streams and the filter efficiency is generally recommended to be 0.9 systems, such as buffers and bio-retention interventions, • Pollution and blockages are visible and easily dealt with rivers by reducing the initial runoff volumes and the as a conservative estimate (Armitage et al. 2013). to form a treatment train. In doing so runoff is retained Limitations associated polluted (Armitage et al. 2013). and dissolved pollutants in stormwater runoff are also Advantages removed. The combination of infiltration and bio- • Usually require a larger land area than conventional kerb and There are two types of stormwater collection and reuse infiltration removes the dissolved pollutants and the channel drainage systems • Can significantly reduce potable water consumption or provides • Not suitable for steep areas or areas with roadside parking systems that are generally applicable to residential, significant amounts of water to those that have no access to larger particles are filtered by the vegetation (Armitage • Risks of blockages in connecting pipe work commercial and industrial uses; namely the pumped potable water et al. 2013). A swale that has been well designed should • Limited removal capabilities for soluble pollutants and fine supply system and the gravity supply system. The water • Reduces pollutant loads that enter nearby watercourses provide reduction in impervious cover, pronouncement sediment • Attenuates flood peaks • Standing water in swales has the potential to result in the collected in the tank from the rooftop is then gravity • Wide variety of storage containers available and generally easy to of the surrounding natural landscape and multiple breeding of mosquitoes and the generation of foul odours fed into specified application points in and around the install aesthetic enhancements, and they should be designed building. The harvesting system could just involve the to meet flow conveyance requirements and effective Limitations Vegetated swales have low capital costs and are cost- collection of rainwater from rooftops via gutters into stormwater pre-treatment (Armitage et al. 2013). They effective in their ability to reduce peak flows and runoff a storage tank where water can then be collected for • Water quality needs to be monitored and is generally such that are usually suitable for road medians, verges, car parking the water can only be used for supplementary purposes volumes and to reduce pollutants. They have medium use. One large tank could be connected to and supply a runoff areas, park and recreational edges. • Rainwater reuse on a domestic scale is relatively expensive with to good amenity potential in that they provide a green number of houses. the storage tanks constituting the most significant cost of the The effective design life of a swale is directly related to alternative to grey infrastructure in urban environments. system There are a number of different types of stormwater the standard of maintenance, particularly in the first two Vegetated swales remove approximately 60-90% of TSS, collection and storage systems that are commercially years during the period of plant establishment which 70-90% of hydrocarbons, 25-50% of total phosphorous, The initial construction costs associated with the 30-90% of total nitrogen and 40-90% of heavy metals. available. An effective system will include strategically rainwater harvesting system are relatively expensive often requires frequent weed control and replanting placed roof gutters and pipes, a filter sock to catch with the tank constituting the most significant cost. leaves/debris, a rainwater storage facility such as a However, maintenance costs can be low and the water tank or barrel, and a UV disinfection device. Storage that the tanks supply to households is an extremely facilities that are child proof, insect and vector proof important benefit, especially in areas where access to should be given preference during the selection process, running water is limited. Figure A1.7 Swales are shallow grassed or vegetated channels used Figure A1.6 Diagram of a rainwater harvesting system. The first picture shows high stormwater runoff with none of the rain being collected to collect and/or move water whereas the second picture shows how rainfall is trapped and collected from the roofs in tanks and the amount of runoff entering (Source:  susdrain, www.susdrain. streams and rivers is significantly reduced. org) PAGE 70 EVALUATING THE POTENTIAL RETURNS TO INVESTING IN GREEN URBAN DEVELOPMENT IN DURBAN EVALUATING THE POTENTIAL RETURNS TO INVESTING IN GREEN URBAN DEVELOPMENT IN DURBAN PAGE 71 A1.3.7 Filter strips different smaller stormwater interventions such as filter Advantages Filter strips are maintained grassed areas of land that strips, temporary pond areas, sand beds, mulch layers are used to manage shallow overland stormwater • Particularly effective in removing suspended solids (TSS) and a wide variety of vegetation (Armitage et al. 2013). • Efficient stormwater management technologies in areas with runoff through several filtration processes in a very limited space as they can be implemented They are particularly effective at managing stormwater similar manner to buffer strips (Armitage et al. 2013). • beneath impervious surfaces runoff from minor and more frequent rainfall events. • They manage stormwater runoff effectively on relatively flat Filter strips are usually gently sloping and provide terrains with high ground water tables where bio-retention Bio-retention areas can manage stormwater runoff on opportunities for slow conveyance and infiltration. They systems are inappropriate • The filtered effluent can be reused for most non-potable a number of sites, such as between residential plots, therefore help to attenuate floods peaks and retain domestic water uses including: toilet flushing, dish washing and alongside parking lots, adjoining roadways and within pollutants. They are commonly designed to accept garden watering; and large landscaped impervious areas. The engineered runoff from upstream development and are usually • May be retrofitted with relative ease into existing impervious developments, constrained urban locations or in series with soil media and the different varieties of vegetation are located between hard-surfaced areas and a receiving conventional stormwater management systems managed to capture and treat a specified water quality stream, surface water collection or treatment system. volume of stormwater runoff and in doing so they They may also be used downstream of agricultural land Limitations reduce runoff quantities and rates whilst improving the to infiltrate and intercept runoff from these areas. • Generally ineffective in controlling stormwater peak discharges quality of stormwater entering watercourses further • Premature clogging is likely to occur in sand filters if they receive downstream (Armitage et al. 2013). Filter strips use vegetative filtering as a primary means excessive sediment carrying runoff, especially from construction of stormwater runoff pollutant removal and if properly Figure A1.8 Filter strips are maintained grassed areas of land that sites and areas with open soil patches designed are able to remove most sediment and other are used to manage shallow overland stormwater • Large sand filters are not generally attractive, especially if they Advantages runoff through several filtration processes. They are are not covered with grass or other vegetation settleable solids such as hydrocarbons (Armitage et al. usually located as strips adjacent to development areas, • Sand filters are expensive to implement and maintain relative to • Reduces runoff volumes and rates, and attenuates flood peaks 2013). Soluble nutrients and heavy metals, however, are roads and waterways. most options technologies effectively often not adequately removed. The pollutant removal • If designed and/or implemented incorrectly, they may fail, • Flexible application means these areas are easily incorporated resulting in standing pools of water which have the potential to into a wide variety of landscapes and water retention characteristics of filter strips is attract nuisances such as mosquitoes and midges. • Very effective at the removal of most stormwater runoff determined by the relationship between the length, A1.3.8 Sand filters pollutants width, slope and soil permeability compared to the • Well-suited for installation in highly impervious areas, provided There are many different forms of sand filters. They the system is well-engineered and adequate space is made stormwater runoff rate and velocity (Armitage et al. A1.3.9 Bio-retention areas usually comprise of a sedimentation chamber that is available 2013). • Good retrofit capability linked to an underground filtration chamber comprising Bio-retention areas, sometimes referred to as ‘rain • Aesthetically pleasing sand or other media through which stormwater runoff gardens’ are landscaped depressions which are typically Advantages can pass (Armitage et al. 2013). The sedimentation Limitations under drained and rely on engineered soils, enhanced • Installation and maintenance costs are relatively low and layout chamber facilitates the removal of suspended vegetation and filtration to remove pollution and reduce • Not suited to areas where the water table is shallower than and design is flexible particulates and heavy metals, whilst the filtration runoff downstream (Armitage et al. 2013). They are 1.8m • Significant removal of suspended solids and hydrocarbons. They chamber removes smaller particulate pollutants. The • Requires frequent landscaping and maintenance to remain trap the pollutants close to source usually employed to manage the runoff from the first aesthetically pleasing             • Infiltration of stormwater runoff helps to attenuate flood peaks removal mechanism is partly through filtration by the 25mm of rainfall by passing runoff through a number • Susceptible to clogging if surrounding landscape is not managed • Integrate well within the natural landscape and can provide open sand bed and partly through microbial action within of natural processes such as filtration, absorption, • Not suitable for areas with steep slope space areas for recreation as well as amenity value the media (Armitage et al. 2013). Sand filters tend to be • Construction costs can be high biological uptake, sedimentation, infiltration and Limitations installed for use in impervious areas that are less than detention. These areas tend to include a number of 8000m2 but may be designed to manage runoff from • Clogging of subsurface drainage media can occur if maintenance larger areas too. is poor • Limited potential for the removal of fine sediments and dissolved pollutants Sand filters are similar to bio-retention areas and other • Stormwater runoff needs to be spread out in order for the strips bio-retention systems, with the only difference being to operate optimally • Minimal stormwater storage capacity and not good at treating that stormwater runoff passes through a linear filter high velocity flows. They are not suitable for steep slopes. medium without vegetation (Armitage et al. 2013). The primary objective for sand filters is water quality Filter strips are designed specifically to control for improvement and they are particularly effective in the nutrients and pollution more so than water quantity and removal of hydrocarbons. They are also used extensively are therefore more efficient at trapping and reducing to remove sediment and other particulate pollutants TSS and pollutants than they are at reducing stormwater from the first flush (Armitage et al. 2013). runoff. Grass filter strips remove approximately 50- 85% of TSS, 70-90% of hydrocarbons, 10-20% of total Sand filters can be expensive to construct and often phosphorous, 10-20% of total nitrogen and 25-40% of require regular maintenace, making them a less cost- heavy metals. effective option. They are highly efficient at removing suspended solids and pollutants. They remove approximately 80-90% of TSS, 50-80% of hydrocarbons, 50-80% of total phosphorous, 25-40% of total nitrogen, Figure A1.9 Bio-retention areas are landscaped depressions employed to manage 40-50% of E.coli and 50-80% of heavy metals from runoff by passing it through stormwater runoff (Armitage et al. 2013). several natural processes. Rain gardens are an example of a bio- retention area. (Source: susdrain, www.susdrain.org) PAGE 72 EVALUATING THE POTENTIAL RETURNS TO INVESTING IN GREEN URBAN DEVELOPMENT IN DURBAN EVALUATING THE POTENTIAL RETURNS TO INVESTING IN GREEN URBAN DEVELOPMENT IN DURBAN PAGE 73 Routine inspections and maintenance are required to tend to be located towards the end of the stormwater 60% of hydrocarbons, 20-70% total phosphorous, 20- saturation level of the wetland and the degree to which ensure that bio-retention areas function effectively. The management train so are used if the extended treatment 60% total nitrogen, 50-70% E.coli and 40-90% of heavy the nutrients and pathogens adhere to other particles design life of these areas, as with most interventions, of runoff is required. The pollutant removal capability metals (Armitage et al. 2013). and sediments (Armitage et al. 2013). is directly related to the quality and frequency in of a detention basin can be improved through the maintenance (Armitage et al. 2013). Maintenance construction of a sediment trap at the entrance to the The strategic positioning of such storage areas in urban Constructed wetlands usually include four distinct zones includes regular inspections, litter and debris removal, basin (Armitage et al. 2013). areas can enrich the urban environment and facilitate (Armitage et al. 2013): replacement of mulch areas, vegetation management maintenance operations. In fact, such areas, given their and sediment removal. The hydraulic and pollution removal performance dimensions, can be easily used as social and recreation ƒƒ The inlet zone which includes a sediment forebay for of detention basins depends on good maintenance. areas, such as play grounds or football fields, or for the removal of the more coarse sediments and litter Bio-retention areas can have high initial construction Regular inspections are needed to check if the clearing agriculture. There is a good example of this in San Paolo, entering the system; costs, making them less cost-effective in terms of cost of accumulated sediment is necessary, especially if the Brazil, where floodplain storage has been applied to per unit reduction of runoff volumes and pollutant loads. basin is being used as a field or common (Armitage et mitigate the flood risk from the Tamanduateí River, as ƒƒ The macrophyte zone which is usually shallow and They remove approximately 50-80% of TSS, 5-80% of al. 2013). Other maintenance includes the management shown in Figure A1. 10b (Giugni et al. 2012). heavily vegetated and facilitates the removal of finer hydrocarbons, 50-60% of total phosphorous, 40-50% of vegetation, inspections after high rainfall events, and particles and the uptake of soluble nutrients such as total nitrogen and 50-90% of heavy metals (Armitage et possible de-silting. nitrogen and phosphorous; al. 2013). Their amenity potential is good. A1.3.11 Constructed treatment wetlands ƒƒ The macrophyte outlet zone which channels cleaner Advantages Wetlands are generally marshy areas of shallow water stormwater runoff downstream; and that are either partially or completely covered in A1.3.10 Detention basins • Able to temporarily store large volumes of stormwater thus attenuating downstream flood peaks aquatic vegetation. Wetlands provide habitat for a wide ƒƒ The high flow bypass channel which protects the inlet, Detention basins or detention ponds are temporary • Relatively inexpensive to construct and easy to maintain variety of fauna and flora and provide aesthetic appeal, outlet and vegetative zones from damage and scour storage facilities that are usually dry but are designed so • Serve multiple purposes during drier seasons, particularly as especially in urban areas where green open space is sports fields, play parks or commons during abnormally high flow events. that they are able to store stormwater runoff for short • If managed regularly, they can add aesthetic value to adjoining limited. Constructed wetlands are man-made systems periods after high rainfall events (Armitage et al. 2013). residential properties as well as presenting fewer safety hazards that are designed to mimic the natural wetland systems Other considerations include litter traps or trash racks The captured stormwater either infiltrates into the than wet ponds due to the absence of a permanent pool of in areas where they were not previously found (Armitage at the inlet to the wetland which prevents litter, debris, water. underlying soil layers or is drained into the downstream et al. 2013). course sediment and other pollutants from entering watercourse at a predetermined rate. Therefore they Limitations the macrophyte zone and from being carried further are effective at regulating the flow in downstream They are able to serve larger catchment areas and are downstream. The selection of the vegetation to be used • Not very good at removing dissolved pollutants and fine material watercourses. Generally detention basins are designed • Generally not as effective in removing pathogens as constructed very useful at removing nutrients and suspended solids in the wetland is important and a number of selection to temporarily store as much water as possible for 24 wetlands from stormwater runoff from residential areas. The most criteria should be considered, such as the speed at which • Siltation can be a problem and the floors of detention ponds can common stormwater pollutant treatment processes that the vegetation establishes itself and grows, the disease – 72 hours whilst aiming to provide a safe and secure become swampy for some time after major rainfall public environment (Armitage et al. 2013). • Not very suitable in areas with a relatively high water table, or wetlands provide are sedimentation, fine particulate or weed risk associated with vegetation, the suitability where the soil is very coarse and there is a risk of groundwater filtration and biological nutrient and pathogen removal of the vegetation for the local climate, the tolerance of Detention basins are typically lined with grass and are contamination (Armitage et al. 2013). The percentage removal of vegetation to becoming water-logged and the pollutant designed to be multifunctional in that they provide pathogens and nutrients depends largely on the removal capacity of the various vegetation types access to recreational area when dry. They are surface Detention basins are relatively inexpensive to construct pollution concentration of the inflow, the rate at which (Armitage et al. 2013). storage basins that provide flow control through the and have low maintenance costs, making them cost- the water is flowing through the wetland, the pollution attenuation of stormwater runoff and also facilitate effective options for control runoff. Detention basins some settling of particular pollutants. Detention basins remove approximately 45-90% of suspended solids, 30- (a) (b) Figure A1.11 Constructed treatment wetlands are man-made systems designed to mimic natural wetland systems (Source:  susdrain, www. Figure A1.10 (a) Lamination effect due to the flood plain storage and (b) Example of flood plain storage in San Paulo, Brazil (Giugni et al. 2012). susdrain.org) PAGE 74 EVALUATING THE POTENTIAL RETURNS TO INVESTING IN GREEN URBAN DEVELOPMENT IN DURBAN EVALUATING THE POTENTIAL RETURNS TO INVESTING IN GREEN URBAN DEVELOPMENT IN DURBAN PAGE 75 A1.4.1 Sweeping and solid waste management ƒƒ In Dar es Salaam, the Mlalakua River Restoration land uses. Riparian buffers play an important role in Advantages Interventions such as street sweeping and proper Project was initiated in 2012 and is a multi-stakeholder improving water quality as well as providing stormwater • Highly efficient at removing pollutants from stormwater runoff removal and disposal of solid waste help to reduced partnership that has focused on implementing infiltration benefits and conservation value. Riparian • May attenuate peak stormwater flows depending on location and design of wetland sediment (and hence pollution) loads entering the measures that enhance healthy living conditions of the buffers are similar to filter strips but differ in that they • Good community acceptability and provides amenity value in drainage system, and help to prevent solid waste from riverine communities, and prevent further pollution on are generally forested and always occur adjacent to river urban environments blocking culverts and reducing the efficiency of the a sustained basis. The Mlalakua River originates from channels. Filter strips tend to be located in urban areas Limitations conveyance system. the Mzinga and Kizinga Rivers and drains into Msasani adjacent to development. Bay in Kinondoni Municipality. The restoration project • Wetlands could potentially attract mosquitos and birds whose forms part of the International Water Stewardship Riparian buffers reduce excess amounts of sediments, faeces can increase the amount of phosphorous in the water organic material, nutrients and pesticides in surface • Limited to relatively flat land A1.4.2 River cleaning and stewardship Programme (IWaSP), an international programme for • Limited depth range for flow attenuation and little reduction in water security managed by the Deutsche Gesellschaft runoff and reduce excess nutrients and other chemicals One approach to keeping rivers clear of litter and in shallow ground water flow (Waidler et al. 2009). run volume fur Internationale Zusammenarbeit (GIZ). Project • Flooding of the wetland may result in water logging of the plants debris and maintaining a healthy river system is to They are also known to reduce pesticide drift entering which may result in die off and a loss in treatment efficiency activities include physical clean-up of the Mlalakua involve communities that live alongside rivers and the water body. With the use of suitable indigenous River, the establishment of sustainable solid waste and streams. Community involvement projects can have vegetation, riparian buffers have the potential to provide wastewater management systems, such as introducing Inspection and maintenance of constructed wetlands multi-sectoral impacts as they generate employment a habitat corridor for wildlife (Armitage et al. 2013). private waste collectors and developing new recycling can be frequent and costly, however can be reduced opportunities, provide awareness, safeguard centres, building capacity of service providers, raising through effective pre-treatment such as litter traps, communities and provide city-wide services such as awareness in communities, improving household Advantages trash racks and sediment forebays at the inlet to the functioning river systems that are clean and clear of sanitation, and implementing effective law wetland (Armitage et al. 2013). Maintaining healthy litter. Sections of rivers or streams are maintained • Relatively low costs involved in planting and establishing buffer enforcement measures. Project partners include the zones vegetation and adequate flow conditions is essential by cooperatives which are responsible for removing Wami River Basin Water Board (WRBWB), National • Significantly improve water quality of streams and rivers to the efficient functioning of the constructed wetland alien vegetation, rubble and any solid waste blocking • Infiltration of stormwater runoff helps to attenuate flood peaks Environment Management Council (NEMC), the local and this requires harvesting of the vegetation, such as the free flow of water down the stream or river. They • Natural intervention that provides amenity and conservation Kinondoni Municipal Council (KMC), Coca-Cola value. papyrus or reeds. Once harvested the vegetation can be are also responsible for maintaining the grass and Kwanza, Nabaki Africa, Nipe Fagio, the Bremen composted and re-used. other vegetation along the banks of the waterway. Limitations Overseas Research and Development Association The cooperatives generally consist of members of the Wetland construction costs can be high when compared (BORDA), and GIZ. Donor funding for the initial phase • Relatively limited potential for the removal dissolved nutrients community that are unemployed and vulnerable and to other interventions, however their ability and of the project was approximately EUR 400 000. In April the project focuses on raising awareness and generating efficiency in removing nutrients and pollutants makes 2016 the multi-stakeholder project came to an end Riparian buffers can be cost-effective in that they require employment. Two examples of such projects include the them relatively cost-effective. They also have the added with the project being handed over to the Mlalakua no major engineering or construction. The costs are Sihlanzimvelo Stream Cleaning Project in Durban and the benefit of providing amenity value. Construction costs Community Change groups which will continue on with associated with the purchasing of seedlings and the Mlalakua River Restoration Project in Dar es Salaam: per hectare of wetland are exponential, meaning the improving the health of the river. labour required to plant them. Riparian buffers are cost per hectare decreases the larger the wetland. ƒƒ In Durban, the Sihlanzimvelo Stream Cleaning Project efficient at removing suspended solids, hydrocarbons Constructed wetlands are estimated to remove has been very successful in areas of the municipality and other pollutants. They are less effective at removing A1.4.3 Riparian buffers approximately 80-90% of suspended solids, 50-80% where a number of rivers were considered critical in dissolved nutrients such as nitrogen and phosphorus. of hydrocarbons, 30-40% of total phosphorous, 30- terms of health and functioning. Approximately 470km A riparian buffer is a vegetated area, or buffer strip, They contribute to the infiltration of stormwater runoff 60% total nitrogen, 50-70% E.coli and 50-60% of heavy of degraded river systems were identified and pilot that is located adjacent to a stream or river channel and and therefore attenuate flood peaks. metals (Armitage et al. 2013). study areas were initiated. Residents of the four is usually forested, which helps to shade and partially communities formed part of the initial pilot study. protect the waterway from the impacts of adjacent They were employed to clean and maintain sections of A1.4 Non-structural interventions the river adjacent to where they live. This includes Non-structural measures do not involve physical unblocking of culverts and the removal of litter and construction but use knowledge, practice or agreement alien vegetation. Grass and vegetation along the to reduce risks and impacts, in particular through riverbed is maintained to a certain height. The results policies and laws, public awareness raising, training have been impressive and rivers have become cleaner, and education (Kundzewicz 2002). These include the risk of flooding has reduced through the removal flood warning systems, land use regulations such of litter and debris and the communities feel safer as as development setbacks which identify where the areas became more accessible and crime has development can and cannot occur, or to what elevation decreased. Through the project, residents have structures should locate their lowest habitable floor to; become more aware of the benefits that are derived flood proofing and retrofitting of buildings may increase from healthy river systems and have an incentive to the strength against flood actions; elevation of buildings keep it clean. The Sihlanzimvelo Stream Cleaning may avoid completely the inundation. Flood insurance Project is funded by the eThekwini Municipality and and relocations also belong to this typology of measure. the South African government’s Expanded Public Some of these measures are described in more detail Works Programme (EPWP) and includes a contractor Figure A1.12 Riparian buffers are located adjacent to streams and river channels. They can either be made up of grasses and smaller plants as in development component. The budget for the project picture (a) or they can be densely vegetated with trees and bushes as in picture (b). They provide a buffer between adjacent land below. uses such as agriculture and residential areas and waterways. is R45 million (approximately US$3 million). Over the course of the project a total of 732 job opportunities have since been created. PAGE 76 EVALUATING THE POTENTIAL RETURNS TO INVESTING IN GREEN URBAN DEVELOPMENT IN DURBAN EVALUATING THE POTENTIAL RETURNS TO INVESTING IN GREEN URBAN DEVELOPMENT IN DURBAN PAGE 77 A1.4.4 Catchment reforestation Catchment reforestation is an important intervention and reduce excess nutrient and sediment loads into that does not differ much from the riparian buffer the rivers and streams (Rutherford et al. 2007, Ouyang intervention. Catchment reforestation focuses on et al. 2013, Opperman 2014). Therefore land cover planting indigenous trees and shrubs within the change, such as deforestation, increases nutrient and greater catchment area, in particular in areas that were sediment loads entering waterways, alters infiltration previously forested. By increasing the number of larger rates, elevates greenhouse gas emissions and leads to trees and shrubs in the catchment the amount of runoff changes in regional and local hydrological cycles (Ouyang entering streams and rivers in reduced through trapping et al. 2013). The latter results in a significant reduction and infiltration. Forested areas are well known for their in floodwater retention and an associated loss of flood ability to reduce runoff as well as reduce nutrient and control (Ouyang et al. 2013). Therefore reforestation pollutant loads entering waterways. Reforestation in and the development of forested floodplain buffers the catchment also increases conservation value and in a catchment can reduce the water discharge and amenity value. sediment load into the rivers and streams and enhance flood attenuation based on catchment characteristics Advantages (Ouyang et al. 2013). Vegetation can have numerous impacts on the amount of rainfall that becomes runoff • Relatively low costs involved in planting and re-establishing and can generally affect flooding in three specific ways: forested areas • Significantly improve water quality of streams and rivers by affecting the size and shape of the stream channel • Infiltration of stormwater runoff helps to attenuate flood peaks (geomorphology), by altering the amount of water that • Natural intervention that provides amenity and conservation reaches the stream channel (hydrology), and by altering value. the resistance to flow (hydraulics) (Rutherford et al. Figure A1.14 Catchment reforestation will aid in runoff infiltration reducing the overall amount of stormwater reaching rivers and streams. Limitations 2007, Opperman 2014). Reforestation will also aid in removing sediments and nutrients. • Relatively limited potential for the removal dissolved nutrients River channels that are forested have a higher roughness which means that the flood arrives later and that the The capital costs involved in catchment reforestation are peak flow is attenuated when compared to channels A1.5 Relative performance of different measures relatively low when compared to other interventions. cleared of vegetation. The response to larger floods Generally a combination of stormwater management that will achieve the outlined objective in the most This is because the intervention involves no engineering generally differs from smaller floods with smaller measures would be applied. There are usually trade-offs cost-effective way. This involves determining the cost or construction work and is based solely on the attenuation of the peak observed in the case of the small among the interventions and finding the correct balance effectiveness of different interventions, i.e. a cost per planting of trees and shrubs. Costs include the buying flood (Rutherford et al. 2007). Revegetating the riparian can be a complex task. The active structural measures, unit reduction of runoff volume (m3) or cost per unit of seedlings and the labour involved in planting them. zone in the Murrumbidgee catchment in Australia had such as floodplain storage interventions will reduce the reduction in pollutant loads (kg), depending on what Catchment reforestation provides numerous benefits a considerable effect on the size and timing of the need for extensive passive measures such as levees or the proposed project is trying to achieve. Outside of such as amenity and conservation value as well as flood peak reaching different outlets (Figure A1. 14; the widening of channels. However, active measures cost-effectiveness, interventions need to be assessed in contributes to providing clean water. Rutherford et al. 2007). At the upstream site (C) the peak alone often will not be able to eliminate flooding terms of any other benefits that they may provide, such is attenuated y 18% and at the larger outlet (A) the peak problems completely and thus generally will need to as conservation value, amenity value or social benefits Trees absorb rainfall, slow down flow velocity, disperse is attenuated by 29% (Figure A1. 14). be implemented in conjunction with some conveyance such as increased water supply. The interventions need surface runoff, offset water discharge, filter pollutants, infrastructure. Generally, the ‘softer’ the intervention to be realistic in terms of what is feasible and practical the less efficient it tends to be in terms of m3 reduction within the designated project area. per unit of space. However, the softer interventions tend to have greater benefits in terms of amenity and Numerous studies have examined the relative ability social value. Therefore it is necessary to develop a sound of different interventions to reduce pollutant loads, methodology for evaluating the interventions based on flow volumes and attenuate peak flows during storm cost-effectiveness, efficiency in removing peak flows, events, and their cost-effectiveness. Estimates for cost- reducing runoff volume and water quality amelioration, effectiveness in terms of cost per unit runoff reduction and providing amenity, conservation and social benefits. ($/m3) and cost per unit reduction in pollutant loads These are discussed in more detail below. ($/kg) were collated from a wide range of stormwater management literature. These estimates tend to be When developing a proposed plan for managing site specific and based on local costs but nonetheless stormwater it is important to link together the various provide a clear indication of which interventions are interventions and the benefits that they provide with generally more cost-effective in terms of their ability to the greatest possible efficiency. That is, identify the reduce peak runoff and remove pollutants. combination of interventions for a specific project site Figure A1.13 The effect of revegetation on discharge upstream and downstream of the Murrumbidgee in Australia (Source: Rutherford et al. 2007) PAGE 78 EVALUATING THE POTENTIAL RETURNS TO INVESTING IN GREEN URBAN DEVELOPMENT IN DURBAN EVALUATING THE POTENTIAL RETURNS TO INVESTING IN GREEN URBAN DEVELOPMENT IN DURBAN PAGE 79 Table A1.1 Measured pollutant removal capacities of selected stormwater management options and technologies (Source: Armitage et al. 2013) Green roofs Pollutant Removal (%) Permeable pavements Rainwater harvesting Hydro- Faecal Heavy Option/Technology TSS TP TN carbons Coliforms Metals Soakaways Source Controls Infiltration trenches Green roofs 60-95 - - - - 60-90 Constructed wetlands Sand filters 80-90 50-80 50-80 25-40 40-50 50-80 Sand filters Underground sand filters 75-90 - 30-60 30-50 40-70 40-80 Bioretention areas Surface sand filters 80-90 - 50-60 30-40 - - Detention Basin Filter drains 50-85 30-70 - - - 50-80 Filter strips Soakaways 70-80 - 60-80 25-60 60-90 60-90 Swales Oil and grit separators 0-40 40-90 0-5 0-5 - - Riparian buffers Modular geocellular structures PS PS PS PS PS PS Catchment reforestation Stormwater collection and reuse PS PS PS PS PS PS 0 50 100 150 200 250 Local controls Cost per unit runoff reduction ($/m 3) Bioretention areas 50-80 50-80 50-60 40-50 - 50-90 Figure A1.15 Comparison of average cost per unit volume of runoff reduction for various stormwater management options, based on data in the literature Filter strips 50-85 70-90 10-20 10-20 - 25-40 Infiltration trenches 70-80 - 60-80 25-60 60-90 60-90 Permeable pavements 60-95 70-90 50-80 65-80 - 60-95 A1.5.1 Average cost effectiveness in terms of peak flow and A1.5.2 Average cost effectiveness in terms of water quality volume reduction amelioration Swales 60-90 70-90 25-80 30-90 - 40-90 Enhanced dry swales 70-90 70-90 30-80 50-90 - 80-90 Cost-effectiveness in terms of runoff reduction ($/ A number of interventions are designed to specifically m3) is shown in Figure A1. 15. These estimates are control and improve the quality of stormwater runoff. Wet swales 60-80 70-90 25-35 30-40 - 40-70 based on reviews and examples from the stormwater Generally their performance is assessed based on their Vegetated buffers* 50-85 70-90 10-20 10-20 - 25-40 management literature (Joksimovic & Alam 2014, Liu pollutant removal capabilities. Some interventions Regional controls et al. 2015, Jiang et al. 2015, Committee for Climate may be more efficient at removing suspended Constructed wetlands 80-90 50-80 30-40 30-60 50-70 50-60 Change 2012, Xiao & McPherson 2002, McPherson et solids and hydrocarbons, whereas others may be al. 1999). From the examples it is clear that green roofs particularly efficient at removing soluble nutrients Extended detention shallow wetland 60-70 - 30-40 50-60 - - and permeable pavements are the least cost-effective, such as phosphorus. Often this means that a number Pocket wetland* 80-90 50-80 30-40 30-60 50-70 50-60 even though they are efficient at reducing runoff, due of interventions are required to achieve a specified Submerged gravel wetland 80-90 - 60-70 55-60 - 85-90 to their higher capital and maintenance costs compared outcome. The capacity for pollutant removal of different Detention ponds* 45-90 30-60 20-70 20-60 50-70 40-90 to other options. Soakaways and infiltration trenches interventions is summarised in Table A1.1. Extended detention ponds 65-90 30-60 20-50 20-30 50-70 40-90 are the most cost-effective of the structural engineering methods as they are cheaper to construct and maintain. Detailed information provided in Armitage et al. (2013) Infiltration basins 45-75 - 60-70 55-60 - 85-90 Constructed wetlands, sand filters, bioretention areas, (Table A1. 1) was used to assess the water quality Retention ponds 75-90 30-60 30-50 30-50 50-70 50-80 detention basins, filter strips and swales generally are all amelioration performance of the various interventions. *Estimated values based on similar stormwater technologies relatively cost-effective. They are however less efficient These data were combined with cost data to estimate TSS – Total Suspended Solids, TP = Total Phosphorous, TN = Total Nitrogen at trapping or attenuating peak flows after a large storm. relative cost-effectiveness for a selected range of Riparian buffers and catchment reforestation represent interventions (Figure A1. 16). the most cost-effective option in that they do contribute significantly to rainwater infiltration, but they also do The most cost-effective options in terms of TSS removal Detention Basin not trap or attenuate peak flows (which is not captured are filter strips, swales, sand filters and detention Constructed wetlands in the $/m3 assessment). The structural engineering basins. Riparian buffers and catchment reforestation are Bioretention areas options are more efficient in this regard. also cost-effective options. Constructed wetlands and Sand filters bioretention areas are among the least cost-effective for Filter strips all three pollutants/nutrients, but they do have higher Swales amenity values when compared to the other options. Riparian buffers Catchment reforestation 0 2 4 0 6 200 400 600 0 2000 4000 6000 Cost per unit TSS removed ($/kg) Cost per unit TN removed Cost per unit TP removed ($/kg) ($/kg) Figure A1.6 Comparison of cost per unit mass of pollutant/nutrient reduction for various stormwater management options PAGE 80 EVALUATING THE POTENTIAL RETURNS TO INVESTING IN GREEN URBAN DEVELOPMENT IN DURBAN EVALUATING THE POTENTIAL RETURNS TO INVESTING IN GREEN URBAN DEVELOPMENT IN DURBAN PAGE 81 A1.5.3 Overall effectiveness, cost-effectiveness and potential co-benefits The relative effectiveness of different interventions in measures to be implemented in combination and/or terms of flood and water quality amelioration, their at scale. Green engineering measures also vary in their cost-effectiveness, and other potential benefits are cost-effectiveness and may not always compete with summarised in qualitative terms in Table A1. 2. This conveyance measures. They do however, also present suggests that while conveyance measures are highly much greater opportunities for delivering co-benefits, effective for reducing flood exposure/risk, they make such as water supply and the provision of recreational little contribution to water quality amelioration, they areas. The latter is particularly the case for the vegetated vary in terms of cost-effectiveness and have relatively options which have greater aesthetic appeal. little in the way of co-benefits. Indeed, they are more likely to lead to externalities such as damage to The protection or restoration of natural systems in aquatic ecosystems or acerbation of flooding further catchment areas contributes to the reduction and downstream. Of the conveyance measures, detention retardation of flows and to water quality amelioration. basins are potentially beneficial in terms of providing Within the flood prone areas, riparian buffers and opportunities for amenity, such as sunken sports fields. functional floodplain areas reduce the exposure to flooding, and further contribute to water quality The “green” engineering measures are generally less amelioration. In all cases, these areas have the potential efficient in flood protection, but are important for water to contribute significantly in terms of other co-benefits. quality. Effective flood protection will require these Table A1.2 Relative merits (indicated by number of “X”) of different measures for stormwater and flood risk management, based on the literature. Measures considered in this study area are marked with an asterisk. Flood Conveyance/ attenuation/ Water Cost- Conser- Reduction of Reduction of quality effective- Water Amenity vation Option/technology exposure flood risk amelioration ness supply potential value Conveyance measures (lower catchment) Swales/drains XX XX Channel enlargement/ XXX X canalisation/levees Hydraulic bypass XXX X ‘Green’ engineering measures (mid-upper catchment) Infiltration trenches XXX XXX XXX X Soakaways XXX XX XX X Permeable pavements* XXX XXX XX X Rainwater harvesting X X X XXX Bio-retention areas XX XXX XX X XX Sand filters XX XXX XX Green roofs* XX XXX X XX Filter strips XX XXX XXX X Vegetated swales XX XXX XXX X Constructed wetlands* XX XXX XX XXX X Detention basins* XXX XX XXX X XXX Non-structural measures Development setbacks X XXX XX X Conventional solid waste XX X XX XX X management* River cleaning programmes* X X XXX XX XX Protection/restoration of XX XXX XXX XX XX XXX catchment forests + wetlands* Protection/restoration of X XX XXX XXX XXX XXX riparian areas, floodplains* Source: US EPA 1999, NRCS Illinois 2005, CIRIA 2007, Fletcher et al. 2003, Stovin & Swan 2007, SNIFFER 2007, UN-Habitat 2008, Dep. Environmental Protection 2010, Michie 2010, FAO 2011, Toronto & Region Conservation Authority 2013, Natural Water Retention Measures Project 2013, Armitage et al. 2013, EA 2015, Hull City Council 2015, Morales Torres et al. 2015, Severn Trent Water 2015, TNC 2015. PAGE 82 EVALUATING THE POTENTIAL RETURNS TO INVESTING IN GREEN URBAN DEVELOPMENT IN DURBAN APPENDIX 2. SCENARIO ASSUMPTIONS A2.1 Sanitation measures A2.1.1 Scenario 1-2 (sanitation backlog) For Scenarios 1 and 2 an adjustment to sanitation had The assumptions for scenarios 1 and 2 were as follows: to be made to incorporate the changes associated with increased development in the catchment. It was ƒƒ Non-point source runoff from newly developed areas assumed that undeveloped land was developed as per will change to be the same as the new landuse type as the eThekwini Scheme Zonation Plans with all other stipulated in zonation plans; areas remaining the same. As a result, sanitation had to be adjusted to include developed areas. Non-point ƒƒ All newly developed residential areas were provided source pollutants in these new areas were changed to with either waterborne sewage or with urine diversion the associated land use as described in the zonation dehydration toilets as per dwelling densities provided plans. It was also assumed that the newly developed in zonation reports and average household size for the residential areas generated additional point source EMA; throughput to the existing WWTWs. ƒƒ The extra sewage generated from the addition of Based on information from the zonation reports it newly developed residential areas will be treated at was assumed that land zoned for general multi-unit existing WWTW in the catchment; residential use would have a density of 25 dwelling ƒƒ Treated effluent will be discharged into the sub- units per ha. With an average household size of 3.2 in catchment from where it is taken; the EMA, this equates to a population density of 80 people per ha. Where land was zoned under the less ƒƒ All effluent at WWTWs will be treated to within formal township establishment act (LFTEA) or the black General Effluent Limits as defined in national communities development act (BCDA), the assumption guidelines (Table A2. 2); of 250 dwelling units and 800 people per ha was applied. Based on information provided by sanitation ƒƒ Servicing and maintenance of existing WWTW and professionals, an average of 150 litres of sewage per wastewater infrastructure will be conducted to meet person per day was assumed in order to calculate the above limits on an ongoing basis; and the total extra throughput to WWTW from the newly developed areas. It was assumed that two thirds of ƒƒ Sanitation in informal settlements remains the same. households in the LFTEA and BCDA areas received urine diversion dehydration toilets1 and one third received access to waterborne sewage. The extra throughput, a total of 28.6 ML, was added as additional output to the already existing WWTW based on the location of each newly developed area in relation to these WWTW in the U60F catchment (Table A2. 1). Table A2.1 Sewage output generated in the newly developed areas Additional output Existing WWTW in U60F (ML per day) Hillcrest 4.5 Umbilo 7.5 Umhlatuzana 16.7 TOTAL 28.6 1 Urine diversion dehydration toilets are dry toilets that collect urine and faeces separately and include a special toilet seat or pan. The faeces are collected in two collection vaults for extended storage in order to dehydrate the faeces for treatment and safe handling (Rieck & von Muench 2011) EVALUATING THE POTENTIAL RETURNS TO INVESTING IN GREEN URBAN DEVELOPMENT IN DURBAN PAGE 83 Table A2.2 General Effluent Limits membrane and drainage, and would be maintained on soakaways are vertical holes dug into the ground. The Substance/Parameter General Limit Substance/Parameter General Limit an ongoing basis. maximum potential extent in the study area is therefore the total residential area (i.e. urban settlement) within Faecal Coliforms (per 100ml) 1 000 Cadmium and its compounds (mg/l) 0.005 Porous paving can be placed in commercial/retail and the catchment (Figure 3.3) and it is assumed that the Chemical Oxygen Demand (mg/l) 75 Dissolved Chromium (VI) (mg/l) 0.05 industrial areas. The potential maximum extent was maximum extent of soakaways installed at the individual pH 5.5-9.5 Dissolved Copper (mg/l) 0.01 determined based on the criteria for soil drainage, slope property level would collect 40% of the erf runoff, and depth to water table (see Table 3.1). There are as used by the eThekwini Municipality (eThekwini Total Ammonia (ionised and un-ionised) as Nitrogen (mg/l) 3 Dissolved Cyanide (mg/l) 0.02 numerous designs for porous paving and these could Municipality 2008). However due to the high cost of this Nitrate/Nitrite as Nitrogen (mg/l) 15 Dissolved Iron (mg/l) 0.300 be considered further – e.g. it could be used at a local measure, the maximum extent was taken to be half of Chlorine as Free Chlorine (mg/l) 0.25 Dissolved Lead (mg/l) 0.01 scale in business premises/shopping centres where site the treated erf area (20%). The volume of the soakaway levelling allows for changes in natural topography and pits was calculated by multiplying the area treated by a Suspended Solids (mg/l) 25 Dissolved Manganese (mg/l) 0.1 other factors such as drainage. The extent of the porous two metre depth for each pit. Electrical Conductivity (mS/m) above intake 70 to 150 mS/m Mercury and its compounds (mg/l) 0.005 paving was based on the assumption that 15% of the Ortho-Phosphate as phosphorus (mg/l) 10 Dissolved Selenium (mg/l) 0.02 total erf area for commercial/retail and industrial areas There were three levels of extent for the implementation is paved. of source controls in the catchment; none, medium or Fluoride(mg/l)) 1 Dissolved Zinc (mg/l) 0.1 maximum (Table A2. 3). Soap, oil or grease (mg/l) 2.5 Boron (mg/l) 1 Infiltration trenches need to be located close to the Dissolved Arsenic (mg/l) 0.02 source of contamination, typically in industrial and The total area of application of the different source commercial/retail areas. They are installed around controls varied with each scenario depending on parking areas and office blocks where permeable the overall extent of other GUD interventions being A2.1.2 Scenarios 3-15 (full sanitation) ƒƒ The equivalent of treated sewage is recycled if any paving is not an option because of space limitations implemented. For example, scenarios with minimum The scenarios with full sanitation assumed that informal human wastes are treated using waterborne sewage. or because of soil and infiltration limitations. Figure conservation areas and no riparian buffers had larger settlement areas and newly developed areas receive 3.3 shows the potential maximum extent of the erf areas of source control application as more buildings Neither sanitation nor wastewater recycling were costed areas that infiltration trenches can be implemented were available for implementation compared to the required access to sanitation which results in a in this study, as these are imperatives that are not under in the catchment. These are the erf areas that could scenarios with medium or maximum conservation zero net increase in WWTW outputs. In other words, scrutiny here. Our study focuses on the added value not be serviced by porous paving. Infiltration trenches areas. The total area of source controls was therefore the wastewater generated in these areas is assumed gains of investing in further green urban development vary in their size based on substrate and size of source less under the compact development scenarios (Table to be removed and recycled using “green” sanitation measures. runoff area. The assumption is that infiltration trenches A2. 4). The implementation of green roofs ranged from measures such as urine diversion dehydration toilets and the recycling of wastewater, with WWTW outputs not slowly infiltrate water received from adjacent hardened 385 – 474 ha and the maximum area of permeable augmented. This approach forms part of the green urban surfaces which was assumed to be 15% of the total erf paving was 70 ha (Table A2. 4). Infiltration trenches were development strategy and is important to consider, A2.2 Stormwater source controls area. constructed in commercial and industrial areas where especially given Durban’s current water shortages. A2.2.1 Design and extent of the different interventions permeable paving was not feasible and had a maximum Soakaway pits can be used to infiltrate runoff from volume of 0.31 million m3 under scenarios 9, 11 and 13 Under the full sanitation scenarios the augmented For each measure, we estimated the potential extent of roofs in residential areas. They collect stormwater and (Table A2. 4). The application of soakaways in residential WWTW outputs included in scenarios 1 and 2, as implementation. In some cases, where costs were very allow it to infiltrate into the surrounding soil, much areas resulted in extensive volumes, with the maximum described above, are removed. Full sanitation scenarios high and full implementation unlikely to be feasible, the like infiltration trenches. However, while infiltration volume reaching almost 52 million m3. also assumed that the compliance and monitoring maximum extent considered in the analysis was less than trenches tend to be long and narrow at the surface, of current WWTW and associated infrastructure is the potential extent. The potential extent of each source improved. Additionally, the full sanitation scenarios control measure is shown in Figure 3.3. were improved further with the placement of treatment Table A2.3 The extent of implementation of source controls wetlands to polish runoff from the existing WWTW. Green roofs can be installed on large flat roof surfaces Source control Type of building Medium extent Maximum extent compatible with commercial/retail and industrial areas. Green roofs Commercial & Industrial X X The assumptions for scenarios 3-15 were as follows: Green roofs could be implemented on the commercial/ Permeable paving Commercial & Industrial X X ƒƒ All households in the catchment, including informal retail and industrial erf areas in the catchment. The Infiltration trenches Commercial & Industrial X X settlements, have access to improved sanitation; total extent of green roofs varied under each scenario. If green roofs were implemented in conjunction with Soakaways Residential X ƒƒ Non-point source runoff from informal settlement the maximum conservation intervention, for example, areas will change to be of the same quality as runoff then the total area suitable for green roofs was less Table A2.4 The total area and volume (ha, m3) of each source control intervention for each of the source control scenarios from formal residential areas; than under the minimum conservation intervention due to the more compact development required under Scenarios with source Green roofs Infiltration trenches Soakaway pits ƒƒ Non-point source runoff from newly developed areas Permeable paving (ha) the maximum conservation scenario. For industrial and control application (ha) (million m3) (million m3) will change to be the same as the new landuse type as commercial buildings, it was assumed that the total Scenario 9 474 70 0.31 51.63 stipulated in zonation plans; roof area was equal to 40% of the erf area (eThekwini Scenario 11 474 70 0.31 - ƒƒ There will be a zero net increase in WWTW outputs; Municipality 2008). For the scenario analysis, it was Scenario 12 386 60 0.24 - assumed that the intervention would be implemented ƒƒ Human wastes will be dealt with from informal on half of the total roof area. It was assumed that the Scenario 13 474 70 0.31 51.63 settlements and newly developed residential areas green roofs would be constructed to include a light Scenario 14 385 68 0.22 39.81 using green sanitation measures such as wastewater layer of vegetation with waterproof membrane, filter Scenario 15 386 60 0.24 36.25 recycling and urine diversion dehydration toilets; and PAGE 84 EVALUATING THE POTENTIAL RETURNS TO INVESTING IN GREEN URBAN DEVELOPMENT IN DURBAN EVALUATING THE POTENTIAL RETURNS TO INVESTING IN GREEN URBAN DEVELOPMENT IN DURBAN PAGE 85 A2.2.2 Cost assumptions that these limits are generally higher than the General A2.4 PC_SWMM model assumptions The costing of GUD measures was based on data The construction costs associated with each scenario Effluent Limits used in South African WWTW permitting, collated from the green urban development literature, varied depending on the combination of source controls with the exception of TP; the South African effluent A2.4.1 Assumptions regarding untreated urban runoff in particular examples pertaining to the construction implemented and their overall extent per scenario limits are 10 mg/L for orthophosphate. quality and implementation of GUD measures in South Africa (Table 4.1 and Table 4.2). Soakaways represented the It was assumed that the wetlands were designed based Runoff from impermeable surfaces are key contributors (Armitage et al. 2013). Costs from the literature were highest construction cost with a maximum of R8 billion on the ‘required’ effluent concentrations (DWS General of sediments and pollutants to rivers and streams. all converted/inflated to 2015 Rands. Determining the for scenarios 9 and 13. The total costs associated with Standards, see DWS General Standards. In order to Characterisation of the untreated runoff quality is overall cost of implementation is difficult as there are a constructing green roofs ranged from R463 million to estimate the extent of the wetland for costing purposes, necessary for determining total nutrient and sediment number of factors that effect SUDS costing. These are R569 million. Total construction costs associated with the removal rate was used to estimate the area required loads flowing into the rivers in the catchment from discussed in more detail in Box 6.1. permeable paving and infiltration trenches was far to meet the calculated concentrations based on the different land use types and is necessary for guiding less than green roofs and soakaways, with permeable given efficiency. Therefore the outflow from each the selection of effective and efficient stormwater The unit costs for the source control measures were paving costing between R185 million and R214 adjusted based on the information provided in Box WWTW was used in conjunction with nutrient and management options. million and infiltration trenches between R86 million sediment concentrations to determine daily loads for 6.1 taking into account the evidence for significantly and R122 million (Table A2. 6). Total source control nutrients and sediments. The removal rate and removal Event Mean Concentration (EMC) data used in water reduced unit costs as a result of economies of scale. The implementation costs were highest for Scenarios 9 efficiency data provided in Table A2. 7 was then used to quality modelling includes Event Mean Concentrations source controls implemented in the catchment cover and 13. calculate the area that would be required to treat these (EMCs), derived from literature and applied in South vast areas; significantly more than the 500 properties estimated daily loads. Based on these assumptions, Africa to the Salt River Stormwater Master Plan study for and 15 ha described as large scale in other studies (see the extent of the treatment wetlands ranged from 2 ha the 1:0.5 year, 5 mm return interval storm, with data for Box 6.1). It was therefore assumed that the unit costs at the Hillcrest WWTW to 16 ha at the Umhlatuzana informal/poorly serviced high density urban settlements for green roofs and soakaways, the two interventions WWTW to 40 ha at the Umbilo WWTW. collected during storm event sampling on the Diep River implemented at the largest scale in the catchment, catchment (Cape Town) and reported in Cerfonteyn & would be reduced by up to 80% and the unit costs for Table A2.7 Annual Average Treatment Performance Capabilities Day (2010). These data can be applied to landuse across permeable paving and infiltration trenches would be for surface flow wetlands (Source: Kadlec & Knight 1995) assuming wetland influent is a “typical municipal the catchment, to derive a fully developed catchment reduced by 50% (see Table A2. 5). effluent” (see Table A2. 8). that accounts for stormwater runoff from different Table A2.5 The area, construction and maintenance costs associated Removal rate Efficiency landuse types. with implementing source controls Parameter (kg/ha/day) (% per ha/day) Table A2.9 Event Mean Concentration (EMC) data used in water Implementation cost Annual maintenance cost (% of BOD 10 67 quality modelling Source control Unit (Rands per unit) construction cost) TSS 10 67 Landuse  TSS BOD TIN Green roofs m2 120 8% NH4-N 4.7 62 description (mg/l) (mg/l) (mg/l) P (mg/l) Permeable paving m2 308 2% TIN 6.9 69 Settlement - 100 15 3.41 0.79 Infiltration trenches m3 390 3% urban TP 0.95 48 Soakaways m3 155 3% Commercial/ Metals 0.1 50 retail/ 166 9 2.1 0.37 Institutional Table A2.6 Source control implementation costs (R millions) Table A2.8 Typical municipal final effluent concentrations on which Industrial / wetland treatment performance outlined in Table 3.2 is 166 9 2.1 0.37 Scenarios with source Green roofs Permeable paving (R Infiltration trenches Soakaways Road & Rail based (after Kadlec & Knight 1995). control application (R millions) millions) (R millions) (R millions) Extractive / mg/L (ug/L) 166 9 2.1 0.37 Scenario 9 569 214 122 8 055 Utility BOD 30 Metals: Farming / Scenario 11 569 214 122 - TSS 30 Cd 10 plantations & 201 4 1.56 0.36 Scenario 12 463 185 93 - woodlots NH3-N 15 Cu 50 Scenario 13 569 214 122 8 055 Recreational TN-N 20 Pb 50 201 4 1.56 0.36 open space Scenario 14 463 209 86 6 211 TP 4 Zn 300 Settlement - Scenario 15 463 185 93 5 655 201 4 1.56 0.36 rural It was assumed that economies of scale had been Natural realised for the construction of treatment wetlands. vegetation 70 6 1.51 0.12 The unit costs were therefore staggered and decreased (D'MOSS) A2.3 Treatment wetlands with the size of the wetland being constructed. The Settlement - 497   22 6.7 informal Constructed treatment wetlands are designed to The treatment wetlands were situated at point source cost of constructing the 2 ha wetland was set at 100% improve polluted runoff and waste water effluent quality pollution outlets in the study area; at the three existing of standard unit costs (R225 per m2), the 16 ha wetland On the basis of design criteria provided in Georgia and provide some limited control of runoff volumes. WWTW. It was assumed that the runoff entering the at 50% unit cost, and for the 40 ha wetland at 20% of (2001), and assuming interventions sized as specified Wetlands can be effective in terms of removal of low wetlands was being treated to general standards and unit costs. Annual maintenance costs were assumed for the relevant volumes, the following stormwater levels of pollutants (i.e. a polishing function) and also that the wetlands further treat runoff to a specific to be 2% of construction costs. Based on the total area treatment interventions would be expected to remove provide buffering functions, aesthetic value and wildlife standard. Table A2. 7 and Table A2. 8 include the annual of 2, 16 and 40 ha for treatment wetlands, the overall the following proportions of the anticipated load – with habitat. Constructed wetlands are located as close to average treatment performance capabilities for surface construction cost was estimated to be R41 million. load being calculated based on landuse type (Table the source of pollution as possible so as to maximise the flow wetlands assuming that the wetland influent is of A2. 10). Removal efficiencies of LIDS measures were impact on improving water quality. a typical municipal final effluent concentration. Note PAGE 86 EVALUATING THE POTENTIAL RETURNS TO INVESTING IN GREEN URBAN DEVELOPMENT IN DURBAN EVALUATING THE POTENTIAL RETURNS TO INVESTING IN GREEN URBAN DEVELOPMENT IN DURBAN PAGE 87 not suitable for larger return periods and therefore water quality treatment first-order decay equations were applied to the GUD measures rather than a simple BMP efficiency. The nutrient removal efficiency of the different GUD measure was estimated in terms of the hydraulic residence time. A higher hydraulic residence time or ‘contact time’ is experienced during low flows (i.e. at the beginning and the end of the hydrograph) and will result in a higher removal rate. Table A2.10 Assumed pollutant reduction based on 1: 0.5 year event (Source: Georgia 2001) % of load in a 0.5 year storm Approach TSS TIN TP Metals Pathogens Dry detention basins 60 30 10 50 Dry enhanced swales 80 50 50 40 Wet enhanced swales 80 40 25 20 Grass channel 50 20 25 30 Gravity oil / grit separator (industrial areas) 40 54 5 Permeable paving 80 50 50 60 Stormwater pond 80 30 50 50 70 Figure A2.1 Applying source controls to a subcatchment in PC-SWMM. Revegetation / reforestation 80 25 50 50 40 Vegetated filter strip 60 20 20 40 A2.4.2.2 Detention basins The water quality treatment first-order decay equations has a specific BMP removal efficiency/treatment. The A SWMM model of flows under the ‘natural’ (original) above parameters were assumed based on information adopted for TIN and P were: process of defining the extent of source controls was land cover for the whole catchment area was executed from the Catchment Management Department at the difficult without the exact erf areas and in order to in order to estimate ‘pre-development’ flows. Ideally eThekwini Municipality, and a recently designed dry define, for example, just the roof area or permeable each detention basin would be sized based on pre- and detention pond that was constructed in the EMA. paving area, the data files had to be accessed and post-development flows at each point. Each detention changed to include the relevant source control and basin was incorporated into the SWMM model by A total of 23 detention ponds were placed at various then imported back into the model. The source control adding a storage node (i.e. converting a junction to a locations within the study area (Figure A2. 3). In practice, was therefore represented by its own subcatchment. storage node) with a standardised capacity of 18 000 detention basins are strategically placed to make use of where TINout and Pout are the outlet concentrations, m3. A 0.2 m diameter outlet pipe was added to allow the natural landscape to minimise construction costs. An For commercial/retail and industrial areas and where TIN and P are the inlet concentrations, is a fitting for the detention pond to drain and for low flows to example of this is the detention basin constructed by the these source controls were specified, it was assumed parameter equal to -0.05 (seconds) and is the flow through. The berm height was set at 3m, allowing EM in the Hillcrest area, as shown in Figure A2. 4. that 20% of the area within each subcatchment was hydraulic residence time (seconds). very high flows to overtop this (Figure A2. 2). The assigned to green roofs and 15% to either permeable The TSS concentration depends on a number of factors paving or infiltration trenches (all depending on the including the settling velocity, water depth and time. allocated placement). For residential areas, 20% of the impervious area was treated using soakaway pits. After source control placement the “percent impervious” and “width” properties of the altered subcatchments were adjusted to compensate for the amount of the original Where TSSout in the outlet concentration, TSS is the inlet subcatchment area that was replaced by source controls concentration, is a fitting parameter related to the (see Figure A2. 1). Both surface and drain outflows from settling velocity and is equal to -0.001 m/s, is the the source controls were routed to the same outlet time step and is the water depth (SWMM, 2009). location assigned to the parent subcatchment. Infiltration trenches and porous paving were treated A2.4.2 Assumptions for stormwater management as pervious surfaces and the Green-Ampt infiltration interventions equations were applied. The dimensions of soakaway A2.4.2.1 Source controls pits, green-roofs and infiltration trenches were taken from SWMM (2009), van Niekerk (2011) and Armitage et The general approach for setting up source controls al. (2013). in PC-SWMM is to define a new landuse that is used exclusively for a source control subcatchment and Figure A2.2 Profile of a detention basin showing the berm and the outlet pipe. PAGE 88 EVALUATING THE POTENTIAL RETURNS TO INVESTING IN GREEN URBAN DEVELOPMENT IN DURBAN EVALUATING THE POTENTIAL RETURNS TO INVESTING IN GREEN URBAN DEVELOPMENT IN DURBAN PAGE 89 A2.4.2.3 Treatment wetlands A2.4.2.4 River buffers Treatment wetlands were located at the outlets of each The dynamics of sediment and nutrient reduction of the three WWTWs in order to improve the quality through the use of river buffers are not fully understood of WWTW effluents entering the river systems. It was and are complex to model. In order to achieve the most assumed that effluent flow from the WWTWs all passed realistic representation of introducing river buffers through the treatment wetland. Evaporative loss was into the model for the required scenarios discharge not taken into account as it was not expected to have a outlet points for each subcatchment were linked to significant effect on the model results. nodes in the drainage systems flowpaths. Generally, overland flow (sheet flow) converges to become channel In the scenarios “with sanitation” it was assumed flow fairly rapidly before entering the main flowpath. that the WWTW effluent entering the wetlands had Therefore, the treatment equations for TSS, TIN and P, as been treated to DWS general standards (see Table provided in section A2.4.1, were applied to the flowpath A2.11). Measured effluent concentrations generally nodes, in order to account for pollutant reduction complied with general effluent limits, except for TSS associated with river buffers. In addition, the hydraulic, concentrations measured at Hillcrest WWTW. The latter soil and water quality parameters were altered to was reduced in the “with sanitation” scenarios from account for the change in landuse. This approach may the mean of 48 to 30 mg/L to meet the general effluent overestimate the reduction due to an accumulative limits. The efficiency ratio was applied to estimate the effect, but the implementation of river buffers would pollutant concentration below the treatment wetland result in the encroachment of vegetation within the river (Table A2. 11) and the new value was applied at each and therefore increased contact area with streamflow. point source. Therefore, this approach is assumed to best account for the overall objective of assessing the impacts of river buffers. Table A2.11 Assumed treated pollutant concentrations of wetland effluent based on removal efficiencies given in Table 3.2 Figure A2.3 Layout of modelled detention basins, shown as green squares, within the U60F quaternary catchment. WWTWs Mean Effluent Wetland Effluent Concentrations General Concentrations (mg/L) (mg/L) Removal Effluent rate (kg/ Efficiency Limit Umhlatu- Umhlatu- Pollutant ha/day) (%) (mg/L) Umbilo zana Hillcrest Umbilo zana Hillcrest TSS 10 67 30 21 7 30 6.93 2.31 9.90 TIN 6.9 69 18 7.9 7.3 6.7 2.45 2.26 2.08 TP 0.95 48 10 3.4 2.1 2.7 1.77 1.09 1.40 Figure A2.4 Example of in-situ detention basin constructed by the EM in the Hillcreast region of the U60F subcatchment. PAGE 90 EVALUATING THE POTENTIAL RETURNS TO INVESTING IN GREEN URBAN DEVELOPMENT IN DURBAN EVALUATING THE POTENTIAL RETURNS TO INVESTING IN GREEN URBAN DEVELOPMENT IN DURBAN PAGE 91 This page intentionally blank. PAGE 92 EVALUATING THE POTENTIAL RETURNS TO INVESTING IN GREEN URBAN DEVELOPMENT IN DURBAN APPENDIX 3. FLOOD MODELLING A3.1 Model setup A3.1.1 Baseline information, software and GIS layers The modelling of landuse change in the catchment Software: required developing a comprehensive complex hydraulic ƒƒ QGIS; surface runoff model. The integration of several GIS layers and post-processing for the hydraulic model input ƒƒ US-EPA SWMM5 interfaced by the PCSWMM GUI; parameters was required for the setup of the model. A summary of the baseline information, software and GIS ƒƒ HecRAS, Hec-Geo-RAS; and landuse layers used in developing the model is provided below. ƒƒ Anaconda – Spyder – Python 2.7 – Data Analysis/ Management. Baseline information: GIS landuse layers: ƒƒ Various GIS datasets from the eThekwini Municipality, including: 0.5m Rasta .IMG files for surface elevation, ƒƒ GIS landuse files (e.g. zoning files, landcover and 2m Contour, Landuse Zonal, Durban Metro Open D’MOSS) collated and reviewed, and concatenated Space System (D’MOSS), High resolution Aerial into one consistent landuse polygon shapefile imagery; ƒƒ Post-processing was conducted to dissolve ƒƒ Shuttle Radar Topography Mission (SRTM) 30m descriptions into a common set of landuse resolution surface elevation data; conventions. The final shapefile was ground truthed using aerial imagery for the EMA ƒƒ Soil type classification maps; ƒƒ Original landuse descriptions from the D’MOSS ƒƒ Geometric HecRAS hydraulic files for EMA rivers landuse file were maintained throughout the landuse (where available); description process and were classified as “Nature and Conservation Areas” (Figure A3. 1) ƒƒ Stormwater networks (where available); ƒƒ The D’MOSS classifications were further discretised to ƒƒ Relevant point source data (e.g. WWTW); provide an indication of the hydraulic parameters required for the hydraulic model, i.e. grassland could ƒƒ Design Rainfall Estimation (HydroRisk, http://ukzn- be described as ‘open_grass’ or ‘open_grass_soil’ iis-02.ukzn.ac.za/unp/beeh/hydrorisk); which indicates a higher soil erodibility (Figure A3. 1) ƒƒ eThekwini Design Rainfall (Smithers 2002); and ƒƒ Water quality parameters and landuse change shapefiles. EVALUATING THE POTENTIAL RETURNS TO INVESTING IN GREEN URBAN DEVELOPMENT IN DURBAN PAGE 93 Figure A3.2 Information required to delineate subcatchments: topographical aerial survey, DEM and river centre lines based on river flow paths Figure A3.1 Landuse Categories and D’MOSS subcategories for all Conservation Areas within the EMA A3.1.2 Subcatchment delineation and flow lines The study area was divided into subcatchments and and bridges), but they did not include the stormwater the outlet points were identified (subcatchment runoff network. Flow paths simulated using a watershed is routed to a single discharge point). Outlet points delineation tool (WDT) were appended to the HecRAS can be defined as nodes of the drainage system or files in order to represent the required study area. they can be routed to other subcatchments. The GIS subcatchment (watershed basins) data derived from EM A shapefile of the stormwater networks was provided flood studies and are in the order of 1km2 and larger. by the eThekwini Municipality. The current available Although appropriate for flood studies, the information stormwater shapefile was incomplete and contained relevant to this scope of works required discretisation numerous errors and inconsistencies (see Figure A3. into smaller, more appropriate subcatchments, in the 3). Invert levels and pipe sizes were often missing and order of 0.2km2 (Figure A3. 2). These new subcatchments connections were incorrect and/or missing. Available were processed from high resolution, .IMG raster files networks were amended where possible, i.e. a standard (DEM files). The raster files were converted to .flt float circular pipe size of 0.375m was allocated to pipes with files which can be used as a TIN (Triangulated Irregular missing geometry and tools were applied to either fill in Network). A spatial analysis tool was then used to missing invert levels (from the DEM) or to apply slopes process out the flow paths, watershed boundaries, and within the network. Where necessary, main pipelines river centre lines. were added to these networks. The pipe profiles were later checked to ensure reasonable slope gradients and The flood models, based on geometric HecRAS files, the continuity of flows. were imported into the PCSWMM model. These files contained some stormwater infrastructure (e.g. culverts Figure A3.3 Current available stormwater shapefile (the red lines represent the flow paths, yellow lines are stormwater conduits and blue dots represent stormwater junctions). PAGE 94 EVALUATING THE POTENTIAL RETURNS TO INVESTING IN GREEN URBAN DEVELOPMENT IN DURBAN EVALUATING THE POTENTIAL RETURNS TO INVESTING IN GREEN URBAN DEVELOPMENT IN DURBAN PAGE 95 The eThekwini Municipality at the time of modelling new shapefiles were imported into the model and Table A3.2 Hydraulic input properties required for each subcatchment was carrying out a Stormwater Management System connected to existing stormwater networks and flow Hydraulic Parameter Description Source (SMS) audit of all stormwater infrastructure. The audit paths. Continuity in these networks were checked and Area (ha) Area of subcatchment GIS tool entailed a visual inspection and assessment of all errors/connections were corrected where necessary. Width (m) Width of overland flow path GIS tool stormwater infrastructure from which shapefiles of The original stormwater shapefile was merged with the existing junctions, stormwater pipes and culverts were new SMS shapefiles and both were connected to the Flow Length (m) Length of overland sheet flow GIS tool generated. Where available (refer Figure A3. 4), these HECRAS and WDT flow paths. Slope (%) Average slope along the pathway of overland flow to inlet locations. GIS tool Imperv. (%) Percent impervious area RGB colour extraction N Imperv Manning’s roughness coefficient, N, for overland flow for impervious area. Rossman 2015 (Table 5) N Perv Manning’s roughness coefficient, N, for overland flow for pervious area. Rossman 2015 (Table 5) Dstore Imperv (mm) Depth of depression storage on impervious areas ASCE 1992 (Table 4) Dstore Perv (mm) Depth of depression storage on pervious areas ASCE 1992 (Table 4) SWMM default setting of 25% Zero Imperv (%) Percent of impervious area with no depression storage based on literature Percent Routed (%) Percent of runoff routed between sub-areas Outfalls Subcatchment areas were measured and the width of the subcatchment defined as the physical width of the overland flow. In an idealised, rectangular catchment, the total width would be twice the length of the drainage channel (assuming both sides of the subcatchment were symmetrical). The most significant input hydraulic parameter is the percentage of impervious area (Imperv. %). There are a number of methods that can be employed to estimate the percent imperviousness of a subcatchment. Ideally the percent imperviousness could be measured accurately from aerial photos or land use maps, however, this can be tedious for large study areas. Two approaches were investigated: 1) a percent impervious area associated with each landuse category based on Figure A3.4 Newly available stormwater network shapefiles from the EM’s SMS audit. standard values for different landuses found in the literature and 2) RGB colour extraction tool applied to differentiate between impervious and pervious areas based on aerial imagery (in the EMA) and Google Earth A3.1.3 Point sources A3.1.4 Hydraulic parameters images (outside of the EMA). The first approach was Average daily abstractions and return flows/discharges The determination of the catchment characteristics were used for this study. Figure A3. 5 shows sections of two were added as point sources at the appropriate estimated using a spatial analyst tool for zonal statics. different areas of contrasting landuse i.e. residential junctions. A list of the wastewater treatment works Raster files were generated to represent the following and industrial. The top value represents the %Imperv (WWTWs) located within the catchment are given in information required for the hydraulic and hydrological using approach 1 and the bottom value represents Table A3. 1. models, with reference to each subcatchment. These the %Imperv using approach 2. The estimation using were used to estimate the many runoff characteristics approach 2 was reasonable except where the colour outlined in Table A3. 2. spectrum was a mixture of green and brown, for Table A3.1 WWTWs located within the study area. example, recently harvest sugarcane and rural sandy areas. WWTW Name Latitude Longitude Design Capacity (Ml/d) Umhlatuzana -29.87713 30.884036 14.8 Estimates of Manning’s roughness coefficient (N values) Umbilo -29.84561 30.891653 23.2 for overland flow, imperviousness and perviousness Hillcrest -29.79410 30.75635 1.2 were taken from literature. A summary from three different sources are given In Table A3. 3. PAGE 96 EVALUATING THE POTENTIAL RETURNS TO INVESTING IN GREEN URBAN DEVELOPMENT IN DURBAN EVALUATING THE POTENTIAL RETURNS TO INVESTING IN GREEN URBAN DEVELOPMENT IN DURBAN PAGE 97 Table A3.3 Estimates of Manning’s roughness coefficient (N values) adjust runoff volumes. Therefore obtaining accurate for overland flow. A summary from three different sources (Source: Rossman 2015) values in the setup may be unnecessary as these value may change during calibration. Depression storage is Source Ground Cover n Range most sensitive for small storms; as the depth increases it Smooth asphalt 0.01 becomes a smaller component of the water budget (EPA, Asphalt and concrete paving 0.014 2015). Crawford Packed clay 0.03 and Table A3.4 Values used for the depression storage based on landuse Linsley Light turf 0.20 (ASCE, 1992) (1966)a Dense turf 0.35 Landuse Depression storage Dense shrubbery and forest Impervious surfaces 1.3 - 2.5 mm 0.4 litter Concrete or asphalt 0.011 0.010-0.013 Lawns 2.5 - 5 mm Bare sand 0.010 0.01-0.016 Pasture 5.0 mm Gravelled surface 0.02 0.012-0.03 Forest Litter 9.6 mm Engman Bare clay-loam, (eroded) 0.02 0.012-0.033 (1986)b Range (natural) 0.13 0.01-0.32 Bluegrass sod 0.45 0.39-0.63 A3.1.5 Soil infiltration Short grass prairie 0.15 0.10-0.20 The largest proportion of rainfall losses over Bermuda grass 0.41 0.30-00.48 pervious areas generally occur due to soil infiltration. Theoretically the Richards equation is the most Smooth asphalt pavement 0.012 0.010-0.015 representative, however its highly nonlinear partial Smooth impervious surface 0.013 0.011-0.015 differential equations make is unsuitable for continuous Tar and sand pavement 0.014 0.012-0.016 long-term simulations. Simpler algebraic infiltration Concrete pavement 0.017 0.014-0.020 models have been developed that represent Rough impervious surface 0.019 0.015-0.023 the dependence of infiltration capacity on soil Smooth bare packed soil 0.021 0.017-0.025 characteristics and the present soil capacity during a storm event. There are five options that can be used in Moderate bare packed soil 0.030 0.025-0.035 SWMM, namely Horton’s method, the modified Horton Rough bare packed soil 0.038 0.032-0.045 method, the Green-Ampt method, the modified Green- Gravel soil 0.032 0.025-0.045 Ampt method and the Curve Number method. With all Mowed poor grass 0.038 0.030-0.045 of these models, the parameters depend on the type Yen Average grass, closely clipped and condition of the soil of interest. 0.050 0.040-0.060 (2001)c sod Pasture 0.055 0.040-0.070 It is worth noting that the Flood Line Delineation studies for EM use the Soil Curve Number (SCN) to represent Timberland 0.090 0.060-0.120 the runoff co-efficient for catchment routing. Although Dense grass 0.090 0.060-0.120 suitable for flood studies (as a conservative approach), Shrubs and bushes 0.120 0.080-0.180 this investigation will use the Green-Ampt method. This Business land use 0.022 0.014-0.035 methods provides a soil memory as opposed to a broad Semi-business land use 0.035 0.022-0.050 brush coefficient approach. Industrial land use 0.035 0.020-0.050 For the Green-Ampt infiltration method, the model Dense residential land use 0.040 0.025-0.060 requires three soil parameters that the user must specify Suburban residential land use 0.055 0.030-0.080 for each of the subcatchments: Parks and lawns 0.075 0.040-0.120 1. Capillary suction head, Ψs (mm); a Obtained by calibration of Stanford Watershed Model b Computed by Engman (1986) by kinematic wave and storage 2. Saturated hydraulic conductivity, Ks (mm/hr); and analysis of measured rainfall-runoff data 3. The maximum available moisture deficit, θdmax c Computed on basis of kinematic wave analysis (volume of dry voids per volume of soil). The depression storage is the volume that must be filled prior to the occurrence of runoff on both pervious and impervious areas. Values for depression storage were taken from the SWMM Manual (EPS 2015 after ASCE, Figure A3.5 Percent impervious area for two different areas of contrasting landuse. The top value represents the %Imperv using 1992; Table A3. 4). In SWMM, depression storage may approach 1 and the bottom value represents the %Imperv using approach 2. be treated as a calibration parameter, particularly to PAGE 98 EVALUATING THE POTENTIAL RETURNS TO INVESTING IN GREEN URBAN DEVELOPMENT IN DURBAN EVALUATING THE POTENTIAL RETURNS TO INVESTING IN GREEN URBAN DEVELOPMENT IN DURBAN PAGE 99 These parameters were taken from the SWMM Manual A3.1.6 Storm design events (Table A3. 5). Figure A3. 6 is a map of the green- The eThekwini Municipality’s Design Storm Generator ampt parameters for the EMA that has recently been was used to determine the distribution of rainfall developed and applied during a current Water Research for different return periods, i.e. a 2-year, 5-year, 10- Commission (WRC) study by Pegram and Sinclair at the year and 20-year design storm. The EM Design Storm University of KwaZulu-Natal (UKZN). These parameters Generator uses a SCS Type II distribution (Figure A3. will be used and the results compared with those using 7) based on historical rainfall data (approximately 20 the parameters given in Table A3. 5. years) to generate a synthetic time distribution of rainfall intensity. The EM procedure of using a 24-hour rainfall Table A3.5 Soil parameters (Source: Rawls 1983) depth for selected return periods was followed. The resultant design rainfall hyetograph was input into the Hydraulic Suction Head Conductivity Initial Deficit Porosity Field Capacity Wilting Point SWMM models using a specified 5-minute interval. Soil Texture Class (mm) (mm/hr) (fraction) (fraction) (fraction) (fraction) Sand 49.02 120.34 0.413 0.437 0.062 0.024 A3.1.7 U60F model Loamy Sand 60.96 29.97 0.39 0.437 0.105 0.047 Sandy Loam 109.98 10.92 0.368 0.453 0.19 0.085 The final model for catchment U60F is shown in Figure A3. 8. Note the denser flowpaths where stormwater Loam 88.9 3.3 0.347 0.463 0.232 0.116 Figure A3.7 SCS 24-hour rainfall distributions (not to scale) (Source: network data were available. The given HEC-RAS model SCS 1984) Silt Loam 169.93 6.6 0.366 0.501 0.284 0.135 and stormwater networks are shown in yellow and Sandy Clay Loam 219.96 1.52 0.262 0.398 0.244 0.136 additional tributaries (in red and green) were added in Clay Loam 210.06 1.02 0.277 0.464 0.31 0.187 using the watershed delineation tool. Silty Clay Loam 270 1.02 0.261 0.471 0.342 0.21 Sandy Clay 240.03 0.51 0.209 0.43 0.321 0.221 Silty Clay 290.07 0.51 0.228 0.479 0.371 0.251 Clay 320.04 0.25 0.21 0.475 0.378 0.265 Figure A3.8 Snapshot of the PCSWMM model for catchment U60F. Figure A3.6 Map of Green-Ampt Parameters developed by UKZN (Source: Sinclair 2015) PAGE 100 EVALUATING THE POTENTIAL RETURNS TO INVESTING IN GREEN URBAN DEVELOPMENT IN DURBAN EVALUATING THE POTENTIAL RETURNS TO INVESTING IN GREEN URBAN DEVELOPMENT IN DURBAN PAGE 101 A3.2 Points of interest for scenario analysis The Umhlatuzana and Umbilo Rivers and other stormwater outfalls within the U60F subcatchment discharge directly into Durban Harbour. In the SWMM model, there are 79 outfalls in total (Figure A3. 9). The peak flows and flow volume were estimated for each scenario for a 2, 5, 10 and 20-year return period. The effectiveness of GUD measures was not estimated for floods above a 20-year return period, as it is known that GUD interventions have little impact on larger floods. Figure A3.10 Thiessen polygons determined for the available rain gauges relevant to the U60F catchment. A3.3.2 Design rainfall generation for U60F Several real time rainfall data sets were applied in order The eThekwini Municipality’s Design Storm Generator to calibrate the models. was used to determine the distribution of rainfall for different return periods, i.e. a 2-year, 5-year, 10-year and A3.3.4 Calibration of flows and water levels 20-year design storm. The EM Design Storm Generator There are currently no measured flow or water level uses a SCS Type II distribution based on historical rainfall data available for the Umbilo and Umhlatuzana Rivers. data (approximately 20 years) to generate a synthetic Emphasis was, therefore, placed on setting up the model time distribution of rainfall intensity. The EM procedure as accurately as possible. Peak flows were compared of using a 24-hour rainfall depth for selected return with peak flows estimated by Jezewski et al. (1984) and periods was followed. The resultant design rainfall Makwananzi & Pegram (2004) who used a HEC-HMS hyetograph was input into the SWMM models using a model (Table A3. 6). The peak flows estimated in the specified 5-minute interval. current study differ to those estimated by Makwananzi Figure A3.9 Aerial image of all of the Durban Harbour outfalls (red triangles). The stormwater network is shown by the yellow lines. and Pegram (2004). Various methods to decrease the A3.3.3 Available measured data flood peaks were employed, however without measured The efficacy of calibration depends entirely on the data this was approached with caution. It is important to availability of measured data. Flow and/or water level note that due to time constraints for this study, the time A3.3 Model calibration data are most useful for the calibration of the hydraulics. allocation for calibration was limited. Real-time data were used for model calibration and A3.3.1 Rainfall selection and application validation. This was done using the data and methods Phase 1 of the calibration and validation was focused on Table A3.6 Data used for calibration of flows and water levels outlined below. While river flows within the EMA tend the hydraulic flows and volumes. Real-time rainfall data Makwananzi & Pegram (2004) to be relatively well monitored, there was no flow data for the EMA were obtained from the EM database. The Source Jezewski (1984) eThekwini Reports CCS Consulting (2016) available for the U60F quaternary catchment. Thiessen polygon method was applied to the rainfall Peak Discharge Q (m /s) for 3 Peak Discharge Q (m3/s) for stations in and around the U60F subcatchment and each each return period each return p eriod Area Area rain gauge was assigned to a certain area (Figure A3. 10). River Area (km2) Tc (h) (km2) 2yr 5yr 10yr 20yr (km2) 2yr 5yr 10yr 20yr Durban Bay 242 6.1 368 649 860 1106 Umbilo 80 79 198 316 451 74 107 209 271 347   Umhlatuzana 118 127 244 334 457 126 222 372 665 763 PAGE 102 EVALUATING THE POTENTIAL RETURNS TO INVESTING IN GREEN URBAN DEVELOPMENT IN DURBAN EVALUATING THE POTENTIAL RETURNS TO INVESTING IN GREEN URBAN DEVELOPMENT IN DURBAN PAGE 103 The EM recently deployed a number of water sensors of baseflows which helped to validate the accuracy in the U60F catchment (in the Umhlatuzana and Umbilo of abstractions and return flows in these rivers. The Rivers) over two separate time periods. However, sensitivity of the model to various hydraulic parameters unusually heavy rainfall was experienced during both was tested using the SRTC tool on PC-SWMM. The periods (including an almost 100-year flood) and most sensitivity of the results was tested for the various of the sensors were washed away. Two sensors were parameters (Table A3. 7). The results were most retrieved - one in the Umbilo River and one in the sensitive to changes in the depression storage and % Umhlatuzana River. This data provided an indication imperviousness. Table A3.7 Sensitivity of runoff volume and peak flow to surface runoff parameters (EPA, 2015). Effect of Typical effect on Effect of increase increase on Parameter hydrograph on runoff volume runoff peak Comments Area Significant Increase Increase Less effect for a highly porous catchment Less effect when pervious areas have low infiltration Imperviousness Significant Increase Increase capacity For storms of varying intensity, increasing the width tends to produce higher and earlier hydrograph peaks, Width Affects shape Decrease Increase a generally faster response. Only affects volume to the extent that reduced width on pervious areas provides more time for infiltration. Same as for width, but less sensitive, since flow is Slope Affects shape Decrease Increase proportional to square root of slope. Roughness Affects shape Increase Decrease Inverse effects as for width. Depression Moderate Decrease Decrease Significant effect only for low-depth storms. storage A3.4 Assumptions and limitations A3.4.3 Stormwater network data A number of assumptions were made during the setup The current stormwater network data were inconsistent of the SWMM model where input data were either and incomplete. Only certain areas of the current SMS insufficient or unreliable. These assumptions may be audit have been completed and were included in the regarded as limitations of the model and therefore model, however inconsistencies and errors were also should be considered when analysing the results. found in these networks. Missing data were entered based on the following assumptions: A3.4.1 The use of design rainfall ƒƒ pipe sizes: a default value of 0.375 m Note that design rainfall assumes that rainfall is equally ƒƒ invert levels: levels were taken from the DEM and the distributed over the whole catchment at the same time. profile was altered in order to acquire a reasonable Realistically, a specific design rainfall does not imply an slope equal design runoff, however this is general practice when performing flood studies. The model was run numerous times in order to resolve flooding and continuity issues resulting from problems with these data. Invert levels were manually adjusted in A3.4.2 Groundwater and baseflows order to correct negative slopes. Groundwater was not included in the modelling. Insufficient data were available to incorporate any accurate representation of groundwater flows. Therefore, measured flow/water level data was used to infer the groundwater as baseflow. Groundwater inputs vary seasonally, therefore the ‘baseflow’ was estimated during summer and winter periods and incorporated into the model as a time series where possible. PAGE 104 EVALUATING THE POTENTIAL RETURNS TO INVESTING IN GREEN URBAN DEVELOPMENT IN DURBAN APPENDIX 4: INFRASTRUCTURE COST ESTIMATE METHOD A4.1 Overview of concrete was estimate at R2400/m3, shuttering was assumed to be R650/m2 and the supplying and fixing This section describes the cost estimation method. The of steel was taken as R12000/t. Concrete blinding was method estimates the existing infrastructure costs from estimated as R1600/m3. A 10-20% allowance on the their dimensions. It then uses a scaling relationship total cost was provided for preliminary and general between flow and the infrastructure dimensions to items (P&G) such as site establishment and supervision. estimate the stormwater requirements under the The individual rates for each infrastructure category are different land use scenarios. The infrastructure required summarised in additional information. to satisfy the various scenarios are then costed. The cost difference between the existing infrastructure and the scenario infrastructure is indicative of the value of the A4.4.1 Bridges natural areas. Bridges have two subcategories: bridge culverts and pipe culverts. Bridges differ from the culvert and pipe categories as they are positioned within watercourses. A4.2 Identifying existing infrastructure To allow for the complications of dealing with water the An inventory of all the stormwater infrastructure is P&G was set to 20%. With the exception of not having identified and categorised into four major categories: manholes the bridge culverts and the pipe bridges bridges; canals; culverts and pipe networks. The bridges are priced the same as the culverts and the pipes are divided into a further two subcategories: bridge respectively. culverts and bridge pipes. The bridge category excludes major bridges as their size is insensitive to flows. A4.4.2 Culverts Culverts are defined as any non-circular structure not A4.3 Assigning rainfall return periods acting as a bridge. The culvert costs are estimated Each infrastructure category is assigned a design return from the cross-sectional area and the length. It is period based on the eThekwini Design Guidelines assumed that all the culverts are constructed from 0.3 (eThekwini Municipality 2008). Table A4. 1 shows m thick insitu reinforced concrete with steel reinforcing the return periods associated with the relevant attributing to 4% of the total volume. It is assumed that infrastructure category. The design rainfalls are then all the ground conditions are the same and that the modelled to estimate the peak flows for each structure. structures are founded on 200 mm of concrete blinding. Excavation quantities are based on 600 mm of cover Table A4.1 Return periods assigned to each of the infrastructure categories and a payment width as defined in Clause 5.2 of SANS 1200DB. Manholes were priced as R25 000 each and one Category Return Period (years) was assumed every 60 m. Bridges Culverts 20 Bridges Bridges Pipes 20 Canals 10 A4.4.3 Canals Culverts 5 Canals are priced the same as culverts except there are Pipes 2 no roof slabs, cover material or manholes. A4.4 Cost estimate of the infrastructure A4.4.4 Pipes Costs are estimated for each of the four categories as The pipe costs are calculated similarly to the culvert each contains different assumptions. Material costs are costs. It is assumed that all the ground conditions are based on 2016 prices with delivery to central Durban. the same and that a Class B bedding (Drawing LB-1, All prices include a 10% mark-up and exclude value SANS 1200LB) is used throughout. Excavation quantities added tax (VAT). Labour rates are legislated for the are based on 600 mm of cover and a payment width as Civil Construction industry and were taken as R27/hr. defined in Clause 5.2 of SANS 1200DB. Manholes were All infrastructure was assumed to be under road ways priced as R25 000 each and one was assumed every and the reinstatement was estimated to be R420/m2. 60 m. Excavation was taken as R90/m3, selected backfill as R80/m3 and backfill at R60/m3. The supply and placing EVALUATING THE POTENTIAL RETURNS TO INVESTING IN GREEN URBAN DEVELOPMENT IN DURBAN PAGE 105 A4.5 Flow vs dimension relationship Table A4.2 Construction rates applied to the bridge culvert category Table A4.5 Construction rates applied to the pipe category To estimate the changes in cost a relationship between Description Unit Rate Description Unit Rate flow and the infrastructure dimensions needs to be Excavation m3 R 90.00 Pipe m Refer to Table A4. 6 established. These flow relationships can be established Selected fill m3 R 80.00 Excavation m3 R 90.00 theoretically for diameter and area from uniform flow where n, the Mannings roughness, is assumed to Backfill m3 R 60.00 Bedding m3 R 320.00 conditions. The relationship is then used to scale the represent concrete at a value of 0.015. a is the cross- Blinding m3 R 1 600.00 Selected fill m3 R 80.00 infrastructure dimensions for the different scenario sectional area, P is the perimeter and S0 is the slope of Shuttering m2 R 650.00 Backfill m3 R 60.00 flows. A scaling relationship exists for open channel the infrastructure. flows and pressurised flows. Both of these flow types are Steel t R 12 000.00 Manholes m2 R 25 000.00 estimated for each flow scenario. If the scenario flow, Q, is less than the threshold flow Concrete m3 R 2 400.00 Reinstatement m2 R 420.00 the open channel scaling is used. If the scenario flow Reinstatement m2 R 420.00 Preliminary and general items % 10 exceeds the threshold flow then the pressurised scaling Preliminary and general items % 20 A4.5.1 Pipe scaling is used. The scaling relationship for pipe diameters in open Table A4.6 The cost of pipes in 2016 delivered to central Durban Two other conditions are included to ensure that the channel flow is Table A4.3 Construction rates applied to the bridge pipe category Diameter Diameter scaling does not artificially inflate the benefit of the (m) Rate (R/m) (m) Rate (R/m) natural areas. If the scenario flow does not exceed the Description Unit Rate 0.3 313.50 0.9 1383.80 threshold flow and the existing flow then no scaling is Pipe m Refer to Table A4. 6 applied. If the scenario flow exceeds the threshold flow 0.375 424.60 1.05 1777.60 Excavation m3 R 90.00 but does not exceed the existing flow then the open Bedding m3 R 320.00 0.45 555.50 1.2 2284.70 and the scaling relationship for pressurised flows is channel scaling is applied. These conditions ensure that 0.525 641.30 1.35 2649.90 Concrete m3 R 2 400.00 scenario flows that are larger than the existing flows but 0.6 1136.30 1.5 3634.40 that do not require larger infrastructure are not scaled. Reinstatement m2 R 420.00 0.75 1060.40 1.8 5195.30 This means that artificial benefits are not attributed to Preliminary and general items % 10 the natural areas. It also ensures that the cost of over 0.825 1204.50 design and future capacity are not penalised. where D is the scaled diameter, D0 is the existing pipe Table A4.4 Construction rates applied to the canal and culvert diameter, Q is the scenario flow and Q0 is the flow under The followings is a summary of all the conditions category the existing conditions (status quo). relevant to the scaling: Description Unit Rate Excavation m3 R 90.00 Condition 1: A4.5.2 Culvert scaling Selected fill m3 R 80.00 Backfill m3 R 60.00 The scaling relationship for culvert area in open channel flow is Blinding m3 R 1 600.00 Condition 2: Shuttering m2 R 650.00 Steel t R 12 000.00 and the scaling relationship for pressurised flows is Concrete m3 R 2 400.00 Reinstatement m2 R 420.00 Condition 3: Preliminary and general items % 10 where A is the scaled area, A0 is the existing culvert area, Q is the scenario flow and Q0 is the flow under the existing conditions (status quo). The culvert width is A4.6 Cost comparison then determined by dividing the scaled area, A, by the The difference between the existing infrastructure costs culverts original height. and the scenario infrastructure costs are the indicative value of the natural areas. A4.5.3 Estimating the flow type To determine which scaling relationship is to be used A4.7 Additional information the flow type needs to be estimated. The flow type is The items included for the costing of each category estimated by calculating the infrastructure’s maximum are shown in Table A4. 2, Table A4. 3, Table A4. 4, open channel flow from the Manning’s equation. Table A4. 5. Table A4. 6 shows the linear meter cost of This flow is referred to as the threshold flow and it is concrete stormwater pipes. calculated from PAGE 106 EVALUATING THE POTENTIAL RETURNS TO INVESTING IN GREEN URBAN DEVELOPMENT IN DURBAN EVALUATING THE POTENTIAL RETURNS TO INVESTING IN GREEN URBAN DEVELOPMENT IN DURBAN PAGE 107 This page intentionally blank. PAGE 108 EVALUATING THE POTENTIAL RETURNS TO INVESTING IN GREEN URBAN DEVELOPMENT IN DURBAN APPENDIX 5. SEDIMENT AND NUTRIENT MODELLING A5.1 Model setup to the different landuse categories across the study area. The data below can be applied in determining the The water quality parameters assessed during the water quality volume to be catered for in stormwater scenario analysis were nitrogen, phosphorous and total management devices (e.g. source controls). suspended solids (TSS). The landuses that generate these pollutants were defined and the pollutant build- The TSS load estimated in PC-SWMM accounts for the up, pollutant washoff and street cleaning parameters total suspended sediments generated due to catchment were assigned to each landuse. The pollutant removal runoff and does not account for the proportion of functions for nodes within the drainage system that sediment transport activated from the river bed (i.e. the contain storage/treatment facilities were also defined. bedload). While bedload transport is the dominant mode The input parameters for each pollutant are as follows: for low velocity flows and/or large grain sizes, suspended load transport is the dominant mode for high velocity ƒƒ the pollutant name; and/or fine grain sizes (Chadwick et al., 2013). In South ƒƒ the concentration units (i.e. mg/L, μg/L, counts/l); Africa, a factor of 1.25 is generally applied to cater for bed load and non-uniformity in suspended sediment ƒƒ concentration in rainfall; concentrations in order to estimate the mean annual sediment load (Msadala et al. 2010, after Rooseboom ƒƒ concentration in groundwater; 1992). Cooper (1993) referenced estimates of the proportion of bedload in KwaZulu-Natal rivers from ƒƒ concentration in direct infiltration/inflow; and other studies to range from 12 to 50%. ƒƒ first-order decay coefficient. A5.1.1 Rainfall Note that no data was available to estimate the pollutant build-up and street cleaning parameters and therefore Real-time rainfall data was applied to the models. The these features were not considered. The pollutant simulations were run from 1 August 2013 until 30 July washoff from a given land use occurs during periods of 2014. Note that sediment yields vary spatially and wet weather and can be characterized in SWMM5 by temporally and therefore a one year simulation is not either using an exponential or rating curve relationship. indicative of the mean annual sediment yield. The annual The Event Mean Concentration is a case of Rating Curve rainfall (572 mm for Durban city central) experienced Washoff where the exponent is 1.0 and the coefficient during this period was below the MAP of Durban (1000 represents the washoff pollutant concentration in mm) and therefore simulated results are conservative. mg/L. In each case build-up is continuously depleted as Note that Rooseboom & Lotriet (1992) suggest that six washoff proceeds, and washoff ceases when there is no years of continuous monitoring is required to obtain a more build-up available. The EMCs were derived from reasonable estimate of the average sediment load of a the literature (Table A5. 1). These data were applied typical South African river. Table A5.1 Event Mean Concentration (EMC) data for different landuse types Landuse  Description TSS (mg/l) BOD (mg/l) TIN (mg/l) P (mg/l) Settlement - urban 100 15 3.41 0.79 Commercial / retail / Institutional 166 9 2.1 0.37 Industrial / road and rail 166 9 2.1 0.37 Extractive / utility 166 9 2.1 0.37 Farming / plantations and woodlots 201 4 1.56 0.36 Recreational open space 201 4 1.56 0.36 Settlement - rural 201 4 1.56 0.36 Natural vegetation (D'MOSS) 70 6 1.51 0.12 Settlement - informal 497   22 6.7 EVALUATING THE POTENTIAL RETURNS TO INVESTING IN GREEN URBAN DEVELOPMENT IN DURBAN PAGE 109 A5.2 Points of interest for scenario analysis Table A5.3 A summary of effluent water quality data at the outfalls of the WWTWs in the U60F catchment The Umhlatuzana and Umbilo Rivers and other Umbilo WWTW Umhlatuzana WWTW Hillcrest WWTW (Umbilo River) (Umhlatuzana River) (Umhlatuzana River) stormwater outfalls within the U60F subcatchment discharge directly into Durban Harbour. As a result, min 1.0 0.9 0.6 Durban Harbour is synonymous with poor water quality. max 42.0 23.0 55.1 In the SWMM model, there are 79 outfalls in total. The TIN (Ammonia, Nitrates + Nitrites average 7.9 7.3 6.7 (mg/L) annual loadings for TIN, TSS and P were simulated over summer 6.6 8.0 6.8 a 1-year period (July 2013 to June 2014). In addition to winter 8.7 2.8 7.1 the outfalls, pollutant load and maximum flows were simulated at a number of water quality monitoring min 0.0 0.0 0.0 stations situated in catchment U60F (Table A5. 2, Figure max 36.0 22.0 22.0 5.8). Note that these are the monitoring stations that Orthophosphates (mg/L) average 3.4 2.1 2.7 monitor nutrients as well as the physico-chemical summer 3.2 2.3 3.1 parameters. winter 3.6 1.9 2.1 Table A5.2 Water quality monitoring stations along the Umhlatuzana and Umbilo rivers min 1 0 0 max 244 214 1996 Sampling station River Location Suspended Solids (mg/L) average 21 7 48 R_Zana_10 Before the Umhlatuzana River meets the Umbilo river summer 25 7 26 R_Zana_28 Umhlatuzana River, downstream of the Umhlatuzana WWTWs winter 21 5 72 R_Zana_29 Umhlatuzana Umhlatuzana River, upstream of the Umhlatuzana WWTWs min 4 2 1 R_Zana_34 Umhlatuzana River, downstream of the Hillcrest WWTWs max 220 4001 4001 R_Zana_35 Umhlatuzana River, upstream of the Hillcrest WWTWs Turbidity (NTU) average 23 32 125 R_Umbilo_04 On the Umbilo River at Bellair Road – downstream of the Umbilo WWTWs summer 24 6 43 R_Umbilo_13 Umbilo Before the Umbilo river meets the Umhlatuzana River R_Umbilo_27 On the Umbilo River at Stapleton Road – upstream of the Umbilo WWTWs winter 24 6 57 R_Mkumbaan_01 Mkumbaan On the Mkumbaan River which is a tributary of the Umbilo River A5.3 Model calibration Monthly water quality data collected by the eThekwini The assumptions and limitations associated with model Water and Sanitation Department were used for calibration include the following: calibration and validation of the model. Most of the rivers within the EMA are monitored. Table A5. 2 ƒƒ The water quality monitoring program takes provides a summary of the sampling stations in the measurements on a monthly basis. This data therefore catchment used for calibration of the model. The water provides a snapshot of the water quality at a point at a quality measured at the outfalls of the WWTWs were specific time. TSS concentrations were inferred from analysed (Table A5. 3). These values were significantly turbidity data collected by the EM. The relationship lower than those provided in the General Effluent Limits. between TSS and turbidity was taken from data Therefore the values were replaced with the average collected by Newman (2015) in the Durban Bay, which TIN, P and TSS values measured at each outfall of the has a high salinity. TSS concentrations are temporally WWTWs. and spatially dependant and therefore this is not a true representation of actual TSS concentrations. We Simulations were run for a one-week period from the recognise that this a major limitation in the estimation 1 – 7 July 2014 at a two second time interval. This of measured TSS loads used for calibration; period was chosen because there was one rainfall event experienced throughout the subcatchment. The ƒƒ TIN, P and TSS concentrations from the WWTWs were results from these simulations are given in Figure A5. averaged and added into the model as a point source 1 and Figure A5. 2. The corresponding water quality at the outfall site of the relevant WWTWs. The parameters measured at the same points are provided discharge rates were assumed to be equal to the in Figure A5. 3. Note that the simulated values are within design capacity and is therefore not a true reasonable range of the measured values. representation of actual discharge rates and quality; and ƒƒ Generalised event mean concentrations (EMC) were Figure A5.1 Measure rainfall and simulated flow, depth, P, TIN, TSS concentrations for monitoring station (R_ZANA_10) on the Umhlatuzana based on values derived in the United States. River. PAGE 110 EVALUATING THE POTENTIAL RETURNS TO INVESTING IN GREEN URBAN DEVELOPMENT IN DURBAN EVALUATING THE POTENTIAL RETURNS TO INVESTING IN GREEN URBAN DEVELOPMENT IN DURBAN PAGE 111 Figure A5.2 Measured rainfall and simulated flow, depth, P, TIN, TSS concentrations for monitoring station (R_UMBILO_13) on the Umbilo River. Figure A5.3 Measured water quality from water sampling stations on the Umbilo and Umhlatuzana Rivers at the same locations as the simulations. PAGE 112 EVALUATING THE POTENTIAL RETURNS TO INVESTING IN GREEN URBAN DEVELOPMENT IN DURBAN APPENDIX 6: RIVER ECOSYSYEM HEALTH ASSESSMENT A6.1 Approach for assessing river condition been used in the hydrological model for a one-year period. The time series dataset was simplified to Changes in river condition associated with each of mean daily concentrations. These were presented in the different scenarios were assessed quantitatively, terms of different river health or condition categories, in terms of modelled instream water quality, and as recommended in South African (national) draft qualitatively, in terms of some of the broad parameters guidelines (DWAF 2008) and interpreted in Table A6. 1. included in considerations of ecosystem health. The range and/or threshold values were also taken The quantitative assessments used modelled from (DWAF 2008), as presented in Table A6. 2. Note concentrations of the three parameters included in that these values include orthophosphate (PO4-P) the hydrological and water quality model namely thresholds rather than total phosphorus. Since only total phosphorous (TP), total suspended solids (TSS) total phosphorus data were available for this study, this and Total Inorganic Nitrogen (TIN). Modelled hourly means that interpretation of data will tend to exaggerate concentrations of these parameters were available phosphorus enrichment slightly in some cases, although for the 10 sites shown in Figure A6. 1. Location of orthophosphate tends to comprise by far the largest water quality sampling sites for which data have portion of total phosphorus in riverine systems. Figure A6.1 Location of water quality sampling sites for which data have been used in the hydrological model EVALUATING THE POTENTIAL RETURNS TO INVESTING IN GREEN URBAN DEVELOPMENT IN DURBAN PAGE 113 Table A6.1 Comparison of different systems for the categorisation of river health/condition data, after DWAF (2008). National guidelines for the Table A6.3 Guidelines to inform Present Ecological State ratings for turbidity/clarity (after DWAF 2008) determination of the ecological reserve with regard to water quality recommend the use of numeric ratings 0-1. Deviation from Rating Deviation from reference conditions A-F Categories Natural-Poor categories reference A Natural Rating condition Environmental clues about the turbidity status 0 No change Pristine river, no known man-made modifications of the catchment, no known concerns about turbidity, A/B 0 No change changes in turbidity appears to be natural and related to natural catchment processes such as rainfall runoff. B Good 1 Small change Some minor man-made modifications to the catchment, changes in turbidity appear to be largely natural and B/C 1 Small change related to natural catchment processes such as rainfall runoff. Very minor effects of silting of habitats, large of a temporary nature and natural river processes clear newly deposited silt soon after the event C 2 Moderate change Moderate changes to the catchment land-use have resulted in unnaturally high sediment loads and high C/D Fair 2 Moderate change turbidity during runoff events. The impacts are however temporary. D Erosion and/or urban runoff processes is a known cause of unnaturally large increases in sediment loads 3 Large change 3 Large change and turbidity, habitat often silted but it is cleared from time to time. Low amounts of periphyton algae or D/E phytoplankton are present. E The catchment is known to have serious erosion problems, increased turbidity levels are present most of the 4 Serious change Poor E/F 4 Serious change time, large silt loads are deposited leading to a serious reduction in habitat. Low amounts of periphyton algae or phytoplankton are present. 5 Extreme change F The catchment is known to have serious erosion problems, increased turbidity levels are present most of the 5 Extreme change time, large silt loads are deposited leading to an almost total loss of habitat, silt loads are so high that fish kills Table A6.2 Threshold values for variables considered in this study, have been attributed to it. using ranges defined for each River Health Category (see The relevance and importance of TSS in assessing Table 5.15). The values shown in each row represent the upper threshold value of that category. water quality impacts on ecosystem health assessment of the impact of the different scenarios on ƒƒ The lack of data for other important water quality Category TSS is an important measure of water quality river condition (Table A6. 4), and in particular, on river parameters means that interacting parameters are not A-F 0-5 PO4-P (mg P/L) TIN (mg N/L) and river ecosystem health. It may be used to water quality, in this study, namely: considered (e.g. the influences of dissolved oxygen, approximate turbidity, and thus provide an pH, temperature and (in the case of some heavy A 0 <0.005 <0.25 ƒƒ The assessments are limited to TP, TSS and TIN and do understanding of water clarity – of importance metals) water hardness); and B 1 <0.015 <0.7 particularly in open water systems such as lakes not take account of other variables such as ammonia- C 2 <0.025 <1 and (in the present case) the harbour, as water nitrogen, various heavy metal concentrations; ƒƒ The lack of data regarding the organic component of D 3 <0.125 <4 clarity often determines dominance of the shallow bacteria, salinity; TSS and the proportions of total ammonia and water environment either by aquatic macrophytes orthophosphate comprising TIN and TP respectively. E 4 <1 <10 F 5 ≥1 ≥10 or phytoplankton. In addition to its links with water clarity and plant growth, suspended inorganic Total suspended solid (TSS) data could not be interpreted material carries an electric charge, and thus Table A6.4 Present Ecological State (PES) metrics and explanations after (DWA 2013), and used in this study to infer qualitative change in river condition as a result of scenarios involving attenuation of runoff and the provision of riparian corridors and buffers in this manner. DWAF (2008) does not in fact quantify provides adsorption sites for nutrients (phosphorus TSS concentrations for different health rating values, on and nitrogen) as well as trace metals and various Metrics Comment the basis that these data are not routinely measured by organic biocides. Suspended material that settles Potential instream Modifications that indicate the potential that instream connectivity may have been changed from the reference. out may smother and abrade riverine plants and habitat Indicators: Physical obstructions (e.g. dams, weirs, causeways). the Department of Water and Sanitation (DWS). Existing continuity modification Flow modifications (e.g. low flows, artificially high velocities, physico-chemical "barriers"). water quality guidelines for TSS are also limited, with animals. Community composition may change, depending on which organisms are best able to Potential riparian/ target ranges for South African aquatic ecosystems being wetland habitat Modifications that indicate the potential that riparian/wetland connectivity may have been changed. specified by DWAF (1996) as limited to a 10% increase cope with this alteration in habitat. Predator-prey continuity modification Indicators: Physical fragmentation, e.g. inundation by weirs, dams; physical removal for farming, mining, etc. in background TSS concentrations at a particular site interactions can also be affected by the impairment Modifications that indicate the potential of instream habitats that may have been changed from the reference. and time, although DWAF (2008) notes the high natural of visibility for predators that hunt by sight. Includes consideration of the functioning of instream habitats and processes, as well as habitat for instream biota variability between systems, making definition of degrees specifically. Potential instream Indicators: Derived likelihood that instream habitat types (runs, rapids, riffles, pools) may have changed in of change difficult in practice. Even when Reference The qualitative assessments carried out in this project habitat frequency (temporal and spatial). Assessment is based on flow regulation, physical modification and sediment (“Natural”) Condition TSS data are available for a particular considered the metrics used in Present Ecological modification activities. changes. Land use/land cover (erosion, sedimentation), abstraction etc. may indicate the likelihood of habitat system, their value is often limited. This is because of modification. The presence of weirs and dams are possible indicators of causes of instream habitat change. Certain State (PES) assessments of turbidity (see Table A6. 3) introduced biota (e.g. carp, crustacea and mollusca) may also cause habitat modification. Eutrophication and the tight links between sediment transport and flow (for which TSS is considered a surrogate value) to resulting algal growth as well as macrophytes may also result in substantial changes in habitat availability. velocity, and the largescale differences in sediment infer qualitative change in river condition as a result Modifications that indicate the potential that riparian/wetland zones may have been changed from the reference transport depending on discharge. In the current situation, of scenarios involving attenuation of runoff and the Potential riparian/ in terms of structure and processes occurring in the zones. Also refers to these zones as habitat for biota. therefore, while modelled TSS concentrations allowed provision of riparian corridors and buffers. Comment wetland zone Indicators: Derived likelihoods that riparian/wetland zones may have changed in occurrence and structure due Modifications to flow modification and physical changes due to agriculture, mining, urbanization, inundation etc. Based on land comparison between different scenarios, and some on the assumed relative influence of different scenarios cover/land use information. The presence and impact of alien vegetation is also included. comment on likely removal rates of other parameters on these issues was provided, with changes in TSS Modifications that indicate the potential that flow and flood regimes have been changed from the reference. likely to be associated with inorganic sediments (e.g. heavy data produced by the model also used to infer (but not Potential flow Indicators: Derived likelihood that flow and flood regimes have changed. metals and total phosphorus) they could not be used to quantify) changes in sediment transport and erosion. modification Assessment based on land cover/land use information (urban areas, interbasin transfers), presence of weirs, dams, infer absolute erosion and sedimentation rates. Guidelines water abstraction, agricultural return flows, sewage releases, etc. for the interpretation of links between turbidity and river In addition to limitations in the applicability of TSS Activities that indicate the potential of physico-chemical conditions that may have changed from the reference. health, as provided in Table A6. 3 (after DWAF 2008) were data in inferring catchment-scale erosion and sediment Potential physico- Indicators: Presence of land cover/land use that implies the likelihood of a change of physico-chemical conditions chemical away from the reference. Activities such as mining, cultivation, irrigation (i.e. agricultural return flows), sewage also drawn on in this regard. processes, the following limitations must also be modification activities works, urban areas, industries, etc. are useful indicators. Algal growth and macrophytes may also be useful considered with regard to the approach taken to response indicators. PAGE 114 EVALUATING THE POTENTIAL RETURNS TO INVESTING IN GREEN URBAN DEVELOPMENT IN DURBAN EVALUATING THE POTENTIAL RETURNS TO INVESTING IN GREEN URBAN DEVELOPMENT IN DURBAN PAGE 115 A6.2 Model outcomes The figures below reflect the modelled concentrations of TSS, total phosphorus and TIN under different scenarios, at key water quality monitoring sites in the catchment (see Figure A6. 1. Location of water quality sampling sites for which data have been used in the hydrological model). Figure A6.2 Effects of different modelled scenarios on total phosphorus concentrations at different monitoring sites in the study area. Figure A6. 2 (cond.) Effects of different modelled scenarios on total phosphorus concentrations at different monitoring sites in the study area. PAGE 116 EVALUATING THE POTENTIAL RETURNS TO INVESTING IN GREEN URBAN DEVELOPMENT IN DURBAN EVALUATING THE POTENTIAL RETURNS TO INVESTING IN GREEN URBAN DEVELOPMENT IN DURBAN PAGE 117 Figure A6.3 Effects of different modelled scenarios on total suspended solids (TSS) concentrations at different monitoring sites in the study Figure A6. 3 (cond) Effects of different modelled scenarios on total suspended solids (TSS) concentrations at different monitoring sites in the area. study area. PAGE 118 EVALUATING THE POTENTIAL RETURNS TO INVESTING IN GREEN URBAN DEVELOPMENT IN DURBAN EVALUATING THE POTENTIAL RETURNS TO INVESTING IN GREEN URBAN DEVELOPMENT IN DURBAN PAGE 119 Figure A6. 3 (cond) Effects of different modelled scenarios on total suspended solids (TSS) concentrations at different monitoring sites in the study area Figure A6.4 Effects of different modelled scenarios on total inorganic nitrogen (TIN) concentrations at different monitoring sites in the study area PAGE 120 EVALUATING THE POTENTIAL RETURNS TO INVESTING IN GREEN URBAN DEVELOPMENT IN DURBAN EVALUATING THE POTENTIAL RETURNS TO INVESTING IN GREEN URBAN DEVELOPMENT IN DURBAN PAGE 121 Figure A6. 4 (cond.) Effects of different modelled scenarios on total inorganic nitrogen (TIN) concentrations at different monitoring sites in the study area Figure A6. 4 (cond.) Effects of different modelled scenarios on total inorganic nitrogen (TIN) concentrations at different monitoring sites in the study area PAGE 122 EVALUATING THE POTENTIAL RETURNS TO INVESTING IN GREEN URBAN DEVELOPMENT IN DURBAN EVALUATING THE POTENTIAL RETURNS TO INVESTING IN GREEN URBAN DEVELOPMENT IN DURBAN PAGE 123