INTEGRATED CATCHMENT MANAGEMENT IN AN URBAN CONTEXT
THE GREAT AND LITTLE LOTUS RIVERS, CAPE TOWN
Report No 846/1/01

EXECUTIVE SUMMARY

The catchment area of the Great and Little Lotus Rivers presents a microcosm of South Africa's urban development, within a population of some 380 000 people. Communities in the catchment range from the fast- growing but desperately poor informal settlements, through low and middle-income communities with high levels of crime and gangsterism, and the economically important horticultural area of Philippi, to a high-income community around the shores of the lake itself. The legacy of apartheid is strongly apparent in the make-up of these communities and to some extent in the political and social structures representing them.

The overarching vision of the Lotus River project was to create a vehicle whereby people from these very diverse communities could come together with a common desire to improve the environment and quality of life within the catchment in which they live. The end result would be to restore the downstream water body, Zeekoevlei, as a recreational amenity for people from all communities in Cape Town to enjoy, as well as to enhance quality of life within the catchment as a whole. Furthermore, the Lotus River project proposed to develop a "blueprint" for managing urban catchments at a local level, based upon sound multi-disciplinary. research, which could then dovetail with the work of the much larger Catchment Management Agencies as set out in the National Water Act of 1998.

The aims and objectives of the Lotus River project were set out as follows:

• To establish a blueprint for urban catchment management in South Africa, focusing upon water quality and ecological stability and improvement, as well as upon hydraulics and flood control
• To identify major sources of pollution in the catchment, both point and non-point sources
• To carry out a detailed flow gauging study and a water quality sampling programme
• To study the ecology of the river systems, including riverine and benthic invertebrates, and macrophytes
• To develop rehabilitation strategies aimed at improving water quality by in stream measures, decanalisation, and the creation of wetlands
• To identify all major stake holders in the catchment and to develop a working model for community management of the catchment.

In the original project proposal, a major objective of much of the research would have been to build a GIS-based urban run-off model. Urban catchment management of the future will be based upon such models. Due to the limitations of time and funding, however, this was removed from the original list of aims and objectives. Indeed, it was found that much of the information necessary to carry out this task would not have been available. However, the work of the project continued to be underpinned by its Geographical Information System and the development of methodologies for the use of GIS in urban catchment management. Secondly, a major modification to the aims of the research, which was introduced during the course of the project, was in terms of the ecology of the river systems. The macro-invertebrate communities were found to be so impoverished, and the biodiversity of the river channels in general so low due to extensive canalisation, that having collected the base SASS data the focus of the project then shifted to looking at the ecological management of the catchment as a whole.

The Lotus River project focused from its inception upon creating a positive working relationship with the local authority structures. At all stages, "buy-in" to the project on the part of the local authority was considered to be essential. This was necessary in order to ensure the sustainability of the project objectives in the long term. Sustainability in this context would be demonstrated after completion of the project, by the major stake holders continuing to further the aims of the project and to manage the catchment effectively, with a focus upon water quality and ecological stability and improvement, as well as upon hydraulics and flood control. Furthermore, this management would be carried out in a manner that is inclusive and participatory , working with the full range of stake holders and communities in the catchment.

The GIS-based approach discussed in this report provides tools for the analysis and assessment of catchment data in an integrated manner. In addition, the GIS database framework has been designed for a multi-scale approach involving the interaction of metropolitan/regional, catchment level and local level databases. The proposed software configuration has been selected to facilitate an effective co-ordination and integration of data from multiple data sources. The proposed database content and analytical capabilities of the software can be used to address the issues of distribution of costs and benefits amongst all stake holders and to define the roles and responsibilities and accountabilities of the stake holders.

The key advantages of a GIS-based approach may be summarised as follows:

• More accurate parameter estimations
• More rapid parameter estimations
• Facilitates potential for utilising existing public databases
• Facilitates potential for grid-based distributed hydrologic modelling
• Facilitates an integrated analysis of the various data sets through the spatial overlay capability
• Facilitates a more realistic representation of data
• Raster backdrops to vector data
• Thematic representation of data
• Facilitates effective communication medium
• Facilitates wide spread diffusion of information
• Facilitates spatial analysis for planning
• Simplified spatial analysis tools
• Complex spatial analysis tools.

The primary objective of the GIS component of the Lotus River project was to integrate existing and acquired geo-spatial data on the catchment onto a single platform. The data collation process was carried out in such a manner as to include data from as many of the key role players in the ICM process as possible. This was for two reasons. Firstly, it helped to determine the practical limitations of "real data sets" on a GIS-based ICM approach. Secondly, by basing the methodology on locally acquired data, it would be possible to ensure that the resulting database design would be compatible with existing local authority databases.

A major constraint on using GIS for a catchment situated in a developing country is the patchy nature of digital data. While South Africa is relatively advanced in having digital cadastral data available for the metropolitan areas, there remain large gaps in this cadastral database, which correspond to the informal settlement areas. This is accompanied by an absence of physical infrastructure data for informal areas, while only a limited amount of subsurface infrastructure data is available for the formal areas. Where paper-based maps are available for the stormwater drainage network in former township areas, these maps tend to be unreliable and out of date. Secondly, the organisational structure within the local authorities is not geared up to the needs of integrated projects, and hence data acquisition becomes very difficult.

Two methodologies were tested for capturing the land use data. The first was a cadastral-based approach, while the second was based on mapping the land uses from digital orthophotos. The cadastral-based approach was followed for a selected sub-catchment and a number of land use parameters were estimated. In practice, the method proved far too time consuming to carry out on a catchment-wide basis. As a result, a more rapid method had to be developed. The merits of rapid land use mapping approaches have been very clearly shown in other non-catchment-based studies. The Rapid Land Use Assessment (RLA) methodology developed by the Planning and Development Corporation (PADCO) in the Philippines has proved to be an efficient approach to land use mapping. This methodology relied on the use of satellite imagery for the mapping process. In the case of the Lotus River catchment, 1:20 000 aerial ortho-photo imagery was used in the mapping process. This was available for the years 1983 and 1996.

A comparison of the mapping detail utilised in the production of maps for other recent catchment studies with the land use maps drafted for the Lotus River catchment study (which can be see on the accompanying CD) reveals that the latter are far more detailed with regard to:

1. the number of classes mapped for each sub-catchment,
2. the total number of classes mapped throughout the catchment, and
3. the size of the smallest land use zone mapped in the catchment.

Whereas previous modelling exercises tend to homogenise sub-catchment areas, the land use analysis mapping component of the present study attempted as far as possible to retain the degree of sub-catchment heterogeneity visible on the aerial photography. The use of high-resolution digital orthophotography (at 0.5 m pixel resolution) enabled land use classes useful for development of GIS-based vegetation management and rehabilitation strategies to be mapped.

The Lotus River catchment is a dynamic, fast developing area situated on the Cape Flats, a flat, sandy tract of land connecting the Cape Peninsula to the mainland of South Africa. Being so flat, the catchment boundaries are now to a large extent artificially defined. Although the Lotus Rivers (both Great and Little) drain south into Zeekoevlei, and ultimately into False Bay, large sections of the catchment drain northwards. Hence the urban run-off from this area could equally well have been routed towards Table Bay. In the longer term, the Cape Flats catchment areas all need to be reassessed as part of a wider metropolitan water management system. However , in the short term the Lotus River catchment makes an excellent case study of a highly impacted urban system, demonstrating the whole spectrum of land uses, socio-economic variations and housing types found in South African cities. It also exhibits acute ecological stress, highly polluted water bodies and significant loss of biodiversity.

The catchment area of the Great Lotus has been analysed in more detail, as this river is much larger and carries a greater pollution load than the Little Lotus. The latter runs primarily through areas of middle-income, formal residential housing, and the resulting water quality problems are not as severe. The Great Lotus catchment area, on the other hand, displays the wide variation in land use mentioned above, and is 4816 hectares in extent. 26% of this catchment is currently cultivated farmland, which is heavily fertilised. Over the 13 years from 1983 to 1996, the impervious area of the catchment grew from 17% to 34% of the entire catchment area, as a result of rapid urbanisation.

The Great Lotus catchment is now home to nearly 380 000 people, only 72% of whom are formally housed. Approximately 24% or 90 000 people reside in informal settlements, while 4% live in informal housing in the site and service areas. The figures for informally housed people exclude those living in backyard shacks within the formal housing areas, who are counted within the formally housed population. Informal settlements are a phenomenon, which is characteristic of cities in the developing world, but due to the former policy of influx control it was relatively unknown in South Africa until the late 1980s. The informal population of the Greater Lotus River catchment is growing especially quickly, due to its desirable location near to job opportunities, and large areas of "municipal open space".

Much of the drainage pattern on the Cape Flats was historically undefined, with high levels of infiltration through the sandy soil and many seasonal wetlands, which evaporated in the dry summer months. T o combat increased runoff due to urbanisation, the original Lotus Rivers (both Great and Little) were extended to newly developing areas as excavated canals. Further upgrading brought about concrete lining, which is efficient in rapidly transporting water yet ignoring problems of water quality and reduced ecological habitat. The Great Lotus River now exists as a multi-stage channel with concrete lined trapezoidal low flow section and grass lined high flow section. The section flowing through the Philippi Horticultural Area is an enlarged earth lined canal, permitting groundwater interaction. Throughout the redesign and upgrading work carried out historically, no flow measurements were ever undertaken to verify the event-based stormwater model for the catchment.

Water quality testing carried out by the municipality was restricted to concentration analysis for selected sites. Without any accompanying flow measurements, no source apportionment or annual loading calculations could be undertaken. Because of these shortcomings in the historical monitoring, a combined water quantity/quality sampling program was thus required aimed at carrying out a detailed flow gauging study and attempting to identify pollutant sources to the canal. Significant tributary inflows were selected for discrete sampling. A continuous datalogger/automated sampler was installed at the Springfield Road culvert, 8.5 km from the source. The probe measurements included flow velocity , pH and dissolved oxygen, and results were logged every 10 minutes. The sampler was also utilised to capture rain events producing runoff over a threshold flow, which allowed the estimation of annual pollutant loading for the various parameters.

A once-off analysis for protozoan parasites detected Giardia oocysts in the samples, but no Cryptosporidium. This is sufficient to constitute a health risk with respect to Giardia. Heavy metal analyses of both water and sediments, and analyses of volatile organic compounds (VOCs) in the water were also undertaken as a baseline assessment. These parameters were present in the Great Lotus canal within accepted guidelines. The VOCs were found to be mainly saturated hydrocarbons, characteristic of fuels and solvents e.g. diesel, motor oil, and kerosene. The Great Lotus River catchment is characterised by few industries and no wastewater treatment plants discharging into the canal. Most pollution contributions are therefore from "non-point" or diffuse sources ie. highway run-off, urban residential pollution entering the canal through the stormwater network and dumping, and the agricultural area contributing groundwater and surface water from drainage ditches.

The Great Lotus River is characterised by very poor water quality , with high nutrient loading, as well as very high faecal coliform counts. This is due to raw sewage effluent overflowing from blocked sewers into the stormwater drains, occasional sewer pump overflows, as well as the inadequate or non-existent sanitation characteristic of the informal settlement areas. Despite recent efforts by the local authorities at sewer upgrading, the microbial counts in the stormwater do not appear to be dropping. Table El summarises the water quality data obtained over the sampling period for the main parameters analysed.

Since sampling was carried out along the length of the Great Lotus River, it was possible to construct profiles of the pollutant concentration (in mg/l) and pollutant mass flux (in mg/s). The effect of each major tributary inflow was found by sampling 20 metres upstream and downstream of each tributary , and also where possible in the tributary itself, together with -a flow measurement. Carrying out a mass balance calculation on the pollutant mass flux then enabled the accuracy of the sampling method to be assessed.

As the numerical analysis in Table E2 shows, the largest contribution to the nitrogen loading came from the poor residential areas in the north of the catchment, which have a high proportion of informal housing, giving a total of 35.8% of the nitrogen flux. On the other hand, these areas contributed only 11.6% to the phosphorus flux, while the largest contributions came from subcatchments in the Philippi Horticultural Area. Also, groundwater flow played a more important role in the transport of phosphorus, in the unlined sections. In the lined sections of the canal, through the built-up residential areas, as expected, there was little groundwater contribution to flow.

To investigate the sources of total phosphorus (TP) coming from the Philippi horticultural area, a subcatchment investigation revealed that TP concentrations for channels adjacent to irrigated fields with intensive cropping are significantly higher than elsewhere in the subcatchment, despite the prevalence of calcareous soils. The ratio of uncultivated to cultivated p concentrations was 0.411, ie. the concentration of phosphorus in drainage ditches adjacent to cultivated lands is more than double than that measured adjacent to uncultivated land. This shows that the agricultural lands contribute high TP concentrations to the underlying subsoil water, through the application of fertiliser. TP concentrations in irrigation ponds were found to be relatively low, compared to values for the drainage ditches, indicating sedimentation taking place.

Storm event monitoring of the main stem channel revealed that both TP and TN mass flux increased during significant events. Maximum flux occurs around flood peak, indicating a seasonal variation in nutrient loading, with most of the load transported during the winter rainfall season. Combined EMC, total storm load (mass) divided by the total runoff volume, for selected storms multiplied by annual runoff at Springfield Road results in a annual TP load at Springfield Road of approximately 5300 kg/a. This is 10% higher than the interpolated load from the discrete sampling program for the same point, indicating that proportion of the load (contributed by winter storms) that is not captured by the discrete sampling programme. Assuming that the relationship between EMC and interpolated load calculation for Springfield Road holds for 7th Avenue, the annual TP load at 7th Avenue (and thus into Zeekoevlei) is approximately 7400 kg/a.

The Lotus River Catchment is a highly modified environment. It has recently become intensively urbanised, and even the remaining vegetated areas and wetlands are very disturbed. Drainage patterns have been highly modified; many of the v lei and marsh areas have been filled in or destroyed, and storm-water canals have been engineered. Nutrient additions to fanning areas and urban run-off have increased the nutrients in the aquatic and terrestrial systems 10- fold, compared to base levels which could be expected in the pristine environment. In addition to this, introduced Australian acacia species have become highly noxious invasive weeds, spreading into all open areas and eliminating the natural vegetation. The aims of this ecological assessment therefore move beyond identification and conservation of the remaining semi-natural/pristine areas of the catchment, towards a more dynamic, positive approach to the environment in an urban context. Rare and threatened habitat types such as seasonal vleis and sandplain fynbos need to be valued and managed, but this has to be done together with the people who live around these areas and bearing in mind the future development of the catchment. Above all, the major focus of ecological management in the catchment is rehabilitation.

Specific aims for the ecological management of the catchment can therefore be summarised as follows:

• Increasing the biodiversity of the catchment, both in terms of species richness and habitat diversity .Conserving unique species and features
• Increasing the sustainability of the ecological systems in the catchment, through reinstating functioning ecosystems
• Improving water quality
• Improving the quality of the available open space (both for human use and in terms of habitat quality for the fauna/flora)
• Ensuring a sufficient representation of habitats and habitat types in the subcatchment (enriching the urban environment)
• Enhancing access by communities to the environmental resources of the catchment.

The open areas in the Lotus River catchment can be classified into three categories, namely formally conserved areas, areas with potential for multiple land use, and degraded/unutilised areas. There are very few existing conservation areas and these are not actively conserved, eg. the Edith Stevens Nature Reserve. On the other hand, Varkensvlei forest reserve, while not officially conserved, is in private ownership and has not been disturbed. It is important that these areas be integrated into plans for the rest of the open space. They are both surrounded by open areas, which are of lesser conservation value but have high rehabilitation potential. The benefits involved in managing the reserves and the surrounding open space together are manifold:

• Active rehabilitation of the areas outside reserves, which are ignored and under-utilised at present, would be encouraged
• Core areas could be protected while still allowing people access to the open areas for recreation .Managing these areas as a single unit would increase both the range of habitats available and the size of the reserves
• If planned correctly it would be possible to create biological corridors and improve dispersal by joining together areas of high conservation value with areas of low conservation value.

Three types of wetland open space in the catchment area with potential for multiple land use and rehabilitation include: detention/retention ponds; farm irrigation ponds; and the river corridor itself. It is important to realise that detention and retention ponds could be examples of a habitat type that is highly threatened on the Cape Flats, ie. seasonal wetlands. They are the urban equivalent of the natural ponding areas that existed prior to urbanisation. Degraded as they are, any plans for improvement or rehabilitation of these areas do not need to be non-destructive. The areas could be engineered to be more accessible to people; water could be diverted into them to encourage nutrient uptake by the plants; possibilities exist for their use for the cultivation of high-value aquatic plants. Their function as detention storage need not be impaired by attempts at rehabilitation, if they are correctly redesigned.

Farm dams are designed purely for water storage; however, a few simple changes could also turn these numerous dams into valuable wetland resources. On a physical level changing the design of the dams so that they have gentle slopes (even just on one side) rather than steep banks would improve the habitat immensely. This would provide a gradient of wetland conditions (wet-dry) and allow for the establishment of aquatic vegetation. Another step, which could be taken, would be fencing off a portion of the wetland to prevent disturbance and trampling from cattle. Further rehabilitation possibilities involve actively planting wetland vegetation. The current use of farm dams does not necessarily clash with their potential to provide acceptable habitat. If this could be implemented then the Philippi horticultural region would contain of a series of islands of suitable wetland habitat scattered within the matrix of an intensively fanned area. Financial subsidy of some kind may be necessary to encourage the farmers to create such areas: this activity would offset to some extent the ecological damage caused by intensive fertilisation.

Within the notion of participatory development there are two important components. The first is the participation component and the second the development project, which should be two distinct, yet integrated processes. This implies that there should be two sets of objectives, one for thc project and the other for community participation. As for any project. there should be clear goals and objectives for the process of participation, which are distinct from those of the project. Thc goal and objectives of the community participation process provide thc basis on which to build and develop a community participation strategy.

The institutional models and mechanisms envisaged to translate the participation strategy into action would be informed by the objectives of the community participation process. The nature of these institutional models and mechanisms finally established should be influenced by the local context, that is, the project environment and the nature of the project. This is because every project environment has unique features, specific to that particular environment or situation, such as the social patterns, inter-community relations, social stability, socio-cultural and power relations in that environment. These features arc very important because they provide key indicators of the nature of structures and mechanisms that are likely to be durable, sustainable, effective and acceptable to the local people.

One of thc major objectives of thc Lotus River Project was to identify all major stake holders in thc catchment and to develop a working model for community management of the catchment. The community participation strategy was designed to involve identified community-based stake holders and representatives of the relevant local government departments directly in the process of developing a model for community involvement in urban catchment management. It was important to bring the community sector to an appropriate level of understanding of the concept, the process and the key components of Integrated Urban Catchment Management, to enable them to negotiate with the local authorities and other stake holders. The specific objectives to the strategy were as follows:

• To identify all stake holders who had an interest in the development of the Lotus River
• To develop a database of these stake holders
• To introduce ICM to thc community sector and their representatives
• To facilitate discussions and co-operation between stake holders
• To identify key issues of importance to stake holders in relation to ICM
• To identify the level of decision making at which the issues would considered
• To gain acceptance of the broad principles of community management
• To explore appropriate mechanisms of community input in the decision-making process.

The strategy was divided into three phases consisting of the database development, local interests and needs analysis and platform building phase. The process started with the identification of organised community groups such as civic organisations, local and school based environmental groups, RDP forums, local Councillors and influential individuals such as business and church leaders in the Lotus River Catchment area. The strategy was implemented through an on-the-ground approach consisting of:

• One-on-one discussions with community representatives
• Group tours to important sites in the catchment and along the river corridor
• Workshops to identify stake holder related interests and needs relevant to the ICM process
• Series of open meetings involving community and local government representatives to discuss issues of principle in terms of community involvement in the ICM process

What makes integrated projects such as ICM important is the inherent ability of such projects to pull together divergent views, needs and interests into generic categories and explore solutions that would address these needs and interest in a holistic manner. During the implementation of the community participation strategy, it was possible to develop specific categories that accommodated the needs and interests that were identified by the different stake holders. These categories are:

• Integrated service delivery and management: stormwater management should be linked to other services such as street cleansing, solid and liquid waste collection and disposal, environmental conservation and improvement. This requires that the planning and implementation of services and functions related to the achievement of a healthy and clean environment should be integrated, through decentralised mechanisms of planning and implementation such as area management
• Integrated development planning and implementation should also facilitate and promote an environmental ethic and sense of ownership. This could be effected through the following:
1. Linking environmental education and management programmes to social and economic development projects such as the provision of housing and appropriate infrastructure, promoting income generation opportunities (short and long-term), and supporting organisational development and skills transfer. The programmes must be designed to broaden the philosophy of environmental conservation and management to include socio-economic and physical elements and to assist people to gradually internalise and understand the connections between the biophysical and the economic elements of development.
2. Catering for the needs and interests of farmers while addressing the impact of their activities on the environment. The plan would therefore have to facilitate the development of appropriate farming systems approach to promote efficient and environmentally sustainable farming methods.
3. Developing long-term environmental awareness programmes that are linked to implementation projects related to local needs and interests. These programmes should be aimed at reaching learners in local schools, people in local organisations, environmental advocacy groups and the general public.
• The existence of urban informal settlements must be acknowledged and upgrading programmes implemented. Any process that is meant to promote an environmental ethic in poor communities will simply fail if people continue to live in squalid conditions, without decent housing and services.
• Community participation and integrated service delivery are two sides of the same coin. Participation should be supported through meaningful capacity building and skills development programmes. It is very important to note that community representatives have constantly raised the issue of tangible benefits resulting from people's participation in ICM, by implementing local projects that improve the physical environment, open opportunities for local employment, and provide capacity building and skill transfer opportunities to local organisations and individuals. Examples include:
1. Outsourcing special services like the collection of garden waste, building rubble and dead pets to local entrepreneurs. Such services are currently not provided in some areas of the catchment.
2. Providing public sanitation and waste disposal facilities in areas such as informal settlements, trading and sport areas, and public transport interchanges.

Conclusions

The future for urban catchment management in South Africa is full of opportunities:

• With advances in modelling, it will be possible to move towards forecasting a range of scenarios with visual GIS-based representation, allowing urban catchment committees to make real choices with regard to stormwater infrastructure. Advanced technology will make it easier, not harder, for non-technical people to understand the consequences of various choices, by being able to see the outcomes represented visually .
• As the concept of catchment management becomes more widely known and accepted among urban professionals and the general public, catchment committees will become able to influence the urban fabric more fundamentally through decisions on land use changes, and through creating multiple land use opportunities.
• With changes in institutional structures, urban catchment management risks losing its financial base, where water supply and wastewater treatment services increasingly become fragmented and privatised. Stormwater management alone has a less secure financial base. Achieving long-term sustainability for urban catchment management can be assured by linking more strongly with quality-of-life considerations, especially the provision of recreational amenities and the conservation and restoration of biodiversity. Here, community-based catchment initiatives as in the example of the USA are crucial to success.
• Conversely, where water services become more integrated institutionally, catchment management can be linked with land care, wastewater reuse and demand minimisation in an overall approach to urban water management. An example of this is the creation of the Sydney Catchment Authority in Australia (July 1999), which owns the water resources for Sydney, has responsibility for 11 catchments, and has strong referral powers with regard to undesirable developments.
• Perhaps the most important recognition to make in catchment management, especially in the urban context, is that social and institutional processes should take precedence over purely technical decisions. This has been shown over and over again in successful catchment management projects worldwide.

The concluding chapter of this report provides some recommendations and guidelines for urban catchment management. The first section of this chapter presents the top 10 lessons learned from watershed protection initiatives in the USA, as summarised by the US Environmental Protection Agency. These provide extremely valuable general pointers for the set-up phase, and for the social and institutional process of urban catchment management. Access to many useful Internet resources is provided in this section. Secondly, best management practices (BMPs) for stormwater management are outlined, together with an analysis of their effectiveness, drawn from international sources. This discusses the importance and priority which should be given to non- structural or "soft" BMPs, and provides a state-of -the-art list of structural or "hard" BMPs to be used in urban catchment management. Thirdly, some general guidelines for GIS-based urban catchment management are presented, which have been drawn from the experiences of the Lotus River project. The final part of the chapter presents specific recommendations for the management of the Lotus River catchment, which have arisen from the work carried out in this project.