Using GIS for optimisation of service delivery by Nthabiseng Motsamai and Chris Munyati, North West University The location of settlements has an important bearing on the effectiveness and feasibility of supply of municipal services. The use of GIS (geographical information systems) could assist in planning for a location that will support service provision and human safety. The demand for housing as population increases leads to unplanned settlements on the edges of cities and towns or what are commonly known as urban-fringes in urban geography. These unplanned settlements place pressure on municipal services and cause backlogs for services such as water, sewerage, road access and waste disposal. GIS technology appears to offer avenues towards solutions. GIS can be used as a tool in municipalities to assess and evaluate urban problems such as the needs of houses and basic services [1]. This is because GIS is a convenient tool which allows the input of different spatial data formats to be analysed at once. Service delivery in South Africa Municipalities fail to function optimally because they are faced with various challenges which render them ineffective in delivering adequate and sufficient services to the people. Settlement planning in South Africa has revolved around political and modernism ideologies for fifty years, in return this has resulted in low levels of services and high levels of inconveniences [2]. According to the Development Action Group (DAG) [3], South African citizens lack access to proper housing and this prevents them from receiving basic services. Although the Integrated Development Planning programme was introduced to contribute towards sustainable development and improving service delivery, the Department of Housing [4] has stated that local government has a lack of direction concerning tools required for settlement management, and that these programmes have spatial components but do not serve as tools in settlement planning. These spatial components make the use of GIS by municipalities vital in planning settlements. Fig. 1: A summary of procedures used for inputting the key requirements for the establishment of a formal settlement into a GIS. Mafikeng as a case study Mafikeng Local Municipality had a backlog of 43 736 houses in 2005, according to Vilakazi [5]. The 2007 Statistics South Africa Community Survey [6] indicated that Mafikeng Local Municipality had 10 884 informal dwellings, which indicates that there is a need for formal housing. The Mafikeng Local Municipality has also recognised this as a problem. The municipality faces challenges such as high service backlogs, growing informal settlements and housing shortages [7]. Optimisation of service delivery needs to be addressed by beating the backlog for housing in formal settlements. In this study GIS analysis was utilised to assess the optimal location of a theoretical new formal settlement in Mafikeng, assuming that the local municipality was to establish one. The analysis in the study opted for medium-density settlements because DAG [3] noted that medium density settlements made a positive contribution to improving basic services and overcoming housing shortages. Medium density settlements mainly consist of multi-unit development, detached triplexes and low-rise apartments in a given area [8]. However, the analysis in this study assumes the new medium density settlement will consist of single (detached) housing units on stands typically 30 x 30 m in sizes, which are common in Mafikeng. It is important for settlement locations to be thoroughly planned using tools such as GIS as 38 PoPositionIT March 2012
because it can clearly define the local shapes of small areas more accurately. Spatial layers such as railway lines, rivers, dolomite geology, dams and wetlands were not used in their original form of size. They were overlaid with the district boundary shapefile in ArcMap and clipped to narrow the analysis to the extent of the Mafikeng town. The spatial layers used for the analysis were rivers, dams and wetlands, the town and outskirt settlements, cultivated lands, main roads, railway line, dolomite geology, landfill site, industrial area, heritage site, conserved area and the game reserve (Table 1). Fig. 2: An ArcMap model summarising the protocol used to derive a layer of undesirable sites. Fig. 3: An ArcMap model summarising the protocol used to derive a layer of unoccupied sites. this could help in supporting service delivery and avoiding areas that are not suitable for developments. Consequently this paper aims to develop an illustrative method that can be adopted during the process of planning a location for a new settlement in Mafikeng. GIS data The implementation of GIS in identifying an optimal location required minimum ideal requirements for a settlement so as to identify spatial layers that would be necessary for the GIS analysis. The requirements for a settlement were obtained through consultations with municipal authorities, the Department of Human Settlement and by referring to appropriate journals as well as internet sources. After obtaining the necessary information that is normally considered at a professional level and in accordance with the municipal standards, a requirements table was formulated (see Table 1). The spatial layers were acquired from digital maps prepared by the Directorate of National Geo-spatial Information, and through on-screen digitisation. Spatial layers which were not available such as the surrounding settlements and Mafikeng Game Reserve were digitised using a 2,5 m spatial resolution SPOT-5 image which was already pan sharpened. The image was acquired on 14 March 2010 and obtained from CSIR. A 2,5 m SPOT-5 image was preferred because of its ability to distinguish small structures on the ground. ArcMap version 9.3 GIS software was used for data analysis. The UTM projection system was adopted for all the spatial layers GIS analysis The ArcMap analysis tools which were applied on the input layers to select an optimal location that would optimise service delivery were overlay, buffering and extract. Dolomite geology was taken into consideration for location planning because settlements located on dolomitic areas could trigger the formation of sinkholes. The spatial layer of dolomite geology was used in the analysis in this study. However the dolomite geology did not affect the study area. Selection of potential sites ModelBuilder was used in ArcMap to create two models, one to derive undesirable areas and another to derive a layer of unoccupied sites. According to Esri [9] the analysis of complex data using ModelBuilder is simple as it creates a suitable environment for models to be created, edited and managed properly. Various buffer zones were allocated around spatial features that required protection and those that were unsafe for the new settlement location. In model one (Fig. 2), the union operation was used to combine all the buffer zones around undesirable areas for a settlement, i.e. rivers and wetlands in order to keep settlements away from potential flood prone areas; railway lines to avoid high noise and air pollution; industrial and landfill sites to avoid areas of bad air quality. This resulted in a map showing areas where a settlement should not be developed. In model two a layer of the study area boundary and the built up areas were combined with the layer showing undesirable sites and the select tool was used to extract the layer showing all unoccupied sites (Fig. 3). The unoccupied sites were identified using PositionIT March 2012 39
Requirements for a settlement Should have a feasible access to public institutions (schools, health services, markets) and public utility services (water provision, sewage removal, storm water disposal, solid waste removal and electricity supply) Should not be located in dolomitic areas Underground municipality water pipes should be avoided Heritage sites and protected areas should be avoided (national parks, wetlands, botanical gardens). Should be within access to public transport systems Low lying areas should be avoided as they are prone to flooding Should be located where there is a good air quality (away from landfill sites, noxious industries). Measures to avoid contaminating the water table should be considered. Soils in the municipal area will help to direct the growth away from the arable land. Shoud not be placed in area where the ground water is within 1 m of the surface Should be placed at least 100 m from the water resources and 50 m from resources requiring lower level of protection. Spatial layer required Schools, town, health services Dolomite geology Ground water pipes National parks, wetlands, botanical gardens, rivers, dams, marshes, heritage graves Railway lines, roads Contour lines Landfill sites, industries Ground water table layer Soil map, cultivated lands Ground water table layer Rivers, dams, wetlands Table 1: List of the requirements for a new formal settlement and spatial layers that were required for GIS input. was performed to find a location with the most favourable capacity to support a large number of formal houses. The average plot size of 900 m 2 (i.e. 30 x 30 m) for a medium-density settlement [11] was used to determine the number of houses that would fit into each of the locations. This was accomplished by using the following equation: The equation, however, did not take the sizes of residential streets and foot path-ways into consideration and was, therefore, merely for indicative purposes with regard to maximum number of housing plots that each site could support. Discussion and conclusion Eight new locations which complied with the requirements of a suitable settlement were identified. The locations were labeled A to H for reference (Fig. 6). None of the new identified locations interfered with the buffers of unsuitable land-uses, that is the new locations were situated away from flood prone areas, polluting land-uses and ecologically sensitive areas. Based on the sizes of the identified locations displayed in Table 2, location C was the most favourable because it could take the largest number of plot units (5380 plot units). The number of informal houses resulting in backlogs of municipal services in Mafikeng could be reduced by developing the new medium density formal settlement at location C. Another location which could be also considered for development is part of location A as it is situated closer to location B. Further analysis should be carried out using a layer of other roads within the Fig. 4: Unsuitable locations and new identified possible locations for a new settlement. the FID values from an auto-generated attribute table [10] of the union layer for undesirable sites, the occupied sites and the boundary of the study area. The map resulting from model one, showing areas where a settlement should not be developed, is shown in Fig. 4. Vacant areas where a settlement could be located are shown in Fig. 5. Polygons were then drawn on unoccupied land based on the researchers point of view (not a GIS tool) to delineate sections of the unoccupied land that could be developed for a settlement (Fig. 6). To display areas which were close to the existing roads a layer of regional roads was used for the analysis. A layer of regional roads was overlaid on top of the layer of the possible locations and a 1 km buffer zone was then applied around the roads (Fig. 7). The method was achieved by using the buffering operation. Further analysis Locations Total area (m 2 ) Number of plot units A 3 621 160 4023 B 1 867 070 2074 C 4 842 310 5380 D 798 555 887 E 925 013 1027 F 3 332 890 3703 G 2 139 850 2377 H 1 411 961 1568 Table 2: The total area sizes for the potential locations and their potential capacity. 40 PositionIT March 2012
Fig. 5: Map showing an extracted layer of unoccupied sites. Fig. 6: Possible locations of new settlement on unoccupied areas. Fig. 7: Proximity of a new settlement location to regional roads. municipality to find out if location B is accessible via existing roads. This will help maximise the total land area that could be developed to improve service delivery by reducing the number of informal settlements. Location D was the least suitable with a total number of 887 houses. The use of sufficient and reliable data is necessary in optimal site selection processes because any error may result in the GIS tool being ineffective. Planning for new locations before time using GIS could also assist in forecasting future demands for houses and in the process minimise the possible inadequacies posed by informal settlements. GIS was also helpful in providing tools for managing complex data and identification of optimal locations from which visualisation maps were produced. On the basis of these results, field trips were able to be conducted. A field trip was conducted to the five locations which were accessible, these being locations B, C, D, E and F for validation. Three of the identified locations (C, D and F) were found vacant while the other two (B and E) were occupied by housing construction works. An investigation will have to be conducted to verify whether the settlers are illegal or not. The factors which affected the accuracy of locating a new settlement were not necessarily caused by the GIS tool but the input data. The results of the study were affected by the age of the SPOT-5 image which was used to map the layer of surrounding settlements in the study area. The image was obtained in 2010 and as a result new building could have emerged after it was taken. Also, the data from the source of spatial layers used may have had digitising errors which could have affected the allocation of buffers around other land uses. The researchers had no control over these errors. From the results recommendations can be made on the use of GIS in settlement planning. GIS should be included in the preliminary stages of Environmental Impact Assessment (EIA) because GIS has tools that can assist in identifying areas that are ideal for development. Other land attributes such as soil, contours and groundwater table layers should be included to make the analysis more accurate. The slope will help avoid locations on flat areas that are prone to flooding, soil conditions will assist in directing settlement away from arable land whereas the groundwater 42 PositionIT March 2012
depth will help show locations of shallow depths to ground water. This is because the developed land could easily affect the quality of ground water. In conclusion the study was also able to establish an illustrative method using GIS to select an optimal location for optimisation of service delivery. References [1] J A Martinez: Evaluating housing needs with the use of GIS. Habitat International, 24 (1), 501-515, (2000). [2] CSIR: Guidelines for human settlement planning and design: The red book. CSIR Building and Construction Technology. 1. www.csir. co.za/built_environment/redbook/ Vol_I/Volume_I_new.pdf Available at: [Accessed on 2011/04/11] (2000). [3] Development Action Group: Development Action Group s Medium Density Housing Programme. Available At: www.dag.org.za/docs/ programmes/mediumdensitydoc.pdf [Accessed on 2011/10/27] (2007). [4] Department of Housing: Commission for sustainable development twelfth session. Available at: www.info.gov. za/view/downloadfileaction?id=70150 [Accessed on 2011/04/20] (2004). [5] F P Vilakazi: Department takes over housing responsibilities from Mafikeng Local Municipality. In: South African Government Information, Available at: www.info.gov.za/ speeches/2005/05121412451002.htm [Accessed on 2011/10/26] (2005). [6] Statistics South Africa: Community Survey 2007: by municipality and type of main dwelling, Available at: www.statssa.gov.za/timeseriesdata/ pxweb2006/dialog/saveshow.asp [Accessed on 2011/10/26] (2007). [7] Mafikeng Local Municipality: Reviewed Integrated Development Plan 2008/9-2011/12. Mafikeng, North West Province, (2010). [8] D Turner, J Hewitt, C Wagner, B Su and K Davies: Best practice in Medium-density housing design, Unitec: New Zealand (2004). [9] Esri: ModelBuilder, http://webhelp. esri.com/arcgisdesktop/9.2/index. cfm?topicname=an_overview_ of_modelbuilder [Accessed on 2011/11/13] (2006). [10] Esri: Union (Analysis), http://help.arcgis.com/en/ arcgisdesktop/10.0/help/index. html#//00080000000s000000.htm [Accessed on 2011/11/07] (2011). [11] K Rampaul (kishore.rampaul@ mafikeng.gov.za): Medium density houses [e-mail] N Motsamai (motsamain@yahoo.com). Sent Wednesday, 19 October 2011: 09:30 AM (2011). Contact Nthabiseng Motsamai, North West University, Tel 076 616-5963, motsamain@yahoo.com PositionIT March 2012 43