Implementing a community- integrated GIS: perspectives from South African fieldwork Trevor M. Harris and Daniel Weiner Chapter 18 18.1 INTRODUCTION A conceptual framework for PPGIS has been well developed during the last decade, and several crucial elements of both the public participation and GIS components have been identified (Abbot et al. 1998; Obermeyer 1998). Until quite recently, however, there were few examples to demonstrate how the implementation of PPGIS might actually proceed (Craig et al. 1999; Talen 1999). The purpose of this chapter is to identify one such approach based on fieldwork undertaken in Mpumalanga Province, South Africa. It was in addressing the complexities of PPGIS implementation that we coined the term community-integrated GIS (CiGIS), intended to represent a slightly different mode of PPGIS implementation than that previously envisaged (Harris and Weiner 1998). In this chapter, we briefly outline the basic con- cepts behind CiGIS, and present an application in support of land and agrarian reform in South Africa. 18.2 GIS, SOCIETY, AND PPGIS Early thoughts about PPGIS implementation envisaged placing a GIS into the hands of communities almost as a counterpart to the systems operated by public and private agencies. In our fieldwork in South Africa, we quickly rejected this approach as infeasible and shifted to a CiGIS orientation. To provide some context for this conceptual and operational change, we now identify several key issues raised in the literature on GIS and society, and explain how they impacted our attempt to connect community participation with a GIS. First, and perhaps foremost, was the issue of how to address or overcome differential access to hardware, software, data, and expertise. Much of the discussion on this topic in the literature was strikingly played out during our work in transitional South Africa (Harris et al. 1995; Weiner et al. 1995). Many communities in the case study area were struggling to acquire © 2002 Taylor & Francis Implementing a community-integrated GIS 247 basic necessities of life such as water, shelter, food, and fuel. Most commun- ities did not have access to electricity, and education had been deliberately withheld. Despite the tremendous enthusiasm and desire of local commun- ities to participate, the chronic and endemic problems of community access to basic resources in Mpumalanga Province necessitated a GIS implemen- tation strategy that drew upon GIS capability, support, and willingness from outside the communities themselves. In this case, a project team from West Virginia University (WVU), in collaboration with the South African Department of Land Affairs, filled this role. Second, the issue of structural knowledge distortion in post-apartheid South Africa became of paramount concern in developing a community response to land and agrarian reform. The major central government and provincial agencies of the apartheid regime had fully embraced the new technologies of GIS and remote sensing, and under government mandates had operated them in support of an oppressive state regime. Digital spatial data were available from these agencies, but they were unreliable and expensive to purchase. Some of the data were also deemed confidential and were not readily available. Some data simply had not been collected. For example, land claims are a major component of the post-apartheid land reform process, yet no official documentation of forced removals exists. The overlapping tribal and community land claims that we encountered suggests several phases of forced removals occurred, none of which were officially recorded or documented. Thus, the data collected by the state, the available geo-spatial databases, and their content were representative of the goals of an apartheid state, and reflected the conceptions of space of an ‘elite’ (predominantly white) sector of society. These factors highlighted a third area of concern frequently discussed in the literature: the desire to complement ‘official’ and ‘expert’ digital spatial information with local knowledge held by members of the community. Seeking to redress structural knowledge distortion through the inclusion of local knowledge held by members of black communities themselves thus became a primary challenge of the project. Much of a community’s know- ledge is heavily qualitative in nature and invariably based on oral history and the experience of having lived in a place for some time. Capturing this knowledge in a GIS that relies heavily on the spatial primitives of point, line, and polygon and the quantitative ordering of information is no easy task. These issues forced us to address both qualitative and quantitative aspects of local knowledge acquisition, and the integration of this knowledge into a GIS. Fourth, it is erroneous to think of local knowledge as homogenous or uniform. In many respects, it would be better to use the term ‘knowledges’ as a way of recognizing that community information is varied and socially differentiated (Mahiri 1998). In seeking to include local knowledge within a GIS, the problems of identifying and incorporating the socially differenti- ated perspectives of community participants had to be confronted. Doing so © 2002 Taylor & Francis 248 T. M. Harris and D. Weiner explicitly acknowledges the very real and constraining problem of differen- tial access to information, and underscores that in many instances the focus of PPGIS may well be to address and ensure that the varying perspectives of community members are incorporated into a GIS. 18.3 CONCEPTUAL FRAMEWORK FOR THE SOUTH AFRICA CiGIS The South Africa CiGIS is guided by three broad conceptual principles: popu- lar community participation; local, social and spatial differentiation; and regional political ecology. Community participation has become a mantra in development planning and field-based academic research. Unfortunately, most participation associated with development planning is essentially participation as legitimization. Community meetings are held, local input is gathered, reports are produced, and top-down planning is maintained. In this context, participation helps to legitimize decisions that are not necessarily ‘popular’ within impacted communities. In the academic world, participation has come to designate a configuration of qualitative methods designed to understand complex social processes better than conventional quantitative or qualitative methods. Efforts to hear the voices of ‘ordinary’ people and ‘capture local knowledge’ are well intentioned, but in many instances these are forms of participation for publication, in which academics under- take research to produce books and journal articles while leaving the subject communities with little (if any) tangible benefits. Popular participation is an attempt to locate community participation in the context of particular local configurations of power within civil society. Participatory processes become part of the structures of everyday life, and ordinary people are able to express their opinions as openly as possible. The South African CiGIS has its roots in a participatory land reform project initiated in 1991 during a period of intense political struggle and violence (Levin and Weiner 1997). As a result of our participation in that project, we are known in the community and viewed as friends and advocates of pop- ular local causes. The participatory process is thus central to our work, and the issues addressed in the CiGIS are community issues that have significant local importance. CiGIS implementation assumes, therefore, that tangible community needs are being addressed and that the project is political by its very nature. Our conceptual and methodological framework for CiGIS development and implementation also assumes social and spatial differentiation. As sug- gested above, communities are not homogenous and GIS can inadvertently maintain unequal development. In the South Africa study, spatial differenti- ation is represented by the inclusion of diverse forms of participant social groups, including land reform organizations, peri-urban former homelands © 2002 Taylor & Francis Implementing a community-integrated GIS 249 groups, farmworkers, large-scale (white) commercial farmers, and local chiefs. Race- and gender-based forms of social differentiation are also included in the South African CiGIS, and age, class, and other forms of difference will be added to the analysis in the future. The South African CiGIS is also guided by an appreciation of regional political ecology. This conceptual framework helps researchers to analyse the social histories and landscape politics of the participant communities, and to reflect on their own academic interests in these areas. Our relation- ship with these South African communities began in 1991 when community elders explained how grand apartheid social engineering had dispossessed them of land, water and biomass resources in their former Lebowa home- land. The contemporary poverty of these groups was clearly linked to the historical geography of forced removals and to the production of local and regional apartheid geographies. 18.4 MPUMALANGA CASE STUDY The Mpumalanga Province is a transitional area between the relatively cool and moist highveld plateau (over 1200 m in altitude) and the hot, dry lowveld (200–600 m in altitude). Mean annual rainfall ranges between 400 and 700 mm in the lowveld and between 1000 and 1500 mm on the escarp- ment and parts of the highveld. These environmental features, combined with the history of forced removals and forced urbanization under colo- nialism and apartheid, have produced a landscape of extreme social and ecological variation. The total population of the Province is over 3 million, of whom one-third live in urban areas and almost half reside in the former homelands. The case study area, the Central Lowveld subregion, is located mainly within the Lowveld Escarpment District of Mpumalanga Province, and includes a small portion of Bushbackridge to the north (Figure 18.1). The latter is disputed territory in Northern Province, and includes portions of the former Lebowa and Gazankulu homelands. Intensive and exotic industrial forest plantations and large-scale commer- cial fruit and vegetable farms dominate the western third of the case study area. Some of these are located on highly arable land. Forestry companies control large tracts of state land, and this raises substantive issues regarding socially and ecologically appropriate land-uses. Forest plantations and large- scale commercial farms thrive because of a highly skewed system of water access. During the apartheid era, the social production of this watershed was centred on a complex system of dams and tributaries that capture valuable water for (mostly white) large-scale commercial farms (Woodhouse 1997). The former homelands of KaNgwane, Gazankulu, and Lebowa are located east of the agriculture and forestry plantations. These bantustans are overcrowded and poorly serviced relics of grand apartheid. Land demand is © 2002 Taylor & Francis 250 T. M. Harris and D. Weiner high, water is in short supply, and the history of forced removals remains fresh in peoples’ memories and imaginations. Historically, political struggles have been connected to the decline in access to land, water, and biomass resources (Levin and Weiner 1997). The Kruger National Park and several private game parks occupy the eastern portions of the case study area. Since 1994, tourism has again become a growth industry, and visitors to the Mpumalanga and Northern Province Lowveld are growing. The use of land for game tourism has generated interesting discussions within the region regarding the potential for community-based range management models. Figure 18.1 The Central Lowveld case study area, South Africa. © 2002 Taylor & Francis Implementing a community-integrated GIS 251 Many of the study participants, however, perceive limited personal benefit from the adjacent game parks. 18.5 CiGIS IMPLEMENTATION Populating the CiGIS database was a central issue in the South African project. Acquiring spatial data is always a challenge for GIS practitioners; however, in CiGIS, this challenge is compounded by the need to draw heav- ily on local knowledge obtained from within communities. Developing the CiGIS database thus focused on obtaining the more traditional GIS cover- ages that detail the physical and cultural infrastructure of a region, and obtaining qualitative knowledge from members of the local community. Incorporating traditional data involved the familiar search for existing data of sufficient quality, attribution, relevance, and scale to meet the needs of the project. Digital spatial data were obtained from official government sources and private data providers, and by scanning and digitizing existing analogue maps. The digital data comprised both vector and image files. Digital raster graphics were generated from the Ordnance Survey 1:50,000 topographic map sheets and were based on mid-1980s source maps. Built- up and peri-urban settlement patterns were obtained in digital form from private vendors and were based on 1997, 1:10,000 orthophoto imagery. Data on land contour, hydrology and dams, roads, railroads, jurisdictional boundaries, and state-owned lands were obtained in digital form from government agencies at a scale of 1:50,000. Land-cover data were obtained at 1:250,000, and land-type data were captured at 1:50,000. CiGIS requires that traditional top-down ‘expert’ information be com- plemented by information garnered from local community groups. In this case study, the latter consisted of local groups within the former homelands with various relationships to the government’s land reform programme. These groups are characterized by a diversity of rural production systems and relations of production, and include Cork Village and Nkuna Tribal Authority; Friedenheim Farmworkers; Masoyi Tribal Authority; Masizakhe Land Redistribution Project; Sitama Impilo Land Redistribution Project (Figure 18.1) and six (white) large-scale commercial farmers in the area. Where appropriate, each of these groups were further subdivided into groups of men, women, and tribal leaders in order to capture the crucial socially differentiated local knowledge each group held. Community workshops were held to compile information from each group based on five broad political ecology concerns: 1 the historical geography of forced removals, 2 identifying and comparing ‘expert’ and local understandings of land potential, © 2002 Taylor & Francis 252 T. M. Harris and D. Weiner 3 perceptions of socially appropriate and inappropriate land-use, 4 access to natural resources, and 5 community views about where land reform should take place. Each community was contacted in advance to arrange the half-day work- shops. Base topographic maps of the areas were created at a variety of scales and brought to the workshops. The maps were overlaid with tracing paper, and each group was asked to record their perspectives on the above questions using colour-coded markers. Each group, comprised of approx- imately eight to ten people, met separately, and included a facilitator. 1 The resulting ‘mental maps’ were digitized, attributed, and incorporated into the GIS database. In addition, photographs, video recordings, and voice record- ings of the interviews were taken and georeferenced to the mental maps. The core of the CiGIS process was to integrate this information into a GIS database. As indicated, this represents a significant challenge because of the qualitative nature of much of the information. Initial work focused on embedding objects within an ArcView GIS, but an Internet-based GIS replaced this system. The Internet-based system provided a more suitable GIS environment within which to link quantitative GIS coverages with qualitative voice, photograph, text, and video data. An Internet-based system also permits ready access to other resources on the web. Perhaps most importantly, an Internet-based GIS provides a means by which to overcome some of the disadvantages of differential access. Although access to the Internet is not widespread in Mpumalanga Province, it provides the poten- tial for greater access to GIS resources in the future. With an Internet-based GIS, maps and information can be downloaded using simple point-and-click procedures. An understanding of GIS concepts or software is not essential, removing another significant obstacle to communities having access to their information. Although we do not wish to minimize the very real obstacles to South African communities gaining access to the Internet, it is remarkable how quickly the Internet is becoming accessible in these regions through state agencies, the private sector, and community telecentres. We believe multi- media Internet GIS is a central component in the development of CiGIS, and that state agencies will have to take responsibility for providing access to resources and technology in order for such a system to be developed and for communities to be incorporated into GIS-based decision-making. 18.6 FIELD RESULTS In this section, we focus on two examples to illustrate the type of informa- tion generated from the fieldwork and some critical issues posed by CiGIS implementation for land reform in South Africa. A more detailed presenta- tion of the data can be found in Weiner and Harris (1999). © 2002 Taylor & Francis Implementing a community-integrated GIS 253 Land types data obtained from the Agricultural Resource Council of the South African Institute of Soil, Climate, and Water are presented in Figure 18.2. Based on the Institute’s soil classification and slope information, land potential categories were established. In the study area, 43 per cent is clas- sified as land of ‘higher’ agricultural potential, 17 per cent as ‘medium’ potential, and 40 per cent as ‘lower’ potential. This database of ‘expert’ knowledge about land potential was focused on the fertile river valleys from which black South Africans tell us they were forcibly displaced. In the course of comparing local knowledge of land potential with the official coverage from the Institute, several anomalies were identified. In several instances the mental maps of local communities, which included both Figure 18.2 The multiple realities of land potential. © 2002 Taylor & Francis 254 T. M. Harris and D. Weiner black and white participants, differed from the Institute’s map. Specifically, local communities identified several areas of land as higher-quality land than was recorded by the Institute. The location of these areas is signific- ant for their proximity to areas under scrutiny for potential reform. In essence, the differences occurred as a result of contrasting scales of data capture and differing perceptions of what constitutes high-quality land. Large portions of land that have a slope sufficient to make them inappro- priate for mechanized agriculture were deemed of low quality by the Institute, but were considered very attractive by small-scale farmers who use animals and hoes. We would expect an operational CiGIS in the region to help locate viable high-slope areas with potential for small-scale agri- cultural production. In the second case study example, there was no ‘official’ data on forced removals in the area. Braum Raubenheimer, former Minister of Water Affairs and a member of cabinet under Prime Minister Verwoerd, was a project participant, and he denied that black South Africans were forced to relocate as a result of white settlement or the actions of the apartheid government and police. A very different story emerged when this issue was discussed with black South African participants. One-quarter of the black population in the study area told us they experienced at least one forced removal in their lifetime (Levin and Weiner 1997). This is why CiGIS participants remain willing, even anxious, to talk about the historical geo- graphy of forced removals. White farmers, however, were reluctant to discuss issues of forced removals. The mental maps of whites and blacks in the subregion are com- pared in Figure 18.3. The maps suggest very different perceptions of subre- gional landscape history. Forced removal mental maps for black participants delineated an extensive area of removals especially in the southwest region of the case study area. The white farmer mental maps acknowledged areas of forced removals that were significantly smaller in size. Unexpectedly, the white farmer responses also identified areas of white farmer removal, most likely for homeland expansion. Furthermore, because the peri-urban black settlement is located on higher potential arable land, black participants indi- cated that the area of settlement located to the immediate south of Hazyview is where the tribal chiefs removed blacks. This was done to enable members of the tribal authority and local black businessmen to gain better access to this high-quality land (Weiner et al. 1995). Mental maps are qualitative representations that must be handled care- fully. However, the mental maps were invaluable in representing the only known record of forced removals and for identifying phases of forced removals in which removed communities were subsequently relocated. These complementary interpretations of historical dispossession have provided the basis for understanding the existence of many overlapping land claims that have contributed to the slow pace of land restitution. © 2002 Taylor & Francis Implementing a community-integrated GIS 255 18.7 CONCLUSION The literature on GIS and society has generated rich conceptual and polit- ical questions for GIS developers, users, and practitioners about issues of database development and use, visualization and representation, and the power relations that affect system access. PPGIS is one outcome of such cri- tiques and PPGIS efforts have, to date, been particularly concerned with issues of community empowerment, disempowerment, and the integration of quantitative and qualitative information. Figure 18.3 The multiple realities of forced removals. © 2002 Taylor & Francis [...]... Cartography and Geographic Information Systems 25(2): 67–76 Harris, T., Weiner, D., Warner, T and Levin, R (1995) ‘Pursuing social goals through participatory geographical information systems: redressing South Africa’s historical political ecology’, in J Pickles (ed.) Ground Truth: The Social Implications of Geographic Information Systems, New York: Guilford Publications, pp 196–222 Hastings and Clark... Africa: problems, challenges and opportunities for co-operation’, International Journal of Geographical Information Systems 5(1): 29–39 Hill, T R and Strydom, J M (2000) ‘A call for participatory GIS as an approach to rural development and spatial data dissemination, with reference to water management’, South African Journal of Surveying and Geo -Information 1(4): 185 –190 Levin, R and Weiner, D (eds.) (1997)... digital landscape: GIS, remote sensing and local knowledge in Kiepersol, South Africa’, Cartography and Geographic Information System 22(1): 30–44 Weiner, D and Harris, T (1999) Community- Integrated GIS for land reform in South Africa’, WVU Regional Research Institute Research Paper # 9907, Morgantown W V (http://www.rri.wvu.edu/wpapers/1999wp.htm) Woodhouse, P (1997) ‘Hydrology, soils and irrigation systems ,... information systems and humanitarian demining’, South African Geographical Journal 82: 56–63 Obermeyer, N (ed.) (1998) ‘Special content: public participation GIS’, Cartography and Geographic Information Systems 25(2): 65–66 Talen, E (1999) ‘Constructing neighborhoods from the bottom up: the case for resident generated GIS’, Environment and Planning B 26: 533–554 Weiner, D., Warner, T., Harris, T and Levin,... basis and for GIS practitioners to seriously engage with the local structures of civil society CiGIS contributes to greater access to, and sharing of, valuable spatial information, and provides multiple representations of past, present, and future landscapes It also highlights the conflictual nature of spatial decision-making and acknowledges GIS practice as both technological and political (Lupton and. ..256 T M Harris and D Weiner The human, financial, and technical resources required for GIS development and operationalization suggest that in-house community GIS are unlikely to broaden access to spatial decision-making, especially in underdeveloped regions and poor communities (Hastings and Clark 1991) Rather, alternative mechanisms for data and GIS delivery that relieve communities... Craig, W., Harris, T and Weiner, D (1999) ‘Empowerment, marginalization and public participation GIS’, Specialist Meeting Report compiled for Varenius: © 2002 Taylor & Francis 258 T M Harris and D Weiner NCGIA’s Project to Advance Geographic Information Science, NCGIA, University of California, Santa Barbara, February Harris, T and Weiner, D (1998) ‘Empowerment, marginalization and communityIntegrated... hardware, software, data, expertise, and maintenance costs will most likely predominate CiGIS seeks to address such concerns in GIS-based decision-making In the future, Internet-based GIS will be a core component of such an access and delivery system, although the social context of Internet access will vary significantly from community to community It is the integration of community viewpoints that has dominated... historical geographies, and future aspirations Black participants’ consciousness about the landscape was formed though historical processes of dispossession, and many are anxious to discuss local natural resource politics and power relations Largescale commercial farmers, on the other hand, were much more reluctant to draw maps and discuss configurations of power within the local and regional landscape It is... 2002 Taylor & Francis Implementing a community- integrated GIS 257 decision-making using CiGIS remains a significant challenge, particularly in the context of socially differentiated knowledge, perceptions about landscape and uneven access to GIS resources Nevertheless, there is considerable interest in South Africa to link community participation with GIS (Hill and Strydom 2000; Mather 2000) To achieve . 1:50,000. Land-cover data were obtained at 1:250,000, and land-type data were captured at 1:50,000. CiGIS requires that traditional top-down ‘expert’ information be com- plemented by information. arable land. Forestry companies control large tracts of state land, and this raises substantive issues regarding socially and ecologically appropriate land-uses. Forest plantations and large- scale. identifying and comparing ‘expert’ and local understandings of land potential, © 2002 Taylor & Francis 252 T. M. Harris and D. Weiner 3 perceptions of socially appropriate and inappropriate land-use, 4