First things first

Stefan Dongus
Axel W. Drescher

Despite the amount of research on urban agriculture, very little is actually known about the extent of farming activities in many cities. After years of ignoring or circumscribing urban agriculture, some cities in developing countries have begun to embrace the phenomenon, as they now see its benefits for poverty reduction and for improving the urban environment. But before municipalities can begin to develop institutional support for urban agriculture, they need to know exactly what is going on, and where. That’s where GIS comes in.

After years of ignoring or circumscribing urban agriculture, some cities in developing countries have begun to embrace the phenomenon, as they now see its benefits for poverty reduction and for improving the urban environment.

Unfortunately, even in those cities that have changed their minds, and in spite of ongoing research, few systematic surveys have been carried out to explore the actual extent of agriculture in inner city areas. Very few data are available regarding the extent, the importance, the development and the output of urban agriculture, as well as its connection to small enterprise development. Basically, no one really knows very much about what goes on where.

Thus, before municipalities can begin to design better regulations that will facilitate urban agriculture and incorporate it into their planning, they need to produce maps of urban farming activities. Many questions remain unanswered. Where are urban agricultural activities concentrated, and why? What kinds of crops are being grown? In what types of soil? By which groups within the urban population? How available is water and what is its quality? How far is it from farm to market? Are there potential health and environmental risks?

Globally, there have been only very limited applications of geographical information systems (GIS) to urban food production activities. In the Dominican Republic, for example, researchers from the University of Santiago have developed a GIS database covering Santiago de los Caballeros. In an EU-funded project in Ouagadougou, Burkina Faso, researchers have mapped urban and peri-urban agricultural areas using IKONOS satellite imagery. To a more limited extent, GIS has been used to map urban agricultural activities in the cities of Accra and Kumasi in Ghana.

Open-space survey

The city with the most comprehensive experience in the use of GIS is Dar es Salaam in Tanzania. In 1999, the German Agency for Technical Cooperation (GTZ), Dar es Salaam city council and the Ministry of Agriculture and Cooperatives collaborated in a survey project that showed that one possible way to close the information gap is to map urban agricultural areas by combining analysis of aerial photographs with field work and GIS. The main motivation for the survey was to create an opportunity for urban farmers to receive additional support from various stakeholders concerning questions of land tenure, water supply, infrastructure, extension services and training. The methodology developed at the time has since been used as a model and adapted for similar projects elsewhere.

Open spaces play an important role in cities and can provide a buffer against food insecurity in times of crisis. In Dar es Salaam the exact locations and areas of open spaces larger than 1000 m² were mapped and integrated into the city council’s GIS database. The actual area used for vegetable production in the city’s open spaces was then assessed, totalling four per cent of the whole surveyed area in 1999. Furthermore, conclusions could be drawn regarding the dynamic development over a seven-year time span. This gave indications about the importance of this type of urban land use, mainly in terms of its viability in view of competing demands for land for housing, infrastructure, and so on.

The knowledge base can be used by town planners, city officials and policy makers for decision-making concerning the place of urban agriculture – especially vegetable production – in the city’s development. The urban agriculture map also offers an entry point for support organizations to contact urban farmers. Further, the evidence created by this map can help raise public awareness and acceptance of urban vegetable production.

Why GIS?

The use of GIS technology offered several benefits, in particular the visualization of spatial data, particularly the distribution of agricultural open spaces in a city, with simple analytical functions to calculate the size of the areas used for farming. The technology also permitted data overlay in order to investigate relations between various relevant factors, such as designated land uses, irrigation water quality, and socio-economic variables. Researchers were also able link vector data (digital data in the form of points, lines and polygons having geographical positions and shapes defined by a set of coordinates) in maps with attribute data such as the type of crop grown, sources of water or number of farmers.

In the future, projects will be able to update the digital maps and extend them to include a greater range of topics and layers. It is also possible to print out hard copies of maps showing any desired topics and areas, at any scale, for discussions with local stakeholders.

GIS is highly flexible. Depending on the local context and available data sources, a wide variety of spatial data can be integrated and combined, including satellite imagery, aerial photography (digital or analogue), topographic or thematic maps at all scales, cadastral maps, and global positioning system (GPS) measurements.


Given that GIS and aerial photography, high-resolution satellite imagery and GPS receivers are now widely available, the methodology used for this project can be used in a similar way in cities anywhere in the world. In Dar es Salaam, an area of 165 km² was surveyed over a period of four months. First, the researchers analyzed analogue aerial photographs (stereo pairs), then carried out the fieldwork, and followed this up with the digitization, visualization and analysis of the results using the GIS.

The analysis of the aerial photographs was the essential step, making it possible to map a large area in a short time, and to produce very precise information on actual land use. The analysis was based on the most up-to-date available set of aerial photographs of the Dar es Salaam region (black and white orthophotos, stereo pairs, at a scale of 1:12,500), taken in 1992. The photographs were analyzed with a stereoscope to identify all open spaces used for agricultural production. The borders of these areas were drawn onto the respective cadastral maps (scale 1:2500) of the Dar es Salaam region. The high resolution of the aerial photographs and the large scale of the cadastral paper maps made it possible to indicate the position of the boundary lines with a high degree of accuracy. Field visits were then made to confirm the boundaries on the ground, and thus to avoid mistakes during this phase.

An alternative option would have been to use digital remotely sensed imagery with a resolution higher than one metre. However, such digital imagery was not available in Dar es Salaam at the time of the survey, and, even if it had been, it would have been less accurate than the analogue aerial photos.

Fieldwork – paper maps or GPS?

Project staff visited all of the areas identified as productive open spaces during the aerial photograph analysis on motorcycle. All sites were checked. However, some sites turned out to be inaccessible – for example, some areas were on land controlled by the army. The extent of farming at each site was compared to the situation in 1992. In almost all cases, this was possible simply by working with the cadastral maps and observation. In case of any changes, the new situation was marked in the maps. In a few cases, GPS measurements were taken to map the newly emerged open spaces, and to measure significant changes to areas inaccessible to farmers. If farmers were found at a site, the researchers asked them to participate in a short questionnaire survey regarding ownership of the site, access to water, and any general problems they had. The researchers then asked them for any further comments they might have.

At the time of the survey in 1999, it was necessary to use differential GPS in order to obtain adequate accuracy (a margin of error less than 3 m). This required the simultaneous use of two GPS receivers, one stationary and the other a mobile unit for use in the field.. The geographical coordinates of the corner points of agricultural open spaces were recorded, saved in the receiver and later downloaded to a computer. Experience showed that it was much more efficient to work without the GPS receivers for the purpose of the study. In most cases, it was sufficient and even more accurate to observe the areas and mark the observations directly onto cadastral maps. An alternative option would have been to use enlarged copies or prints of the aerial pictures on which to mark the results of the field observations. With paper maps it was easy to show and explain the purpose of the work to community leaders and members, whereas the use of GPS equipment in the field was sometimes problematic as it caused suspicious reactions from city residents who were not familiar with this technology. Today, GPS as a technique has become much more accurate, and the receivers are a lot easier to use. However, the researchers recommend the use of paper maps or printed aerial pictures as a basis for field mapping, and the use of GPS only if orientation in the field is not possible otherwise.


Finally, the results of the analysis of the aerial photographs and the fieldwork were digitized using MapInfo, a software package that the Dar es Salaam city council was using as their standard GIS at the time of the survey. The borders of open spaces drawn onto the cadastral maps were digitized as polygons with a digitizing tablet. In cases where the areas were measured by GPS, the downloaded coordinates were read into the GIS software and connected manually to form polygons. A digital base map of Dar es Salaam (including the positions of roads, railway lines, rivers and the coastline) was already available. Attribute data collected in the questionnaire survey was then entered and linked to the respective polygons.

The database that resulted from this work can be used by town planners for further analysis and planning purposes, and to make overlays with other relevant spatial data.
The developed methodology is entirely applicable to other cities and can be expanded to incorporate other topics of interest. Indeed, the methodology is currently being used as part of a research project on the use of wastewater in urban and peri-urban agriculture in India and Pakistan. The same methodology is also being used in a participatory mapping project to investigate malaria hazards in relation to Dar es Salaam’s urban agriculture.

Urban agriculture in Dar es Salaam, as in many cities in the developing world, may well be manifestly apparent as a form of land use, but without a detailed mapping of the precise extent of such activities, municipal authorities cannot begin to develop institutional support for it. Integrated into local government planning processes, the GIS data can be used to help raise public awareness of the benefits of urban farming, to improve regular access to land, aid extension services, and can be used by town planners for further analysis and planning purposes. Participatory mapping is the first step on the road to the normalization and promotion of urban agriculture.

Stefan Dongus ( ) is a geographer and mathematician, and Axel W. Drescher ( ) is professor of geography at the Department of Physical Geography, Section on Applied Geography of the Tropics and Subtropics, University of Freiburg, Germany.

24 November 2006

Copyright © 2014, CTA. Technical Centre for Agricultural and Rural Cooperation (ACP-EU)