AG:TCP/URT/4556 TECHNICAL COOPERATION PROGRAMME ESTABLISHMENT OF A LAND INFORMATION SYSTEM THE UNITED REPUBLIC OF TANZANIA Terminal Statement the Food and Agriculture Organization of the United Nations Rome, 1998 |
AG:TCP/URT/4556 TECHNICAL COOPERATION PROGRAMME ESTABLISHMENT OF A LAND INFORMATION SYSTEM THE UNITED REPUBLIC OF TANZANIA Terminal Statement the Food and Agriculture Organization of the United Nations Rome, 1998 |
Tanzania, with a total land area of approximately one million km2 and a population of about 36 million people, the majority of whom are small-scale subsistence farmers and rural workers, is heavily dependent on its productive land. The agricultural development strategy of the Government is geared towards increasing the domestic production of strategic staple food crops and traditional export crops. It is also stimulating diversification into non-traditional crops for export in order to increase the income of the rural populations.
As the population grows, the demand for food increases and poor land-husbandry practices (including soil conservation), the opening-up of marginal lands for rainfed cropping in semi-arid areas and encroachment into the remaining forest areas are becoming reasons for concern. Increasing soil loss through accelerated erosion, sub-optimal land use owing to the widespread growth of crops which are unsuitable to the prevailing agro-ecological conditions, damage from flooding and the deposition of silt in reservoirs and the coastal waters are some of the negative environmental effects which have been reported.
The Government has become increasingly concerned about the problem of land degradation and attaches great importance to the formulation of a soil and land-use policy as a prerequisite for the successful implementation of its policies for sustainable food production.
Among the issues urgently requiring attention are the protection of forests in the upper catchment areas, sound land use and sustainable agricultural practices, the identification of optimal crop mix and a balanced development among the various sub-sectors in agriculture. Adequate information is also needed to guide agricultural investments.
The Government has realized that the long-term success of the agricultural sector depends on a sound knowledge and management of the country's land resources, in accordance with their potentials and constraints. The formulation of soil and land-use policies requires reliable and up-to-date information on these aspects of the land in the various agro-ecological zones of the country. The identification of the production potential of each of the major types of land use in Tanzania, notably various crops and livestock systems, based on a systematic and quantitative land suitability and productivity assessment, is therefore urgently required. However, the information required to carry out such an exercise is scattered and not readily accessible. There is also a lack of appropriate tools to enable the rapid integration and analysis of the various data sets and to assess land potentials for a variety of purposes. The inadequate supply of information (notably in the quantity and quality of land resources and farming systems information, uncheckable data and variations in the standards or interpretations of data) has compounded this problem.
The Government of Tanzania therefore requested FAO assistance in establishing a land information system to provide the basis for planning and policy formulation at sub-national level. The project would provide specialized technical assistance and essential hands-on training and experience to local staff in the establishment and development of a computerized land information system and its applications. It was to establish a basic system capable of producing in a rapid and efficient manner land-resources information products for various purposes, notably for agricultural land-use planning at the district level, and to demonstrate its use. The land information system was expected to contribute towards sound land-use planning and policy formulation.
This assistance was approved by FAO under the Technical Cooperation Programme project TCP/URT/4556, "Establishment of a Land Information System". The Project Agreement was signed by the Government and FAO on 6 December 1995, with a total budget of $US 260 000 and a duration of 18 months. The project started officially on 1 January 1996, with the Soil Conservation and Land Use Planning Section (SCLUPS) of the Ministry of Agriculture and Cooperatives designated the government agency responsible for project implementation.
After delays in the installation of hardware in the new premises of the SCLUPS at the Mikocheni Agricultural Research Station near Dar-es-Salaam, the original project completion date was extended to 31 December 1997. The project lasted for a period of 7-8 months. The phased approach foreseen in the Project Agreement was revised accordingly. The original four visits by the international consultants were reduced to two, while the four remaining phases of the project were reduced by combining phases II and III and phases IV and V. This speeded up project implementation but reduced the time available for local staff training in the use of land information systems, notably the Geographic Information System (GIS) and automated land evaluation. The three-month period available for a national consultant in land-use planning was divided into 1.5 m/m for a national consultant in land resources and another 1.5 m/m for a national consultant in crop/forestry production systems.
The objective of the project was to create the technical basis to support the rational development and conservation of the country's soil and land resources. This was to be achieved by means of the following specific outputs:
- a computerized natural resources information system, consisting of thematic databases covering agroclimatic factors, soils, topography and physiography, vegetation and present land use, and linked to a GIS able to display combinations of this and other data, and maps showing different combinations of the information;
- land evaluation procedures to identify land-use options for each identified land unit, in terms of potential performance or yield, cost/benefit ratios and long-term environmental impact. This would include land-evaluation procedures for the assessment of land suitability for a number of forestry species;
- a demonstration of the capacity of the installed system to assist in identifying the best mix of land uses (including forestry) for a given land unit, using both national and regional objectives;
- the training of local staff to operate the system; and
- the provision of the relevant computer hardware and software.
A GIS-based pilot land information system was established and made operational within the SCLUPS. The system was an important step in the introduction of computer technologies to the Ministry of Agriculture and Cooperatives. Its capacity to process land resources information in different ways for different purposes makes it a powerful tool in land-use planning and land resources management. The SCLUPS has now become a service centre capable of processing information from different sections of the Ministry and other governmental and semi-governmental agencies. Spatial information can be handled and made readily available to users in a rapid and efficient manner.
The system greatly enhances SCLUPS capability in agricultural and forestry land-use planning at the district level. It facilitates the identification of both constraints and potentials for production, as well as areas that should be earmarked primarily for conservation and protection.
The land information system consists of three basic components: a set of geographically referenced databases (including knowledge databases); a land evaluation programme (containing various land evaluation models); and an application programme for mapping both physical and economic land suitability at two management/input levels.
The geographically referenced data and knowledge databases were created for Morogoro District, which was selected by the Government as a pilot area for system development. Data were compiled and organized in the database as GIS coverages, including digital maps and tabular databases on the following themes:
- topographic information (district boundary, railroads, major roads, streams and rivers and relief using contour lines with 200 m intervals);
- agroclimate (following an analysis of historical climatic records stored in a separate agro-climatic database containing mainly monthly rainfall data: agroclimatic zones were identified, based on the length of growing period, mean annual rainfall, mean annual temperature and mean temperature during the growing period);
- soils/terrain (soil classification, slope class and numerous soil characteristics were stored in a soil/physiography database);
- plant environmental requirements: climatic and edaphic requirements for both forestry (11 tree-species) and rainfed cropping (36 crops);
- farming-production systems data including crop name, utilization, market orientation, management units, agronomic practices and cropping characteristics, such as labour requirements, land preparation, propagation, recommended varieties, planting, fertilizing, weed control, etc., pests and diseases, yields, detailed costs of input and output prices; and
- forestry-production systems data.
The layout of all the maps produced to show the geographic distributions of the first three themes above was standardized to a scale of 1:250 000 (although they can be presented at any scale) and legend compilation was automated as far as possible.
Land mapping units were created by overlaying climatic, edaphic and physiographic characteristics. These units consisted of two or more land units (i.e., geographically referenced data pertaining to the physical environment) and formed the basic unit for land-use planning. For Morogoro District a total of 167 land mapping units consisting of 636 land units was identified.
Land-unit data from the databases was imported into the Automated Land Evaluation System (ALES), containing land evaluation models to assess land suitability for different kinds of land use. The results generated by the ALES were exported back into the GIS and computer programs were prepared to show ALES outputs in map form in terms of physical and economic land suitability. One example of land-use planning was elaborated, including the identification of areas best suited for various types of land use.
Two major types of land use were considered in the land evaluation exercise: forestry (physical land suitability assessment only) and rainfed cropping (both physical and economic land suitability for two levels of management/inputs: low and medium corresponding to subsistence and semi-commercial farming).
The physical land suitability assessment for rainfed cropping covered 36 single crops or land utilization types. Ten land qualities composed of 18 major land characteristics were used. For forestry, 11 tree-species were considered and six land qualities composed of nine land characteristics were used. The physical land suitability assessment classified the land into one of five suitability classes varying from very suitable (expected yields in the range of 80% to 100% of maximum) to unsuitable (below 20% of maximum). Wherever conditions were less than very suitable, the limiting factors were indicated (e.g., insufficient soil nutrients, inadequate drainage or sub-optimal temperature regime), facilitating appropriate management actions.
For the economic land suitability assessment, data on yields, production costs and produce prices were derived from the production systems database and entered in the ALES. For perennial crops a planning horizon of 15 years (from initial planting) was used, while for annuals the planning horizon was limited to one year. Gross-margin analyses were then carried out for each land unit having a certain degree of physical land suitability for a particular crop and results were presented in terms of net returns and divided into four economic suitability classes (based on selected income threshold values). The ALES also indicated expected yields, total production costs and gross returns for each land utilization type/land unit combination.
The land evaluation results for Morogoro District were prepared in the form of disk files (for further processing and archiving) and hard copies (for ease of reference). The following maps and associated statistical printouts were produced:
- for forestry: physical land suitability classes and limiting factors for each of the 11 tree-species considered;
- for rainfed cropping (for both low and medium management/input levels): physical land suitability classes and limiting factors for each of the 36 crops considered, sorted in decreasing order of suitability; economic land suitability (based on gross-margin analyses) in terms of net returns for each land unit, sorted by crop in decreasing order of net returns; economic land suitability (based on gross-margin analyses) in terms of economic suitability classes based on preselected threshold values for each of the 24 selected crops.
To demonstrate the use of the system, the project prepared an application in land-use planning at the district level. This application involved dividing areas into three categories. Areas that were identified as unsuitable for both rainfed cropping and forestry based on major terrain characteristics (such as very steep slopes or shallow, rocky and stony soils) were classified as "areas for conservation", to be protected and preserved under their natural vegetation cover.
Following an overall assessment, other land units were identified as suitable for rainfed cropping under both subsistence farming (low management/input level) and semi-commercial farming (medium management/input level). This assessment was based on the overall land suitability of each land unit for each of the 16 crops considered under each model. Physical land suitability was used to assess overall land suitability for subsistence farming, while for semi-commercial farming the overall assessment was based on economic land suitability. Areas with agricultural potential were divided into three categories, based on the land suitability ratings obtained for each of the 16 crops considered, and classified as "high, moderate or low potential areas for rainfed cropping".
For land units that could not be classified in the above categories the potential for forestry development was assessed in terms of physical land suitability for the tree species considered. Land units falling into this category were classified as "areas of moderate potential for forestry".
Any remaining land was classified as belonging to the category of "other types of land use".
The project carried out four types of training: on-the-job training, formal training, a study tour and a workshop.
Three staff members from the SCLUPS were assigned to the project initially, but in August this number was reduced to two. They were trained in the use of all or some of the system components (as appropriate). SCLUPS staff members are now capable of operating and maintaining the system, although they will need further training to ensure that they can use the system to its full capacity. Currently they are able to enter and update data and perform standard processing procedures and are familiar with the main GIS functions, such as digitizing, editing and querying coverages and generating map outputs. They are capable of running land evaluations in the ALES but will need further guidance in building computerized land evaluation models.
A one-week training course in ARCView was provided, attended also by a staff member of the Tanzania Resource Information Centre (TANRIC). At the end of the course, certificates were awarded to those who had passed a formal examination. The National Project Director and two staff members went to South Africa for some weeks to familiarize themselves with database and GIS applications in land-use planning and natural resources management.
Finally, a two-day workshop was held in December 1997. The purpose of the workshop was to present the system to representatives from relevant government agencies, to explain the methodologies and procedures involved, to show the outputs obtained and applications in district-level land-use planning, to demonstrate its functioning and to discuss follow-up activities. The workshop, which was attended by 22 participants, proved to be timely and successful and strongly stimulated interest and discussions in land information systems.
A follow-up project proposal was prepared by the National Project Coordinator and reviewed by participants during the workshop. After a number of suggestions for improvements had been made, it was strongly recommended that the proposal be elaborated by the Government into a final project proposal to be submitted to donors for funding. The objectives of such a project would be to develop the core land-resources information system installed by the project into a fully-fledged system capable of handling all major types of land use encountered in the country (incorporating livestock and extensive grazing models, irrigated cropping models, etc.), to elaborate and enhance system applications and to provide additional training to staff operating the system.
The introduction of computer-assisted technology, methodologies, approaches and procedures in land-use planning, staff training and the compilation of databases pertaining to the country's physical environment and crop production systems has strengthened the position of the SCLUPS in district-level land-use planning. The fact that SCLUPS staff members have acquired skills in computerized operations has generated interest among the staff of various governmental and semi-governmental organizations in GIS/land-information-system-related staff training and advice on equipment. SCLUPS' central role in providing information on natural resources and land-use issues has been reinforced and SCLUPS staff members are now in a better position to advise their colleagues in the district offices.
The following recommendations focus on applying the system, disseminating the results to users and enhancing the capacity of the system through the development of further applications in agricultural land management and use planning, crop diversification, land-use optimization, policy formulation and protection of the natural resource base.
It will be necessary to upgrade both hardware and software as new products become available. The lack of hard-disk memory is already becoming a constraint to complex GIS operations/data processing.
Computers, plotter, digitizer, printers, software and databases are subject to intensive use. Maintenance should therefore be provided on a regular basis to keep the land information system operational. Such maintenance will include regular back-ups and the removal of inessential system files. Checking for viruses should be carried out regularly and the copying of diskettes should be strictly controlled. It is strongly recommended that the Ministry provide appropriate budget allocations to cover expenses on computer consumables.
System outputs, such as maps, should not be provided free of charge. It is recommended that the Government set a realistic price for information products to recover at least the costs of the materials used. These funds should remain with the SCLUPS in order to maintain the system.
Regular updating of the information contained in the system is essential. This applies to all database components, although more frequent updating will be required for the production systems databases, as costs related to inputs and outputs change continually. Less frequent updating will be needed for the land resources data. It is recommended that the crop production systems database be updated at least on a semi-annual basis, that the climatic database be updated at least on an annual basis to include, for example, new monthly rainfall records and that the digital information on land use/land cover be obtained as rapidly as possible from TANRIC in order to compare actual land suitability and land use with potential land suitability and land use.
The geographically referenced databases can be used for such purposes as land-use planning, natural resources management, land suitability assessments, crop diversification programmes and environmental impact assessments. They should be updated and revised as new information becomes available. The system can be applied to assess the potential for new crops, to identify constraints in crop and forestry production, to identify high, low and moderate potential areas for rational and sustainable agricultural development, to reassess current land use versus actual land suitability and optimal land use, to determine areas for intervention, to provide guidance on adequate fertilizer applications, to identify areas for conservation and protection of the natural resources and to serve as a benchmark for monitoring future changes in land use/vegetation cover and identifying potential land-use conflicts.
Although at present the system includes only rainfed cropping and forestry, it could easily be expanded to incorporate other aspects of the agricultural sector, such as irrigated cropping or livestock production. This would require the development of separate land evaluation models for these categories. Ideally, such models need to be designed in order to obtain most information from databases compiled particularly for this purpose.
For agricultural diversification purposes the plant requirement database should be expanded to include other crops of interest, particularly those which are not currently grown within a district but for which physical and socio-economic conditions appear to be favourable. Other kinds of land use, such as extensive grazing by livestock and wildlife should be incorporated. Land suitability assessments should be carried out and results presented in map and tabular formats. This will be of particular interest to private investors in the agricultural sector.
The forestry production systems database needs to be completed and the necessary data compiled as soon as possible in order to allow a full assessment of the potential for forestry land-use types. Once this has been completed, an economic assessment model needs to be appended to the existing physical land suitability assessment model. This will further enhance the applications of the system in land-use planning. Enhancement of the system, by incorporating evaluation models for other important sectors of the economy with an influence on land use, may contribute to wider system applications in land-use planning. It is therefore strongly recommended that the database be expanded with information relevant to livestock and wildlife production systems, forestry, land tenure and land-use data, etc., as and when available, and that appropriate land evaluation models be designed to incorporate these major types of land use.
In order to facilitate the development and applications of the system, it is essential that additional staff training be provided in computer operations, particularly basic programming to prepare interfaces, build evaluation models and process ALES outputs. Alternatively, a system manager post could be created within the SCLUPS. This would also facilitate system maintenance and enhance the overall use of the system.
Further training will also be required in the use of information outputs and results.
The Government should provide the resources required to maintain and apply the system and assist various ongoing district-level development activities, such as the Dutch-funded District Rural Development Programme and the Swedish International Development Authority-funded Land Management Programme. The land information system could also be made relevant in matters related to food security (subsistence and semi-commercial farming) and environmental issues (district-level natural resources management and protection). A possible linkage and close collaboration with TANRIC should also be explored. It is recommended, finally, that the Government finalize a follow-up proposal focusing on district-level system applications and identify a donor as soon as possible.