B.M. Mati, F.N. Gichuki R.P.C. Morgan, J. N. Quinton, T. Brewer and H.P.Liniger
GIS DATA FOR EROSION ASSESSMENT IN THE UPPER EWASO NGIRO NORTH BASIN, KENYA

GIS DATA FOR EROSION ASSESSMENT IN THE UPPER EWASO NGIRO NORTH BASIN, KENYA

A simple Geographic Information Systems database is prepared for the Upper Ewaso Ngiro North basin of Kenya, using limited technological and financial resources. The basin forms part of the greater Ewaso Ngiro River drainage basin, the largest and least studied river basin in Kenya. The database was necessary as a first step in the data acquisition for the assessment of soil erosion. The main outputs were thematic maps of the basin, covering mean annual rainfall, agro-climatic zones, soils, land use/land cover, and the drainage system. Due to the scale and the facilities used, the maps are suitable for resource assessment at reconnaissance scale, and therefore, applications at higher resolutions would require some further detailed studies. The GIS software used in this work were PC Arc-Info and Arc-View.


1.0 INTRODUCTION

Sustainable development of natural resources in a developing country such as Kenya is hampered by material and technological limitations, and the gaps in the basic infrastructure needed to get started. This problem becomes especially manifest in the arid and semi-arid lands (ASAL) of Kenya, where the physical, environmental and socio-economic structures for the development of a reliable database are either lacking, or in very poor state. Yet the ASALs take up about 80% of the total land area in Kenya. They are mainly hot and dry, with highly variable rainfall (in space and time), and evaporation rates which are twice the annual rainfall. The soils have low organic matter content due to low vegetation density and microbial activities. Soil-water storage is rather limited and the soils are very susceptible to degradation. Soil erosion and land degradation are widespread in these areas. Data acquisition and management, especially georeferenced information, plays a vital role in the planning of conservation strategies. The GIS database for the Upper Ewaso Ngiro basin was prepared especially for use in soil erosion monitoring and management under such circumstances.

1.1 The Upper Ewaso Ngiro North Basin

The Ewaso Ngiro North makes up the largest drainage basin in Kenya, covering a total of 210,226 km2 which is predominantly ASAL. It lies north to north east of Mt. Kenya and the Nyandarua (Aberdare) range. Although the main river originates from the Nyandarua range, the tributaries originating from Mt. Kenya supply most of the flow. Whereas the surface flow from the Ewaso Ngiro river disappears into the Lorian Swamp in Kenya, subsurface flows continue eastwards to recharge rivers inside Somalia, which eventually drain into the Indian Ocean.

The Upper Ewaso Ngiro North basin (or simply Upper Ewaso Ngiro basin), is the upstream section of the greater Ewaso Ngiro River basin, bounded by the natural topographic divide, and controlled downstream at Archers' Post. The basin covers an area of 15,634 sq.km between latitudes 0o 20' south and 1o 00' north and longitudes 36o 15' east and 38o 00' east (Figure 1). The biggest town in the region is Nanyuki, situated 200 km north of Nairobi. Although the basin traverses a diverse topography and climatic zones I -VI (Sombroek et al, 1980), about 70% of the basin comprises what is known as the "Laikipia Plateau".

The Laikipia Plateau is a zone of transition from the wetter to drier part of the eastern Kenya highlands. A large proportion of the central region is under large scale ranches, while wheat and barley are grown on the higher (wetter) altitudes. Small scale subsistence settlements are also spreading. Pastoralists inhabit the northern region, an area having a harsh and fragile environment, where widespread overgrazing, soil erosion and general land degradation are quite rampant.

Location of the Upper Ewaso Ngiro North Basin in Kenya

1.2 Use of GIS in Kenya

In Kenya, the use of GIS is rather limited. The reasons for this include the high cost of computer hardware and software necessary to set up a GIS station (Yimbo, 1992) and the lack of trained manpower coupled with the high cost of the training courses. Acquisition of georeferenced data is also an expensive undertaking, including the data management and dissemination. Another limitation is poor consumer awareness which means less demand for the products and services of GIS.

These limitations notwithstanding, GIS has been used in Kenya for several projects with good result, for instance, in compiling the National Water Master Plan (Republic of Kenya, 1992). The Kenya Wildlife Services (Kariuki, 1992) uses GIS for managing the large volumes of data they acquire relating to wildlife census, vegetation and landuse dynamics, infrastructure, security and planning of operations. The Department of Resource Surveys and Remote Sensing makes use of GIS and Remote Sensing (Ottichilo, 1986) spatial data management and handling, for natural resource inventories (forest cover, wildlife and livestock populations, environmental parameters) and information on land use, crop cover and yield, including crop production forecasts. GIS has also been used to prepare the National Environment Action Plan (Ministry of Environment and Natural Resources, 1994), and to monitor a development programme in Laikipia District (Hoesli, 1995). Many of these projects make use of Arcinfo software.

2.0 COMPILATION OF THEMATIC DATA

In the case of the Upper Ewaso Ngiro Basin, GIS data was required for assessment of soil erosion. In general, GIS data for the basin is also required for resource monitoring (soils, topography, wildlife, rainfall, vegetation cover, land use characteristics and crop monitoring), and for planning and management of projects such as soil conservation, water supply, irrigation and drainage, wildlife and range management. Achieving such a GIS database would require computer, financial and material resources currently unavailable. Therefore, the basic thematic maps to be used as a first step in soil erosion assessment were prepared using PC Arc-Info version 3.1.4d and Arcview version 3.0. These are maps of rainfall, agro-climatic zone map, soils, landuse/land cover and drainage.

2.1 Rainfall map

Rainfall data from 94 geo-referenced stations in and surrounding the Upper Ewaso Ngiro Basin were used. The rainfall records were obtained from diverse sources including government ministries, individual farmers, and research and other non-governmental organizations (Berger, 1989, Thomas, 1994). The periods used for the calculation of the mean were therefore varied ranging from about 5 to over 50 years.

The mean annual rainfall data was put in a spreadsheet with the GPS of each station. As the basin lies in the rain shadow of two mountains, six stations outside the basin boundary and on the windward side of the mountains, that were causing overestimation of rainfall on the lee-ward side were removed. Extra data was added in the northern areas where only a few stations were available, by interpolation of rainfall maps from Jaetzold &Schimdt (1983). The file was saved as comma delimited text. The text file was then imported into Autocad and the data interpolated in using the contour function. The interpolated data was exported as a DXF file. This DXF file was imported into PC Arc-Info and the basin boundary used to clip the final map. The result was the rainfall map (mean annual average) shown in figure 2.

Rainfall Distribution in the UpperEwaso Ngiro North Basin Rainfall in the Upper Ewaso Ngiro Basin varies with altitude, and since the Basin lies in the lee slopes of Mt. Kenya and the Nyandarua range, the area is generally dry. Rainfall ranges from 365 mm per annum at Archer's Post to over 2000 mm on the Nyandarua range. However, the average annual precipitation for most of the Basin is about 700 mm. A varied rainfall distribution is found in the Basin. In the western and north western parts, rains occur in a single season, between April and August. The eastern side has a clear bimodal distribution with rainfall maxima in April and October. The central region is a transition zone, where the two patterns overlap. Temperatures are relatively low, with mean annual temperatures ranging about 18-200C (Jaetzold & Schmidt, 1983).

2.2 Agro-climatic zone map

In the Upper Ewaso Ngiro basin, land use, crop performance, range management and other social and economic activities are more closely associated with the agro-climatic zones than with rainfall amount. This is because these zones also show the effects of temperature and evaporation, which are important for agricultural production. The agroclimatic zone map of Kenya (Sombroek et al 1980) was digitized to obtain the agro-climatic zone map. This was due to lack of more recent data that covers the entire basin. The map preparation was done in three stages as follows:

First the map of the moisture availability index, which is a ratio of annual rainfall to potential evaporation, was digitized. Next the temperature zone map was digitized. The agro-climatic zone map was then obtained by overlaying the two maps in Arc-Info using the identity function. The resulting agro-climatic zone map is shown in figure 3.

The agroclimatic zone map shows that high moisture and low temperature gradients are associated with increasing relief. Moisture gradient increases from zone I to VI, while temperature decreases with zone 1 being hottest (about 33oC) and zone 9 coldest. Therefore, zone VI-1 is very hot and very dry, while zone I-9 is very cool and very wet. The other zones lie between these two extremes. The central part of the basin, which forms the Laikipia Plateau lies in moisture zones IV-VI, and temperature zone 4-5. This indicates a relatively cool dry region, which is normally unsuitable for rainfed crop production. The hot, dry areas are found in the north of the basin.

Agroclimatic Zone Map of the Upper Ewaso Ngiro North Basin

2.3 Soils map

Soils maps are important in land and water management. Activities such as conducting feasibility studies for land use planning, assessment of erosion, agriculture and environmental conservation rely on a good knowledge of the soils resource. The soils data for the Upper Ewaso Ngiro basin were obtained from the Exploratory soil map of Kenya (Sombroek et al, 1980), which was digitized. The resulting soil map is shown in figure 4.

There were limitations in getting a more recent and higher resolution soils map, because although some recent soil survey have been done, they cover just parts of the basin and the whole basin is covered at exploratory scale. As the data required for erosion assessment at this stage was for reconnaissance studies, this scale was adequate. However, for more detailed applications, it might be necessary to consult published soil survey reports such as those by Mainga and Mbuvi (1994) and Kironchi et al (1992).

The soils of the Upper Ewaso Ngiro basin are diverse and they vary with the topography and the geology. Figure 4 shows that about 36 major soil types were identified based on the FAO soil classification (FAO, 1987). The dominant soils include the following (Sombroek et al, 1980);

* Soils on mountains and major scarps. These are developed on older volcanoes and they include, haplic phaeozems lithosols, eutric regosols (FAO Classification).

* Soils on plateaus are developed on tertiary basic igneous rocks. They include; ortho-vertic phaeozems, vertisol and planosols.

* Soils on dissected and non-dissected plains. These are developed from basalts. They include chromic luvisols, ortho-luvic phaeozems and chromic cambisols.

In the Northern areas of Wamba, there are rock outcrops and soils derived from basalts. They include eutric regosols, arenosols, calcic cambisol and haphic xerosols.

Soils of  the Upper Ewaso Ngiro North Basin

2.4 Land use/cover map

The land use/cover assessment for the Upper Ewaso Ngiro basin was done by interpretation of 10 prints of multispectral SPOT 1 HRV imagery of 30 metre resolution. All the prints were taken in the dry-season between 1987 and 1988. Due to technical and financial constraints, it was not possible to purchase more recent or digital data. Image interpretation was done on the prints using training areas representative of different thematic categories selected from known field conditions. This information was subsequently ground truthed in reconnaissance studies. Additional information was consistently added to update the original interpretation. The imagery were then digitized and merged into one coverage showing the land use/cover of the basin. The resulting map, whose classification scheme is described below, is shown in figure 5.

The land use/land cover assessment adopted the Pratt and Gwynne (1977) classification system for East African Rangelands. This classification scheme fitted quite well with the conditions of the Upper Ewaso Ngiro basin. A few modifications were included such as the addition of cropland classes. Eleven major land use/cover classes were identified as follows:

Moorland: Is the afro-Alpine vegetation found at altitudes exceeding 3,500 m above sea level. This consist mainly of grasses and moorland shrubs such as Lobelia keniensis, Dendrosenecio brassica, Carex monostachya and Alchemilla johnstonii.

Forest: Most trees are 7-40 m or taller with crowns often interlocking. Evergreen forests are characterized by individual trees that may shed leaves, but the canopy as a whole remains green throughout the year (e.g. composed of trees such as Olea africana). Deciduous forests are characterized by trees that lose their leaves during the dry season (e.g. Acacia and Combretum spp.).

Bushland: Bushlands consist of woody plants which often have multiple stems, with most of the plants not exceeding 10 m in height. Crowns are often interlocking and canopy cover is over 20%. Trees are scattered but conspicuous. The herbaceous understorey is usually sparse.

Bush Grassland: Bush grassland consists of grassland with scattered trees and shrubs having a combined canopy cover less than 20%.

Grassland: Grasses or sedges dominate these communities. Woody plants are either lacking or are dwarfed and inconspicuous. Woody plants compose less than 2% of the canopy cover.

Shrub Grassland: Shrub grasslands are grassland with scattered shrubs which have a canopy cover less than 20% .

Shrubland: Shrubland consists of woody plants about 6 m tall and without a significant presence of trees. Canopy cover is more than 20%. The herbaceous understorey is usually sparse.

Scarpline shrubs: Vegetation and conditions of the minor scarps characterized by scattered shrubs less than 6 m tall, sparse or no herbaceous vegetation and rocky ground cover.

Small Scale farms: Small scale cultivated lands and mixed farms with varying levels of grass, trees, shrubs, fallow and crop covers.

Large scale farms: Mechanised large scale farms mostly growing wheat or barley.

Swamps: Land covered by permanent standing water and supporting various plant communities including reeds, sedges, rushes, sometimes trees or shrubs and aquatic species.

The land use and natural vegetation types in the Upper Ewaso Ngiro basin depend on the altitude, climate and soils. On mountain slopes, moist montane forests dominate. There are also riverline forests especially in higher altitude, and dry forests in the drier highlands such as Mathews Range. Shrub grasslands and bush grasslands occupy much of the Laikipia Plateau area, while in the plains of Isiolo and Samburu, shrubland is dominant (Thurow and Herlocker, 1993).

The predominant economic activity is animal husbandry, much of it on large commercial ranches, and on communal grazing lands or group ranches. Livestock is also found on small scale farms as part of mixed farming. Wildlife is found in most parts of the basin and there are privately-owned as well as on government-run game reserves. Range management therefore plays a very important role in the sustainable management of the natural resources in the Basin. However, arable farming, especially on small scale farms, is common on highlands and mountain footslopes, where the majority of the population is found.

Land Use/Cover of the Upper Ewaso Ngiro North Basin

2.5 Drainage map

The preparation of the drainage map of the Upper Ewaso Ngiro basin was important for hydrological studies and to be used at a later stage for developing topographic models. It also gives an indication of the terrain characteristics. Drainage was digitized from 31 topographic maps of scale 1:50 000, that cover the basin. Care was taken to digitize in the direction of the streamflow so that future use for development of digital elevation models and slope classification would be accurate. All the watercourses, including minor dry stream channels were digitized, resulting in the drainage map shown in figure 6. Although the basin is traversed by many stream channels, many of them are ephemeral, and others are large gully beds.

Drainage of the Upper Ewaso Ngiro North Basin

3.0 CONCLUSIONS

The GIS database for the Upper Ewaso Ngiro basin was a humble beginning, in the build up of a limited number of thematic maps, with which soil erosion could be evaluated. These are general thematic maps which can also be used for planning and management of other resources. The preparation of the database was achieved with limitions in the technical and financial resources. Some of the limitations included material resources to get higher resolution data, especially in the northern areas, where the physical and environmental infrastructure is very poor. Technical support was scarce as only a few people are trained in GIS in Kenya. Time was a constraint and computer facilities as only PC Arc-Info was available. Powerful hardware and software would be required to prepare digital elevation models, and run large files. The basin itself was large causing computer memory problems, especially with the drainage file. However, despite these limitations, the GIS data that was obtained provides a useful tool for soil erosion assessment in the Upper Ewaso Ngiro basin.

The thematic maps presented here can be used bearing in mind the limitations associated with the data acquisition and map preparation. They were prepared especially for resource assessment at reconnaissance scale, and therefore, applications at higher resolutions would require some further detailed studies. Still, more thematic data is required for the Upper Ewaso Ngiro basin including geology, ground water, population densities and dynamics, infrastructure, weather patterns including even digital terrain models that cover the whole Basin at a suitable resolution. Improvements should be made to bring the current database up-to-date technologically and in scope. These improvements would involve upgrading the available technological and human resources. This would require financial and administrative support, including training and re-training of staff. Also, a commitment by all concerned to face the challenges of the ever-changing technological demands in the dynamic field of GIS.

4.0 ACKNOWLEDGMENTS

The authors appreciate the material and administrative support provided by the Natural Resource Monitoring, Modelling and Management Project (NRM3) of Nanyuki, Kenya; the Rockefeller Foundation, Nairobi Office, and Silsoe College, Cranfield University, United Kingdom. We would also like to thank Mr. J. Ndungu, Mr. J.K. Mitugo, Mr. Kungu and Mr. Githu all of NRM3 for their assistance with the field work; Mr. K. Gitari and Mr. C. Mwangi of NRM3 for their support in acquisition of some of the data.

5.0 REFERENCES

Berger, P. 1989. Rainfall and agroclimatology of the Laikipia Plateau, Kenya. Institute of Geography, University of Berne, Switzerland.

FAO, 1987. Guidelines for soil profile descriptions. FAO. Rome.

Hoesli, T. 1995. GIS based impact monitoring of a development programme. Laikipia-Mount Kenya Papers. No.18.

Jaetzold, R. and Schmidt, H. 1983. Farm management handbook of Kenya. Vol. II/B. Central Kenya. Ministry of Agriculture, Nairobi.

Kariuki, A. 1992. Applications of geographic information systems in the management of the wildlife resource. In: Applications of geographical information systems for efficient data storage and handling in Kenya. Okoth, P.F. (ed.). Proceedings of a symposium. Kenya Soil Survey, Nairobi. Pp.10-19.

Kironchi, G., Mbuvi, J.P. and Liniger, H.P. 1992. Soil type and land use effects on infiltration in Sirima and Mukogodo Catchments, Laikipia District. Laikipia-Mount Kenya Papers. D-15.

Liniger, H. 1991. Water conservation for rainfed farming in the semi-arid footzone North West of Mt. Kenya (Laikipia Highlands). PhD. thesis. Laikipia-Mount Kenya Papers. D-3.

Mainga, P.M. and Mbuvi, J.P. 1994. Preliminary soil conditions of Embori, Kalalu and Mukogodo sites. Laikipia-Mt. Kenya Papers. Laikipia Research Programme. B-10.

Ministry of Environment and Natural Resources. 1994. The Kenya National Environment Action Plan. Summary. Nairobi, Kenya.

Ottichilo, W.K. 1986. Food production, famine and early warning system: The Kenyan experience. In: Proceedings of the 20th International Symposium on Remote Sensing of the Environment. Nairobi, Kenya. Environmental Research Institute of Michigan, Ann Arbor, Michigan, U.S.A. Pp. 177-180.

Republic of Kenya. 1992. The study on the National Water Master Plan. Sectoral Report (S). GIS based analysis. Ministry of Water Development. Nairobi.

Sombroek, W.G., Braun, H.M.H. and Van der Pouw, B.J.A. 1980. The exploratory soil map and agro-climatic zone map of Kenya. Report No. E1; Kenya Soil Survey, Nairobi.

Thomas, M.K. 1994. Development of a streamflow model for rural catchments in Kenya. Laikipia-Mount Kenya Papers. D-11.

Thurow, T. L. and Herlocker, D.J. 1993. Range Management Handbook of Kenya. Volume III,5. Users Guide. Republic of Kenya. Ministry of Agriculture, Livestock Development and Marketing. Nairobi.

Yimbo, S. 1992. Economics of GIS. In: Applications of geographical information systems for efficient data storage and handling in Kenya. Okoth, P.F. (editor). Proceedings of a symposium. Kenya Soil Survey, Nairobi. Pp. 27-31.


Bancy M. Mati
PhD student, School of Agriculture, Food and Environment, Cranfield University, U.K/
Research Associate,
Natural Resource Monitoring, Modelling and Management
P.O. BOX 144
Nanyuki, Kenya

Roy P.C. Morgan
Professor, Soil Erosion Control
School of Agriculture, Food and Environment, Cranfield University, U.K
. MK45 4DT
United Kingdom

John N. QUINTON
Lecturer, Soil and Water Management
School of Agriculture, Food and Environment, Cranfield University, U.K.
MK45 4DT
United Kingdom

Tim Brewer
Lecturer, GIS
School of Agriculture, Food and Environment, Cranfield University, U.K
. MK45 4DT
United Kingdom

Francis N. Gichuki
Senior Lecturer, Department of Agricultural Engineering
University of Nairobi
P.O. BOX 30197,
Nairobi, Kenya.

Hanspeter Liniger
Group for development and Environment
Institute of Geography, Berne University
Hellerstrasse 12
CH-3012 BERNE
SWITZERLAND