Spatial stratification of western Kenya as a basis for research on soil fertility management

Spatial information is increasingly used in applied agricultural research for planning and priority setting. Considerable inter est has been shown in the potential use of Geographic Information Systems for generalisation of results and agricultural technology,. This paper provides an example of an as yet neglected al ea of enquiry in agricultural research, that of synthesis. A case study from western Kenya is presented, in which a spatial stratification is constructed, drawing on diverse environmental and human variables, to orientate soil fertility management research. It begins with a conceptual model of the factors thought to influence soil management in the region, and uses this to select appropriate variables. Multivariate statistical techniques are employed to construct the stratification, although it is emphasised that simpler methods can be equally effective. The resulting spatial strata are described and the logic of the classification is explained with reference to the conceptual model. Gaps in the model are identified by this process. The utility of the strata for future soil management research in western Kenya is explained. (C) 1997 Published by Elsevier Science Ltd.