Characterization of the spatial distribution of farming systems in the Kenyan Highlands J.A. van de Steeg a , P.H. Verburg b, * , I. Baltenweck a , S.J. Staal a a International Livestock Research Institute, P.O. Box 30709, Nairobi, Kenya b Department of Environmental Sciences, Wageningen University, PO Box 47, 6700 AA Wageningen, The Netherlands Keywords: Land use Farming systems Spatial distribution Classification Mapping Kenya abstract Land cover change maps are not sufficient to identify subtle changes in land use and farming systems. This paper describes a method that is developed to identify the spatial distribution of farming system types without the need to extensively map all farming systems across a large region. Moreover, it explains differences between farming systems based on spatial variation in environmental and socio-economic conditions. In the study area farming systems were characterized and classified based on the criteria area under cultivation of both food and cash crops, milk production and the usage of fertilizers. Logit models were fitted to explain differences in farming system using location factors and household characteristics. A model based on an integrated set of household and location factors best described the diversity of farming systems across the region. However, the location factors alone also described a larger part of the diversity. The spatial variation in location factors and household characteristics were used to determine the likelihood of occurrence of the different farming systems across the study area. By assigning the farming system to a location that best fits the local conditions based on the logit model a regional level farming systems map for the Kenyan Highlands was created. The methodology provides a tool of analyzing spatial variation in farming systems complementary to the analysis of farming systems at the household level and gives insight in the spatial determinants of these systems. Ó 2009 Elsevier Ltd. All rights reserved. Introduction In recent years much research efforts have been made to better understand and model land use and land cover change (LUCC) (Kok, Verburg, & Veldkamp, 2007; Lambin, Geist, & Lepers, 2003; Rindfuss, Walsh, Turner, Fox, & Mishra, 2004; Turner, Lambin, & Reenberg, 2007). LUCC deals both with changes in land cover, i.e. the complete replacement of vegetation that covers the earth surface, and land use that includes modification of land cover due to environmental change and/or human management (Turner, 2003). In practice, most regional LUCC research deals with land cover only, due to the inability to observe land use practices by remote sensing, the main source of land cover data (Verburg, van de Steeg, Veldkamp, & Willemen, 2009). The collection of land use information including human management practices requires intensive field campaigns that are often not feasible for large regions. The focus on land cover disregards the importance of land management changes that have important effects on environment and human wellbeing (Duvernoy, 2000; Ellis & Ramankutty, 2008). Agricultural land use and management is often studied from the perspective of farming systems. A farming system is defined as a population of individual farm systems that have broadly similar resource bases, enterprise patterns, household livelihoods and constraints, and for * Corresponding author. Tel.: þ31 317 485208; fax: þ31 317 419000. E-mail address: peter.verburg@wur.nl (P.H. Verburg). Contents lists available at ScienceDirect Applied Geography journal homepage: www.elsevier.com/locate/apgeog 0143-6228/$ – see front matter Ó 2009 Elsevier Ltd. All rights reserved. doi:10.1016/j.apgeog.2009.05.005 Applied Geography 30 (2010) 239–253