The role of population in understanding Honduran land use patterns.

Land use patterns are usually influenced by large variety of factors that act over a broad range of scales. Biophysical, climatic, and socioeconomic factors are important and need to be considered, when distribution of land use is to be understood. The main objective of this study is to test this hypothesis using a statistical analysis at 'supra-local' level. Regression analysis is used to describe land use patterns in Honduras, selected because of its rare combination in Latin America of high population growth and poor biophysical conditions. Furthermore, the aim of the analysis is to specifically highlight two aspects, the effect of spatial and temporal scale and the influence of population density: to determine the influence of spatial and temporal scale, six spatial resolutions at two points in time (1974 and 1993) were included. To determine the role of population density and population growth, this factor was singled out; an analysis of migration patterns was performed; and a measure for technological development was calculated. Multiple regression equations indicate the importance of soil-related, climatic and demographic factors for most of the land uses. Relations appear to be stable in space and time. Rural population density dominates as driver over the whole range of resolutions and for both years, especially for maize where it explains up to 80% of the variation. The strong constant relationship between population and agricultural area could be caused by a lack of technological development. An analysis of yield development confirms that for most annual crops yield increases lag behind area growth. Besides, the strong correlation could be explained by assuming rural population density to be a proxy for a range of other factors, like labour costs, or accessibility that are the direct drivers of land use change. In any case, this study suggests that for a specific--relatively coarse--window of temporal and spatial scale, land use patterns can be described with very simple relationships, with a strong contribution of population density. More local studies are needed to test the hypothesis that rural population density is a proxy for other variables.