Quantified soil dynamics and spatial fragmentation within the shifting agricultural landscape in southern Cameroon

摘要

This research used scientific modeling tools to produce quantitative information on the effects of shifting agriculture on soil and spatial pattern of landscape dynamics in the area. An analysis of farming system led to the development of a conceptualmodel of the spatio-temporal dynamics of shifting agriculture, including transition matrices of rotational cycles. The study of soil variability showed that 30-35% of the total variance of some topsoil (0-20 cm) properties was due to the influence landuse practices. A robust quantitative multi-criteria method was then developed that selected five soil properties (pH, calcium, available phosphorus, bulk density and organic carbon) that are the most sensitive to these agricultural practices. Empirical models of linear/quadratic fractional rational functions were successfully fitted to time series soil variables to derive quantitative measures on temporal changes in soil with land use. Multi-spectral satellite imagery was able to map with 80% accuracy the extension front of shifting agricultural landscape and the most dynamic land cover types (crop fields, young fallows), which shift every season and every year. Data and methods produced are useful for soil quality assessment and spatio-temporal dynamic simulation in order to guide decision-making for sustainable land-use planning and forest resources management.