Detection and Mapping of Desertification

The extent and severity of desertification is an important issue that affects economic development, and is significant in the context of global climate change and the global carbon cycle. The need for a measure of desertification that can also be used to map the extent and severity of the problem has long been recognized, but there is a lack of any readily measured, objective indicators applicable at a regional. Amongst the definitions of desertification many make clear the central role of the carbon cycle. Satellite remote sensing has partly supplied the need for data by providing techniques that can detect variables related to net primary production (NPP). We have developed procedures for mapping desertification using satellite data and a carbon cycle model (GLO-PEM) that allows monitoring the actual and potential NPP. We will discuss work in southern Africa where we have shown that negative deviations from the potential (undesertified) NPP provide an effective indicator of desertification. Three approaches to measurement using reduced NPP will be presented. The first is the Rain Use Efficiency, the ratio of NPP to rainfall over a five-year period. Deviations from the conservative value of RUE provides a useful index of degradation, independent of the rainfall. A second, more mechanistic technique using a comprehensive biogeochemical model to estimate potential NPP is used to overcome the limitations of estimating potential NPP with rainfall alone. The index of degradation in this case is the difference between potential and actual NPP. One problem with both RUE and Potential-Actual NPP is a limitation on spatial resolution caused by data; meteorological stations are sparsely distributed and local rainfall can be spatially highly variable in semi-arid regions. To avert this problem, a third technique has been developed, Local NPP Scaling (LNS), in which the NPP of each pixel is expressed as a proportion of the maximum observed in all land falling into the same terrain type. Stratification by terrain type allows climate, soil and land cover differences to be normalized and degradation to be detected relative to the maximum observed NPP. Cultural factors can also be included, such as commercial or communal land tenure. The results suggest that the deviation of NPP was quite variable throughout the region during the period of study. Thus, in parts of the region, NPP seems to be significantly less than the potential productivity, and it is these that we have identified as desertified.