Mapping regions of high temporal variability in Africa

Kurzfassung

Long term observation of space-borne remote sensing data provides a means to explore temporal variation on the Earth’s surface. This improved understanding of variability is required by numerous global change studies to explain annual and interannual trends and to separate those from individual events. This knowledge also can be included into budgeting and modeling for global change studies. The study employs daily 8km NOAA AVHRR data of the Pathfinder program to study changes in the annual variability of the African continent between 1982 and 2000. The daily data were processed to improved 10-day composites using an iterative approach including metadata and robust statistical techniques. Seasonality analysis using harmonics and its explained variance is combined with cross correlation between years with identical seasonality to account for temporal shifts. Deserts and the inner-tropical rain forest with none and moderate variability, respectively, are circled by a zone of transition assigned with high variability and changing seasonality. In between there are two gradual units of stability, usually identified as grassland or woodland by continental classifications. This relatively zonal pattern is altered in the eastern tropics and Namibia by varying topography and oceanic influences, respectively. The results can provide a basis for spatial distributed modeling of dynamic hotspots such as the transition zones bordering the savannas and for linking vegetation dynamics with continental climate models.