Vol. 4, No. 2, 157-170, 2005

Abstract
The development of a vulnerability index for the evaluation of environmental changes in north-western Namibia is a central research topic within the interdisciplinary research project ACACIA (Arid Climate Adaptation and Cultural Innovation in Africa). The calculation of aboveground phyto­mass (= biomass) and its seasonal productivity by using a mesoscale biosphere model based on satellite data is of specific interest. The research area, Kaokoveld (north-western Namibia), is characterised by a strong hydro-climatic gradient with an annual precipitation range from 380mm/a in the north-eastern part of the research area to 50 mm/a at the border of the Namib Desert. Small-scale vegetation patterns with fractions of savannah, woody savannah, open and closed shrub land and grassland reflect this climatic gradient and the heterogeneous relief. The research area is partly used by local herders of the Himba people as pasture ground for their livestock. This causes problems such as overgrazing and degradation of the vegetation. Together with the impact of climate change the known ecological gradients have aggravated during the past decade.

With the remote sensing based regional biosphere model (RBM Kaokoveld) quantitative information about biomass changes and pasture ecology can be determined. The growth and the reduction of biomass can be observed using the theory of Monteith (1972) and Running et al. (2000). In this modelling approach, biomass production is derived from the combination of incoming solar radiation, normalized difference vegetation index (NDVI), resulting from MODIS data, and a biophysical conversion factor, which describes the ability of plants for net primary production (NPP). The regional biosphere model allows detailed information on an area-wide biomass balance to be extracted using remote sensing. This balance describes the production as well as the consumption of biomass by cattle and game and its natural decomposition. The modelling approach runs on medium temporal and spatial scales with a decadal time step and a spatial resolution of 1 by 1 km². The results of this modelling approach have been checked and evaluated in three different ways. Thus, the model provides reliable data.

The model uses a four-year time series of MODIS data from 2000 to 2003, with biomass changes and degradation areas as results. In the detailed result for the years 2001 and 2002, in wide areas of Kaokoveld, a reduction of biomass production by more than 10 g/m² can be observed. These changes may be explained by the different rainfall patterns between the two observed years 2001 and 2002.

ISSN 1729-3782