RANGELAND VEGETATION DYNAMICS AND MOISTURE AVAILABILITY IN TUNISIA - AN INVESTIGATION USING SATELLITE AND METEOROLOGICAL DATA

Improved knowledge of the interactions between regional climatic patte rns and vegetation dynamics are necessary for predicting the future im pacts of climate change on vegetation and biogeochemical processes. Th is paper describes how Normalized Difference Vegetation Index (NDVI) i mages generated from Advanced Very High Resolution Radiometer (AVHRR) satellite data were used to investigate the dynamics of rangeland vege tation in Tunisia. The NDVI images provided information about intra-an d inter-annual variations in vegetation over nine growing seasons (198 3-1992). Comparison of the NDVI data with field-collected ecological p arameters for nine individual field sites indicated a strong relations hip between the NDVI and percentage vegetation cover. The relationship between biomass measurements and NDVI was, however, less strong. Rain fall and NDVI data for each field site were compared, and significant relationships were found between the two. These indicated that there w as a delay in the vegetation response to rainfall. In addition, the ND VI data showed that the vegetation at some of the field sites remained active throughout the summer although there was no rainfall during th is period. TuMERT (Tunisian Model to Estimate Rangeland Transpiration) , a simple water-balance model, was developed to estimate the amount o f rainfall available for use by the vegetation during transpiration. T he estimates of actual transpiration derived from TuMERT were found to be more strongly correlated with the AVHRR-NDVI measurements than the rainfall data.