Global changes in dryland vegetation dynamics (1988-2008) assessed by satellite remote sensing: Comparing a new passive microwave vegetation density record with reflective greenness data

Abstract

Drylands, covering nearly 30% of the global land surface, are characterized by high climate variability and sensitivity to land management. Here, two satellite-observed vegetation products were used to study the long-term (1988-2008) vegetation changes of global drylands: the widely used reflective-based Normalized Difference Vegetation Index (NDVI) and the recently developed passive-microwave-based Vegetation Optical Depth (VOD). The NDVI is sensitive to the chlorophyll concentrations in the canopy and the canopy cover fraction, while the VOD is sensitive to vegetation water content of both leafy and woody components. Therefore it can be expected that using both products helps to better characterize vegetation dynamics, particularly over regions with mixed herbaceous and woody vegetation. Linear regression analysis was performed between antecedent precipitation and observed NDVI and VOD independently to distinguish the contribution of climatic and non-climatic drivers in vegetation variations. Where possible, the contributions of fire, grazing, agriculture and CO2 level to vegetation trends were assessed. The results suggest that NDVI is more sensitive to fluctuations in herbaceous vegetation, which primarily uses shallow soil water, whereas VOD is more sensitive to woody vegetation, which additionally can exploit deeper water stores. Globally, evidence is found for woody encroachment over drylands. In the arid drylands, woody encroachment appears to be at the expense of herbaceous vegetation and a global driver is interpreted. Trends in semi-arid drylands vary widely between regions, suggesting that local rather than global drivers caused most of the vegetation response. In savannas, besides precipitation, fire regime plays an important role in shaping trends. Our results demonstrate that NDVI and VOD provide complementary information and allow new insights into dryland vegetation dynamics. © Author(s) 2013.