Assessing vegetation response to climate variability via time series of NDVI, precipitation and soil moisture content

The forcing role of climate variability on vegetation was examined over East and Central Africa, using time series of satellite images and climate data. In search of a meteorological variable that best explains the temporal evolution of vegetation over multiple years, lagged correlations were tested between time series (1998-2011) of ten daily SPOT-VGT NDVI images and series of precipitation and soil moisture content from the ECMWF ERA-Interim re-analysis. First, it was found that NDVI displays an area-wide near-instantaneous response to soil moisture content as opposed to a regionally differing lagged response to precipitation. Second, interannual anomalies of NDVI correlate increasingly better with anomalies of soil moisture content in deeper soil layers, in a more spatially coherent pattern. This may be ascribed to the function of the soil as a buffering reservoir, reflected in relatively smooth series of soil moisture content with a gradual decline after the seasonal peak. Third, upon removal of seasonality from the series, the areas with the best NDVI-soil moisture response shift towards the semi-arid areas. Further investigation of dominant soil and vegetation patterns is needed to explain the observed spatial patterns. Furthermore, implementation of more detailed soil information is deemed necessary. Overall, the potential of soil moisture content as a climate descriptor for explaining vegetation dynamics was demonstrated in this study.International EARSeL Workshop on Temporal Analysis of Satellite Images: 22-27. Organised on 24 - 25 May 2012, Mykonos Island, Greece, in conjunction with the 32nd EARSeL Symposium "Advances in Geosciences", 21-24 May 2012. Available online at http://www.earsel.org/SIG/timeseries/pdf/proceedings/proceedings_003.pdf .status: publishe