Statistical relationship between drought indices and NDVI at regional scale

Abstract

The potential contribution of easily accessible satellite data to the detection and quantification of regional droughts, in the absence of reliable meteorological data, is the objective of this study. The former statement can be investigated by identifying the characteristics of vegetation response to dry-wet cycles. The relationships between several "palmer-based" drought indices that are derived from meteorological data, and the concurrent time series of a coarse resolution departure of Normalized Difference Vegetation Index were used to investigate the feasibility of using remote sensing data in monitoring and observing drought conditions at spatial resolutions higher than the currently used in drought monitoring methods. Methods of investigation included crosscorrelation between drought indices and both regional averaged and land-cover stratified departure NDVI, as well as spectral analysis of each type of time series. The analysis showed a positive relationship between drought indices used and most vegetation covers and even a stronger relationship for particular vegetation covers such as open shrublands and needleleaf forest. Spectrum analysis detected a similar signal from both series that is related to ENSO events. This signal is observed at low frequency variability in departure NDVI and drought indices.