Using satellite based soil moisture to quantify the water driven variability in NDVI: A case study over mainland Australia

Abstract

Soil moisture is crucial in regulating vegetation productivity and controlling terrestrial carbon uptake. This study aims to quantify the impact of soil moisture on vegetation at large spatial and long-term temporal scales using independent satellite observations. We used a newly developed satellite-derived soil moisture product and the Normalized Difference Vegetation Index (NDVI) to investigate the impact of soil moisture on vegetation across mainland Australia between 1991 and 2009. Our approach relied on multiple statistical methods including: (i) windowed cross correlation; (ii) quantile regression; (iii) piecewise linear regression. We found a strong positive relationship between soil moisture and NDVI, with NDVI typically lagging behind soil moisture by one month. The temporal characteristics of this relation show substantial regional variability. Dry regions with low vegetation density are more sensitive to soil moisture for the high end of the distribution of NDVI than moist regions, suggesting that soil moisture enhances vegetation growth in dry regions and in the early stage in wet regions. Using piecewise linear regression, we detected three periods with different soil moisture trends over the 19years. The changes in NDVI trends are significant (p<0.01) with turning points of soil moisture in the beginning of 2000 and the end of 2002. Our findings illustrate the usefulness of the new soil moisture product by demonstrating the impacts of soil moisture on vegetation at various temporal scales. This analysis could be used as a benchmark for coupled vegetation climate models.