PS 43-165 - Bidirectional dependency between vegetation and terrestrial water storage in China

Abstract: Changes in water availability strongly influence vegetation growth, and vegetation can also affect terrestrial water storage (TWS) by altering surface hydrological processes.

The goal of this study is to examine the interactions between vegetation change and terrestrial water resources conditions in the vegetation cover of China.

Our approach to addressing our goal was analysing the trends in TWS and normalized difference vegetation index (NDVI) changes by Mann-Kendall trend test, quantifying the time lag effect between TWS and NDVI each other use Spearman's partial correlation analysis, revealing the interaction between TWS and NDVI by non-linear Granger causality.

Here we show: Most of the vegetation in China underwent significant greening from January 2003 to December 2020, whereas TWS showed a spatial distribution pattern of increasing in the south and decreasing in the north. More than 27% of vegetation-covered areas showed a significant increasing trend in NDVI and TWS simultaneously, and approximately 20% of the vegetation-covered areas showed a significant increasing trend in NDVI and a significant decreasing trend in TWS. Combining the time lag effect and partial correlation analysis, we found that both TWS and NDVI had a significant mutual lag response. After superimposing the time lag effect, the proportion of area significantly correlated between TWS and NDVI increased from 42% to 63%; NDVI in 32% of vegetation-covered areas was significantly positively correlated with TWS, and NDVI in 31% of vegetation-covered areas was significantly negatively correlated with TWS. The combined non-linear Granger causality indicated that NDVI was generally influenced by TWS, and there was a unidirectional causality of TWS on NDVI in 58% of the vegetation-covered areas. There were two different ways for NDVI changes to regulate TWS: vegetation greening drove an increase in TWS in humid regions dominated by forest cover, and vegetation greening drove a decrease in TWS in arid regions dominated by grassland and cropland cover. The influence of vegetation dynamics on surface hydrological processes might be exacerbated with increasing aridity. In China, over 64% of vegetation-covered areas showed unidirectional dependence of TWS on NDVI.

In conclusion, this study provides a preliminary comprehension of the possible non-linear causal relations between vegetation productivity and terrestrial hydrological conditions.