Double disaggregation of the decline of terrestrial water storage for a highly cultivated dryland partially covered by glaciers

Dramatic declines of terrestrial water storage (TWS) have been found in global drylands, the home to more than a third of the world’s population. TWS variations can be internally disaggregated into changes in hydrological components or externally disaggregated into impacts of climate change and human activities. This study proposed an innovative double disaggregation framework to improve the explanations of TWS depletion in a highly cultivated dryland partially covered by glaciers, i.e., the northern slope of the Tianshan Mountains (NSTM). A widespread and significant decline of TWS was detected in NSTM. Besides the Tianshan Mountains, TWS also declined significantly downstream where it should have increased given the substantive glacier meltwater supply from upstream, implying that the evolution of TWS in NSTM has probably deviated from natural manners. Pixel-wise internal disaggregation indicated that groundwater storage was the predominant hydrological component leading to TWS depletion in most of the NSTM except for glacier-covered areas. Additionally, basin-averaged external disaggregation revealed a more dramatic TWS depletion rate induced solely by human activities compared to GRACE observations, suggesting that human activities have dominated TWS decline. To be specific, substantial withdrawal of groundwater for irrigation enhanced regional evapotranspiration, which subsequently accelerated the dissipation of TWS, and therefore counteracted and even reversed the potential increase in TWS downstream. The double disaggregation framework facilitated the holistic explanations of TWS decline in NSTM, and was expected to serve as a useful tool for attributing TWS variations in other drylands worldwide. • The proposed framework facilitated the holistic explanations of TWS decline. • GWS was the foremost component leading to TWS depletion in most of the NSTM. • TWS depletion was predominantly attributed to human activities in NSTM. • GlacierWS must be considered when estimating GWS in glacier-covered areas.