Identifying hotspots and representative monitoring area of groundwater changes with time stability analysis

Groundwater is a most accessible freshwater resource for human beings, and it is increasingly important as an alternative to surface water under the threat of climate change. However, its complex spatio-temporal dynamic remains unattended from management perspective. Past studies on groundwater management were stalled by a relative dearth of high-quality data and a lack of synthetic analysis on both spatial and temporal information. Thanks to NASA's launch of Gravity Recovery and Climate Experiment (GRACE) satellite mission, our study has solved these two problems by innovatively applying time stability analysis to GRACE-based groundwater data. Taking the Yellow River Basin (YRB) as an example, we employed GRACE satellite data to obtain monthly changes of groundwater tables from Jan. 2003 to Dec. 2016 in 1.0-degree grid of spatial resolution. Then we identified hotspots (which indicated severe groundwater declines and fluctuations over time) and representative monitoring areas (which stably represented the spatial average over time) using time stability analysis. Time stability employs multiple coefficients to identify the spatial relations between local variables and global variables overtime, thus showing the overall effect of spatial-wise and temporal-wise factors but never used in groundwater studies before. Based on this innovative method, we further identified management categories across the YRB using multivariate cluster analysis. As a result, the YRB has been divided into five zones for different management strategies. We identified the hotspots in west-most and east-most areas of the YRB, where we suggest a strengthened groundwater protections and risk response system. The northern part of the middle reach in the YRB was also identified as the representative monitoring areas. With these knowledge, decision-makers can have a clearer regional plan for groundwater protection, monitoring, and risk response system. This new method enables a quick decision on the prioritized areas for different groundwater management strategies while not losing the scope of spatio-temporal heterogeneity.