Forecasting groundwater nitrate vulnerability dynamics through spatiotemporal lagged bivariate local indicators of...

As a crucial water resource, groundwater faces a growing threat from nitrate pollution. Therefore, it is important to identify potential risk areas through an assessment of groundwater nitrate vulnerability. Nevertheless, most studies on the vulnerability of groundwater nitrate are limited to static evaluations, and dynamic trend prediction is lacking. Therefore, this study proposes a novel framework combining bivariate local indicators of spatial association and temporal dynamics. We define four types of vulnerability dynamics with clear implications for management: Persistent (areas with long-term high vulnerability), Decline (areas with a decline in vulnerability), Stable (areas maintaining long-term low vulnerability), and Emerging (areas with a potential for increasing vulnerability). Taking the shallow groundwater of the North Weifang Plain as the empirical area, multi-year nitrate vulnerability assessment (2008, 2013, 2015, 2016, 2017, 2018) and multi-time span prediction (Δt = 1 year and Δt = 5 years) were carried out. The results showed that the accuracy between the predicted dynamic type and the actual vulnerability level change was 82%-88% (within statistically significant areas), and the change in nitrate concentration was substantially consistent with the prediction of the dynamic trend of vulnerability. In addition to short-term predictions (1 year), the method is reliable for medium-term trend analysis (5 years). It can provide scientific tools for managers to proactively formulate differentiated strategies (such as continuous focus in the Persistent area and setting priority on the prevention and control in the Emerging area) to facilitate the transition of groundwater governance from reactive to proactive management.