Groundwater (GW) depletion is a significant global problem, particularly in semi-arid regions like North China. Overexploitation, driven by population growth, economic development, and climate variability, threatens water and food security. The North China Plain (NCP) shows some of the highest GW depletion rates globally, impacting urban water supply, agriculture, and ecosystems. Beijing, in particular, faces severe water scarcity, with GW levels declining significantly. One strategy to mitigate this is large-scale water diversion projects, such as the South-to-North Water Diversion (SNWD), which began in 2002. This study investigates the impact of the central SNWD route on GW storage recovery in Beijing, considering climate change and other water management policies. Previous studies have evaluated the impact of SNWD, but often omitted the effects of climate extremes and water policies on GW storage recovery. This research uniquely quantifies the contributions of water diversion, altered irrigation practices, and climate variability to GW recovery in Beijing, providing a framework for managing water resources in urban areas under environmental stress.

Numerous studies highlight global groundwater depletion (GWD) in various regions, including Northwest India, California’s Central Valley, the U.S. High Plains, and the Middle East. The NCP is a particularly affected area, with GWD rates among the highest globally. Many studies have documented the severe water scarcity faced by large cities worldwide due to GW over-extraction. Existing research on the SNWD’s impact on Beijing’s GW has primarily used GW flow models or coupled surface water/GW models. However, these models often lacked comprehensive consideration of climate variability, water policies (like reduced irrigation), and complex interactions within the aquifer system. This study addresses this gap by integrating these factors into the analysis.

The study employs a multi-faceted approach to quantify the impact of the SNWD on Beijing's groundwater storage (GWS). It combines analysis of observed data, statistical methods, and hydrological modeling. The key data sources include in situ measurements of GW depth and specific yield, monthly water use data from various sectors (domestic, agricultural, industrial, and environmental), and precipitation data. A high-resolution Community Water Model (CWM) coupled with MODFLOW is used to simulate GWS changes under various scenarios. The model calibration was conducted using observed monthly GWS changes from 2006–2010, and validation was performed using data from 2010–2014. Three scenarios were simulated to isolate the effects of different factors. Scenario I removed the SNWD's impact, Scenario II also removed the impact of reduced agricultural water use, and Scenario III further removed the influence of precipitation variability. The STL (Seasonal and Trend decomposition using LOESS) method was used to decompose the time series of GWS anomalies to separate seasonal variability from long-term trends. Groundwater recharge was adjusted in the model to account for the dominance of preferential flow in the aquifer. The model was calibrated and validated using statistical metrics, aiming to balance simulated trends and seasonal variability against observed data.

The analysis reveals that water diverted to Beijing through the SNWD significantly reduced cumulative GWD by approximately 3.6 km³ (40%) from 2006 to 2018. Increased precipitation and policies that led to reduced irrigation also contributed significantly to GWS recovery (-2.7 km³ and -2.8 km³, respectively, each representing ~30%). The study uses a hydrological model to quantify the relative contributions of each factor. Scenario analysis shows that without the SNWD, groundwater depletion would have been significantly greater. GW withdrawal from self-contained wells decreased by 30% (2006-2016), largely due to reduced agricultural use. GW withdrawal from water-source wells declined by 22% due to the increased availability of diverted surface water. The use of reclaimed wastewater also increased, mainly for environmental purposes. The model projections suggest that GWS in Beijing is likely to continue recovering over the next decade, provided that sufficient water is diverted, even under various precipitation scenarios. The study also explored various scenarios of precipitation and groundwater withdrawal for 2019–2030, predicting potential GWS changes based on different policies.

The findings demonstrate the significant role of the SNWD in mitigating GWD and promoting GWS recovery in Beijing. The integration of water diversion with policies aimed at reducing irrigation significantly enhanced the effectiveness of the project. The study highlights the interconnectedness of climate variability, water management policies, and large-scale engineering projects in influencing GW resources. The model simulations, encompassing different scenarios, strengthen the robustness of the findings, indicating that even under various climate conditions, the positive impact of the SNWD is likely to persist. The results underscore the importance of integrated water resource management strategies that combine large-scale infrastructure projects with effective policies to achieve sustainable groundwater use.

The SNWD has played a crucial role in stabilizing and recovering Beijing's groundwater levels, significantly reducing groundwater depletion. The combined effects of water diversion, reduced irrigation, and increased precipitation have led to substantial GWS improvements. Future research should focus on more detailed assessments of the SNWD's long-term effects, including potential ecological consequences, and the development of further refined models that can better account for complex aquifer behavior. Continuous monitoring and adaptive management strategies are vital for ensuring the sustainable use of groundwater in Beijing and other water-stressed regions.

The study's model simplifies the complex groundwater system in Beijing, using a single-layer confined aquifer representation. This simplification might underestimate the complexity of GW flow and interactions within the aquifer. The model also relies on available data, and uncertainties in data quality or spatial coverage might influence the results. Future studies could benefit from incorporating more detailed aquifer characterizations and higher-resolution data to improve model accuracy.