Objective Groundwater over−exploited areas in North China, represented by Hengshui City, encounter notable technical bottlenecks because the core technical parameters for deep groundwater recharge lack systematic engineering verification. Taking the Changcun Test Site in Hengshui City as the research area, this study systematically identifies the single−well recharge rate threshold for deep groundwater, aiming to provide a scientific basis for parameter optimization, operational stability control, and ecological restoration of regional groundwater recharge projects.

Methods Two 320−m−deep recharge wells (HⅢ4 and HⅢ5) with an interval of 10 m, along with 15 monitoring wells and three layered land subsidence monitoring benchmarks, were deployed at the test site. Pulsed controlled single−well recharge experiments were conducted, and real−time dynamic data of recharge rate and water table depth were continuously monitored and collected. SPSS software was adopted to analyze data characteristics and correlation relationships. Combined with scatter plots and linear regression models, the three−point rolling correlation coefficient method was used to identify abrupt positive−negative transition points of correlation coefficients, so as to quantify the effective threshold range of single−well recharge rate.

Results The results show that the effective single−well recharge rate thresholds are 14.65~32.77 m³/h for Well HⅢ4 and 14.15~21.10 m³/h for Well HⅢ5. Within this threshold range, recharge water infiltrates uniformly, and the recharge rate has a significant negative correlation with water table depth (r < −0.8 for HⅢ4, r < −0.6 for HⅢ5). Exceeding the threshold will cause abnormal infiltration and reverse the correlation to strongly positive (r > 0.8). Significant spatial heterogeneity is found in recharge responses: Well HⅢ4 exerts a prominent influence on surrounding monitoring wells, with an absolute correlation coefficient higher than 0.8. while Well HⅢ5 only affects the immediate near−well area, with an absolute correlation coefficient of 0.37. Furthermore, the three−point rolling correlation coefficient method is verified to accurately capture the critical transition threshold of groundwater recharge states.

Conclusions This study clarifies the effective recharge rate threshold ranges of the two deep groundwater recharge wells and reveals the inherent response mechanism: a negative correlation corresponds to stable and effective recharge, whereas a positive correlation indicates abnormal recharge conditions. The reliability and applicability of the three−point rolling correlation coefficient method are fully validated via field tests. This study fills the gap in field engineering verification of core technical parameters for deep groundwater recharge in over−exploited areas of North China and provides solid technical support for the optimal design and safe operation of regional groundwater recharge projects.