Skip to main content
NCBI home page
Scientific Reports logoLink to Scientific Reports
. 2026 Feb 6;16:7583. doi: 10.1038/s41598-026-38451-5

Ecological water replenishment effects on groundwater recovery in the largest shallow groundwater depression cone of the North China Plain

Wen Lu 1,2, Chuiyu Lu 1,2,✉, Yangwen Jia 1,2, Yong Zhao 1,2, Xin He 1,2, Qingyan Sun 1,2, Chu Wu 1,2, Wenjia Zhang 1,2, Zhenjiang Wu 3
PMCID: PMC12936216  PMID: 41651941

Abstract

The North China Plain is among the world’s most severely groundwater-overexploited regions. Since 2018, large-scale ecological water replenishment (EWR) has been applied, yet its long-term impacts on groundwater dynamics and efficient EWR strategies remain poorly understood, particularly in the Shijiazhuang Plain, which hosts the largest shallow groundwater depression cone in the region. In this study, the coupled surface–groundwater model MODCYCLE was applied to investigate these issues in the Shijiazhuang Plain. Results reveal that large-scale EWR has effectively reversed groundwater depletion, recovering the average groundwater table by 1.96 m and storage by 1.76 billion m3 by 2022. Contributions were dominated by EWR from the Hutuo River (82.5%), followed by the Sha-Zhulong River (12%) and other Rivers (5.5%), with significant recovery (> 0.5 m, > 50,000 m3/km2) extending approximately 24.5 km, 12.5 km, and 5.5 km from these rivers. The highest-recovery zones were mainly in the middle and lower reaches of the EWR rivers. Our findings highlight that, due to its favorable infiltration path length and groundwater flow conditions, the Hutuo River should serve as the primary EWR channel, with stable, moderate flows rather than the “greater is better” approach, to achieve both efficient groundwater storage recovery and widespread groundwater table recovery. Optimal flows of 14.64 m3/s during droughts and ≥ 8.4 m3/s under normal precipitation efficiently mitigate and reverse groundwater depletion, respectively. This study provides scientific support for the continued fight against groundwater overexploitation in the North China Plain and offers insights for optimizing EWR strategies to efficiently recover groundwater in globally overexploited aquifers.

Keywords: Ecological water replenishment, Groundwater recovery, MODCYCLE, North China Plain, Shijiazhuang Plain

Subject terms: Environmental sciences, Hydrology

Introduction

Approximately 30% of regional aquifers worldwide have experienced accelerated declines in groundwater heads over the past four decades, with rapid declines (> 0.5 m/year) becoming increasingly widespread in the twenty-first century, especially in arid regions and areas with extensive agricultural irrigation1. These declines have resulted in a range of hydrogeological and ecological consequences, including seawater intrusion2, land subsidence3, streamflow depletion4, and the drying of wells5. In response, the application of Managed Aquifer Recharge (MAR) has been expanding globally at an average annual rate of approximately 5% as a strategy to mitigate or even reverse groundwater depletion1,6. Common MAR techniques include infiltration ponds, excess irrigation, recharge wells, river infiltration, and river infiltration is the most widely used, accounting for 52% of global MAR applications7,8.

The North China Plain is one of the most severely groundwater-overexploited regions in China and the world9,10. Between the 1980 s and 2014, shallow groundwater heads experienced a dramatic decline of approximately 20 to 60 m11. To mitigate groundwater depletion, efforts to apply ecological water replenishment (EWR) to rivers in the North China Plain began in the early twenty-first century, providing artificial groundwater recharge through infiltration. For example, between 2007 and 2018, 2.78 × 108 m3 of reclaimed water were used to replenish the Chaobai River in the Beijing section, with groundwater heads along the river increasing by approximately 3 m over this period12. Since September 2018, large-scale and systematic EWR has been continuously implemented through coordinated operations of mountain reservoirs, the South-to-North Water Diversion Project, and other initiatives. This involves EWR for 48 rivers and lakes, including the Hutuo River, Yongding River, Chaobai River, and Baiyangdian Lake, helping to recover groundwater and alleviate associated ecological and environmental issues13,14. By the end of 2022, the total EWR volume had reached 24 billion cubic meters, with the estimated infiltration recharge volume based on groundwater table dynamics reaching 10.22 billion cubic meters15. Despite the significant recharge provided to the groundwater system, studies on the groundwater table and groundwater storage recovery effects of EWR in the North China Plain are still understudied16,17, with several issues remaining. First, most studies focus on single EWR events, offering little insight into long-term, regional-scale impacts17–19. Second, analyses based solely on groundwater table observations cannot disentangle the effects of EWR from other drivers, such as pumping and precipitation18,20,21. Third, research has been concentrated on the EWR activities of the Chaobai River and the Yongding River in the Beijing Plain of the North China Plain12,16,22–25, largely due to Beijing’s status as the capital, while other key EWR regions remain understudied. Shijiazhuang City is the capital of Hebei Province, China, and the Shijiazhuang Plain is the largest shallow groundwater depression cone in the North China Plain. By the end of 2022, the accumulated EWR volume in the Shijiazhuang Plain accounted for about 21% of the total EWR in the North China Plain, but has received little attention. Finally, despite the scarcity of surface water in the North China Plain, insights on optimizing EWR strategies for efficient groundwater recovery are still emerging. These issues constrain further progress in groundwater recovery in the North China Plain.

Numerical models represent one of the fundamental tools for investigating hydrological processes. Representative examples include the SWAT model, which emphasizes surface hydrological processes26,27; the MODFLOW and FEFLOW models, which simulate groundwater dynamics28,29; and coupled models such as SWAT-MODFLOW, MIKE SHE, and MODCYCLE, which integrate surface and groundwater processes into a unified framework30–32. The essence of EWR in supporting groundwater recovery lies in the integrated management of the surface–groundwater hydrological continuum. Consequently, stand-alone surface or groundwater models remain inadequate for accurately characterizing the system27. In contrast, coupled surface–groundwater models are widely applied to analyze the spatial patterns and dynamic impacts of surface–groundwater interactions at regional scales, offering a physically based framework for investigating groundwater dynamics under the influence of EWR17,23. Nevertheless, such coupled models demand more extensive data support and involve a more complex and time-consuming calibration process23.

MODCYCLE is a tightly coupled model developed by the China Institute of Water Resources and Hydropower Research32, implemented in C + +, and has been successfully applied in various regions of China, including the North China Plain and the Northeast China Plain33–35. In this study, it was applied to quantify the long-term and regional-scale impacts of EWR on groundwater table and storage in the Shijiazhuang Plain. The results provide scientific support for developing optimized EWR strategies to promote efficient groundwater recovery in the North China Plain and offer broader implications for groundwater management in water-stressed regions worldwide.

Materials and methods

Study area

As shown in Fig. 1a, the Shijiazhuang City is located in the central and southern parts of Hebei Province, with a total area of 14,088.35 km2. The terrain slopes from northwest to southeast, with the western part belonging to the Taihang Mountain area (7,507.44 km2), and the eastern part to the Hutuo River alluvial plain (6,580.91 km2). Major rivers in Shijiazhuang include the Hutuo River, Sha-Zhulong River, Ci-Mudao River, Beisha-Huai River, Xiao River, and Wu River, with the main canals being the South-to-North Water Diversion Middle Route Project (SNWD-MR) and its associated canals. Major reservoirs include four large ones: Gangnan (GN), Huangbizhuang (HBZ), Koutou (KT), and Hengshanling (HSL), as well as eight medium-sized reservoirs: Honglingjin (HLJ), Yanchuan (YC), Xiaguan (XG), Baicaoping (BCP), Zhanghewan (ZHW), Nanpingwang (NPW), Bayi (BY), and Chenjiayuan (CJY). The city lies in a temperate semi-humid, semi-arid continental monsoon climate zone, characterized by low and unevenly distributed precipitation. From 2004 to 2022, the average annual precipitation was 538.2 mm, with approximately 75% occurring between June and September. From 2008 to 2022, the average annual water supply and use in Shijiazhuang City was 3.11 billion m3, with surface water accounting for about 33.4% and groundwater for about 66.6%.

Fig. 1.

Fig. 1

Open in a new tab

(a) Terrain and surface water network of Shijiazhuang city, and (b) hydrogeological cross-section of the Shijiazhuang Plain.

Shijiazhuang City is located in China’s main grain-producing region, with an annual grain production consistently exceeding 4.3 million tons36. Over 70% of the land in the Shijiazhuang Plain is used for agriculture, and over 80% of the groundwater extraction is used for agricultural irrigation. Agricultural irrigation is the primary cause of the long-term overexploitation of groundwater in the Shijiazhuang Plain, further leading to ecological issues such as river drying up and lake shrinkage37,38. Despite efforts in recent years to promote drought-resistant crops and water-saving agricultural irrigation39,40, groundwater extraction in the Shijiazhuang Plain remains high in order to ensure food security (2022 amounted to 1.08 billion m3). To mitigate groundwater depletion and restore the ecological environment, large-scale EWR has been continuously implemented in the Shijiazhuang Plain since September 201841, with a cumulative volume exceeding 5 billion m3 by the end of 2022. The EWR rivers in this region include the Hutuo River, Sha–Zhulong River, Beisha–Huai River, Xiao River, and Wu River (Fig. 1a).

As shown in Fig. 1b, the thickness of the Quaternary strata in the Shijiazhuang plain increases from several dozen meters in the mountain front areas to about 500 m in the southeastern part. Groundwater is mainly distributed in sand-gravel and sand aquifers. From west to east, the sedimentary particle size gradually decreases, the number of aquifers increases, but the thickness of each aquifer decreases, and the thickness of the aquitards gradually increases. Based on the lithology and hydraulic characteristics of the sediments, as well as the current status of groundwater development and utilization, the Quaternary strata in the Shijiazhuang plain are typically divided into four aquifer groups: I, II, III, and IV36,42. The lower boundary elevations of aquifer groups I, II, and III range from −20 to 50 m, −200 to −30 m, and −380 to −120 m, respectively. The upper two aquifer groups are unconfined, with close hydraulic connections between them, and many mixed pumping wells penetrate these two groups. Except for areas near Xinji County, there is no continuous, thick aquitard between the aquifer group III and the upper two groups, making it a semi-confined system, with confined conditions near Xinji County. Aquifer group IV is confined. Generally, the upper two aquifer groups are classified as the shallow aquifer system, whereas the lower two groups are classified as the deep aquifer system. The groundwater depth increases from over 10 m near the mountain front to approximately 50 m in the southeast, with groundwater generally flowing from the northwest to the southeast.

MODCYCLE model

As shown in Fig. 2, in terms of horizontal structure, MODCYCLE first discretizes the simulation domain into subbasins and hydrological response units (HRUs) based on underlying surface information, with each subbasin having exactly one main river channel. The subbasins located in mountainous and plain areas are designated as mountainous subbasins and plain subbasins, respectively. The entire region of plain subbasins constitutes the groundwater numerical simulation domain. The finite-difference grid, consisting of multiple layers, rows, and columns of cells, is used to spatially discretize the groundwater numerical simulation domain, establishing a spatial mapping relationship between grid cells and plain subbasins, HRUs, main river channels, and canals (Fig. 2). The main river channel segments and canal segments located within the cells are referred to as river and canal reaches (Fig. 2). The model can also simulate water storage nodes such as reservoirs.

Fig. 2.

Fig. 2

Open in a new tab

Surface–groundwater coupling simulation method of the MODCYCLE model.

MODCYCLE is a daily-scale simulation model. The Green-Ampt equation is used to simulate runoff generation and precipitation infiltration, the Muskingum method is applied for river flow simulation, and the Penman–Monteith equation is used for evapotranspiration simulation32. A vegetation growth model, coupled with HRUs, tracks daily soil-moisture dynamics during the crop growing season. When soil moisture drops below a predefined threshold, an irrigation event is triggered, and a specified amount of water is withdrawn from the designated source34. Multiple events may occur within a growing season, enabling the simulation of agricultural irrigation usage. The groundwater system of each mountainous subbasin is simplified as an independent linear reservoir, with no groundwater exchange occurring between mountainous subbasins. In the groundwater numerical simulation domain, the boundaries of the plain subbasins do not block lateral groundwater flow (Fig. 2), and groundwater flow follows the finite-difference flow equation, which is also the governing equation used in MODFLOW-200528. The groundwater systems of the mountainous and plain areas are treated as separate systems, with no water exchange between them. Lateral recharge from the mountain side to the plain side is represented by the infiltration of runoff from the mountain front.

The surface–groundwater coupling occurs at the subbasin scale in mountainous areas and at the grid cell scale in the plain areas. Meteorological data from the weather station closest to the centroid of each subbasin in the mountainous area and of each grid cell in the plain area are used to drive the simulations. In the mountainous area, daily exchange fluxes between the main river channel and groundwater are explicitly calculated based on the groundwater head from the previous day. In the plain area, daily exchange fluxes between the river reaches, canal reaches, and groundwater are treated as unknown sources and sinks in the finite-difference flow equation and are implicitly solved through iteration until convergence of river stage, canal stage, and groundwater heads is achieved.

MODCYCLE setup

Simulation period and data

The simulation period spans from 2018 to 2022. Starting in September 2018, large-scale EWR implementation began in the Shijiazhuang Plain. Therefore, the period from January to August 2018 is designated as the model warm-up period, used to initialize the state variables of MODCYCLE. The period from September 2018 to December 2022 was used for model calibration and validation. The main modeling and calibration data are shown in Table 1, including geographic data, hydrogeological data, water supply and usage data, crop structure and irrigation quotas, runoff data, and groundwater head data, etc.

Table 1.

Model data.

Type Sources Resolution
Digital elevation model (DEM) Resource and Environmental Science Data Platform (https://www.resdc.cn/) 90 m × 90 m
Landuse Resource and Environmental Science Data Platform (https://www.resdc.cn/) 1 km × 1 km
Soil Harmonized World Soil Database (https://www.fao.org/soils-portal/en/) 1 km × 1 km
River network and canal network Hai River Basin Statistical Yearbook 1:500,000
Meteorological data from 24 stations National Meteorological Center (https://data.cma.cn/) Day
Crop structure, crop growing period, and irrigation quotas Hebei Province Rural Statistical Yearbook Year
Reservoir storage and discharge parameters Literature Review43–46 /
Observed runoff from 3 stations Hai River Basin Statistical Yearbook Month
Hydrogeological Data, including hydrogeological zoning, groundwater heads, specific yield, etc Hebei Provincial Hydrogeological and Engineering Geology Survey Institute 1:500,000, month
Water supply and usage, including EWR, agricultural usage, industrial usage, domestic usage, etc Hebei Provincial Department of Water Resources and Shijiazhuang Water Resources Bulletin Year/month
Open in a new tab

The EWR water sources and volumes for each river during the simulation period are shown in Table 2. For the period from 2018 to 2020, only annual-scale EWR data were collected; these were refined to a monthly scale through literature investigation and reference to the runoff data of upstream and downstream runoff stations (Fig. 3). Monthly-scale EWR data were collected for the years 2021–2022. The total EWR volume from 2018 to 2022 reached 5.06 billion m3, with the Hutuo River, Sha-Zhulong River, and other Rivers (Beisha-Huai River, Wu River, Xiao River) accounting for 74.59%, 22.64%, and 2.77%, respectively (Fig. 1a).

Table 2.

EWR volumes in the Shijiazhuang Plain from 2018 to 2022.

Rivel channel Water source EWR volume (108 m3)
2018 2019 2020 2021 2022
Hutuo river GN and HBZ reservoirs 0.34 1.01 0.72 11.68 8.63
SNWD-MR 3.86 3.62 3.13 3.23 1.55
Sha-Zhulong river SNWD-MR 0 0.40 2.28 4.84 3.94
Beisha-Huai river BCP reservoir 0 0 0 0 0.13
SNWD-MR 0 0 0.50 0 0.28
Wu river SNWD-MR 0 0 0.02 0 0.39
Xiao river SNWD-MR 0 0 0 0 0.08
Total volume (108 m3) 4.20 5.03 6.65 19.75 15.00
Open in a new tab
Fig. 3.

Fig. 3

Open in a new tab

Model discretization, meteorological stations, runoff stations, and groundwater head observation wells.

Model discretization

Based on the DEM and actual river network, Shijiazhuang City is divided into 516 mountainous subbasins and 583 plain subbasins (Fig. 3), considering the storage and release operations of 12 large- and medium-sized reservoirs, as well as the water diversion operations of SNWD-MR (Fig. 1a). For groundwater numerical simulation, the plain subbasins are further subdivided into 1 km × 1 km grid cells, with 125 rows and 111 columns, resulting in a total of 6,913 single-layer variable-head grid cells (Figs. 3). The upper aquifer groups I and II of Shijiazhuang Plain are generalized into two-layer grid cells for simulation. Since the current exploitation of the aquifer group IV is limited, it is combined with the aquifer group III as a single-layer grid cell.

Boundary conditions for groundwater numerical simulation

The northern, eastern, and southern boundaries of the Shijiazhuang Plain are specified as time-variant specified-head boundaries (Fig. 3). The time-variant specified-head for each grid cell is obtained by interpolating observed groundwater head data from observation wells. The western boundary of the plain, where it meets the mountainous area, is conceptually a lateral recharge boundary condition. However, since the MODCYCLE model performs surface–groundwater coupling simulations, the infiltration of runoff from the mountain front already accounts for the lateral recharge effect, so there is no need to specifically set boundary conditions. Other boundary conditions, such as soil water deep percolation and river leakage, are calculated by MODCYCLE model and do not require specification.

Model calibration, validation, and parameterization

MODCYCLE is calibrated in three aspects: runoff, agricultural irrigation usage, and groundwater head (Fig. 3). The following criteria were used to select the runoff stations: First, the station should ideally be located upstream of a reservoir; otherwise, the observed runoff would mainly reflect the downstream discharge from the reservoir rather than the natural flow in the river. Second, the runoff of the major EWR rivers in the Shijiazhuang Plain should be calibrated. Third, stations with relatively large catchment areas should be selected. Based on these criteria, the Xiajue (XJ) and Pingshan (PS) stations in the mountainous area, and the Beizhongshan (BZS) station in the plain area, were chosen (Fig. 3). The XJ and PS stations are located upstream of the Hutuo River, with catchment areas of 579.4 km2 and 1,622.8 km2, respectively, accounting for 7.7% and 21.6% of the total mountainous area. The BZS station is located downstream of the Hutuo River, with a catchment area of 737.87 km2 in the plain area, accounting for 11.2% of the total plain area. Since runoff data for 2022 at the XJ and PS stations were not collected, the comparison period for the simulated and observed runoff at these two stations is from 2018 to 2021. The period from September 2018 to December 2020 serves as the model runoff calibration period at the three stations. The period from January 2021 to December 2022 is used as the validation period for the BZS station, while the period from January 2021 to December 2021 is used for the XJ and PS stations. For agricultural irrigation usage, model calibration was performed by comparing simulated values with statistical data from Shijiazhuang City. Since some months in 2018 were within the warm-up period and the agricultural irrigation statistics are provided on an annual scale, the comparison period for the simulated and statistical values was set from 2019 to 2022. For groundwater heads in Shijiazhuang Plain, the simulated groundwater heads were calibrated using monthly observed groundwater head data from 79 unconfined aquifer wells and 11 confined aquifer wells in Xinji County (Fig. 3). Due to the availability of higher temporal resolution EWR data during 2021–2022, this period is used for calibrating the groundwater head, while the period from September 2018 to December 2020 is used for validation. The Nash–Sutcliffe Efficiency coefficient (NSE) and the correlation coefficient (R) were used to evaluate the model performance23.

Based on the real-world hydrogeological conditions in the Shijiazhuang Plain (Fig. 1b), the hydrogeological parameters were calibrated for different zones. Since aquifer groups I and II are hydraulically connected and have similar hydraulic properties, the simulation parameters for the first and second layers of grid cells are the same. Additionally, based on observed data, the groundwater head in the semi-confined area of aquifer group III (outside of the Xinji County) is similar to that of aquifer groups I and II. Therefore, except for the Xinji County, the storage coefficient for the third layer of grid cells is set to a very small value, while the vertical leakance between the second and third layer is set to a very high value, ensuring that the simulated groundwater head in the semi-confined area of the third layer remains similar to that in the first and second layers. The calibrated hydrogeological parameters are presented in Fig. 4 and are generally consistent with previous findings47, which reported that the specific yield in the Shijiazhuang Plain ranges from 0.05 to 0.25, and that the hydraulic conductivity of the shallow aquifers (layers 1 and 2) and deep aquifers (layer 3) ranges from 10–50 m/d and 1–100 m/d, respectively. The calibration results of other key parameters in the MODCYCLE model are presented in Table 3 and are also supported by previous studies48,49, which found that SOL_AWC in the Hutuo River basin of the Shijiazhuang area ranges from 0.14 to 0.20, and SOL_K ranges from 3.4 to 26 mm/hr.

Fig. 4.

Fig. 4

Open in a new tab

Hydrogeological parameters of the Shijiazhuang Plain.

Table 3.

Key parameter calibration results for MODCYCLE.

Category Parameter Description Calibration values
Surface water network CH_K Hydraulic conductivity in main river channel alluvium in the mountainous area (mm/hr) 1.50
RIVER RECAH_K Leakance between the river reach and the phreatic aquifer (mm/hr/m) 0.2–12
CANAL REACH_K Leakance between the canal reach and the phreatic aquifer (mm/hr/m) 0.04–0.365
SURLAG Surface runoff lag coefficient 4
Soil ESCO Soil evaporation compensation factor 0.9
SOL_AWC Available water capacity of the soil layer 0.15–0.18
SOL_K Saturated hydraulic conductivity (mm/hr) 1–18
Groundwater in mountainous area ALPHA_BF Baseflow recession constant (day) 0.25
GWDMN Threshold depth of shallow groundwater that triggers baseflow (m) 10
GW_SPYLD Specific yield of the shallow aquifer 0.08
Open in a new tab

Evaluation framework for EWR recovery effect

The MODCYCLE model is first calibrated to establish the benchmark model, and then several scenarios are evaluated: Scenario 1, where all EWR is canceled; Scenario 2, where the EWR of the Hutuo River is canceled; and Scenario 3, where the EWR of both the Hutuo River and Sha-Zhulong River is canceled. The simulated results from each model are compared to assess the recovery effects of total and individual river EWR on groundwater table and storage in the Shijiazhuang Plain. For example, comparing the benchmark model with Scenario 2 reveals the effect of the Hutuo River EWR, and comparing Scenario 2 with Scenario 3 reveals the effect of the Sha-Zhulong River EWR.

By comparing the river leakage across different models, the EWR-induced leakage from each river can be determined. Assuming that at time t, the cumulative leakage induced by a river’s EWR is Lt, and the cumulative EWR for that river at the same time is EWRt, the overall leakage rate of the river’s EWR at this time (EWRLR, t, %) is defined as:

graphic file with name d33e1014.gif 1

Assuming that at this time, the cumulative groundwater storage recovery induced by the river’s EWR is GSRt, the overall groundwater storage recovery efficiency of the river’s EWR at this time (EWRGSRE, t, %) is defined as:

graphic file with name d33e1035.gif 2

Results

Benchmark model performance

Runoff

As shown in Fig. 5a, the monthly simulated runoff at XJ and PS stations fits well with the observed runoff, with NSE and R values above 0.85 and 0.94, respectively, during both the calibration and validation periods. For the BZS station, the NSE and R during the calibration period were −0.83 and 0.53, respectively, while during the validation period, the NSE and R were 0.74 and 0.94. The simulated runoff at BZS generally follows the same trend as the observed runoff. The poor fit during the calibration period is mainly due to the fact that the collected EWR data for 2018–2020 were annual-scale data, lacking accurate monthly-scale dynamic information. In contrast, the EWR data for 2021–2022 were on a monthly scale, leading to a better fit. From September 2018 to December 2022, the total simulated and observed runoff at BZS station were 2.08 billion m3 and 2.16 billion m3, respectively, which are quite close. Therefore, the benchmark model effectively captures the runoff process and EWR behavior in the Shijiazhuang city.

Fig. 5.

Fig. 5

Open in a new tab

Benchmark model performance, (a) runoff, (b) agricultural irrigation usage, (c) groundwater heads at 90 wells, (d) dynamic groundwater heads of 7 unconfined aquifer wells along the EWR rivers and 2 confined aquifer wells in Xinji County.

Agricultural irrigation usage

Agricultural irrigation accounts for approximately half of the total water usage in Shijiazhuang City and is one of the major fluxes in the regional water cycle, making it essential for calibration. Figure 5b shows the comparison between simulated and statistical agricultural irrigation usage from 2019 to 2022. It can be observed that both datasets generally exhibit a decreasing trend. In 2019, which had below-average precipitation (427.58 mm), the simulated agricultural irrigation usage was 169 million m3 higher than the statistical data. In 2020 (576.14mm), 2022 (582.22mm), and 2021 (891.26mm), which experienced normal and above-normal precipitation, the simulated agricultural irrigation usage was lower by 60 to 118 million m3 compared to the statistical data. From 2019 to 2022, the average statistical agricultural irrigation usage was 1.48 billion m3 per year, while the simulated value was 1.47 billion m3 per year, which are close. Therefore, the benchmark model effectively captures the agricultural irrigation behavior in the Shijiazhuang City.

Plain groundwater heads

As shown in Fig. 5c, from September 2018 to December 2022, the monthly simulated plain groundwater heads for 79 unconfined aquifer wells and 11 confined aquifer wells (Fig. 3) achieved an R of 0.99 with the observed groundwater heads. The simulated groundwater heads were close to the observed ones, and the overall fit was good.

The groundwater head dynamics for 7 unconfined aquifer wells (PAC, PT, GCCG, XHL, LC, DZ, and XDZ wells) along the EWR rivers and 2 confined aquifer wells (HY and GX wells) in Xinji County (Fig. 3) are shown in Fig. 5d. The trend of simulated and observed groundwater heads for these nine wells was generally consistent, especially during the calibration period, with all R values exceeding 0.79. In the validation period, wells PT, GCCG, XHL, situated in the middle reaches of the Hutuo River, exhibited negative R values, whereas all other wells had R values greater than 0.51. The model performance during the calibration period was obviously better than during the validation period, this discrepancy is also mainly due to a lack of accurate monthly-scale EWR dynamics from 2018 to 2020. For the PT, GCCG, and XHL wells, the differences between simulated and observed groundwater heads during the validation period were mostly within 1.5 m, with the average values for each well also closely matching: 27.90 m vs. 27.54 m for PT; 13.50 m vs. 14.27 m for GCCG; and 10.81 m vs. 11.51 m for XHL, indicating that the simulation results remained acceptable. In conclusion, the benchmark model could capture groundwater head fluctuations in the Shijiazhuang Plain.

Plain groundwater balance and key hydrological components

From September 2018 to December 2022, the variation in groundwater storage of the Shijiazhuang Plain simulated by the benchmark model was 1.51 billion m3, indicating an increase in groundwater storages (Table 4). The total recharge to the groundwater system was 8.64 billion m3, with the largest source being river leakage, accounting for approximately 46.48% of the total recharge. This indicates that EWR is a key driving force in groundwater storage increasement. The total discharge of the groundwater system was 7.14 billion m3, with agricultural irrigation extraction accounting for 58.21%. Groundwater depth in the Shijiazhuang Plain typically exceeds 30 to 40 m, resulting in minimal phreatic evaporation, and there is no discharge of groundwater into the rivers. Only unidirectional leakage from the rivers to the groundwater exists.

Table 4.

Comparison of water balance for the Shijiazhuang Plain groundwater system from September 2018 to December 2022 for different scenarios.

Groundwater balance
components (108 m3)		 Benchmark model Scenario 1 Scenario 2 Scenario 3
Recharge Soil water deep percolation 20.07 19.97 19.98 19.97
Agricultural water leakage 3.88 3.88 3.88 3.88
Canal leakage 2.66 2.49 2.52 2.49
River leakage 40.18 19.8 23.60 20.67
Specified-head inflow 19.64 21.03 20.78 21.03
Total recharge 86.43 67.17 70.76 68.04
Discharge Phreatic evaporation 0.01 0.01 0.01 0.01
Agricultural water extraction 41.54 41.54 41.54 41.54
Non-agricultural water extraction 5.92 5.92 5.92 5.92
Specified-head outflow 23.89 22.23 22.74 22.23
Total discharge 71.36 69.69 70.21 69.70
Variation of groundwater storage 15.07 −2.52 0.55 −1.66
Groundwater balance error 0.00 0.00 0.00 0.00
Open in a new tab

The key hydrological components of the Shijiazhuang Plain are shown in Fig. 6. Prior to October 2021, despite EWR, groundwater storage continued to decline slowly (Fig. 6a). After October 2021, groundwater storage began to increase, coinciding with a significant increase in EWR. The Sha-Zhulong River receives water from the SNWD-MR (Fig. 1a), resulting in relatively high leakage in the downstream river of this project (Fig. 6b). The Hutuo River reaches along the mountain front, as well as the middle and lower sections, exhibit favorable riverbed permeability and groundwater flow conditions (Fig. 4b). Therefore, the leakage in these areas is higher, with the cumulative leakage in individual grid cells reaching over 30 million m3 (Fig. 6b). Despite the fact that the areas along the EWR rivers are predominantly farmland with high groundwater extraction (Fig. 6c), groundwater storage along these rivers still shows a obvious increase (Fig. 6d), further highlighting the important role of EWR in recovering groundwater storage.

Fig. 6.

Fig. 6

Open in a new tab

From September 2018 to December 2022: (a) cumulated EWR volume in the Shijiazhuang Plain and cumulated variation of groundwater storage simulated by the benchmark model, (b) spatial distribution of cumulated river leakage, (c) agricultural groundwater extraction, and (d) groundwater storage variation simulated by the benchmark model.

Effects of EWR on groundwater recovery in the Shijiazhuang Plain

Effects of EWR on groundwater table recovery

As shown in Fig. 7a, from September 2018 to December 2022, EWR recovered the average groundwater table in the Shijiazhuang Plain by 1.96 m, averaging 0.45 m per year. EWR recovered the monthly average groundwater table by 0.06–1.96 m, with the recovery extent increasing over the majority of the time, reflecting the cumulative recovery effect of EWR. The monthly increment in groundwater table recovery more closely followed the monthly increment of EWR-induced leakage than the monthly EWR volume (Fig. 7a), suggesting that not all EWR directly contributed to groundwater recharge. Although EWR increased significantly after 2021, the increments in both EWR-induced leakage (1.02 vs. 1.04 billion m3) and groundwater table recovery (0.86 m vs. 1.10 m) were smaller than those in the period before 2021. The EWR-induced leakage totaled 2.04 billion m3, with high-value zones distributed along the mountain front and in the middle and lower reaches of the Hutuo River, the middle-lower reaches of the Shahe-Zhulong River, and the middle reaches of the Beisha–Huai River (Fig. 8a). These regions also experienced the most pronounced long-term groundwater table recovery (Fig. 8b1).

Fig. 7.

Fig. 7

Open in a new tab

(a) Comparison of monthly average groundwater table between the benchmark model and Scenario 1 simulations for the Shijiazhuang Plain, along with monthly EWR volume, monthly increment of EWR-induced leakage, and groundwater table recovery effect; (b) comparison of monthly groundwater storage variation between the benchmark model and Scenario 1 simulations, along with the monthly increment of groundwater storage recovery effect.

Fig. 8.

Fig. 8

Open in a new tab

Spatial distribution of (a) EWR-induced leakage, (b) groundwater table recovery, and (c) groundwater storage recovery, along with the impact radius of EWR for each river.

Comparison of scenario simulation results shows that by the end of 2022, long-term EWR led to groundwater table recovery across the Shijiazhuang Plain ranging from 0 to 13.02 m, with an average of 1.96 m and significant recovery (> 0.5 m) observed over 4000 km2, accounting for 60.78% of the plain (Fig. 8b1). EWR from the Hutuo River induced recoveries of 0–13.02 m, averaging 1.61 m, with local gains exceeding 5 m and 0.5 m within maximum radii of 9.5 km and 24 km, covering 3155 km2 (47.94% of the plain, Fig. 8b2). The Sha-Zhulong River EWR led to recoveries of 0–7.87 m, averaging 0.23 m, with local gains above 5 m and 0.5 m within 4.5 km and 12 km, covering 498 km2 (7.57%, Fig. 8b3). EWR from other Rivers resulted in recoveries of 0–11.44 m, averaging 0.12 m, with local gains exceeding 5 m and 0.5 m within 2 km and 5.5 km, covering 277 km2 (4.21%, Fig. 8b4). Overall, the contributions of the Hutuo River EWR, Sha-Zhulong River EWR, and other Rivers’ EWR to long-term groundwater table recovery were 82.04% (1.61m/1.96m), 11.66% (0.23m/1.96m), and 6.30% (0.12m/1.96m), respectively, with the Hutuo River EWR being the dominant contributor.

Effects of EWR on groundwater storage recovery

As shown in Fig. 7b, from September 2018 to December 2022, without EWR, the groundwater storage variation was −0.25 billion m3, compared to 1.51 billion m3 in the benchmark model. In other words, EWR helped reverse the overexploitation and depletion of groundwater storage in the plain. A total of 1.76 billion m3 of groundwater storage was recovered by all river EWRs, averaging 0.41 billion m3 per year. The monthly increment in groundwater storage recovery also closely tracked that of EWR-induced leakage (Fig. 7), with the cumulative amount being higher before 2021 (0.96 billion m3 vs. 0.8 billion m3). Spatially, areas with high EWR-induced leakage also exhibited the most pronounced long-term recovery of the groundwater storage (Fig. 8a,c1).

The scenario simulation results show that by the end of 2022, long-term EWR led to significant groundwater storage recovery (> 50,000 m3/km2) across 4,076 km2 of the Shijiazhuang Plain, accounting for 61.94% of the total plain area (Fig. 8c1). EWR from the Hutuo River induced recoveries of 1.45 billion m3, with local gains exceeding 500,000 m3/km2 and 50,000 m3/km2 within maximum radii of 14.5 km and 24.5 km, covering 3,225 km2 (49.01% of the plain, Fig. 8c2). The Sha-Zhulong River EWR led to recoveries of 0.22 billion m3, with local gains above 500,000 m3/km2 and 50,000 m3/km2 within 6 km and 12.5 km, covering 526 km2 (7.99%, Fig. 8c3). EWR from other Rivers resulted in recoveries of 0.09 billion m3, with local gains exceeding 500,000 m3/km2 and 50,000 m3/km2 within 2.5 km and 5.5 km, covering 278 km2 (4.22%, Fig. 8c4). Overall, the contributions of the Hutuo River EWR, Sha-Zhulong River EWR, and other Rivers’ EWR to long-term groundwater storage recovery were 82.52% (1.45/1.76 billion m3), 12.59% (0.22/1.76 billion m3), and 4.89% (0.09/1.76 billion m3), respectively, with the Hutuo River EWR being the dominant contributor.

Along with the increase in EWR volume, there was not only a significant rise in river leakage and groundwater storage but also a certain increase in soil water deep percolation, canal leakage (SNWD-MR and its associated water diversion canals), and specified-head outflow, while specified-head inflow decreased to some extent (Table 4). These flux changes are closely related to the dynamic groundwater table. For example, the increase in EWR volume leads to a rise in the groundwater table, which in turn results in a decrease in specified-head inflow and an increase in specified-head outflow. An elevated groundwater table reduces the vadose zone thickness, thereby enhancing deep percolation of soil water to the groundwater table. These demonstrate that EWR impacts the groundwater storage of the Shijiazhuang Plain through a complex surface water–soil water–groundwater cycle.

Discussion

Groundwater recovery efficiency

The EWR, EWRLR, and EWRGSRE for each river are shown in Fig. 9. In 2018, the Hutuo River had a large monthly EWR volume, resulting in relatively low EWRLR and EWRGSRE values, around 40%. From then until mid-2021, the monthly EWR volume decreased and stabilized, leading to a gradual increase in EWRLR and EWRGSRE to around 60%. However, after mid-2021, with a significant increase in EWR volume, both EWRLR and EWRGSRE quickly dropped back to around 40%. After mid-2021, the EWRLR and EWRGSRE of the Sha-Zhulong River and other Rivers also rapidly decreased. Corresponding, despite the EWR volume before 2021 being only about half of that after 2021 (Table 2), the EWR-induced leakage increment before 2021 remained slightly higher (1.04 vs. 1.02 billion m3, Fig. 7a). Meanwhile, the significant increase in EWR after 2021 would lead to substantial groundwater outflow from the plain (Table 4). These two factors combined resulted in higher increments in the recovery of groundwater table (1.10 vs. 0.86 m) and storage (0.96 vs. 0.80 billion m3) before 2021 than after.

Fig. 9.

Fig. 9

Open in a new tab

EWR, EWRLR, and EWRGSRE of the Hutuo River (a), Sha-Zhulong River (b), and other Rivers (c).

Long-term EWR strategy for the Shijiazhuang Plain

The Sha-Zhulong River is the shortest EWR river in the Shijiazhuang Plain (Fig. 1a), and thus, by the end of 2022, its EWRLR and EWRGSRE were the lowest, at 25.64% and 19.32%, respectively (Fig. 9). Despite contributing about 12% to the groundwater table and storage recovery but occupying 22.64% of the total EWR, this indicates that it should not be considered the main EWR river. The EWR to other Rivers is small, so their EWRLR and EWRGSRE are the highest, at 61.49% and 61.33%, respectively (Fig. 9). These rivers contribute about 5.5% to the groundwater table and storage recovery, occupying 2.77% of the total EWR. However, the groundwater flow conditions in the Beisha-Huai River region are relatively poor (Fig. 4b), which results in the groundwater table and storage recovery from EWR being difficult to diffuse, limiting the affected area (Fig. 8b4,c4). Therefore, these rivers should also not be considered the primary EWR rivers. The EWRLR and EWRGSRE of the Hutuo River were moderate, at 43.89% and 38.44%. It contributes about 82.5% to the groundwater table and storage recovery, occupying 74.59% of the total EWR, with a maximum impact radius extending up to approximately 24.5 km (Fig. 8b2,c2). This relatively high EWRGSRE and wide impact range are partly due to the long length of the Hutuo River, which favors the infiltration of EWR (Fig. 1a), and partly due to the relatively high hydraulic conductivity along the river, facilitating the diffusion of EWR effects (Fig. 4b). For long-term, large-scale EWR in the Shijiazhuang Plain in the future, the Hutuo River is suggested to be the primary EWR river, achieving a relatively high groundwater storage recovery efficiency while ensuring widespread recovery of the groundwater table across the plain.

Shijiazhuang City faces severe scarcity of surface water resources, with the average annual volume from 2018 to 2022 being only 1.06 billion m3, corresponding to approximately 100 m3 per capita per year. Moreover, the majority of the plain is agricultural land (Fig. 6c), which produces little to no surface water, with nearly all surface water originating from the mountainous areas. To efficiently achieve the recovery of the plain’s groundwater system while meeting the needs of socioeconomic water demands, two strategies are suggested: First, use external water sources such as the SNWD-MR for EWR (Fig. 1a); second, improve the EWRLR and EWRGSRE, which requires utilizing the storage and regulation capacity of reservoirs, canals, and other infrastructure to compensate for the EWR flow in dry periods by utilizing the surplus water during wet periods, maintaining a relatively stable and moderate EWR flow rather than follow the “greater is better” approach (Fig. 9). In simulations for the calibration and validation periods, driven by precipitation data from 2019 (a dry year, 427.58mm), 2020 (a normal year, 576.14mm), and 2021 (a wet year, 891.26mm), respectively, while keeping other settings of the benchmark model unchanged, it was found that in the dry period, the Hutuo River EWR is insufficient to maintain the groundwater recharge-discharge balance in the Shijiazhuang Plain. When the Hutuo River EWR flow is 14.64 m3/s (460 million m3/year), the EWRLR and EWRGSRE are high, about 83% and 77%, respectively. At this point, the average groundwater table in the Shijiazhuang Plain dropped by 0.28 m per year, and the groundwater storage decreased by 164 million m3 per year. Further increases in Hutuo River EWR flow rate would quickly reduce the EWRLR and EWRGSRE. Therefore, during the dry period, additional EWR should be conducted in areas such as the Sha-Zhulong River and Beisha-Huai River to maintain stable groundwater table and storage. During the normal period, when the Hutuo River EWR reaches 8.4 m3/s (265 million m3/year) or more, the average groundwater table and storage in the Shijiazhuang Plain will rise. During the wet period, no EWR is needed, and the average groundwater table rises by 1.68 m per year, with an annual increase of 1.28 billion m3 in groundwater storage.

Now, the EWR rivers are located in the northern (Shahe-Zhulong River), central (Hutuo River), and southwestern (Beisha-Huai River, Xiao River, Wu River) parts of the Shijiazhuang Plain. In these areas, groundwater storages have gradually increased (Fig. 6d). However, in the southeastern part of the plain, groundwater storages continue to decline (Fig. 6d). Owing to the imperative of ensuring grain production in the Shijiazhuang Plain, groundwater extraction is expected to remain at a high level over the long term. Therefore, in the future, it is important to select suitable rivers in the southeastern part of the plain for EWR to help slow down or even reverse the decline in groundwater storages in that area.

Uncertainties and limitations

The simulation results shows that large-scale, long-term EWR has significantly recovered the groundwater table and groundwater storage in the Shijiazhuang Plain, which is consistent with findings from studies across the North China Plain17,23,50. Nevertheless, uncertainties and limitations remain in several aspects. First, the absence of accurate monthly EWR data from 2018 to 2020 introduces uncertainty into the short-term simulation results; therefore, this study focuses on assessing the multi-year recovery effects of large-scale EWR rather than short-term effects at the monthly scale. Second, the estimates of suitable EWR flows for dry, normal, and wet precipitation periods, derived from meteorological data for 2019–2021, are subject to certain uncertainties. For example, even in a year classified as hydrologically dry based on low annual precipitation, if the majority of rainfall occurs during the irrigation seasons (March–June and November in Shijiazhuang Plain), groundwater abstraction could be substantially reduced. In such cases, groundwater tables and storage might remain stable without the need for large-scale EWR. Third, the choice of different models may also introduce potential uncertainties in the assessment of recovery effects. Future research should prioritize the acquisition of higher-resolution EWR datasets to further improve model accuracy and interpretability. In addition, accounting for the intra-annual distribution of precipitation will be critical for a more comprehensive assessment of appropriate EWR schemes in the Shijiazhuang Plain.

Conclusions

In this study, the surface–groundwater coupled model MODCYCLE was used to assess the long-term recovery effects and efficient strategies of large-scale EWR on the Shijiazhuang Plain groundwater system, the largest shallow groundwater depression cone in the North China Plain. The main findings are as follows:

  1. Large-scale EWR helped reverse groundwater overexploitation and depletion in the Shijiazhuang plain. From September 2018 to December 2022, EWR recovered the average groundwater table in the Plain by 1.96 m and recovered groundwater storage by 1.76 billion m3, averaging 0.45 m and 0.41 billion m3 per year, respectively. Contributions from the EWR in the Hutuo River, the Sha-Zhulong River, and other Rivers accounted for approximately 82.5%, 12.0%, and 5.5%, respectively.

  2. From September 2018 to December 2022, EWR in the Hutuo River, Sha-Zhulong River, and other Rivers led to significant recovery of groundwater table and storage along the rivers (> 0.5 m; > 50,000 m3/km2), with maximum recovery within radial distances of approximately 24.5 km, 12.5 km, and 5.5 km, respectively. The highest-recovery zones were distributed along the mountain front and in the middle and lower reaches of the EWR Rivers.

  3. For long-term, large-scale EWR in the Shijiazhuang Plain, the Hutuo River should serve as the primary replenishment channel due to its favorable infiltration path length and groundwater flow conditions.The EWR flow rate is suggested to remain relatively stable and moderate, rather than following a “greater is better” approach, in order to achieve both efficient groundwater storage recovery and widespread groundwater table recovery. During drought periods, the EWR flow in the Hutuo River of 14.64 m3/s (460 million m3/year) would efficiently mitigate groundwater depletion, but EWR in additional rivers is still necessary to maintain stable groundwater table and storage. Under normal precipitation conditions, a flow of 8.4 m3/s (265 million m3/year) or greater is sufficient to reverse groundwater depletion.

These findings provide scientific support for the continued fight against groundwater overexploitation in the North China Plain and offer insights for optimizing EWR strategies to efficiently recover groundwater in globally overexploited aquifers.

Author contributions

Wen Lu: Conceptualization, Methodology, Software, Data curation, Writing—original draft, Writing—review & editing, Investigation. Chuiyu Lu: Conceptualization, Methodology, Software, Investigation, Funding acquisition, Visualization, Supervision. Yangwen Jia: Methodology, Writing—review & editing. Yong Zhao: Methodology. Xin He: Writing—review & editing. Qingyan Sun: Data curation, Writing—review & editing. Chu Wu: Data curation, Writing—review & editing. Wenjia Zhang: Data curation, Writing—review & editing. Zhenjiang Wu: Writing—review & editing.

Funding

This work was supported by the National Key Research and Development Program of China (Grant No. 2025YFC3215100).

Data availability

The figure and table source data that support the findings of this study are publicly available at 10.5281/zenodo.18278147. Further data are available from the corresponding author upon reasonable request.

Declarations

Competing interests

The authors declare no competing interests.

Ethical approval

The authors confirm that the submission is original work and not published elsewhere.

Consent to participate

All authors consented to participate in this study.

Consent for publication

All authors consented to publish this study.

Footnotes

Publisher’s note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

References

  • 1.Jasechko, S. et al. Rapid groundwater decline and some cases of recovery in aquifers globally. Nature.625(7996), 715–721. 10.1038/s41586-023-06879-8 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 2.Werner, A. D. et al. Seawater intrusion processes, investigation and management: Recent advances and future challenges. Adv. Water Resour.51, 3–26. 10.1016/j.advwatres.2012.03.004 (2013). [Google Scholar]
  • 3.Herrera-García, G. et al. Mapping the global threat of land subsidence. Science.371(6524), 34–36. 10.1126/science.abb8549 (2021). [DOI] [PubMed] [Google Scholar]
  • 4.de Graaf, I. E. M., Gleeson, T., van Beek, L. P. H., Sutanudjaja, E. H. & Bierkens, M. F. P. Environmental flow limits to global groundwater pumping. Nature.574(7776), 90–94. 10.1038/s41586-019-1594-4 (2019). [DOI] [PubMed] [Google Scholar]
  • 5.Jasechko, S. & Perrone, D. Global groundwater wells at risk of running dry. Science.372(6540), 418–421. 10.1126/science.abc2755 (2021). [DOI] [PubMed] [Google Scholar]
  • 6.Dillon, P. et al. Sixty years of global progress in managed aquifer recharge. Hydrogeol. J.27(1), 1–30. 10.1007/s10040-018-1841-z (2019). [Google Scholar]
  • 7.Sufyan, M., Martelli, G., Teatini, P., Cherubini, C. & Goi, D. Managed aquifer recharge for sustainable groundwater management: New developments, challenges, and future prospects. Water. 16, 3216. 10.3390/w16223216 (2024). [Google Scholar]
  • 8.Zhang, H., Xu, Y. & Kanyerere, T. A review of the managed aquifer recharge: Historical development, current situation and perspectives. Phys. Chem. Earth.118–119, 102887. 10.1016/j.pce.2020.102887 (2020). [Google Scholar]
  • 9.Yang, W. et al. Human intervention will stabilize groundwater storage across the North China Plain. Water Resour. Res.58(2), e2021WR030884. 10.1029/2021WR030884 (2022). [Google Scholar]
  • 10.Zhao, X. et al. A novel framework for heterogeneity decomposition and mechanism inference in spatiotemporal evolution of groundwater storage: Case study in the North China Plain. Water Resour. Res.60(12), e20230WR036102. 10.1029/2023WR036102 (2024). [Google Scholar]
  • 11.Yang, H. et al. Evolution of groundwater level in the North China Plain in the past 40 years and suggestions on its overexploitation treatment. Geology in China. 48(4), 1142–1155. 10.12029/gc20210411 (2021). [Google Scholar]
  • 12.He, Z. et al. Variations of groundwater dynamics in alluvial aquifers with reclaimed water restoring the overlying river, Beijing, China. Water.13, 806. 10.3390/w13060806 (2021). [Google Scholar]
  • 13.Li, Y., Wu, Y. & Zhu, J. Effectiveness of ecological recovery of rivers and lakes in groundwater over-exploitation areas of North China. Haihe Water Resources.2024(12), 13–20. 10.3969/j.issn.1004-7328.2024.12.003 (2024). [Google Scholar]
  • 14.Nan, T. et al. Evaluation of shallow groundwater dynamics after water supplement in North China Plain based on attention-GRU model. J. Hydrol.625, 130085. 10.1016/j.jhydrol.2023.130085 (2023). [Google Scholar]
  • 15.Wang, X., Ma, H. & Li, Y. Practice and reflection of ecological environment recovery of rivers and lakes in the Haihe River Basin. Haihe Water Resources.2023(10), 29–32. 10.3969/j.issn.1004-7328.2023.10.007 (2023). [Google Scholar]
  • 16.Nan, T. & Cao, W. Effect of ecological water supplement on groundwater restoration in the Yongding River based on multi-model linkage. Water.15, 374. 10.3390/w15020374 (2023). [Google Scholar]
  • 17.Sun, K. et al. Quantifying the contribution of ecological water replenishment on aquifer recovery using a refined groundwater model. Sci. Total Environ.857, 159216. 10.1016/j.scitotenv.2022.159216 (2023). [DOI] [PubMed] [Google Scholar]
  • 18.Hu, L., Guo, J., Zhang, S., Sun, K. & Yang, Z. Response of groundwater regime to ecological water replenishment of the Yongding River. Hydrogeology & Engineering Geology.47(5), 5–11. 10.16030/j.cnki.issn.1000-3665.20200802 (2020). [Google Scholar]
  • 19.Li, Z. et al. Ecological water replenishment and groundwater response in Chaobai River in spring. South-to-North Water Transfers and Water Science & Technology.20(2), 375–384. 10.13476/j.cnki.nsbdqk.2022.0038 (2022). [Google Scholar]
  • 20.Ma, R., Chen, F., Cheng, G., Shi, W. & Zhang, L. Effect of ecological water replenishment on groundwater recharge in the Beijing-Hangzhou Grand Canal. South-to-North Water Transfers and Water Science & Technology.22(3), 462–472. 10.13476/j.cnki.nsbdqk.2024.0048 (2024). [Google Scholar]
  • 21.Sun, C. et al. Analysis of groundwater change around Yongding River since ecological water replenishment. Journal of Beijing Normal University (Natural Science).60(5), 717–723. 10.12202/j.0476-0301.2024059 (2024). [Google Scholar]
  • 22.Chen, J. et al. Influence of ecological water replenishment on groundwater level and water quality along the Yongding River. Earth and Environment.51(3), 266–273. 10.14050/j.cnki.1672-9250.2022.50.113 (2023). [Google Scholar]
  • 23.Lyu, K. et al. Data-driven and numerical simulation coupling to quantify the impact of ecological water replenishment on surface water–groundwater interactions. J. Hydrol.649, 132508. 10.1016/j.jhydrol.2024.132508 (2025). [Google Scholar]
  • 24.Sun, K. et al. Enhancing the understanding of hydrological responses induced by ecological water replenishment using improved machine learning models: A case study in Yongding River. Sci. Total Environ.768, 145489. 10.1016/j.scitotenv.2021.145489 (2021). [DOI] [PubMed] [Google Scholar]
  • 25.Xu, C. et al. Study on the processes influencing and importance of ecological water replenishment for groundwater resources: A case study in Yongding River. Water. 14, 828. 10.3390/w14050828 (2022). [Google Scholar]
  • 26.Arnold, J. G., Srinivasan, R., Muttiah, R. S. & Williams, J. R. Large area hydrologic modeling and assessment. Part I: Model development. J. Am. Water Resour. Assoc.34(1), 73–89. 10.1111/j.1752-1688.1998.tb05961.x (1998). [Google Scholar]
  • 27.Sánchez-Gómez, A., Schürz, C., Molina-Navarro, E. & Bieger, K. Groundwater modelling in SWAT+: Considerations for a realistic baseflow simulation. Groundw. Sustain. Dev.26, 101275. 10.1016/j.gsd.2024.101275 (2024). [Google Scholar]
  • 28.Harbaugh, A. W. MODFLOW 2005, the U. S. Geological Survey Modular Ground-Water Model — The Ground-Water Flow Process. USGS Tech. Met.10.3133/tm6A16 (2005). [Google Scholar]
  • 29.Trefry, M. G. & Muffels, C. FEFLOW: A finite-element ground water flow and transport modeling tool. Ground Water.45(5), 525–528. 10.1111/j.1745-6584.2007.00358.x (2007). [Google Scholar]
  • 30.Bailey, R. T., Wible, T. C., Arabi, M., Records, R. M. & Ditty, J. Assessing regional-scale spatio-temporal patterns of groundwater–surface water interactions using a coupled SWAT–MODFLOW model. Hydrol. Process.143, 103662. 10.1002/hyp.10933 (2016). [Google Scholar]
  • 31.Hughes, J. D. & Liu, J. MIKE SHE: Software for integrated surface water/ground water modeling. Ground Water.46, 797–802. 10.1111/j.1745-6584.2008.00500.x (2008). [Google Scholar]
  • 32.Lu, C., Qin, D., Zhang, J. & Wang, R. MODCYCLE—an object-oriented modularized hydrological model I: Theory and development. Journal of Hydraulic Engineering.43(10), 1135–1145. 10.13243/j.cnki.slxb.2012.10.002 (2012). [Google Scholar]
  • 33.Luan, Q. et al. Agricultural drought evaluation based on a soil moisture index coupled hydrological model in North China Plain. Ecol. Indic.166, 112473. 10.1016/j.ecolind.2024.112473 (2024). [Google Scholar]
  • 34.Sun, Q. et al. Response of groundwater to different water resource allocation patterns in the Sanjiang Plain, Northeast China. J. Hydrol. Reg. Stud.42, 101156. 10.1016/j.ejrh.2022.101156 (2022). [Google Scholar]
  • 35.Zhang, B. et al. Using storage of coal-mining subsidence area for minimizing flood. J. Hydrol.572, 571–581. 10.1016/j.jhydrol.2019.03.044 (2019). [Google Scholar]
  • 36.Shu, Y., Villholth, K. G., Jensen, K. H., Stisen, S. & Lei, Y. Integrated hydrological modeling of the North China Plain: Options for sustainable groundwater use in the alluvial plain of Mt. Taihang. J. Hydrol.464, 79–93. 10.1016/j.jhydrol.2012.06.048 (2012). [Google Scholar]
  • 37.Li, Y. et al. Investigation of quality and pollution characteristics of groundwater in the Hutuo River Alluvial Plain, North China Plain. Environ. Earth Sci.75, 581. 10.1007/s12665-016-5366-2 (2016). [Google Scholar]
  • 38.Zhang, R. et al. Replenishment impacts on hydrogeochemistry and water quality in the Hutuo River Plain. Water. 15, 3326. 10.3390/w15193326 (2023). [Google Scholar]
  • 39.Sun, Z. et al. Effects of manure on topsoil and subsoil organic carbon depend on irrigation regimes in a 9-year wheat–maize rotation. Soil Tillage Res.205, 104790. 10.1016/j.still.2020.104790 (2021). [Google Scholar]
  • 40.Yan, Z. et al. Grain yield and water productivity of winter wheat controlled by irrigation regime and manure substitution in the North China Plain. Agric. Water Manag.295, 108731. 10.1016/j.agwat.2024.108731 (2024). [Google Scholar]
  • 41.Li, Y., Zheng, L., Li, J. & Zhang, B. Ecological water replenishment effect in Hutuo River and driving factors. Chinese Journal of Ecology.43(11), 3423–3432. 10.13292/j.1000-4890.202411.021 (2024). [Google Scholar]
  • 42.Meng, R. et al. Optimizing groundwater recharge plan in North China Plain to repair shallow groundwater depression zone, China. J. Groundw. Sci. Eng.11(2), 133–145. 10.26599/JGSE.2023.9280012 (2023). [Google Scholar]
  • 43.Li, X., Ye, S., Wang, L. & Zhang, J. Tracing groundwater recharge sources beneath a reservoir on a mountain-front plain using hydrochemistry and stable isotopes. Water Science and Technology-Water Supply.17(5), 1447–1457. 10.2166/ws.2017.036 (2017). [Google Scholar]
  • 44.Tian, W., Han, S., Guo, Y. & Wei, S. Management measures and outcomes in the upstream small watershed of the Baicaoping Reservoir. Soil And Water Conservation In China.2014(8), 18–19. 10.14123/j.cnki.swcc.2014.08.037 (2014). [Google Scholar]
  • 45.Xu, G., Yang, F. & Liang, S. Application of numerical simulation in the dam reinforcement of Hengshanling Reservoir. South-to-North Water Transfers and Water Science & Technology.7(02), 93–95. 10.3969/j.issn.1672-1683.2009.02.030 (2009). [Google Scholar]
  • 46.Yao, J. et al. 3-D flood mapping from SWOT observations during extreme rainfall: A case study of Gangnan Reservoir. Int. J. Digit. Earth.18(1), 2544916. 10.1080/17538947.2025.2544916 (2025). [Google Scholar]
  • 47.Zhang, X. Research on forecast of agricultural water-limited irrigation based on groundwater numerical simulation technology-Taking Shijiazhuang Plain as an example. Master’s thesis, China University of Geosciences Beijing, Beijing, China.10.27493/d.cnki.gzdzy.2021.000309 (2021). [Google Scholar]
  • 48.Liu, B. et al. Analysis of runoff according to land-use change in the upper Hutuo River Basin. Water. 15, 1138. 10.3390/w15061138 (2023). [Google Scholar]
  • 49.Omer, A., Wang, W., Basheer, A. K. & Yong, B. Integrated assessment of the impacts of climate variability and anthropogenic activities on river runoff: A case study in the Hutuo River Basin, China. Hydrol. Res.48(2), 416–430. 10.2166/nh.2016.229 (2017). [Google Scholar]
  • 50.Long, D. et al. Unprecedented large-scale aquifer recovery through human intervention. Nat. Commun.16(1), 7296. 10.1038/s41467-025-62719-5 (2025). [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Data Availability Statement

The figure and table source data that support the findings of this study are publicly available at 10.5281/zenodo.18278147. Further data are available from the corresponding author upon reasonable request.

Articles from Scientific Reports are provided here courtesy of Nature Publishing Group

RESOURCES