SEARCH

SEARCH BY CITATION

Keywords:

  • North China Plain;
  • groundwater depletion;
  • groundwater storage;
  • groundwater sustainability;
  • groundwater modelling;
  • groundwater recharge

Abstract

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Study Site and Hydrogeologic Data
  5. 3. Groundwater Storage Change From Water Level Data
  6. 4. Simulation of the Groundwater Flow System
  7. 5. Results and Discussion
  8. 6. Approaches Toward More Sustainable Groundwater Management
  9. 7. Model Limitations and Future Studies
  10. 8. Conclusions
  11. Acknowledgments
  12. References
  13. Supporting Information

[1] The North China Plain (NCP) is one of the global hotspots of groundwater depletion. Currently, our understanding is limited on spatiotemporal variability in depletion and approaches toward more sustainable groundwater development in this region. This study was intended to simulate spatiotemporal variability in groundwater depletion across the entire NCP and explore approaches to reduce future depletion. Simulated predevelopment groundwater recharge (∼13 km3/yr) primarily discharged as base flow to rivers and evapotranspiration. Initial groundwater storage was estimated to be 1500 km3 of drainable storage in shallow aquifers and 40 km3 of compressive storage in deep aquifers. Simulated groundwater depletion from 1960s to 2008 averaged ∼4 km3/yr. Cumulative depletion was 50 km3 (∼20% of pumpage) in the piedmont district, 103 km3 (∼20%) in the central plain, and 5 km3 (12%) in the coastal plain. However, depletion varied with time: ∼2.5 km3/yr in the 1970s, ∼4.0 in the 1980s, ∼2.0 in 1990–1996; ∼7.0 in 1997–2001, and ∼4.0 in 2002–2008. Recharge also varied spatially, averaging ∼120 mm/yr and concentrated in the piedmont district (200–350 mm/yr) while lower in the central and coastal plains (50–100 mm/yr). Simulation of several alternatives, including managed aquifer recharge, increased water use efficiency, brackish water use, and interbasin water transfer, indicated that the combination of these strategies could be used to recover groundwater storage by 50 km3 over a 15-year period. This study provides valuable insights for developing more sustainable groundwater management options for the NCP; the methods are useful for managing other depleted aquifers.

1. Introduction

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Study Site and Hydrogeologic Data
  5. 3. Groundwater Storage Change From Water Level Data
  6. 4. Simulation of the Groundwater Flow System
  7. 5. Results and Discussion
  8. 6. Approaches Toward More Sustainable Groundwater Management
  9. 7. Model Limitations and Future Studies
  10. 8. Conclusions
  11. Acknowledgments
  12. References
  13. Supporting Information

[2] Groundwater is an essential source of water for agricultural, industrial, and environmental uses as well as for drinking water supply due to its generally good quality and widespread occurrence. Over the decade between 1990 and 2000, groundwater contributed about 20% of human's fresh water needs globally [Kinzelbach et al., 2003], and estimated total groundwater withdrawals ranged from ∼750 to 800 km3/yr [Shah et al., 2003; Falkenmark, 2005; Wada et al., 2010]. Groundwater exploitation has facilitated economic development and food production. Global crop yield per unit area increased by a factor of 2.3, and total crop yield increased by a factor of 2.4 between 1961 and 2004 [Oki and Kanae, 2006]; with respect to China, maize and wheat yields have increased by factors of 1.5 and 2.4, respectively, over the last four decades [Piao et al., 2010]. These increases would not have been possible without major increases in irrigation, a significant share of which was derived from groundwater. Currently, groundwater supplies ∼40% of irrigation water globally [Siebert et al., 2010], and ∼25% of irrigation water in north China [Chen and Ma, 2002].

[3] However, negative environmental impacts of groundwater depletion from overexploitation are well known, e.g., groundwater quality deterioration, ecosystem degradation, and in some cases, land subsidence and/or seawater intrusion and sea level rise [Sophocleous, 2002; Konikow and Kendy, 2005; Fogg and LaBolle, 2006; Konikow, 2011]. During the past 50 years, excessive groundwater depletion has become a global problem, affecting major regions of North Africa, the Middle East, South and Central Asia, North China, North America, and Australia [Shah et al., 2000, 2003; Konikow and Kendy, 2005; Wada et al., 2010]. The North China Plain (NCP) is one of global hotspots of groundwater depletion [Alley et al., 2002; Zheng et al., 2010].

[4] Groundwater sustainability is commonly defined as groundwater development in a manner that can be maintained for an indefinite time without causing unacceptable environmental, economic, or social consequences [Alley et al., 1999]. Groundwater depletion refers to sustained loss of groundwater from storage [Hurd et al., 1999; Kinzelbach et al., 2003]. As discussed by Theis [1940], some loss of water from aquifer storage always occurs in response to any increase in withdrawal from wells. The groundwater budget of an aquifer system that is being pumped can be expressed as [Bredehoeft, 2002; Zhou, 2009]:

  • display math(1)

where R is natural recharge, ΔR is increased recharge induced by pumping, D is natural discharge, ΔD is decreased discharge caused by pumping, P is pumping, and ΔS is change in groundwater storage.

[5] With the initial phase of groundwater pumping, all pumpage water is derived from aquifer storage; however, with time, pumpage can be increasingly balanced by increased recharge and/or decreased discharge. A new equilibrium can prevail when all pumpage is balanced by increases in recharge and or decreases in discharge. If the aquifer is initially being pumped but the pumping rate is increased, the same principle applies, except that a new equilibrium will not be attained until the total initial discharge has decreased and/or the recharge has increased, by amounts which sum to the increase in pumpage. Except in cases where recharge can be increased, such as from increases in precipitation, land use change, or irrigation return from surface water irrigation, developing sustainable pumpage generally relies on capturing groundwater discharge to surface water (i.e., groundwater evapotranspiration (ET) from declining water tables or groundwater discharge to streams) [Johnston, 1997]. Should the pumping rate increase exceed the sum of the total reduction in initial discharge and the total increase in recharge which can be achieved, storage loss will continue until ultimately declining water levels will force a reduction in the pumping rate.

[6] Persistent, multiyear declines in groundwater levels are usually an indication that pumping has in fact reached levels that cannot be balanced by increases in recharge and/or decreases in prior discharge, and thus that the level of pumping is not sustainable. However, careful analysis of the initial flow regime, the changes imposed by pumping, and the changes likely under continued pumping are generally required to test this interpretation. A numerical flow model may be the most efficient and effective tool to conduct these analyses, gain reasonable information on relationships among the components of the water budget, and simulate the impacts of proposed groundwater management alternatives. Sustainability assessments depend on such information and must consider spatiotemporal variability in pumpage and recharge [Alley et al., 1999; Foster, 2000; Alley and Leake, 2004]. Given the complexity and heterogeneity of aquifer systems, and the ability of numerical modeling to integrate different data and to evaluate regional groundwater dynamics, simulation is clearly an excellent tool for use in sustainability assessments [Luckey and Becker, 1999; Kinzelbach et al., 2003; Konikow and Kendy, 2005; Welsh, 2006].

[7] In the NCP, several numerical models have been built to evaluate groundwater resources and flow dynamics. Zhang et al. [2008] assessed groundwater budgets during predevelopment and estimated total inflow of 15.7 km3, including recharge from precipitation of 13.6 km3, mountain front recharge of 1.8 km3, leakage from rivers of 0.3 km3, and total outflow including discharge to rivers of 8.8 km3, ET of 5.0 km3, discharge to the Bohai Sea of 1.5 km3, and discharge to the inner flooded area of the Yellow River of 0.2 km3. S. Wang et al. [2008] evaluated groundwater depletion in the NCP from January 2002 to December 2003 and estimated total inflow of 49.4 km3 relative to total outflow of 56.5 km3, resulting in a budget deficit of 7.1 km3. Zhang [2007] evaluated flow budgets in the shallow part of the aquifer system and estimated an average annual depletion rate of 2.5 km3 from 1980 to 2000. Cui et al. [2009] used groundwater modeling to evaluate effects of groundwater pumpage reductions in response to the south-to-north water transfer (SNWT) (middle route) in 10 years of prediction period and estimated average groundwater level recovery rates of 2.1 m/yr in the shallow aquifer in Shijiazhuang in the Piedmont region and 0.8–1.5 m/yr in the deep aquifer in Dezhou in the central plain. Xue et al. [2010] used a numerical flow model in the shallow aquifer calibrated using hydraulic data from 1965 to 2005 to evaluate groundwater storage recovery as a result of the SNWT project and estimated that by the end of 2030 groundwater storage will recover to the state in 2000. Additionally, some of the existing models were restricted to local areas of the NCP. Jia and Liu [2002] used numerical modeling to evaluate groundwater level recovery in Luancheng in the piedmont region and estimated that reducing groundwater pumpage by 50 and 100 mm in the simulation period (January 1990 to December 1990) would result in water level recoveries of 0.25 and 0.56 m, respectively. Liu et al. [2008] used a model to simulate flow dynamics from 1960 to 2004 to evaluate the effects of urbanization of rural areas in the vicinity of Shijiazhuang City and the SNWT project and concluded that the SNWT project can only mitigate depletion in local areas, however, urbanization can result in water level recoveries of 3–15 m. Hu et al. [2010] integrated a crop-growth model and a groundwater model to evaluate the effects of reduction in irrigation in Shijiazhuang and estimated that ∼140 mm reduction in irrigation could stop groundwater depletion in this area.

[8] Some of these previous models were developed to assess groundwater budgets during predevelopment or to evaluate flow dynamics over a relatively short period after the groundwater had been highly developed. Models with relatively long simulation periods focused primarily on the shallow part of the aquifer system; however, the deep part of the aquifer system plays an important role in water supply. Therefore, these models limited understanding of the groundwater balance and the impact of groundwater development across the entire NCP since predevelopment. A further limitation of most existing models is that groundwater recharge is calculated by multiplying precipitation by a coefficient, and this coefficient is assumed to be constant over time; this approach does not allow simulation of impacts of water level declines on recharge. The objective of this study was to use groundwater modeling to simulate spatiotemporal variability in groundwater budgets, focusing on depletion, and to examine potential approaches to reduce depletion in the future. The numerical groundwater flow model for the entire NCP was used to simulate both predevelopment (1960s) and postdevelopment (1960s–2008) conditions. This is the first time that groundwater dynamics are simulated across the NCP from predevelopment to present. The calibrated flow model provides an invaluable tool for comprehensive assessment of the long-term, large-scale groundwater balance that can be used to evaluate approaches toward more sustainable management.

2. Study Site and Hydrogeologic Data

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Study Site and Hydrogeologic Data
  5. 3. Groundwater Storage Change From Water Level Data
  6. 4. Simulation of the Groundwater Flow System
  7. 5. Results and Discussion
  8. 6. Approaches Toward More Sustainable Groundwater Management
  9. 7. Model Limitations and Future Studies
  10. 8. Conclusions
  11. Acknowledgments
  12. References
  13. Supporting Information

2.1. Study Site

[9] The NCP, as defined in this study, ranges from the Yanshan Mountains in the north, the Bohai Sea on the east, the Yellow River in the south, and the Taihang Mountains on the west (Figure 1). The total area of the NCP as defined is ∼140,000 km2. Some studies define the NCP as extending to the Huaihe River in the south with a much larger area [Song et al., 2010]. The NCP covers the entire plains of the Beijing Municipality, the Tianjin Municipality, the Hebei Province, and the northern parts of plains of the Shandong and Henan Provinces. The population of the NCP, based on 2000 county census data, is ∼105 million [National Bureau of Statistics of China, 2003]. The average population density is ∼800/km2, ranging from ∼ 400/km2 in rural areas to ∼12,000/km2 in the Beijing metropolis. The region accounts for ∼15% of China's total gross domestic product and ∼10% of total grain production based on 2009 statistical data [National Bureau of Statistics of China, 2010].

image

Figure 1. Location and topography of the North China Plain. (1) Boundaries of geomorphologic zones: (I) piedmont plain, (II) alluvial fan and flood plain, and (III) coastal plain [Wu et al., 1996]; (2) A-A′ line of cross section in Figure 2; and (3) locations of meteorological stations and observation wells used in water level analysis for model calibration.

Download figure to PowerPoint

[10] The NCP is divided into four main hydrogeological zones from the Taihang Mountains in the west to the Bohai Sea in the east; these zones are based on the geomorphology of palaeochannels, in the sense that geomorphology has defined their sedimentary character and relative geographic positions, and include the piedmont plain, alluvial fan, alluvial plain (central plain), and coastal plain [Wu et al., 1996]. In turn, the sedimentary character and geographic position determine the hydrogeologic roles of the four zones: mountain and piedmont plain as regional groundwater recharge zone, parts of the alluvial plain and coastal plain as regional groundwater discharge zone, and alluvial fan and part of alluvial plain as an intermediate zone [Chen et al., 2004]. The alluvial fan and alluvial plain, which were formed through the depositional processes of the Yellow River, Hai River, Luan River and their tributaries [Wu et al., 1996; Xu et al., 1996] constitute the main part of the NCP. The topography is relatively flat with an average elevation of ∼30 m above sea level (a.s.l.) and slopes from ∼100 m a.s.l., at the base of the Taihang Mountains in the west, to sea level at the coast (Figure 1). The piedmont plain is the primary area of grain production in the NCP, where the percentage of wheat cultivation is much higher than the mean value across the NCP and most of the irrigation is from groundwater [Zhang et al., 2010]. According to data on crop water deficits from 1990 to 2000, groundwater consumed by the wheat and maize is ∼5000 m3/ha in the piedmont plain and ∼3000 m3/ha in the central plain [Xu et al., 2005].

[11] The climate in the NCP is continental semiarid, with mean annual temperature of 12°C –13°C [Z. Y. Chen et al., 2003]. Mean annual precipitation is ∼560 mm (1951–2008), and decreased from ∼600 mm in the 1950s to ∼500 mm in 2000s based on precipitation at 23 meteorological stations from the China Meteorological Administration (http://cdc.cma.gov.cn, accessed January 2011). Precipitation, influenced by the East Asia monsoon, decreases from southeast to northwest, and ∼70–80% of precipitation occurs during the monsoon season from June to September [Chen, 1999]. Mean annual pan evaporation ranges from 1100 to 2000 mm [Chen et al., 2005].

[12] Winter wheat and summer maize are the main crops in the NCP rotation system, accounting for ∼80% of total cultivated area and ∼90% of grain yield in the NCP, based on statistical data from 2009 [National Bureau of Statistics of China, 2010]. Winter wheat is sown at the beginning of October and harvested in mid-June the following year, and summer maize is planted immediately after wheat harvest and harvested at the end of September the same year. Long-term lysimeter measurements show that annual water consumption of the two crops without a water deficit exceeds precipitation by ∼300 mm in the piedmont area [Liu et al., 2002]. As a result, supplemental irrigation is required for the two crops. The monsoon season coincides with the growing season of maize, whereas only 20%–30% of annual precipitation falls during the wheat growing season [Liu et al., 2001]. Consequently, most irrigation is applied to winter wheat, which is grown during the dry season. In recent years, groundwater has provided ∼70% of the total water supply to support grain production and ∼50% of the total urban water supply [Han, 1998]; 70%–80% of groundwater exploitation is used for irrigation [Zhang et al., 2009b, 2010].

[13] The Quaternary aquifer of the NCP is traditionally divided into two aquifer zones referred as “shallow” and “deep” [Fei, 1988]. Extensive groundwater exploitation from the two aquifer zones has greatly impacted the hydrodynamic and hydrogeochemical system and caused many environmental problems, such as groundwater depletion, groundwater contamination and salinization, and land subsidence [Liu et al., 2001; Zheng et al., 2010]. Therefore, continuing groundwater depletion in both shallow and deep aquifer zones is a common concern.

2.2. Hydrogeologic Data

[14] Data availability in the study area was an important consideration in model development. The most direct approach to monitor groundwater storage change is to use groundwater level monitoring data, multiplying the water level change by specific yield in the unconfined aquifer area [Konikow and Kendy, 2005]. Historical groundwater levels were collected by the China Institute of Geo-Environmental Monitoring (CIGEM). The data set includes 230 monthly water level monitoring time series from 1993 to 2008 for 178 wells (including 38 multilevel wells) belonging to the national groundwater level monitoring network (http://www.cigem.gov.cn, accessed January 2011). The distribution of monitoring wells is fairly sparse, particularly in some parts of the central plain and coastal plain, because shallow groundwater is brackish and nonpotable in these areas.

[15] In addition to the groundwater level monitoring data set, historical water level contour maps are available from the China Geological Survey (CGS): maps of the shallow aquifer zone are available for 1959, 1984, 2001, and 2002 [Zhang et al., 2009a]; maps of the deep aquifer zone are available for 1980 and from 2001 to 2004 [Zhang et al., 2009a]; maps of each aquifer zone in the Hebei Plain are available for 1959, 1975, 1985, and 1992 [Zhang et al., 2000]; and a map of the deep aquifer zone in the Hebei Plain is available for 1969 (http://water.cgs.gov.cn, accessed January 2011). The groundwater level monitoring data set and these historical groundwater level contour maps were used to estimate groundwater storage change and to calibrate the flow model. These best available data at measurement frequencies sufficient for estimation of monthly and seasonal changes in groundwater storage, however, were rarely used in previous studies to calibrate flow models or calculate water storage changes.

[16] Hydraulic conductivity, specific yield, and storage coefficient are the principal hydraulic parameters required for the groundwater model of the NCP. Studies to estimate the hydraulic parameters across the NCP began in the 1970s, and most of the parameters were estimated in the 1980s [Chen, 1990]. The hydraulic conductivity and storage coefficients were estimated using pumping test data. No detailed information from these pumping tests was collected for this study, but values of these hydraulic parameters were summarized for the different lithologies across the NCP by Chen [1999] and Zhang et al. [2009a]. Horizontal hydraulic conductivities, specific yields, and storage coefficients were mapped across the entire NCP by the CGS [Zhang et al., 2009a]. Horizontal hydraulic conductivity of gravel generally ranges from 150 to 700 m/d in the shallow aquifers and 50–100 m/d in the deep aquifers. Horizontal hydraulic conductivity of sand generally ranges from 5 to 200 m/d in the shallow aquifers and 5–60 m/d in the deep aquifers. Specific yield ranges from ∼3% for clay to ∼30% for gravel. Storage coefficients in the deep aquifers decrease from 0.001 to 0.008 in the alluvial plain to 0.0004–0.0005 in the coastal plain.

3. Groundwater Storage Change From Water Level Data

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Study Site and Hydrogeologic Data
  5. 3. Groundwater Storage Change From Water Level Data
  6. 4. Simulation of the Groundwater Flow System
  7. 5. Results and Discussion
  8. 6. Approaches Toward More Sustainable Groundwater Management
  9. 7. Model Limitations and Future Studies
  10. 8. Conclusions
  11. Acknowledgments
  12. References
  13. Supporting Information

[17] The groundwater level contour map for 1959 was used to calculate the initial groundwater in storage in the NCP. Groundwater storage volumes were calculated from drainable volumes in the shallow aquifers (aquifers I and II) and recoverable compressive storage in the deep confined aquifers (aquifers III and IV). Drainable volumes representing conditions when the confined aquifers became unconfined were also calculated. The bottom elevation of the Quaternary aquifer across the NCP is not clearly defined because most boreholes do not penetrate the entire section of Quaternary strata [Zhang et al., 2009a]. The saturated thickness used to estimate the initial storage was calculated by subtracting a contour map interpolated from the bottom elevation of boreholes from the water level contour map; therefore, initial water storage is a lower bound of actual thickness. All historical water level contour maps were digitized and then converted to a grid using 2 × 2 km cells, which is the same grid scale as was used in the flow model.

[18] Groundwater level changes in the shallow aquifer zone were converted to storage changes by multiplying by the specific yield across the NCP. Based on data from 22 pumping tests, the mean storage coefficient of the deep aquifer is 0.00125, which is approximately 2% of the area-weighted specific yield of the shallow aquifer (0.075) [Chen, 1999]. The maximum water level declines in the primary cones of depression are 40–50 m in the shallow aquifer zone and 60–80 m in the deep aquifer zone, respectively [Fei et al., 2009]. Therefore, the amount of groundwater released from storage in the deep aquifers is much less than that in the shallow aquifer, and is ignored in the storage change calculation. Cumulative groundwater storage changes were then estimated using historical water level contour maps and the more recent monitoring time series. Land subsidence accompanying substantial pumpage from the deep aquifer zone indicates that large amounts of water released from storage are derived from compaction of low-permeability clay layers (“inelastic storage”), and this storage cannot be recovered. However, lack of land subsidence monitoring data, particularly seasonal monitoring of compaction, constrains calculation of inelastic storage changes in the deep aquifer zone. Groundwater storage calculated using specific yield and elastic storage coefficient, therefore, refers to recoverable storage, rather than total storage (product of saturated volume and porosity).

[19] The primary challenge for estimating groundwater storage changes using water level changes directly in the NCP is accounting for significant gaps in groundwater level monitoring time series. Estimated storage changes would be highly questionable if these gaps were retained in the calculation process. Preliminary tests showed that different gap-filling regression methods resulted in large variations in estimated groundwater depletion. In this study, instead of using a regression method to fill the gaps, two groundwater level contour maps were first generated by inverse distance weighting (IDW) interpolation using monitoring locations where both monitoring data were available for the two data sets, then the groundwater level change was calculated. A total of 117 monitoring wells in the shallow aquifer (Figure 1) were used to calculate groundwater storage changes. It is difficult to verify the reliability of the storage change estimates because the water level decline in the large cones of depression around cities may be underestimated through data interpolation due to the sparse distribution of observation wells; however, some monitoring wells may be affected by pumpage and may result in overestimation of depletion.

4. Simulation of the Groundwater Flow System

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Study Site and Hydrogeologic Data
  5. 3. Groundwater Storage Change From Water Level Data
  6. 4. Simulation of the Groundwater Flow System
  7. 5. Results and Discussion
  8. 6. Approaches Toward More Sustainable Groundwater Management
  9. 7. Model Limitations and Future Studies
  10. 8. Conclusions
  11. Acknowledgments
  12. References
  13. Supporting Information

4.1. Conceptual Model of the Groundwater Flow System

[20] In the first regional hydrogeological survey conducted in the1970s by the Hebei Bureau of Geology and Minerals, the Quaternary aquifers in the NCP were divided into four major aquifer units (Figure 2, I–IV). The geologic units corresponding to these aquifers are Holocene series (Q4) and late (Q3), middle (Q2), and early (Q1) Pleistocene series [Foster et al., 2004; Zhang et al., 2009a]. Aquifer units I and II in the early work correspond to the shallow aquifer zone in this study, and aquifer units III and IV correspond to the deep aquifer zone. A total of 38 multilevel monitoring wells in the CIGEM monitoring nests show a pervasive downward hydraulic gradient and a generally increasing trend in the magnitude of that gradient over the monitoring period. The magnitude of the vertical hydraulic gradient ranges from a minimum of 0.04 in areas with little pumping to a maximum of 0.3 within cones of depression. No detailed information on depths of pumping wells is available for the NCP.

image

Figure 2. Hydrogeological cross section of the NCP along A-A′ in Figure 1 (modified from Chen et al. [2005]). Aquifer units I and II correspond to the shallow aquifers in this study, and deep aquifers refer to aquifer units III and IV.

Download figure to PowerPoint

[21] Precipitation is the dominant source of natural groundwater recharge in the NCP; other sources of natural recharge include lateral flow from the Taihang and Yan mountains and leakage from the lower reach of the Yellow River [Zhang et al., 2008]. Irrigation return flow from areas irrigated with surface water from the reservoirs to the west toward the base of mountains and the Yellow River is another recharge source to groundwater in the NCP. Mountain front lateral flow is a significant source of recharge, but is not well understood or quantified [Kendy et al., 2004]. Recharge rates from lateral flow, used in previous regional water resource evaluations, ranged from 1.0 to 2.0 km3/yr [Zhu et al., Assessment of groundwater source in North China Plain, unpublished report, 1990; Zhang et al., 2009a], and were assumed to be restricted to the shallow aquifer [Liu and Wei, 1989]. Some researchers measured percolation rates at selected sites along the Taihang Mountain front [J. Y. Chen et al., 2003; Li et al., 2007; Song et al., 2007]. A limitation of the approach is that the total surface area undergoing recharge must be estimated to calculate regional recharge rates. The amount of lateral flow from the mountain area depends on the horizontal permeability, thickness of the saturated zone, and local hydraulic gradient. Estimated mountain front recharge using Darcy's law based on water level contour maps in this study is 2.8, 2.1, and 3.1 km3 in 1959, 1984, and 2001, respectively.

[22] The lower reach of the Yellow River, flowing from the Taohuayu hydrological station to the Bohai Sea, serves as the southern boundary of the NCP defined in this study. In this section, accumulated sediment deposits have raised the riverbed several meters above the ground, forming an “above-ground river.” Leakage from this reach of the Yellow River to the NCP was estimated to have decreased from ∼0.6 km3/yr in the 1980s to ∼0.4 km3/yr in the 1990s [Jiao and Duan, 2005; Ma and Duan, 2005] because of the severe flow cutoff of the lower reach in the 1990s [Cai and Rosegrant, 2004; Liu and Xia, 2004]. Analyses of groundwater dynamics and isotope data indicate that the area affected by seepage from the Yellow River is limited to 10–20 km wide [Zhao et al., 2003; Shao et al., 2003]. More than 1000 reservoirs have been constructed in the NCP since the 1950s; these reservoirs in the piedmont region control more than 90% of the upstream area. Recharge from reservoir leakage was combined into the total recharge determined in the model calibration process, and was not calculated separately in this study. Under predevelopment conditions, groundwater discharged to rivers; therefore, rivers were gaining. Rivers only became losing after significant groundwater depletion since the 1980s. However, by this time, most of the surface water was dammed. Due to over-development of surface water and reductions in precipitation, surface water flow in the NCP has decreased significantly. At present, more than 40% of the river channels have dried up or become ephemeral streams [Xia, 2006]. Mean annual discharge to rivers from 1980 to 2000 is ∼0.032 km3 [Ren, 2007], which is much less than pumpage, and leakage to groundwater is also limited.

[23] Groundwater is the primary source of water for irrigation in this area [Zhang et al., 2003]. The area irrigated by groundwater accounts for ∼70% of the total irrigation area [Liu et al., 2004]. Water for the remaining 30% irrigation area is derived from canals from upstream reservoirs to the west and the Yellow River to the south. Irrigation derived from surface water decreased with time because of rapidly increasing water demand for large cities. For example, the irrigated area with diverted water from upstream reservoirs has been reduced by 50% by 2000. Irrigation return flow is significant because flood and furrow irrigation are still the primary techniques for both surface water-fed and groundwater-fed irrigation [Jiang et al., 2009]. Detailed information about irrigated areas and applied amounts is not available. In this study, natural recharge from precipitation and recharge from irrigation return flow were estimated as total recharge through model calibration. Therefore, the results should reflect recharge from the sum of precipitation and total irrigation, whether the irrigation is originally derived from surface water or groundwater. Natural discharge from the NCP aquifer occurs to streams as baseflow, ET from areas of shallow water tables, and to the Bohai Sea from the shallow aquifer. Runoff data are unavailable for streams during predevelopment. In the steady-state flow model constructed for the predevelopment by Zhang et al. [2008], the riverpackage of MODFLOW was used to estimate groundwater discharge to streams by mean monthly stream stages in 1965. Natural discharge to streams during predevelopment was estimated to be ∼9.0 km3/yr by this method; this estimate provided water budget information to calibrate the steady-state model in this study. Little is known about discharge to the Bohai Sea, but the amount is thought to be small because hydraulic conductivity is relatively low for the coastal aquifer and because the horizontal hydraulic gradient in the coastal plain is quite small.

[24] Artificial discharge from the aquifers in the NCP occurs through pumping by numerous wells. The 1970s was the time of most rapid agricultural development in the NCP, and corresponds to rapidly increasing groundwater pumpage. In the 1980s and 1990s, groundwater pumping increased further, but at a slower rate, as agricultural and industrial development continued. Since 2000, problems caused by excessive exploitation of groundwater have attracted much attention on environmental problems, and agricultural water saving measures have been promoted; as a result, groundwater pumping has shown a slight decreasing trend [Zhang et al., 2009b].

4.2. Flow Model Construction

[25] A multilayer, heterogeneous and anisotropic model was built to simulate the flow system in the NCP using MODFLOW-2000 [Harbaugh et al., 2000]. The finite difference grid consisted of 325 rows and 300 columns, with a regular cell size of 2 × 2 km. In the three-dimensional reconstruction of the aquifer system, the model grid was discretized into three layers (Table 1). Aquifer units I and II were combined and represented by the first layer of the grid. The thickness of layer 1 (aquifer units I and II) ranges from ∼40 to ∼250 m, with a mean of ∼160 m. Layer 2 (aquifer unit III) is ∼50–250 m in thickness, with a mean thickness of ∼130 m. The thickness of layer 3 (aquifer unit IV) ranges from ∼30 to ∼300 m, with a mean of ∼110 m. Layer 1 represents the shallow aquifer system, and layers 2 and 3 represent the deep aquifer system. Steady-state predevelopment flow was first simulated and followed by transient postdevelopment flow from the 1960s to 2008. The transient model was divided into annual stress periods. Pumpage and other source/sink terms were considered to be constant in each stress period.

Table 1. Summary of Hydrologic Parameters Used in the Flow Model
Model LayerAquifer UnitThickness/Hydrologic Parameters Min-Max (Geometric Mean)
Thickness (m)Hydraulic Conductivity (m/d)Storage Parameters
  1. a

    No data are available for hydraulic parameters in aquifer IV; therefore, parameter values for aquifer III were used for aquifer IV.

1Aquifers I AND II40–250 (160)3–200 (16.7)0.04–0.25 (0.075)
2Aquifer III50–250 (130)5–60 (9.6)0.0001–0.005 (0.0012)
3aAquifer IV30–300 (110)5–60 (9.6)0.0001–0.005 (0.0012)

[26] Lateral boundary conditions for layer 1 included specified head along the Bohai Sea and specified flow along the mountains to the west and along the Yellow River to the south. Lateral boundaries for layers 2 and 3 were assumed to be no flow. Recharge from precipitation and irrigation return flow (including irrigation from surface water and groundwater, leakage from water diversion canals) were combined to represent total areal recharge. Pumpage was simulated by one “virtual” pumping well in each active model cell, which accounted for all the simulated withdrawal in that cell. The amount of pumpage for different model layers was initially allocated according to the reported groundwater pumping data in the shallow and deep aquifer systems [Zhang et al., 2009b], and were adjusted during model calibration. Because river recharge represents a small fraction of total input/output, rivers were not simulated explicitly in this study, and discharge rates were combined with the pumpage. ET was simulated with a uniform extinction depth of 4 m across the NCP (see the Supporting Information).

4.3. Model Calibration

[27] The calibration process was divided into three parts: calibration of the steady-state model using groundwater level contour maps for 1959; calibration from 1960 to 1992 using historic water level contour maps, and detailed calibration from 1993 to 2008 of the transient model. Contour maps for 1959 may not accurately reflect predevelopment conditions because irrigation became important in the NCP in the 1950s after the foundation of the People's Republic of China [Kendy et al., 2003]. However, crop quotas limited grain production in this period and crop production was mostly restricted to low-yielding cotton [Kendy et al., 2003]. Groundwater pumping was relatively low in the 1950s (∼2.0 km3/yr) compared to current groundwater withdrawals (∼20 km3/yr), and groundwater was primarily discharged through ET and base flow to rivers. Agriculture was rainfed dominated and most irrigation was derived from river water diversions and groundwater carried from shallow stone/brick wells. Irrigation should have little effect on natural recharge. Therefore, the water level distribution in 1959 (the earliest measured water levels available across the NCP) may represent the average flow dynamics under natural conditions, and could be used to calibrate the steady state model. The trial-and-error method and inverse modeling using the automated Parameter ESTimation (PEST) code [Doherty, 2003] were implemented in the calibration process of the steady-state model. The recharge and pumpage distributions were first adjusted to roughly match the overall water level distribution patterns, and then the PEST code was used to obtain the optimal values for recharge rates. Using classical groundwater model calibration or inverse modeling to estimate recharge from information on water levels and hydraulic conductivity only provides information on the ratio of recharge to hydraulic conductivity, and the reliability of the recharge estimates depends on how well we know the hydraulic conductivity [Scanlon et al., 2002]. In this study, the hydraulic parameters were assumed to be representative and were not adjusted during model calibration, and spatial and temporal distributions of recharge rates were the primary calibration parameters. Recharge rates and their temporal distribution were adjusted in the model calibration to optimize the fit between observed and simulated water levels. The ratio between vertical and horizontal hydraulic conductivity (Kh/Kv) was also adjusted during calibration, and a uniform value of 10,000 was used. This Kh/Kv ratio is considered representative given the average Kh of 3–200 m/d in model parameter zonations with previously reported Kv of 0.00003–0.0003 m/d for clay, and 0.005–0.007 m/d for the sand in aquitards [Chen, 1990] in the NCP.

[28] Using the water level distributions calculated by the steady-state flow model as the initial condition for the transient flow model, the transient model was first roughly calibrated to 1992 data only using water level contour maps in 1975, 1984, and 1992 [Zhang et al., 2000; Zhang et al., 2009a], and then calibrated from 1993 to 2008 using monitoring water level time series. In this two-part calibration process, total areal recharge in each county administrative region was the primary calibration parameter, and the spatial distribution of pumping was also adjusted through the trial-and-error calibration process. Because the majority of wells were monitored manually, the water level monitoring time series data were prefiltered to remove obvious erroneous data prior to being used for model calibration. Due to lack of detailed information on distribution of irrigated fields or an irrigation density map, the province-level pumpage was first redistributed by county or city by comparing simulated with measured water level changes, and then was divided equally into each active cell in the county or city.

[29] The water level contour map shows that the simulated water level contours are similar to the regional flow pattern of hand-contoured measured water levels for 1959 and 1984 (Figure 3). The goodness of fit between simulated and measured water levels in the monitoring wells from 1991 to 2008 is presented in Figure 4. The mean error (ME) between simulated and measured water levels is −3.76 m and the root–mean-square error (RMSE) is 7.29 m for the 105 shallow observation wells (in model layer 1). The ME and RMSE are −2.38 m and 15.12 m, respectively, for 72 deep observation wells (in model layers 2 and 3), totaling 871 observations. These errors are small relative to the maximum groundwater level variations of ∼60 and ∼70 m in the shallow and deep aquifer zones, respectively; therefore, the model calibration is considered reasonably reliable.

image

Figure 3. Comparison of observed and simulated water level contours (in meter above sea level, with an interval of 10 m) in the shallow aquifer zone in (a) 1959 and (b) 1984. Solid lines are simulated water level contours, and dotted lines are observed water level contours.

Download figure to PowerPoint

image

Figure 4. Comparison between observed and simulated annual water levels from 1993 to 2008 at (a) 105 shallow observation wells (in model layer 1), total 1231 observations, and (b) 72 deep observation wells (in model layers 2 and 3), total 871 observations. Dotted lines indicate the 95% prediction intervals.

Download figure to PowerPoint

5. Results and Discussion

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Study Site and Hydrogeologic Data
  5. 3. Groundwater Storage Change From Water Level Data
  6. 4. Simulation of the Groundwater Flow System
  7. 5. Results and Discussion
  8. 6. Approaches Toward More Sustainable Groundwater Management
  9. 7. Model Limitations and Future Studies
  10. 8. Conclusions
  11. Acknowledgments
  12. References
  13. Supporting Information

5.1. Groundwater Flow Dynamics

[30] The simulated predevelopment water table elevation (a.s.l) decreases gradually from 90 m in the piedmont area of the Taihang Mountain to ∼5 m in the coastal plain (Figure 3a). Hydraulic gradients are up to 30% in the piedmont area of the Taihang and Yan Mountains and relatively flat in the central and eastern parts of the plain, resulting in limited discharge to the sea. Groundwater is discharged mainly through ET and base flow to rivers. In cross-sectional view, groundwater flows from the mountain area into the plain and then discharges to the Bohai Sea. Vertically, the deep aquifer zone receives leakage from the shallow aquifer zone in the piedmont area and then discharges to the shallow aquifer in the central and coastal plains. Estimated predevelopment total groundwater recharge is ∼12 km3/yr (corresponding to ∼90 mm/yr), and varies from 1.8 km3/yr (∼80 mm/yr) in the piedmont region to 7.8 km3/yr (∼80 mm/yr) in the central plain, and 2.1 km3 (∼110 mm/yr) in the coastal plain. Total simulated recharge in this study is close to the simulated recharge in B. G. Wang et al. [2008]. Simulated maximum recharge rates per unit area are ∼250 mm/yr in the piedmont region, ∼200 mm/yr in the central plain, and ∼150 mm/yr in the coastal plain, reflecting higher infiltration in coarser soils in the piedmont region.

[31] In 1959 and the 1960s, the water table depth was 0–3 m below the land surface in most places. Increasing groundwater withdrawals since the 1970s caused continuing declines in groundwater levels (Figure 5). Simulated annual water table depths show a nearly linear declining trend (R = 0.97), with a slope of 0.3 m/yr. Most of the cones of depression in the shallow aquifer zone are beneath cities in the piedmont region east of the Taihang and Yan Mountains. Cones of depression in the deep aquifer beneath the eastern part of the plain merged, forming an extensive, regional cone of depression. Model results show that by 2008, the area with water table depths ≥10 m below the land surface is ∼60,000 km2, or ∼45% of the entire plain; and the area with the potentiometric surface elevation of the deep aquifers below sea level is ∼90,000 km2, covering ∼70% of the entire plain. Similar results were obtained using more monitoring data in recent field investigations [Fei et al., 2009].

image

Figure 5. Total groundwater pumpage and simulated annual average water table depth in the NCP. Annual groundwater pumpage before 1980 was estimated.

Download figure to PowerPoint

5.2. Groundwater Storage Change Affected by Pumpage

[32] Groundwater storage in 1959 was estimated to be 1500 km3 of drainable storage in the shallow aquifer zone and 40 km3 of compressive storage in the deep aquifer zone. Groundwater drainable storage is 220 km3 in the piedmont region, 1090 km3 in the central plain, and 190 km3 in the coastal plain. However, if the confined aquifers transitioned to unconfined conditions, total drainable storage in the aquifer system would be 3960 km3.

[33] Groundwater pumpage depleted storage, particularly in the piedmont region, because of much higher groundwater withdrawal density, i.e., pumping rate per unit area (Figure 6a). Uneven distribution of groundwater pumpage is mainly caused by considerable variations in the density of grain crop cultivation. According to the 2008 statistical data [Hebei Provincial Bureau of Statistics, 2009], grain crop cultivation in the NCP ranges from ∼15 to ∼110 ha/km2, and the majority of the counties with higher density than the mean value (70 ha/km2) are in the piedmont region. The area with water table declines ≥10 m in the shallow aquifers is 67,000 km2, covering about one half of the NCP (Figure 6b). Average water level declines in the shallow aquifers were 19 m in the piedmont region, 12 m in the central plain, and 3 m in the coastal plain. Average water levels in the deep aquifers decreased by 25 m in the piedmont region, 31 m in the central plain, and 29 m in the coastal plain. These water level declines correspond to groundwater depletion of 50 km3 in the piedmont region, 103 km3 in the central plain, and 5 km3 in the coastal plain. By the end of 2008, simulated cumulative recoverable groundwater storage depletion is approximately 160 km3, which is ∼10% of total initial recoverable groundwater storage.

image

Figure 6. Spatial distribution of (a) average annual groundwater withdrawal, i.e., the pumping rate per unit area and (b) simulated water level change in the shallow aquifers from 1970 to 2008. The thin gray lines in Figure 6a indicate boundaries of municipalities and counties.

Download figure to PowerPoint

[34] According to recent land subsidence monitoring across the NCP in 2005, the cumulative land subsidence volume was approximately 40.0 km3, which is equivalent to the initial compressive storage. The average annual land subsidence volume was 1.1 km3, assuming all land subsidence occurred after 1970 [Shi et al., 2006]. This indicates that an additional 1.1 km3/yr of storage depletion is derived from aquifer compaction (unrecoverable aquifer storage). The cumulative storage depletion accompanying land subsidence was 1.3 km3 in the piedmont region, 27.4 km3 in the central plain, and 10.5 km3 in the coastal plain.

[35] By the end of 2008, cumulative total groundwater storage depletion (sum of recoverable and unrecoverable depletion) is ∼200 km3, including 51.3 km3 in the piedmont region, 130.4 km3 in the central plain, and 15.5 km3 in the coastal plain. However, depletion rates varied across the NCP (Figure 7). Storage depletion per unit area, in equivalent saturated water thickness, was highest in the piedmont area with an average of 60 mm/yr, and decreased to 30 mm/yr in the central plain and 17 mm/yr in the coastal plain. Comparisons of storage depletion with groundwater pumpage distribution (Figure 6a) suggest that depletion in the NCP is generally driven by irrigation pumpage. Almost all depleted storage in the piedmont region is recoverable; however, much of the depletion in the central and coastal plains is irrecoverable.

image

Figure 7. Spatial distribution of average annual groundwater storage depletion from 1970 to 2008 by summing depletion simulated by the model and land subsidence data assuming a constant land subsidence rate.

Download figure to PowerPoint

[36] Seasonal groundwater storage based on monitoring groundwater level data from 1993 to 2008 is lowest in spring (March–May) corresponding to the grain-filling stage of winter wheat (April and May) in the NCP (Figure 8). Groundwater is replenished mostly during the summer/rainy season (July–September) (Figure 8), which corresponds to the maize cropping season. Generally, maize cultivation does not have much impact on groundwater depletion, except during drought years, such as 1997, 1999, and 2002, when average annual precipitation was less than 400 mm. From October to the beginning of irrigation of winter wheat in the following March, which is the cultivation and vernalization period of winter wheat, groundwater storage gradually increases. Water levels in the NCP are influenced by precipitation, well pumping, and irrigation return flow. Moreover, only 130 monitoring wells were selected for storage estimation, which accounts for approximately one well per 1000 km2. Therefore, aquifer storage could not be estimated accurately by water level change analysis, and the results in Figure 8 using water level change data only represent relative storage change and were not used to calculate cumulative storage change.

image

Figure 8. Monthly average groundwater storage changes in the NCP calculated using water level data (in terms of equivalent saturated water storage thickness averaged across the plain).

Download figure to PowerPoint

[37] Model results indicate that mean recoverable groundwater depletion was ∼4 km3/yr from 1970 to 2008. However, depletion rates varied with time (Table 2). Calculated mean annual recoverable storage depletion was ∼2.5 km3 in the 1970s, and increased to ∼4.0 km3 in the 1980s. Mean depletion decreased to ∼2.0 km3/yr in 1990–1996 corresponding to a period of increasing precipitation. Mean depletion was highest in 1997–2001 (∼7.0 km3/yr), corresponding to lowest precipitation. Depletion decreased to ∼4.0 km3/yr in 2002–2008 because of reduction in pumpage. Total storage depletion (recoverable plus irrecoverable storage depletion) was obvious except in 1973, 1977, and 1990 when annual precipitation exceeded 700 mm and in 1996 when the NCP was affected by a severe flood event (Figure 9). High negative correlation between storage depletion and precipitation within a given year (R = −0.92) and low correlation with total pumpage (R = 0.38) indicates that precipitation dominates temporal distribution of storage depletion at annual timescales.

image

Figure 9. (a) Model simulated annul groundwater recoverable storage depletion plus additional irrecoverable storage depletion accompanying land subsidence and precipitation and (b) simulated cumulative groundwater storage depletion with 1959 as the base year in different regions in the NCP, 1970–2008.

Download figure to PowerPoint

Table 2. Simulated Water Budget and Storage Depletion for the NCP Aquifer for Steady-State (Predevelopment) and Transient Simulations (Unit: km3/yr)
PeriodRechargeaLateral FlowbSpecified River LeakageETPumpingStorage Changec
  1. Positive values represent inputs and negative values outputs. Negative storage values represent depletion.

  2. a

    Simulated recharge represents areal recharge from precipitation and irrigation return flow.

  3. b

    Lateral flow represents specified mountain front recharge, and river leakage is derived from the Yellow River.

  4. c

    Storage change only represents depleted recoverable storage, excluding irrecoverable compressive storage changes.

Predevelopment11.601.57−8.00−5.170.000.00
1970–197920.531.86−5.00−5.96−13.93−2.50
1980–198916.331.90−0.67−2.94−18.71−4.09
1990–199617.631.882.10−3.13−20.59−2.11
1997–200114.421.921.38−2.22−22.83−7.33
2002–200818.711.910.30−2.58−22.71−4.36

5.3. Sources of Water for Groundwater Pumpage

[38] During the nearly 50-year simulation period, groundwater storage depletion represented ∼25% of total pumpage (Table 3). Lateral flow representing mountain front recharge and leakage from the Yellow River contributed ∼9% of total pumpage. ET capture only contributed ∼2% of total pumpage. The rest of the pumpage came from recharge from precipitation and irrigation return flow. However, contributions of different sources of pumpage varied. In the piedmont region, almost all storage depletion is recoverable, and the contribution of compressive storage and ET capture can be ignored. A larger amount of compressive storage depletion (∼5% of pumpage, ∼20% of total storage depletion) in the central plain is attributed to higher pumpage from the deep aquifer zone (∼30% of total pumpage in this region). In this region, lateral flow primarily from the piedmont region is a noticeable source of pumpage. In the coastal plain, contributions to pumpage from compressive storage depletion and lateral flow exceeded that from recoverable storage depletion, and ET capture is an important source of pumpage.

Table 3. Primary Sources of Groundwater Pumpage in the Shallow and Deep Aquifers in the Piedmont Region, Central Plain, Coastal Plain, and Across the Entire NCP During the Simulation Period of Nearly 50 Years (1960–2008; Units, km3)a
RegionShallow PumpageDeep PumpageRecoverable Storage DepletionCompressive Storage DepletionLateral FlowET Capture
  1. a

    Numbers in brackets are percentages of total pumpage.

Piedmont2441450 (19)1 (0.4)15 (6)3 (1)
Central365138103 (20)27 (5.4)51 (10)11 (2)
Coastal16255 (12)11 (27)10 (24)2 (5)
Total625178158 (20)39 (4.9)76 (9)15 (2)

[39] Development of the deep groundwater has resulted in increasing downward leakage from the shallow aquifer zone, particularly in the central and coastal plains. Calculated leakage from the shallow aquifer zone accounted for ∼70% of total pumpage from the deep aquifer zone, and increased to ∼90% in 2008. This increasing trend of downward leakage is consistent with water levels in the shallow aquifer zone becoming generally much higher than those in the deep aquifer zone. However, assuming all compressive storage depletion (not represented in the flow model) occurs in the deep aquifer zone, contributions from downward leakage would be overestimated by ∼20% (percentage of pumpage in the deep aquifer zone).

[40] The total contribution of storage depletion, lateral flow, and ET capture represents ∼30%, 40%, and 70% of pumpage in the piedmont region, central plain, and coastal plain, respectively, and indicates that recharge from precipitation and irrigation return flow was still the largest contributor to pumpage across the NCP. The calibrated mean annual spatially averaged groundwater recharge rate across the NCP from precipitation and irrigation is 120 mm. The calibrated mean groundwater recharge is consistent with previous point recharge estimates using environmental tracer methods and numerical model results [Kendy et al., 2004; B. G. Wang et al., 2008; Lu et al., 2011]. Overestimation of areal recharge from ignoring comprehensive storage depletion in the flow model is about 8 mm/yr. High correlation between model-calibrated annual groundwater recharge and annual precipitation plus irrigation (r2 = 0.86) implies that an average of ∼20% of the total precipitation and irrigation (assuming 70% of groundwater pumpage is for irrigation) in a year reaches the aquifer as recharge.

[41] There is considerable spatial variability in calibrated recharge over the NCP (Figure 10). Calibrated recharge is highest (200–350 mm/yr) in the piedmont region, and decreases to 50–100 mm in the central and coastal plain areas. Disagreement between overall patterns of groundwater recharge and spatial distribution of mean annual precipitation (increasing from northwest to southeast) indicates that surface boundary conditions, such as soil properties, may also be a controlling factor on spatial distribution of groundwater recharge.

image

Figure 10. Spatial distribution of calibrated average annual recharge and precipitation contours interpolated using 23 meteorological stations.

Download figure to PowerPoint

6. Approaches Toward More Sustainable Groundwater Management

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Study Site and Hydrogeologic Data
  5. 3. Groundwater Storage Change From Water Level Data
  6. 4. Simulation of the Groundwater Flow System
  7. 5. Results and Discussion
  8. 6. Approaches Toward More Sustainable Groundwater Management
  9. 7. Model Limitations and Future Studies
  10. 8. Conclusions
  11. Acknowledgments
  12. References
  13. Supporting Information

[42] A sustainable groundwater management alternative will supply water of sufficient quantity and quality, and will not cause continuous groundwater depletion, particularly in the deep aquifers. In practice, sustainability need not require fully recovery of groundwater levels; a management alternative that provides a supply of groundwater to achieve a new equilibrium over a relatively short time period may be considered sufficiently sustainable. In such a groundwater pumping configuration, most groundwater pumpage should come from recharge to minimize groundwater depletion. It is clear that the current level of groundwater development in the NCP is unsustainable, and numerous studies have been conducted to explore effective and feasible approaches to mitigate groundwater depletion. Some of these studies are reviewed by Liu et al. [2008] and Zheng et al. [2010]. Considering that groundwater depletion is mainly caused by winter wheat production, the most effective means of reducing irrigation demand is by reducing irrigated acres of winter wheat [Foster et al., 2004; Kendy et al., 2007]. However, more than 70% of the food supply for local farmers is currently from wheat [Yang et al., 2002]. Importing food from outside is impractical. Moreover, considering that the wheat is self-sufficient while the maize supply has relied in part on import, substantial reduction of wheat acreage would conflict with the Chinese government's emphasis on food security and agricultural policies [Foster et al., 2004]; moreover, a simple change in agricultural crop rotation would be difficult. The most reasonable strategies involve augmentation and conservation of water resources. The simulated average annual recharge (17 km3) was used as the future recharge base input, and the estimated groundwater demand [Cui et al., 2009; Zhang et al., 2009b] was used as the groundwater pumpage input. The calibrated model was run from 2015 (the year after the completion of the first phase of the middle route of the SNWT project) to 2030 to evaluate the effects of the following potential approaches to reduce groundwater depletion.

6.1. Managed Aquifer Recharge

[43] Historically, a common response to floods and droughts in the NCP was to construct reservoirs. Reservoir construction in the Hebei Province began in the 1950s. By the end of 1988, there was a total of ∼1200 reservoirs in Hebei, including 17 large reservoirs with single storage capacity ≥0.1 km3 [Hebei Department of Water Resources, 1995]. The majority of the large reservoirs are located in the mountain front areas of the Taihang and Yan Mountains, and control more than 90% of the upstream area. Recharge from the rivers has become seasonal. The central plain has no suitable locations for dams and reservoirs. Depleted aquifers can be transformed into underground “reservoirs” by artificially recharging excess runoff when available. Urban water reuse and excessive runoff diversion from these mountain front reservoirs may provide a potential approach to conduct managed aquifer recharge (MAR) [Xu et al., 2009; Currell et al., 2012]. Increased recharge in the seven regions identified by Xu et al. [2009] for MAR was estimated to be ∼1.6 km3/yr. Average annual groundwater pumpage in recent years was approximately 22 km3, and the flow model results indicate that average annual groundwater recharge is ∼17 km3, and simulated ET in 2008 is 3.0 km3. Considering an additional ∼1.5 km3/yr of lateral flow from mountain front recharge and leakage from the Yellow River, it is possible for the groundwater system to reach a new equilibrium. The time required for the flow model to reach approximately steady-state (storage change approaching zero) is ∼300 years. By 2030, storage depletion still contributes ∼18% of pumpage.

6.2. Utilization of Brackish Water

[44] In the NCP, brackish water (total dissolved solids >2 g/L) occurs in the shallow aquifer zone in most of the central and coastal plains, where fresh water is pumped from the deep aquifer zone. Utilization of brackish water provides an alternative to reduce deep groundwater pumpage in these regions. Pumping shallow brackish water to replace deep fresh groundwater for irrigation will mitigate overexploitation of the deep aquifer zone and slow down the rate of the downward migration of the brackish/fresh water interface [Duan and Xiao, 2003; Foster et al., 2004]. Simulated shallow water table (≤10 m deep) in 2008 covers an area of 58,000 km2, mostly in the brackish water region. In the coastal plain, simulated average annual evaporation from the shallow groundwater was ∼1.5 km3/yr and recharge was 2.0 km3/yr. However, groundwater pumpage (primarily from deep aquifers) was ∼1.0 km3/yr. Reducing deep groundwater pumpage by utilizing shallow brackish water and capturing evaporation makes it possible for the groundwater system in this region to adjust to a new equilibrium state. By 2030, storage depletion still contributes ∼24% of pumpage with pumping brackish water at 0.6 km3/yr.

6.3. Improvement of Water Use Efficiency

[45] Agriculture is the largest water user in the NCP. Therefore, increasing the efficiency of agricultural water use is a major focus of water conservation efforts. Although the water use efficiency (WUE, defined by crop yield/ET in this study) for wheat and maize on well-managed experimental sites can reach 1.0–1.9 kg/m3, which is comparable with global average values, the WUE in farmers' fields is much lower than on experimental sites [Fang et al., 2010]. Intuitively, the effective way to improve irrigation efficiency is to replace flood irrigation with water-saving irrigation techniques, such as low pressure pipe irrigation and sprinkler irrigation [Blanke et al., 2007]. However, Kendy et al. [2007] brought up the counterintuitive point that these “water saving” technologies do not benefit the groundwater balance for the NCP because they are likely to increase ET and decrease groundwater recharge. Therefore, technologies that can reduce ET are critical for increasing WUE in the NCP. Duan and Xiao [2003] estimated that groundwater pumpage would be reduced by ∼4.0 km3/yr if the irrigation quota was decreased from ∼4,000 to 2,500 m3/ha. This pumpage reduction approaches the annual groundwater depletion. However, by 2030, storage depletion still contributes ∼16% of pumpage.

6.4. Implementation of the SNWT Project

[46] The previous approaches focused on reducing water demand to enhance sustainability; however, another approach would be to increase water supply. Importing surface water from river basins outside of the NCP would reduce stress on groundwater. The Chinese government is heavily relying on the SNWT) project designed to transfer water from the water-rich south, mainly the Yangtze River, to the drought-prone north. It consists of three routes: the eastern, the central, and the western, with the central and eastern routes impacting the NCP [Liu and Zheng, 2002]. Construction on the central route began in 2003, and ∼0.3 km3/yr of water has been transferred locally from reservoirs in the Hebei Province to Beijing since 2008. The first phase of the SNWT was scheduled to be completed in 2010, but environmental concerns and resettlement issues have pushed the completion date to 2014. The total volume of water projected to be transferred into the NCP is 7.0–7.5 km3/yr, including ∼1.2 km3/yr into Beijing, ∼1.0 km3/yr into Tianjian, and ∼3.0 km3/yr into Hebei [Duan and Xiao, 2003; Cui et al., 2009].

[47] A reduction of 6.0 km3/yr of groundwater pumpage was simulated to assess the impacts of the SNWT project, which is approximately 2.5 km3/yr greater than the current deficit between groundwater pumpage and recharge. Local groundwater levels in the piedmont region, particularly in Beijing, recover significantly and groundwater level declines in the central and coastal plain are reduced (Figure 11a). However, model results did not indicate recovery of groundwater storage by the end of 2030 because of continuous water level declines in the central plain and ET increases in the piedmont region. Simulated storage depletion would still contribute ∼3% of pumpage by 2030. The model was run for ∼150 years to approach a new equilibrium state. Therefore, the SNWT project could provide a critical water supplement in the piedmont region and some other local areas in the plains, but will not result in full recovery of groundwater storage throughout the NCP over a short time period.

image

Figure 11. Simulated water level change by 2030 after (a) implementation of the SNWT project and (b) the integrated effects of MAR, utilization of brackish water, improvement of WUE, and the SNWT project during the simulation period of 2015–2030. Negative values indicate water level recovery.

Download figure to PowerPoint

6.5. Combination of Potential Management Strategies

[48] To mitigate water shortage over the entire NCP, combinations of previous strategies may be more effective. The comprehensive impacts of previously mentioned potential groundwater management strategies were evaluated using the calibrated flow model which was run to 2030. Simulated average annual recharge (17 km3) was used as the future recharge base input and was derived from estimates of recharge increases through the MAR in the piedmont region (1.6 km3/yr). The offset of deep groundwater pumpage through shallow brackish water pumpage in the coastal plain was assumed to be 0.6 km3/yr. Groundwater recharge is ∼7 km3/yr higher than the predicted total pumpage. Model results indicated a more rapid groundwater level recovery rate throughout the NCP (Figure 11b). By the end of 2030, water level recovery could cause ET to increase by ∼3 km3/yr, and the groundwater storage was increased by ∼50 km3. These results indicate that an additional 50 yr would be required for groundwater to recover.

7. Model Limitations and Future Studies

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Study Site and Hydrogeologic Data
  5. 3. Groundwater Storage Change From Water Level Data
  6. 4. Simulation of the Groundwater Flow System
  7. 5. Results and Discussion
  8. 6. Approaches Toward More Sustainable Groundwater Management
  9. 7. Model Limitations and Future Studies
  10. 8. Conclusions
  11. Acknowledgments
  12. References
  13. Supporting Information

[49] There are a few potential discrepancies between model results and historical hydrologic conditions, including discrepancies between calculated and historical groundwater recharges, and between simulated water levels and mapped water levels, and distributions of cones of depression. The simple input method used to represent historical groundwater and surface water interactions with data from the literature, implementation of hydraulic properties adapted from previous studies without further verification, and coarse resolution groundwater pumping data can explain these discrepancies. Future studies will simulate interactions between surface water and groundwater using available river gage data and will calibrate hydraulic properties further using water levels and groundwater isotopic age data. With respect to storage changes, temporal variations in compressive storage change were not simulated by the flow code; however, future simulations will include inelastic water storage changes accompanying land subsidence.

8. Conclusions

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Study Site and Hydrogeologic Data
  5. 3. Groundwater Storage Change From Water Level Data
  6. 4. Simulation of the Groundwater Flow System
  7. 5. Results and Discussion
  8. 6. Approaches Toward More Sustainable Groundwater Management
  9. 7. Model Limitations and Future Studies
  10. 8. Conclusions
  11. Acknowledgments
  12. References
  13. Supporting Information

[50] Groundwater level monitoring data and regional groundwater flow modeling in this study provide valuable information about groundwater depletion in the NCP. The developed model is a useful tool to evaluate potential management strategies for more sustainable groundwater development in this region. Estimated groundwater storage during the predevelopment period includes 1500 km3 of drainable storage in shallow aquifers and 40 km3 of compressive storage in deep aquifers. Simulated recoverable groundwater storage depletion averaged ∼4 km3/yr (∼30 mm/yr) from 1970 to 2008. Total recoverable groundwater depletion varied spatially from 50 km3 in the piedmont region to 103 km3 in the central plain and 5 km3 in the coastal plain, corresponding to approximately 20%, 20%, and 12% of pumpage in their respective regions. Depletion also varied with time: ∼2.5 km3/yr in the 1970s, ∼4.0 km3 in the 1980s, ∼2.0 km3/yr in 1990–1996, ∼7.0 km3/yr in 1997–2001, and ∼4.0 km3/yr in 2002–2008. Simulated irrecoverable groundwater depletion accompanying aquifer compaction (land subsidence) was ∼40.0 km3, which is equivalent to the initial compressive storage. The corresponding average annual compressive storage depletion would be 1.1 km3, assuming all subsidence occurred after 1970. Groundwater depletion is constrained principally by groundwater pumpage at seasonal scales, but is dominated by precipitation fluctuations at interannual to decadal timescales.

[51] Current groundwater pumpage (∼22 km3/yr) is almost twice the predevelopment recharge rate (∼13 km3/yr). Spatially averaged recharge from the transient model, estimated via model calibration, is ∼120 mm/yr (17 km3/yr), corresponding to ∼20% of precipitation and irrigation. Irrigation pumpage is supplied by groundwater storage, groundwater recharge, including lateral flow from mountain front recharge, leakage from the Yellow River, and reduction in natural discharge, including ET and base flow to rivers. Spatiotemporal variability in groundwater depletion indicates that the primary control on depletion is irrigation pumpage. Municipal pumpage also contributes to depletion in and adjacent to cities. The largest storage declines are found in the piedmont region and in the vicinity of major cities. Storage declines in the piedmont region and central plain are from recoverable drainable and irrecoverable compressive storage, but storage declines are dominated by irrecoverable storage accompanying aquifer compaction in the coastal plain.

[52] Various water management strategies to reduce depletion, including (1) MAR in the piedmont region, (2) utilization of brackish water in the central and coastal plains, (3) improvement of WUE by reducing ET, and (4) implementation of the SNWT project, were evaluated using modeling analyses. Simulations of future pumpage and recharge to 2030 with implementation of the different strategies indicate that storage will continue to decline until 2030. By the end of 2030, storage contributes 18%, 24%, 16%, and 2% of pumpage for the above four scenarios, respectively. The SNWT project, which may be most effective in reducing groundwater pumpage, only results in storage recovery in the piedmont region and some local areas in the central and coastal plains. Combining all four strategies resulted in increases in groundwater storage by ∼50 km3 by 2030 but would require an additional 50 years for aquifer storage to fully recover in the NCP. These management strategies are essential for mitigating groundwater depletion within the NCP; they also provide valuable information for developing sustainable groundwater management strategies in other global hotspots of aquifer depletion. It is essential that the long timescales required for groundwater storage recovery be considered while implementing groundwater management strategies in any aquifer system.

Acknowledgments

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Study Site and Hydrogeologic Data
  5. 3. Groundwater Storage Change From Water Level Data
  6. 4. Simulation of the Groundwater Flow System
  7. 5. Results and Discussion
  8. 6. Approaches Toward More Sustainable Groundwater Management
  9. 7. Model Limitations and Future Studies
  10. 8. Conclusions
  11. Acknowledgments
  12. References
  13. Supporting Information

[53] This work was supported by the National Natural Science Foundation of China (grants 40911130505 and 41140017) and China Geological Survey (grant 1212011121174). We appreciate the support from Hao Wang and Yangwen Jia of the China Institute of Water Resources and Hydropower Research. We also thank Leonard Konikow, Prabhakar Clement, and two anonymous reviewers for their helpful comments and suggestions which have led to significant improvement of this paper.

References

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Study Site and Hydrogeologic Data
  5. 3. Groundwater Storage Change From Water Level Data
  6. 4. Simulation of the Groundwater Flow System
  7. 5. Results and Discussion
  8. 6. Approaches Toward More Sustainable Groundwater Management
  9. 7. Model Limitations and Future Studies
  10. 8. Conclusions
  11. Acknowledgments
  12. References
  13. Supporting Information
  • Alley, W. M., and S. A. Leake (2004), The journey from safe yield to sustainability, Ground Water, 42(1), 1216.
  • Alley, W. M., T. E. Reilly, and O. L. Franke (1999), Sustainability of ground-water resources, U.S. Geol. Survey Circular 1186, U.S. Geol. Survey, Denver, Colo.
  • Alley, W. M., R. W. Healy, J. W. LaBaugh, and T. E. Reilly (2002), Flow and storage in groundwater systems, Science, 296(5575), 19851990.
  • Blanke, A., S. Rozelle, B. Lohmar, J. Wang, and J. Huang (2007), Water saving technology and saving water in China, Agr. Water Manage., 87(2), 139150.
  • Bredehoeft, J. D. (2002), The water budget myth revisited: Why hydrogeologists model, Ground Water, 40(4), 340345.
  • Cai, X., and M. W. Rosegrant (2004), Optional water development strategies for the Yellow River Basin: Balancing agricultural and ecological water demands, Water Resour. Res., 40(8), W4SW8S, doi:10.1029/2003WR002488.
  • Chen, W. (1990), Evaluation of water resources and hydraulic parameters for the shallow groundwater in the Hebei Plain [in Chinese], Hydrogeol. Eng. Geol., 6, 1922.
  • Chen, W. (1999), Groundwater in Hebei Province [in Chinese], Seismological Press, Beijing.
  • Chen, M., and F. Ma (2002), Groundwater Resources and the Related Environment in China [in Chinese], Earthquake Press, Beijing.
  • Chen, J., C. Tang, Y. Sakura, A. Kondoh, J. Yu, J. Shimada, and T. Tanaka (2004), Spatial geochemical and isotopic characteristics associated with groundwater flow in the North China Plain, Hydrol. Process., 18(16), 31333146.
  • Chen, J. Y., C. Y. Tang, Y. J. Shen, Y. Sakura, A. Kondoh, and J. Shimada (2003), Use of water balance calculation and tritium to examine the dropdown of groundwater table in the piedmont of the North China Plain (NCP), Environ. Geol., 44(5), 564571.
  • Chen, Z. Y., J. X. Qi, J. M. Xu, J. M. Xu, H. Ye, and Y. J. Nan (2003), Paleoclimatic interpretation of the past 30 ka from isotopic studies of the deep confined aquifer of the North China plain, Appl. Geochem., 18(7), 9971009.
  • Chen, Z., Z. Nie, Z. Zhang, J. Qi, and Y. Nan (2005), Isotopes and sustainability of ground water resources, North China Plain, Ground Water, 43(4), 485493.
  • Cui, Y., Y. Wang, J. Shao, Y. Chi, and L. Lin (2009), Research on groundwater regulation and recovery in North China Plain after the implementation of south-to-north water transfer [in Chinese with English abstract], Resour. Sci., 31(3), 382387.
  • Currell, M. J., D. Han, Z. Chen, and I. Cartwright (2012), Sustainability of groundwater usage in northern China: Dependence on palaeowaters and effects on water quality, quantity and ecosystem health, Hydrol. Process., 26(26), 40504066.
  • Doherty, J. (2003), PEST—Model-Independent Parameter Estimation. User's Manual, 5th ed., Watermark Numerical Computing, Brisbane, Australia.
  • Duan, Y., and G. Xiao (2003), Sustainable utilization of groundwater resources in Hebie Plain [in Chinese], Hydrogeol. Eng. Geol., 30(1), 28.
  • Falkenmark, M. (2005), Water usability degradation, Water Int., 30(2), 136146, doi:10.1080/02508060508691854.
  • Fang, Q. X., L. Ma, T. R. Green, Q. Yu, T. D. Wang, and L. R. Ahuja (2010), Water resources and water use efficiency in the North China Plain: Current status and agronomic management options, Agr. Water Manage., 97(8), 11021116.
  • Fei, J. (1988), Groundwater resources in the North China Plain, Environ. Geol. Water Sci., 12(1), 6367, doi:10.1007/BF02574828.
  • Fei, Y., J. Miao, Z. Zhang, Z. Chen, H. Song, and M. Yang (2009), Analysis on evolution of groundwater depression cones and its leading factors in North China Plain [in Chinese with English abstract], Resour. Sci., 31(3), 394399.
  • Fogg, G. E., and E. M. LaBolle (2006), Motivation of synthesis, with an example on groundwater quality sustainability, Water Resour. Res., 42(3), W03S05, doi:10.1029/2005WR004372.
  • Foster, S. (2000), Sustainable groundwater exploitation for agriculture: Current issues and recent initiatives in the developing world, Uso intensivo de las aguas subterráneas, Aspectos éticos, tecnológicos y económicos, Series A, no. 6, Papeles del Proyecto Aguas Subterrfineas, Papeles del proyecto aguas subterráneas.
  • Foster, S., H. Garduno, R. Evans, D. Olson, Y. Tian, W. Zhang, and Z. Han (2004), Quaternary aquifer of the North China Plain—Assessing and achieving groundwater resource sustainability, Hydrogeol. J., 12(1), 8193.
  • Han, Z. S. (1998), Groundwater for urban water supplies in northern China—An overview, Hydrogeol. J., 6(3), 416420, doi:10.1007/PL00010968.
  • Harbaugh, A. W., E. R. Banta, M. C. Hill, and M. G. McDonald (2000), MODFLOW-2000, The U. S. Geological Survey modular ground-water model-user guide to modularization concepts and the ground-water flow process, Open-File Rep. 00-92, U.S. Geol. Survey, Reston, Va.
  • Hebei Department of Water Resources (1995), Hebei Water Conservancy [in Chinese], Hebei People's Publishing, Shijiazhuang, China.
  • Hebei Provincial Bureau of Statistics (2009), Hebei Agricultural Statistical Yearbook 2008, China Statistics Press, Beijing.
  • Hu, Y. K., J. P. Moiwo, Y. H. Yang, S. M. Han, and Y. M. Yang (2010), Agricultural water-saving and sustainable groundwater management in Shijiazhuang Irrigation District, North China Plain, J. Hydrol., 393(3–4), 219232.
  • Hurd, B., N. Leary, R. Jones, and J. Smith (1999), Relative regional vulnerability of water resources to climate change, J. Am. Water Resour. Assoc., 35(6), 13991409.
  • Jia, J., and C. Liu (2002), Groundwater dynamic drift and response to different exploitation in the North China Plain: A case study of Luancheng County, Hebei Province [in Chinese with English abstract], Acta Geogr. Sin., 57(2), 201209.
  • Jiang, X., H. Li, S. Cai, L. Ji, T. Cheng, and N. Xu (2009), Investigation and analysis of farmers'irrigation actions in North China Plain well irrigation district [in Chinese with English abstract], Anhui Agr. Sci. Bull., 15(11), 152154.
  • Jiao, H., and Y. Duan (2005), Sustainable utilization of groundwater resource in the Yubei plain [in Chinese], Hydrogeol. Eng. Geol., 1, 6166.
  • Johnston, R. H. (1997), Sources of water supplying pumpage from regional aquifer systems of the United States, Hydrogeol. J., 5(2), 5463.
  • Kendy, E., D. J. Molden, T. S. Steenhuis, C. Liu, and J. Wang (2003), Policies drain the North China Plain: Agricultural policy and groundwater depletion in Luancheng County, 1949–2000, Research Rep. 71, International Water Management Institute, Colombo, Sri Lanka.
  • Kendy, E., Y. Zhang, C. Liu, J. Wang, and T. Steenhuis (2004), Groundwater recharge from irrigated cropland in the North China Plain: Case study of Luancheng County, Hebei Province, 1949–2000, Hydrol. Process., 18(12), 22892302.
  • Kendy, E., J. X. Wang, D. J. Molden, C. Zheng, C. M. Liu, and T. S. Steenhuis (2007), Can urbanization solve inter-sector water conflicts? Insight from a case study in Hebei Province, North China Plain, Water Policy, 9(suppl. 1), 7593.
  • Kinzelbach, W., P. Bauer, T. Siegfried, and P. Brunner (2003), Sustainable groundwater management—Problems and scientific tool, Episodes, 26(4), 279284.
  • Konikow, L. F. (2011), Contribution of global groundwater depletion since 1900 to sea-level rise, Geophys. Res. Lett., 38(17), L17401, doi:10.1029/2011GL048604.
  • Konikow, L. F., and E. Kendy (2005), Groundwater depletion: A global problem, Hydrogeol. J., 13(1), 317320.
  • Li, F., X. Song, C. Tang, C. Liu, J. Yu, and W. Zhang (2007), Tracing infiltration and recharge using stable isotope in Taihang Mt., north China, Environ. Geol., 53(3), 687696.
  • Liu, C., and J. Xia (2004), Water problems and hydrological research in the Yellow River and the Huai and Hai River basins of China, Hydrol. Process., 18(12), 21972210.
  • Liu, C., and Z. Wei (1989), Agricultural Hydrology and Water Resources in the North China Plain [in Chinese], Science Press, Beijing.
  • Liu, C., and H. Zheng (2002), South-to-north water transfer schemes for China, Int. J. Water Res. Dev., 18(3), 453471.
  • Liu, C., J. Yu, and E. Kendy (2001), Groundwater exploitation and its impact on the environment in the North China Plain, Water Int., 26(2), 265272.
  • Liu, C., X. Zhang, and Y. Zhang (2002), Determination of daily evaporation and evapotranspiration of winter wheat and maize by large-scale weighing lysimeter and micro-lysimeter, Agr. Forest Meteorol., 111(2), 109120.
  • Liu, C., J. Chen, and J. Qiao (2004), A survey of irrigation status in the well irrigation district in the North China Plain [in Chinese], Water Res. Dev. Res., 10, 3841.
  • Liu, J., C. Zheng, L. Zheng, and Y. Lei (2008), Ground water sustainability: Methodology and application to the North China Plain, Ground Water, 46(6), 897909.
  • Lu, X., M. Jin, M. T. van Genuchten, and B. Wang (2011), Groundwater recharge at five representative sites in the Hebei Plain, China, Ground Water, 49(2), 286294, doi:10.1111/j.1745-6584.2009.00667.x.
  • Luckey, R. L., and M. F. Becker (1999), Hydrogeology, water use, and simulation of flow in the High Plains aquifer in northwestern Oklahoma, southeastern Colorado, southwestern Kansas, northeastern New Mexico, and northwestern Texas, Water-Resources Inv. Rep. 99-4104, U.S. Geol. Survey, Denver, CO.
  • Ma, Z., and Y. H. Duan (2005), Sustainable utilization of groundwater resource in the Lubei plain of Shandong Province [in Chinese], Hydrogeol. Eng. Geol., 2, 110.
  • National Bureau of Statistics of China (2003), China Population County Census Data 2000, China Statistics Press, Beijing.
  • National Bureau of Statistics of China (2010), China Statistical Yearbook 2010, China Statistics Press, Beijing.
  • Oki, T., and S. Kanae (2006), Global hydrological cycles and world water resources, Science, 313(5790), 10681072, doi:10.1126/science.1128845.
  • Piao, S., P. Ciais, Y. Huang, Z. Shen, S. Peng, J. Li, L. Zhou, H. Liu, Y. Ma, and Y. Ding (2010), The impacts of climate change on water resources and agriculture in China, Nature, 467(7311), 4351, doi:10.1038/nature09364.
  • Ren, X. (2007), Water Resources Assessment of the Haihe River Basin [in Chinese], China Water Power Press, Beijing.
  • Scanlon, B. R., R. W. Healy, and P. G. Cook (2002), Choosing appropriate techniques for quantifying groundwater recharge, Hydrogeol. J., 10(1), 1839.
  • Shah, T., D. Molden, R. Sakthivadivel, and D. Seckler (2000), The global groundwater situation: Overview of opportunities and challenges, International Water Management Institute, Colombo, Sri Lanka.
  • Shah, T., A. D. Roy, A. S. Qureshi, and J. Wang (2003), Sustaining Asia's groundwater boom: An overview of issues and evidence, Nat. Resour. Forum, 27, 130141.
  • Shao, J., Y. Cui, Y. Zhao, and G. He (2003), Groundwater numerical modeling and its application in affected zone of The Down-yellow River (Henan Section) [in Chinese], J. Changchun Univ. Sci. Technol., 33(1), 5155.
  • Shi, J., J. Guo, Y. Sun, Y. Sun, and Y. Chen (2006), Spatial analysis of the relation between deep groundwater exploitation and land subsidence in Beijing-Tianjin-Hebei-Dezhou Plain Area, Geol. Rev., 52(6), 804809.
  • Siebert, S., J. Burke, J. M. Faures, K. Frenken, J. Hoogeveen, P. Döll, and F. T. Portmann (2010), Groundwater use for irrigation—A global inventory, Hydrol. Earth Syst. Sci., 14(10), 18631880.
  • Song, X., F. Li, C. Liu, C. Tang, Q. Zhang, and W. Zhang (2007), Water cycle in Taihang Mountain and its recharge to groundwater in the North China Plain [in Chinese with English abstract], J. Nat. Resour., 22(3), 398408.
  • Song, Z. W., H. L. Zhang, R. L. Snyder, F. E. Anderson, and F. Chen (2010), Distribution and trends in reference evapotranspiration in the North China Plain, J. Irrig. Drain. Eng., 136(4), 240247.
  • Sophocleous, M. (2002), Interactions between groundwater and surface water: The state of the science, Hydrogeol. J., 10(1), 5267.
  • Theis, C. V. (1940), The source of water derived from wells, Civil Eng., 10(5), 277280.
  • Wada, Y., L. P. H. van Beek, C. M. van Kempen, J. W. T. M. Reckman, S. Vasak, and M. F. P. Bierkens (2010), Global depletion of groundwater resources, Geophys. Res. Lett., 37, L20402, doi:10.1029/2010GL044571.
  • Wang, B. G., M. G. Jin, J. R. Nimmo, L. Yang, and W. F. Wang (2008), Estimating groundwater recharge in Hebei Plain, China under varying land use practices using tritium and bromide tracers, J. Hydrol., 356(1–2), 209222.
  • Wang, S., J. Shao, X. Song, Y. Zhang, Z. Huo, and X. Zhou (2008), Application of MODFLOW and geographic information system to groundwater flow simulation in North China Plain, China, Environ. Geol., 55(7), 14491462.
  • Welsh, W. D. (2006), Great Artesian Basin transient groundwater model, Bureau of Rural Sci., Dept. of Agric., Fisheries and Forestry, Australia.
  • Wu, C., Q. H. Xu, X. Q. Zhang, and Y. H. Ma (1996), Palaeochannels on the North China Plain: Types and distributions, Geomorphology, 18(1), 514.
  • Xia, J. (2006), Water Security and ground water problem to changing environment in North China, paper presented at International Symposium on Groundwater Sustainability, Alicante, Spain.
  • Xu, Q. H., C. Wu, X. Q. Zhu, and X. L. Yang (1996), Palaeochannels on the North China Plain: Stage division and palaeoenvironments, Geomorphology, 18(1), 1525.
  • Xu, Y., X. Mo, Y. Cai, and X. Li (2005), Analysis on groundwater table drawdown by land use and the quest for sustainable water use in the Hebei Plain in China, Agr. Water Manage., 75(1), 3853.
  • Xu, G., L. Liu, Y. Fei, D. Li, and Y. Qian (2009), Research on the adjustment of groundwater storage in North China Plain [in Chinese with English abstract], Resour. Sci., 31(3), 375381.
  • Xue, L., W. Li, W. Yang, J. Li, and X. Zhu (2010), Numerical modelling on exploited shallow groundwater flow in the plain of Haihe river basin [in Chinese with English abstract], Geotech. Invest. Survey., 3, 5055.
  • Yang, Y. H., M. Watanabe, Y. Sakura, C. Y. Tang, and S. Hayashi (2002), Groundwater-table and recharge changes in the Piedmont region of Taihang Mountain in Gaocheng City and its relation to agricultural water use, Water SA., 28(2), 171178.
  • Zhang, G., Z. Liu, Y. Fei, Y. Lian, and M. Yan (2010), The relationship between the distribution of irrigated crops and the supply capability of regional water [in Chinese with English abstract], Acta Geophys. Sin., 31(1), 1722.
  • Zhang, X. (2007), Numerical simulation of the shallow groundwater in the Haihe Basin [in Chinese], Haihe Water Resour., 3, 5254.
  • Zhang, X. Y., D. Pei, and C. S. Hu (2003), Conserving groundwater for irrigation in the North China Plain, Irrigation Sci., 21(4), 159166.
  • Zhang, X., L. Xue, Q. Zhang, and J. Li (2008), Simulation of groundwater exploitation at early stage and analysis on water budget in Haihe River Basin [in Chinese], J. Arid Land Resour. Environ., 9, 102107.
  • Zhang, Z., Z. Shen, Y. Xue, and F. Ren (2000), Groundwater Environment Evolution in the North China Plain [in Chinese], Geological Publishing House, Beijing.
  • Zhang, Z., et al. (2009a), Investigation and Assessment of Sustainable Utilization of Groundwater Resources in the North China Plain [in Chinese], Geological Publishing House, Beijing.
  • Zhang, Z., G. Luo, Z. Wang, C. Liu, Y. LI, and X. Jiang (2009b), Study on sustainable utilization of groundwater in North China Plain [in Chinese with English abstract], Resour. Sci., 31(3), 355360.
  • Zhao, Y., J. Shao, Z. Yan, Y. Cui, H. Jiao, and G. He (2003), Identification of area boundary receiving leakage from the lower reaches of the Yellow River [in Chinese], Yellow River, 25(1), 35.
  • Zheng, C., J. Liu, G. Cao, E. Kendy, H. Wang, and Y. Jia (2010), Can China cope with its water crisis? Perspectives from the North China Plain, Ground Water, 48(3), 350354.
  • Zhou, Y. (2009), A critical review of groundwater budget myth, safe yield and sustainability, J. Hydrol., 370(1–4), 207213, doi:10.1016/j.jhydrol.2009.03.009.

Supporting Information

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Study Site and Hydrogeologic Data
  5. 3. Groundwater Storage Change From Water Level Data
  6. 4. Simulation of the Groundwater Flow System
  7. 5. Results and Discussion
  8. 6. Approaches Toward More Sustainable Groundwater Management
  9. 7. Model Limitations and Future Studies
  10. 8. Conclusions
  11. Acknowledgments
  12. References
  13. Supporting Information

All Supporting Information may be found in the online version of this article.

FilenameFormatSizeDescription
wrcr011899-sup-0001-SupplD1.dat1466KDataset 1
wrcr011899-sup-0002-SupplD2.dat1466KDataset 2
wrcr011899-sup-0003-SupplD3.dat1466KDataset 3
wrcr011899-sup-0004-SupplD4.ba6573KDataset 4
wrcr011899-sup-0005-SupplD5.chd6KDataset 5
wrcr011899-sup-0006-SupplD6.dis6070KDataset 6
wrcr011899-sup-0007-SupplD7.ets1359KDataset 7
wrcr011899-sup-0008-SupplD8.lpf13521KDataset 8
wrcr011899-sup-0009-SupplD9.nam1KDataset 9
wrcr011899-sup-0010-SupplD10.oc5KDataset 10
wrcr011899-sup-0011-SupplD11.pcg0KDataset 11
wrcr011899-sup-0012-SupplD12.wel173168KDataset 12
wrcr011899-sup-0013-txts01.doc99KText file in Doc format for the supplemental model information about additional technical details regarding the model construction are provided.

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.