Environ Earth Sci (2016) 75:1131 DOI 10.1007/s12665-016-5913-x

ORIGINAL ARTICLE

Theory and methodology of critical control levels of groundwater: a case study of Tianjin, China Zhiqiang Wang1 • Jingli Lu1 • Caiyun Hu1 • Wei Zhang2 • Ruiying Dong2 Hua Li2

•

Received: 11 March 2015 / Accepted: 19 July 2016 / Published online: 2 August 2016 Ó Springer-Verlag Berlin Heidelberg 2016

Abstract Groundwater is one of the most important freshwater resources worldwide. However, long-term overabstraction of groundwater has resulted in deteriorating groundwater functions, including gradually decreasing resources function and geo-environmental function and subsequent adverse geo-environmental issues. In this study, the concept of critical control levels of groundwater including blue line levels and red line levels, which play an important role in groundwater resource management, is presented for the first time. In addition, the theory of critical control levels of groundwater is established based on the groundwater geo-environmental function and resources function. A coupled numerical model of groundwater flow field and land subsidence is then established for Tianjin based on the conceptual model and analysis of the hydrological conditions of the study area. Finally, the critical control levels of groundwater for the subarea are presented based on results obtained from a numerical model by the model space division and time discretization, hydrogeological parameters processing, model calibration and validation. The results presented herein will be useful for groundwater resource management.

& Zhiqiang Wang [email protected] 1

School of Energy and Environmental Engineering, Hebei University of Technology, Tianjin 300401, People’s Republic of China

2

Hydrology and Water Resources Survey and Management Center of Tianjin, Tianjin 300061, People’s Republic of China

Keywords Management of groundwater resource  Groundwater abstraction  Groundwater level  Land subsidence

Introduction Groundwater is one of the most important sources of potable water in the world. However, groundwater in many parts of the world is currently being over exploited, which has resulted in continuously declining groundwater levels. This has caused negative effects on the geo-environment such as groundwater level depression and land subsidence (Jesu´s et al. 2013). There are many causes of land subsidence, but groundwater abstraction is most common (Shi et al. 2011). Groundwater level fluctuations can be used to precisely quantify groundwater abstraction (Song et al. 2011). Several models have been applied to investigate a wide variety of hydrogeological conditions, and the results have helped identify aquifer properties and analyze changes in groundwater levels, as well as groundwater flow dynamics (Yang et al. 2008). Therefore, groundwater levels simulated by mathematical modeling have been useful in management of groundwater resources (Gatot et al. 2013). In addition, groundwater level fluctuations are closely related to several natural hazards, including land subsidence (Uma 2012). So studies of groundwater level fluctuations have become important since they can be used to predict environmental hazards (Varouchakis and Hristopulos 2013). Groundwater levels provide more important information regarding groundwater resource management than quantification of groundwater abstraction (Leonard and Eloise 2005); therefore, they have important applications in

123

1131 Page 2 of 13

groundwater resource management. To enable use of groundwater level for management of groundwater resources, it is essential that their critical control levels be determined accurately. Tianjin is one of the most developed areas in North China. This region has suffered from severe land subsidence caused by groundwater over-abstraction since the 1960s. The critical control levels of groundwater presented by the theory established in this paper will be useful for groundwater resource management in groundwater over exploited areas worldwide.

Concept of critical control levels of groundwater The critical control levels of groundwater refer to a series of groundwater levels thresholds that can specifically indicate the states of groundwater abstraction. These groundwater levels have two distinct features. Specifically, they are characterized by fluctuation and are not static values but influenced by hydrology, meteorology and the abstraction of groundwater. Additionally, they are characterized by instruction, consisting of a series of expected groundwater levels set for implementation of groundwater resource management objectives in different periods by the groundwater resource administrative departments (Li et al. 2013). The critical control levels of groundwater can be classified by blue line levels and red line levels according to the degree of groundwater level dynamics, ecological function and geo-environmental function. The blue line levels of groundwater may be defined as a series of groundwater levels or thresholds, which indicates that the groundwater circulation is in the equilibrium of recharge and discharge when the groundwater levels are at or above the blue line levels. They can also be defined as a series of groundwater levels or thresholds set for implementation of groundwater management objectives. When the groundwater levels are at or above these blue line levels, the groundwater resources and ecological and geo-environmental functions will function normally (Chen et al. 2014). The red line levels of groundwater may be defined as a series of groundwater levels or thresholds, which indicates that the groundwater system equilibrium is destructed and groundwater abstraction will result in a continuous decline in groundwater levels and land subsidence when groundwater levels are at or below the red line levels. The red line levels of groundwater can also be defined as a series of groundwater levels or thresholds set for implementation of groundwater management objectives (Yan et al. 2012). Corresponding to the different critical control levels of groundwater, the groundwater resource management strategy can be made, which is shown in Table 1.

123

Environ Earth Sci (2016) 75:1131 Table 1 Critical control levels and management strategies Critical levels

Management grade

Management strategies

Blue lines

Blue area

Reasonable abstraction

Yellow area

Restrictive abstraction

Red area

Prohibited abstraction

Red lines

Theory and methodology of critical control levels of groundwater Theory of critical control levels of groundwater based on groundwater geo-environmental functions Great decreases in groundwater levels primarily occur because of unreasonable abstraction, which may cause land subsidence, ground fissures and irreversible ground surface collapse. A theory of critical control levels of groundwater based on geo-environmental functions can be established based on control of geo-environmental issues as inducers of constrained conditions to control the groundwater level. The land subsidence caused by the groundwater overabstraction was a complex process in which the seepage interacts with the stress field. In Tianjin areas with a high degree of deep groundwater abstraction, the regional deep groundwater level declines, land subsidence increases, leakage recharge of shallow groundwater and released water of compaction in clayey soil are important components of the deep groundwater abstraction (Shi et al. 2011). The models describing land subsidence due to groundwater abstraction are all well based on theoretical considerations. The mathematical model of land subsidence generally includes 2D groundwater seepage models and a one-dimensional soil skeleton deformation model, but the following quasi-3D model is currently widely applied (Maheswaran and Rakesh 2013):     o oHi o oHi oHi T T þ Qileak þ Qisub þ ¼S ox oy ox oy ot  Qipump þ Qirrg ð1Þ where Hi is the water head of the aquifer; Qileak is leakage; Qisubk is water release by soil skeleton deformation or water soakage by soil skeleton rebound; Qipump is the pumpage amount; Qiirrg is irrigation amount; S is the storage coefficient; and T is the coefficient of hydraulic conductivity. Theory of critical control levels of groundwater based on groundwater resource The level of groundwater is closely related to groundwater abstraction, so the level of groundwater will fluctuate when

Environ Earth Sci (2016) 75:1131

Page 3 of 13

the groundwater is abstracted. Groundwater flow in the 3D space aquifer media can be expressed as the control equation for the fluctuation of groundwater levels, in which the variable parameter is the value of the fluctuation in groundwater level (Li et al. 2012).       o oh o oh o oh oh Txx Tyy Tzz þ þ ¼S RþL ox ox oy oy oz oz ot ð2Þ

1131

abstraction, groundwater level, and land subsidence was established to control groundwater level fluctuations (Yang et al. 2012). The coupled numerical model can be generalized as an unsteady three-dimensional numerical groundwater flow model coupled with a one-dimensional soil consolidation model, which can be expressed as follows (Yan et al. 2012; Li et al. 2014):

8       oh o oh o oh o oh > > > K K Kz S ¼ þ þ þeþq x; y; z 2 X; t  0 > > ot ox ox oy oy oz oz > >       > 2 2 2 > oh oh oh oh oh > > > > > l ot ¼ K ox þK oy þKz oz  oz ðKz þ pÞ þ p x; y; z 2 C0 ; t  0 > > < hðx; y; z; tÞj ¼ h x; y; z 2 X; t  0 0 t¼0 hðx; y; z; tÞj ¼ H x; y; z 2 C1 ; t  0 > C1 >  > >  oh > > > x; y; z 2 C2 ; t  0 Kn  ¼ qðx; y; z; tÞ > > ~ C2 on > >  > > oh > ðhr  hÞ > >  Kn  ¼ 0 x; y; z 2 C3 ; t  0 : r oz C3

8 oh > > q1 ¼  Ssk > > ot > > > < Db ¼ S b oh sk ot0 0 2 0 > > > o h ¼ Ss oh > > 2 > Kv0 ot > : oz Db0 jt¼0 ¼ 0

x; y; z 2 C4 ;

t0

x; y; z 2 C4 ;

t0

x; y; z 2 C5 ;

t0

x; y; z 2 X;

t0

ð3Þ

where Ti (i = xx, yy, zz) is the coefficient of hydraulic conductivity; h is the water head of the aquifer; R is a term describing sources and sinks, which includes pumpage and rainfall infiltration recharge; L is leakage; S is the storage coefficient of aquifer media; and t is the model time. Boundary conditions: ah þ b oh on ¼ f ðx; y; z; tÞ Constant water head boundary: b ¼ 0; a ¼ 1; h ¼ f1 ðx; y; z; tÞ Constant flow boundary: a ¼ 0; b ¼ KrFðx; yÞ; b oh on ¼ f2 ðx; y; z; tÞ Mixed boundary: Lðh  h0 ÞjrF j ¼ K oh on rF Initial conditions: h ¼ h0 ðx; y; z; 0Þ Model of coupled groundwater flow field and land subsidence Because groundwater level is a direct indicator of groundwater abstraction, dynamic fluctuations of groundwater level have a direct and close relationship with groundwater abstraction. In addition, groundwater overabstraction can cause land subsidence; accordingly, correlations exist among groundwater abstraction, groundwater level fluctuation and land subsidence (Ajay 2013). Therefore, a coupled numerical model of groundwater

where X is the computational seepage zone, h is the groundwater level elevation, K is the horizontal hydraulic conductivity, Kz is the vertical hydraulic conductivity, e is the source and convergence, x, y, and z are coordinate variables, q is the total compression released water amount, q1 is the anhysteretic compression released water amount, h0 is the initial groundwater level elevation of the groundwater system, S is the specific aquifer storage rate, l is the specific yield of the confined aquifer, p is the intensity of evaporation and precipitation recharge of the confined aquifer surface, C0 is the upper boundary of the computational seepage zone, C1 is the first boundary of the seepage zone, C2 is the second boundary of the seepage zone, C3 is the third boundary of the seepage zone, C4 is the anhysteretic compression aquifer zone, C5 is the hysteretic compression clay zone, Kn is the permeability coefficient in normal directions of the boundary, n is the external normal direction of the boundary, hr is the water level elevation of the third boundary, r is the resistance coefficient of the third boundary, Db is the amount of anhysteretic compression, b is the thickness of anhysteretic compression, Ssk is the water storage rate of the aquifer skeleton, h0 is the water level elevation of the hysteretic cohesive soil layer, S0s is the water storage rate of hysteretic cohesive soil layer, Kv0 is the vertical permeability

123

1131 Page 4 of 13

coefficient of the hysteretic cohesive soil layer and Db0 is the initial land subsidence.

Case study Geology and hydrogeology of the study area The study area is the entire plain of Tianjin, which covers an area of about 1.1 9 104 km2. The study area extends from 116°420 0500 E to 118°030 3100 E Longitude and 38°330 5700 N to 40°000 0700 N Latitude. A map of the study area is provided in Fig. 1.

Fig. 1 Location map of the study area

123

Environ Earth Sci (2016) 75:1131

The geological structure of Tianjin is very complex, with neotectonics controlling the deposits of Cenozoic strata, as well as the groundwater storage, recharge and drainage. Because the storage of groundwater in Tianjin is controlled by the geological structure, hydrology and ancient geographical conditions, the hydrogeological structure shows obvious horizontal zoning from the piedmont plain to the coastal plain. On the horizontal level, the groundwater distribution in Tianjin is divided by the Baodi fracture, which stretches from Baodi to Ninghe. In the northern part of the fracture, the bedrock is buried at less than 300 m and with storage of the Quaternary pore groundwater. In the southern portion

Environ Earth Sci (2016) 75:1131

of the fracture, it has a thick Cenozoic aquifer with pore groundwater. Vertically, the groundwater is distributed among the following six aquifers: Aquifer 1: the depth of the bottom is between 20 and 30 m. The aquifer media changes from coarse sand in the mountain front plain to fine sand in the coastal plain. Aquifer 2: the depth of the bottom is between 180 and 228 m. The aquifer media is primarily fine sand (65 % of the total area), while the rest of the media is silt (30 %) and medium sand (5 %). Aquifer 3: the depth of the bottom is between 290 and 315 m. The aquifer media is mainly fine sand (about 65 % of the total area), while silt and medium sand comprise 20 and 15 % of the total area, respectively. Aquifer 4: the depth of the bottom is between 370 and 429 m. The aquifer media consists of fine sand distributed throughout the entire area, which accounts for more than 60 % of the total media. Aquifer 5: the depth of the bottom is between 480 and 500 m. The aquifer media is mainly composed of fine sand (43 %) and silt (57 %). Aquifer 6: the depth of the bottom is between 580 and 650 m. The aquifer media is primarily medium sand with a loose structure and poor cementation. Aquifer 1 is an unconfined aquifer that receives groundwater recharge from rainfall and loses water via

Page 5 of 13

1131

evaporation. Aquifers 2–6 are confined aquifers composed of multilayers of fine or medium sand that receive water by seepage from boundary areas and aquifer media compression and lose water by groundwater abstraction. A crosssectional hydrogeological map of the study area is shown in Fig. 2. The geochemical properties of groundwater vary greatly from the piedmont plain to the coastal plain, with concentrations of bicarbonate ion (HCO3-), calcium ion (Ca2?) and magnesium ion (Mg2?) decreasing gradually, and those of chloride ion (Cl-) and sodium ion (Na?) increasing gradually. In addition, the total dissolved solids (TDS) of groundwater increased from 0.5 g/L in the piedmont plain to 15.0 g/L in the coastal plain. In the piedmont plain (Ji County), the chemical constituents of groundwater were mainly HCO3-CaMg, and the TDS ranged from 0.5 to 1.5 g/L. In the central plain (from Baodi district to Jinghai and Jinnan district), the chemical constituents of the unconfined aquifer groundwater were mainly ClSO4–NaMg and Cl– Na. In this region, the TDS increased gradually from 1.5 g/L in the NW to 10 g/L in the SE. In some subareas, the chemical constituents of the confined aquifer (from aquifers 2 to 6) were primarily HCO3–CaMg, HCO3ClSO4– Na, ClHCO3SO4–Na and HCO3Cl–Na. In the coastal plain (Hangu, Tanggu and Dagang district), the chemical constituents of unconfined aquifer groundwater were mainly ClSO4–NaMg and Cl–Na, and

Fig. 2 Hydrostratigraphy along cross-sectional line A–A0 in the study area

123

1131 Page 6 of 13

the TDS was more than 10.0 g/L, reaching as high as 30.0 g/L in some subareas. The chemical constituents of the confined aquifer groundwater were ClHCO3–Na, but the TDS was only between 0.5 and 1.4 g/L. Conceptual model The groundwater system in the conceptual model consists of six aquifers distributed vertically that are each subjected to different conditions of groundwater abstraction. Aquifer 1 is an unconfined aquifer, while Aquifers 2–6 are fully confined. The upper surface of the unconfined aquifer is the upper boundary of the model, through which the groundwater system exchanges water with the outer systems vertically via infiltration by atmospheric precipitation, irrigation and evaporation drainage. The bottom boundary of the model is impermeable, with a maximum depth of 600 m in the vertical direction. The lateral boundaries of aquifer 1 and aquifer 2 are the same boundaries shown in Fig. 3. The east boundary to the Bohai Sea is the constant head boundary (blue line in Fig. 3), and the north boundary to the border of the mountains the flow rate boundary (green line in Fig. 3),

Fig. 3 Lateral boundaries of aquifers 1 and 2

123

Environ Earth Sci (2016) 75:1131

from which the groundwater system receives rainfall infiltration inflow, and the rest boundary is the general head boundary (red line in Fig. 3). Geology and hydrogeology of the study area show that the storage of groundwater is controlled by the geological structure. The lateral boundaries of aquifers 3–6 are the same boundaries shown in Fig. 4. The north boundaries of aquifers 3–6 are all the impermeable boundaries (purple line in Fig. 4), while the rests are the general head boundaries (red line in Fig. 4), which allow flow across the boundaries. Space division and time discretization The MODFLOW software was used to simulate the groundwater flow field. The study area (or space) was divided into six layers with 346 rows and 248 columns, giving 514,848 rectangular units using the finite difference method. Among these, each cell was 500 m 9 500 m and the number of active units was 236,860. The principle of simulation periods or identification periods is required to have the following characteristics.

Fig. 4 Lateral boundaries of aquifers 3–6

Environ Earth Sci (2016) 75:1131

Page 7 of 13

The data should be abundant in the simulation period and rainfall should be representative (normal flow year or dry year). Additionally, the simulation period should be the latest special years, and the data should be as abundant as possible. It is also best to have more than one complete calendar year or hydrological year as the simulation period or identification period. According to these principles, the period from January 1998 to December 2008 was selected as the model identification period and the period from January 2007 to December 2008 was selected as the model validation period based on the groundwater abstraction data, in which 1 month is used as a stress period and a total of 132 stress periods are present (Li et al. 2014). The measured groundwater levels of December 1997 were used as the basic values, and the Kriging interpolation method was used to obtain the initial groundwater levels. Hydrogeological parameters processing Identification of hydrogeological parameters is one of the main tasks of the model. To accomplish this, parameters that characterize the permeability and groundwater storage capacity of aquifers were selected (e.g., hydraulic conductivity, specific yield and water storage rates of each aquifer). After the initial values of the parameters of the regions were input into the model, identification of the hydrogeological parameters could be started. Each parameter zoning value was determined by fitting flow fields of each aquifer that can identify the parameters (Li et al. 2014). 1.

Groundwater abstraction data

The initial groundwater abstraction was calculated by the quota method from 1998 to 2005 and by statistical analysis from 2006 to 2008. 2.

Rainfall infiltration recharge

The rainfall infiltration recharge was calculated as follows Singh and Woolhiser (2002): Qki ¼ 103 ai Pki Fi

ð4Þ

where Qki is the precipitation infiltration recharge in the ith subregion during the kth month; ai is the rainfall infiltration coefficient in the ith subregion; Pki is the amount of rainfall infiltration in the ith subregion; and Fi is the ith subregion area. The amount of rainfall is monitored by the 52 meteorological stations and distributed to each district by the Thiessen polygon rule. 3.

Irrigation infiltration recharge

The irrigation infiltration recharge is mainly infected by the irrigation quota, irrigation mode, agricultural acreage

1131

and lithology of the aeration zone, which can be calculated as follows: Qki ¼ 103 ai Pki Fi

ð5Þ

where ai is the irrigation infiltration coefficient in the ith subregion, and the other variables are the same as those in Eq. (4). 4.

Evaporation drainage

The evaporation drainage is mainly influenced by the depth of the groundwater table and climate. The evaporation drainage was calculated by the Avi Jan Andrianof formula in this study (Yang et al. 2009). 8 9 <0  =  n d [ d0 d Eg ¼ ð6Þ 0  d  d0 ; : E0 1  d0 where Eg is the intensity of groundwater evaporation; E0 is the intensity of groundwater table evaporation; d0 is the maximum depth of groundwater evaporation, which is related to the lithology of the aeration zone and from 4.0 to 5.0 m in the model; d is the depth of the groundwater table; and n is the empirical coefficient, which is 1.0 in the model. 5.

Inflow and outflow from lateral boundary

The inflow and outflow from the lateral boundaries of the study area can be estimated by Darcy’s law (Wu et al. 2003). Qc ¼ 0:1k  I  B  M  DT

ð7Þ

where Qc is inflow and outflow from the lateral boundaries, with a positive value referring to inflow and negative to outflow; k is the permeability coefficient; I is hydraulic gradient of the vertical section; B is the width of the section; M is the thickness of the aquifer; and DT is the simulation time. 6.

Land subsidence compression release

Groundwater released from the aquitard can be calculated by Eq. (8) (Leonard and Eloise 2005): X Qr ¼ hs  F  106 ð8Þ where Qr is the groundwater released from the aquitard; hs is the amount of land subsidence; and F is the calculated unit acreage. Equilibrium analysis of groundwater recharge and discharge using the aforementioned models revealed the groundwater system states. On average, the total recharge amount was 15.13 9 108 m3/a and the total discharge amount was 18.25 9 108 m3/a, resulting in a difference between recharge and discharge of -3.12 9 108 m3/a. These

123

1131 Page 8 of 13 Table 2 Groundwater system balance

Environ Earth Sci (2016) 75:1131 Recharge amount (9108 m3/a)

Discharge amount (9108 m3/a)

Rainfall

9.42

Abstraction from unconfined aquifer

Irrigation

1.63

Abstraction from confined aquifers

6.58

Compression releasing

2.30

Outflow from lateral boundary

0.52

Lateral inflow

1.68

Evaporation

10.36

Leakage recharge

0.10 Total

18.25

Total

findings indicated that the groundwater system was losing water, as shown in Table 2. Model calibration and validation

December 2009. The calibration was achieved by adjusting several spatially distributed and sensitive soil hydraulic input parameters, such as the hydraulic conductivity and permeability coefficient. The precision and accuracy of the model calibration and verification were determined by trial and error (Zhang et al. 2011). The average groundwater levels and land subsidence observed for the Jinghai district of Tianjin were taken as an example to calibrate and verify the model (Table 2; Fig. 6).

a LSMP01

70

Land subsidence/mm

About 35,000 abstraction wells distributed throughout the study area were used to monitor groundwater levels. The values observed in these wells were used to calculate the mean groundwater levels of the study area. Moreover, there were about 1272 land subsidence monitoring points throughout the study area (Fig. 5). The model was calibrated using groundwater level and land subsidence data measured between October 2008 and

15.13

0.79

Monitored values Simulated values

60 50 40 30 20 1998

2000

2002

2004

2006

2008

2006

2008

Time/a

b LSMP02

Land subsidence/mm

70

Monitored values Simulated values

60 50 40 30 20 1998

2000

2002

2004

Time/a

Fig. 5 Location of land subsidence monitoring points

123

Fig. 6 Comparison of monitored and simulated land subsidence

Environ Earth Sci (2016) 75:1131 Table 3 Error criteria of groundwater levels for model calibration and validation

Page 9 of 13

Aquifer

Simulated levels (m)

Monitored levels (m)

Relative error (%)

Aquifer 1

-4.87

-4.50

Aquifer 2

-37.63

-35.51

5.63

Aquifer 3

-44.85

-45.33

-7.06

Aquifer 4

-58.02

-60.24

-3.82

Aquifer 5

-67.99

-65.61

3.51

Aquifer 6

-66.61

-68.55

-2.92

As shown in Table 3, the relative errors of all six aquifers were within ±10 %, and the maximum error was only 7.69 %, indicating that the coupled model is sufficiently accurate for the groundwater level simulation. Two typical land subsidence monitoring points LSMP01 (116°150 4800 E, 38°450 0000 N) and LSMP02 (117°100 0000 E, 38°380 1500 N) in Jinghai district were selected to calibrate and verify the model. The cumulative land subsidence values could be simulated by inputting the identified hydrogeological parameters into the model. These values are compared with the monitored data in Fig. 6. As shown in Fig. 6, the simulated land subsidence curves basically matched the monitored cumulative curves, within a mean relative error of ±10 % in Jinghai district. Accordingly, the model has good simulation accuracy and can be used to predict land subsidence.

1131

7.69

groundwater levels of the subarea of Jinghai are shown in Appendix Fig. 8. Correlations of groundwater abstraction versus land subsidence The determination approach of the correlation of groundwater abstraction versus land subsidence is almost the same as that of the correlation of groundwater abstraction versus groundwater levels. The land subsidence values were obtained by running the model when the groundwater abstraction data were input into the model, after which the correlation equations of groundwater abstraction versus land subsidence could be obtained. The graphs of groundwater abstraction versus predicted land subsidence of the subarea of Jinghai are shown in Appendix Fig. 9. Critical control levels of groundwater

Results and discussion Correlations of groundwater abstraction versus groundwater levels Because groundwater level fluctuation has a direct relationship with groundwater abstraction, the groundwater abstraction amount can be calculated by the monitored groundwater levels. Therefore, the correlations of groundwater levels versus groundwater abstraction can be established by the inverse simulation. The correlation equations enabled calculation of the mean groundwater levels and corresponding abstraction of each aquifer, and the relationships for the study area to be obtained. In these analyses, each subarea is regarded as a hydrogeological unit and the sources and sinks are distributed to the grid cell averagely. The groundwater abstraction data of the mean annual abstraction from 1998 to 2008 were input into the model, while the other hydrogeological parameters remained unchanged, then the correlation equations of the groundwater abstraction versus groundwater levels of each aquifer and each subarea could be obtained by running the model. The graphs of groundwater abstraction versus predicted

Groundwater abstraction in Tianjin has occurred over the long term, but was not monitored until the 1980s. As a result, it was difficult to determine healthy and unhealthy levels of groundwater based on historically monitored groundwater level data alone. Therefore, we introduced the concept of critical control levels of groundwater. Constrained conditions of critical control levels of groundwater The key issue of critical control levels of groundwater is to determine the constrained conditions. The groundwater levels can reflect groundwater abstraction and have close relationship with land subsidence. The first groundwater abstraction well in Tianjin was drilled in 1907, and there were 9522 wells by 1981. Based on the groundwater abstraction from 1981 to 2008, the average abstraction was 7.37 9 108 m3/a over the past 30 years (Fig. 7). Because of the large scale and unplanned groundwater abstraction, land subsidence has begun in areas subjected to abstraction. The process of land subsidence could be divided into three stages according to the velocity of land

123

1131 Page 10 of 13

Environ Earth Sci (2016) 75:1131

Determination of critical control levels of groundwater

Groundwater abstraction amount/108m3a-1

12.00 10.00 8.00 6.00 4.00 2.00 0.00

1985

1990

1995

2000

2005

2010

Time/year

Fig. 7 Statistics of groundwater abstraction amount

subsidence, the formation stage, developing stage and decreasing stage. The average velocity of land subsidence was 30 mm/a during the formation stage (1961–1970), 120 mm/a in the developing stage (1971–1985), and 40 mm/a during the decreasing stage (1986–2008). Large scale groundwater abstraction has occurred for over 30 years, resulting in serious geo-environmental problems and land subsidence in Tianjin; accordingly, groundwater abstraction is considered a constrained condition for determining the critical control levels of groundwater. The current geo-environmental situation, including land subsidence, groundwater abstraction and groundwater levels of Jinghai district in Tianjin is shown in Table 4. Table 4 Geo-environmental situation of Jinghai subarea in 2008

Table 5 Blue line levels of groundwater of Jinghai subarea in the near period of 2020

123

Long-term groundwater over-abstraction has resulted in declining groundwater levels and serious land subsidence. The objective of groundwater resource management is to recover the levels of groundwater via groundwater resource management planning by the local administrative departments. The prohibition of groundwater abstraction can recover groundwater levels and greatly reduce the velocity of land subsidence. Blue line levels of groundwater are defined as the levels under the conditions of prohibition of groundwater abstraction. These values will facilitate the recovery of groundwater resource function in Tianjin, even though the blue line levels obtained by the method are expected values. Taking the subarea of Jinghai District as an example, the blue line levels of groundwater in the near period of 2020 could be obtained by the correlations of groundwater abstraction versus groundwater levels and the correlations of groundwater abstraction versus land subsidence shown in Appendix Figs. 8 and 9, respectively, in which the groundwater abstractions to be zero are taken as the constrained conditions. The blue line levels of groundwater of Jinghai subarea in the near period of 2020 are shown in Table 5. As shown in Table 5, if groundwater abstraction is prohibited, the groundwater levels will rise gradually and the velocity of land subsidence will decrease by 2020 compared with the present geo-environmental situation.

Groundwater levels (m)

Aquifers

Land subsidence (mm/a)

Groundwater abstraction (104 m3/a)

Aquifer 1

4.77

8659.25

-4.50

Aquifer 2

6.45

1866.65

-35.51

Aquifer 3

7.76

2780.25

-45.33

Aquifer 4

8.55

913.50

-60.24

Aquifer 5

5.89

392.04

-65.61

Aquifer 6

3.23

315.31

-68.55

Aquifers

Land subsidence (mm/a)

Groundwater abstraction (104 m3/a)

Blue line levels (m)

Aquifer 1 Aquifer 2

-10.17 -9.76

0 0

-1.52 -27.32

Aquifer 3

-6.95

0

-41.32

Aquifer 4

-3.69

0

-50.43

Aquifer 5

-1.81

0

-61.50

Aquifer 6

-0.52

0

-65.78

Environ Earth Sci (2016) 75:1131 Table 6 Red line levels of groundwater of Jinghai subarea in the near period of 2020

Page 11 of 13

Red line levels (m)

Aquifers

Land subsidence (mm/a)

Groundwater abstraction (104 m3/a)

Aquifer 1

-0.77

1707.85

-3.54

Aquifer 2

-3.45

273.13

-33.75

Aquifer 3

6.76

532.05

-44.86

Aquifer 4

8.55

142.72

-57.01

Aquifer 5

2.89

58.01

-65.27

Aquifer 6

0.23

42.66

-68.12

These findings indicate that the critical control levels of groundwater are helpful for groundwater resource management, and the strategy of groundwater management will achieve the objective of groundwater resource functions. The red line levels of groundwater are designed to enable efficient utilization of groundwater resources without adversely affecting the geo-environment; therefore, these levels can be defined as the thresholds set for implementation of the groundwater management objective. The total allowable abstraction is 1.54 9 108 m3/a for the entire area of Tianjin in the near period according to the groundwater resource management perspective planned by the local administrative departments, that is about 10 % of recharge, whereas currently the groundwater abstraction is 7.37 9 108 m3/a. To distribute the allowable abstraction to all subareas and every aquifer according to the proportion of groundwater abstraction in past years, the allowable abstraction of every subarea and aquifer can be obtained. The groundwater abstraction of Jinghai district was calculated to be 0.28 9 108 m3/a for the near period shown in Table 6. Taking the subarea of Jinghai District as an example, the red line levels of groundwater in the near period of 2020 can also be obtained by the correlations of groundwater abstraction versus groundwater levels and the correlations of groundwater abstraction versus land subsidence shown in Appendix Figs. 8 and 9, respectively, in which the allowable abstractions are taken as the constrained conditions. The red line levels of groundwater of the subarea of Jinghai District are shown in Table 6. As shown in Table 6, if the groundwater abstraction is restricted according to the groundwater management strategy, the utilization of groundwater will be sustainable, and groundwater resource function and geo-environmental

1131

function will be recovered, which is the objective of groundwater resource management.

Conclusions 1.

2.

3.

The concept of critical control levels of groundwater is presented for the first time, and the theoretical control levels of groundwater were established based on geoenvironmental function and groundwater resource function. A coupled numerical model of groundwater flow field and land subsidence was established for Tianjin area based on the conceptual model and analysis of the hydrological conditions of the study area. The correlation equations of groundwater levels versus groundwater abstraction, groundwater abstraction versus land subsidence, and land subsidence versus groundwater levels were determined based on the coupled numerical model. Taking the groundwater abstraction as the constrained conditions, the critical control levels of groundwater (the blue line levels and the red line levels) were determined for the subarea of Jinghai district in Tianjin, which is helpful and meaningful for the groundwater resource management.

Acknowledgments The authors acknowledge financial support provided by the Tianjin Research Program of Application Foundation and Advanced Technology (Grant No. 14ZCZDSF00002). The authors also gratefully acknowledge Edanz Editing for editing the paper and providing valuable comments.

Appendix See Figs. 8 and 9.

123

1131 Page 12 of 13

Environ Earth Sci (2016) 75:1131 4

3 -1

0

2000

4000

6000

8000

10000

-2.0

-26

Groundwater levels/m

-1.5

Groundwater levels/m

b

Groundwater Abstraction/10 m a

Aquifer1

-2.5 -3.0 -3.5 -4.0

0

3 -1

Groundwater Abstraction/10 m a 500 1000 1500 2000

c

-28

Aquifer2

-30 -32 -34

-1

-41 -42

Aquifer3

-43 -44

-46

4

-100 0

3 -1

Groundwater Abstraction/10 m a 100 200 300 400 500 600 700 800

e -61

4

0

3 -1

f

Groundwater Abstraction/10 m a 100 200 300 400

-65.5

Aquifer4

-52 -54 -56 -58

-62

Groundwater levels/m

-50

Groundwater levels/m

Groundwater levels/m

3

Groundwater Abstraction/10 m a 500 1000 1500 2000 2500 3000

-45

-36

-4.5

d

4

0

Groundwater levels/m

4

a

Aquifer5

-63 -64 -65

4

3 -1

Groundwater Abstraction/10 m a -50

0

50

100 150 200 250 300 350

-66.0 Aquifer6

-66.5 -67.0 -67.5 -68.0 -68.5

-60 -66

-69.0

Fig. 8 Graphs of groundwater abstraction versus predicted groundwater levels

4000

a Groundwater abstraction/104m3a-1

Groundwater abstraction/10 4m3a-1

25000

20000

Aquifer1

15000

10000

5000

0

b Aquifer2 Aquifer3 Aquifer4 Aquifer5 Aquifer6

3500 3000 2500 2000 1500 1000 500 0 -500

-10

-5

0

5

-10

10

-5

0

5

10

Land subsidence/mm

Land subsidence/mm Fig. 9 Graphs of groundwater abstraction versus predicted land subsidence

References Ajay S (2013) Groundwater modeling for the assessment of water management alternatives. J Hydrol 481:220–229 Chen XX, Luo ZJ, Zhou SL (2014) Influences of soil hydraulic and mechanical parameters on land subsidence and ground fissures caused by groundwater abstraction. J Hydrodyn 26(1):155–164. doi:10.1016/S1001-6058(14)60018-4 Gatot ES, Koichi Y, Tsuyoshi I (2013) Modeling groundwater level fluctuation in the tropical peatland areas under the effect of El Nino. Procedia Environ Sci 17:119–128

123

´ O, Mario Jesu´s PM, Martı´n HM, Burbey ThomasJ, Norma GC, Jose´ A EZ, Alfredo SP (2013) Land subsidence and ground failure associated to groundwater abstraction in the Aguascalientes Valley, Me´xico. Eng Geol 164:172–186 Leonard FK, Eloise K (2005) Groundwater depletion: a global problem. Hydrogeol J 13:317–320 Li WY, Cui YL, Su C (2012) An integrated numerical groundwater and land subsidence model of Tianjin. J Jinlin Univ (Earth Sci Ed) 42(3):805–813 (in Chinese) Li FW, Feng P, Zhang W, Zhang T (2013) An integrated groundwater management mode based on control indexes of groundwater

Environ Earth Sci (2016) 75:1131 amount and level. Water Resour Manag 27:3273–3292. doi:10. 1007/s11269-013-0346-8 Li WY, Zhang W, Zhao TY, Li H, Li Z, Su DM, Wei JS, Zhang ZQ, Tong Y, Shi JG, Wang ZQ (2014) Report on the controlling groundwater levels of Tianjin, China. Hydrology and Water Resources Survey and Management Center of Tianjin (in Chinese) Maheswaran R, Rakesh K (2013) Long term forecasting of groundwater levels with evidence of non-stationary and nonlinear characteristics. Comput Geosci 52:422–436 Shi J, Wang Z et al (2011) Assessment of deep groundwater overexploitation in the North China Plain. Geosci Front 2(4):593–598 Singh VP, Woolhiser DA (2002) Mathematical modeling of watershed hydrology. J Hydrol Eng 7(4):270–292 Song YX, Fang Y, Qian H, Zhang XD (2011) Research and application of groundwater numerical simulation-a case study in Balasu Water Source. Procedia Environ Sci 8:146–152 Uma S (2012) Impacts of reservoir on groundwater level and quality in a saline area, Nakhon Panom Province, Thailand. APCBEE Procedia 4:16–21 Varouchakis EA, Hristopulos DT (2013) Improvement of groundwater level prediction in sparsely gauged basins using physical laws

Page 13 of 13

1131

and local geographic features as auxiliary variables. Adv Water Resour 52:34–49 Wu Q, Zheng XX, Xu H (2003) Relative sea-level rising and its control strategy in coastal regions of China in the 21st century. Sci China (D) 46(1):74–83 Yan XJ, Zhou YP, Zhang W, Wang ZQ (2012) Study on the management modes of groundwater level-quality two indicators in abstraction. J Hebei Univ Technol 41(2):65–68 (in Chinese) Yang FG, Cao SY, Liu XN, Yang KJ (2008) Design of groundwater level monitoring network with ordinary kring. J Hydrodyn 20(3):339–346 Yang ZP, Lu WX, Long YQ, Li P (2009) Application and comparison of two prediction models for groundwater levels: a case study in Western Jilin Province, China. J Arid Environ 73:487–492 Yang QC, Liang J, Yang ZP (2012) Numerical modeling of groundwater flow in Daxing (Beijing), China. Energy Procedia 14:1671–1676 Zhang FZ, Wei Y, Qian H, Zhang XD (2011) Numerical simulation of the groundwater in Bulang River-Red Stone Bridge Water Source. Procedia Environ Sci 8:140–145

123