Regional land subsidence simulation in Su-Xi-Chang area and Shanghai City, China
Introduction
Land subsidence caused by excessive groundwater withdrawal can be explained through the principle of the effective stress (Poland and Davis, 1969, Galloway et al., 1999). In the recent 30 years, it has been extensively investigated quantitatively and qualitatively by many previous researchers (Gambolati and Freeze, 1973, Gambolati et al., 1974, Helm, 1975, Helm, 1976, Neuman et al., 1982, Bravo et al., 1991, Gambolati et al., 1991, Shearer, 1998, Larson et al., 2001).
The Su-Xi-Chang area (including Suzhou, Wuxi and Changzhou cities) and Shanghai City are located in the Southern Yangtze Delta in the eastern part of China. The area is about 12,000 km2 and 5000 km2 for the Su-Xi-Chang area and Shanghai City, respectively. It is an area of intensive groundwater development for domestic and industrial uses. Groundwater has been extracted for over 100 years in the area, which resulted in severe land subsidence. Currently, excessive exploitation of groundwater across the provincial boundary forms a huge regional cone of depression. Consequently, the land subsidence cone is also regional, centering in the downtowns of Shanghai and Su-Xi-Chang areas.
Shanghai is the first city in which land subsidence was found and reported in 1921 and its effect and damage were the greatest. In 2002, the maximum cumulative subsidence of Shanghai and Su-Xi-Chang area were 2.63 m and 2.00 m, respectively (Sun, 2002). Merely in Su-Xi-Chang, the area with subsidence larger than 0.2 m has reached 5000 km2 in 2000 (NCCGS et al., 2003). With the regional scale of the groundwater exploitation, the existence of cone of depression and land subsidence at present, thus it is necessary to break through the provincial boundary for studying land subsidence.
An integrated numerical groundwater and land subsidence model for land subsidence simulation is proposed. In the previously developed land subsidence models, either the study area was usually tens to hundreds of square kilometers, the simulation involved only one or two hydrostratigraphic units, or the same soil deformation feature was assumed throughout the study area. The model in this paper deals with the regional land subsidence in tens of thousands of square kilometers and involves several deformation features of the whole aquifer system. This model is improved as to the previous ones (Gambolati et al., 1974, Helm, 1975, Helm, 1976, Larson et al., 2001, Chen et al., 2003).
Some new issues are involved in this model. The first and the most important issue is that not only the complicated geological conditions, but also the complex deformation features of different hydrostratigraphic units are considered due to the large study area. An identical unit may also present different deformation characteristics, such as elasticity, elasto-plasticity, and visco-elasto-plasticity, at different sites of the cone of depression or in different periods (Ye et al., 2005, Shi et al., 2007, Shi et al., in press, Zhang et al., 2007). Different subsidence models are employed according to the deformation features in different locations and over different periods when establishing the regional land subsidence model. Many previous land subsidence models assumed the same deformation models for all the hydrostratigraphic units, such as subsidence model based on elastic deformation (Chen et al., 2003), subsidence model based on visco-elastic deformation (Corapcioglu and Brutsaert, 1977), subsidence models based on elasto-plastic deformation (Gambolati and Freeze, 1973, Gambolati et al., 1974, Helm, 1975, Helm, 1976, Neuman et al., 1982, Gambolati et al., 1991), and subsidence model based on visco-elasto-plastic deformation (Gu and Ran, 2000).
Secondly, the parameters in most of the previous subsidence models are constant (Gambolati and Freeze, 1973, Gambolati et al., 1974, Helm, 1975, Gambolati et al., 1991, Li et al., 2000). In fact, the hydrogeological parameters, such as the hydraulic conductivity and specific storage, changed along with the compression of hydrostratigraphic units. Rudolph and Frind (1991) investigated the transient hydraulic behavior of highly compressible aquitards through both numerical analysis and field studies. They found the hydraulic parameters of an aquitard varied during consolidation. Chen et al. (2003) also found the reduction in the hydraulic properties using the numerical model for Suzhou City. In order to represent groundwater flow under the condition of land subsidence, the flow model should have variable parameters.
Thirdly, coupling the flow model and the subsidence model to represent the impact of subsidence to the hydrological parameters is a new issue. The ground water flow model and subsidence model should be solved together in theory, which is called fully coupled land subsidence model, such as Biot model. However, the fully coupled model is hard to be used because of too many parameters and the complication when considering non-elastic deformation. Hence the so called ‘Two Steps’ model or ‘uncoupled model’ was developed, which calculated the hydraulic heads firstly and then calculates the deformation (Gambolati and Freeze, 1973). Since the deformation of hydrostratigraphic units and the changes in hydraulic heads actually occur simultaneously, the ‘two steps’ model cannot accurately describe the physical mechanism of land subsidence. Another coupled method should be applied to overcome the disadvantages of ‘fully coupled’ model and ‘two steps model’. In this research, a ‘coupled two step’ model (Chen et al., 2003) is used. The flow model and the land subsidence model are coupled by changing parameters as functions of the hydraulic heads and deformation, and then the hydraulic heads and deformation are calculated by two steps in the model.
Last but not least, for the regional land subsidence simulation, the horizontal scale is much more than the vertical scale. The vertical thickness of each hydrostratigraphic unit is generally not larger than tens meters and the thinnest is only a few meters, while the layer extends tens of thousands of square kilometers in the horizontal scale. If we discretize the aquifer units by finer mesh according to the vertical scale, a surprisingly large CPU time and computational memory are required in the conventional finite element method. Otherwise, if we discretize the aquifer units by coarser mesh according to the horizontal scale, the very thin and deformed elements should be faced, which would increase the errors of results.
The main objectives of this paper are to introduce the development of regional land subsidence model based on complex different deformation features and to calibrate the model using observed data including compression of individual strata from groups of extensometers and groundwater levels from observation wells. The objectives can be accomplished through the following three steps. (1) Study on the ground water flow models based on elastic, elasto-plastic, visco-elastic and visco-elasto-plastic constitutive laws. (2) Study on the subsidence model based on elastic, elasto-plastic, visco-elastic and visco-elasto-plastic constitutive laws. (3) Apply the multiscale finite element method (MsFEM) to solve the land subsidence model.
Section snippets
The geographic location
The study area is located in the Taihu Lake plain in the south of the Yangtze Delta. It borders Mogan Moutain and Maodong Plain in the west, the East Sea and the Yellow Sea in the east, the Yangtze River in the north, and Zhejiang Province in the south, as illustrated in Fig. 1. The study area includes Suzhou, Wuxi and Changzhou cities in Jiangsu Province (except Taihu Lake) and Shanghai City (except Chongming, Changxing and Hengsha islands). The elevation decreases from the west to the east.
The hydrogeological condition
Groundwater exploitation and resultant change in piezometric level
Groundwater extraction from the aquifer system below the Southern Yangtze Delta began in 1860 and became extensive in the 1950s. In the beginning, several separate cones of depression were centered in the areas of pumping overdraft. Then the cones extended and exceeded the boundaries of provinces when the groundwater pumpage increased. Currently, they have joined together and formed a large and regional coalesced cone of depression in the Southern Yangtze Delta. The groundwater pumping
The hydrogeological conceptual model
The simulated aquifer system is loose Quaternary deposit. Because of the influence from the basal structures and paleotropography, the thickness of the Quaternary deposits increases from the western part to the southeastern part. It is thinner than 100 m in Changzhou City and greater than 300 m in the north of Shanghai. There are more than 250 observation wells in the study area. They provide fairly detailed information on the spatial distribution of each hydrostratigraphic unit and are an
The flow model based on different constitutive laws
The general expression of flow governing equation (Gambolati and Freeze, 1973) is:where H is hydraulic head.
- K
is hydraulic conductivity.
- γ
is volume weight of water.
- β
is volume compressibility of water.
- ϕ
is porosity.
- ε
is the vertical strain.
The concrete expression of second term in the right hand of Eq. (1) is different in condition of different constitutive laws. The simplest constitutive law is elastic. According to elastic stress–strain relationship and the effective stress
Model calibration and verification
The solution of a regional land subsidence model is one of the difficulties in land subsidence simulation. The horizontal scale of the study area is much larger than the vertical scale and the thickness distribution of the aquifer and aquitard are non-uniform. Some hydrostratigraphic units are very thin. If the traditional finite element method (FEM) were adopted, the fine discretization should be used to avoid the distorted elements. It would cost much computer resource and a very long
Conclusion
Based on the regional subsidence simulation in the Su-Xi-Chang area and Shanghai City, the following conclusions can be obtained:
- 1.
Both the spatial heterogeneity of the hydrogeological units and the temporal and spatial groundwater level variations determine the complexity of the deformation features in the regional land subsidence. The modified Merchant model can describe visco-elasto-plastic deformation of different hydrostratigraphic units. The results indicate that modified model has
Acknowledgments
The authors would like to express appreciation to the editors and two anonymous reviewers for their valuable comments and suggestions. The paper is financially supported by the National Nature Science Foundation of China grants No. 40335045, 40702037 and 40725010.
References (36)
- et al.
A multiscale finite element method for elliptic problems in composite materials and porous media
Journal of computational physics
(1997) - et al.
Prediction of optimal safe ground water yield and land subsidence in the Los Banos-Kettleman City area, California, using a calibrated numerical simulation model
Journal of Hydrology
(2001) - et al.
Deliminitation of ground failure zones due to land subsidence using gravity data and finite element modeling in the Queretaro valley, Mexico
Engineering Geology
(2006) A numerical model to calculate land subsidence applied at Hangu in China
Engineering Geology
(1998)Stress and strain analysis of the fourth artesian aquifer in Shanghai
The Hydrology and Engineering Geology
(2002)- et al.
Characteristics of aquifer system deformation in the Southern Yangtze Delta, China
Engineering Geology
(2007) Dynamics of fluids in porous media[M]
(1972)Land subsidence and cracking due to groundwater depletion
Ground Water
(1977)- et al.
Analysis of groundwater level fluctuations and borehole extensometer data from the Baytown Area, Houston, TX
- et al.
Land subsidence caused by groundwater exploitation in Suzhou City, China
Hydrogeology Journal
(2003)
Viscoelastic aquifer model applied to subsidence due to pumping
Water Resources Research
Soil Mechanics[M]
Land subsidence in the United States: U. S. Geological Survey Circular 1182
Mathematical simulation of the subsidence of Venice: 1. Theory
Water Resources Research
Mathematical simulation of the subsidence of Venice: 2. Results
Water Resources Research
Mathematical simulation of the subsidence of Ravenna
Water Resources Research
A 3-D coupled model with consideration of rheological properties
One-dimensional simulation of aquifer system compaction near Pixley, California: 1. Constant parameters
Water Resources Research
Cited by (103)
Impacts of aquifer heterogeneity on subsidence resulting from engineering dewatering in the Lower Yangtze River plains
2024, Journal of Hydrology: Regional StudiesGoverning factors for actively heated fiber optics based thermal response tests
2023, Renewable EnergyThermo-hydro-mechanical (THM) coupled simulation of the land subsidence due to aquifer thermal energy storage (ATES) system in soft soils
2023, Journal of Rock Mechanics and Geotechnical EngineeringViscoelastic-viscoplastic damage analysis of transversely isotropic soft soils
2022, Engineering Geology