Environ Earth Sci DOI 10.1007/s12665-015-4131-2

ORIGINAL ARTICLE

Groundwater-derived land subsidence in the North China Plain Haipeng Guo • Zuochen Zhang • Guoming Cheng Wenpeng Li • Tiefeng Li • Jiu Jimmy Jiao

•

Received: 27 August 2014 / Accepted: 29 January 2015 Ó Springer-Verlag Berlin Heidelberg 2015

Abstract Land subsidence usually bears strong relationship to abstraction of underground fluid. Land subsidence occurs commonly in the North China Plain (NCP) and has become a major environmental factor hindering regional sustainable development. This paper focuses on issues associated with mechanism of land subsidence in the NCP. The analysis shows that multi-layer aquifer systems with deep confined aquifers and the thick normally consolidated or unconsolidated compressible clay layers are the key of geological and hydrogeological conditions favorable for the development of land subsidence in the NCP. Groundwater withdrawal results in an increase in the distribution of effective stress within the strata and the compression of the aquifers and the confining layers, and then triggers land subsidence. In the middle-east plain of the NCP, the land subsidence volume approximately represents the amount of water released from compression of deep aquifers and aquitards in that land subsidence is primarily caused by excessive groundwater withdrawal in deep aquifer system. The percentage of water released from compression of aquifers and aquitards in deep groundwater abstraction is

H. Guo (&)  Z. Zhang  G. Cheng China Institute of Geo-Environment Monitoring, Beijing 100081, China e-mail: [email protected] W. Li Center for Hydrogeology and Environmental Geology Survey, Baoding 071051, Hebei, China T. Li China Geological Survey, Beijing 100037, China J. J. Jiao Department of Earth Sciences, University of Hong Kong, Hong Kong, China

significant but distinct in Cangzhou City and in the whole middle-east plain. This is due to the difference in local lithological structure and recharge and discharge conditions of deep groundwater system. The hysteresis of land subsidence is also discussed in typical areas and the results reveal that the time for completing the primary consolidation ranges from less than one year to tens of years, and that the rate of secondary compression tends to increase with the moisture content. Keywords Land subsidence  Mechanism  Groundwater withdrawal  The North China Plain

Introduction Land subsidence can be defined as the sinking of the ground surface with respect to surrounding terrain or sea level. The causes of land subsidence include natural factors and human activities. The former includes, for example, tectonic motion or sea level rise, and the latter includes withdrawal of groundwater, oil, and gas (Abidin et al. 2013; Zeitoun and Wakshal 2013; Galgaro et al. 2014; Papadaki 2014). At present, there have been 50 countries where land subsidence occurs, including USA, Japan, Mexico, Italy, Thailand, and China (Hu et al. 2004). Anthropogenically induced land subsidence and its subsequent impact on geological environment have been observed in many countries of the world (Adrian et al. 1999; Chatterjee et al. 2006; Phien-wej et al. 2006). Land subsidence has been observed for many years and explained by previous researchers who investigated this phenomenon by conceptual discussion, field and laboratory observations. Chen et al. (2003) conducted a three-dimensional finite difference numerical model representing the multi-layered

123

Environ Earth Sci

aquifer system to study the land subsidence in response to groundwater abstraction in Suzhou City, Jiangsu Province, China. They found that the area with maximum drawdown was not necessarily the area with maximum ground settlement due to the occurrence of the soft mud layer. They also concluded that a reallocation in pumping rates based on the spatial distribution of the thick mud layer could greatly reduce land subsidence. Chen et al. (2010) examined the elevation changes on ground surface in the Choshuichi Alluvial Fan of central Taiwan and described the land subsidence phenomena quantitatively. They concluded that the highest correlations existed between elevation changes and groundwater level variations in the land subsidence area. Tsai and Jang (2014) developed a poroelastic consolidation model to investigate the combined deformation effect of porosity variation on soil consolidation caused by groundwater table decline. They concluded that a reliable analysis of soil consolidation must simultaneously account for the variations in hydraulic conductivity, Young’s modulus, and body force. Hernandez-Marin et al. (2014) conducted evaluation and analysis of surface deformation in west Chapala basin, central Mexico. From April to November in 2012, the maximum accumulated vertical displacement in the basin was calculated as approximately 7.6 cm (0.89 cm/month), which was relatively high compared with other subsidence areas around the world. The investigation revealed that the likely key factor leading to this large rate of deformation was the likelihood of the presence of high compressibility volcaniclacustrine sediments combined with large groundwater withdrawal rates. Over the last decades, Interferometric Synthetic Aperture Radar (InSAR) and global positioning system (GPS) have become very useful tools for accurately measuring the spatial and temporal evolution of land subsidence over broad areas (Stramondo et al. 2008; Khan et al. 2013; Tomas et al. 2014). Land subsidence has been commonly observed in the NCP (Fig. 1), and has become the main factor that impacts regional sustainable economic and social development. The NCP has more than six cities and municipalities where disastrous land subsidence has occurred, for instance, in Tianjin, Beijing, Cangzhou, Hengshui, Tangshan, Dezhou, and others. The first occurrence of land subsidence in the NCP took place in Tianjin in the 1920s, and by the 1960s it became a severe disaster in this city. Since the 1980s, the land affected by subsidence in NCP extended from cities to rural areas. Ground cracks or fissures are observed in areas where large differential subsidence occurs. Land subsidence in Beijing was first noticed in 1935, and the maximum subsidence reached up to 1.1 m in Beijing plain during the period 1955–2007 (Zhang et al. 2014). The subsidence in Cangzhou and Dacheng covers 9,363 km2 and the accumulative subsidence at the

123

Fig. 1 Sketch map of the NCP showing distribution of the major land subsidence areas

Fig. 2 Changes of accumulative subsidence in several subsidence centers in the NCP

subsidence center exceeded 2 m (Xue et al. 2005). Land subsidence has also been observed in Dezhou and Binzhou of Shandong Province. The common characteristics of land subsidence in the NCP are slow, accumulative, and irreversible. Figure 2 demonstrates temporal changes of the accumulative subsidence in several subsidence centers from 1975 to 1995, with the maximum accumulative subsidence of more than 800 mm during these twenty years.

Environ Earth Sci

Land subsidence creates a permanent damage to the land surface and may lead to many problems such as localized flooding (Zuo et al. 1993), decline of groundwater storage capacity (Rudolph and Frind 1991), and failure of well casings and changes in channel gradient (Holzer 1989). Many of these problems have also been observed in the NCP. Wang et al. (2013) conducted soil compression and rebound tests to analyze the unloading and rebound regularity of deep soil on the fringes of three typical land subsidence regions in the NCP. The results revealed that only 14.7 % of the land subsidence deformation in the study areas can be restored and earlier land subsidence treatment can result in better effects. The economic loss due to land subsidence in the NCP has exceeded 200 billion yuan (Ye and et al. 2006). The percentage of indirect loss in the overall loss is much greater than the direct loss. Evaluation results revealed that the economic loss of land subsidence in Tianjin was 119 billion RMB as of 2003, and the ratio of direct loss to indirect loss was 1:11. The general aim of this paper is to present a brief summary on mechanism of land subsidence in the NCP. At first, the paper will give a description on the geological and hydrogeological conditions of the NCP and then discuss the influencing factors of land subsidence, including local geological environment and human factors. A GIS spatial analysis method was employed to study the quantitative relationship between land subsidence and the deep groundwater exploitation. Finally, a general description of the characteristics of primary consolidation and secondary compression is presented in typical areas of the NCP, which is based on previous documents and lab results.

Physical geography, hydrologic characteristics, and regional hydrogeology The NCP is located in the eastern part of China (Fig. 1) and has a total area of about 140,000 km2. This region comprises the plain area of Beijing Municipality, Tianjin Municipality and Hebei Province, and the plain area of Henan and Shandong provinces to the north of the Yellow River. The NCP is a typical plain landscape and the elevation is less than 100 m. The topography inclines from north, west, and southwest to the Bohai Bay. The slope gradient is 1–0.2 % in the piedmont plain and decreases to 0.1 %– 0.2 % in the coastal plain. The NCP belongs to continental arid climate, and the perennial annual precipitation ranges from 500 to 600 mm. The evaporation is about 1,100–2,000 mm, and reaches a maximum from June to September. Most rivers in NCP have become seasonal or even dry due to limited precipitation and the reservoirs construction in the upstream reaches.

The NCP comprises three distinct hydrogeological settings within the Quaternary sediments (Fig. 1): the piedmont plain and associated major alluvial fans, the alluvial plains with many abandoned river channels, and the coastal plain strip around the margin of the Bohai Sea (Foster et al. 2004). The main stratigraphy of this region is composed of unconsolidated Quaternary sediments with the thickness from 200 m to over 600 m, consisting of unconsolidated pebble, gravel, sand, silt, and clay. Generally, the soil particles become finer from the piedmont area to the coastal area. The major groundwater type in the NCP is pore water in loose and unconsolidated sediments of Quaternary. From the piedmont plain to alluvial and coastal plains in the middle-east, the aquifer systems usually change from a single aquifer of sandy gravel to multiple aquifers of sand separated by silt or clay layers. The stratigraphy in the middle-east plain can be divided into four aquifer groups. The depth of the bottom of the first aquifer group is generally less than 50 m, and that of the second aquifer group ranges from 120 to 210 m. Here groundwater in the first and the second aquifer groups is defined as shallow groundwater, which has unconfined or semi-confined hydraulic characteristics. The lower boundary of the third aquifer group is located between 250 and 310 m below the ground, and that of the fourth aquifer group is the base of the Quaternary strata. Groundwater in the third and the fourth aquifer groups is confined and defined as deep groundwater. Areally extensive and thick aquitards exist in the Quaternary strata.

Influencing factors of land subsidence As the factors affecting land subsidence in the NCP are very complicated, the area of influence and the development rate of land subsidence are very different with time and space. Some main factors that may favor the occurrence of land subsidence include geological conditions, excessive abstraction of groundwater, oil, and gas, tectonism, and natural settlement of under-consolidated soils. Geologically, the soil underlying the NCP is composed mainly of a multi-layered system of Quaternary sediments, consisting of coarse-grained materials, mainly unconsolidated sands as well as fine-grained deposits, which are predominantly silts and clays that have medium to high compressibility. These alternating sequences of sand and compressible clay layers are very susceptible to deformation (Fig. 3). The subsidence due to neotectonic movement is about 1–5 mm/year in the NCP (Hu et al. 1993; Yang et al. 2001). Subsidence due to natural compaction of unconsolidated soil is about 1 mm/year in coastal areas in Tianjin (Tianjin Institute of Geo-Environmental Monitoring 2013).

123

Environ Earth Sci

Fig. 3 Hydrogeological cross section A–A0 (the location shown in Fig. 1)

Sea level rise is not a direct cause for land subsidence, but is an important influencing factor augmenting the damage of land subsidence in the social-economic system. Based on field real-time monitoring and data analysis, the sea level in Tanggu area of Tianjin City shows an fluctuated upward trend with an average rising rate of about 3.3 mm/a from 1951 to 2012, and to 7.0 mm/a from 1980 to 2012. Similar sea level rising rates are observed in Huanghua area of Cangzhou City (Fig. 4), where the rising rates are 2.7 and 5.2 mm/a in the periods of 1951–2012 and 1980–2012, respectively. Note that the datum of the measured sea levels is the ground surface. In the period 1994–2007, the global sea level rising rate was calculated as 2 mm/year, which is equal to the sum of the average subsiding rate of -8 mm/year in Tanggu and the rising rate of 10 mm/year for the local sea level. The data of land subsidence were obtained by leveling measurement. As discussed above, tectonism and natural settlement could not be the major causes leading to land subsidence. Excessive withdrawal of underground fluids such as groundwater, oil, and gas may be the important driving condition for land subsidence, as has been observed in most land subsidence areas in the NCP. The types of fluid abstraction for major areas of land subsidence in distinct topography and geomorphology settings are listed in Table 1. Shallow groundwater abstraction may be the main driving condition for land subsidence in the piedmont area and associated alluvial fans (e.g., Beijing and Handan). In the alluvial and coastal plains, land subsidence may be commonly caused by withdrawal of deep

123

Fig. 4 Distribution of the depression cones of shallow and deep groundwater, and major land subsidence area in the NCP

groundwater, but in few areas where large volumes of shallow groundwater, oil, and gas are extracted, land subsidence is also observed.

Environ Earth Sci Table 1 The types of fluid extraction in the NCP Topography and geomorphology

Major areas of land subsidence

Type of fluid extraction

Piedmont area and associated alluvial fans

Beijing, Handan

Shallow groundwater

Baoding

Shallow and deep groundwater

Alluvial plain with many abandoned channels

Suning Dongguang, Hengshui, Nangong, Pingxiang, Langfang Wuqing Xiqing, Jinghai, Urban area of Cangzhou, Dezhou, Binzhou

Shallow groundwater Deep groundwater

Renqiu, Hejian, Xian County

Deep groundwater, oil, and gas

Most coastal areas of Tianjin, Some coastal areas of Cangzhou

Deep groundwater

Coastal plain

Land subsidence in the NCP may be caused primarily by extensive pumping of groundwater (Table 1), which leads to drop of groundwater levels and therefore an increase in the distribution of effective stresses within the strata. Most cities in NCP have undergone overextraction of groundwater due to the demand for water associated with municipal, industrial and agricultural needs, which has resulted in long-term depressurization of the aquifer system and subsequent local subsidence bowls. Large-scale groundwater development in Beijing plain has supported rapid growth of economic and urban population in last 30 years, while overexploitation of groundwater has caused serious land subsidence in this region. Land subsidence in Tianjin was first noticed in 1923. The thicknesses of the Quaternary deposit range from several hundred meters to 1,000 m in the plain area. Land subsidence developed slowly from 1923 to 1958, and evolved rapidly from 1959 to 1985 when the maximal exploitation of groundwater reached 120 9 106 m3/a and the maximal rate of subsidence exceeded 100 mm/a (Xue et al. 2005). Due to the efforts to reduce groundwater abstraction, the annual quantity of extracted groundwater was limited in 15 - 20 9 106 m3/a in urban district and the subsiding rate decreased to 10–15 mm/a from 1986 to 2002 (Xue et al. 2005; Cui et al. 2002). Excessive groundwater withdrawal in Hebei plain has resulted in nine huge groundwater depression cones and therefore nine main subsidence bowls. Shallow groundwater depression cones are mostly distributed in the piedmont area and associated alluvial fans (Fig. 4), and so far the total area of the depression cones has exceeded 9700 km2. Most of these shallow depression cones formed in the 1970s, and expanded greatly in the 1980s. Groundwater levels have recovered in some depression cones in the 21st century. Figure 5 shows variations of water level depths in the depression cone centers in Baoding, Shijiazhuang, Ningborong, and Handan. Overall the depth increases in the 1970s, 1980s, and 1990s, but decreases in the last 7 years except in the depression cone

Fig. 5 Change of water level depths in centers of several major shallow groundwater depression cones in the NCP

of Ningborong. The largest measured water level decline occurs in the depression cone of Ningborong located southwest of the NCP, where a maximum accumulative water level drop of 60 m was measured in the period 1975–2010. Deep groundwater depression cones are primarily distributed in the alluvial and coastal plains, showing three great depression cone groups, namely Langfang-Tianjin, Cangzhou, and Hengshui-Dezhou (Fig. 4). Figure 6 shows variations of water level depths in depression cone centers of Langfang, Dacheng, Cangzhou, and Jizaoheng. The level depth fluctuates but overall has an increasing trend from the 1970s to the 1990s, which follows a similar trend as observed in shallow groundwater depression cones (Fig. 5). The groundwater levels tend to recover in the last few years. The most significant water level drop is located in the depression cone of Dacheng, with a maximum accumulative water level drop of 75 m. Groundwater overdraft in deep confined aquifers may have become the major triggering factor to induce land subsidence in middle-east plain of the NCP, since land subsidence is commonly observed in the areas where depression cones of deep groundwater occur (Fig. 4).

123

Environ Earth Sci

Fig. 6 Change of water level depths in centers of several major deep groundwater depression cones in the NCP

In summary, the effects of neotectonic movement and natural settlement of under-consolidated soils are minor, and the conditions highly favorable for land subsidence in the NCP include the local geological condition and the dynamic conditions. Based on local geological conditions especially hydrogeological conditions, it can be found that, loose multi-layer aquifer systems with deep confined aquifers and thick normally consolidated or unconsolidated and compressible and thick clay layers are most vulnerable for land subsidence. From the perspective of the alteration of geo-stress within soil layers, the significant and continuous decrease in the water level is a necessary prerequisite of the land subsidence.

Land subsidence and deep groundwater abstraction The correlation between land subsidence and its impacting factors can be described by the Pearson correlation coefficient. Pearson’s correlation coefficient between two variables is defined as the covariance of the two variables divided by the product of their standard deviations. This coefficient could be used to measure the strength of a linear association between two variables X and Y. An alternative formula for the sample Pearson correlation coefficient is as follows: P P P xi yi  ð xi Þð yi Þ=n ffi: r ¼ qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi ð1Þ P P P P ð x2i  ð xi Þ2 =nÞð y2i  ð yi Þ2 =nÞ In the case of normally consolidated clay, the consolidation settlement can be defined as (Craig 2004) SC ¼ mV Dr0 H0

ð2Þ

where SC is the consolidation settlement; mV is the coefficient of volume compressibility; Dr0 is the effective vertical stress increment, and H0 is the thickness of the

123

compressible layer. Equation (2) indicates that the consolidation settlement increases with mV, Dr0 , and H0. If the soil is in an over consolidated state, the same conclusion could be made, but the same stress increase would give a much lower settlement in that the over consolidated soil is less compressible. The effective vertical stress increment is primarily caused by groundwater abstraction in the NCP. In the piedmont plain and associated alluvial fans of the NCP (cross section I–I0 in Fig. 7a), the correlation coefficient between the accumulative land subsidence and the total decrease of shallow groundwater levels from 1975 to 2009 is a positive value of 0.59, suggesting that land subsidence tends to increase with the total drop of shallow groundwater levels. Hence, shallow groundwater abstraction may be the main factor to trigger the subsidence in the piedmont plain. The accumulative subsidence also has good positive relation with the thickness of Quaternary clay (r = 0.79) in this region. In the measurement period from 1975 to 2009, the average thickness of Quaternary clay and the average shallow groundwater level drop are 260 and 21 m, respectively, and the average land subsidence is 24 cm. Figure 4 shows that most serious subsidence areas in the NCP lie in the alluvial plain and in the coastal strip around the margin of the Bohai Bay, which coincide well with the distribution of depression cones of deep groundwater. In the alluvial and coastal plains shown in the cross section II–II0 (Fig. 7), the accumulative subsidence has poor relation with the total decrease of shallow groundwater levels from 1975 to 2009, but has clear positive relation with the total drop of deep groundwater levels (r = 0.64). This implies that the subsidence in these areas may be primarily derived by groundwater withdrawal in deep aquifers. The average deep groundwater level drop from 1975 to 2009 in the cross section II–II0 is 34 m, and the average land subsidence is 56 cm, both are much greater than those in cross section I– I0 . The positive correlation coefficient between land subsidence and the clay thickness of the third aquifer group is as great as 0.71 in that the cross section II–II0 traverses the subsidence area of Cangzhou, where groundwater overdraft in the third aquifer group contributes the main amount of settlement to the total subsidence on the ground. Hence, the big thickness of clay layers in the aquifer system of groundwater production makes up the geological conditions favorable for the development of land subsidence, and extensive pumping of deep groundwater has become the major triggering factor for land subsidence in the alluvial and coastal plains of the NCP. Theoretical analysis of relation between land subsidence and deep groundwater abstraction Terzaghi introduced the basic principle of effective stress that

Environ Earth Sci Fig. 7 Variations of land subsidence, groundwater levels, and thicknesses of clay in the strata in cross sections I–I0 (a) in the piedmont plain and II–II0 (b) from west to east (locations of the cross sections shown in Fig. 4)

r0 ¼ r  u

ð3Þ

where r0 is effective stress, r is the geostatic stress (total stress), and u is the pore pressure or hydrostatic stress. When no tectonic stresses are present, the geostatic stress (total stress) can be expressed as (Poland and Davis 1969; Zeitoun and Wakshal 2013) r ¼ rload þcm dm þcs ds

ð4Þ

cm ¼ cg ð1  nÞ þ nw cw

ð5Þ

cs ¼ cg  nðcs  cw Þ

ð6Þ

and the hydrostatic stress can be expressed as u ¼ ds c w

ð7Þ

where rload is the loading on the soil surface; cm is the unit weight of moist sediments above the water table; cg is the unit weight of sediments grains; cs is the unit weight of saturated sediments below the water table; cw is the unit weight of water; n is the porosity; nw is the moisture

content of sediments in the unsaturated zone, as a fraction of total volume; dm is the depth below land surface in the unsaturated-zone interval, land surface (z = 0) to the water table (z = zwt); and d0 is the depth of interest in the saturated zone. With Eqs. (3)–(7), the variation of effective stresses in response to water level changes in the aquifers can be summarized in Table 2. Assuming rload = 0, stress diagrams for a water level drop of 40 m in an unconfined aquifer and a confined aquifer are made with Eqs. (3)–(7), respectively (Fig. 8). Figure 8a demonstrates the relationships between geostatic, hydrostatic, and effective stresses and the changes in these stresses after a water table drop of 40 m in an unconfined aquifer when fluid-pressure equilibration is achieved. The geostatic stress decreases slightly below the depth of the original water table. The unconfined aquifer below the depth of the active water table, the confining unit, and the confined aquifer are affected equally by a 10.8 m decrease in geostatic stress. The hydrostatic stress

123

Environ Earth Sci

decreases by 40 m in the unconfined aquifer and remains unchanged in the confined aquifer. Therefore, the effective stress in the unconfined aquifer increases 29.2 m and that in the confined aquifer decreases by 10.8 m. Figure 8b illustrates change of geostatic, hydrostatic, and effective stresses as water head drops 40 m in the confined aquifer. The geostatic stress remains unchanged in the unconfined aquifer below the depth of the water table, the confining unit, and the confined aquifer. The hydrostatic stress is unchanged in the unconfined aquifer and decreases in the confined aquifer by 40 m. Thus, the effective stress in the unconfined aquifer remains unchanged and that in the confined aquifer increases by 40 m. Table 2 demonstrates that the effective stresses increases in the unconfined aquifer while decreases slightly in the confined aquifer when water table drops. As hydraulic head in the confined aquifer drops, the effective stresses remains unchanged in the unconfined aquifer, while increases in the confined aquifer. Note that the increase in effective stresses of the unconfined aquifer caused by water table drop is reduced by a factor (1-n-nw) compared with the effective stress increment in the

confined aquifer caused by an equivalent water level decline in the confined aquifer. Additionally, the water level drop in confined aquifers is much greater than that in unconfined aquifers when equivalent amount of groundwater is abstracted from the aquifer system. Moreover, recharge condition for confined aquifers is commonly much poorer than unconfined aquifers. As a result, groundwater abstraction can cause land subsidence much more easily in confined than unconfined aquifers. In the middle-east plain of the NCP (Fig. 4), groundwater in deep confined aquifers is excessively pumped, which has led to serious land subsidence. In the piedmont plain, however, groundwater is mainly pumped from shallow unconfined aquifers; thus, the effective stress increments in aquifers and aquitards are relatively low. In addition, coarse sediments that make up the piedmont plain have low compressibility, while fine clays and silts with organic matter in confined aquifers of the middle-east plain typically demonstrates high compressibility. Hence, land subsidence in the middle-east plain is much more severe than the piedmont plain in the NCP. Quantitative relationship between deep groundwater exploitation and land subsidence in the NCP

Table 2 Variation of effective stresses due to water level changes in the aquifers Water level change

Variation of effective stresses Unconfined aquifer

Confined aquifer

Dhwt

cw ð1  n þ nw ÞDhwt

cw ðn  nw ÞDhwt

Dhc

0

cw Dhc

Dhwt and Dhc denote water level changes in the unconfined and confined aquifers, respectively

The hydraulic head drop in the aquifer will eventually result in commensurate amount of head decline in the confining layer. If the solid and fluid are assumed to be incompressible, then, the volume of the fluid removed is equal to that of the subsidence in the confining layers (Narasimhan et al. 1984; Sun et al. 1998). The amount of deep groundwater exploitation in the NCP mainly consists of leakage from shallow aquifers, released water from

Fig. 8 Stress diagrams for a water table drop of 40 m in an unconfined aquifer (a), and for a head drop of 40 m in a confined aquifer (b), when cg = 2.6 9 104 N/m3, cw = 9.8 9 103 N/m3, n = 0.35, and nw = 0.08

123

Environ Earth Sci

compressibility of deep aquifers and aquitards, and lateral recharge. The accumulative subsidence volume represents the amount of released water from compressibility of aquifers and the aquitards. Some factors such as hysteresis effects of subsidence may be ignored in the short-term time series analysis of the relation between subsidence and groundwater abstraction; therefore, from the perspective of water resources evaluation, the long-term series of data should be used for calculation. With GIS spatial analysis method, subsidence data from field observations were used to construct subsidence contours to help study the quantitative relationship between deep groundwater abstraction and land subsidence. The middle-east plain of the NCP (Fig. 4), with an area of 7.6 9 104 km2, consisting of Hebei plain, Tianjin plain, and Dezhou region. The annual subsidence volume in this region is calculated as 1073 9 106 m3 until 2010 and the annual deep groundwater abstraction is 2,450 9 106– 2,675 9 106 m3/a. Therefore, the land subsidence volume accounts for 40.1–43.8 % of deep groundwater yield in the whole middle-east plain. Cangzhou, with an area of 1.34 9 104 km2, is located in eastern Hebei, immediately to the south of Tianjin, and near the coast of the Bohai Sea (Fig. 4). Cangzhou is located within the middle-east plain of the NCP, and bordering prefecture-level cities are Hengshui to the southwest, Baoding to the west, and Langfang to the north. The percentage distribution of cumulative subsidence area and volume for different ranges of cumulative subsidence is shown in Fig. 9. This percentage denotes the ratio of subsidence area (or volume) in a range of cumulative subsidence to the total subsidence area (or volume). In the last forty years, the percentage of the subsidence area with the cumulative subsidence greater than 140 cm has amounted to 0.9 %. Note that the percentage of cumulative subsidence area is greater than cumulative subsidence volume as shown in Fig. 9, implying that the subsidence bowls have a large size up top and a small on bottom. The annual average subsidence volume in Cangzhou was calculated as 198 9 106 m3 during the measurement period 1970–2008. Results indicate that the total amount of deep groundwater exploitation over the period 1972–2008 is 12,734 9 106 m3 and translates to a yearly average abstraction rate of about 344 9 106 m3/a. Then the ratio of the accumulative land subsidence volume to the amount of deep groundwater exploitation can be calculated as 57.6 %,which is much greater than the ratio for the whole middle-east plain. This is probably attributed to the differences of lithologic structure and recharge and discharge conditions in distinct areas. Land subsidence is induced when various factors such as lithological structure and groundwater abstraction combine favorably. The lithological structure determines susceptible

Fig. 9 Percentage of cumulative subsidence area and volume in Cangzhou City during 1970–2008

degree, and the degree of land subsidence is dependent on the degree of groundwater development. The hydrogeological conditions have important impact on the development of subsidence in Cangzhou area. The deformation of different soil layers is due not only to its compressibility but also to its thickness. Compared with areas in margins of alluvial-proluvium fans such as Hengshui in Hebei (Fig. 1), the lithological conditions in Cangzhou are more vulnerable to produce subsidence. The thickness of clay in subsidence center of Cangzhou amounts to 214 m, 84 % of which is located in the major groundwater production ranges. Lateral recharge from surrounding areas is poor in that serious groundwater depression cones have also formed in these surrounding areas. Thus, released water from compressibility of the deep aquifers and aquitards may be the main source of deep groundwater exploitation and belongs to the mining of the storage. By taking the whole middle-east plain as the research object, however, relatively good lateral recharge from the piedmont plain occurs, and in some areas close to the piedmont plain, large leakage from shallow aquifers may be triggered as groundwater is excessively abstracted. Therefore, the ratio of the land subsidence volume to the amount of deep groundwater withdrawal in Cangzhou is much larger than the whole middle-east plain.

Hysteresis of land subsidence The soil layers in the subsidence areas of the NCP contain a large fraction of high-plastic clay and silty clay, which favorably lead to permanent land subsidence. The observed data have proved characteristic of delay for land subsidence, suggesting that the subsidence still continues even though the groundwater level decline stops. Figure 10 shows the relation between the water level depth of the third aquifer group and the average subsidence rate on the ground in a major subsidence center of Cangzhou, where land subsidence is mainly triggered by groundwater

123

Environ Earth Sci Fig. 10 Relation between water level depth of the third aquifer group and average subsidence rate in a major subsidence center of Cangzhou (according to Xing et al. 2004)

overdraft in the third aquifer group. The average subsidence rate tends to increase significantly since 1980, but the rate of water level drop declines greatly after this time point, suggesting strong hysteresis of land subsidence. Hence, the water level depth of 70 m in 1980 could be taken as a critical water level for inducing significant subsidence. Land subsidence in this region initiates as minor deformation on the ground and eventually evolves into an evident subsidence bowl as is observed today. Primary consolidation When a confined layer is bounded above and below by aquifers, the degree of consolidation and the expected time to accomplish the consolidation can be obtained from Domenico and Schwartz (1998): t tKv ¼  T ðH=2Þ2 S0s

ð8Þ

where t is the time required to accomplish the consolidation, T* is termed the time constant, Kv is the vertical conductivity of the confined layer, H is thickness of the confined layer, and S0s denotes the specific storage of the confining layer. The consolidation achieves almost 95 % as t=T  ¼ 1:0. The hysteresis of the consolidation increases with H and S0s , and decreases with Kv . In the land subsidence center of Cangzhou, the estimated thicknesses of clay and mild clay amount to 214 and 131 m, respectively, and the thickness of clay in the groundwater production range of 200–400 m below ground also reaches 179 m. Test results of soil samples in the laboratory indicate that the value of the coefficient of uniformity is high for clay samples, which may lead to large range of particle sizes and therefore high compressibility of the soil. High compression coefficients of the clay may have resulted in great hysteresis of land

123

Fig. 11 Changes of soil hydraulic conductivities with time during a seepage test with a high confining pressure triaxial seepage equipment (the depths for soil samples no. 1–9 are 226, 248, 297, 385, 403, 414, 426, 433, and 444 m, respectively)

subsidence, as inferred by Eq. (8). Test results also indicate that the compression coefficient is a variable which decreases with pressure; thus, it is important to accurately estimate this parameter during analytical or numerical calculations. Nine clay samples extending to depths of up to 450 m were collected from the subsidence center at Cangzhou, then seepage tests were carried out by the high confining pressure triaxial seepage equipment for determining the hydraulic conductivities.With a confining pressure of 2.5 MPa and a seepage pressure of 2.0 MPa, the changes of hydraulic conductivities with time were investigated (Fig. 11). Results indicate that the hydraulic conductivities for samples taken from deep strata decrease rapidly together with some fluctuations during the first 2–4 days due to a large stress relief. In the early period of the seepage

Environ Earth Sci

test, the microstructure properties of clay samples have changed a lot and the pore radius and throat of clay expand so that short-circuit-type seepage is dominant. Then the piston-type seepage is dominant due to effects of the confining pressure, the hydraulic pressure, and clay minerals. The hydraulic conductivities become stable in the later period of the test. A pumping test was conducted in the aquifer 250–350 m below the ground, which has yielded the aquifer parameter value for hydraulic conductivity to be 2.86 m/d that is about six orders of magnitude greater than those of the clay samples. This fact suggests that the hydraulic conductivities of the confining layers are significantly lower compared with the aquifers; therefore, strong hypothesis of land subsidence may occur after enormous volumes of groundwater are pumped from the aquifer, as inferred by Eq. (8). The time for 90 % consolidation of a soil layer can be derived by Eq. (8): t2 ¼ t1

H22 H12

ð9Þ

where t2 is the time required to accomplish the consolidation of a soil layer, t1 is the time required to accomplish the consolidation of a soil sample within the soil layer, H1 is the thickness of the soil sample, and H2 is thickness of the soil layer. With Eq. (9), the time for 90 % consolidation of soil layers was calculated as from less than one year to tens of years in subsidence areas of Cangzhou and Tianjin. Secondary compression Secondary compression is thought to be caused by the gradual readjustment of the clay particles into a more stable configuration following the structural disturbance induced by the decrease in void ratio, especially if the clay is laterally confined. An additional factor is the gradual lateral displacements which take place in thick clay layers subjected to shear stresses (Craig 2004). During the secondary compression some compression of soil takes place at slow rate even after the reduction of hydrostatic pressure. Settlement caused by secondary compression is negligible in sand but may be significant in clay. The study in Tianjin area shows that the subsidence caused by secondary compression can be as great as more than 10 mm per year (Tianjin Institute of Geo-Environmental Monitoring 2013). The rate of secondary compression (Ca ) in the oedometer test can be defined by the slope of the final part of the compression–log time curve, measured as the unit compression over one decade on the log time scale. Experimental results in Tianjin show that the rate of

Fig. 12 The relations among rate of secondary compression, moisture content, and clay content in coastal areas of Tianjin

secondary compression tends to increase with the moisture content (Fig. 12). Although the rate of secondary compression seems to have no clear relation with the clay content as shown in Fig. 12, a possible correlation may exist between them if more data are used to produce the scatter chart.

Summary and conclusions The main factors leading to land subsidence in the NCP are represented by the geological condition such as highly compressible thick loose sediments and the dynamic conditions of the groundwater system, e.g., groundwater abstraction. The sediments beneath the NCP represent alternating sequences of sand and highly compressible clay layers. Long-term excessive withdrawal of groundwater has led to significant decrease of groundwater levels and therefore a large amount of subsidence in this region. Generally, land subsidence in most areas of the piedmont plain is caused by shallow groundwater abstraction, while land subsidence in the alluvial and coastal plains bears strong relation with deep groundwater withdrawal and consequent decline of groundwater levels in deep confined aquifers. More specifically, the formation of land subsidence is greatly controlled by the compressibility and the thickness of clay layers in the aquifer system of groundwater production, as well as the abstraction amount. Groundwater abstraction in confined aquifers can cause land subsidence more easily compared with unconfined aquifers, which has been supported by both theoretical analysis and field observations. Most land subsidence in middle-east plain of the NCP is caused by abstraction of deep confined groundwater. The cumulative subsidence volume represents the amount of released water from compressibility of aquifers and the aquitards. The ratio of the cumulative land subsidence volume to the amount of deep groundwater abstraction was estimated as

123

Environ Earth Sci

40.1–43.8 % in middle-east plain of the NCP, and 57.6 % in Cangzhou City. This difference of the ratio is controlled by the local lithological structure and groundwater recharge and discharge conditions. The hydrogeological conditions in middle-east plain of the NCP determine that the water released from compression of deep aquifers and aquitards may be the main source of deep groundwater abstraction, and belongs to the depletion of the storage. The time for 90 % consolidation was calculated as from less than one year to tens of years in subsidence areas of Cangzhou and Tianjin. Secondary compression should not be overlooked especially in the coastal plain. Experimental results indicate that the rate of secondary compression tends to increase with the moisture content. The area of subsidence in the NCP keeps extending and the accumulative subsidence increasing though some controlling measures have been taken. The subsidence of the NCP has a direct relationship with groundwater exploitation. On the other hand, both population and economic growth will be heavily dependent upon groundwater resource. Therefore, an important task is to seek a balance between controlling land subsidence and ensuring normal water supply so that the NCP can be developed in a sustainable and harmonious manner. Despite the fact that land subsidence and groundwater levels have been measured over the last few decades, some mechanisms underlying the subsidence are still not fully known in the NCP. For example, efficient methods have not been found to quantify the settlements caused by extraction of deeply buried oil, natural gas, or geothermal water resources. Additionally, the relations between land subsidence patterns and the distribution of soil discontinuities in the form of cracks or fissures are still not fully understood. Investigation, monitoring, and related synthetic study of land subsidence will be important if more accurate understanding of the subsidence mechanism is to be obtained to speculate on the potential trend of the subsiding rates and possibly on the location of new subsidence areas in the future. Mitigative measures could be better provided for controlling land subsidence if mechanism of land subsidence in the NCP is fully known. Acknowledgments This work was supported by the projects: China Geological Survey (No. 12120113011700), Comparative research of groundwater management in the coastal areas in Southeast Asia (No. RETA 6498), and National Basic Research Program of China (973 Program, 2010CB428806).

References Abidin HZ, Gumilar I, Andreas H et al (2013) On causes and impacts of land subsidence in Bandung Basin, Indonesia. Environ Earth Sci 68(6):1545–1553

123

Adrian OG, Rudolph DL, Cherry JA (1999) The analysis of long term land subsidence near Mexico City: field investigations and predictive modeling. Water Resour Res 35:3327–3341 Chatterjee RS, Fruneau B, Rudant JP et al (2006) Subsidence of Kolkata (Calcutta) city, India during the 1990s as observed from space by differential synthetic aperture radar interferometry (DInSAR) technique. Remote Sens Environ 102:176–185 Chen CH, Pei SP, Jiao JJ (2003) Land subsidence caused by groundwater exploitation exploitation in Suzhou City, China. Hydrogeol J 11:275–287 Chen CH, Wang CH, Hsu YJ, Yu SB, Kuo LC (2010) Correlation between groundwater level and altitude variations in land subsidence area of the choshuichi alluvial fan, Taiwan. Eng Geol 115:122–131 Craig (2004) Soil mechanics, 7th edn. Spon Press, Taylor & Francis Group, London and New York Cui XD, Niu XJ, Bai JW (2002) Prevention and control of land subsidence in Tianjin (in Chinese). In: Wei ZX, Li QF (eds) Proceedings of the national symposium on land subsidence, Shanghai Institute of Geology Survey, Shanghai, pp 338–345 Domenico and Schwartz (1998) Physical and chemical hydrology, 2nd edn. Wiley, New York Foster S, Garduno H, Evant R et al (2004) Quaternary Aquifer of the North China Plain—assessing and achieving groundwater resource sustainability. Hydrogeol J 12:81–93 Galgaro A, Boaga J, Rocca M (2014) HVSR technique as tool for thermal-basin characterization: a field example in N–E Italy. Environ Earth Sci 71(10):4433–4446 Hernandez-Marin M, Pacheco-Martinez J, Ramirez-Cortes A (2014) Evaluation and analysis of surface deformation in west Chapala basin, central Mexico. Environ Earth Sci 72(5):1491–1501 Holzer TL (1989) State and local response to damaging land subsidence in united states urban areas. Eng Geol 27:449–466 Hu H, Huang L, Yang G (1993) Recent vertical crustal deformation in the coastal area of eastern China. Scientia Geologica Sinica 28(3):270–278 (in Chinese) Hu RL, Yue ZQ, Wang LC, Wang SJ (2004) Review on current status and challenging issues of land subsidence in China. Eng Geol 76:65–77 Khan AS, Khan SD, Kakar DM (2013) Land subsidence and declining water resources in Quetta Valley, Pakistan. Environ Earth Sci 70(6):2719–2727 Narasimhan TN, Kanehiro BY, Witherspoon A (1984) Interpretation of earth tides response to three deep, confined aquifers. J Geophys Res 89(B3):1913–1924 Papadaki ES (2014) Monitoring subsidence at Messara basin using radar interferometry. Environ Earth Sci 72(6):1965–1977 Phien-wej N, Giao PH, Nutalaya P (2006) Land subsidence in Bangkok, Thailand. Eng Geol 82:187–201 Poland JF, Davis GH (1969) Land subsidence due to withdrawal of fluids. In: Varnes DJ, Kiersch G (eds) Reviews in engineering geology II. Geological Society of America Inc., Boulder Rudolph DL, Frind EO (1991) Hydraulic response of highly compressible aquitards during consolidation. Water Resour Res 27(1):17–30 Stramondo S, Bozzano F, Marra F et al (2008) Subsidence induced by urbanisation in the city of Rome detected by advanced InSAR technique and geotechnical investigations. Remote Sens Environ 112:3160–3172 Sun H, Grandstaff D, Shagam R (1998) Land subsidence due to groundwater withdrawal: potential damage of subsidence and sea level rise in southern New Jersey, USA Tianjin Institute of Geo-Environmental Monitoring (2013) Land subsidence monitoring and remediation in Tianjin of the North China Plain, China Geological Survey Report. (in Chinese)

Environ Earth Sci Tomas R, Romero R, Mulas J (2014) Radar interferometry techniques for the study of ground subsidence phenomena: a review of practical issues through cases in Spain. Environ Earth Sci 71(1):163–181 Tsai TL, Jang WS (2014) Deformation effects of porosity variation on soil consolidation caused by groundwater table decline. Environ Earth Sci 72(3):829–838 Wang XY, Sun L, Wang ZL (2013) An analysis of the resilience capacity of soils in North China: a study on land subsidence treatment. Bull Eng Geol Environ. doi:10.1007/s10064-0130561-9 Xing Z, Li H, Zhang S, Gao H (2004) Surface subsidence and its countermeasures in Cangzhou City. Geol Surv Res 27(3):157–163 Xue YQ, Zhang Y, Ye SJ, Wu JC, Li QF (2005) Land subsidence in China. Environ Geol 48(6):713–720

Yang G, Wang M, Han Y (2001) Crustal deformation in northern part of North China Associated with the Zhangbei earthquake. (in Chinese) Ye XB, He QC et al (2006) Estimation of economic loss due to land subsidence in the North China Plain, China Land Press Zeitoun DG, Wakshal E (2013) Land subsidence analysis in urban areas: the Bangkok metropolitan area case study. Springer Environ Sci Eng. doi:10.1007/978-94-007-5506-22 Zhang YQ, Gong HL, Gu ZQ, Wang R, Li XJ, Zhao WJ (2014) Characterization of land subsidence induced by groundwater withdrawals in the plain of Beijing city, China. Hydrogeol J 22:397–409 Zuo H, Liu TZ, Lin XH (1993) Economic benefit risk assessment of land subsidence in Shanhai. Environ Geol 21:208–211

123