4 Mitigation of Land Subsidence by Water Injection

The simplest, most straightforward action toward mitigating land subsidence caused by fluid withdrawal would seem to be artificial fluid injection. It goes without saying that other strategies can help to prevent land subsidence, including the policy of requiring withdrawal limits, permits, fees, taxes, metering, and enforcement control on groundwater pumping as exercised by central and local authorities. Freeze (2000) conveys the general recommendation that land subsidence should act as a guiding factor when defining a groundwater exploitation management strategy, along with more traditional factors, such as water table decline, saltwater intrusion, and avoidance of groundwater contamination.

Generally speaking, when land subsidence has occurred and/or is still occurring, methods used to control, mitigate, or arrest it include reduction of pumping rates, artificial aquifer recharge from the land surface, re‑pressurization of depleted layers by way of injection wells, creation of a hydraulic barrier to stop advancement of the cone of depression, and generation of an overpressure in geological units unaffected by pumping in order to build a structural obstacle to the migration of in‑depth compaction to the ground surface. A combination of any of the above methods can be used as well, consistent with a cost/benefit analysis. An example of conservative mitigation strategy is one whereby the effective stress within the depleted formation does not increase beyond the stress level experienced to date. A more aggressive strategy might dictate a decrease in the effective stress and/or the active involvement of overlying formations through the use of fluid injection. Injecting water into a geological formation generates an increase in pore pressure, a decrease in effective stress, and hence an expansion of the injected formation. Part of the latter may migrate to the ground surface, giving rise to an anthropogenic land rebound and/or uplift.

While anthropogenic land subsidence is a well‑known process, the reverse, namely artificial land uplift, is a much less observed and recognized event, even though the practice of injecting fluids underground is more than a half a century old. Injection technology has been advancing continuously since it came into wide use in the 1950s and 1960s in order to reinject the formation water extracted along with hydrocarbons, or to dispose of industrial wastes. The number of injection wells has grown exponentially, to the point that EPA (the United States Environmental Protection Agency) has identified approximately 400,000 injection boreholes in the USA alone (USEPA, 2002). The injection of water‑based solutions, hydrocarbons, CO2 or N2 to enhance oil production (EOR) started in the 1940s and soon became an accepted technique for recovering additional oil from reservoirs that were already depleted or water flooded. Thermal recovery processes by vapor injection, used in reservoirs containing heavy (viscous) oil or bitumen, are generally accompanied by an noticeable uplift (for example, locally recorded up to 30 cm). Examples include the Cold Lake (Stancliffe and van der Kooij, 2001), Shell Peace River (Du et al., 2008), and Athabasca oil sands (Collins, 2007), in Canada.

In the Krechba gas field, Algeria, land rebound was caused by the reinjection of CO2 separated from the produced gas (Vasco et al., 2010). Storing gas underground may generate measurable land uplift as well (Teatini et al., 2011a). The aquifer systems underlying Tokyo and Osaka, Japan (Sreng et al., 2011) and Taipei, Northern Taiwan (Chen et al., 2007), experienced a natural flow field recovery after cutting the water pumpage, and significant land rebound as well. There are also examples of water being pumped into an oil field to mitigate land subsidence caused by oil production, including the case of Long Beach, California, USA. Here the mitigation program was carefully controlled and monitored (Pierce, 1970; Rintoul, 1981; Colazas and Strehle, 1995). Water injection started on a major scale in 1958 using appropriately treated seawater collected from shallow wells 30–120 m deep, later mixed with formation wastewaters produced with the oil. Eleven years later, when 2 m3/s was being pumped into the oil field, the settling area had been reduced from 58 to 8 km2, with local land surface rebound of 30 cm.

Land motion related to subsurface fluid injection went unnoticed for a long time in the vast majority of cases. There are a number of reasons for this. First, in most cases, the disposal of fluids occurred in deserted or sparsely inhabited areas where measuring surface displacements was not a priority, in part due to the large cost of leveling surveys. In other instances, uplift was so slight that no environmental hazards were created, no monitoring was installed, or the area involved was quite limited, with no damage to engineered structures or infrastructures reported or even expected. Only in recent times has satellite technology offered a relatively inexpensive, spatially distributed, accurate methodology for detecting ground movements worldwide. It has revealed anthropogenic uplifts of some interest in terms of magnitude, size of the area involved, and time of occurrence. The use of SAR‑based techniques has grown rapidly over the last decade, facilitating the detection and measurement of rising areas. This is particularly true for surface movements connected with natural fluctuations of the groundwater head and in areas of aquifer storage and recovery (ASR), which have been systematically monitored by the United States Geological Survey: including, among others, Santa Clara Valley, California, USA (Schmidt and Burgmann, 2003), Santa Ana basin, California, USA (Galloway and Hoffmann, 2007), and Las Vegas Valley, Nevada, USA (Hoffmann et al., 2001; Bell et al., 2008). Measured uplift amounted to 4 cm from 1992 to 1999 in Santa Clara Valley and 3 cm from 2003 to 2005 in Las Vegas Valley. In addition, surface and borehole tiltmeters have been widely used in recent years to monitor ground heave within relatively small areas (Du et al., 2008). Teatini et al. (2011b) provide a recent thorough review of areas uplifted anthropogenically by injecting fluid underground.

As far as soil compressibility is concerned, the value of cb in the first loading cycle is to be used if the aquifer is pumped, and in unloading/reloading when the aquifer is recharged/repressurized. The ratio cb,loading/cb,unloading decreases with depth and may approach 1 order of magnitude for very shallow silty/clayey sediments (Teatini et al., 2011b).

Because of their low elevation and location on the sea, Shanghai in China and Venice in Italy represent two special cases where land subsidence mitigation is of paramount importance. Shanghai, a coastal city situated in the southern part of the Yangtze Delta, China, has experienced a large land subsidence (Table 1) due mainly to excessive long‑term groundwater withdrawal and, secondarily, to the rapid development of the area. Groundwater extraction in Shanghai dates back to 1860. The pumped water volume was quite small before 1949 and then it increased rapidly, especially in the late 1950s. The yearly pumping rate reached its peak of 200×106 m3/year in 1963 (Figure 27a). Intensive groundwater extraction has caused severe land subsidence. During the period 1957–1961, the maximum yearly rate of subsidence peaked up to 17 cm/year as shown in Figure 27a (Zhang et al., 2015, Ye et al., 2016).

a) History of groundwater pumping, artificial recharge, and average land subsidence in Shanghai. Pumping and recharge wells in b) aquifer A2 and c) aquifer A4 in Shanghai. d) Representative hydrogeological section of the Shanghai aquifer system along the cross‑section I‑I' shown in (b).

Figure 27 a) History of groundwater pumping, artificial recharge, and average land subsidence in Shanghai (modified after Zhang et al., 2015). Pumping and recharge wells in b) aquifer A2 and c) aquifer A4 in Shanghai. The white‑to‑red triangles represent pumping wells (negative values), white‑to‑blue circles represent recharge wells (positive values). Symbol size is proportional to the average yearly rate from 1980 to 1996 (modified from Ye et al., 2016). d) Representative hydrogeological section of the Shanghai aquifer system along the cross‑section I‑I’ shown in (b) (modified from Ye et al., 2016).

In order to control land subsidence, a series of measures were implemented by the Shanghai government beginning in the 1960s. These measures include:

  • reduction of groundwater withdrawal;
  • exploitation of deeper producing layers; and,
  • artificial recharge of aquifers.

Pumping gradually moved from the second (A2) and third (A3) confined aquifer to the fourth (A4) and fifth (A5) aquifers, with decreasing yearly pumpage since 1998. Artificial recharge started in 1966 and slightly increased in the following years. The recharge has been carried out using properly treated tap water taken from the Huangpu River. Over the period 1983–1989, the yearly injection rate was nearly constant at 30×106 m3/year. After that, it decreased slowly and increased year after year since 2003 (Figure 27a).

Figures 27b and 27c show the discharge and recharge wells in aquifers A2 and A4, along with the average annual discharge and recharge rates over the period 1980‑1996. The number of recharge wells was much larger than the number of pumping wells in aquifer A2. Conversely, several pumping wells with high flow rates and a few recharge wells were active in aquifer A4. Because of the implementation of the above mitigation measures, land subsidence recently decreased to about 1 cm/year.

In Venice, land uplift is predicted with the aid of a finite element (FE) model (Figure 28a). An upheaval of the city induced by seawater injection into deep saline aquifers could significantly reduce the frequency of the high tides that periodically flood Venice. A recent exceptional high tide on November 12, 2019, peaked at 187 cm above datum and severely damaged the city (https://www.voanews.com/europe/venice-mayor-declares-disaster-city-hit-2nd-worst-high-tide). Early numerical studies based on a simplified lithostratigraphy of the Venetian subsurface (Comerlati et al., 2004) suggested that the city might be raised by pumping seawater into deep aquifers through 12 wells located on a 10 km diameter circle. Using a more accurate 3‑D reconstruction of the Quaternary deposits, developed very recently from about 1050 km of multichannel seismic profiles and eight exploration wells, along with a more accurate representation of the injection boreholes, new FE predictions were performed (Teatini et al., 2011c). The new model simulates the lithostratigraphy of the lagoon subsurface and allows for a reliable assessment of the water volumes injected into the geologic formations based on the measured bottomhole overpressures, which vary both in space and time. Selection of the best hydraulic conductivity is discussed by Teatini et al. (2010), while rock compressibility in the unloading condition has been derived in agreement with Comerlati et al. (2004) and Ferronato et al. (2013). Pumping is planned along two Pleistocene sequences originating from the Alps and Apennines sedimentation and terminating just south and north of Venice, respectively, and the shelf portion of a rather continuous Pliocene sequence below the central lagoon, with arenite layers as deep as 1000 m below mean sea level. With a proper tuning of the injection pressure, the model (Teatini et al., 2011c) allows for prediction of a fairly uniform 25‑30 cm uplift over 10 years after the initiation of injection (Figure 28b).

a) Axonometric view of the tetrahedral mesh used to predict the anthropogenic uplift of Venice by seawater injection into saline aquifers. b) Predicted uplift (cm) at Venice after 10 years of injection into saline aquifers 650–1000 m deep below the lagoon.

Figure 28 a) Axonometric view of the tetrahedral mesh used to predict the anthropogenic uplift of Venice by seawater injection into saline aquifers. The mesh has 1,905,058 elements and 328,215 nodes. b) Predicted uplift (cm) at Venice after 10 years of injection into saline aquifers 650–1000 m deep below the lagoon. The injection wells are marked in red (modified after Teatini et al., 2011c).

A pilot experiment has been designed to verify the feasibility of the project for uplifting Venice (Castelletto et al., 2008). The pilot experiment plan foresees three boreholes located at the vertices of a triangle with sides 1 km long, in a lagoon area to be selected in the vicinity of Venice’s historical center. The aim would be (1) to obtain further detailed lithostratigraphy of the underground lagoon; (2) to perform an injection test with (treated) seawater and measure the overpressure generated in the injected formation; (3) to monitor continuously and in real time land uplift in the area, with the aid of high‑precision leveling, GPS, and satellite interferometry; and (4) to set up and experiment with a procedure of optimal control; for instance, the uniformity of uplift may be checked with the aid of sensor feedback automatically accommodating the injection rate in each single well. A detailed description of the project for anthropogenic uplift of Venice, its major environmental impact, and expected cost is provided by Gambolati and Teatini (2014).


Land Subsidence and its Mitigation Copyright © 2021 by Giuseppe Gambolati and Pietro Teatini. All Rights Reserved.