7 Conclusions
Anthropogenic land subsidence as related to subsurface fluid production has been known for almost a century. Groundwater withdrawal is the primary cause worldwide. Although overall damage today is estimated at billions of dollars a year (for example, Borchers and Carpenter, 2014), it is expected to increase due to population and economy growth. Land subsidence is still a problem that is under‑evaluated by both governments and public opinion, especially in developing countries. Impacts include the loss of conveyance capacity in canals, streams and rivers, diminished effectiveness of levees, damage to roads, bridges, buildings, water wells, pipelines and other surface structures and infrastructures, increasing vulnerability of aquifers to saltwater intrusion, contamination of shallow aquifers through ground ruptures, and the flooding of coastal and inland urban areas (for example, New Orleans, Louisiana, USA and Mexico City, Mexico). The environmental impact of land subsidence has shifted over the last decade from rural and industrial sites (for example, the Antelope Valley, California, USA, or the Po River delta, Italy) to urban centers (for example, Shanghai, China and Mexico City, Mexico) because of increasing population and growth of mega‑cities. Whereas in 1950, New York was the only urban area totaling more than 10 million people, presently more than 30 cities in the world exceed this impressively large number, most of them located on the coasts of developing countries.
When estimating the impacts of land subsidence over horizontal multi‑aquifers, an initial approach uses 1‑D vertical movement with ground settlement η = Δp cb s0 (that is, equal to the compaction η of the pumped unit with initial thickness s0 and uniaxial vertical compressibility cb subject to the pore pressure decline (Δp). If the fraction of clayey/silty soils (namely aquitards/confining beds and intersperse clayey/silty lenses) is important, compaction, hence land subsidence, can be delayed in time relative to the compaction of the pumped sandy formations. As a major consequence land subsidence can still continue after well pumping ceases. Rebound due to aquifer recharge and aquitard re‑pressurization (either natural or artificial) can make up for only a small fraction of the overall land subsidence as cb in expansion is generally significantly smaller than virgin cb in compression, and especially so in clayey/silty units.
The mechanisms underlying the basic process are well understood and universally accepted, and the mathematical modeling of past events and expected future cases is also well established. Modern computer technology allows for the simulation of complex geology and geometry in subsiding basins, of arbitrarily distributed pumping rates, of heterogeneity, anisotropy and non‑linearity of the porous media properties, with a degree of accuracy inconceivable until only a few years ago. Measuring and monitoring anthropogenic land subsidence is presently at a very advanced stage, especially with the aid of satellite technology. Scientists can help support decision makers toward predicting, preventing or at least mitigating land subsidence successfully, although certain specific areas may still require more in‑depth investigation. These include the 3‑D deformation and stress fields correlating to groundwater pumping, uplift caused by water injection, and inverse modeling calibration. Land subsidence rates have been drastically reduced in several places around the world, for example, in Venice, Italy; Tokyo, Japan; and more recently Shanghai, China by exploiting water resources other than groundwater. However, for the majority of other mega‑cities this target is not within easy reach. This is why land subsidence was recently mentioned as one of the most urgent threats to sustainable development, in the latest UNESCO International Hydrological Programme VIII (2014‑2020).
Major research advancements are needed to better predict earth fissuring, hydraulic fracturing, fault activation, and induced seismicity. Modeling these processes require approaches developed in the field of discontinuous mechanics, approaches only partially assimilated in geosciences so far. Significant progress has been made in understanding theoretical mechanisms. However, monitoring their occurrence, characterizing their rheological properties, and developing reliable, robust, and accurate numerical models still pose major challenges for research efforts in the near future.