Groundwater exploitation is a major cause of land subsidence, which in coastal areas poses a flood inundation hazard that is compounded by the threat of sea-level rise (SLR). In the lower Mekong Delta, most of which lies <2 m above sea level, over-exploitation is inducing widespread hydraulic head (i.e., groundwater level) declines. The average rate of head decline is ∼0.3 m yr −1 , based on time-series data from 79 nested monitoring wells at 18 locations. The consequent compaction of sedimentary layers at these locations is calculated to be causing land subsidence at an average rate of 1.6 cm yr −1 . We further measure recent subsidence rates (annual average, 2006-10) throughout the Delta, by analysis of interferometric synthetic aperture radar (InSAR), using 78 ALOS PALSAR interferograms. InSAR-based subsidence rates are 1) consistent with compaction-based rates calculated at monitoring wells, and 2) ∼1-4 cm yr −1 over large (1000s of km 2 ) regions. Ours are the first mapped estimates of Delta-wide land subsidence due to groundwater pumping. If pumping continues at present rates, ∼0.88 m (0.35-1.4 m) of land subsidence is expected by 2050. Anticipated SLR of ∼0.10 m (0.07-0.14 m) by 2050 will compound flood inundation potential. Our results suggest that by mid-century portions of the Mekong Delta will likely experience ∼1 m (0.42-1.54 m) of additional inundation hazard.

Many major river deltas in the world are subsiding and consequently become increasingly vulnerable to flooding and storm surges, salinization and permanent inundation. For the Mekong Delta, annual subsidence rates up to several centimetres have been reported. Excessive groundwater extraction is suggested as the main driver. As groundwater levels drop, subsidence is induced through aquifer compaction. Over the past 25 years, groundwater exploitation has increased dramatically, transforming the delta from an almost undisturbed hydrogeological state to a situation with increasing aquifer depletion. Yet the exact contribution of groundwater exploitation to subsidence in the Mekong delta has remained unknown. In this study we deployed a delta-wide modelling approach, comprising a 3D hydrogeological model with an integrated subsidence module. This provides a quantitative spatially-explicit assessment of groundwater extraction-induced subsidence for the entire Mekong delta since the start of widespread overexploitation of the groundwater reserves. We find that subsidence related to groundwater extraction has gradually increased in the past decades with highest sinking rates at present. During the past 25 years, the delta sank on average ∼18 cm as a consequence of groundwater withdrawal. Current average subsidence rates due to groundwater extraction in our best estimate model amount to 1.1 cm yr−1, with areas subsiding over 2.5 cm yr−1, outpacing global sea level rise almost by an order of magnitude. Given the increasing trends in groundwater demand in the delta, the current rates are likely to increase in the near future.

Deep aquifers in South and Southeast Asia are increasingly exploited as presumed sources of pathogen- and arsenic-free water, although little is known of the processes that may compromise their long-term viability. We analyze a large area (>1,000 km 2 ) of the Mekong Delta, Vietnam, in which arsenic is found pervasively in deep, Pliocene–Miocene-age aquifers, where nearly 900 wells at depths of 200–500 m are contaminated. There, intensive groundwater extraction is causing land subsidence of up to 3 cm/y as measured using satellite-based radar images from 2007 to 2010 and consistent with transient 3D aquifer simulations showing similar subsidence rates and total subsidence of up to 27 cm since 1988. We propose a previously unrecognized mechanism in which deep groundwater extraction is causing interbedded clays to compact and expel water containing dissolved arsenic or arsenic-mobilizing solutes (e.g., dissolved organic carbon and competing ions) to deep aquifers over decades. The implication for the broader Mekong Delta region, and potentially others like it across Asia, is that deep, untreated groundwater will not necessarily remain a safe source of drinking water.

The Vietnamese Mekong delta is subsiding due to a combination of natural and human-induced causes. Over the past several decades, large-scale anthropogenic land-use changes have taken place as a result of increased agricultural production, population growth and urbanization in the delta. Land-use changes can alter the hydrological system or increase loading of the delta surface, amplifying natural subsidence processes or creating new anthropogenic subsidence. The relationships between land use histories and current rates of land subsidence have so far not been studied in the Mekong delta. We quantified InSAR-derived subsidence rates for the various land-use classes and past land-use changes using a new, optical remote sensing-based, 20-year time series of land use. Lowest mean subsidence rates were found for undeveloped land-use classes, like marshland and wetland forest (~6-7mmyr), and highest rates for areas with mixed-crop agriculture and cities (~18-20mmyr). We assessed the relationship strength between current land use, land-use history and subsidence by predicting subsidence rates during the measurement period solely based on land-use history. After initial training of all land-use sequences with InSAR-derived subsidence rates, the land-use-based approach predicted 65-92% of the spatially varying subsidence rates within the measurement error range of the InSAR observations (RMSE=5.8mm). As a result, the spatial patterns visible in the observed subsidence can largely be explained by land use. We discuss in detail the dominant land-use change pathways and their indirect, causal relationships with subsidence. Our spatially explicit evaluation of these pathways provides valuable insights for policymakers concerned with land-use planning in both subsiding and currently stable areas of the Mekong delta and similar systems.

No abstract

Rice production in Cambodia, essential to food security and exports, is largely limited to the wet season. The vast majority (96%) of land planted with rice during the wet season remains fallow during the dry season. This is in large part due to lack of irrigation capacity, increases in which would entail significant consequences for Cambodia and Vietnam, located downstream on the Mekong River. Here we quantify the extent of the dry season "deficit" area in the Cambodian Mekong River catchment, using a recent agricultural survey and our analysis of MODIS satellite data. Irrigation of this land for rice production would require a volume of water up to 31% of dry season Mekong River flow to Vietnam. However, the two countries share an aquifer system in the Mekong Delta, where irrigation demand is increasingly met by groundwater. We estimate expansion rates of groundwater-irrigated land to be >10% per year in the Cambodian Delta using LANDSAT satellite data and simulate the effects of future expansion on groundwater levels over a 25-year period. If groundwater irrigation continues to expand at current rates, the water table will drop below the lift limit of suction pump wells, used for domestic supply by >1.5 million people, throughout much of the area within 15 years. Extensive groundwater irrigation jeopardizes access for shallow domestic water supply wells, raises the costs of pumping for all groundwater users, and may exacerbate arsenic contamination and land subsidence that are already widespread hazards in the region.

Groundwater exploitation is rising in the Mekong Delta, Vietnam, potentially exacerbating arsenic contamination from natural sources. We investigate trends and controls on contamination patterns throughout the Delta's multi-aquifer system as observed in a spatially exhaustive data set of arsenic measured in >40,000 wells, 10.5% of which exceed the WHO drinking water standard for arsenic (10 μg/L). We relate strong trends in the distribution of contamination among well samples to explanatory variables derived from 3D ancillary physicochemical data sets using logistic regression models. Parsimonious models describe much of the observed variability in arsenic occurrence, which differs considerably between subsets of wells tapping shallow versus deeper aquifer groups. In the shallowest Holocene-Pleistocene aquifers, arsenic occurrence is best described by distance to the Mekong river channels and delta front, depth, and location within fault-bounded zones of the region. The same model, however, fails to explain observations in the deeper group of Pliocene-Miocene aquifers. Among these deeper units, arsenic occurrence is rare except among older wells in near-river, heavily pumped areas. Our analysis is the first to examine both natural and anthropogenically mediated contributions to the distribution of arsenic throughout the Mekong Delta's multi-aquifer system, with implications for management of similarly affected basins throughout Southeast Asia.

No abstract