Abstract. Over the last decades, economic developments in the Vietnamese Mekong Delta have led to a sharp increase in groundwater pumping for domestic, agricultural and industrial use. This has resulted in alarming rates of land subsidence and groundwater salinization. Effective groundwater management, including strategies to work towards sustainable groundwater use, requires knowledge about the current groundwater salinity distribution, in particular the available volumes of fresh groundwater. At the moment, no comprehensive dataset of the spatial distribution of fresh groundwater is available. To create a 3D model of total dissolved solids (TDS), an existing geological model of the spatial distribution and thickness of the aquifers and aquitards is updated. Next, maps of drainable porosity for each aquifer are interpolated based on the sedimentological description of the borehole data. Measured TDS in groundwater, inferred TDS from resistivity measurements in boreholes and soft incomplete data (derived from measurements in boreholes and data from domestic wells) are combined in an indicator kriging routine to obtain the full probability distribution of TDS for each (x,y,z) location. This statistical distribution of TDS combined with drainable porosity yields estimates of the volume of fresh groundwater (TDS < 1 g L−1) in each aquifer. Uncertainty estimates of these volumes follow from a Monte Carlo analysis (sequential indicator simulation). Results yield an estimated fresh groundwater volume for the Mekong Delta of 867 billion cubic metres with an uncertainty range of 830–900 billion cubic metres, which is somewhat higher than previous assessments of fresh groundwater volumes. The resulting dataset can for instance be used in groundwater flow and salt transport modelling as well as aquifer storage and recovery projects to support informed groundwater management decisions, e.g. to prevent further salinization of the Mekong Delta groundwater system and land subsidence, and is available at https://doi.org/10.5281/zenodo.4441776 (Gunnink et al., 2021).
Read full abstract