Implementation of satellite-based data for improving predictions of arsenic contamination in groundwater in the Red River Delta in Vietnam

Natural arsenic contamination of groundwater aquifers is globally widespread, and particularly poses a problem in regions where groundwater is the main source of drinking and cooking water. Arsenic poisoning can lead to a myriad of serious health effects such as diseases of blood vessels, diabetes and cancers. The aquifers of the Red River Delta in Vietnam are highly contaminated with arsenic and it has been estimated that in this area, around 3 million people are affected by high arsenic concentrations (> 10 µg/L, WHO guideline value; Winkel et al., 2011). Previously, predictions of arsenic contamination in the Red River Delta were established via geospatial modelling using arsenic measurements, as well as surface and 3D-geology. Based on these predictions, probability maps of arsenic at specific depths were created. By comparing these depthresolved probabilities to measured arsenic concentrations, a drawdown of arsenic-enriched waters from Holocene aquifers to previously uncontaminated Pleistocene aquifers was observed. This finding indicated that arsenic contamination has been exacerbated by excessive groundwater pumping rates (Winkel et al., 2011). Furthermore, in a study conducted in the Mekong delta, it was hypothesized that groundwater extraction causes interbedded clays to compact, thereby releasing water containing dissolved arsenic that is subsequently transported to deeper aquifers (Erban et al., 2013). Such human-induced changes cannot be captured by the previous predictive models based on natural predictive parameters mentioned above, leading to erroneous predictions of the arsenic content in areas affected by urbanization, especially in deeper aquifers. To improve predictions in human-affected regions we are using satellite data and remote sensing techniques that enable detection of changes of urban and suburban extents (Nghiem et al., 2009) and vertical build-up (Mathews et al., 2019). Those data and techniques in combination with geochemical and environmental data can help in i) resolving mechanisms behind arsenic mobilization in aquifers due to increased pumping rates and ii) making predictions of arsenic contamination more accurate, especially in areas characterized by increased groundwater pumping.