Decadal trends in groundwater quality observed in national groundwater monitoring wells - assessment of climate change effects using machine learning.

As pressures on water resources are expected to increase due to climate change and population growth, ensuring sufficient quantity and quality of drinking water emerges as a global challenge. Climate change can affect groundwater quantity and quality through changes in chemical equilibria, reaction kinetics, and soil processes induced by shifting temperature, precipitation, and evapotranspiration. However, the impact of climate change on groundwater quality has not been studied thoroughly and thus is identified as an important scientific challenge and knowledge gap. To address this challenge, we analyzed high-quality long-term datasets, spanning from 1993 to 2022 of several environmental and hydroclimatic factors, as well as groundwater quality and quantity, available at national scale in Denmark. Initial results from around 200 groundwater monitoring wells distributed across Denmark show a decrease in pH and oxygen content in most of the wells for the period 1993-2022. Expected results will show the temporal change of selected geogenic compounds and other major chemicals and physical properties and their trends for this climatic period. Machine learning analysis will be applied in future work to identify the main drivers of change in concentrations of selected geogenic compounds, oxygen, and pH, and to create baseline maps of the recent period (2017-2022/23). The baseline maps, representing current conditions, will be derived by geospatial machine learning modelling frameworks linking covariate maps with borehole scale information of water quality parameters. How to distinguish the impacts of climate change from human-induced changes such as pumping, as well as link observed trends in the past, current baseline maps, and expected future hydroclimatic changes to investigate groundwater quality patterns under future conditions still needs to be studied. As some geogenic compounds are harmful to human health and the environment, decrease drinking water quality and increase purification costs, a better understanding of the linkages between climate change and groundwater chemistry will be vital for future groundwater management in Denmark. The developed machine learning model and its potential for global upscaling could contribute to sustainable groundwater management worldwide.