Basin-scale evaluation of current and future climate influences on groundwater variations using satellite and model observations

Abstract

Study region: The Murray Darling Basin (MDB), Southeast Australia Study focus: An integrated approach involving satellite and model observations is applied to understand groundwater trends in drought-susceptible regions. This research advanced the understanding of hydrological and climatic factors responsible for the groundwater behaviour over the MDB using Gravity Recovery and Climate Experiment (GRACE), Global Land Data Assimilation System (GLDAS) data and in-situ groundwater records. Also, a 5-year forecast of groundwater trends over this region is made using feedforward neural network algorithm. New hydrological insights for the region: With the use of new data from the global land water storage model, GLDAS-CLSM-F2.5, and the Australian Water Outlook (AWO) model, an improved understanding of groundwater behaviour over the MDB is revealed. We found that deep drainage (DD), evapotranspiration (ET), and the Oceanic Nino Index (ONI) were major contributors to the drought period witnessed in the region from 2003 to 2009 and from 2013 to 2019, while runoff, El Nino Southern Oscillation (ENSO) and the Pacific Decadal Oscillation (PDO) contributed significantly to the wet periods witnessed from 2010 to 2012 and from 2020 to 2024. Also, we observed that the interaction between surface and groundwater storages in the region was minimal during drought episodes but was strongest during the wet seasons, which means that strong connections exist between them. This means that the MDB’s groundwater resources contribute to sustaining streams and surface water during the dry seasons and benefit from infiltration and recharge from surface water resources during the wet seasons. The feedforward neural network algorithm used to predict groundwater in the region showed that the MDB will witness an upward groundwater trend of 0.32 mm/month between 2024 and 2029 based on a 95 % confidence interval.