Comparing physics-based and data-driven models to predict groundwater levels in the Wairau Aquifer

Abstract

Observations of groundwater heads are frequently used to monitor the storage of aquifers and as decision variables for groundwater management. Groundwater levels can exhibit a strong seasonality and can be highly dynamic in alluvial gravel aquifers that are essentially extensions of the rivers that recharge them. The sustainable management of such aquifers is challenging and requires dynamic forecasting models to test their capabilities in history-matching and o evaluate their performance for extreme events that fall outside their respective calibration ranges.