Artificial intelligence helps forecasting groundwater levels

In a new project the Federal Institute for Geosciences and Natural Resources (BGR) and the Karlsruhe Institute of Technology (KIT) are developing methods for a nationwide forecast of groundwater levels. The researchers are focussing on methods of artificial intelligence (AI).

Artificial neural networks (ANN) are applied. Their suitability for modelling and predicting groundwater levels has already been proven frequently. Compared to conventional approaches such as process-based numerical modelling, the technology has advantages especially for cross-regional considerations. New developments in the area of deep learning give hope to significant improvements of existing approaches to predictions.

Input data used by an artificial neural network to predict groundwater levels Source: KIT & BGR

A reliable prediction of the groundwater level is, among others, the basis for the assessment of water availability for drinking water supply and irrigation in agriculture. It is also important for the delineation of potential zones of land subsidence due to extremely low groundwater levels in connection with droughts and water abstractions, as well as areas of potential high groundwater levels with impacts on the transport infrastructure, buildings and agricultural land. In this way, suitable avoidance and adaptation strategies can be designed.

The aim of the project is to develop a method that enables nationwide forecasts of groundwater levels and spring discharges, particularly with regard to extreme events such as droughts, heat waves or extreme precipitation. The periods considered in the forecast extend up to the year 2100 in order to be able to estimate the development of the groundwater levels in the course of climate change based on scenarios defined by the Intergovernmental Panel on Climate Change. In addition to groundwater data from the monitoring networks in the Federal States, the information base includes also weather and climate forecasts and climate projections from the Deutscher Wetterdienst (DWD, Germany's National Meteorological Service).

The forecasts will be made available for specific user groups through a specialist application on the BGR website.