Predicting groundwater levels using machine learning and time series on measurements taken from observation wells

Nowadays, sustainable management of all resources in the world has gained importance and a wide variety of efforts have been initiated globally, such as the 17 Goals of Sustainable Development carried out by the United Nations and Sustainable Development in our country. The sustainability of water, the indispensable source of life, which is increasingly under threat due to the increasing population and climate changes, is also critical. In this context, accurate estimation of water levels in aquifers, which constitute a large portion of fresh water resources, is of vital importance for the effective management and protection of water resources and for early recognition of water scarcity problems and taking precautions. Machine learning is a powerful tool for learning complex patterns and relationships from large data sets. Time series analysis allows analyzing data points over time and predicting future trends. In this study, ways to model the behavior of aquifers and predict water levels in aquifers more accurately and effectively using machine learning and time series analysis are discussed. As a result of the studies, it has been determined that applying data-based techniques to time series analyzes is very effective in estimating water levels in aquifers. These techniques can play an important role in developing water management policies, sustainable use of water resources and taking measures against possible water shortages, and can contribute to shaping future water security strategies by providing valuable insights to decision makers and researchers on water resources management.