Improving Results of Existing Groundwater Numerical Models Using Machine Learning Techniques: A Review

This paper presents a review of papers specifically focused on the use both numerical and machine learning methods for groundwater level modelling. In reviewed papers, models (also called data-driven models) are used to improve prediction or speed process existing When long runtimes inhibit models, can be valid alternative, capable reducing time model development calibration without sacrificing accuracy detail in forecasting. The results this highlight that do not offer complete representation physical system, such as flux estimates total water balance and, thus, cannot substitute large study areas; however, they affordable tools predictions at specific observation wells. Numerical successfully complementary each other powerful management tool. techniques whereas allow us understand system select proper input variables models. Machine integrated decision-making processes when rapid effective solutions need considered. Finally, computationally efficient correct head error