Yet, groundwater data are often incomplete or nonexistent. Therefore, it is a challenge to achieve a groundwater potential assessment.

This study developed methods to produce reliable groundwater potential maps with only digital elevation model-derived data as inputs. A case study area in Iran was selected to achieve this objective, and 13 factors were extracted from the digital elevation model.

A spring location dataset was obtained from the water sector organizations and fed into machine learning algorithms along with the non-spring locations.

In the absence of critical groundwater-related factors such as land use, lithology, soil, and fault-related parameters, the employed machine learning algorithms succeeded in extracting relationships among the digital elevation model-derived characteristics and groundwater potential.

We conclude that the developed methodology can produce initial information for groundwater exploitation in areas facing data scarcity. The proposed method can also generate necessary information for water sector managers and groundwater professionals to implement proper water resources plans. 

Application of extreme gradient boosting and parallel random forest algorithms for assessing groundwater spring potential using DEM-derived factors
(Journal of Hydrology, 589, p.125197.)

Photo by Tim Mossholder on Unsplash