Detecting groundwater resources potential in Isfahan Lenjanat region using weights-of-evidence model

Identifying groundwater potential is a key requirement in groundwater conservation and management and is especially useful for engineers and designers looking for suitable locations for the exploitation of these resources. GIS-based methods have increased the accuracy and speed of groundwater surveys. Many researchers have focused on the use of GIS-based methods, remote sensing and machine learning, and statistical methods. These methods identified latent patterns in groundwater and obtain a nonlinear relationship between affecting variables. In this regard, the final groundwater potential map is determined using 12 effective variables based on ArcGIS and Weight-Of-Evidence (WOE) model for Lenjanat area, Isfahan Province.

Lenjanat study area is located in the southwest of Isfahan Province. It has a variable climate that is always affected by the central semi-arid and semi-humid areas of Chaharmahal and Bakhtiari Province. WOE model is based on standard deviation and used to predict the occurrence of accidents when there is enough data available to assess the relative importance of the layers using the statistical average. The Receiver Operating Characteristic (ROC) curve related to the groundwater potential mapping in the Lenjanat study area was prepared using SPSS software. The Area Under Curve (AUC) indicates the accuracy degree of the final map. The extent area of each curve indicates the model’s prediction ability. The value below the graph would be equal to unity in the best case and ideal scenario. This index is a suitable criterion to evaluate the model accuracy.

The areas with low elevation and low slope due to the high level of infiltration and water movement in the soil had a higher potential of groundwater, which is consistent with the results of the previous researchers. These areas had a higher groundwater potential in the east and flat terrains than in other directions. From the land use point of view, arable lands and areas with dense vegetation had the highest groundwater potential. The shorter the distance from the waterway and the higher the drainage density leads to the greater the groundwater potential. Areas with Aridsols soil have higher groundwater potential compared to the Entisol and Inceptisols categories. Soft alluvial sediments have a higher potential than other sediments. It is expected that the areas with high groundwater potential will be more in the eastern part of the region due to the high concentration of factors affecting the creation of potential groundwater conditions. The final groundwater potential map also determined these areas as high groundwater potential regions.

The results showed a higher sensitivity to land use and slope variables, which play an important role in water penetration. More than 60% of Lenjanat has moderate to very high groundwater potential. Identifying areas with high groundwater potential is important in determining areas for implementing management programs. It is suggested to use the WOE model with more input variables and other models to prepare a high accuracy potential map should be evaluated in future studies.