Groundwater recharge favorability modelling by diffuse logic paradigm

Abstract

Geographic information is uncertain, which means that the boundaries between different phenomena are blurred or there is heterogeneity within a class, due to differences between geological, pedological, geomorphological, vegetal features and so on. Methods based on artificial intelligence (AI) provide specific solutions to the fuzzy nature of the real world based on expert-knowledge. The uncertain nature of the processes that control groundwater recharge in watersheds allows these methods to be applied in groundwater management, supporting planning and decision-making related with water use and protection of vulnerable areas. The aim of this work was to define favourable areas for groundwater recharge from variables related variables samples near monitoring wells in a watershed in an outcrop area of the Guarani Aquifer System (GAS). Fuzzy logic was used to define an inference system capable of spatially extrapolating the point data for the entire watershed. The output was a map of favourability to recharge based on variables related to the texture and management of soil, terrain features and vegetation. The synthesis map support both planning and decision making on land use considering hydrological processes in its surface and subsurface interfaces. From the results achieved, the discussion on the importance of ethical choices in the hydrogeology deci-sion-making processes related to the use of AI-based methods is extended.