Hydrologic connectivity assessment among natural hydrosystem components using emerging entropy and evolutionary computing methods

Understanding the interactions among the natural components of hydrological systems is essential for managing water resources, addressing environmental concerns, and mitigating the impacts of natural events such as floods and droughts. In this study, long-term (1993–2019) hydrologic connectivity among precipitation, humidity, evaporation, surface runoff, minimum/maximum temperature, and their interactions with groundwater level (GWL) in the southeastern Caspian Sea region was assessed using mutual information theory and the state-of-the-art jittered genetic programming approach. While the former was used to find dominant components, their effective time delay, and the power of potential nonlinear interactions, the latter was utilized to extract an explicit relation between the GWL and the most influential components. The data were gathered from several piezometers, meteorological stations, and a stream gauge available in the study area. The results showed an overall positive trend in the GWL with an increasing rate since 2007 that reflects the influence of artificial recharge infrastructures built in the study area. Statistical connectivity analyses demonstrated that historical precipitation and streamflow series have the least impact on the temporal variation of the average GWL.