Assessment of Groundwater-Level Monitoring Network in Irrigated Regions with a Complex Aquifer System Using Information Theory

Abstract

The present study focuses on the assessment of groundwater level monitoring networks (GWLMN) by examining the statistical relationship between groundwater level (GWL) uncertainty and sustainable groundwater indicators (SGI). In this study, four SGIs are considered that are the key groundwater dynamics drivers, which include aquifer properties, a command area, a noncommand area, and the cropping season. The entropy-based theory has been innovatively used to estimate the uncertainty in random variables with SGI and existing observation wells (OWs). Entropy measures estimated in the analysis are marginal entropy (ME), joint entropy (JE), mutual entropy or transinformation (TE), and conditional entropy (CE). The applicability of the entropy-based theory to complex aquifer systems with varying hydrological characteristics is analyzed, estimating the optimum distance between two OWs, temporal frequencies, and priority zones for monitoring. The proposed framework has been used to analyze data from 30 OWs obtained from the Wainganga subbasin with an area of 3,320 km2 in India. Criteria considered for selecting suitable OWs for monitoring are (1) the lowest uncertainty (ME), and (2) the representative monitoring location based on the SGI. OWs are considered significant for monitoring if they fail to meet Criterion 1 or 2. The results show that out of the 30 OWs in the existing network, 28 OWs are found significant for monitoring purposes. The outcome of priority zoning was compared and examined using groundwater recharge and irrigation wells density. The results of this study will provide useful insights for computing the spatial and temporal uncertainty in GWLMNs and can help to assess the suitability of the network.