Abstract: Knowledge of the natural baseline quality of groundwaters is an essential prerequisite for understanding pollution and for imposing regulatory limits. The natural baseline of groundwaters may show a range of concentrations depending on aquifer mineralogy, facies changes, flow paths and residence time. The geochemical controls on natural concentrations are discussed and an approach to defining baseline concentrations using geochemical and statistical tools is proposed. The approach is illustrated using a flowline from the Chalk aquifer in Berkshire, UK where aerobic and anaerobic sections of the aquifer are separately considered. The baseline concentrations for some elements are close to atmospheric values whereas others evolve through time-dependent water–rock interaction. Certain solutes (K, NH4+), often considered contaminants, reach naturally high concentrations due to geochemical controls; transition metal concentrations are generally low, although their concentrations may be modified by redox controls. It is recommended that the baseline approach be incorporated into future management strategies, notably monitoring.

Abstract: Predicting groundwater levels is pivotal in curbing overexploitation and ensuring effective water resource governance. However, groundwater level prediction is intricate, driven by dynamic nonlinear factors. To comprehend the dynamic interaction among these drivers, leveraging machine learning models can provide valuable insights. The drastic increase in computational capabilities has catalysed a substantial surge in the utilisation of machine learning-based solutions for effective groundwater management. The performance of these models highly depends on the selection of hyperparameters. The optimisation of hyperparameters is a complex process that often requires application-specific expertise for a skillful prediction. To mitigate the challenge posed by hyperparameter tuning’s problem-specific nature, we present an innovative approach by introducing the automated machine learning (AutoML-GWL) framework. This framework is specifically designed for precise groundwater level mapping. It seamlessly integrates the selection of best machine learning model and adeptly fine-tunes its hyperparameters by using Bayesian optimisation. We used long time series (1997-2018) data of precipitation, temperature, evaporation, soil type, relative humidity, and lag of groundwater level as input features to train the AutoML-GWL model while considering the influence of Land Use Land Cover (LULC) as a contextual factor. Among these input features, the lag of groundwater level emerged as the most relevant input feature. Once the model is trained, it performs well over the unseen data with a strong correlation of coefficient (R = 0.90), low root mean square error (RMSE = 1.22), and minimal bias = 0.23. Further, we compared the performance of the proposed AutoML-GWL with sixteen benchmark algorithms comprising baseline and novel algorithms. The AutoML-GWL outperforms all the benchmark algorithms. Furthermore, the proposed algorithm ranked first in Friedman’s statistical test, confirming its reliability. Moreover, we conducted a spatial distribution and uncertainty analysis for the proposed algorithm. The outcomes of this analysis affirmed that the AutoML-GWL can effectively manage data with spatial variations and demonstrates remarkable stability when faced with small uncertainties in the input parameters. This study holds significant promise in revolutionising groundwater management practices by establishing an automated framework for simulating groundwater levels for sustainable water resource management.

Abstract: In coastal salinized groundwater systems, contamination from various nitrate (NO3) inputs combined with complex hydrogeochemical processes make it difficult to distinguish NO3 sources and identify potential NO3 transformtation processes. Effective field-based NO3 studies in coastal areas are needed to improve the understanding of NO3 contamination dynamics in groundwater of such complex coastal systems. This study focuses on a typical Mediterranean coastal agricultural area, located in Tunisia, experiencing substantial NO3 contamination from multiple anthropogenic sources. Here, multiple isotopic tracers (δ18OH2O, δ2HH2O, δ15NNO3, δ18ONO3, and δ11B) combined with a Bayesian isotope MixSIAR model are used (i) to identify the major NO3 sources and their contributions, and (ii) to describe the potential NO3 transformation processes. The measured NO3 concentrations in groundwater are above the natural baseline threshold, suggesting anthropogenic influence. The measured isotopic composition of NO3 indicates that manure, soil organic matter, and sewage are the potential sources of NO3, while δ11B values constrain the NO3 contamination to manure; a finding that is supported by the results of MixSIAR model revealing that manure-derived NO3 dominates over other likely sources. Nitrate derived from manure in the study area is attributed to organic fertilizers used to promote crop growth, and livestock that deposit manure directly on the ground surface. Evidence for ongoing denitrification in groundwaters of the study area is supported by an enrichment in both 15N and 18O in the remaining NO3, although isotopic mass balances between the measured and the theoretical δ18ONO3 values also suggest the occurrence of nitrification. The simultaneous occurrence of these biogeochemical processes with heterogeneous distribution across the study area reflect the complexity of interactions within the investigated coastal aquifer. The multiple isotopic tracer approach used here can identify the effect of multiple NO3 anthropogenic activities in coastal environments, which is fundamental for sustainable groundwater resources management.