Evaluation of machine learning algorithms for groundwater quality modeling

Abstract

Groundwater quality is measured through water sampling, and lab analysis. The eld-based measurements are costly and time-consuming when applied over a large domain. In this study, we developed a machine learning-based framework to map groundwater quality in an uncon ned aquifer in the north of Iran. Groundwater samples were provided from 248 monitoring wells across the region. The groundwater quality index (GWQI) in each well was measured and classi ed into four classes of Very poor, Poor, Good, and Excellent according to their cut-off values. Factors affecting groundwater quality, including distance to industrial centers, distance to residential areas, population density, aquifer transmissivity, precipitation, evaporation, geology, and elevation, were identi ed and prepared in the GIS environment. Six machine learning classi ers, including extreme gradient boosting (XGB), random forest (RF), support vector machine (SVM), arti cial neural networks (ANN), k-nearest neighbor (KNN), and Gaussian classi er model (GCM), were used to establish relationships between GWQI and its controlling factors. The algorithms were evaluated using the receiver operating characteristic curve (ROC) and statistical e ciencies (overall accuracy, precision, recall, and f-1 score). Accuracy assessment showed that ML algorithms provided high accuracy in predicting groundwater quality. However, RF was selected as the optimum model given its higher accuracy (overall accuracy, precision, and recall = 0.92; ROC = 0.95). The trained RF model was used to map GWQI classes across the entire region. Results showed that the Poor GWQI class is dominant in the study area and Good GWQI can be found in southwest. An area of Very Poor GWQI was observed in the north. Findings indicated that the distance to industrial locations is the main factor affecting groundwater quality in the area. The study provides a cost-effective methodology in groundwater quality modeling that can be duplicated in other regions with similar hydrological and geo-logical settings.

Enhancing Groundwater Quality Evaluation Using Associative Rule Mining Technique with Random Forest Split Gini Indexing Algorithm for Nitrate Concentration Analysis

“…Further, the outcomes of the study have been suggested that the ensemble ML algorithms has been utilized to estimate and determine the WQI with accuracy. A ML-based framework has been developed in [25], to map the quality of groundwater in an released aquifer in North Iran. In this study, groundwater samples have been conferred from 248 monitoring wells through the area of North Iran.…”

Predicting Rate Constants of Reactive Chlorine Species toward Organic Compounds by Combining Machine Learning and Quantum Chemical Calculation

“…The advancement of machine learning (ML) methods has garnered significant attention due to their robust algorithms, which can greatly enhance the predictive capabilities of models. − Among ML approaches, tree-based algorithms, such as XGBoost, CatBoost, and LightGBM, have been successfully applied to develop QSAR models in environmental fields. , These algorithms have been applied in various fields, such as to retrieve daily PM2.5 concentrations, evaluate groundwater quality, and design adsorbent materials . Additionaly, they have been used for predicting the k of SO 4 •– , HClO, O 3 , and ClO 2 .…”