Groundwater potentiality mapping using ensemble machine learning algorithms for sustainable groundwater management

Abstract

PurposeThe present study aims to construct ensemble machine learning (EML) algorithms for groundwater potentiality mapping (GPM) in the Teesta River basin of Bangladesh, including random forest (RF) and random subspace (RSS).Design/methodology/approachThe RF and RSS models have been implemented for integrating 14 selected groundwater condition parametres with groundwater inventories for generating GPMs. The GPM were then validated using the empirical and bionormal receiver operating characteristics (ROC) curve.FindingsThe very high (831–1200 km2) and high groundwater potential areas (521–680 km2) were predicted using EML algorithms. The RSS (AUC-0.892) model outperformed RF model based on ROC's area under curve (AUC).Originality/valueTwo new EML models have been constructed for GPM. These findings will aid in proposing sustainable water resource management plans.

A national-level study on groundwater potentiality mapping using a hybrid machine learning models under the scenario of climate change

“…Because of the over-extraction of groundwater, unregulated surface water contamination, and climate change, the country has a year-round water crisis (Ahmad et al, 2014;Abedin et al, 2019). Only in the last few years has the demand increased for groundwater, and continued exploitation in Bangladesh piqued the interest of experts and planners (Hasan et al, 2021;Sarkar et al, 2022). As a result, several studies on groundwater depletion, contamination, and potentiality mapping have been done (Nowreen et al, 2020;Hofste et al, 2019;Garner et al, 2017;Sarkar et al, 2022).…”

“…Only in the last few years has the demand increased for groundwater, and continued exploitation in Bangladesh piqued the interest of experts and planners (Hasan et al, 2021;Sarkar et al, 2022). As a result, several studies on groundwater depletion, contamination, and potentiality mapping have been done (Nowreen et al, 2020;Hofste et al, 2019;Garner et al, 2017;Sarkar et al, 2022).…”

“…Since nonlinear connections between groundwater and conditioning factors of a dynamic nature cannot be predicted using normal statistical approaches (Sahoo et al, 2017;Guillaume et al, 2016). Advanced statistical methods and machine learning approaches have both been introduced by the researchers into the GWP models (Sahoo et al, 2017;Sarkar et al, 2022;Tran et al, 2021;Hasanuzzaman et al, 2022). Several machine learning techniques like neuro-fuzzy, decision trees, ANN and SVM, etc., have been applied for the mapping of GWP (Panahi et al, 2020;Khosravi et al, 2018;Sarkar et al, 2022;).…”

“…Advanced statistical methods and machine learning approaches have both been introduced by the researchers into the GWP models (Sahoo et al, 2017;Sarkar et al, 2022;Tran et al, 2021;Hasanuzzaman et al, 2022). Several machine learning techniques like neuro-fuzzy, decision trees, ANN and SVM, etc., have been applied for the mapping of GWP (Panahi et al, 2020;Khosravi et al, 2018;Sarkar et al, 2022;). Evidence-based methodologies like evidentiary belief function, logistic regression (LR), probability-frequency ratio (FR), etc., have also been used by several studies for GWP mapping (Rizeei et al, 2019;Park et al, 2014;Pradhan et al, 2021).…”

Soil Infiltration Rate Impact on Water Quality Modeled Using Random Forest Regression

“…Throughout the past several decades, many prediction algorithms effectively used civil and environmental engineering [5,6]. Numerous linear regression techniques were used to estimate and forecast infiltration rates.…”