GIS-based groundwater potential mapping using boosted regression tree, classification and regression tree, and random forest machine learning models in Iran

Abstract

Groundwater is considered one of the most valuable fresh water resources. The main objective of this study was to produce groundwater spring potential maps in the Koohrang Watershed, Chaharmahal-e-Bakhtiari Province, Iran, using three machine learning models: boosted regression tree (BRT), classification and regression tree (CART), and random forest (RF). Thirteen hydrological-geological-physiographical (HGP) factors that influence locations of springs were considered in this research. These factors include slope degree, slope aspect, altitude, topographic wetness index (TWI), slope length (LS), plan curvature, profile curvature, distance to rivers, distance to faults, lithology, land use, drainage density, and fault density. Subsequently, groundwater spring potential was modeled and mapped using CART, RF, and BRT algorithms. The predicted results from the three models were validated using the receiver operating characteristics curve (ROC). From 864 springs identified, 605 (≈70 %) locations were used for the spring potential mapping, while the remaining 259 (≈30 %) springs were used for the model validation. The area under the curve (AUC) for the BRT model was calculated as 0.8103 and for CART and RF the AUC were 0.7870 and 0.7119, respectively. Therefore, it was concluded that the BRT model produced the best prediction results while predicting locations of springs followed by CART and RF models, respectively. Geospatially integrated BRT, CART, and RF methods proved to be useful in generating the spring potential map (SPM) with reasonable accuracy.

Geophysical evaluation of groundwater potential in part of southwestern Basement Complex terrain of Nigeria

“…Several statistical methods can also be adopted for groundwater mapping where adequate information on different influencing parameters to groundwater accumulation and movement are available. These include frequency ratio (Davoodi et al 2013), multi-criteria decision evaluation (Murthy and Mamo 2009;Kumar et al 2014), logistic regression model (Ozdemir 2011), weights-of-evidence model (Ozdemir 2011;Pourtaghi and Pourghasemi 2014), random forest model (Rahmati et al 2016Naghibi et al 2016, maximum entropy model (Rahmati et al 2016), boosted regression tree (Naghibi et al 2016;Naghibi and Pourghasemi 2015), classification and regression tree (Naghibi et al 2016), multivariate adaptive regression spline model (Zabihi et al 2016), certainty factor model (Zabihi et al 2016), evidential belief function (Pourghasemi and Beheshtirad 2015;Naghibi and Pourghasemi 2015), and generalized linear model (Naghibi and Pourghasemi 2015). These information are lacking in many third world country hence proper understanding of hydrogeological characteristics for successful exploitation of groundwater in basement areas depend largely on geophysical methods.…”

GIS-based model of groundwater occurrence using geological and hydrogeological data in Precambrian Oban Massif southeastern Nigeria

“…There are various statistical methods that were adopted by various authors for groundwater potential mapping elsewhere such as: frequency ratio (Guru et al 2017;Al-Zuhairy et al 2017;Razandi et al 2015), Analytical hierarchical process Chowdhury et al 2009;Razandi et al 2015), binary logistic regression method (Ozdemir 2011a), weight of evidence model (Ghorbani et al 2017;Tahmassebipoor et al 2016;Ozdemir 2011b), k-nearest neighbor (Naghibi and Dashtpagerdi 2017), Demster-Shafer model , machine learning model/artificial neural network (Lee et al 2012), boosted regression tree BRT (Naghibi and Pourghasemi 2015;Naghibi et al 2016), multivariate adaptive regression spline (Zabihi et al 2016), maximum, entropy model ), generalized adaptive model (Falah et al 2016), random forest model (Rahmati et al 2018), and other GIS-based models such as earth observation and entropy weighted linear aggregate novel approach (Bandyopadhyay et al 2007;Al-Abadi et al 2016). In the above studies, themes such as vegetation, land use/land cover, hydrogeomorphology, drainage, lithology, subsurface lithology, structure, and slope were interpreted to infer groundwater potential.…”

Evaluation of Flashover Voltage Levels of Contaminated Hydrophobic Polymer Insulators Using Regression Trees, Neural Networks, and Adaptive Neuro-Fuzzy

“…In [20], a fuzzy Logic (FL) model has been applied in order to estimate the critical flashover voltage on polluted insulators. Recently, the least squares regression tree has been used to describe the groundwater potential mapping [21].…”