Tree-Based Modeling Methods to Predict Nitrate Exceedances in the Ogallala Aquifer in Texas

Abstract

The performance of four tree-based classification techniques-classification and regression trees (CART), multi-adaptive regression splines (MARS), random forests (RF) and gradient boosting trees (GBT) were compared against the commonly used logistic regression (LR) analysis to assess aquifer vulnerability in the Ogallala Aquifer of Texas. The results indicate that the tree-based models performed better than the logistic regression model, as they were able to locally refine nitrate exceedance probabilities. RF exhibited the best generalizable capabilities. The CART model did better in predicting non-exceedances. Nitrate exceedances were sensitive to well depths-an indicator of aquifer redox conditions, which, in turn, was controlled by alkalinity increases brought forth by the dissolution of calcium carbonate. The clay content of soils and soil organic matter, which serve as indicators of agriculture activities, were also noted to have significant influences on nitrate exceedances. Likely nitrogen releases from confined animal feedlot operations in the northeast portions of the study area also appeared to be locally important. Integrated soil, hydrogeological and geochemical datasets, in conjunction with tree-based methods, help elucidate processes controlling nitrate exceedances. Overall, tree-based models offer flexible, transparent approaches for mapping nitrate exceedances, identifying underlying mechanisms and prioritizing monitoring activities.Water 2020, 12, 1023 2 of 27 worldwide [5,6]. Intensification of agricultural activities for both food and energy will further increase the risks of nitrate contamination in aquifers across the world [7][8][9].Nitrate is mobile and fairly recalcitrant, especially in shallow groundwater systems that typically tend to be under oxidizing conditions. Nitrate exhibits the ability to spread over large areas and cannot be treated in-situ using conventional plume scale treatment technologies [10]. Therefore, individual homeowners are often required to install costly point-of-use treatment systems to mitigate nitrate risks arising from the ingestion of contaminated groundwater [11,12]. However, as nitrate is colorless and odorless, many people do not realize the risk of nitrate contamination and are unwittingly exposed to elevated levels of nitrate over long periods of time [13]. Therefore, nitrate contamination must be prevented through proper land management practices. Additionally, areas with a high susceptibility to nitrate pollution must be carefully delineated, with the goal of increasing public awareness regarding elevated health risks arising from nitrate exposures. Such an effort is also useful to prioritize monitoring activities and ensure that the limited fiscal and logistic resources are being used in a prudent manner.Mapping the susceptibility of aquifers to nitrate contamination is an essential step in mitigating and managing nitrate contamination. Multi-criteria decision making (MCDM) methods, such as DRASTIC [14], have been widely used to map aquifer vulnerabili...

Examining Relationships Between Groundwater Nitrate Concentrations in Drinking Water and Landscape Characteristics to Understand Health Risks

“…Here, CART analysis of groundwater nitrate concentrations in small watersheds of Michigan's lower peninsula, a midwestern glacial landscape, identified aquifer recharge, SOC, agricultural nitrogen sources, and soil drainage properties (hydraulic conductivity) as key factors. This is generally consistent with other decision‐tree‐based studies throughout the United States that find variables related to sources, soil properties, and recharge as important factors (Burow et al., 2010; Messier et al., 2019; Pennino et al., 2020; Ransom et al., 2017; Tesoriero et al., 2017; Uddameri et al., 2020). Burow et al.…”

“…Soil organic matter, a similar variable not quantified in this study, was also identified as an important variable correlated to agricultural activity and clay content by Uddameri et al. (2020).…”

“…Only those soils within the bottom 20% of SOC levels (generally <745-775 gC/m 2 in the top 5 cm of soil) had higher probabilities of detection or elevated groundwater nitrate. Soil organic matter, a similar variable not quantified in this study, was also identified as an important variable correlated to agricultural activity and clay content by Uddameri et al (2020).…”

Comparative Analysis of Artificial Intelligence Models for Accurate Estimation of Groundwater Nitrate Concentration

“…Although Noi et al [ 30 ] stated that Cubist regression and random forest algorithms have a good performance in estimating daily air surface temperature from dynamic combinations of MODIS LST data, Bayesian ANN notes to developing standard networks with posterior inference to regard a probability distribution of weights instead of a single set of weights [ 31 ]. Sahoo et al [ 32 ] put forward Bayesian methods for water quality assessment and presented that the quality of water was improved during dry seasons more than during wet seasons owing to the dilution of pollutants [ 33 , 34 , 35 , 36 ].…”

Neural Network and Random Forest-Based Analyses of the Performance of Community Drinking Water Arsenic Treatment Plants

“…Furthermore, random forest (RF) is an ensemble machine learning technique created by the fusion of bootstrapping aggregation and the algorithms of classification and regression trees (CART) [17]. RF is used in various fields, such as model prediction [16,18,19], attribute classification [20], and important feature selection [17,21]. The application of these machine learning tools to the large datasets of field-scale plants may be considered an efficient tool for performance evaluation and decision making.…”