Human-induced arsenic pollution modeling in surface waters - An integrated approach using machine learning algorithms and environmental factors.

In recent years, assessment of sediment contamination by heavy metals, i.e., arsenic, has attracted the interest of scientists worldwide. The present study provides a new methodology to better understand the factors influencing surface water vulnerability to arsenic pollution by two advanced machine learning algorithms including boosted regression trees (BRT) and random forest (RF). Based on the sediment quality guidelines (Effects range low) polluted and non-polluted arsenic sediment samples were defined with concentrations >8 ppm and <8 ppm, respectively. Different conditioning factors such as topographical, lithology, erosion, hydrological, and anthropogenic factors were acquired to model surface waters' vulnerability to arsenic. We trained and validated the models using 70 and 30% of both polluted and non-polluted samples, respectively, and generated surface vulnerability maps. To verify the maps to arsenic pollution, the receiver operating characteristics (ROC) curve was implemented. The results approved the acceptable performance of the RF and BRT algorithms with an area under ROC values of 85% and 75.6%, respectively. Further, the findings showed higher importance of precipitation, slope aspect, distance from residential areas, and slope length in arsenic pollution in the modeling process. Erosion, lithology, and land use maps were introduced as the least important factors. The introduced methodology can be used to define the most vulnerable areas to arsenic pollution in advance and implement proper remediation actions to reduce the damages.

How do data-mining models consider arsenic contamination in sediments and variables importance?

Arsenic (As) is one of the most important dangerous elements as more than 100 million of people are exposed to risk, globally. The permissible threshold of As for drinking water is 10 µg/L according to both the WHO's drinking water guidelines and the Iranian national standard. However, several studies have indicated that As concentrations exceed this threshold value in several regions of Iran. This research evaluates an As-susceptible region, the Tajan River watershed, using the following data-mining models: multivariate adaptive regression splines (MARS), functional data analysis (FDA), support vector machine (SVM), generalized linear model (GLM), multivariate discriminant analysis (MDA), and gradient boosting machine (GBM). This study considers 12 factors for elevated As concentrations: land use, drainage density, profile curvature, plan curvature, slope length, slope degree, topographic wetness index, erosion, village density, distance from villages, precipitation, and lithology. The susceptibility mapping was conducted using training (70%) and validation (30%). The results of As contamination in sediment showed that classifications into 4 levels of concentration are very similar for two models of GLM and FDA. The GBM calculated the areas of highest arsenic contamination risk by MARS and SVM with percentages of 30.0% and 28.7%, respectively. FDA, GLM, MARS, and MDA models calculated the areas of lowest risk to be 3.3%, 23.0%, 72.0%, 25.2%, and 26.1%, respectively. The results of ROC curve reveal that the MARS, SVM, and MDA had the highest accuracies with area under the curve ROC values of 84.6%, 78.9%, and 79.5%, respectively. Land use, lithology, erosion, and elevation were the most important predictors of contamination potential with a value of 0.6, 0.59, 0.57, and 0.56, respectively. These are the most important factors. Finally, these data-mining methods can be used as appropriate, inexpensive, and feasible options to identify As-susceptible areas and can guide managers to reduce contamination in sediment of the environment and the food chain.