monitoring lead contamination by integrating environmental indices and random forest-based digital soil mapping in shiraz urban watershed, iran

The rapid increase in population and economic expansion has resulted in the infiltration of environmental pollutants, particularly heavy metals, into the soil, presenting a substantial threat to public well-being and reliability of food security. consequently, awareness and evaluation of these elements are key to assessing soil quality and related risks. in this study, machine learning modeling (random forest model) and digital mapping were employed to quantify and model lead (pb) contamination using various environmental indices in a section of the urban watershed of shiraz. for this purpose, 148 soil samples were systematically gathered from a depth of 0 to 20 cm utilizing a randomized sampling approach. after sample preparation, the total pb content in the soil was determined applying standard analytical methods. pb contamination risk assessment was conducted using three environmental indices: geo-accumulation index (igeo), enrichment factor (ef), and contamination factor (cƒ). the results indicated that all analyzed samples exhibited total pb concentrations (mean: 7.78 mg/kg) below the recommended standard levels for iran. based on the igeo (range: 1.54–4.72), the samples were categorized as moderately to severely contaminated. the ef (range: 4.35–39.65) classified the samples as moderately, highly, and extremely enriched, while the cƒ (range: 3.37–29.63) placed the samples in the high to very high contamination category. the interpretation of environmental indices confirmed low to moderate levels of pb contamination, primarily influenced by anthropogenic activities. therefore, to ensure sustainable food security, continuous monitoring of pb concentration variations in the studied soils is essential.