Spectroscopic Based Quantitative Mapping Of Contaminant Elements In Dumped Soils Of A Copper Mine

Possibility of mapping the distribution of Arsenic and Chromium in a mining area was investigated using combination of (VNIR) reflectance spectroscopy and geostatistical analysis. Fifty five soil samples were gathered from a waste dump at Sarcheshmeh copper mine and VNIR reflectance spectra were measured in a laboratory. Savitzky-Golay first derivative was used as the main pre-processing method before developing Genetic Algorithm Partial Least Squares Regression (GA-PLSR) and PLSR models for predicting toxic elements concentrations. Physicochemical mechanism that allows the prediction with reflectance spectroscopy was also investigated and it was found that, elements sorption by spectrally active Fe and clay contents of soil was the major mechanism helping the prediction of spectrally featureless As and Cr. Positive relationships were observed between performance of predicting models and iron and clay contents of the samples. Comparing to PLSR, higher prediction performances of both toxic elements concentrations were obtained by applying GA-PLSR model. Furthermore, similar spatial patterns for soil pollution hotspots were observed by geostatistical interpolation (kriging) of chemically measured and models' predicted values. Results demonstrated that the amount and spatial variability of arsenic and chromium can be determined using VNIR spectroscopy and geostatistics in Sarcheshmeh mine's waste dump.