Clean and accurate soil quality monitoring in mining areas under environmental rehabilitation in the Eastern Brazilian Amazon
- 20 March 2024
- journal article
- research article
- Published by Springer Nature in Environmental Monitoring and Assessment
- Vol. 196 (4) , 1-18
Abstract
Soil quality monitoring in mining rehabilitation areas is a crucial step to validate the effectiveness of the adopted recovery strategy, especially in critical areas for environmental conservation, such as the Brazilian Amazon. The use of portable X-ray fluorescence (pXRF) spectrometry allows a rapid quantification of several soil chemical elements, with low cost and without residue generation, being an alternative for clean and accurate environmental monitoring. Thus, this work aimed to assess soil quality in mining areas with different stages of environmental rehabilitation based on predictions of soil fertility properties through pXRF along with four machine learning algorithms (projection pursuit regression, PPR; support vector machine, SVM; cubist regression, CR; and random forest, RF) in the Eastern Brazilian Amazon. Sandstone and iron mines in different chronological stages of rehabilitation (initial, intermediate, and advanced) were evaluated, in addition to non-rehabilitated and native forest areas. A total of 81 soil samples (26 from sandstone mine and 55 from iron mine) were analyzed by both traditional wet-chemistry methods and pXRF. The available/exchangeable contents of K, Ca, B, Fe, and Al, in addition to H+Al, cation exchange capacity at pH = 7, Al saturation, soil organic matter, pH, sum of bases, base saturation, clay, and sand were accurately predicted (R2 > 0.70) using pXRF data, with emphasis on the prediction of Fe (R2 = 0.93), clay content (R2 = 0.81), H+Al (R2 = 0.81), and K+ (R2 = 0.85). The best predictive models were developed by RF and CR (86%) and when considering pXRF data + mining area + stage of rehabilitation (73%). The results highlight the potential of pXRF to accurately assess soil properties in environmental rehabilitation areas in the Amazon region (yet scarcely evaluated under this approach), promoting a more agile and cheaper preliminary diagnosis compared to traditional methods.Keywords
Funding Information
- CNPq
This publication has 64 references indexed in Scilit:
- Bioavailability of copper and nickel in naturally metal-enriched soils of Carajás Mining Province, Eastern Amazon, BrazilEnvironmental Monitoring and Assessment, 2021
- Assessment of iron-rich tailings via portable X-ray fluorescence spectrometry: the Mariana dam disaster, southeast BrazilEnvironmental Monitoring and Assessment, 2021
- Prediction of multiple soil fertility parameters using VisNIR spectroscopy and PXRF spectrometrySoil Science Society of America Journal, 2021
- Active rehabilitation of Amazonian sand mines converges soils, plant communities and environmental status to their predisturbance levelsLand Degradation & Development, 2019
- Source and background threshold values of potentially toxic elements in soils by multivariate statistics and GIS-based mapping: a high density sampling survey in the Parauapebas basin, Brazilian AmazonEnvironmental Geochemistry and Health, 2019
- Natural concentrations and reference values of heavy metals in sedimentary soils in the Brazilian AmazonEnvironmental Monitoring and Assessment, 2018
- Synthesis of proximal sensing, terrain analysis, and parent material information for available micronutrient prediction in tropical soilsPrecision Agriculture, 2018
- Soil organic matter doubles the cation exchange capacity of tropical soil under no‐till farming in BrazilJournal of the Science of Food and Agriculture, 2018
- Support Vector Machine: Principles, Parameters, and Applications2017
- Advances in Portable X-ray Fluorescence (PXRF) for Environmental, Pedological, and Agronomic Applications2014