| Autor: |
Ferreira GWD; Department of Forest Sciences, Federal University of Lavras, Minas Gerais State, Doutor Sylvio Menicucci Avenue, Lavras, 37200-900, Brazil. gferreira@uga.edu.; Savannah River Ecology, University of Georgia, P O Drawer E, SC, Aiken, 29802, USA. gferreira@uga.edu., Ribeiro BT; Department of Soil Science, Federal University of Lavras, Minas Gerais State, Doutor Sylvio Menicucci Avenue, Lavras, 37200-900, Brazil.; Department of Plant and Soil Science, Texas Tech University, Bayer Plant Science Building, Room 211A, 2911 15th Street, Lubbock, TX, 79409, USA., Weindorf DC; Department of Plant and Soil Science, Texas Tech University, Bayer Plant Science Building, Room 211A, 2911 15th Street, Lubbock, TX, 79409, USA.; Department of Earth and Atmospheric Sciences, Central Michigan University, Mount Pleasant, MI, 48859, USA., Teixeira BI; Department of Forest Sciences, Federal University of Lavras, Minas Gerais State, Doutor Sylvio Menicucci Avenue, Lavras, 37200-900, Brazil., Chakraborty S; Agricultural and Food Engineering Department, Indian Institute of Technology , Kharagpur, West Bengal, 721302, India., Li B; Department of Experimental Statistics, Louisiana State University, Baton Rouge, LA, 70802, USA., Guilherme LRG; Department of Soil Science, Federal University of Lavras, Minas Gerais State, Doutor Sylvio Menicucci Avenue, Lavras, 37200-900, Brazil., Scolforo JRS; Department of Forest Sciences, Federal University of Lavras, Minas Gerais State, Doutor Sylvio Menicucci Avenue, Lavras, 37200-900, Brazil. |
| Abstrakt: |
On November 5, 2015, the Fundão dam collapsed and released > 60 million m 3 of iron-rich mining sediments into the Doce river basin, covering >1000 ha of floodplain soils across ~80 km from the rupture. The characterization of alluvial mud covering and/or mixed with native soil is a priority for successful environmental rehabilitation. Portable X-ray fluorescence (pXRF) spectrometry was used to (1) assess the elemental composition of native soils and alluvial mud across impacted riparian areas; and 2) predict fertility properties of the mud and soils that are crucial for environmental rehabilitation and vegetation establishment (e.g., pH, available macro and micronutrients, cation exchange capacity, organic matter). Native soils and alluvial mud were sampled across impacted areas and analyzed via pXRF and conventional laboratory methods. Random forest (RF) regression was used to predict fertility properties using pXRF data for pooled soil and alluvial mud samples. Mud and native surrounding soils were clearly differentiated based on chemical properties determined via pXRF (mainly SiO 2 , Al 2 O 3 , Fe 2 O 3 , TiO 2 , and MnO). The pXRF data and RF models successfully predicted pH for pooled samples (R 2 = 0.80). Moderate predictions were obtained for soil organic matter (R 2 = 0.53) and cation exchange capacity (R 2 = 0.54). Considering the extent of impacted area and efforts required for successful environmental rehabilitation, the pXRF spectrometer showed great potential for screening impacted areas. It can assess total elemental composition, differentiate alluvial mud from native soils, and reasonably predict related fertility properties in pooled heterogeneous substrates (native soil + mud + river sediments). |
Full text from SpringerLink