Assessment of soil quality in a heavily fragmented micro-landscape induced by gully erosion

• DOI: 10.1016/j.geoderma.2023.116369
• Publication Date: 2023-02-09T18:35:42Z
• Authors: Xin Chen; Xin Zhang; Yujie Wei; Shu Zhang; Chongfa Cai; Zhonglu Guo; Junguang Wang
• Keywords: Silt; Soil Quality; Elevation (ballistics)

Soil quality degradation induced by erosion significantly inhibits sustainable development worldwide. For assessment of soil variations in an area with a heavily fragmented micro-landscape gully erosion, 16 indicators were tested laboratory settings and selected principal component analysis (PCA). Meanwhile, prediction was conducted the random forest (RF) model its derived from 3-dimensional structure landscape (resolution, 0.01 m) obtained unmanned aerial vehicle (UAV). During RF modelling, 80 % Quality Indices (SQIs) estimated PCA randomly as training data, remaining used to validate result. The optimal SQIs shown include Mnd, bulk density, silt content, cation exchange capacity (CEC). Additionally, PCA-calculated SQI ranging 0.33 0.85 decreased decreasing elevation erosional area. Moreover, spatial predicted satisfied accuracy (R2 = 0.83 ∼ 0.86; RMSE 0.03 0.04) comparable SQI. Overall, variation attributed (13.4 24.1 %), slope gradient (8.0 13.4 relief amplitude (9.8 12.9 terrain roughness index (10.3 11.9 %). This study confirmed excellent performance for prediction, also indicated that ultra-high-resolution (0.01 (UAV) competent tool areas complicated microtopography limited availability sampling.