Revealing the 2D–scale, location–specific variations of soil properties in the coal mining area of Changhe watershed, China. (27th June 2020)

• Record Type: Journal Article
• Authors: Zhu, Hongfen; Bi, Rutian; Sun, Ruipeng; Xu, Zhanjun; Lv, Chunjuan; Yang, Jason
• Is Part Of: Land degradation & development. Volume 31:Number 18(2020)
• Page Start: 2775
• Page End: 2788
• Publication Date: 2020-06-27
• Subjects: bidimensional intrinsic mode function (BIMF) -- coal mining area -- land use type -- soil organic matter (SOM) -- two‐dimensional empirical mode decomposition (2DEMD); Land degradation -- Periodicals; Soil conservation -- Periodicals; Reclamation of land -- Periodicals; Land use -- Periodicals; Economic development -- Environmental aspects -- Periodicals; 333.7315
• DOI: 10.1002/ldr.3639
• Languages: English
• ISSNs: 1085-3278

Abstract

Abstract: Many researches have been conducted to understand the scale–location specific variations of soil properties in one‐dimensional soil samples. However, these results did not enable us to properly understand the 2D (two‐dimensional) distributions of soil characteristics, especially in the coal mining areas. Therefore, the objective of this study is to reveal the scale–location dependent variation of two‐dimensional patterns for soil properties and to develop a soil organic matter (SOM) prediction based on those scale–location effects in the coal mining area. To this end, SOM, porosity, and silt were measured in the top (0–20 cm), middle (20–40 cm), and bottom (40–60 cm)layers of soil in Changhe watershed, China. The scale‐specific spatial patterns of soil properties were extracted using two‐dimensional empirical mode decomposition (2DEMD), and SOM was predicted using stepwise multiple linear regression (SMLR) and 2DEMDGS . It was found that the spatial distributions of SOM residue (>24 km) at middle layer and porosity residue at top and bottom layers could reveal the differentiation between coal mining and non‐coal mining areas. The scale‐effects on the relationships between SOM and its covariates was stronger than that of land uses, and the relationships of SOM with the covariates in different land use types were distinctive at top layer, and were similar at bottom layer. 2DEMDGS performed better than SMLR on SOM prediction for its involvement of scale and location effects. Therefore, the relationship between SOM and influencing factors is a function of scale, location, and land uses activities, and the scale‐ and location‐dependent method should be considered for SOM prediction.