Spatial variability of reconstructed soil properties and the optimization of sampling number for reclaimed land monitoring in an opencast coal mine

详细信息    查看全文

• 作者：Jinman Wang ; Ruixuan Yang ; Yu Feng
• 关键词：Land reclamation ; Soil quality ; Mining activities ; Geostatistics ; Loess Plateau
• 刊名：Arabian Journal of Geosciences
• 出版年：2017
• 出版时间：January 2017
• 年：2017
• 卷：10
• 期：2
• 全文大小：
• 刊物类别：Earth and Environmental Science
• 刊物主题：Earth Sciences, general;
• 出版者：Springer Berlin Heidelberg
• ISSN：1866-7538
• 卷排序：10

文摘

Drastically disturbed soils caused by opencast mining can result in the severe loss of soil structure and increase in soil compactness. To assess the effects of mining activities on reconstructed soils and to track the changes in reclaimed soil properties, the variability of soil properties (soil particle distribution, penetration resistance (PR), pH, and total dissolved salt (TDS)) in the Shanxi Pingshuo Antaibao opencast coal-mine inner dump after dumping and before reclamation was analyzed using a geostatistics method, and the number of soil monitoring points after mined land reclamation was determined. Soil samples were equally collected at 78 sampling sites in the study area with an area of 0.44 km2. Soil particle distribution had moderate variability, except for silt content at the depth of 0–20 cm with a low variability and sand content at the depth of 20–40 cm with a high variability. The pH showed a low variability, and TDS had moderate variability at all depths. The variability of PR was high at the depth of 0–20 cm and moderate at the depth of 20–40 cm. There was no clear trend in the variance with increasing depth for the soil properties. Interpolation using kriging displayed a high heterogeneity of the reconstructed soil properties, and the spatial structure of the original landform was partially or completely destroyed. The root-mean-square error (RMSE) can be used to determine the number of sampling points for soil properties, and 40 is the ideal sampling number for the study site based on cross-validation.