Environmental variables controlling soil organic carbon in top- and sub-soils in karst region of southwestern China. (July 2018)

• Record Type: Journal Article
• Authors: Zhu, Jie; Wu, Wei; Liu, Hong-Bin
• Is Part Of: Ecological indicators. Volume 90(2018)
• Page Start: 624
• Page End: 632
• Publication Date: 2018-07
• Subjects: Soil organic carbon content -- Soil horizon -- Environmental variable -- Classification and regression tree -- Importance; Environmental monitoring -- Periodicals; Environmental management -- Periodicals; Environmental impact analysis -- Periodicals; Environmental risk assessment -- Periodicals; Sustainable development -- Periodicals; 333.71405
• DOI: 10.1016/j.ecolind.2018.03.073
• Languages: English
• ISSNs: 1470-160X

Abstract

Highlights: SOC content varied greatly in vertical dimensions. CARTs were used to explore the key variables affecting SOC content in horizons. Topography had the greatest impact on SOC content in the A and B horizons. Climate had the greatest impact on SOC content in the C horizon. Soil moisture-related variables were the primary factors influencing SOC content. Abstract: Detailed information about the variability of soil organic carbon (SOC) content in different horizons is vital for the effective management of soil fertility and can improve understanding of SOC accumulation and decomposition. However, few studies have explored the variations of SOC content of cultivated lands in different horizons at the regional scale, particularly in a karst region in southwestern China. In this work, 235 soil samples were collected from the A to C horizons at 80 sites over the study area. Classical statistics and classification and regression trees (CARTs) were applied to investigate the critical environmental variables (topography, climate and parent material) that control SOC content variability in various horizons. Results indicated that SOC content decreased and its variability increased with an increase in horizon depth. The SOC content in the three horizons was predicted well by the CART models. The mean absolute prediction error, root mean square error and coefficient of determination ranged from 2.38 to 3.85 g kg −1, 3.23 to 5.2 g kg −1 and 0.5837 to 0.856, respectively. The importance of the environmental variables in SOC content varied with the horizons. In the A horizon, topographical variables, such as terrain wetness index (TWI), terrain ruggedness index and slope, were the key factors affecting SOC content variability. In the B horizon, the topography still showed the primary influence by elevation and TWI, meanwhile the importance of the parent material strengthened. Climate variables had the greatest impact on SOC content in the C horizon. TWI and precipitation that directly influenced soil moisture were the primary factors controlling SOC content in the three horizons.