Spatial Variation of Soil Organic Carbon Stocks in Relation to Land Management Practices: A Case Study From Henan Province, China

A better understanding of spatial variation of soil organic carbon stocks (SOCS) is indispensable for sustainable soil management. This study is to discuss the spatial variation of SOCS and its drivers in Xixia County within Henan Province. A set of 447 soil samples is included for Random Forest models calibration in 2008 and 2013. The results show that land management practices accounted for 67% (2008) and 55% (2013) in SOCS prediction, and the land management practices also significantly improve the prediction accuracy of SOCS based on R2, RMSE and ME values. The high SOCS values are concentrated in the northern and southern part of Xixia County in 2008 and 2013, and there is an average of about 0.59 Mg ha−1 of SOCS decrease over 5 years. The SOCS in forest soils decreases from 24.8 to 22.5 Mg ha−1 from 2008 to 2013, and that of grassland decreases as well from 24.1 to 23.9 Mg ha−1. The decline of SOCS is mainly due to climate and land management practices, which are controlled by mean annual evapotranspiration, mean annual precipitation and phosphorus fertiliser and straw turnover. However, the SOCS of cultivated land increases 9.6% due to the fertilisation and straw turnover from 2008 to 2013, which likely suppresses the priming effect on SOC. We find that proper fertilisation and straw turnover can maintain a stable soil carbon pool in cultivated land, which is important for agricultural ecosystem sustainable development. This study provides guidance for regulating land management practices toward increasing SOC storage, and also supplies methodologies for SOCS prediction.