Spatio-termporal Characteristics and Pedotransfer of Soil Penetration Resistance in Eroded Sloping Farmland of Black Soil Region, China

Abstract

[Objective] To delve deeper into the spatiotemporal characteristics and pivotal driving factors underlying the hardness of black soil in eroded sloping farmlands. [Methods] Based on the dynamic monitoring of soil penetration resistance across 110 sample points in typical water erosion sloping farmlands, coupled with measurements of pertinent factors, the factors affecting soil penetration resistance were analyzed by multivariate linear regression (MLR) and random forest (RFR) models, and its transfer function was further construct. [Results] The spatiotemporal variability of soil penetration resistance was influenced by a combination of soil depth, moisture conditions, and agricultural practices. Notably, the spatial heterogeneity was significantly lower in the topsoil layer compared to the sub-surface layer, with coefficients of variation of 17.4% and 26.3%, respectively. As the soil dried out and the duration since the last tillage increases, spatial heterogeneity tended to intensify. Soil erosion played a significant role, with higher soil penetration resistance observed in areas of intense erosion compared to deposition zones, particularly under wet conditions where the difference is most pronounced (p<0.05). Furthermore, 73% of the sampled points in the studied sloping farmlands exhibited increased soil penetration resistance compared to natural forestlands, especially in regions of intense erosion. Key factors that influenced soil penetration resistance included soil moisture content, bulk density, and soil organic carbon (SOC), with Pearson coefficients of -0.69, 0.58 and -0.54, respectively, collectively explaining 88% of its variability. In terms of predictive soil penetration resistance, the RFR model demonstrated superior accuracy compared to the MLR model, achieving an R2 of 0.91 and an RMSE of 91.2 kPa for mean predictions across the growing season. [Conclusion] This study enhances our understanding of the spatial heterogeneity characteristics of black soil hardness under erosion, providing a theoretical foundation for effective management of soil compaction in black soil regions.

Keywords

• black soil
• soil compaction
• soil erosion
• pedotransfer
• random forest model