Development of pedotransfer functions for predicting the bulk density in the critical zone on the Loess Plateau, China

Abstract

The bulk density (BD) is an important physical property of soil, which is used to estimate soil carbon/nutrient reserves, and it is an important parameter in various predictive and descriptive models. However, BD data are lacking due to the difficulty of measuring it directly. Pedotransfer function (PTF) may provide an alternative method for estimating BD indirectly based on easily measured soil properties. The Loess Plateau in China (620,000 km2) has deep loess deposits (50–200 m), which makes it difficult to obtain BD values for the deep soil layer, and thus, a PTF is needed for estimating BD. In this study, multiple linear regression (MLR) and artificial neural network (ANN) methods were used to develop BD PTFs for the deep layer of the Loess Plateau based on the soil organic carbon, texture, and depth. In total, 534 undisturbed soil cores were obtained by soil core drilling from five typical sites, ranging from the top of the soil profile to the bedrock. The BD values all exhibited low variation (CV < 10%). Pearson’s correlation coefficient analysis showed that BD had significant correlations with the sand, silt, clay, soil organic carbon (SOC), and depth (P < 0.01). The performance of MLR was similar to that of the ANN method. The soil depth and clay were also important input variables for the BD PTF. The PTF developed in this study performed better than existing BD PTFs. In this study, we developed the first BD PTF for the deep layer (50–200 m) of the Loess Plateau.