Improve the prediction of soil bulk density by cokriging with predicted soil water content as auxiliary variable

Soil bulk density (SBD) is a key soil physical property affecting the transport of water and solutes, which is essential to estimating soil carbon and nutrient reserves. However, it is considered to be time consuming, labor intensive, and expensive to obtain in situ SBD data. Therefore, it is important to estimate SBD at an acceptable level of accuracy using auxiliary variables with limited field-measured data. The Heqing Dam agriculture area in Yunnan Province, Southwest China, was selected as the study area. Sampling points of 114 were obtained based on a uniform 1 km x 1 km grid methodology, which were broken into training data and test data using ArcGIS software. The training data (91 soil samples) were used for prediction and the test data (23 soil samples) were used for validation. Predicted soil water content (SWC) was estimated using kriging with high accuracy. The predicted SWC was used as auxiliary data to improve the prediction of SBD by cokriging. The correlation coefficient (r) between field-measured and predicted values, mean error (MAE), and root mean square error (RMSE) were used to validate the performance of the geostatistics. The SBD was highly correlated with measured (r = -0.73, p < 0.01) and predicted SWC (r = -0.73, p < 0.01). The two parameters MAE and RMSE showed an improvement after introducing the predicted SWC. The MAE and RMSE decreased from 0.111 and 0.032 g cm(-3) by kriging to 0.070 and 0.018 g cm(-3) by cokriging with predicted SWC, respectively. The r value increased from 0.449 by kriging to 0.852 by cokriging. Compared with kriging, the application of cokriging with predicted SWC resulted in a relative improvement of 45.5 %. This study demonstrates that predicted SWC used as auxiliary data can improve the prediction of SBD. Therefore, when the predicted data have been demonstrated to be of high accuracy and are highly correlated with the dependent variable, they have the potential to be a good source of auxiliary data for cokriging.