基于经验贝叶斯克里金的微尺度植烟田土壤有机质空间变异性.

【Objective】Accurate spatial information of soil properties in tobacco planting fields is essential to precision fertilization for high quality tobacco production. 【Method】In this paper, 54 topsoil samples were collected in a long-term tobacco planting fields in Neixiang county, Henan Province. The spatial variability of soil organic matter in tobacco planting fields was analyzed by using the experience bayesian method（EBK） to simulate the distribution of the content of soil organic matter in the study areas on micro sampling scale based on geostatistics. Its performance was compared with that of ordinary kriging（OK）, RBF net work model.【Result】The results indicated that the the mean of soil organic matter contents in the study areas was 14.81 g·kg-1, which had intermediate variability and the coefficient of variability was 13.74 % by traditional statistics.The spatial variability of soil organic matter contents in tobacco planting fields exhibited semi-variance structure and its semi-variance function fitted expontial model well. The analysis showed that the spatial correlation in surface soil organic matter contents in tobacco planting fields was strong on micro sampling scale. The predictions of soil organic matter contents in simulation using EBK was better than those using RBF net work predication model and ordinary kriging method.【Conclusion】Thus, the spatial heterogeneity information of soil organic matter could be achieved exactly and quickly by the method of EBK which used spatial coordinates and neighbor samples information as inputs of net. The results provide a theoretical basis and technical support for predicting soil property spatial distribution accurately, using fertilizer properly and mitigating non-point source pollution of agriculture in tobacco planting fields. [ABSTRACT FROM AUTHOR]