基于环境因子和联合概率方法的土壤有机质空间预测

<正>土壤连续属性(如土壤中养分含量、重金属含量等)的空间分布特征和定量分布信息是进行土壤质量评价和区域环境综合评估的基础。精准农业战略的实施和各种区域生态评价均需要更详细更精确的土壤属性信息作为依据1-2]。因此,土壤属性空间预测一直是土壤学研究的热点问题。经典地统计学以各种克里格插值法为代表,是土壤属性空间预测中的常用方法。但该方法缺乏对辅助信息(如环境信息)的有效利用3-4],导致预测精度降低5]。而土壤景观定量模型的理论依据就是土壤与环境的关系,但该法忽略了采样点之间的空间相

Prediction of spatial distribution of soil organic matter based on environmental factors and a joint probability method

The spatial distribution of soil organic matter is useful information to precision agriculture. To make full use of spatial correlation of sample sites and the impact on predicted soil properties from the environmental factors (such as terrain attributes), in this study, soil attribute values of the sample sites and environmental factors data was used to establish the soil nutrients spatial distribution prediction model based on probability distributions obtained by the soil-landscape model and ordinary Kriging under the framework of the joint probability and to predict the spatial distribution of soil organic matter in Shayang County, Hubei Province. Results indicates that this method improves the prediction accuracy of soil organic matter compared to the regression Kriging method, the accuracy improvement can up to 7.202%. Prediction accuracy was closely related to what environmental factors are selected, their resolution and the way to use the them. Although this method is still preliminary, the results of this study show that this method is a flexible and effective approach to improve the spatial distribution of soil properties prediction accuracy using the environmental factors (such as topography factors).