基于协同克里格的耕层土壤速效钾空间异质性研究

摘要

基于万年县配方施肥项目野外采集的耕层 (0～20 cm) 土壤数据, 运用常规性统计和地统计学方法定量刻画土壤速效钾的空间异质性.结果表明, 研究区土壤速效钾值域为33.46～164.84 mg·kg-1, 均值为82.04 mg·kg-1, 属于4级 (第二次土壤普查土壤养分分级标准), 变异系数为30.49％, 呈中等程度变异.成土母质、土壤类型 (亚类和土属) 、高程、p H和阳离子交换量 (CEC) 对研究区耕层土壤速效钾空间变异影响极显著 (P<0.01) .交叉验证的结果表明, 加入辅助变量的协同克里格法较普通克里格法可以提高模拟精度, 辅助变量个数越多, 模拟精度越高, 空间异质性描述越详细.通过插值得出土壤速效钾的空间分布图, 可为当地土壤钾素管理和精准农业提供科学依据.%The spatial heterogeneity of soil available potassium (AK) was quantified by using general statistics combined with geostatistics based on soil data collected from field surveys (0-20 cm) in the formula fertilization project in Wannian County.The results showed that soil AK ranged from 33.46 mg·kg-1to 164.84 mg·kg-1with an average of 82.04 mg·kg-1in the study area, and was in the fourth grade (classification standard of soil nutrients in the second soil surveys) with a coefficient variation of 30.49％.Parent materials, soil group (subgroup and soil family), elevation, p H, and cation exchange capacity (CEC) had significant (P<0.01) effects on the spatial variation of soil AK.The results of cross-validation showed that co-Kriging with auxiliary variables could improve accuracy compared to ordinary Kriging, the more auxiliary variables, the higher the simulation accuracy, and the more detailed spatial heterogeneity description.The spatial distribution map of soil AK could offer scientific basis for local soil AK management and precision agriculture.