For analyzing the spatial variability and correlation of corn yield, the combined method of traditional statistical methods and geostatistics methods were compared. Firstly, the reasonable sampling number was sampled and its statistical characteristics were analyzed. The crop yield information was obtained by traditional statistical methods. Then, the spatial variability and correlation of crop yield at county-scale were studied by geostatistics methods. The regional spatial distribution map of crop yield was made by kringing interpolation methods. The results showed that the corn yield and its components had intermediate variability and the coefficient of variability was 10% ~ 100% by traditional statistics. However, the spatial distribution of corn yield was found that it had relatively strong spatial correlation by geostatistics analysis. Further analysis showed that the spatial variability of corn yield was depended on its consistent factors rather than random factors and the spatial heterogeneity scales was about 3 400 m.%采用传统统计学和地统计学相结合的方法,研究县级尺度作物产量信息的空间变异特征及其空间相关性.在确定合理采样数的基础上,利用传统统计学方法研究样本的统计特征,获得传统方法估计的总体产量信息分布概况,利用地统计学理论分析其空间变异性特征和克里格插值对作物产量的空间格局进行分析.传统统计学分析表明,玉米产量及其构成因子的空间变异系数均在10％ ～ 100％之间,属于中等变异性,而利用地统计学方法进行最佳半方差函数拟合发现,县域尺度上的玉米产量信息表现为强烈的空间相关性,说明产量的空间变异性主要是由其结构性因素决定的,随机性因素对该尺度下的变量空间变异性影响较小,且相关范围较大,在3 400 m左右.
展开▼
