论文部分内容阅读
本文利用地统计学理论,对棉花二因子3水平的随机区组试验资料进行了空间方差分析.土壤趋势因子导致半方差函数分析表明,产量偏差(对均值)在12.6m的范围内存在着空间相关.传统方差分析否定了任何处理效应而空间方差分析则指明在0.05水平,播期对棉花产量有着显著的影响.这是因为传统方差分析忽视试验误差的空间依变性,使得田间存在的趋势因子掩盖了处理效应.应用空间方差分析可以分离出趋势因子,相对地降低误差项,从而得出真正的处理效应.因此,空间方差分析是田间试验中评价处理效应的有用方法.
In this paper, we use the geostatistics theory to analyze the spatial variance of randomized block test data of two-factor-three cotton. The analysis of the semivariance function resulting from the soil trend factor shows that there is a spatial correlation between the yield deviation (for the mean) over a range of 12.6 m. Traditional analysis of variance negates any treatment effects and spatial variance analysis indicates that at 0.05 the sowing date has a significant effect on cotton yield. This is because traditional ANOVA ignores the spatial variability of experimental errors, making the trend factors in the field obscure the treatment effect. Spatial variance analysis can be used to isolate the trend factor and reduce the error term relatively to get the true treatment effect. Spatial variance analysis is therefore a useful method of evaluating treatment effects in field trials.