玉米品种多环境测试数据的空间插值分析

Spatial Interpolation Research of Multi-environment Trials Data for Maize

在地统计学和地理信息系统支持下,采用多种插值方法对玉米品种多环境测试数据进行空间插值研究。多环境测试数据为东华北、黄淮海试验点的多年平均值,利用几种空间插值方法对数据各个表型性状进行插值分析,比较各个插值方法的均方根预测误差,选取精度最高的插值方法,得出各个表型性状的空间分布情况。结果表明,平均单产适合普通克里格插值方法,百粒重、穗行数、穗位高、穗长、株高、倒伏率、倒折率、纹枯病、玉米螟适合使用反距离加权插值法,单穗粒重、秃尖长适合使用简单克里格插值法,空秆率适合使用径向基函数插值法。

Based on the geostatistical method and GIS, the paper used several interpolation methods to estimateMaize＇s phenotype via multi-environment trails data. The data came from the average of the trims which planted sev- eral years in Northeast and North and the Huang-Huai-Hai plain. For each phenotypic trait, the highest precision in- terpolation method was chosen as the suitable method by comparing the root mean square, and then it＇s spatial distri- bution was displayed. The results showed that the most suitable method for yield was Ordinary Kriging, for grain weight, rows per ear, ear height, ear length, plant height, stem down, stem breakage, maize sheath blight and corn bor- er was Inverse Distance Weighting, for grain weight per ear and length of barren ear tip was Simple Kriging and for barrenness was Radial Basis Functions. The study results could contribute to analyze the variety performance under the interactions, optimize trials distribution and forecast the performance of variety extension.