Abstract

The results of the principal component analysis application for estimating the average regional winter wheat yield based on the integration of satellite and ground meteorological information for the southern regions of the Russian Federation are presented. The satellite indices such as NDVI (Normalized Difference Vegetation Index), VCI (Vegetation Condition Index), and satellite product LAI (Leaf Area Index) were used. Meteorological information was represented by temperature, humidity deficit, total precipitation, and the Selyaninov hydrothermal coefficient. The parameters that have the greatest impact on the yield were selected. The components with the largest eigenvalues were extracted from this set of parameters using the principal component analysis. Equations for the dependence of the winter wheat yield on the extracted components were calculated. To check the equations, the expected yield for the period from 2012 to 2017 was calculated. The relative error varied within 5–12%. In all cases, the yield calculation error when using the principal component analysis is smaller than when using correlation and regression dependences.

中文翻译：

---

基于星地信息融合的主成分分析法估算冬小麦产量

摘要

介绍了俄罗斯联邦南部地区基于卫星和地面气象信息整合估算区域冬小麦平均产量的主成分分析应用程序的结果。使用了诸如 NDVI（归一化差异植被指数）、VCI（植被状况指数）和卫星产品 LAI（叶面积指数）等卫星指数。气象信息由温度、湿度不足、总降水量和 Selyaninov 热液系数表示。选择对产量影响最大的参数。使用主成分分析从这组参数中提取具有最大特征值的成分。计算了冬小麦产量对提取成分的依赖性的方程。为了检查方程，计算了 2012 年至 2017 年期间的预期收益率。相对误差在 5-12% 范围内变化。在所有情况下，使用主成分分析时的收益率计算误差都小于使用相关性和回归相关性时的误差。