Application of GGE biplot in spring wheat yield stability analysis in rainfed areas of China

GGE (genotype main effect and genotype-environment interaction) biplot analysis is a method based on principal component analysis (PCA) for the effective exploration of multi-environment trials (METs). It allows visual understanding of genotype-environment interaction (G×E-interaction), cultivar yield stability for each mega-environment, and test-location suitability. Grain yields of 9 spring wheat genotypes (cultivars or lines) tested at 17 national spring wheat regional test sites under rainfed conditions in 2005 were analyzed via GGE biplot and AVONA methods. The results show that G×E-interaction effect is 5.4 times of genotypic effects for grain yield. The stability of different genotypes varies greatly, and genotypes with high yield and yield stability account for only 11% of the total tested genotypes, though some genotypes are specifically adaptable to certain environments. The 17 test-sites are broadly classified into three environments. Accordingly, “8821-1-1” and “longchun 9143” spring wheat are the best genotypes for arid regions of the Loess Plateau, “qingchun193” and “wumai 7” are the best genotypes for arid and cold regions of the Tibetan Plateau, and arid regions of North China, respectively. Based on discrimination ability and environmental representativeness, two ideal test-sites, Yuzhong County of Gansu Province and Huzhu County of Qinghai Province are proposed for spring wheat variety regional test. Grain yield is more closely correlated with precipitation during growing season than the other six environment factors.