Abstract
In Ethiopia, sweet potato is the 2nd and the most important root crop after Enset [Ensete ventricosum (Welw) Cheesman]. Even though widely cultivated in Ethiopia, the occurrences of wide agro-ecological variability are the key challenge for the selection of high yield and stable orange-fleshed sweet potato genotypes due the high interaction of genotype by environment effect (GEI). Until to date, the research reports on GEI and stability of orange-fleshed sweet potato genotypes under Southwest and West of Ethiopian conditions is very limited. Due to this fact, the study was conducted in main sweet potato growing of Southwest and West of Ethiopia since the 2019-2020 growing seasons. Nine Orange-fleshed sweet potato genotypes were evaluated in four different ecologies in completely randomized block design with three replications. The mainly indicative data and related variables were collected and analyzed by using ANOVA, AMMI, and GGE biplots. The ANOVA for total storage root yield revealed significant difference in the environments, genotypes, and the interactions of both (P < 0.001). The genotypes NASPOT-12 was the mostly performed one over all the rest with total fresh storage root yield of 55.88 ton ha(-1) while Kulfo was performed as the lowest. The AMMI1 and AMMI2 biplot analysis revealed that, the genotypes NASPOT-12 was resulted in an above average mean of total fresh storage root yield and the genotype Koka-12, NASPOT-13, NASPOT-12, and Kulfo were far apart from the biplot origin which indicated the genotypes were the more responsive and largely contributed to the interaction component and thus considered as specifically adapted genotypes respectively. The GGE biplot revealed that the total variations of 79% with the PC1 accounted about 55% and the PC2 affect about 24% of variations approximately on the total fresh storage root yield of the tested orange-fleshed sweet potato genotypes. The "which won where" identified the three growing mega environments. In the comparison and ranking of the genotypes by the GGE biplot analyses, the genotype NASPOT-12 was more desirable than the rest genotypes. In the comparison and ranking of the environments by this analysis, the environment Jimma-1 was the most representative followed by Agaro-2 in the total fresh storage root yield of orange-fleshed sweet potato genotypes. Above all the genotype NASPOT-12 was responded well in most of the environments hence, recommended genotype for multipurpose advantage.