Partitioning Genotypic and Environmental Variance of Fresh Root Yield in Provitamin A Cassava Genotypes Using Mixed Models

This study employed a multi-model statistical approach to evaluate the yield performance and environmental responsiveness of 42 provitamin A cassava genotypes across two contrasting cropping seasons. Four linear mixed models spanning BLUP-based random effects, fixed-effect genotype comparisons, genotype-by-environment interaction analysis, and variance partitioning were used to dissect phenotypic variation in fresh root yield. Months After Planting (MAP) emerged as the most influential factor, accounting for 19.82% of total variance, while genotypic effects were modest (H{superscript 2} = 4.4%) and seasonal effects reached statistical significance when accessions were treated as fixed. Residual variance dominated (70.98%), suggesting the influence of unmeasured factors such as soil properties, pest pressure, and microclimatic variation. Rainfall data revealed a stark contrast between seasons-127.38 mm in 2019/2020 versus 8.45 mm in 2020/2021 highlighting an environmental stress pattern that shaped genotype performance. Accessions IITA-TMS-IBA980581(WChk-White Check), IITA-TMS-IBA180017, and IITA-TMS-IBA180146 maintained high yields across both seasons, indicating broad adaptability, while IITA-TMS-IBA180182 and IITA-TMS-IBA180065 performed better under low-moisture conditions, suggesting drought responsiveness. These findings underscore the importance of integrating rainfall patterns and genotype-specific environmental sensitivity into cassava breeding pipelines. Future multilevel models incorporating locations and climatic data could enhance selection precision and support the development of resilient varieties for variable agroecologies.