Adaptation and Agronomic Characterization Trials of 14 Rice Genotypes in Contrasting High-Altitude Marshes of Gisha and Akagoma, Ngozi Province, Burundi
Privat Rwasa *
Food Security and Human Nutrition Programme, Faculty of Agricultural and Veterinary Sciences, University of Ngozi, P.O Box 137, Ngozi, Burundi and Centre for Research in Agriculture and Rural Development (CERADER), University of Ngozi, P.O Box 137, Ngozi, Burundi.
Félicien Niyibitanga
Centre for Research in Agriculture and Rural Development (CERADER), University of Ngozi, P.O Box 137, Ngozi, Burundi and Rural Development Programme, Faculty of Agricultural and Veterinary Sciences, University of Ngozi, P.O Box 137, Ngozi, Burundi.
Marc Ngendahayo
Centre for Research in Agriculture and Rural Development (CERADER), University of Ngozi, P.O Box 137, Ngozi, Burundi and Faculty of Agronomy and Agribusiness, CERADER, University of Ngozi, P.O Box 137, Ngozi, Burundi.
Alexis Ndayizeye
University Centre for Research for Economic and Social Development (CURDES), P.O. Box 1049 Bujumbura, Burundi and Doctoral School of the University of Burundi, avenue de l’ UNESCO No2, P.O. Box 1550, Bujumbura, Burundi.
*Author to whom correspondence should be addressed.
Abstract
Rice is a strategic crop for food security and sustainable development in Burundi, yet its productivity in high-altitude marshes remains severely constrained by low nighttime temperatures (<15°C), which induce spikelet sterility during reproductive stages. Despite breeding efforts by ISABU, FACAGRO, and IRRI, the availability of high-yielding, stable varieties adapted to these medium-altitude ecological conditions remains limited. Understanding genotype × environment (G×E) interaction is essential for identifying superior genotypes that combine high yield potential with stability across diverse environments, thereby supporting targeted breeding programs. This study evaluated the agronomic performance and stability of 14 rice varieties, 11 from ISABU preliminary trials and 3 locally cultivated, in the high-altitude marshes of Gisha and Akagoma (Ngozi Province) during the 2013–2014 growing season. Using a randomized complete block design with three replications, nine agronomic and yield-related traits were analyzed through two-way ANOVA, multiple regression, and the AMMI model to dissect G×E interaction. Multiple regression analysis revealed that effective tillering, thousand-grain weight, spikelet sterility rate, and total grains per panicle significantly influenced paddy yield (R² = 0.512; p < 0.001), highlighting these components as key selection criteria for direct breeding improvement. ANOVA demonstrated highly significant differences between sites and varieties for plant height, grains per panicle, thousand grain weight, sterility rate, and yield (p < 0.001). The AMMI analysis identified varieties UNG, 74, 71, and 46 as exhibiting general adaptation across both sites, with UNG combining the highest mean yield (6.4 t/ha) with superior stability. Varieties 54 and 12 achieved exceptional yields (7.58 and 6.02 t/ha, respectively) but displayed specific adaptation to Akagoma conditions. These findings provide evidence-based recommendations for varietal dissemination and selection criteria to enhance rice productivity in Burundi's high-altitude zones. Multi-season and multi-location testing is recommended to further validate stability patterns and support breeding strategies for climate-resilient rice varieties.
Keywords: Oryza sativa, high-altitude marshes, genotype × environment interaction, AMMI model, yield stability