Fifty years of semi-dwarf spring wheat breeding at CIMMYT: Grain yield progress in optimum, drought and heat stress environments

Abstract

Grain yield progress over 50 years of spring wheat breeding at the International Maize and Wheat Improvement Center (CIMMYT) was determined in field trials conducted during five crop seasons (2013–2017) at Norman E. Borlaug research station near Ciudad Obregon, Mexico. The trials included 30 varieties (24 bread wheat and 6 durum wheat) released between 1965–2014 and were sown under managed optimum, drought, and heat stress conditions. The optimum irrigated environment had 3 management systems, flat sowing with weekly drip irrigation (FDI), bed sowing with flood irrigation (BFI), and flat sowing with flood irrigation (FFI). The drought environment had 2 management systems, flat sowing with reduced irrigation (FRI) and flat sowing under severe drought stress (FSD). The heat stress environment was sown in beds (HFI) three months later than the normally sown irrigated and drought environments. The rate of grain yield progress was estimated relative to Sonalika released in 1965 and Mexicali C75 released in 1975 for bread and durum wheat, respectively. Grain yield progress per year for bread wheat was, 31.2 kg ha−1, 35.3 kg ha−1, and 24.7 kg ha−1 in irrigated environments FDI, BFI, and FFI, respectively. In the stress environments, bread wheat grain yield progress was estimated as 25.6 kg ha−1, 17.7 kg ha−1, and 18.1 kg ha−1 per year in FRI, FSD, and HFI, respectively. For durum wheat, the grain yield progress was estimated as 29.6 kg ha−1, 48.1 kg ha−1, 18.8 kg ha−1, and 29.8 kg ha−1, per year in FDI, BFI, FFI, and HFI, respectively. Trait linkage graph analysis using LASSO regularized graphical model estimated that biomass, harvest index, and grains per meter square (GNM) were linked to grain yield progress in all environments. Thousand kernel weight was associated with grain yield progress under optimum and heat stress conditions, whereas grain weight per tiller (GWT) associated with progress under drought. Results also show that the highest yielding varieties in each environment however, had different trait attributes, with some varieties having higher GNM and tillers per meter square compensating for low GWT, while others had high GWT to compensate for reduced GNM. In conclusion, CIMMYT’s wheat breeding program has continued to show progress in grain yield in different environments/management systems, and while certain traits have consistently improved over the years, the varieties developed have employed different trait strategies to achieve final grain yield.