Realization of rainfed wheat and barley production environment based on drought patterns in the northeast Iran.

High spatiotemporal climate variability in the northeast Iran has led to increased production risk for rainfed crops. Hence, to explore potential opportunities for reducing the probable risk, it is necessary to characterize the main rainfed growing regions based on drought patterns. The CSM-CERES-Wheat and -Barley models were applied to simulate yield and to achieve input for estimating water supply demand ratio (WSDR) in 10 reference weather stations (RWSs) for the period 1980–2009. Simulated WSDR was averaged for every 100°Cd to and from flowering for each location (each RWS and each season). The environmental classes and their optimum number for each RWSs and the region were derived by k-means clustering. The soil diversity combined with spatiotemporal variations in rainfall caused rainfed wheat and barley to be exposed to drought stresses, which were very different across locations and among seasons. Nevertheless, four main environmental classes could represent the variability at local and regional scales. The first environmental class corresponded to environments with low drought stress with a frequency of 1.10 and 12.01% for wheat and barley, respectively. For wheat, environmental classes 2, 3, and 4 with a frequency of 22.62, 51.46, and 24.82% begin from 1100, 900, and 700°Cd (degree days) before flowering, respectively, and lasted until maturity. For barley, these three environmental classes with the frequency of 26.50, 32.86, and 28.62%, respectively, corresponded to sever drought stress levels with different degrees that begin from 500°Cd before flowering and was maximized at about 300°Cd after flowering which coincided with the grain filling. The occurrence frequency of environmental classes differed temporally and spatially. Over time from 1980 to 2009, environmental class 4 increased and environmental class 1 decreased for wheat and barley, respectively. In most locations, wheat and barley yield tended to decline from environmental classes 1 to 4. Such characterization of production environment can be used to help breeders focus on genes and traits adapted to target production environments. [ABSTRACT FROM AUTHOR]