Abstract
Between 2002 and 2011, 3 varieties of wheat, ‘Norin61’ (spring type), ‘Shirogane Komugi’ (spring type), and ‘Satonosora’ (winter type), were grown in the upland and lowland fields in Tsukuba, Ibaraki Prefecture, Japan. Phenological development was divided into 5 phases: phase 1, sowing to emergence; phase 2, emergence to jointing; phase 3, jointing to heading; phase 4, heading to flowering; and phase 5, flowering to maturation. The duration of phase 2 varied with the sowing date and cultivar. In ‘Satonosora’, the jointing stage was later than that in the other cultivars in early sowing, but this difference diminished at late sowing. Phenological development was expressed using the developmental rate (DVR) and developmental index, which is the cumulative DVR. DVR was expressed as a function of daily mean temperature and photoperiod in phase 2, but as a function of temperature in the other phases. These models achieved a good fit for the jointing, heading, and maturity dates of each cultivar measured in a wide area in Japan, with root mean square error=4–6 days. However, a simpler 2-phase model (sowing to heading and heading to maturation) could provide better estimates for the prediction of the jointing day alone.
