The flowering dates of Japanese apricot trees (Prunus mume) at 15 meteorological stations in western Japan are estimated, considering processes both of rest breaking and developing of flower bud. As the basis of estimation, accumulation model using an exponential function of daily mean temperature, so called DTS (the number of days transformed to standard temperature) method, is used for expressing development of flower bud. In this study, it is assumed that flower bud with more advanced rest break obtains greater responsiveness to DTS value. The degree of rest break on each day is quantified as the accumulations of values computed by another exponential function of hourly chilling temperature, and expressed in an index ranged from 0 to 1. Daily DTS value multiplied by the index of degree of rest break is accumulated, and the date, on which its accumulations reach to their mean value required for flowering, is determined as the estimated flowering date.
Due to estimate flowering date accurately, it is necessary to determine the most suitable combination of three parameters in the exponential function of chilling temperature at each station. These parameters represent the point when chilling temperature starts to affect in breaking the rest of flower bud, the optimum temperature to break the rest and the point when rest completes, respectively. Most suitable combination of parameters for each station shows that the rest break of flower bud initiates in late-October and almost completes during the period since late-Novemver until mid-December.
By taking account of the process of rest break into the estimation of flowering date using DTS-method, the part of errors attributed by yearly variation in the succession of rest break process is decreased considerably. The root mean square errors in estimations with corrections considering rest break processes fall within the range of 5-10 days, whereas those without corrections correspond to the range of 7-12 days. Estimation methods with considerations of rest break processes generally account for 70-90% of the variances in actual flowering dates of Japanese apricot.