Calculating Chilling Hours and Chill Units from Daily Maximum and Minimum Temperature Observations

Abstract

Models that require hourly weather data as inputs are being constructed to simulate development of insects and plants. Most agricultural and, climatological weather observation stations, however, do not include the capability to record hourly data. Thus, methods, for simulating hourly observations need to be developed if the wealth of data collected in such stations is to be used in these models. One application for simulated temperature data is in the peach industry. This industry uses chilling hours and chill units to determine progression of fulfillment of chilling requirements through the winter season. In the past, chill hours and chill units have been calculated at a limited number of locations using either manually recorded hourly observations or thermograph charts. Richardson (1974) proposed a very simple model for determining temperatures from which chill units were calculated. He used a straight line with 12 hr between maximum and minimum temperature. Richardson recognized the limitations of this model, especially when assuming the daily temperature curve was symmetrical about 12 hr. He suggested that the time between maximum and minimum temperature may need to be changed to better represent daylength changes during the year. McFarland et al. (1987) compared various mathematical representations of the daily heating wave. In their analysis, most published models used a modified sine function to describe heating during daytime hours. A sine function should be a close representation of the curve because daytime temperatures follow the daily solar cycle. The nighttime cooling curve, however, is not as simple. Cooling depends on many factors, including moisture content of the air, cloud cover, and soil heat flow. One form of the cooling curve suggested by Sutton (1953) depends on the square root of time since sunset. Parton and Logan (1981) used an exponential cooling rate to describe nighttime cooling. Eckersten (1986) modified Parton and Logan’s model to a sine-sine ex-