2014 ASHS Annual Conference

Growing degree-days (GDD) has been widely used for heat summation in region classification and phenology prediction. It is usually calculated from the mean of daily minimum and maximum temperatures with a pre-determined base temperature. The present study compares GDD with growing degree hours (GDH) to examine the source and magnitude of errors in heat summation calculation. Heat summation was calculated as a simple rectangular integration of the mean or average temperatures above 10°C at an interval of one day for GDD or one hour for GDH. Daily GDD was calculated from the mean of daily maximum and minimum temperatures (GDDmm) or daily average of 1440 minutely measurements (GDDavg). Daily GDH was calculated as the sum of 24 GDH values from hourly average of 60 minutely measurements for a given day (GDHavg). Cumulative and total GDD or GDH was obtained by adding up daily values during the growing season. GDDmm exhibits the most error, under- or over-estimating heat summation most of the time. GDDavg, on the other hand, only underestimates heat summation when temperatures are low. Errors in heat summation calculations are associated with the asymmetrical distribution of diurnal temperatures and the base temperature, dependent upon the resolution of the temperature data. The asymmetrical distribution results in the mean of daily minimum and maximum temperatures to be unequal to the daily average temperatures at a higher resolution. The error due to asymmetrical distribution cannot be fixed in GDDmm, due to the vast variation and the lack of consistent patterns in temperature distribution. Base temperature as a lower threshold, in combination with the asymmetrical distribution, also contributes to the error of heat summation calculation especially in GDDmm or GDDavg. When temperatures are higher than the base temperature during only part of the day but the daily mean or average temperatures are equal or lower than the base temperature, GDH accumulates growing degree, but GDD does not, leading to an underestimate of heat summation. GDHavg improves the estimation of heat summation over GDDmm or GDDavg by using temperature data at a much greater resolution (24x), allowing the inclusion of all hourly temperatures higher than the base temperature in heat summation calculation. GDH estimates heat summation much more accurately and should be used instead of GDD, especially GDDmm, when hourly temperatures are available.