Heat Units (Standard Deviation) - Summer

The concept of the heat unit (or degree day), known since the mid-18th century already, revolves around the development of a plants or organism`s being dependent upon the accumulated heat to which it was subjected during its lifetime, or else during a certain developmental stage. This measure of accumulated heat is known as physiological time. In general it holds that the lower the temperature, the slower the rate of growth and development of plants and invertebrate animals and the amount of heat required to complete a given organism`s development does not vary. Physiological time is usually expressed and approximated in units of degree days, also popularly called heat units, where these are an accumulation of mean temperatures above a certain lower threshold value and below an upper developmental threshold (above which growth is considered to remain static or even decline), over a period of time. The degree day concept is not without its limitations. These include that, a linear relationship is assumed between growth and temperature, threshold temperatures may change during the life cycle of a crop or Pest, temperatures exceeding the upper threshold may, in fact, have a detrimental/negative effect on development. Derivation over South Africa of heat units were done by calculating degree days for the daily temperature fluctuations that occur in nature. These range from degree hours, sine and triangular methods based on hourly temperatures, as well as so-called cut-off methods when the upper threshold of temperature is reached. To estimate heat units (HUs) over South Africa, the 50 year time series of daily maximum and minimum temperatures (Tmxd, Tmnd), generated at each of the 429 700 raster points covering the area at 1` x 1` latitude/longitude, was used (Schulze and Maharaj, 2004). The generation of this time series is described in Section 2a. HUs was computed for each day for a threshold temperature of 10 degree Celsius, i.e. HU = [(Tmxd + Tmnd) / 2 - 10] for HU ≥ 0.0 This method is a considerable advance on the computation of HUs in the previous version of this Atlas (Schulze, 1997), in which HUs were calculated from regression derived monthly means of Tmx and Tmn. From these daily values, monthly as well as seasonal and annual HUs and their standard deviations, could be calculated and mapped.