Frequency of Frost Occurrences

Neither the median dates of first and last frosts nor the median duration of the frost period give any indications on the number of years within the 50 year time series of temperature values used in the study (Schulze and Maharaj, 2004) in which frost actually occurred, nor on the number of frost occurrences within a year, nor on the variability of frost occurrences from year to year. In the mapping statistics of frost over South Africa, the temperature database used in this study contained quality controlled time series of daily maximum and minimum temperatures from > 970 qualifying temperature stations, the records of which had been infilled and extended (where necessary) to a common 50 year time period 1950 - 1999 by techniques described in Schulze and Maharaj (2004). These daily station records were then used with regionally and seasonally determined temperature lapse rates to generate 50 years of daily maximum and minimum temperatures at each of 429 700 one arc minute (1.7 x 1.7 km) raster points covering the RSA, Lesotho and Swaziland. At each raster point a minimum temperature of ≤ 0 degree Celsius in the 50 year daily time series was designated a `heavy frost day`. Any raster point where < 5 years out of the 50 had experienced least one heavy frost day that point was deemed to be `frost free`.