Development of a Database of Gridded Daily Temperatures for Southern Africa
Report No 1156/2/04
Dec 2004

EXECUTIVE SUMMARY

BACKGROUND AND OBJECTIVES

Temperature is a basic climatological parameter used frequently as an index of the energy status of the environment.

Because temperature measurement is a simple procedure and because long term temperature data are relatively abundant, it is not surprising that temperature parameters have been used as input variables in a wide range of climatological, hydrological and agricultural applications, from the estimation of solar radiation, relative humidity or potential evaporation to the concepts of heat and chill units or the delimitation of frost zones and agroclimatically optimum growth areas of crops.

In light of its importance, as highlighted above, temperature was selected as the one multiple-use hydroclimatic variable, other than rainfall, for detailed study in the Water Research Commission (WRC) Project K5/1156 titled

“Development of a Revised Spatial Database of Annual, Monthly and Daily Rainfall and Other Hydroclimatic Variables for South Africa”.

More specifically, this temperature related component of Project K5/1156, which is reported separately from the rainfall component, sought to

• develop procedures to infill missing daily maximum and minimum temperature values and extend temperature records to a pre-selected common 50 year base period (1950-1999) for stations qualifying on the strength of existing records, and to
• develop procedures to generate daily time series of temperature at any one of the 429 700 one arc minute of a degree grid points making up southern Africa, defined here as the RSA plus Lesotho and Swaziland.

In order to achieve these two objectives

• a database of temperature records in southern Africa first had to be collated and quality controlled (Chapter 2),
• temperature lapse rates had to be established for each month of the year for both maximum and minimum temperatures (Tmx, Tmn) for different lapse rate regions and sub-zones within southern Africa (Chapter 3),
• the temperature records at qualifying control stations then had to be infilled and extended to a common 50 year base period, and the procedures verified (Chapter 4), before
• techniques were developed for daily temperatures to be generated at any location (i.e. at any one of the 429 700 grid points) in southern Africa by making use of the extended/infilled control station values, lapse rates and methods to adjust for spatio-temporal lapse rate discontinuities (Chapter 5).
• Finally, conclusions were drawn and recommendations made for future research in this field (Chapter 6).

The specific objectives of this study relate clearly to daily temperature estimates. Why?

In a previous publication by the senior author (the “South African Atlas of Agrohydrology and -Climatology”, WRC 1997), long term monthly means of temperature parameters were used in developing equations for computations of solar radiation, relative humidity, potential evaporation and crop yields. However, neither inter-annual nor intra-seasonal variabilities, nor probabilities of occurrence could be computed, neither could any trends over time be evaluated, because only long term monthly means were used as forcing functions. The generation of daily temperature time series at any location in southern Africa would, however, facilitate computation of all these higher order statistics.