Trends in probabilities of temperature records in the non‐stationary climate of South Africa

Abstract

AbstractThe temporal and spatial distribution of the frequency of breaking highest and lowest daily maximum and minimum temperature records over the period of 1951–2019 has been investigated for South Africa. Temperature records are station specific and defined as either larger or smaller than any previous values in a time series of specific year‐days. Daily maximum and minimum temperatures from homogenized time series of 25 weather stations in South Africa were analysed. Aspects considered to influence the frequency of record‐breaking events were the warming trend and variance. The study found that the record‐breaking frequencies of the highest daily maximum (high Tmax) and some high daily minimum temperature (high Tmin) records were higher than the theoretically expected number in a stationary climate. This was particularly apparent near the end of the analysis period. The ratio of highest maximum to lowest minimum temperature records was almost an equal 1:1 ratio near the start of the analysis period and increased to a about 4:1 in the last decade of the period. Focusing on the last decade, that is, 2010–2019 the study found that there is a different spatial pattern between the occurrence of high Tmax and high Tmin records. For high Tmax records the highest number were mostly recorded by stations over the central parts of the country (e.g., Kimberley, Glen College and Bloemfontein). In contrast, the highest number of high Tmin records were less confined spatially. Even when considering the general warming due to climate change, many more high temperature records are broken than expected in certain regions and on average. We deduce that the higher than expected numbers of high Tmax and high Tmin records in the latter part of the analysis period were mainly due to the variability in the warming trend with acceleration in the last decade.