Temporal variations in temperature over Northeastern Africa

Abstract

Research in climate change requires the use of long, continuous, relatively homogeneous time series. Unlike the extra-tropics, these series are not widely available in the tropics and sub-tropics. Influences due to nonclimatic factors are common in most series. These influences can arise from either abrupt or gradual discontinuities. Therefore, in order to minimize non-climatic influences, the inhomogeneities must be identified through either station histories, a statistical technique, or both. Once these inhomogeneities are identified the original series is adjusted. Typically, globally averaged or hemispherically averaged annual mean land temperature series are used for analysis of thermal trends. These series are generally from the extra-tropics. Early studies generally focused on mean annual temperatures, but more recent studies have considered maximum and minimum temperature series. However, most of these studies are on data from the extra-tropics from 1950 to present. With access to previously unavailable data from Northeastern Africa from 1900 to 1950, the data are analyzed to determine whether thermal patterns are identifiable in maximum and minimum temperatures over smaller scales in the tropics and sub-tropics. First, the Alexandersson Test for homogeneity is applied to the series to ensure relative homogeneity in the series. Averages of the first ten and last ten years are then calculated, and a difference is obtained. A t-test is used to determine if the differences in the ten-year means are significant. A F-test is also used to compare the variances. The results show that there is significant warming in the minimum and mean temperatures across the region. The diurnal temperature range is decreasing, while changes in the maximum temperature vary. When series from large urban stations are compared to smaller rural stations, possible trends due to urbanization can be identified. The results show that no unusual warming is 'indicated in the Egyptian stations, but warming is greater at Khartoum, Sudan than the surrounding stations. Finally, the problems that arise in a typical tropical area are addressed. Because most series in this region lack continuity, the methods commonly used in climate studies in the extra-tropical regions cannot be applied without some modification.