资源环境科技发展态势分析平台

To establish precise climate trend analyses, highly reliable and accurate homogenous historical climate data are required. To this end, we undertake a robust quality control and homogenization process of daily T-max and T-min data (1916-2013) for the Western Cape Province, South Africa, using RClimDex, ProClimDB and Anclim software. Inhomogeneities were detected using the pairwise method in AnClim, suggesting possible artificial shifts in the time series. The adjustment of time series utilized ProClimDB software to create reference series using highly correlated or nearest neighbour stations. Given few available long-term data sets, the study is limited to eight suitable stations. The modified Mann Kendall test in XLSTAT 2015 software identified annual and seasonal trends in the newly homogenized monthly T-max and T-min data. Annual T-max and T-min adjusted data over the Western Cape region indicate statistically significant increasing temperature trends over the period 1916-2013, with the exception of an insignificant decreasing T-max trend at Cape St Blaize. A statistically significant increasing trend (0.13 degrees C/decade) for all stations used in this study is recorded for the common period 1937-2001. The seasonal trends also support significant increasing trends, with the exception of T-max trends for summer (-0.03 degrees C/decade) and autumn (0 degrees C/decade) at Kirstenbosch.