Statistics

Statistics have been calculated for the modelled climate parameters of daily mean temperature, precipitation and sunshine duration in order to quantify uncertainties in future climate projections. The statistics refer to the change of mean (with respect to 1961–1990), standard deviation and skewness of the multi-year distribution for each calendar month over 2071–2100. Thirty years of data have been analysed to estimate the statistical parameters for each modelling experiment. For the PRUDENCE HC-HadRM3P A2 scenario case there are three runs (adhfa, adhfe & adhff), which have been analysed both individually (30 years) and in combination (90 years; denoted as "adhfaef"). For the ENSEMBLES transient climate experiments the 2071–2100 trend has also been calculated, whereas standard deviation and skewness have been calculated for original and detrended time series. Note that only 29 years (2071–2099) were analysed for HC-HadRM3Q0.

In the tabulated lists of statistics, for PRUDENCE/SMHI-RCAO & CNRM-Arpège, PRUDENCE/HC-HadRM3P and ENSEMBLES experiments, the file name ending denote the statistical parameter calculated as follows:

• change: difference in multi-year mean between 2071–2100 and 1961–1990
• sdev: standard deviation of original time series over 2071–2100
• skew: skewness of original time series over 2071–2100
• trend: linear trend of original time series over 2071–2100
• sdev_dt: standard deviation of detrended time series over 2071–2100
• skew_dt: skewness of detrended time series over 2071–2100

Each statistical file comprises 12 monthly two-dimensional arrays (overall dimension 100 × 80 × 12 for PRUDENCE and 216 × 168 × 12 for ENSEMBLES experiments) written in the same format as the mean-climate files compiled. Maps of visualised statistics are also provided in postscript format.