Key points are not available for this paper at this time.
Development, testing and example applications of the pattern-scaling approach for generating future climate change projections are reported here, with a focus on a particular software application called “ClimGen”. A number of innovations have been implemented, including using exponential and logistic functions of global-mean temperature to represent changes in local precipitation and cloud cover, and interpolation from climate model grids to a finer grid while taking into account land-sea contrasts in the climate change patterns. Of particular significance is a new approach for incorporating changes in the inter-annual variability of monthly precipitation simulated by climate models. This is achieved by diagnosing simulated changes in the shape of the gamma distribution of monthly precipitation totals, applying the pattern-scaling approach to estimate changes in the shape parameter under a future scenario, and then perturbing sequences of observed precipitation anomalies so that their distribution changes according to the projected change in the shape parameter. The approach cannot represent changes to the structure of climate timeseries (e.g. changed autocorrelation or teleconnection patterns) were they to occur, but is shown here to be more successful at representing changes in low precipitation extremes than previous pattern-scaling methods.
Osborn et al. (Wed,) studied this question.