SPAGETTA Weather Generator Linked with Climate Models May Produce Weather Series for Future Climate 

Abstract

<p>Stochastic weather generators (WGs) are tools for producing weather series, which are statistically similar to the real world weather series. The synthetic series may represent both present and changed (not only the future) climate. In the latter case, WG parameters derived from the observed weather series are modified with climate change scenario, which is typically based on RCM or GCM simulations. As the GCM/RCM simulations are very demanding on computer resources, the numbers of simulations made for individual possible emission scenarios are limited, especially for some (mostly the less probable ones) emission scenarios (e.g. RCP 2.6). Still, many climate change impact studies try to give projections of the CC impacts assuming uncertainties coming from all possible sources, including the modeling uncertainty and  uncertainties in emissions & climate sensitivity. To allow generation of weather series fitting the projection of any GCM forced by any emission scenario, we use a pattern scaling approach, in which the standardized climate change scenario (consisting of changes in climatic characteristic related to 1ºC change in global mean temperature) derived from a given GCM is multiplied by a change in global mean temperature (dTg) projected (for a selected emission scenario and climate sensitivity) by a simple climate model MAGICC.</p><p>In our contribution, we will demonstrate the use of the generator (using SPAGETTA WG, which is our multi-site multi-variate parametric daily WG) in probabilistic projection of future changes in selected climatic characteristics of temperature (T) and precipitation (P); we will focus on spatial hot/cold/dry/wet/hot-dry/hot-wet/cold-dry/cold-wet spells). Standardized climate change scenarios will be derived from multiple GCMs (taken from CMIP5 database) and scaled by dTg projected by MAGICC. Effects of the three above-named sources of uncertainty, as well as the effects of changes in individual statistical characteristics (the means & the site-specific variabilities & the characteristics of the temporal and spatial variability of both T and P) will be assessed.</p><p>Acknowledgements: Projects GRIMASA (Czech Science Foundation, project no. 18-15958S) and SustES (European Structural and Investment Funds, project no. CZ.02.1.01/0.0/0.0/16_019/0000797).</p>