SPAGETTA: a Multi-Purpose Gridded Stochastic Weather Generator

0484699 - ÚFA 2018 US eng A - Abstrakt
Dubrovský, Martin - Huth, R. - Rotach, M. W. - Dabhi, H.
SPAGETTA: a Multi-Purpose Gridded Stochastic Weather Generator.
AGU Fall Meeting. s. l.: American Geophysical Union, 2017. NG41A-0115.
[AGU Fall Meeting 2017. 11.12.2017-15.12.2017, New Orleans]
Institucionální podpora: RVO:68378289
Klíčová slova: SPAGETTA * stochastic weather generator * climate change scenarios * validation
Kód oboru RIV: DG - Vědy o atmosféře, meteorologie
https://agu.confex.com/agu/fm17/meetingapp.cgi/Paper/244257

SPAGETTA is a new multisite/gridded multivariate parametric stochastic weather generator (WG). Site-specific precipitation occurrence and amount are modelled by Markov chain and Gamma distribution, the non-precipitation variables are modelled by an autoregressive model conditioned on precipitation occurrence, and the spatial coherence of all variables is modelled following the Wilks’ (2009) approach.
SPAGETTA may be run in two modes. Mode 1: it is run as a classical WG, which is calibrated using weather series from multiple sites, and only then it may produce arbitrarily long synthetic series mimicking the spatial and temporal structure of the calibration data. To generate the weather series representing the future climate, the WG parameters are modified according to the climate change scenario, typically derived from GCM or RCM simulations. Mode 2: the user provides only basic information on the temporal and spatial auto-correlation structure of the weather variables and their mean annual cycle; the generator itself derives the parameters of the underlying AR model, which produces the multi-site weather series.
The contribution consists of following parts: (a) Model of the WG. (b) Validation of WG in terms of the spatial temperature and precipitation characteristics, including characteristics of spatial hot/cold/dry/wet spells. (c) Results of the climate change impact experiment, in which the WG parameters representing the spatial and temporal variability are modified using the climate change scenarios and the effect on the above spatial validation indices is analysed. In this experiment, the WG is calibrated using the E-OBS gridded daily weather data for several European regions, and the climate change scenarios are derived from the selected RCM simulations (CORDEX database). (d) The second mode of operation will be demonstrated by results obtained while developing the methodology for assessing collective significance of trends in multi-site weather series.
Trvalý link: http://hdl.handle.net/11104/0279835