Accounting for uncertainty in climate projections and the premise of decadal climate predictions Didier Swingedouw & Giovanni Sgubin
Outlines • News from the global warming • What is a climate projection? • Decadal prediction • Downscaling methods • Results from our group => Giovanni Sgubin
News from the global warming Source: NASA-GISS
Climate models CMIP: coupled model intercomparison project Þ 6 th phase is on- going Þ CMIP5
Detection-attribution
Climate projections IPCC 2013
RCP26 RCP8.5 IPCC 2013
Hawkins & Sutton 2009 Sources of uncertainties Fraction de variance expliquée en % Anomalie de température (1971-2000) en °C Hawkins & Sutton 2009 Années Fraction de variance expliquée en %
Decadal variations at the regional scale Spring minimum temperature in Arcachon Years
Decadal predictions Weather Seasonal Decadal Centennial Ice age cycle prediction prediction projection day week month year decade century millennium
Downscaling v Statistical downscaling: e.g. Analog method => create statistical relationship between large scale predictors and regional scale predictant SAFRAN v Dynamical downscaling : data e.g. CORDEX project: regional model with higher resolution nested in a global model
Debiasing model data v Removing potential drift in climate predictions v Correct the modelled distribution of a given variable with the observed one: quantile-quantile approach v Link (EU projects) with the Climate Data Factory, a company dedicated to develop climate services and who is applying the methods from Matthieu Vrac (LSCE) on various field. v Use of cumulative Distribution Function-transform
THANK’S!
Recommend
More recommend