• English 
  • Spanish 

Evaluation and projection of extreme temperature percentiles by means of statistical and dynamical downscaling methods

Conferencia: 8º Congreso Internacional AEC
Año: 2012
Tipo de contribución: Oral
Archivo PDF: 2012_Casanueva_PublicacionAEC.pdf
Poster: 2012_Casanueva_AEC_Charla.pdf

The study of extreme events has become of great interest in the recent years due to their direct impact on society. Extremes can be evaluated by using either extreme value statistics or extreme indicators, the latter being based in order statistics on the tail of the probability distribution (typically percentiles). In this study we analyze the highest (95p) and the lowest (5p) percentiles in maximum and minimum temperatures, respectively, derived from different downscaling methods (statistical and dynamical) in the Iberian Peninsula. In particular, we analyze the results of the esTcena and ESCENA projects, two strategic actions of Plan Nacional de I+D+i 2008-2011 funded by the Spanish government, which contributed to the new version of the regional climate change scenarios program Escenarios-PNACC 2012 within Plan Nacional de Adaptación al Cambio Climático.
First, the skill of the downscaling methods to reproduce extreme percentiles is tested in present climate conditions, using reanalysis-driven simulations. The comparison among the different methods is performed in terms of the seasonal bias, considering the public gridded dataset Spain02, a new regular (approximately 20km) daily gridded precipitation and temperature dataset covering continental Spain and Balearic Islands.
Secondly, we analyze future projections in different climate change scenarios to check the increments and the uncertainty of the results up to the mid of the century. We also study the effect of nesting the methods to different Global Circulation Models (GCMs), using the 20C3M historical scenario as reference. By analyzing these changes, we are able to extract differences due to the downscaling method and to the driving GCM.

Palabras clave: