“Comparisons Between Observed And Modeled Precipitation And Temperature Extremes In South America During The XX Century (Ipcc 20c3m). Part I: Mean, Variability and MSE”. Matilde Rusticucci (UBA), José Marengo (CPTEC), Olga Penalba (UBA), Madeleine Renom (UR)

Extreme indices used Extreme indices were derived from IPCC models (IPCC web site) and from observations in South America [based on Frich et al. 2002, Vincent et al. 2005, Haylock et al. 2005)]

Objective: how IPCC global models simulate climate extreme indices over South America. we compare simulated indices with station data indices, for –Mean –Standard Deviation and –Mean Square Error: diff between the nearest station to the grid point.

Frost Days (FD) Mean MIROC3.2 hires CCSM GFDL 2.0GFDL 2.1

Frost Days (FD) StD GFDL 2.0GFDL 2.1 CCSM3 MIROC3.2 hires

MIROC3.2hires GFDL 2.1 GFDL 2.0 CCSM3 Frost Days (FD) MSE

Warm Nights (Tn 90) Mean

Warm Nights (TN90) StD.

Warm nights (Tn90) MSE

Consecutive Dry Days (CDD) Mean

Consecutive Dry Days (CDD) StD

Consecutive Dry Days (CDD) MSE

Max. 5 Days precipitation (R5d) Mean

Max. 5 Days precipitation (R5d) Std.D

Max. 5Days Precipitation (R5d) MSE GFDL 2.0 GFDL 2.1 MIROC3.2hires

Heavy Precipitation Days (R10) Mean

Heavy Precipitation Days (R10) StdD

Heavy Precipitation Days (R10) MSE

Summary GFDL better reproduces Fd mean and interanual variability. Tn90 is the best modeled extreme. CDD interannual variability is better represented by MIROC but mean by CCSM3 R10 is well modeled. Days over threshold (R10) are better represented than precip amount (R5d) Southeastern South America Temperature extremes are best modeled by GFDL and precip by CCSM3 and MIROC.