Data: 17/09/2014 à 19/09/2014
Local: São Paulo - Brazil
Evaluation Cmip5 Models Monthly Precipitation Maximum at Southeast Atlantic Region of Brazil Through Statistical Moments and L-Moments (PAP014908)
MARCIO TAVARES NOBREGA, SAULO AIRES DE SOUZA
Floods in a changing climate
Climate change is thought to impact flood management in many unpredictable ways, being foreseen by IPCC an increase in magnitude and frequency of flood events. Such effects are investigated here for the whole range of Global Circulation Models (GCMs) available from CMIP5, focusing on the behavior of monthly maximum precipitation, using as basis for comparison not only observed data as well as Climate Research Unit (CRU) data. Southeast Atlantic region in Brazil was depicted as study case. CRU grid data (half a geographical degree resolution) was chosen as referential, being both observed and GCMs data transformed to it using IDW methodology. In the former case (observed data), radius was restrict to 30km, while in the later one, GCMs, as restriction only the 4 nearest stations were featured. For each and every grid point both statistical moments and l-moments were investigated through coefficient of variation (Cv), coefficient of skewness (Cs) and coefficient of kurtosis (Ck) statistics. These empirical distribution functions were obtained through bootstrap technique, considering each a thousand samples generation for observed data. Meanwhile it was performed a Monte-Carlo simulation to generate a thousand samples for each grid point of the CRU and GCMs, as their available series is longer enough to such, in contrast to shorter observed data. The distributions of calculated statistics were then evaluated, comparing CRU and GCMs curves obtained to the observed one through chi-squared test to a 5% significance level, cumulating the acceptance throughout the whole geographic region. Moreover, GCMs were ranked in accordance with the summary of chi-squared for each one of the considered statistics (Cv, Cs, Ck, and l-Cv, l-Cs, l-Ck ), allowing the selection of the most suitable for the region.