Análise de eventos extremos de precipitação no Nordeste do Brasil

Abstract

Extreme weather events are becoming increasingly frequent in Brazil, causing floods, drought, forest fires, landslides, river flows, cold or hot flashes, among others. This events have major implications for society, especially health, agriculture and water resources. Given this scenario, having knowledge of the behavior and frequency with which extreme values occur is of great importance to society. However, a major difficulty factor in performing these analyzes is the quality of the data series used, especially those on the daily scale. Therefore, the objectives of this study are to evaluate the quality of multiple imputation methods for filling in missing daily rainfall data, to investigate the ability of Tropical Rainfall Measuring Mission (TRMM) Multisatellite Precipitation Analysis (TMPA) 3B42 product to estimate occurrence and intensity of daily and extreme precipitation events, to estimate the return level of extreme daily precipitation events and to classify the microphysical parameters of the extreme precipitation generating clouds for Northeaster of Brazil (NEB). This region has high spatial and temporal variability of precipitation and is vulnerable to extremes of daily precipitation. For this purpose, it made use of daily rainfall data from rainfall during the period from 01/01/1986 to 12/31/2015 and from products 3B42, 2ACLIM and 2A12 from the TRMM satellite during the period from 01/01/2000 to 12/31/2015. Data analysis was performed by means of statistical measures and methods such as mean standard deviation, bias, mean square error, correlation, multiple imputation, t-student test, sensitivity analysis, cluster analysis, extreme value theory, analysis of variance, F test, Tukey test, among others. The main results indicated that multiple data imputation using the Bootstrap EM algorithm method could be a tool that corroborates the reconstruction of historical series of daily precipitation data. Overall, TRMM 3B42 product data performed well when estimating precipitation values for NEB. Its quality varies according to the location and time scale at which precipitation occurs. Estimates of the return level indicate that the expected daily extreme rainfall intensity depends on the seasonal period and where it occurs. The East of the NEB stands out as the region where the highest intensities of extreme precipitation are expected. Parameters such as amount of liquid water and integrated ice, height of the freezing top and type of rain were characterized, generating extreme precipitation clouds in relation to different NEB regions.