ABSTRACT
Nonstationary trends in hydrological time series have aroused the interest of experts in recent decades. Consequently, conventional methods used for frequency analysis and quantification of risk associated with the occurrence of extreme climate events require adjustment. This study aims to perform the frequency analysis and the quantification of the risk of precipitation for a ten-day period in Tarauacá - Acre, Brazil. The results have shown that nonstationary Gumbel distribution with time-dependent location and scale best fits the observed data, thus allowing the most reliable measurements for risk of rare events, and determining reference quantiles associated with planning horizons or design of hydraulic structures.
Keywords:
Frequency analysis; Nonstationary; Design live level; Tarauacá