SEVECLIM: SOFTWARE TO ANALYZE PROBABILISTIC MODELS TO CHARACTERIZE EXTREME CLIMATE VARIABLES

Abstract

Extreme Value Theory is a discipline that develops modern techniques to describe the least common events, which in itself makes it a “unique” discipline. The goal of this theory is to analyze observed extreme values and to estimate values for different return periods. This discipline is of high importance for the National Meteorological Institute and especially to the Clime Center due to the characterization of the extreme climatological variables, this is the reason for the development of a software to provide these services obtaining relevant information as well as an economical gain. The knowledge of the principal climatic variables associated with extreme events, as well as the historical data in series with a observational time period of the highest temporal longitude is also important, for example in construction. In this line of work is required a specific safety margin to elaborate measures to prevent social and economic affectations as well as to maintain the safety of the population. SEVECLIM was developed with the purpose to facilitate the work and to reduce calculation errors. The software allows a user to introduce the extreme variables data. The goal is to characterize through different probability distributions the different extreme climatological variables to obtain the model that best fit the data, to estimate the return periods of 1%, 2 %, 5 % y 10 %, as well as the main statistics. It is show the proposed work methodology to a better understanding in the software utilization. With the
model that best fit the data, it is possible to be closer to reality thus providing a better knowledge of the regime of the studied variable; it will also allow us to offer these services to other institutions to obtain economic gains.