Bibliographic citations
Fernández, C., (2015). Frecuencia de precipitaciones máximas según los L-momentos en la cuenca hidrográfica del lago Titicaca [Tesis, Universidad Nacional Agraria La Molina]. https://hdl.handle.net/20.500.12996/2161
Fernández, C., Frecuencia de precipitaciones máximas según los L-momentos en la cuenca hidrográfica del lago Titicaca [Tesis]. : Universidad Nacional Agraria La Molina; 2015. https://hdl.handle.net/20.500.12996/2161
@mastersthesis{renati/248040,
title = "Frecuencia de precipitaciones máximas según los L-momentos en la cuenca hidrográfica del lago Titicaca",
author = "Fernández Palomino, Carlos Antonio",
publisher = "Universidad Nacional Agraria La Molina",
year = "2015"
}
Extreme precipitation often leads to the occurrence of flood events with consequent loss of life, agricultural production and infrastructure in the Vertiente of Titicaca Lake (VT). However, an understanding of the frequency of occurrence of extreme events, either for design purposes or disaster planning, is often limited by poor availability and quality of data in the desired temporal and spatial scale. Therefore, this thesis develops the regional frequency analysis (ARF) of extreme precipitation in 24 hours, using L-moments and the index flood proceeding for estimating quantiles of precipitation in VT (33 raingauge stations selected). Initially detailed quality analysis and verification of the assumptions of frequency analysis of the series of maximum precipitation were performed, applying for this different verification tests for outliers data, trend, homogeneity and independence of the time series. The method Lmoments through the measure of heterogeneity, allows to define objectively that Vertiente of Titicaca Lake forms one hydrologically homogeneous region in terms of the frequency of maximum precipitation. According to diagram Lmoment ratios and Z-statistic, generalized extreme value (GEV) and generalized normal (GNO) distributions are the best choices for estimates of regional quantiles. However, a robustness analysis approach based on Monte Carlo simulations reveals that GNO is slightly more robust than GEV distribution to estimate quantiles in large design return periods (50 years to more). It is found that the application at VT of the Gumbel distribution function in the local frequency analysis of maximum precipitation leads to an overestimation of the estimated quantiles. Finally, the index flood map was generated, valid for index flood estimates in sites without measuring precipitation in VT.
This item is licensed under a Creative Commons License