A variância de krigagem na compreensão do comportamento da precipitação pluviométrica no Brasil subtropical

• Ano de defesa: 2020
• Autor(a) principal: Pisoni, Alana
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade Federal de Santa Maria; Brasil; Engenharia Ambiental; UFSM; Programa de Pós-Graduação em Engenharia Ambiental; Centro de Tecnologia
• Palavras-chave em Português: Estações pluviométricas; Variância da krigagem; GAMLSS; Pluviometric stations; Kriging variance; CNPQ::ENGENHARIAS
• Link de acesso: http://repositorio.ufsm.br/handle/1/24904

The pluviometric precipitation presents a large spatial variability and it is a phenomenon characterized by its irregular distribution. Its data are available from a limited number of stations that provide only punctuations mediations. Being the setting of a pluviometric stations’ net it is one of the most important factor to data accuracy, is necessary the net’s optimization. This work aims to analyze the kriging variance as an indicator to determine possible places that need new pluviometric stations in Rio Grande do Sul and to understand the aspect that influence the variability of the pluviometric precipitation. The National Water Agency (ANA) and National Institute of Meteorology (INMET) are organs that make the pluviometric monitoring at Rio Grande do Sul State and available these data. For the study were selected only the stations that had available data for 30 years period (from 1989 to 2018), and containing 5 years or more of consecutive data, aiming to have more than 10% of the years and the minimum of data reliability in order to do not lose the spatial representativeness of raining variability. Thereby it were selected 259 ANA stations and 18 INMET stations, totalizing 277 pluviometric stations. After obtaining the daily precipitation data, the monthly climatological average was performed. From the kriging interpolation method estimated the kriging variance to verify regions that present larger variability, where could be inserted more stations becoming better the sampling area. It were done systemic reductions of the stations from 100% to 95%, 90% and 50% intending to modulate the maximum kriging variance in function of the stations number and the months of the year. The modeling by generalized additive models of position, scale and shape (GAMLSS) was used to verify the relationship of the maximum kriging variance as a function of the number of seasons and the months of the year. Through the kriging variance it was possible to observe the places with larger variability that need of pluviometric stations every month of the year. And through the GAMLSS modulations, it was verified that there isn’t evident relation of the kriging maximum variance with the stations number, so that not necessarily the increase of measurement stations, without the reorganization of the spatial mesh, would cause a significant improvement in predictions.