Fechar
Metadados

@MastersThesis{Freire:2014:CaAvSi,
               author = "Freire, Julliana Larise Mendon{\c{c}}a",
                title = "Calibra{\c{c}}{\~a}o e avalia{\c{c}}{\~a}o de um sistema 
                         regional de previs{\~a}o clim{\'a}tica sazonal para a 
                         Am{\'e}rica do Sul durante o per{\'{\i}}odo mar{\c{c}}o-maio",
               school = "Instituto Nacional de Pesquisas Espaciais (INPE)",
                 year = "2014",
              address = "S{\~a}o Jos{\'e} dos Campos",
                month = "2014-05-16",
             keywords = "calibra{\c{c}}{\~a}o, downscaling, nordeste do Brasil, 
                         previs{\~a}o sazonal, calibration, downcaling, northeast of 
                         Brazil, seasonal forecast.",
             abstract = "As previs{\~o}es clim{\'a}ticas sazonais t{\^e}m 
                         contribu{\'{\i}}do efetivamente para o planejamento de 
                         atividades de diversos setores da sociedade, tais como, energia, 
                         agricultura e sa{\'u}de. A necessidade de um maior detalhamento 
                         das previs{\~o}es, atrav{\'e}s da diminui{\c{c}}{\~a}o da 
                         escala (ou aumento da resolu{\c{c}}{\~a}o) espacial da 
                         informa{\c{c}}{\~a}o clim{\'a}tica, tem concentrado nos 
                         {\'u}ltimos anos esfor{\c{c}}os cient{\'{\i}}ficos em 
                         v{\'a}rias regi{\~o}es do globo. A aplica{\c{c}}{\~a}o da 
                         t{\'e}cnica de redu{\c{c}}{\~a}o de escala atrav{\'e}s da 
                         modelagem din{\^a}mica regional, conhecida como 
                         \emph{downscaling} din{\^a}mico, vem sendo utilizada com esse 
                         objetivo. Desta forma, o presente trabalho tem como objetivo 
                         calibrar e avaliar a destreza do modelo regional atmosf{\'e}rico 
                         \emph{Brazilian developments on the Regional Atmospheric Modeling 
                         System} (BRAMS) aninhado unidirecionalmente ao Modelo de 
                         Circula{\c{c}}{\~a}o Geral Atmosf{\'e}rico (MCGA) do Centro de 
                         Previs{\~a}o de Tempo e Estudos Clim{\'a}ticos (CPTEC) em prever 
                         a variabilidade clim{\'a}tica dos padr{\~o}es atmosf{\'e}ricos 
                         dominantes sobre a Am{\'e}rica do Sul (AS), para o trimestre 
                         mar{\c{c}}o-abril-maio (MAM), que representa o pico da 
                         esta{\c{c}}{\~a}o chuvosa da regi{\~a}o Nordeste do Brasil 
                         (NEB) durante o per{\'{\i}}odo de 1979 a 2010. As previs{\~o}es 
                         com o BRAMS foram produzidas utilizando um conjunto de 10 membros 
                         na resolu{\c{c}}{\~a}o espacial de aproximadamente 30 km de 
                         latitude por 30 km de longitude, tendo como condi{\c{c}}{\~o}es 
                         de contorno a temperatura da superf{\'{\i}}cie do mar persistida 
                         e a umidade do solo climatol{\'o}gica. Este trabalho trata-se de 
                         uma iniciativa pioneira de avalia{\c{c}}{\~a}o sistem{\'a}tica 
                         e comparativa de um modelo global e um modelo regional para a 
                         Am{\'e}rica do Sul, ambos os modelos configurados em modo de 
                         previs{\~a}o clim{\'a}tica sazonal por conjunto (10 membros) 
                         rodados por um longo per{\'{\i}}odo (32 anos, 1979-2010), 
                         reproduzindo assim as mesmas condi{\c{c}}{\~o}es encontradas no 
                         ambiente operacional de previs{\~a}o clim{\'a}tica. Ap{\'o}s o 
                         processo de calibra{\c{c}}{\~a}o, o modelo BRAMS conseguiu 
                         capturar o padr{\~a}o de precipita{\c{c}}{\~a}o associada aos 
                         eventos clim{\'a}ticos extremos, como a redu{\c{c}}{\~a}o 
                         (aumento) da precipita{\c{c}}{\~a}o principalmente no leste da 
                         regi{\~a}o Norte e NEB e o aumento (redu{\c{c}}{\~a}o) de 
                         precipita{\c{c}}{\~a}o na regi{\~a}o Sul durante o evento de El 
                         Niņo de 1983 (La Niņa de 1989). Os resultados mostraram que 
                         atrav{\'e}s da t{\'e}cnica de \emph{downscaling} o modelo 
                         regional BRAMS conseguiu reproduzir o padr{\~a}o espacial da 
                         precipita{\c{c}}{\~a}o e temperatura e as principais 
                         caracter{\'{\i}}sticas de circula{\c{c}}{\~a}o, durante o pico 
                         da esta{\c{c}}{\~a}o chuvosa do NEB. As an{\'a}lises das 
                         composi{\c{c}}{\~o}es de El Niņo e La Niņa mostraram que o 
                         modelo regional representou razoavelmente bem o padr{\~a}o 
                         an{\^o}malo das vari{\'a}veis estudadas quando comparado aos 
                         padr{\~o}es observados. O uso do modelo regional BRAMS resultou 
                         em melhorias na previs{\~a}o do sinal das anomalias de 
                         precipita{\c{c}}{\~a}o sobre os pa{\'{\i}}ses da costa Oeste e 
                         Norte da AS (Bol{\'{\i}}via, Peru, Col{\^o}mbia e Venezuela), 
                         centro-leste da Argentina e sobre o Estado do Acre. Sobre a 
                         regi{\~a}o Nordeste do Brasil ambos os modelos (BRAMS e MCGA) 
                         apresentaram desempenho semelhante nas previs{\~o}es de anomalias 
                         de precipita{\c{c}}{\~a}o. As correla{\c{c}}{\~o}es de 
                         anomalia de temperatura utilizando o conjunto de previs{\~o}es do 
                         BRAMS possuem valores mais expressivos do que as 
                         correla{\c{c}}{\~o}es resultantes do MCGA, principalmente sobre 
                         a regi{\~a}o Norte do Brasil e sobre o Peru. Erros 
                         sistem{\'a}ticos foram verificados em algumas regi{\~o}es da AS, 
                         associados as parametriza{\c{c}}{\~o}es f{\'{\i}}sicas e 
                         condi{\c{c}}{\~o}es inicias e de contorno aqui utilizados, 
                         sugerindo a necessidade de realizar ajustes nas 
                         configura{\c{c}}{\~o}es e parametriza{\c{c}}{\~o}es do modelo 
                         em futuras vers{\~o}es para a melhoria da qualidade das 
                         previs{\~o}es. ABSTRACT: Seasonal climate predictions have 
                         effectively contributed to activities planning in various societal 
                         sectors, such as energy, agriculture and health. The need for more 
                         detailed forecasts by increasing climate information spatial 
                         resolution has focused research efforts in various regions of the 
                         globe in recent years. The application of the downscaling 
                         technique through regional dynamic modeling, known as dynamical 
                         downscaling, has been used for this purpose. This study aims to 
                         calibrate and evaluate the skill of the regional model Brazilian 
                         developments on the Regional Atmospheric Modeling System (BRAMS) 
                         nested unidirectionally with the Atmospheric General Circulation 
                         Model (AGCM) of the Center for Weather Forecast and Climate 
                         Studies (CPTEC), in predicting climate variability of the dominant 
                         climate patterns over South America (SA). The study was performed 
                         for the March-April-May (MAM) period, which represents the peak of 
                         the rainy season in Northeast Brazil (NEB) during the period from 
                         1979 to 2010.The predictions with the BRAMS were produced using a 
                         set of 10 members with spatial resolution of 30 km by 30 km in 
                         latitude and longitude, having the persisted sea surface 
                         temperatures and climatological soil moisture as boundary 
                         conditions. This work is a pioneering initiative performing a 
                         systematic comparative evaluation of a global model and a regional 
                         model for South America. Both models were configured in ensembles 
                         (10 members) seasonal climate forecast mode and integrated during 
                         a long period (32 years, 1979-2010), thus reproducing the same 
                         conditions found in the climate prediction operational 
                         environment. After the calibration process BRAMS was able to 
                         capture rainfall patterns associated to extreme climate events, 
                         such as the decrease (increase) in precipitation over the eastern 
                         portion of North region and over the Northeast region of Brazil, 
                         and increase (decrease) in precipitation over the Southern region 
                         of Brazil during the 1983 El Niņo (1989 La Niņa) event. The 
                         results showed that BRAMS was able to reproduce the spatial using 
                         the technique of downscaling the patterns precipitation and 
                         temperature and the main circulation features over South America, 
                         during the peak of NEB rainy season. El Niņo and La Niņa composite 
                         analyses showed that the regional model represented reasonably 
                         well the anomalous patterns of the studied variables when compared 
                         to the observed. The use of the regional model BRAMS resulted in 
                         improvements in precipitation anomaly forecast signal over of 
                         western and northern coast of SA (Bolivia, Peru, Colombia and 
                         Venezuela), over east-central Argentina and the over State of Acre 
                         over the Northeast region of Brazil both models (BRAMS and AGCM) 
                         showed similar performance when forecasting precipitation 
                         anomalies. As for temperature anomaly forecast BRAMS showed higher 
                         performance, given by the correlation coefficient between forecast 
                         and observed anomalies over the period 1979-2010, than the AGCM, 
                         particularly over northern Brazil and Peru. Systematic forecast 
                         errors were found in some regions associated to the physical 
                         parameterizations and initial and boundary conditions used in this 
                         study suggesting the need for adjustments in the model 
                         configuration and parameterization in future model version with 
                         the aim of improving the quality of forecasts.",
            committee = "Herdies, Dirceu Luis (presidente) and Freitas, Saulo Ribeiro de 
                         (orientador) and Coelho, Caio Augusto dos Santos (orientador) and 
                         Rivero, Silvio Nilo Figueroa and Souza, Enio Pereira de",
         englishtitle = "Calibration and evaluation of a regional seasonal climate forecast 
                         system for Sul America during the march-may period",
             language = "pt",
                pages = "131",
                  ibi = "8JMKD3MGP8W/3G5UGHP",
                  url = "http://urlib.net/rep/8JMKD3MGP8W/3G5UGHP",
           targetfile = "publicacao.pdf",
        urlaccessdate = "21 jan. 2021"
}


Fechar