@MastersThesis{Silva:2013:AvMoWR,
author = "Silva, Adaiana Francisca Gomes da",
title = "Avalia{\c{c}}{\~a}o do modelo WRF ao perfil do vento no Centro
de Lan{\c{c}}amento de Alc{\^a}ntara",
school = "Instituto Nacional de Pesquisas Espaciais (INPE)",
year = "2013",
address = "S{\~a}o Jos{\'e} dos Campos",
month = "2013-02-27",
keywords = "engenharia do vento, radiossondagem, simula{\c{c}}{\~a}o
num{\'e}rica, lan{\c{c}}amento de foguetes, micrometeorologia,
wind engineering, radiosounding, numerical simulation, rocket
launching, micrometeorology.",
abstract = "Este trabalho tem como objetivo avaliar a capacidade de um modelo
num{\'e}rico regional de mesoescala (no caso, o modelo WRF) em
representar o escoamento atmosf{\'e}rico da regi{\~a}o do Centro
de Lan{\c{c}}amento de Alc{\^a}ntara (CLA), possibilitando seu
uso para previs{\~a}o do vento quando em ocasi{\~o}es de
lan{\c{c}}amentos naquela base, fazendo tamb{\'e}m uma
an{\'a}lise sazonal ao comportamento do modelo. A
motiva{\c{c}}{\~a}o deste estudo {\'e} o impacto direto e
profundo que as caracter{\'{\i}}sticas do vento podem ter na
trajet{\'o}ria de um foguete em lan{\c{c}}amento, podendo
desvi{\'a}-lo, causar imprevistos e at{\'e} mesmo acidentes.
Previs{\~o}es do vento de 72 horas foram feitas a partir de
condi{\c{c}}{\~o}es iniciais fornecidas pelo \textit{Global
Forecasting System} (GFS), do {National Centers for Environmental
Prediction} (NCEP). As rodadas foram configuradas com tr{\^e}s
dom{\'{\i}}nios aninhados e resolu{\c{c}}{\~o}es horizontais
de 9, 3 e 1 km. Simula{\c{c}}{\~o}es foram iniciadas {\`a}s 00
e {\`a}s 12 UTC para o per{\'{\i}}odo de 16 a 26 de setembro de
2008, esta{\c{c}}{\~a}o seca, e de 19 a 25 de mar{\c{c}}o de
2010, esta{\c{c}}{\~a}o chuvosa da regi{\~a}o,
per{\'{\i}}odos para os quais se disp{\~o}e de alta densidade
de dados observados (radiossondagens). As sa{\'{\i}}das do
modelo foram confrontadas contra observa{\c{c}}{\~o}es
atrav{\'e}s de tr{\^e}s {\'{\i}}ndices estat{\'{\i}}sticos:
vi{\'e}s, erro quadr{\'a}tico m{\'e}dio e {\'{\i}}ndice de
concord{\^a}ncia de Willmott (d), calculados a cada 6 horas.
Foram realizados testes de sensibilidade nos quais foram variados
e comparados diferentes par{\^a}metros din{\^a}micos e
f{\'{\i}}sicos, sejam eles: tamanhos de dom{\'{\i}}nios,
n{\'u}mero de n{\'{\i}}veis verticais, \textit{spin-up time} e
parametriza{\c{c}}{\~o}es de Camada Limite Planet{\'a}ria. Os
testes mostraram pouca sensibilidade do modelo com
rela{\c{c}}{\~a}o ao tamanho dos dom{\'{\i}}nios e ao
\textit{spin-up time}, com diferen{\c{c}}as na ordem da segunda
ou terceira casa decimal para d e da primeira ou segunda casa
decimal para os erros. Mas, com rela{\c{c}}{\~a}o ao n{\'u}mero
de n{\'{\i}}veis verticais, diferen{\c{c}}as de at{\'e} 1 m/s
nos erros foram favor{\'a}veis {\`a} utiliza{\c{c}}{\~a}o de
mais n{\'{\i}}veis verticais, no caso 42. 6 horas ficaram
definidas como tempo de ajuste, descartadas. Nas
parametriza{\c{c}}{\~o}es, as diferen{\c{c}}as entre os
{\'{\i}}ndices dos quatro esquemas testados (YSU, ETA, ACM2 e
MYNN2,5) tamb{\'e}m foram muito pequenas (primeira ou segunda
casa decimal) e os mais adequados foram definidos a partir de uma
avalia{\c{c}}{\~a}o te{\'o}rica. Em fun{\c{c}}{\~a}o do
{\'{\i}}ndice de Willmott, o modelo se mostrou razo{\'a}vel
para representar o perfil vertical do vento, dentro de suas
limita{\c{c}}{\~o}es, n{\~a}o mostrando diferen{\c{c}}a em seu
desempenho entre as esta{\c{c}}{\~o}es seca ou chuvosa, com
valores m{\'a}ximos na ordem de 0,90 e m{\'e}dios da ordem de
0,70. N{\~a}o foi poss{\'{\i}}vel obter um intervalo de horas
padr{\~a}o de todas as simula{\c{c}}{\~o}es em que a qualidade
de previs{\~a}o fosse melhor ou pior. ABSTRACT: This study aims
to assess the ability of a regional mesoscale numerical model (in
this case, the WRF model) to represent the atmospheric flow in the
region of Alcantara Launching Center (CLA), allowing its use for
predicting the wind on occasions of rocket launching, also doing
an analysis to the seasonal behavior of the model. The motivation
of this study is the direct and profound impact that the
characteristics of wind can have on the trajectory of a rocket
launch, as deflect it and even cause unforeseen accidents. 72
hours wind forecasts were made from initial conditions provided by
the Global Forecasting System (GFS), of the National Centers for
Environmental Prediction (NCEP). The configuration of the runs was
three grid nested with a horizontal resolution of 9, 3 and 1 km.
Simulations were initiated at 00 and 12 UTC for the period from 16
to 26 September 2008, dry season, and from 19 to 25 March 2010,
the rainy season in the region, periods for which we have high
data density observed (radiosonde). The model outputs were
compared against observations using three statistical indices:
bias, mean squared error and Willmott agreement index, calculated
every 6 hours. Sensitivity tests were made, in which were varied
and compared different physical and dynamic parameters, they are:
domain sizes, number of vertical levels, spin-up time and
Planetary Boundary Layer parameterizations. Tests showed little
sensitivity of the model with respect to the domain size and to
the spin-up time, with differences on the order of the second or
third decimal place for d and the first or the second decimal
place for errors. But about the number of vertical levels,
differences of up to 1 m/s on errors were in favor of using more
vertical levels, in the case 42. 6 hours were defined as spin-up
time, discarded. In the parameterizations, the differences between
the indices of the four tested schemes (YSU, ETA, ACM2 and
MYNN2.5) were also very small (first or second decimal place) and
the most suitable were defined from a theoretical evaluation.
According to the Willmott index, the model was reasonable to
represent the vertical wind profile, within its limitations,
showing no difference in performance between the dry or rainy
seasons, with maximum values on the order of 0.90 and, on the
average, on the order of 0.70. It wasn´t possible to obtain a time
interval pattern of all simulations where quality prediction was
better or worse.",
committee = "Oyama, Marcos Daisuke (presidente) and Fisch, Gilberto Fernando
(orientador) and Leal Junior, Jo{\~a}o Bosco Ver{\c{c}}osa",
copyholder = "SID/SCD",
englishtitle = "WRF model assessment of the wind profile at the Alcantara
Launching Center",
language = "pt",
pages = "98",
ibi = "8JMKD3MGP7W/3DGJ992",
url = "http://urlib.net/ibi/8JMKD3MGP7W/3DGJ992",
targetfile = "publicacao.pdf",
urlaccessdate = "09 maio 2024"
}