@MastersThesis{Reis:2024:AnViCo,
author = "Reis, Bruna Larissa Rabelo dos",
title = "An{\'a}lise de vi{\'e}s e corre{\c{c}}{\~a}o das estimativas
de vento em superf{\'{\i}}cie do modelo WRF a partir de dados de
escater{\^o}metros na regi{\~a}o oeste do oceano Atl{\^a}ntico
Tropical e Sul",
school = "Instituto Nacional de Pesquisas Espaciais (INPE)",
year = "2024",
address = "S{\~a}o Jos{\'e} dos Campos",
month = "2024-08-05",
keywords = "corre{\c{c}}{\~a}o de vi{\'e}s, vento na superf{\'{\i}}cie do
mar, WRF, ASCAT, HSCAT, bias correction, sea surface wind, WRF,
ASCAT, HSCAT.",
abstract = "O campo de vento na superf{\'{\i}}cie do mar a 10m {\'e}
fundamental para determinar padr{\~o}es e fen{\^o}menos ligados
a processos de intera{\c{c}}{\~a}o oceano-atmosfera, e obter
informa{\c{c}}{\~o}es a respeito da sua variabilidade
espa{\c{c}}o-temporal contribui para algumas
aplica{\c{c}}{\~o}es, como a Previs{\~a}o Num{\'e}rica de
Tempo e Clima (PTC). A modelagem num{\'e}rica para a PTC carece
de uma boa representa{\c{c}}{\~a}o de processos de r{\'a}pida
evolu{\c{c}}{\~a}o como os de mesoescala, que s{\~a}o
fundamentais para resolver processos de maiores escalas nos
modelos. Os escater{\^o}metros s{\~a}o sensores orbitais
desenvolvidos para estimar o vento a partir da rugosidade da
superf{\'{\i}}cie do mar, e a utiliza{\c{c}}{\~a}o dos seus
dados na modelagem num{\'e}rica {\'e} fundamental para melhorar
a precis{\~a}o das previs{\~o}es do U10. Perante isso, a
principal motiva{\c{c}}{\~a}o desta pesquisa foi fazer uma
estimativa e corre{\c{c}}{\~a}o do vi{\'e}s do campo de U10 do
modelo Weather Research and Forecasting (WRF) com base no
m{\'e}todo de corre{\c{c}}{\~a}o com escater{\^o}metro
(Scatterometer Correction - SC) na regi{\~a}o ocidental do oceano
Atl{\^a}ntico Tropical e Sul. Neste estudo foram utilizados dados
dos sensores escater{\^o}metros Advanced SCATterometer (ASCAT) e
HY SCATterometer (HSCAT), de diferentes miss{\~o}es espaciais,
coletados na {\'a}rea de estudo. Inicialmente, os dados dos
sensores foram compilados em um mosaico (SCAT) a fim de aumentar a
amostragem espacial. Posteriormente foi feita a
valida{\c{c}}{\~a}o pontual do mosaico e das sa{\'{\i}}das do
WRF, sem (WRFi) e com a corre{\c{c}}{\~a}o (WRF*), com dados
observacionais medidos in situ pelo projeto Prediction and
Research Moored Array in the Atlantic (PIRATA) para as componentes
zonal (u10) e meridional (v10) do U10, atrav{\'e}s das
m{\'e}tricas de correla{\c{c}}{\~a}o linear (r), Ra{\'{\i}}z
do Erro Quadr{\'a}tico M{\'e}dio (RSME) e vi{\'e}s calculadas
para o per{\'{\i}}odo de 01/04/2022 a 31/03/2023. A partir do
m{\'e}todo SC aplicado aos dados validados foi gerado o produto
de corre{\c{c}}{\~a}o do vi{\'e}s do modelo (WRF*). Os valores
estat{\'{\i}}sticos para as componentes u10 e v10 indicaram que
o WRF* (r entre 0,49 e 0,86, RSME entre 1,11 e 1,72 m/s e
vi{\'e}s entre -0,46 e 0,33 m/s) teve um desempenho superior ao
WRFi (r entre 0,45 e 0,81, RSME entre 1,21 e 1,96 m/s e vi{\'e}s
entre -0,22 e 1,17 m/s), ficando abaixo apenas do SCAT (r entre
0,72 e 0,89, RSME entre 0,79 e 1,25 m/s e vi{\'e}s entre -0,45 e
0,21 m/s). Atrav{\'e}s da an{\'a}lise de dois estudos de caso de
eventos meteorol{\'o}gicos, verificou-se que o produto gerado
apresentou melhorias na representa{\c{c}}{\~a}o de processos
f{\'{\i}}sicos de mesoescala como a forma{\c{c}}{\~a}o de
frentes e convec{\c{c}}{\~a}o. O SC mostrou que o WRFi
apresentou mais falhas na captura dos processos em
rela{\c{c}}{\~a}o {\`a} v10. Conclui-se que o WRF* {\'e}
{\'u}til para compor as condi{\c{c}}{\~o}es iniciais e de
contorno em experimentos num{\'e}ricos para estudos de
intera{\c{c}}{\~a}o oceano-atmosfera. ABSTRACT: The wind field
at the sea surface at 10m is fundamental for determining patterns
and phenomena related to ocean-atmosphere interaction processes,
and obtaining information about its spatial-temporal variability
contributes to some applications, such as Numerical Weather
Prediction (NWP). Numerical modeling for NWP requires a good
representation of rapidly evolving processes such as mesoscale,
which are essential for resolving largerscale processes in models.
Scatterometers are orbital sensors designed to estimate wind based
on the roughness of the sea surface, and the use of their data in
numerical modeling is fundamental for improving the accuracy of
U10 forecasts. Given this, the main motivation of this research
was to estimate and correct the bias of the U10 field from the
Weather Research and Forecasting (WRF) model based on the
Scatterometer Correction (SC) method in the western region of the
Tropical and South Atlantic ocean. In this study, data from the
Advanced SCATterometer (ASCAT) and HY SCATterometer (HSCAT)
sensors from different space missions, collected in the study
area, were used. Initially, the sensor data were compiled into a
mosaic (SCAT) to increase spatial sampling. Subsequently, point
validation of the mosaic and WRF outputs, without (WRFi) and with
correction (WRF*), was performed using observational data measured
in situ by the Prediction and Research Moored Array in the
Atlantic (PIRATA) project for the zonal (u10) and meridional (v10)
components of U10, using the metrics of linear correlation (r),
Root Mean Square Error (RMSE), and bias calculated for the period
from 04/01/2022 to 03/31/2023. Based on the SC method applied to
the validated data, the model bias correction product (WRF*) was
generated. The statistical values for the u10 and v10 components
indicated that WRF* (r between 0.49 and 0.86, RMSE between 1.11
and 1.72 m/s, and bias between -0.46 and 0.33 m/s) performed
better than WRFi (r between 0.45 and 0.81, RMSE between 1.21 and
1.96 m/s, and bias between -0.22 and 1.17 m/s), falling just below
SCAT (r between 0.72 and 0.89, RMSE between 0.79 and 1.25 m/s, and
bias between -0.45 and 0.21 m/s). Through the analysis of two case
studies of meteorological events, it was found that the generated
product showed improvements in the representation of mesoscale
physical processes such as front formation and convection. SC
showed that WRFi had more failures in capturing processes related
to v10. It is concluded that WRF* is useful for composing initial
and boundary conditions in numerical experiments for
ocean-atmosphere interaction studies.",
committee = "Souza, Ronald Buss de (presidente) and Oliveira, Natalia Rudorff
(orientadora) and Parise, Cl{\'a}udia Klose",
englishtitle = "Analysis of bias and correction of WRF model surface wind
estimates using scatterometer data in the western Tropical and
South Atlantic ocean region",
language = "pt",
pages = "109",
ibi = "8JMKD3MGP3W34T/4BQKAUE",
url = "http://urlib.net/ibi/8JMKD3MGP3W34T/4BQKAUE",
targetfile = "publicacao.pdf",
urlaccessdate = "2026, Oct. 11"
}