@InProceedings{DiasNetoCoel:2015:AvPrOp,
author = "Dias Neto, Jos{\'e} and Coelho, Simone Marilene Sievert da
Costa",
affiliation = "{Instituto Nacional de Pesquisas Espaciais (INPE)} and {Instituto
Nacional de Pesquisas Espaciais (INPE)}",
title = "Avalia{\c{c}}{\~a}o da profundidade optica dos Aeross{\'o}is
obtidas pela nova gera{\c{c}}{\~a}o de sat{\'e}lites polar da
NOAA sobre o territ{\'o}rio brasileiro",
booktitle = "P{\^o}steres",
year = "2015",
organization = "Simp{\'o}sio Internacional de Climatologia, 6. (SIC)",
keywords = "Aerossois, Sat{\'e}lite, S-NPP, Aerosol, Satellite, S-NPP.",
abstract = "Os esfor{\c{c}}os para obter a Profundidade {\'O}ptica dos
Aeross{\'o}is (AOT) tem crescido com o decorrer dos anos e
est{\'a} associado a necessidade de entender a
intera{\c{c}}{\~a}o com a radia{\c{c}}{\~a}o e o melhorar os
modelos num{\'e}ricos. Os aeross{\'o}is atuam como controladores
da radia{\c{c}}{\~a}o e das nuvens. S{\~a}o classificados como
naturais (ex. poeira, sal marinho) ou antropog{\^e}nico (ex.
polui{\c{c}}{\~a}o, fuligem). No Brasil, a maioria das fontes de
aeross{\'o}is s{\~a}o da queima de biomassa. Devido a grande
varia{\c{c}}{\~a}o espacial e temporal, dados de sat{\'e}lites
s{\~a}o usados como a principal forma de obter AOT em escala
global. Em continuidade do programa de Observa{\c{c}}{\~a}o do
Sistema Terrestre, o Suomi National Polar-orbiting Partnership
(S-NPP) foi lan{\c{c}}ado em 2011, sendo o primeiro da nova
gera{\c{c}}{\~a}o de sat{\'e}lites do Sistema Nacional de
Sat{\'e}lites Ambientais de Orbita Polar. O S-NPP {\'e} composto
por cinco sensores, um deles {\'e} o Visible Infrared Imaging
Radiometer Suite (VIIRS) com vinte e duas bandas, das quais nove
s{\~a}o empregadas para calcular AOT. Desde 2013, a
DSA/CPTEC/INPE tem processado AOT do VIIRS/NPP permitindo
monitorar o aerossol no Brasil. O objetivo deste trabalho {\'e}
apresentar a valida{\c{c}}{\~a}o do AOT/S-NPP Produto
Intermedi{\'a}rio (resolu{\c{c}}{\~a}o de 0.750 x 0.750km) por
compara{\c{c}}{\~a}o com dados de superf{\'{\i}}cie
provenientes de nove esta{\c{c}}{\~a}o da rede de
observa{\c{c}}{\~a}o AERONET. As esta{\c{c}}{\~o}es est{\~a}o
distribu{\'{\i}}das pelo Brasil, uma na regi{\~a}o nordeste,
uma na regi{\~a}o sudeste, uma na regi{\~a}o sul, tr{\^e}s na
regi{\~a}o norte e tr{\^e}s na regi{\~a}o centro-oeste.
Adicionalmente. AOT/MODIS do sat{\'e}lite Aqua {\'e} usado na
compara{\c{c}}{\~a}o devido ser um produto bem estabelecido
funcionando por mais de dez anos. A metodologia de an{\'a}lise
{\'e} baseada em Ichoku et al.,(2002) no qual AOT do
sat{\'e}lite {\'e} a m{\'e}dia espacial de uma caixa de 40x40km
centrada na esta{\c{c}}{\~a}o da AERONET. O AOT/AERONET {\'e}
calculado como a m{\'e}dia temporal em uma hora, centrada no
tempo da passagem do sat{\'e}lite. A concord{\^a}ncia entre o
sat{\'e}lite e AERONET {\'e} quantificada aplicando as seguintes
m{\'e}tricas estat{\'{\i}}sticas: R2, desvio padr{\~a}o e
calculando a curva de ajuste como y=ax+b. O par{\^a}metro 'a' foi
calculado para cada esta{\c{c}}{\~a}o usando AOT/AERONET x
AOT/MODIS e AOT/AERONET x AOT/VIIRS. Se |1-a|>0.5, para
compara{\c{c}}{\~a}o com o AOT/MODIS, a esta{\c{c}}{\~a}o foi
classificada como suspeita. Os resultados mostram grande
dispers{\~a}o do AOT/VIIRS, sendo que o R2 calculado para todas
esta{\c{c}}{\~o}es foi de 0.20 e aumentou para 0.38 quando as
esta{\c{c}}{\~o}es suspeitas foram negligenciadas. Significantes
mudan{\c{c}}as ocorreram para o coeficiente angular 'a', o qual
foi 0.55 para todas as esta{\c{c}}{\~o}es e aumentou para 0.86
negligenciando as esta{\c{c}}{\~o}es suspeitas. A mesma analise
foi realizada para AOT/MODIS, onde R2 foi 0.59 para todas as
esta{\c{c}}{\~o}es e 0.83 quando as esta{\c{c}}{\~o}es
suspeitas foram removidas. Na an{\'a}lise por esta{\c{c}}{\~a}o
do AOT/S-NPP, sem as esta{\c{c}}{\~o}es suspeitas, foram obtidos
R2=0.48 e a=0.82 para Cuiaba-Miranda e R2=0.31 e a=0.65 para
Itajub{\'a}. ABSTRACT: The effort to retrieve Aerosol Optical
Thickness (AOT) has increased along of years and is associated to
the need understand interaction with radiation and to improve
numerical models. Aerosols act as radiation and cloud controller.
They can be classified as natural (e.g. dust, marine salt) or
anthropogenic (e.g. pollution, soot). In Brazil, most aerosol
source is from biomass burning. Due to high spatial and temporal
variability, satellite information is one the main tools to
retrieve AOT value in global scale. In continuity of remote Earth
Observation System program, the Suomi National Polar-orbiting
Partnership (S-NPP) was launched in 2011, being the first of new
generation satellites from National Polar-orbiting Operational
Environmental Satellite System. S-NPP carry five instruments, one
of them is Visible Infrared Imaging Radiometer Suite (VIIRS) with
twenty-two bands From this twenty-two bands nine are used to
derive AOT. Since 2013, DSA/CPTEC/INPE has been processing AOT
from VIIRS/NPP enabling to monitor aerosol over Brazil. This work
aim is to present the validation of the AOT/S-NPP Intermediate
Product by comparing satellite data with ground-based data from
nine AERONET stations. The analyses are performed on daily bases
from January to December 2014. The AERONET stations are spread out
over Brazil where one station is situated in Northeast, Southeast
and South regions and three are placed in North and West-center
regions. Additionally, the AOT/MODIS from Aqua Satellite is used
in the comparison since it is a well established product working
for more than 10 years. The analysis methodology is based in
Ichoku et al.,(2002) which satellite AOT value is the spatial mean
data centered box (40 × 40km) in AERONET station. AERONET AOT data
value is calculated as temporal average of one hour data centered
in satellite passage time. The agreement between satellite and
AERONET data was quantified applying the following statistics
metrics: R2, standard deviation and calculating of linear fitting
curve y=ax+b. For each station angular coefficient 'a' was
calculated using AOT/AERONET x AOT/MODIS and AOT/AERONET x
AOT/VIIRS. If |1-a|>0.5, for AOT/MODIS comparison, the station was
classified like suspected. Results showed larger AOT/VIIRS
dispersion. R2 calculated for all stations was 0.20 and increased
to 0.38 when neglecting suspected stations. A significant changes
occurred for angular coefficient 'a'. The 'a' values founded for
all stations was 0.55 and it increases to 0.86 neglecting
suspected stations. The same analysis was made using AOT/MODIS,
where R2 was 0.59 for all station while was 0.83 when suspected
was removed. These results shows that AOT/S-NPP has similar
performance than AOT/MODIS. Analysis for AOT/S-NPP by station,
without suspected group, gave R2 = 0.48 and a = 0.82 to
Cuiaba-Miranda and R2 = 0.31 and a = 0.65 to Itajub{\'a}.",
conference-location = "Natal, RN",
conference-year = "13-16 out.",
urlaccessdate = "27 abr. 2024"
}