@MastersThesis{Gurgel:2000:VaEsTe,
author = "Gurgel, Helen da Costa",
title = "Variabilidade espacial e temporal do NDVI sobre o Brasil e suas
conex{\~o}es com o clima",
school = "Instituto Nacional de Pesquisas Espaciais (INPE)",
year = "2000",
address = "S{\~a}o Jos{\'e} dos Campos",
month = "2000-03-03",
keywords = "{\'{\i}}ndice de vegeta{\c{c}}{\~a}o por diferen{\c{c}}a
normalizada (NDVI), clima, cobertura vegetal, an{\'a}lise por
principais componentes, AVHRR, normalized difference vegetation
index (NDVI), climate, vegetation cover, principal components,
AVHRR.",
abstract = "Este trabalho avalia as conex{\~o}es entre o {\'{\I}}ndice de
Vegeta{\c{c}}{\~a}o por Diferen{\c{c}}a Normalizada (NDVI) e as
variabilidades clim{\'a}ticas anual e interanual sobre o Brasil.
As an{\'a}lises foram feitas utilizando-se NDVI e
Radia{\c{c}}{\~a}o de Ondas Longas (ROL), derivados do National
Oceanic and Atmospheric Administration - Advanced Very High
Resolution Radiometer (NOAA-AVHRR). O estudo foi baseado no
comportamento estat{\'{\i}}stico destes dados para o
per{\'{\i}}odo de janeiro de 1982 a dezembro de 1993. Foram
utilizadas as t{\'e}cnicas de estat{\'{\i}}stica multivariada
de an{\'a}lise por componentes principais (ACP) e an{\'a}lise de
agrupamento, bem como m{\'e}dias mensais e sazonais e anomalias.
A ACP foi aplicada nos dados de ROL e de NDVI; j{\'a} a
an{\'a}lise de agrupamento foi feita apenas nos dados de NDVI. Os
resultados obtidos mostram que as m{\'e}dias e anomalias de ROL,
assim como a ACP, possibilitaram a caracteriza{\c{c}}{\~a}o
clim{\'a}tica espacial e temporal dos principais sistemas de
tempo que atuam sobre o Brasil. Em rela{\c{c}}{\~a}o {\`a} ACP,
observou-se que a primeira componente representa o padr{\~a}o
m{\'e}dio da atividade convectiva; a segunda e quarta componentes
est{\~a}o associadas aos modos ver{\~a}o/inverno e
primavera/outono do ciclo anual; a terceira componente mostra as
varia{\c{c}}{\~o}es moduladas pelos eventos de El Niņo, e; a
quinta componente est{\'a} associada ao ciclo semi-anual. A
an{\'a}lise de agrupamento realizada nos dados de NDVI,
identificou sete grandes grupos de vegeta{\c{c}}{\~a}o que
ocorrem no Brasil e revelou a variabilidade anual e interanual e o
tempo de resposta a precipita{\c{c}}{\~a}o destes grupos. A ACP
aplicada ao NDVI demonstrou a potencialidade desta vari{\'a}vel
em caracterizar as grandes fei{\c{c}}{\~o}es da
vegeta{\c{c}}{\~a}o que predominam no territ{\'o}rio
brasileiro. Al{\'e}m disso, caracterizou-se a variabilidade anual
e interanual destas tipologias de vegeta{\c{c}}{\~a}o
relacionadas com a variabilidade clim{\'a}tica (componentes 1, 2
e 3). A ACP tamb{\'e}m revelou os principais fatores {"}n{\~a}o
clim{\'a}ticos{"}, tais como ocorr{\^e}ncia de queimadas,
mudan{\c{c}}a de hor{\'a}rio da passagem dos sat{\'e}lites NOAA
e trocas dos sat{\'e}lites NOAA 9 e 11 (componentes 4, 5, 6 e
10), que ocasionam dist{\'u}rbios no NDVI. Os altos valores de
correla{\c{c}}{\~o}es obtidos entre os autovetores, resultantes
da ACP da ROL e do NDVI, que correspondem aos modos
ver{\~a}o/inverno (r = 0,91) e primavera/outono (r = 0,70),
indicam que a variabilidade anual da cobertura vegetal brasileira
{\'e} modulada em grande parte pelo regime clim{\'a}tico. Os
resultados encontrados tamb{\'e}m sugerem que as t{\'e}cnicas
estat{\'{\i}}sticas utilizadas neste trabalho, mostraram-se ser,
interessantes ferramentas para caracterizar eventos de
mudan{\c{c}}as peri{\'o}dicas ou aperi{\'o}dicas em dados de
s{\'e}ries temporais obtidos atrav{\'e}s de sensoriamento
remoto. ABSTRACT: This work studies the connections between
Normalized Difference Vegetation Index (NDVI) and annual and
interanual climatic variabilities over Brazil. For this purpose,
NDVI and Outgoing Longwave Radiation (OLR) data, obtained from
National Oceanic and Atmospheric Administration - Advanced Very
High Resolution Radiometer (NOAA-AVHRR) between January/82 and
December/93, were analyzed. To accomplish this study, Cluster and
Principal Component Analysis (PCA) were the statistical tools
used, as well as monthly and seasonal means, and annual anomalies.
PCA was done at both NDVI and OLR data, in contrast to Cluster
analysis, which was applied only to NDVI data. The obtained
results point that the means and anomalies at OLR data and the
PCA, allow a climatic characterization in space and time of the
main climatic systems that act on Brazil. In relation to the PCA,
it was observed that the first component represents the mean
pattern of convective activity; second and fourth components are
associated to the summer/winter and spring/autumn modes of the
annual cycle; the third component shows the variations modulated
by El Niņo's events, and the fifth component is associated with
the semi-annual cycle. The cluster analysis of NDVI data, sorted
seven great vegetation groups that appear in Brazil and showed the
annual and interanual variation, as well as these groups' response
time to precipitation. The PCA of NDVI demonstrated the
potentiality of this variable in characterizing the main features
of vegetation that prevail in the Brazilian territory. In
addition, annual and interannual variability of these vegetation
groups related to the climatic variability (components 1, 2 and
3), was characterized. PCA also revealed the main {"}non
climatic{"} factors that disturbs NDVI, which are: the occurrence
of burns, NOAA?s satellites period delays and the changes of
NOAA?s satellites 9 and 11 (components 4, 5, 6 and 10). The high
values of correlations obtained among the eigenvectors, resultants
from PCA of OLR and NDVI data, that correspond to summer/winter (r
= 0.91) and spring/autumn (r = 0.70) modes, indicate that the
annual variability of Brazilian vegetation cover is largely
modulated by the climatic regime. The obtained results suggest
that the statistical techniques used in this work, proved to be
useful tools to characterize events of periodic and non-periodic
changes in temporal series data collected by remote sensing.",
committee = "Ferreira, Nelson Jesus (orientador/presidente) and Freitas, Corina
da Costa and Batista, Get{\'u}lio Teixeira and Shimabukuro, Yosio
Edemir and Dias, Pedro Leite da Silva",
copyholder = "SID/SCD",
englishtitle = "Spatial and temporal variability of NDVI over Brazil and its
connections with the climate.",
language = "pt",
pages = "118",
ibi = "6qtX3pFwXQZ3P8SECKy/y5cy7",
url = "http://urlib.net/ibi/6qtX3pFwXQZ3P8SECKy/y5cy7",
targetfile = "publicacao.pdf",
urlaccessdate = "15 jun. 2024"
}