@PhDThesis{Gouveia:2024:QuIn,
author = "Gouveia, Carolina Daniel",
title = "Avalia{\c{c}}{\~a}o de extremos clim{\'a}ticos na Am{\'e}rica
do Sul para v{\'a}rios n{\'{\i}}veis de aquecimento global
gerados pelos modelos do HELIX: quantifica{\c{c}}{\~a}o de
incertezas",
school = "Instituto Nacional de Pesquisas Espaciais (INPE)",
year = "2024",
address = "S{\~a}o Jos{\'e} dos Campos",
month = "2023-11-17",
keywords = "mudan{\c{c}}as clim{\'a}ticas, proje{\c{c}}{\~o}es,
incertezas, n{\'{\i}}veis de aquecimento espec{\'{\i}}fico,
climate changes, projections, uncertainties, specific warming
levels.",
abstract = "As mudan{\c{c}}as no clima s{\~a}o evidentes e as
proje{\c{c}}{\~o}es indicam aumento nos valores m{\'e}dios e
nos extremos clim{\'a}ticos durante o s{\'e}culo XXI.
Regi{\~o}es como a Am{\'e}rica do Sul (AS) ser{\~a}o as mais
afetadas pelas mudan{\c{c}}as, principalmente relacionadas aos
extremos, uma vez que s{\~a}o marginalizadas e desestruturadas
para suportar essas mudan{\c{c}}as. As proje{\c{c}}{\~o}es
provenientes de modelos, por mais que possuam grades refinadas e
boa representa{\c{c}}{\~a}o dos processos que comp{\~o}em o
Sistema Terrestre, possuem incertezas associadas, fazendo-se
necess{\'a}rio conhec{\^e}las. Assim, este trabalho tem como
objetivo analisar as proje{\c{c}}{\~o}es de extremos
clim{\'a}ticos dos modelos pertencentes ao High-End cLimate
Impacts and eXtremes (HELIX) para a Am{\'e}rica do Sul nos
n{\'{\i}}veis de aquecimento espec{\'{\i}}fico (SWLs) de 1,5
°C, 2 °C e 4 °C, quantificando as incertezas e impactos
associados. A avalia{\c{c}}{\~a}o dos modelos mostra que o
ensemble ECEARTH3- HR superestima dias secos consecutivos (CDD) na
regi{\~a}o central do Brasil e mostra vieses positivos sobre os
Andes para dias de precipita{\c{c}}{\~a}o intensa acima de 10 mm
(R10mm), dias muito {\'u}midos (R95p) e m{\'a}ximo de
precipita{\c{c}}{\~a}o em 5 dias {{(Rx5Day);}} enquanto as
maiores subestimativas est{\~a}o sobre a regi{\~a}o equatorial
para os extremos de precipita{\c{c}}{\~a}o (exceto CDD). Para o
ensemble HadGEM3, as subestimativas tamb{\'e}m est{\~a}o sobre a
faixa equatorial e Amaz{\^o}nia (AMZ) para os mesmos
{\'{\i}}ndices, e sobre o semi{\'a}rido nordestino para
{{CDD;}} j{\'a} as superestimativas de R95p e Rx5Day s{\~a}o
observadas sobre o Sul/Sudeste (S/SEB) e pa{\'{\i}}ses
adjacentes. Em rela{\c{c}}{\~a}o aos {\'{\i}}ndices de
temperatura, as rean{\'a}lises do ERA-5 e do MERRA-2 s{\~a}o
melhor representadas pelo EC-EARTH3-HR em todo o territ{\'o}rio
(exceto AMZ), sendo os maiores vieses para HadGEM3. A habilidade
dos modelos em simular as observa{\c{c}}{\~o}es e
rean{\'a}lises foi obtida atrav{\'e}s do Diagrama de Taylor e
mostra maiores acur{\'a}cias das rodadas do EC_EARTH3-HR em
representar todos os {\'{\i}}ndices, exceto CDD e noites muito
frias (TN10p). Os {\'{\i}}ndices de temperatura t{\^e}m os
menores vieses (n{\~a}o-significativos), apresentando desempenhos
mais satisfat{\'o}rios em simular o presente quando comparados
aos {\'{\i}}ndices de precipita{\c{c}}{\~a}o (vieses
significativos). Em rela{\c{c}}{\~a}o {\`a}s
proje{\c{c}}{\~o}es para o s{\'e}culo XXI, os resultados
mostram mudan{\c{c}}as mais intensas e abrangentes para HadGEM3
em compara{\c{c}}{\~a}o ao ECEARTH3- HR. Para extremos de
precipita{\c{c}}{\~a}o, as proje{\c{c}}{\~o}es evidenciam
escassez de chuvas no Nordeste (NEB) e leste da AMZ, aumento de
CDD no central do Brasil e Andes, e eleva{\c{c}}{\~a}o das
chuvas sobre o oeste da AMZ, Sul do Brasil e norte da Argentina.
Por{\'e}m, tais mudan{\c{c}}as s{\~a}o n{\~a}o-significativas
e abrangem quase a totalidade da {\'a}rea nos 3 SWLs. J{\'a}
para os extremos de temperatura, as proje{\c{c}}{\~o}es
evidenciam aumento dos extremos quentes e diminui{\c{c}}{\~a}o
dos extremos frios, em particular sobre a regi{\~a}o AMZ. Tais
mudan{\c{c}}as sofrer{\~a}o intensifica{\c{c}}{\~a}o e
expans{\~a}o na dire{\c{c}}{\~a}o sul da AS com o aumento de
SWL, sendo significativamente robustas para todos os
{\'{\i}}ndices (exceto dura{\c{c}}{\~a}o de ondas de calor -
WSDI) quando aquecido 4 °C. Adicionalmente, as incertezas
apresentaram maior variabilidade interna de CDD sobre o Sudeste,
Centro-Oeste e litoral norte do NEB, e no S/SEB e Centro-Oeste
para {\'{\i}}ndices {{{\'u}midos;}} enquanto a variabilidade
dos modelos aumenta com o n{\'{\i}}vel de aquecimento, se
tornando dominante no total de incertezas em SWL 4.0 (90% do
total). Para os extremos de temperatura, ambas as variabilidades
s{\~a}o mais pronunciadas na regi{\~a}o tropical e menos sobre a
Argentina, sendo as incertezas dos modelos tamb{\'e}m
predominantes no total, por{\'e}m essa variabilidade diminui com
SWL (exceto para dias muito quentes - TX90p - e WSDI). Por fim, a
rela{\c{c}}{\~a}o Sinal-Ru{\'{\i}}do indica alta
confiabilidade das proje{\c{c}}{\~o}es de dias de
precipita{\c{c}}{\~a}o intensa acima de 30 mm (R30mm) para
grande parte da AS, enquanto h{\'a} alta confiabilidade de
diminui{\c{c}}{\~a}o de TN10p e dias muito frios (TX10p) sobre
toda AS, e aumento de noites muito quentes (TN90p) para a faixa
central e restante do continente, quando o incremento for de 4 °C,
evidenciando assim a import{\^a}ncia de limitar o aquecimento
global em 2 °C. ABSTRACT: Climate changes are evident and
projections indicate an increase in average values and climate
extremes during the 21st century. Regions such as South America
(SA) will be the most affected by changes, mainly related to
extremes, since they are marginalized and unstructured to support
these changes. Projections from models, even though they have
refined grids and good representation of the processes that make
up the Earth System, have associated uncertainties, making it
necessary to know them. Therefore, this work aims to analyze the
projections of climate extremes from models belonging to the
High-End cLimate Impacts and eXtremes (HELIX) for South America at
specific warming levels (SWLs) of 1.5 °C, 2 °C and 4 °C,
quantifying the associated uncertainties and impacts. The model
evaluation shows that the ECEARTH3- HR ensemble overestimates
consecutive dry days (CDD) in the central region of Brazil and
shows positive biases over the Andes for days with intense
precipitation above 10 mm (R10mm), very humid days (R95p ) and
maximum precipitation in 5 days {{(Rx5Day);}} while the largest
underestimates are over the equatorial region for precipitation
extremes (except CDD). For the HadGEM3 ensemble, the
underestimates are also over the equatorial belt and the Amazon
(AMZ) for the same indices, and over the northeastern semi-arid
region for {{CDD;}} overestimates of R95p and Rx5Day are observed
over the South/Southeast (S/SEB) and adjacent countries. Regarding
temperature indices, the ERA-5 and MERRA-2 reanalyses are best
represented by EC-EARTH3-HR across the entire territory (except
AMZ), with the largest biases being for HadGEM3. The ability of
the models to simulate observations and reanalysis was obtained
through the Taylor Diagram and shows greater accuracies of the
EC-EARTH3-HR runs in representing all indices, except CDD and very
cold nights (TN10p). Temperature indices have the smallest biases
(non-significant), presenting more satisfactory performances in
simulating the present when compared to precipitation indices
(significant biases). Regarding projections for the 21st century,
the results show more intense and comprehensive changes for
HadGEM3 compared to EC-EARTH3-HR. For precipitation extremes,
projections show a lack of rainfall in the Northeast (NEB) and
east of the AMZ, an increase in CDD in central Brazil and the
Andes, and an increase in rainfall over the west of the AMZ,
southern Brazil and northern Argentina. However, such changes are
non-significant and cover almost the entire area in the 3 SWLs. As
for temperature extremes, projections show an increase in hot
extremes and a decrease in cold extremes, particularly over the
AMZ region. Such changes will intensify and expand towards the
south of SA with the increase in SWL, being significantly robust
for all indices (except heat wave duration - WSDI) when heated by
4 °C. Additionally, the uncertainties presented greater internal
variability of CDD over the Southeast, Central-West and northern
coast of the NEB, and in the S/SEB and Central-West for humid
{{indices;}} while the variability of the models increases with
the level of warming, becoming dominant in the total uncertainties
in SWL 4.0 (90% of the total). For temperature extremes, both
variabilities are more pronounced in the tropical region and less
over Argentina, with model uncertainties also predominant in
total, but this variability decreases with SWL (except for very
hot days - TX90p - and WSDI). Finally, the Signal-to-Noise ratio
indicates high reliability of projections of days with intense
precipitation above 30 mm (R30mm) for a large part of SA, while
there is high reliability of a decrease in TN10p and very cold
days (TX10p) over all of SA, and an increase in very hot nights
(TN90p) for the central region and the rest of the continent, when
the increase is 4 °C, thus highlighting the importance of limiting
global warming to 2 °C.",
committee = "Chan, Chou Sin (presidente) and Orsini, Jos{\'e} Antonio Marengo
(orientador) and Fisch, Gilberto Fernando and Calheiros,
S{\^a}mia Regina Garcia and Ambrizzi, T{\'e}rcio",
englishtitle = "Evaluation of climate extremes in South America for various global
warming levels generated by HELIX models: quantification of
uncertainty",
language = "pt",
pages = "133",
ibi = "8JMKD3MGP3W34T/4A9TNBP",
url = "http://urlib.net/ibi/8JMKD3MGP3W34T/4A9TNBP",
targetfile = "publicacao.pdf",
urlaccessdate = "03 jun. 2024"
}