Mak Sic 2022

Palaeogeography, Palaeoclimatology, Palaeoecology 585 (2022) 110717
Contents lists available at ScienceDirect
Palaeogeography, Palaeoclimatology, Palaeoecology

journal homepage: www.elsevier.com/locate/palaeo
Brazilian biomes distribution: Past and future

J. Maksic a, *, I.M. Venancio b, M.H. Shimizu b, C.M. Chiessi c, P. Piacsek d, G. Sampaio b,
Francisco W. Cruz e, F.F. Alexandre b
a
Division for Impacts, Adaptation and Vulnerabilities, National Institute for Space Research (INPE), São José dos Campos, São Paulo, Brazil
b
Center for Weather Forecasting and Climate Studies (CPTEC), National Institute for Space Research (INPE), Cachoeira Paulista, Brazil
c
School of Arts, Sciences and Humanities, University of São Paulo, São Paulo, Brazil
d
Post-Graduate Program in Geosciences (Environmental Geochemistry), Fluminense Federal University, Niteroi, Brazil
e
Geosciences Institute, University of São Paulo, São Paulo, Brazil
A R T I C L E I N F O A B S T R A C T
Editor: Dr. Paul Hesse The Last Glacial Maximum (LGM, 26.5–19 ka) was marked by atmospheric cooling, in contrast to the current
warming climate, which will probably continue in the coming decades, according to climate models projections.
Keywords: The LGM to pre-industrial transition provides an opportunity to test the vegetation response to a very large
Last Glacial Maximum temperature change that can then be applied to project pre-industrial to end-of-century changes. In order to
Simulations
explore the changes in Brazilian biomes due to temperature change, we projected potential vegetation for both
Biomes
past and future scenarios. We compared biome projections with a compilation of 149 published LGM re
Future scenario
Palaeorecords constructions of climate and vegetation within Brazil and adjacent areas. In addition, we evaluated the particular
effects that changes in precipitation, temperature and CO2 had on vegetation by performing sensitivity exper
iments. Our results suggest that biomes in the western and central portions of the Amazon forest remained largely
unchanged during the LGM mainly due to negative temperature anomalies, while a decrease in past precipitation
was responsible for the shift from tropical evergreen forest to tropical seasonal forest in the eastern portion of the
Amazon. These results are consistent with proxy reconstructions. LGM model projections and proxy re
constructions suggest expansion of grassland in the southern Brazilian highlands. Under future warming sce
narios, biome changes are mostly forced by decreasing precipitation and increasing temperatures, which
counteract potential biomass gain from the positive CO2 fertilization effect. Under future warming, our simu
lations show an expansion of Savanna/Cerrado and a reduction of tropical seasonal forest and Caatinga, with
potential large impacts over biodiversity and regional climate.
1. Introduction particularly vulnerable biome to climate change (Seddon et al., 2016). In

the case of the Amazon, a potential replacement of the forest by other
Global warming is expected to affect mean precipitation as well as its types of drought-resilient vegetation (i.e., tropical savanna) has been
seasonality, the length of dry and wet seasons, the frequency and previously hypothesized (Nobre et al., 1991; Oyama and Nobre, 2003).
magnitude of extreme events in different regions of South America The consequences of one such biome change would not only have a
(Duffy et al., 2015; Hoegh-Guldberg et al., 2018). These changes in negative effect on biodiversity but would also affect the climate of other
hydroclimate, together with increases in global temperature and carbon South American regions, potentially affecting global climate (Lenton
dioxide (CO2) concentrations will certainly affect vegetation distribu et al., 2008; Lovejoy and Nobre, 2019). Simulations of past climate and
tion and structure, possibly disrupting ecosystem services and biome distribution, validated by palaeoclimate and palaeoecological
decreasing biodiversity (Root et al., 2003; Gonzalez-Orozco et al., 2016; records, provide unique insights into biome responses to climate change
Aguirre-Gutiérrez et al., 2020). South America presents unique ecosys and improve the understanding of future biome projections (Braconnot
tems with enormous biodiversity and differing vulnerability to future et al., 2012; Nolan et al., 2018).
climate change (Magrin et al., 2014). In Brazil, besides the Amazon During the Last Glacial Maximum (LGM, 26.5–19 ka) (Clark et al.,
forest (Trisos et al., 2020), the caatinga has been identified as a 2009), global climate was significantly different from modern (i.e., pre-
* Corresponding author.
E-mail address: [email protected] (J. Maksic).
https://doi.org/10.1016/j.palaeo.2021.110717
Received 4 June 2021; Received in revised form 24 September 2021; Accepted 13 October 2021
Available online 19 October 2021
0031-0182/© 2021 Elsevier B.V. All rights reserved.
J. Maksic et al. Palaeogeography, Palaeoclimatology, Palaeoecology 585 (2022) 110717
industrial) climate. Since the LGM, the atmosphere warmed and given section were used to force the Center for Weather Forecasting and
climate models projections, warming will certainly continue in the Climate Studies Potential Vegetation Model version 2 (CPTEC-PVM2;
coming decades. The global mean warming since the LGM is of a similar Lapola et al., 2009) and produce equilibrium vegetation projections.
magnitude to the upper range of Coupled Model Intercomparison Project This model can reproduce the main South American biomes like tropical
phase 5 (CMIP5) projections for the end of the twenty-first century forests over Amazonia and the Atlantic coastal region, savannas over
(Collins et al., 2013). This provides the rare opportunity to test the central Brazil (‘cerrado’), dry shrublands (‘caatinga’) over northeastern
vegetation response to a temperature increase of approximately 4 ◦ C Brazil and the Chaco region, grasslands over the Pampas, and semi-
from the LGM to the pre-industrial, and then use this knowledge to desert vegetation over Patagonia (Olson et al., 2001). Model resolu
project vegetation response on an additional 4 ◦ C of potential future tion in this study is 1◦ lat × 1◦ lon. To perform the potential vegetation
warming. Hereafter, the LGM will also be referred to as the “− 4 ◦ C" simulations for the − 4 ◦ C and +4 ◦ C scenarios as well as sensitivity
scenario and the end-of-century scenario will also be referred to as experiments for each scenario, we used the climatology from CMIP5
“+4 ◦ C". Here we simulated the potential vegetation in Brazil for both models described in the previous section. To filter out the effect of the
scenarios in order to verify the most affected biomes and also explore the systematic errors from the different models (Mueller and Seneviratne,
extent of biome changes with temperature increase in each scenario. We 2014), the CPTEC-PVM2 model was initialized with the 1961–1990
also ran sensitivity experiments for both scenarios, where we isolated observed climatology (Willmott and Matsuura, 2001) and then we
the influence of changes in CO2, precipitation and surface temperature added the anomalies (i.e., experiment minus present) following the
over vegetation. The sensitivity experiments allowed us to determine the anomaly coupling procedure (Kutzbach et al., 1998) for each scenario.
main drivers behind vegetation changes in both scenarios. In addition, Both single model and multi-model ensemble mean climatologies are
we compared our biome projections for the past with a compilation of considered, but as the multi-model ensemble means of present-day
149 published reconstructions of climate and vegetation within Brazil climate agrees better with observations than any single model
and adjacent areas. Our results show that changes in monsoon intensity (Tebaldi and Knutti, 2007), and in order to decrease the range of climate
during the LGM cannot be used as the main driver for vegetational models uncertainty, we based our final analyses of biome projections
changes/stability across the Amazon biome, while for the future +4 ◦ C from the multi-model ensemble mean anomalies input. Regarding the
scenario biome shifts will mostly depend on precipitation changes. concentration of CO2, the − 4 ◦ C and the +4 ◦ C scenarios used 180 ppm
and 900 ppm, respectively (Kageyama et al., 2017) (Table 1). To
2. Data and methods determine the influence of each climate parameter on biome distribu
tion, we ran sensitivity experiments for both − 4 ◦ C and +4 ◦ C scenarios,
In this section, we present a description of the model data selected for considering the anomalies of CO2, precipitation and temperature
the temperature and precipitation analyses, as well as the methods used separately.
for vegetation projections and compilation of proxy data from paleo To perform the potential vegetation simulations for the Present
climate archives used for model validation. scenario, the CPTEC-PVM2 model was initialized and calibrated with
the 1961–1990 observed climatology of Willmott and Matsuura (2001)
2.1. Past and future climate simulations (hereafter WM), that is: (precWM + tempWM) with CO2 = 350 ppm. For
the potential vegetation simulations of the − 4 ◦ C and + 4 ◦ C scenarios,
With the aim of analyzing the maximum number of models compiled the CPTEC-PVM2 model was initialized with the observed climatology
in the Coupled Model Intercomparison Project phase 5 (CMIP5)/Palae plus the field of multi-model ensemble mean anomalies (MME) for each
oclimate Modeling Intercomparison Project phase 3 (PMIP3) (Taylor scenario: (precWM + MME− 4) + (tempWM + MME− 4) with CO2 = 180
et al., 2012), we used all the models that were available at the moment ppm for the LGM, or (precWM + MME+4) + (tempWM + MME+4) with
of analysis at the British Atmospheric Data Centre (BADC) and had both CO2 = 900 ppm for the future. In the case of the sensitivity experiments
past and future scenario outputs. Climate scenarios for both past and only one parameter was changed. For example, in performing sensitivity
future were extracted from the following models: CCSM4, CNRM-CM5, simulations for the LGM precipitation the setup was (precWM + MME-
FGOALS-g2, GISS-E2-R, IPSL-CM5A-LR, MIROC-ESM, MPI-ESM-P, and 4) + (tempWM) with CO2 = 350 ppm; or, in performing sensitivity
MRI-CGCM3. Models references and a brief description of these exper simulations for the LGM CO2 the setup was (precWM+ tempWM) with
iments are given in the Supplementary information Table S1. As models CO2 = 180 ppm. How biome projections and sensitivity experiments
differ in type, configuration, parameterization and resolution, pro were performed in each case are given in Table 2.
jections for some regions differ significantly among them (Knutti and
Sedláček, 2013). In order to diminish those differences, the model data
2.3. Compilation of proxy data from palaeoclimate archives
were converted to the same spatial resolution, and the multi-model
ensemble mean was calculated. To evaluate nearly opposite scenarios,
In this study we updated the compilation of 107 published
we used the LGM and RCP8.5 CMIP5/PMIP3 experiments and analyzed
the surface air temperature and precipitation in comparison with the
Historical experiment, both single model and multi-model ensemble Table 1
mean. For the − 4 ◦ C scenario we used the LGM experiment and calcu Inputs used for the Center for Weather Forecasting and Climate Studies Potential
Vegetation Model version 2 (CPTEC-PVM2) experiments.
lated the 30-year average monthly mean for temperature and precipi
tation. For the +4 ◦ C scenario we used the RCP8.5 experiment and Experiment CO2 Precipitation Temperature
(ppm)
determined the respective +4 ◦ C period for each model, where the +4 ◦ C
period is defined as the time when the global mean temperature reaches Present 350 Willmott and Matsuura Willmott and Matsuura
+4 ◦ C compared to the Historical run (1850–2005) (as in Vautard et al., (2001) (WM) (2001) (WM)
− 4 ◦ C (LGM) 180 WM + MME anomalies WM + MME anomalies
2014). This method was adopted to reduce uncertainty originating from from scenario − 4 ◦ C from scenario − 4 ◦ C
the different climate sensitivity of a model and analyze +4 ◦ C warmer +4 ◦ C 900 WM + MME anomalies WM + MME anomalies
global climate in terms of temperature and precipitation distribution, (Future) from scenario +4 ◦ C from scenario +4 ◦ C
independently of the period of the century when it could occur. Note: For the Present experiment the model was forced with surface temperature
and precipitation from the Willmott and Matsuura (2001)(WM) data set. For the
2.2. Past and future biomes distribution − 4 ◦ C scenario (representing the Last Glacial Maximum (LGM)) and the +4 ◦ C
scenario (representing +4 ◦ C Future) we added the anomalies from multi-model
The outputs of the climate simulations described in the previous ensemble mean (MME).
2
Table 2
Description of sensitivity experiments performed with the Center for Weather Forecasting and Climate
Studies Potential Vegetation Model version 2 (CPTEC-PVM2).
Sensitivity CO₂ CO₂ CO₂ PREC TEMP PREC TEMP PREC TEMP
180 350 900 from from WM + WM + WM + WM +
experiment ppm ppm ppm WM WM MME MME MME MME
( -4°C) ( -4°C) ( +4°C) ( +4°C)
-4°C (LGM)
CO2
-4°C (LGM)
PREC
-4°C (LGM)
TEMP
Present
+4°C (Future )
CO2
+4°C (Future )
PREC
+4°C (Future )
TEMP
Note: For each sensitivity experiment only fields marked in yellow are considered. Willmott and Matsuura
(2001) data set (WM); PREC = precipitation; TEMP = temperature; MME = multi-model ensemble mean
anomalies.
hydroclimate records from tropical South America made by Zhang et al. precipitation reduction of more than 14 mm/month has been projected
(2016). We considered all their records that cover the LGM and added for the northwestern sector of Amazonia. Multi-model ensemble mean
new ones, published since then, that showed vegetation, hydroclimate anomaly also suggests increased precipitation, higher than 10 mm/
and environmental reconstructions. Our compilation of 149 published month, over the southeastern portion of the South American Monsoon
vegetation and hydroclimate records with their respective references is System region (SAMS) (50◦ W–40◦ W, 10◦ S–28◦ S) (Fig. 1a). For the +4 ◦ C
available in Supplementary Information, Proxy Compilation Table S2. scenario a precipitation decrease of ~14 mm/month is projected for the
Original chronologies of all palaeorecords were used. To evaluate the eastern equatorial Brazil (i.e., the region around the Marajó island).
dating quality of the compiled records, we applied a chronological Central Brazil shows a weak positive anomaly of 2 to 4 mm/month,
reliability index (CRI). The CRI is based on age model properties and while precipitation increase (~20 mm/month) is the highest over
sampling resolution of each record, where higher CRI values indicate southern Brazil, below 20◦ S (Fig. 1b). Temperature projections for both
more reliable palaeorecords (Prado et al., 2013; Zhang et al., 2016). This scenarios are more uniform, with more than 70% of agreement among
index simply involves the computation of an arithmetic mean, where the models over the entire Brazil (Fig. 1c, d). Temperature anomalies show
same weight is given to calibration (C), sampling resolution of each generally higher values inland than in the coastal regions, with regional
record (R) and dating (D): variations of approximately 2 ◦ C. Single model anomalies can be found
A detailed description is given in Supplementary Information. The in Supplementary Information (Fig. S1 and S2).
temporal proximity of the LGM to two Heinrich Stadials (namely
Heinrich Stadial 2 and Heinrich Stadial 1) make the CRI particularly 3.2. Validation of the climate simulations for the − 4 ◦ C scenario using
necessary, since Heinrich Stadials are thought to have substantially palaeoclimate proxies
altered the distribution of Brazilian biomes (e.g., Wang et al., 2004;
Bouimetarhan et al., 2018; Pinaya et al., 2019). To validate our multi-models ensemble mean for the − 4 ◦ C scenario,
In order to compare the results from the compilation with model we compared it with our proxy compilation (Supplementary Informa
simulations, we defined categories for precipitation and vegetation tion, Proxy Compilation Table S2). The simulated temperature reduction
(biome) changes based on the original interpretations of the authors. In is broadly consistent with reconstructions from South America. The
the case of precipitation, anomalies are expressed as the difference be simulated precipitation anomaly, on the other hand, shows some
tween LGM and present time, and the categories are “drier”, “wetter” discrepancy in palaeo-precipitation reconstructions (Supplementary
and “unclear”. Also for vegetation anomalies are expressed as the dif Information, Fig. S1). Higher precipitation simulated along the Andes is
ference between LGM and present time, and the categories are “change”, consistent with lake and ice core records. The simulated anomaly over
“no change” and “unclear”. Although this approach has been used in equatorial South America corroborates the findings from 39 records
other studies, we are aware that this simplification can result in an between 10◦ N and 10◦ S that suggest a drier LGM. Over the SAMS region,
agreement between the proxy record and the model scenario even in simulated precipitation increase is in agreement with 4 high-resolution
case when they show opposite signals, so we specify a description of speleothem-based records. However, 10 other reconstructions, mostly
change and change evidence in the compilation. estimations from biological proxies between 10◦ S and 20◦ S, appear to
contradict the increased activity of the SAMS. Still, the comparison with
3. Results 20 palaeorecords south of 20◦ S shows high consistency and indicates a
reduction in precipitation (for full details about proxies see Supple
3.1. Climate simulations mentary Information, Proxy Compilation Table S2).
Fig. 1 shows the annual mean precipitation and temperature anom 3.3. Simulated biome changes for − 4 ◦ C and +4 ◦ C scenarios
aly fields from the multi-model ensemble calculated for the − 4 ◦ C and
the +4 ◦ C scenarios in comparison with the Historical multi-model Fig. 2 shows CPTEC-PVM2 vegetation equilibrium solutions based on
ensemble. For the − 4 ◦ C scenario our results show a negative precipi long-term mean monthly climate variables for the − 4 ◦ C scenario, Pre
tation anomaly over the tropical region (0◦ –10◦ S) (Fig. 1a). The highest sent (1961–1990) climate, and the +4 ◦ C scenario. The simulation of the
3
Fig. 1. Annual mean precipitation (mm/month) (a-b) and temperature (◦ C) (c-d) fields from the multi-model ensemble mean differences calculated between the
− 4 ◦ C scenario (a, c) and the +4 ◦ C scenario (b, d) in comparison with Historical simulations. Stippling represents areas where at least 70% of the models agree with
the sign of the difference.
− 4 ◦ C scenario indicates a distinct difference in southern Brazil palaeorecords in our compilation for each biome. Their original in
(60◦ W–40◦ W, 20◦ S–35◦ S) in comparison with the Present simulation terpretations of precipitation or vegetation/biome changes are compiled
(Fig. 2a, b) suggesting grassland expansion. In northeastern Brazil, open separately in Supplementary Information (for full details see Supple
shrubland/caatinga is rearranged and its central part is transformed in mentary Information, Proxy Compilation Table S2). 77 out of the 149
semi-desert (Fig. 3a). While western Amazonia persisted as a tropical records in our compilation discuss vegetation/biome changes and we
forest with no substantial changes, eastern Amazonia was transformed identified that 49 out of the 77 match the changes suggested by the
into tropical seasonal forest. Tropical seasonal forests with savanna model. However, while palaeorecords from southern and eastern Ama
patches are also simulated in northern Amazonia (65◦ W–55◦ W, 5◦ N–0◦ ). zonia are in accordance with the model output, four palaeorecords from
The simulation of the +4 ◦ C scenario shows an expansion of savanna/ the core of the Amazon forest contradict model projections.
cerrado, replacing most of the open shrubland/caatinga simulated from
the Present conditions (Fig. 2b, c). The eastern part of the tropical
evergreen forest of Amazonia is replaced with patches of savanna that 3.5. Sensitivity experiments
reach 55◦ W, while the +4 ◦ C scenario conditions favor tropical ever
green forest expansion over southern Brazil. Sensitivity experiments were performed to evaluate the specific in
fluence of precipitation, temperature and CO2 over vegetation. Fig. 4
shows that CO2 and/or precipitation reduction of the − 4 ◦ C scenario are
3.4. Validation of the biome simulations for the − 4 ◦ C scenario using able to induce ample changes in a large part of Amazonia and north
palaeovegetation proxies eastern Brazil. The CO2 reduction turns the tropical evergreen forest into
a less productive tropical seasonal forest and enlarges the occurrence of
Most regions where the CPTEC-PVM2 model projected biome open shrubland/caatinga vegetation (Fig. 4a). Changes in precipitation
changes for the − 4 ◦ C scenario are in high agreement with palae from the − 4 ◦ C scenario drive savanna/cerrado expansion and produce
ovegetation proxies (Fig. 3). We combine as many as possible different great reduction of tropical evergreen forest in the eastern and southern
4
Fig. 2. Distribution of Brazilian biomes projected by the Center for Weather Forecasting and Climate Studies Potential Vegetation Model version 2 (CPTEC-PVM2) for
a) the − 4 ◦ C scenario (representing the Last Glacial Maximum), b) Present (1961–1990) climate, and c) the +4 ◦ C scenario (representing RCP8.5 experiment period
when the 30-year average global mean temperature reaches +4 ◦ C compared to the Historical run). The color scale displays different biomes.
substantially expand their occurrence with elevated CO2. Future pre

cipitation anomalies cause reduction in tropical evergreen forests to the
north of the equator and the occurrence of semi-desert vegetation in
northeastern Brazil. Future temperature increase of 4 ◦ C causes tropical
evergreen forest to be completely replaced by less productive biomes,
while large extensions of savanna/cerrado turns into open shrubland/
caatinga, and the occurrence of semi-desert vegetation over north
eastern Brazil is increased.
4. Discussion
4.1. Biome simulations: from the − 4 ◦ C scenario to the present
Prior to discussing the biome simulations, it is worth noting that

although CMIP5/PMIP3 models improved the representation of physical
processes and resolution if compared to earlier phases, some climatic
processes are still simulated with high uncertainties (e.g., precipitation,
clouds, aerosols, air–sea interactions), and some regional and seasonal
climatology biases persist (Kumar et al., 2014; Wang et al., 2014; Adam
et al., 2018; Shimizu et al., 2020). Models have clearly succeeded in
reproducing the main features of past periods, but not all models resolve
the topography and regional-scale features of the South American
climate realistically. As simulated climatologies are used as the input for
vegetation simulations, their validation for past conditions is of extreme
Fig. 3. Differences between biome projection for the − 4 ◦ C (representing the importance.
Last Glacial Maximum) scenario and Present. Grey fields represent areas where Models in general agree about the sign and magnitude of tempera
the Center for Weather Forecasting and Climate Studies Potential Vegetation
ture change for LGM, but due to the complexity in simulating precipi
Model version 2 (CPTEC-PVM2) projection for the − 4 ◦ C scenario and for the
tation and model differences (in special physical parameterizations and
Present differ from each other. The approximate location of compiled vegeta
tion proxy sites is presented by circles. Red(black) circles indicate vegetation
vertical and horizontal resolution), substantially different precipitation
change (no change), brown circles indicate unclear signals. The sizes of the patterns are projected for LGM. Simulations of precipitation have shown
circles are defined by the Chronological Reliability Index (CRI), indicating the even opposite signs, and great regional variations in some regions
reliability of the palaeorecord (the higher the CRI value, the more reliable the (Supplementary Information, Fig. S1). On the proxy side, there is a small
palaeorecord). For full details about proxies see Supplementary Information, number of well dated, long and continuous records from South America
Proxy Compilation Table S2. (For interpretation of the references to color in this (Flantua et al., 2015; Deininger et al., 2019). Most pollen records have
figure legend, the reader is referred to the web version of this article.) interpreted vegetation changes as function of precipitation changes
(Whitney et al., 2011). In some regions, such as central-western Brazil,
Amazonia (Fig. 4b). Past temperature decrease in the -4 ◦ C scenario, in the grassland expansion is interpreted as weaker monsoon precipitation
turn, mostly reduces the occurrence of open shrubland/caatinga, with a during the LGM (Whitney et al., 2011), while wet conditions are pointed
positive effect in the occurrence of tropical evergreen forest (Fig. 4c). out by oxygen isotope speleothem records from the same region (Nov
In the case of the future +4 ◦ C scenario, tropical evergreen forests ello et al., 2017). These discrepancies in interpretations between
5
Fig. 4. Sensitivity simulations for biome distri

bution projected by the Center for Weather
Forecasting and Climate Studies Potential Vege
tation Model version 2 (CPTEC-PVM2) for a-c)
the − 4 ◦ C scenario (representing the Last Glacial
Maximum) and d-f) the +4 ◦ C scenario (repre
senting RCP8.5 experiment period when the
average global mean temperature reaches +4 ◦ C
compared to the Historical run). Anomalies of the
CO2, precipitation and temperature were consid
ered separately for the − 4 ◦ C scenario (upper
panels) and + 4 ◦ C conditions (lower panels). The
color scale displays different biomes. Details
about sensitivity experiments are given in
Table 2.
speleothem and pollen records are possibly associated with a major contradictory evidence also seems to exist (Colinvaux et al., 1996;
control of low atmospheric CO2 and temperatures in the grassland Haberle and Maslin, 1999; Mosblech, 2012; Baker and Fritz, 2015). The
expansion than precipitation during glacial period (Novello et al., 2019; projected stronger SAMS is in agreement with high-resolution speleo
Azevedo et al., 2021). Discrepancies between the results from biological them- and marine-based records (e.g., Cruz Jr et al., 2006; Cruz et al.,
and geochemical proxies in some regions complicate the debate if 2009; Novello et al., 2017; Hou et al., 2020). However, precipitation
tropical South America experienced wet or dry conditions during the reconstructions from biological proxies from the southern border of the
LGM (i.e. − 4 ◦ C scenario). One way to overcome this issue is the use of a Amazon and from southeastern Brazil appear to contradict the increased
multi-model ensemble mean instead of a single model output (Tebaldi activity of the SAMS (e.g., (Salgado-Labouriau et al., 1997; Behling,
and Knutti, 2007), together with different palaeorecords for its 2002; Burbridge et al., 2004; Mayle et al., 2000; Whitney et al., 2011;
validation. Fornace et al., 2016). The reasons for discrepancies between the results
Here we validate computed LGM temperature and precipitation of biological reconstructions on the one side and multi-model ensemble
anomalies with our compilation of palaeorecords covering the LGM mean and geological reconstructions on the other side could lie in
within South America. Based on our review of the existing literature we edaphic constraints (Rossetti et al., 2019). The improved spatial reso
set the referent cooling of LGM at − 4 ◦ C and future warming at +4 ◦ C for lution of climate models or the use of regional models, in addition with
the sake of making a symmetric comparison. The computed temperature better soil representation (Quesada et al., 2010) would probably clarify
reduction showed here (Fig. 1a) is consistent and within the range of these discrepancies. Another source of inconsistency may come from the
several estimates for the LGM over South America (Stute et al., 1995; interpretation of Poaceae pollen abundance increase as an indicator of
Colinvaux et al., 2000; Van Der Hammen and Hooghiemstra, 2000; Wille regional aridity, since these taxa can be influenced by other factors (e.g.,
et al., 2001; Bush and Silman, 2004; Urrego et al., 2005; Chiessi et al., increase in pollen grains related to the extension of periodically flooded
2015; Tierney et al., 2019). environments) (Bush, 2002; da Silva et al., 2020). A noteworthy study of
Our multi-model ensemble mean for the − 4 ◦ C scenario shows a Scheff et al., 2017, calls for caution when paleo hydrologic conditions
negative precipitation anomaly over the equatorial region with the are inferred from plant-based proxies, as changes in vegetation can be
highest precipitation reduction projected for the northwestern part of driven by the low glacial CO2 and temperatures, and not water scarcity,
the Amazon Basin. Validation with proxies, shows that computed which is generally assumed. The influence of low CO2 transcends many
climate coincides with glacial conditions inferred from the paleorecords levels, with physiological effects on individual plants to changes in
available in the region (Fig. S1a). High precipitation along the Andes is ecosystem functioning. The direct effect of low CO2 on plant physio
consistent with cave, lake, marine and ice core records (e.g., Grosjean, logical processes is favoring C4 plants, while C3 plants show an average
1994; Gosling et al., 2008; Niemann and Behling, 2008; Urrego et al., reduction in photosynthesis and biomass production (Gerhart and Ward,
2010; Cheng et al., 2013; Govin et al., 2014). Precipitation lower than 2010; Prentice et al., 2011; Novello et al., 2019; Azevedo et al., 2021).
present over tropical South America to the east of the Andes and north of There is also the possibility of a no-analog climate for large portions of
the Equator are in agreement with palaeorecords (e.g., Absy et al., 1991; South America during the LGM (i.e. extended dry season plus extreme
Häggi et al., 2017; Cruz et al., 2009; Rossetti et al., 2012), but runoff events) as well as no-analog plant communities (Ledru, 2002;
6
Bush and Silman, 2004). Still, a comparison between climate proxies Brazilian highlands, which reveal an expansion of Araucaria forest at the
and multi-model ensemble mean precipitation shows high consistency expense of grassland vegetation in the Late Holocene (Ledru et al., 1996;
over southern Brazil (Fig. S1a). Thus, we conclude that the computed Ledru et al., 2009; Behling et al., 2004; Gu et al., 2017). It is important to
temperature and precipitation anomalies over Brazil are reliable as in highlight that the LGM grassland biome seen in records is reproduced by
puts for the potential vegetation model. our model only when all forcings are taken into account, as none of them
Once the reliability of inputs for the potential vegetation model is separately is sufficient for producing adequate conditions necessary for
confirmed, we evaluate the relative biome distribution (Fig. 5). Relative the grassland biome vegetation (Fig. 2).
biome distribution (%) for the − 4 ◦ C scenario shows that tropical Over northeastern Brazil, with warming, open shrubland/caatinga is
evergreen forest decreased by 10%. This is the result of the trans rearranged and its central part modeled as semi-desert for the − 4 ◦ C
formation of tropical seasonal forest on the eastern portion of the scenario is transformed into open shrubland/caatinga (Fig. 3a). This is in
Amazon Basin, which is consistent with vegetation reconstructions agreement with marine sediment records that suggest vegetation
(Absy et al., 1991; Sifeddine et al., 2001; Behling, 2002; Anhuf et al., changes during the LGM (Behling et al., 2000). Bouimetarhan et al.
2006; Häggi et al., 2017) as well as transformation of grassland and (2018) infer the presence of open shrubland/caatinga vegetation in
tropical seasonal forest in southern Brazil (e.g., Behling et al., 2004; Gu periods of expansion of tropical rainforest and tree ferns during the
et al., 2018). Younger Dryas, indicating the coexistence of a variety of vegetation
The performed sensitivity experiments suggest that with CO2 types in northeastern Brazil and challenging the idea that caatinga is the
reduction, from 350 ppm (Present scenario) to 180 ppm (LGM scenario), result of long-term climatic stability (Werneck et al., 2011). Previously,
there is an almost complete disappearance of tropical evergreen forest (De Oliveira et al., 1999) found abundant cold and humid-adapted
biome. In the photosynthesis model used by CPTEC-PVM2 (Farquhar montane rainforest taxa during the Pleistocene/Holocene transition in
et al., 1980) the response on changes in CO2 concentration is linear, due northeastern Brazil. Montane rainforests expanded and reached lower
to the term applied in photosynthesis formulation which avoids an un levels during LGM, afterwards decreasing gradually until caatinga was
realistic steep decay of photosynthesis rates at low CO2 concentrations established at 4.2 ka BP (Utida et al., 2020). Today, with semi-arid
(Lapola et al., 2009). Nevertheless, halving CO2 in LGM induced a sig climate, rainforest enclaves with different phylogeographic histories
nificant reduction of net primary production and shifted tropical ever are isolated on scattered ancient peaks where the higher elevation in
green forest biome to less productive tropical seasonal forest. creases humidity (Silveira et al., 2019). The difficulty in finding per
Lower LGM temperature counteracts the effect of lower CO2 and manent lakes, bogs or stable locations for soil profiles made vegetational
prevents a more extensive forest replacement by savanna (Fig. 4a, c). It and climate evolution history within the caatinga domain very limited.
is estimated that about 32% of share Amazonian rainfall originates from It is important to remember that local climate, topography and soil
evapotranspiration within the basin, two thirds of which originate as conditions are likely to influence vegetation structure (e.g. changes in
tree transpiration (Staal et al., 2018). Thus, under decreased CO2 and biomass or canopy cover) and dynamics (e.g. mortality and productiv
water vapor in the atmosphere, vegetation water use efficiency is lower ity) within the biomes. It is also worth noting that here we evaluate only
(Beerling and Chaloner, 1993), implying a higher stomatal conductance the LGM time slice, and that millennial fluctuations, before and after this
and an increase in transpiration. That effect could be balanced with period, could have a larger impact on biomes, as no two time slices had
temperature decrease and improved soil moisture (reduced evaporation) identical biome patterns (Pinaya et al., 2019; Allen et al., 2020). Another
(Scheff et al., 2017) and consequently reduced prevalence of fire, which potential limitation is that our model is not dynamic and does not take
is an important element for savanna establishment, promoting the previous vegetation history into account and this could introduce some
presence of tropical evergreen forest (Staver et al., 2011). On the other discrepancies between the model results and some paleovegetation data.
hand, for savanna's C4 grasses, the cerrado biome, which evolved under To summarize, in the context of 4 ◦ C climate warming since the LGM
low CO2 conditions and tolerates droughts and fire, LGM conditions to the present times, the relative distribution of savanna/cerrado and
created an opportunity for advancement into forested areas. Therefore, tropical seasonal forest decreased, tropical evergreen forest and open
alteration in the eastern Amazon could be mainly the result of insuffi shrubland/caatinga increased, while grassland biome virtually dis
cient soil moisture due to LGM precipitation reduction. Our model does appeared. Our results indicate that the absence of evidence for climatic
not consider soil types and variations in soil fertility, which may play an effects on vegetation could be understood as evidence of resilience.
important role in explaining species composition and forest productivity Amazon forest and caatinga showed resilience to temperature increase
across the Amazon Basin (Quesada et al., 2012), besides the climatic of 4 ◦ C since LGM. Whether their resilience will continue with another
east-west gradient presented here. Given that the exact magnitude of step in forcing of the same magnitude or they will be significantly
precipitation change remains to be determined (Baker et al., 2020), it disturbed in the future requires further analysis.
cannot be ignored that some precipitation biases persist in computed
multi-model ensemble means and influence the relative biome 4.2. Biome simulations: from the present to a future +4 ◦ C scenario
distribution.
Hence, our results suggest that the western and central Amazon Future conditions show an expansion of savanna/cerrado, with
forest was largely maintained during the LGM mainly due to negative projected coverage 10% higher than the present, and a decrease of both
temperature anomalies, while a decrease in past precipitation was tropical evergreen forest and tropical seasonal forest (Figs. 2b, c, 5). Our
responsible for biome changes in the eastern portion of the Amazon. Our results suggest that an even greater savanna advancement into forests is
simulations reaffirm that changes in the monsoon intensity cannot be prevented by the “CO2-fertilization” effect (Fig. 4d-f). It is worth noting
used as the main driver for vegetational changes/stability across the that the CO2-fertilization effect, where increased atmospheric CO2 tends
Amazon biome (Gerhart and Ward, 2010; Prentice et al., 2011; Scheff to increase rates of photosynthesis and increase leaf areas, but reduce
et al., 2017) and that lower temperatures in combination with sub evapotranspiration, is the largest and most uncertain of all ecosystem
stantially lower CO2 are most probably the main controlling factors over feedbacks (Arneth et al., 2010; Huntzinger et al., 2017). While some
changes in plant communities here (Mayle et al., 2000). Thus, according observed data are consistent with the projected net effect of elevated
to our simulations, the synergistic effect of temperature, precipitation CO2 (Hubau et al., 2020; Baccini et al., 2017; Lewis et al., 2004), others
and CO2 increase since LGM positively affected the Amazon forest. show no growth stimulation or even point to a decline (Van Der Sleen
Synergy of rising temperature and CO2 also induced advancement of et al., 2015; Brienen et al., 2015; Peñuelas et al., 2011). One possible
forests over grassland in southern Brazil, represented in our simulations explanation is that there is still a very limited number of field and species
with complete disappearance of the grassland biome. This change has observations relative to the large spatial and compositional heteroge
been seen in pollen records and isotopic soil profiles from the southern neity (Ter Steege et al., 2020). Simultaneously, conclusions obtained
7
analyzing satellite vegetation indices, such as leaf area (Forzieri et al., in our study.
2017) or “greening” trends (Zhu et al., 2016), relate observed changes in Although the direct impact of warming on vegetation patterns is not
the Amazon Basin with increased CO2. Zhu et al. (2016) attributed 70% easy to determine, as none of the evaluated driving factors (i.e., tem
of Amazon greening to the 46 ppm increase in CO2 in the period between perature, precipitation and CO2) can be considered in isolation, our
1982 and 2009. However, the long-term persistance and uniformity of results indicate that two completely different biomes, both resilient to
this effect in the future can be modulated by stress responses above the warming since the LGM, will be transformed in the future. In contrast to
optimal temperature (Piao et al., 2020), mortality rates (Esqui the warming of 4 ◦ C since the LGM, when tropical evergreen forests and
vel-Muelbert et al., 2019; Brienen et al., 2020), forest demography caatinga increased, with additional warming those two biomes will be
(Körner et al., 2005; Hubau et al., 2020), mycorrhizal association (Terrer pushed to their thresholds. This is consistent with Ciemer et al. (2019)
et al., 2016) and nutrient limitations (Terrer et al., 2019). Another way which claim higher resilience to climatic disturbances in regions
of looking at the topic of CO2 fertilization is exploring historical global exposed to a higher rainfall variability during their long-term past. Our
gross primary production (Campbell et al., 2017), which is significantly results suggest that changes in the monsoon intensity cannot be used as
higher than current projections, suggesting an underestimate by current the main driver for biomes changes in the past (Figs. 2a, b, 4a-c), but we
models of potential CO2 removal in the future (Haverd et al., 2020). The also find that with additional 4 ◦ C warming, changes and direction of
most recent findings indicate that tropical secondary forests, which are biomes shifts in the eastern Amazon forest and caatinga, will mostly
highly productive and resilient, sequester carbon up to 20 times faster depend on future precipitation changes. At the same time precipitation
than old-growth forests (Heinrich et al., 2021). In addition, potential decrease will induce decrease in biomass production and with that
aboveground biomass and carbon accumulation in naturally regrowing provoke additional local warming, preventing regeneration and adap
forests in the tropics over the next 30 years is estimated to be 54% higher tation (Perugini et al., 2017).
than previously assumed (Cook-Patton et al., 2020). Thus, in spite of the Precipitation changes in the last 30 years already increased the
uncertainties (Walker et al., 2020) considering CO2 fertilization esti abundance of drought-tolerant genera and increased mortality of wet-
mates in humid tropical areas under future climate change scenarios is affiliated trees in the areas where the drying trend was stronger across
important for promoting forest protection, regeneration and restoration the Amazon forest (Esquivel-Muelbert et al., 2019), which indicates that
of degraded areas through low-impact agriculture systems (Nobre and this biome change is already underway. This relatively slow process of
Nobre, 2020), due to potentially major consequences for the global compositional change could continue in the future, and if it happens
carbon cycle. over larger spatial or temporal scales, there is potential for adaptation to
In our future simulation (i.e., +4 ◦ C scenario), although savanna/ changing conditions. However, under future global change, large
cerrado expanded, sensitivity experiments show that with additional changes in abiotic drivers in a short time-frame bring question whether
4 ◦ C even savanna/cerrado is pushed to a critical threshold (Fig. 4f) and ecosystems will have enough time to adapt to new warming conditions
future precipitation changes will be critical in determining the fate of (Etterson and Shaw, 2001; Hoffmann and Sgro, 2011). On the other
this biome (Fig. 4e). The ecotonal regions of Amazonia projected to hand, even if compositional change happens that will lead to lower di
change in the future, had experienced shifts throughout LGM (Fig. 3a, versity, changes in evapotranspiration and forest resilience will likely
b), showing that forest–savanna transitions are dynamic and particularly decrease (Aguirre-Gutiérrez et al., 2020).
sensitive to change. Compositional changes, together with more common droughts and
Our model also suggests that the relative distribution of caatinga will fires, especially in the drier portions of the Amazon, could abruptly shift
be ~50% lower. However, several issues have to be considered before forests to new states (Brando et al., 2014). The spread of forest fires
interpreting this trend. First, with CO2 = 900 ppm in our simulation, the during 21st-century droughts suggests that roads, logging, deforestation
modeled fertilization effect is strong enough to overweight temperature and conversion of rainforests to pastures, make the Amazon forest more
and precipitation impact and shift open shrubland/caatinga vegetation flammable and vulnerable to wildfires (Aragão et al., 2008; Aragão
towards the more productive savanna/cerrado biome (Fig. 4d-f). Pre et al., 2018). The increased susceptibility of forests to fires, through
vious experiments by Lapola et al. (2009) showed that halving the effect fragmentation and edge effects associated with longer recovery time (De
of CO2 fertilization yields small changes in tropical forests, but consid Faria et al., 2021), will cause long-lasting damage or permanent
erable changes and higher permanence of shrubland in the caatinga degradation. The results presented by Brando et al. (2020) indicate that
region. Moreover, Zhu et al. (2016) using three long-term satellite leaf projected climate changes will double the area burned by wildfires,
area index records and ten global ecosystem models showed that for affecting up to 16% of the Amazon's forests by 2050. Contrastingly, fire
northeastern Brazil the key driver in vegetation change is climate and suppression in transitional savanna zones and within cerrado caused
not elevated CO2. Second, high precipitation uncertainties for north woody encroachment, a serious threat to cerrado biodiversity but can
eastern Brazil should also be taken into account (Fig. S1). Our sensitivity enhance carbon sequestration (Rosan et al., 2019). The herein projected
experiments indicate that with joined precipitation reduction and tem expansion of tropical evergreen forest in southern Brazil (Fig. 2b-c), that
perature increase, caatinga vegetation, which is naturally highly could possibly counteract forest loss due to savanna/cerrado expansion
adapted to water deficits (Dombroski et al., 2011; Lima and Rodal, in eastern Amazonia, is already prevented by land use with regional high
2010), could be replaced by semi-desert biome in its central domain rates of conversion to pastures and croplands (Lapola et al., 2014; Song
(Fig. 4e, f). Nevertheless, knowing that models do not agree about the et al., 2018). Despite substantial agricultural intensification in recent
sign of precipitation anomaly (Supplementary Information, and Figs. S1 years (VanWey et al., 2013), cropland extensification is evident in every
and S3), and with the possibility that the effect of CO2 fertilization does biome, except the caatinga. In the Amazon biome, ~30% of new crop
not fully happen, the simulated caatinga replacement by savanna in the land during 2000–2014 was the result of forest conversion (Zalles et al.,
future (+4 ◦ C scenario) must be treated with caution (Fig. 2b, c). In case 2019). Increased temperatures, changes in rainfall patterns and reduced
of temperature increase and taking into account changes in frequency water availability, are the most important causes of declining cropland
and intensity of droughts as well as aridization in northeastern Brazil productivity. The general trend of clearing rates in Brazil since 2012 has
(Marengo et al., 2017), it is more likely that the caatinga central domain been increasing (Carvalho et al., 2019). It is estimated that under the
will shift in the opposite direction and be replaced with semi-desert most pessimistic climate scenario southern Brazil may lose nearly 5
vegetation. In both cases (savannization or desertification), the caa million hectares of optimal agricultural area in 2030 (Assad et al., 2013).
tinga biome will experience an overall loss of biodiversity (Silva et al., Thus, projected trends in loss of optimal agricultural land may have a
2019). It is worth noting that our model experiments are not able to positive feedback loop, leading to a further acceleration of vegetation
identify potential impacts of the rate of climate change and how vege conversion and global warming. Land-use and land-cover change will
tation could adapt to the accelerated warming remains an open question also alter the projected vegetation patterns over Brazil by weakening the
8
hydrological cycle even more (Leite-Filho et al., 2019). In this study we Appendix A. Supplementary data
do not model land use and land-cover changes due to human activity,
but there is a need to integrate the suite of negative effects from human Supplementary data to this article can be found online at https://doi.
activity that could lead to a cascade of tipping points in combination to org/10.1016/j.palaeo.2021.110717.
the scrutinized driving factors.
References
5. Conclusions
Absy, M.L., Cleef, A., Fournier, Marc, Martin, Louis, Servant, Michel,
Sifeddine, Abdelfettah, da Silva, M. Ferreira, et al., 1991. Mise en évidence de quatre
Because neither available palaeorecords nor state-of-the-art models phases d’ouverture de la forêt dense dans le Sud-Est de l’Amazonie au cours des 60
alone are conclusive, especially under extremely heterogeneous and 000 dernières années: première comparaison avec d’autres régions tropicales. In:
complex environmental conditions, it is crucial to employ both while Comptes Rendus de l’Académie des Sciences. Série 2: Mécanique..., 312(6),
pp. 673–678.
reconstructing and projecting past and future Brazilian biomes. Adam, O., Schneider, T., Brient, F., 2018. Regional and seasonal variations of the double-
Before exploring the extent of Brazilian biome changes in the past ITCZ bias in CMIP5 models. Clim. Dyn. 51 (1–2), 101–117.
(− 4 ◦ C scenario, standing for the LGM) and in the future (+4 ◦ C sce Aguirre-Gutiérrez, J., Malhi, Y., Lewis, S.L., et al., 2020. Long-term droughts may drive
drier tropical forests towards increased functional, taxonomic and phylogenetic
nario, standing for the end of the twenty-first century), we computed homogeneity. Nat. Commun. 11, 3346.
precipitation anomalies over Brazil and validated them with an updated Allen, J.R.M., Forrest, M., Hickler, T., Singarayer, J.S., Valdes, P.J., Huntley, B., 2020.
compilation of 149 published reconstructions. Computed LGM anoma Global vegetation patterns of the past 140,000 years. J. Biogeogr. 2020, 1–18, 00.
Aragão, L.E.O., Malhi, Y., Barbier, N., Lima, A., Shimabukuro, Y., Anderson, L.,
lies showed decreased precipitation over equatorial (0◦ -10◦ S) Brazil.
Saatchi, S., 2008. Interactions between rainfall, deforestation and fires during recent
Over the SAMS region precipitation increased. In comparison with the years in the Brazilian Amazonia. Philos. Trans. R. Soc. B 363 (1498), 1779–1785.
compilation of published reconstructions we conclude that the Anhuf, D., Ledru, M.P., Behling, H., Da Cruz Jr., F.W., Cordeiro, R.C., Van der
computed climate anomalies over Brazil are reliable and valid as inputs Hammen, T., Karmann, I., Marengo, J.A., De Oliveira, P.E., Pessenda, L.,
Siffedine, A., 2006. Paleo-environmental change in Amazonian and African
for the potential vegetation model for the -4 ◦ C scenario. rainforest during the LGM. Palaeogeography, Palaeoclimatology, Palaeoecology 239
Our results are consistent with proxy reconstructions and suggest (3–4), 510–527.
expansion of grassland in the southern Brazilian highlands and Amazon Aragão, L.E., Anderson, L.O., Fonseca, M.G., Rosan, T.M., Vedovato, L.B., Wagner, F.H.,
Silva, C.V., Junior, C.H.S., Arai, E., Aguiar, A.P., Barlow, J., 2018. 21st Century
rainforest persistence through colder and drier conditions during the drought-related fires counteract the decline of Amazon deforestation carbon
LGM. The results from the vegetation model reveal that the western and emissions. Nat. Commun. 9 (1), 1–12.
central Amazon forest were largely maintained during the LGM mainly Arneth, A., Harrison, S.P., Zaehle, S., Tsigaridis, K., Menon, S., Bartlein, P.J., Feichter, J.,
Korhola, A., Kulmala, M., O’donnell, D., Schurgers, G., 2010. Terrestrial
due to negative temperature anomalies, while a decrease in past pre biogeochemical feedbacks in the climate system. Nat. Geosci. 3 (8), 525–532.
cipitation was responsible for tropical evergreen forest to be substituted Assad, E., Pinto, H.S., Nassar, A., Harfuch, L., Freitas, S., Farinelli, B., Lundell, M.,
by tropical seasonal forest and savanna/cerrado in its eastern portion. Bachion, L.C., Fernandes, E.C.M., 2013. Impactos das Mudanças Climáticas na
Produção Agrícola Brasileira (Washington DC).
We reaffirmed that changes in monsoon intensity cannot be used as the Azevedo, V., Strikis, N.M., Novello, V.F., Roland, C.L., Cruz, F.W., Santos, R.V.,
main driver for vegetational changes/stability across the Amazon Vuille, M., Utida, G., De Andrade, F.R.D., Cheng, H., Edwards, R.L., 2021.
biome, and that lower temperatures in combination with substantially Paleovegetation seesaw in Brazil since the late Pleistocene: a multiproxy study of
two biomes. Earth Planet. Sci. Lett. 563, 116880.
lower CO2 are important controlling factors during the LGM. As our
Baccini, A., Walker, W., Carvalho, L., Farina, M., Sulla-Menashe, D., Houghton, R.A.,
results are broadly consistent with independent reconstructions of 2017. Tropical forests are a net carbon source based on aboveground measurements
vegetation, we conclude that the omission of these LGM vegetation of gain and loss. Science 358 (6360), 230–234.
changes in the experimental design for LGM climate simulations could Baker, P.A., Fritz, S.C., 2015. Nature and causes of Quaternary climate variation of
tropical South America. Quat. Sci. Rev. 124, 31–47.
induce underestimation of important climate-vegetation feedback. Baker, P.A., Fritz, S.C., Battisti, D.S., Dick, C.W., Vargas, O.M., Asner, G.P., Martin, R.E.,
In contrast, we find that for the future +4 ◦ C scenario, biome shifts Wheatley, A., Prates, I., 2020. Beyond Refugia: new insights on Quaternary climate
will mostly depend on future precipitation changes. Under future variation and the evolution of biotic diversity in tropical South America. In:
Neotropical Diversification: Patterns and Processes. Springer, Cham, pp. 51–70.
warming, our simulations show an expansion of savanna/cerrado, with Beerling, D.J., Chaloner, W.G., 1993. Evolutionary responses of stomatal density to
projected coverage 10% higher than Present. Further, future warming global CO2 change. Biol. J. Linn. Soc. 48 (4), 343–353.
will trigger a reduction of Amazon forest, tropical seasonal forest and Behling, H., 2002. Late Quaternary vegetation and climate dynamics in southeastern
Amazonia inferred from Lagoa da Confusao in Tocantins State, northern Brazil.
caatinga (savannization or desertification). We conclude that the +4 ◦ C Amazoniana 17 (1/2), 27–39.
scenario will impact local biodiversity and regional climate, but both Behling, H., Arz, H.W., Pätzold, J., Wefer, G., 2000. Late Quaternary vegetational and
will be much more strongly influenced by human activity than climate climate dynamics in northeastern Brazil, inferences from marine core GeoB 3104-1.
Quat. Sci. Rev. 19 (10), 981–994.
or CO2 and thus suggest addressing the human driven pressures as an Behling, H., Pillar, V.D., Orlóci, L., Bauermann, S.G., 2004. Late Quaternary Araucaria
additional driver. forest, grassland (Campos), fire and climate dynamics, studied by high-resolution
pollen, charcoal and multivariate analysis of the Cambará do Sul core in southern
Brazil. Palaeogeogr. Palaeoclimatol. Palaeoecol. 203 (3–4), 277–297.
Declaration of Competing Interest
Bouimetarhan, I., Chiessi, C.M., González-Arango, C., Dupont, L., Voigt, I., Prange, M.,
Zonneveld, K., 2018. Intermittent development of forest corridors in northeastern
None. Brazil during the last deglaciation: climatic and ecologic evidence. Quat. Sci. Rev.
192, 86–96.
Braconnot, P., Harrison, S.P., Kageyama, M., Bartlein, P.J., Masson-Delmotte, V., Abe-
Acknowledgements Ouchi, A., Otto-Bliesner, B., Zhao, Y., 2012. Evaluation of climate models using
palaeoclimatic data. Nat. Clim. Chang. 2 (6), 417–424.
J.M. acknowledges the financial support from FAPESP (grant 2018/ Brando, P.M., Balch, J.K., Nepstad, D.C., Morton, D.C., Putz, F.E., Coe, M.T., Silvério, D.,
Macedo, M.N., Davidson, E.A., Nóbrega, C.C., Alencar, A., 2014. Abrupt increases in
23522-6). CAPES financially supported I.M.V. with a scholarship (grant Amazonian tree mortality due to drought–fire interactions. Proc. Natl. Acad. Sci. 111
88887.156152/2017-00 and 88881.161151/2017-01). C.M.C. ac (17), 6347–6352.
knowledges the financial support from FAPESP (grant 2018/15123-4), Brando, P., Macedo, M., Silvério, D., Rattis, L., Paolucci, L., Alencar, A., Coe, M.,
Amorim, C., 2020. Amazon wildfires: scenes from a foreseeable disaster. Flora 268,
CAPES (grants 564/2015 and 88881.313535/2019-01), CNPq (grant 151609.
302607/2016-1) and the Alexander von Humboldt Foundation. This Brienen, R.J., Phillips, O.L., Feldpausch, T.R., Gloor, E., Baker, T.R., Lloyd, J., Lopez-
study was financed in part by CNPq Project RAiN (grant 406322/2018- Gonzalez, G., Monteagudo-Mendoza, A., Malhi, Y., Lewis, S.L., Martinez, R.V., 2015.
Long-term decline of the Amazon carbon sink. Nature 519 (7543), 344–348.
0). The data reported in this paper will be archived in Pangaea (https://d Brienen, R.J., Caldwell, L., Duchesne, L., Voelker, S., Barichivich, J., Baliva, M.,
oi.org/10.1594/PANGAEA.927527). Ceccantini, G., Di Filippo, A., Helama, S., Locosselli, G.M., Lopez, L., 2020. Forest
carbon sink neutralized by pervasive growth-lifespan trade-offs. Nat. Commun. 11
(1), 1–10.
9
Burbridge, R.E., Mayle, F.E., Killeen, T.J., 2004. Fifty-thousand-year vegetation and Gerhart, L.M., Ward, J.K., 2010. Plant responses to low [CO2] of the past. New Phytol.
climate history of Noel Kempff Mercado National park, Bolivian Amazon. Quat. Res. 188 (3), 674–695.
61 (2), 215–230. Gonzalez-Orozco, C.E., Pollock, L.J., Thornhill, A.H., Mishler, B.D., Knerr, N., Laffan, S.
Bush, M.B., 2002. On the interpretation of fossil Poaceae pollen in the lowland humid W., Miller, J.T., Rosauer, D.F., Faith, D.P., Nipperess, D.A., Kujala, H., 2016.
neotropics. Palaeogeogr. Palaeoclimatol. Palaeoecol. 177 (1–2), 5–17. Phylogenetic approaches reveal biodiversity threats under climate change. Nat.
Bush, M.B., Silman, M.R., 2004. Observations on Late Pleistocene cooling and Clim. Chang. 6 (12), 1110–1114.
precipitation in the lowland Neotropics. J. Quat. Sci. 19 (7), 677–684. Gosling, W.D., Bush, M.B., Hanselman, J.A., Chepstow-Lusty, A., 2008. Glacial-
Campbell, J.E., Berry, J.A., Seibt, U., Smith, S.J., Montzka, S.A., Launois, T., Belviso, S., interglacial changes in moisture balance and the impact on vegetation in the
Bopp, L., Laine, M., 2017. Large historical growth in global terrestrial gross primary southern hemisphere tropical Andes (Bolivia/Peru). Palaeogeogr. Palaeoclimatol.
production. Nature 544 (7648), 84–87. Palaeoecol. 259 (1), 35–50.
Carvalho, W.D., Mustin, K., Hilário, R.R., Vasconcelos, I.M., Eilers, V., Fearnside, P.M., Govin, A., Chiessi, C.M., Zabel, M., Sawakuchi, A.O., Heslop, D., Hörner, T., Zhang, Y.,
2019. Deforestation control in the Brazilian Amazon: a conservation struggle being Mulitza, S., 2014. Terrigenous input off northern South America driven by changes
lost as agreements and regulations are subverted and bypassed. Perspect. Ecol. in Amazonian climate and the North Brazil current retroflection during the last 250
Conserv. 17 (3), 122–130. ka. Clim. Past 10 (2), 843–862.
Cheng, H., Sinha, A., Cruz, F.W., Wang, X., Edwards, R.L., d’Horta, F.M., Ribas, C.C., Grosjean, M., 1994. Paleohydrology of the Laguna Lejía (north Chilean Altiplano) and
Vuille, M., Stott, L.D., Auler, A.S., 2013. Climate change patterns in Amazonia and climatic implications for late-glacial times. Palaeogeogr. Palaeoclimatol. Palaeoecol.
biodiversity. Nat. Commun. 4 (1), 1–6. 109 (1), 89–100.
Chiessi, C.M., Mulitza, S., Mollenhauer, G., Silva, J.B., Groeneveld, J., Prange, M., 2015. Gu, F., Zonneveld, K.A., Chiessi, C.M., Arz, H.W., Pätzold, J., Behling, H., 2017. Long-
Thermal evolution of the western South Atlantic and the adjacent continent during term vegetation, climate and ocean dynamics inferred from a 73,500 years old
termination 1. Clim. Past 11 (6). marine sediment core (GeoB2107-3) off southern Brazil. Quat. Sci. Rev. 172, 55–71.
Ciemer, C., Boers, N., Hirota, M., Kurths, J., Müller-Hansen, F., Oliveira, R.S., Gu, F., Chiessi, C.M., Zonneveld, K.A., Behling, H., 2018. Late Quaternary environmental
Winkelmann, R., 2019. Higher resilience to climatic disturbances in tropical dynamics inferred from marine sediment core GeoB6211-2 off southern Brazil.
vegetation exposed to more variable rainfall. Nat. Geosci. 12 (3), 174–179. Palaeogeogr. Palaeoclimatol. Palaeoecol. 496, 48–61.
Clark, P.U., Dyke, A.S., Shakun, J.D., Carlson, A.E., Clark, J., Wohlfarth, B., Mitrovica, J. Haberle, S.G., Maslin, M.A., 1999. Late Quaternary vegetation and climate change in the
X., Hostetler, S.W., McCabe, A.M., 2009. The last glacial maximum. Science 325 Amazon Basin based on a 50,000 year pollen record from the Amazon Fan, ODP Site
(5941), 710–714. 932. Quat. Res. 51 (1), 27–38.
Colinvaux, P.A., De Oliveira, P.E., Moreno, J.E., Miller, M.C., Bush, M.B., 1996. A long Häggi, C., Chiessi, C.M., Merkel, U., Mulitza, S., Prange, M., Schulz, M., Schefuß, E.,
pollen record from lowland Amazonia: forest and cooling in glacial times. Science 2017. Response of the Amazon rainforest to late Pleistocene climate variability.
274 (5284), 85–88. Earth Planet. Sci. Lett. 479, 50–59.
Colinvaux, P.A., De Oliveira, P.E., Bush, M.B., 2000. Amazonian and neotropical plant Haverd, V., Smith, B., Canadell, J.G., Cuntz, M., Mikaloff-Fletcher, S., Farquhar, G.,
communities on glacial time-scales: the failure of the aridity and refuge hypotheses. Woodgate, W., Briggs, P.R., Trudinger, C.M., 2020. Higher than expected CO2
Quat. Sci. Rev. 19 (1–5), 141–169. fertilization inferred from leaf to global observations. Glob. Chang. Biol. 26 (4),
Collins, M., Knutti, R., Arblaster, J., Dufresne, J.L., Fichefet, T., Friedlingstein, P., 2390–2402.
Gao, X., Gutowski, W.J., Johns, T., Krinner, G., Shongwe, M., 2013. Long-term Hoegh-Guldberg, O., Jacob, D., Bindi, M., Brown, S., Camilloni, I., Diedhiou, A.,
climate change: projections, commitments and irreversibility. In: Climate Change Djalante, R., Ebi, K., Engelbrecht, F., Guiot, J., Hijioka, Y., 2018. Impacts of 1.5 C
2013-The Physical Science Basis: Contribution of Working Group I to the Fifth global warming on natural and human systems. In: Global Warming of 1.5◦ C. An
Assessment Report of the Intergovernmental Panel on Climate Change. Cambridge IPCC Special Report.
University Press, pp. 1029–1136. Hoffmann, A.A., Sgro, C.M., 2011. Climate change and evolutionary adaptation. Nature
Cook-Patton, S.C., Leavitt, S.M., Gibbs, D., Harris, N.L., Lister, K., Anderson-Teixeira, K. 470 (7335), 479–485.
J., Briggs, R.D., Chazdon, R.L., Crowther, T.W., Ellis, P.W., Griscom, H.P., 2020. Hou, A., Bahr, A., Raddatz, J., Voigt, S., Greule, M., Albuquerque, A.L., Chiessi, C.M.,
Mapping carbon accumulation potential from global natural forest regrowth. Nature Friedrich, O., 2020. Insolation and greenhouse gas forcing of the South American
585 (7826), 545–550. Monsoon System across three glacial-interglacial cycles. Geophys. Res. Lett. 47 (14)
Cruz Jr., F.W., Burns, S.J., Karmann, I., Sharp, W.D., Vuille, M., 2006. Reconstruction of (p.e2020GL087948).
regional atmospheric circulation features during the late Pleistocene in subtropical Hubau, W., Lewis, S.L., Phillips, O.L., Affum-Baffoe, K., Beeckman, H., Cuní-Sanchez, A.,
Brazil from oxygen isotope composition of speleothems. Earth Planet. Sci. Lett. 248 Daniels, A.K., Ewango, C.E., Fauset, S., Mukinzi, J.M., Sheil, D., 2020. Asynchronous
(1–2), 495–507. carbon sink saturation in African and Amazonian tropical forests. Nature 579 (7797),
Cruz, F.W., Vuille, M., Burns, S.J., Wang, X., Cheng, H., Werner, M., Edwards, R.L., 80–87.
Karmann, I., Auler, A.S., Nguyen, H., 2009. Orbitally driven east–west antiphasing of Huntzinger, D.N., Michalak, A.M., Schwalm, C., Ciais, P., King, A.W., Fang, Y.,
South American precipitation. Nat. Geosci. 2 (3), 210–214. Schaefer, K., Wei, Y., Cook, R.B., Fisher, J.B., Hayes, D., 2017. Uncertainty in the
da Silva, E.F., Lopes, K.D.S., Alves, R., Carreira, L.M.M., Silva, D.F.D., Romeiro, L.D.A., response of terrestrial carbon sink to environmental drivers undermines carbon-
Batista Júnior, W.F., Rodrigues, T.M., Secco, R.D.S., Guimarães, J.T.F., 2020. climate feedback predictions. Sci. Rep. 7 (1), 1–8.
Hydroclimate influences on modern pollen rain of upland southeastern Amazonia. Kageyama, M., Albani, S., Braconnot, P., Harrison, S.P., Hopcroft, P.O., Ivanovic, R.F.,
The Holocene 30 (5), 721–732. Lambert, F., Marti, O., Peltier, W.R., Peterschmitt, J.Y., Roche, D.M., 2017. The
De Faria, B.L., Marano, G., Piponiot, C., Silva, C.A., Dantas, V.D.L., Rattis, L., Rech, A.R., PMIP4 contribution to CMIP6-part 4: scientific objectives and experimental design of
Collalti, A., 2021. Model-based estimation of Amazonian forests recovery time after the PMIP4-CMIP6 Last Glacial Maximum experiments and PMIP4 sensitivity
drought and fire events. Forests 12 (1), 8. experiments. Geosci. Model Dev. 10 (11), 4035–4055.
De Oliveira, P.E., Barreto, A.M.F., Suguio, K., 1999. Late Pleistocene/Holocene climatic Knutti, R., Sedláček, J., 2013. Robustness and uncertainties in the new CMIP5 climate
and vegetational history of the Brazilian caatinga: the fossil dunes of the middle São model projections. Nat. Clim. Chang. 3 (4), 369–373.
Francisco River. Palaeogeography, palaeoclimatology, palaeoecology 152 (3–4), Körner, C., Asshoff, R., Bignucolo, O., Hättenschwiler, S., Keel, S.G., Peláez-Riedl, S.,
319–337. Pepin, S., Siegwolf, R.T., Zotz, G., 2005. Carbon flux and growth in mature
Deininger, M., Ward, B.M., Novello, V.F., Cruz, F.W., 2019. Late Quaternary variations in deciduous forest trees exposed to elevated CO2. Science 309 (5739), 1360–1362.
the South American monsoon system as inferred by speleothems—new perspectives Kumar, D., Kodra, E., Ganguly, A.R., 2014. Regional and seasonal intercomparison of
using the SISAL database. Quaternary 2 (1), 6. CMIP3 and CMIP5 climate model ensembles for temperature and precipitation. Clim.
Dombroski, J.L.D., Praxedes, S.C., de Freitas, R.M.O., Pontes, F.M., 2011. Water relations Dyn. 43 (9–10), 2491–2518.
of Caatinga trees in the dry season. S. Afr. J. Bot. 77 (2), 430–434. Kutzbach, J., Gallimore, R., Harrison, S., Behling, P., Selin, R., Laarif, F., 1998. Climate
Duffy, P.B., Brando, P., Asner, G.P., Field, C.B., 2015. Projections of future and biome simulations for the past 21,000 years. Quat. Sci. Rev. 17 (6–7), 473–506.
meteorological drought and wet periods in the Amazon. Proc. Natl. Acad. Sci. 112 Lapola, D.M., Oyama, M.D., Nobre, C.A., 2009. Exploring the range of climate biome
(43), 13172–13177. projections for tropical South America: the role of CO2 fertilization and seasonality.
Esquivel-Muelbert, A., Baker, T.R., Dexter, K.G., Lewis, S.L., Brienen, R.J., Feldpausch, T. Glob. Biogeochem. Cycles 23 (3).
R., Lloyd, J., Monteagudo-Mendoza, A., Arroyo, L., Álvarez-Dávila, E., Higuchi, N., Lapola, D.M., Martinelli, L.A., Peres, C.A., Ometto, J.P., Ferreira, M.E., Nobre, C.A.,
2019. Compositional response of Amazon forests to climate change. Glob. Chang. Aguiar, A.P.D., Bustamante, M.M., Cardoso, M.F., Costa, M.H., Joly, C.A., 2014.
Biol. 25 (1), 39–56. Pervasive transition of the Brazilian land-use system. Nat. Clim. Chang. 4 (1), 27–35.
Etterson, J.R., Shaw, R.G., 2001. Constraint to adaptive evolution in response to global Ledru, M.P., 2002. Late Quaternary history and evolution of the cerrados as revealed by
warming. Science 294 (5540), 151–154. palynological records. In: The Cerrados of Brazil: Ecology and Natural History of a
Farquhar, G.D., von Caemmerer, Berry, J.A., 1980. A biochemical model of Neotropical Savanna, pp. 33–50.
photosynthetic CO 2 assimilation in leaves of C 3 species. Planta 149 (1), 78–90. Ledru, M.P., Braga, P.I.S., Soubiès, F., Fournier, M., Martin, L., Suguio, K., Turcq, B.,
Flantua, S.G., Hooghiemstra, H., Grimm, E.C., Behling, H., Bush, M.B., González- 1996. The last 50,000 years in the Neotropics (Southern Brazil): evolution of
Arango, C., Gosling, W.D., Ledru, M.P., Lozano-García, S., Maldonado, A., Prieto, A. vegetation and climate. Palaeogeogr. Palaeoclimatol. Palaeoecol. 123 (1–4),
R., 2015. Updated site compilation of the Latin American pollen database. Rev. 239–257.
Palaeobot. Palynol. 223, 104–115. Ledru, M.P., Mourguiart, P., Riccomini, C., 2009. Related changes in biodiversity,
Fornace, K.L., Whitney, B.S., Galy, V., Hughen, K.A., Mayle, F.E., 2016. Late Quaternary insolation and climate in the Atlantic rainforest since the last interglacial.
environmental change in the interior South American tropics: new insight from leaf Palaeogeogr. Palaeoclimatol. Palaeoecol. 271 (1–2), 140–152.
wax stable isotopes. Earth Planet. Sci. Lett. 438, 75–85. Leite-Filho, A.T., de Sousa Pontes, V.Y., Costa, M.H., 2019. Effects of deforestation on the
Forzieri, G., Alkama, R., Miralles, D.G., Cescatti, A., 2017. Satellites reveal contrasting onset of the rainy season and the duration of dry spells in southern Amazonia.
responses of regional climate to the widespread greening of Earth. Science 356 J. Geophys. Res. 124 (10), 5268–5281.
(6343), 1180–1184.
10
Lenton, T.M., Held, H., Kriegler, E., Hall, J.W., Lucht, W., Rahmstorf, S., Schellnhuber, H. Rossetti, D.F., Bertani, T.C., Zani, H., Cremon, E.H., Hayakawa, E.H., 2012. Late
J., 2008. Tipping elements in the Earth’s climate system. Proc. Natl. Acad. Sci. 105 Quaternary sedimentary dynamics in Western Amazonia: implications for the origin
(6), 1786–1793. of open vegetation/forest contrasts. Geomorphology 177, 74–92.
Lewis, S.L., Phillips, O.L., Baker, T.R., Lloyd, J., Malhi, Y., Almeida, S., Higuchi, N., Rossetti F, D, Moulatlet M, G, Tuomisto, H, Gribel, R, Toledo M, P, Valeriano M, M,
Laurance, W.F., Neill, D.A., Silva, J.N.M., Terborgh, J., 2004. Concerted changes in Ruokolainen, K, Cohen C, M, Cordeiro L, C, Renno D, C, Coelho S, L, 2019. White
tropical forest structure and dynamics: evidence from 50 South American long-term sand vegetation in an Amazonian lowland under the perspective of a young
plots. Philos. Trans. R. Soc. Lond. B 359 (1443), 421–436. geological history. Anais da Academia Brasileira de Ciências 91, 1–21.
Lima, A.L.A., Rodal, M.J.N., 2010. Phenology and wood density of plants growing in the Salgado-Labouriau, M.L., Casseti, V., Ferraz-Vicentini, K.R., Martin, L., Soubiès, F.,
semi-arid region of northeastern Brazil. J. Arid Environ. 74 (11), 1363–1373. Suguio, K., Turcq, B., 1997. Late Quaternary vegetational and climatic changes in
Lovejoy, Thomas E., Nobre, Carlos, 2019. Amazon Tipping Point: Last Chance for Action cerrado and palm swamp from Central Brazil. Palaeogeography, Palaeoclimatology,
(eaba2949). Palaeoecology 128 (1–4), 215–226.
Magrin, G.O., Marengo, J.A., Boulanger, J.-P., Buckeridge, M.S., Castellanos, E., Scheff, J., Seager, R., Liu, H., Coats, S., 2017. Are glacials dry? Consequences for
Poveda, G., Scarano, F.R., Vicuña, S., 2014. Central and South America. In: Barros, V. paleoclimatology and for greenhouse warming. J. Clim. 30 (17), 6593–6609.
R., Field, C.B., Dokken, D.J., Mastrandrea, M.D., Mach, K.J., Bilic, T.E., Seddon, Alistair W.R., et al., 2016. Sensitivity of global terrestrial ecosystems to climate
Chatterjee, M., Ebi, K.I., Estrada, Y.O., Genova, R.C., Birma, B., Kissel, E.S., Levy, A. variability. Nature 531 (7593), 229–232.
N., MacCracken, S., Mastrandrea, P.R., White, L.L. (Eds.), Climate Change 2014: Shimizu, M.H., Sampaio, G., Venancio, I.M., Maksic, J., 2020. Seasonal changes of the
Impacts, Adaptation, and Vulnerability. Part B: Regional Aspects. Contribution of South American monsoon system during the Mid-Holocene in the CMIP5
Working Group II to the Fifth Assessment Report of the Intergovernmental Panel on simulations. Clim. Dyn. 54 (5), 2697–2712.
Climate Change. Cambridge University Press, Cambridge, pp. 1499–1566. Sifeddine, A., Martin, L., Turcq, B., Volkmer-Ribeiro, C., Soubiès, F., Cordeiro, R.C.,
Marengo, J.A., Torres, R.R., Alves, L.M., 2017. Drought in Northeast Brazil—past, Suguio, K., 2001. Variations of the Amazonian rainforest environment: a
present, and future. Theor. Appl. Climatol. 129 (3–4), 1189–1200. sedimentological record covering 30,000 years. Palaeogeogr. Palaeoclimatol.
Mayle, F.E., Burbridge, R., Killeen, T.J., 2000. Millennial-scale dynamics of southern Palaeoecol. 168 (3–4), 221–235.
Amazonian rain forests. Science 290 (5500), 2291–2294. Silva, J.L.S.E., Cruz-Neto, O., Peres, C.A., Tabarelli, M., Lopes, A.V., 2019. Climate
Mosblech, N.A.S., 2012. Central Andean Vegetation Response to Rapid Paleoclimate change will reduce suitable Caatinga dry forest habitat for endemic plants with
Changes. Florida Institute of Technology. disproportionate impacts on specialized reproductive strategies. PLoS One 14 (5) (p.
Mueller, B., Seneviratne, S.I., 2014. Systematic land climate and evapotranspiration e0217028).
biases in CMIP5 simulations. Geophys. Res. Lett. 41 (1), 128–134. Silveira, M.H.B., Mascarenhas, R., Cardoso, D., Batalha-Filho, H., 2019. Pleistocene
Niemann, H., Behling, H., 2008. Late Quaternary vegetation, climate and fire dynamics climatic instability drove the historical distribution of forest islands in the
inferred from the El Tiro record in the southeastern Ecuadorian Andes. J. Quat. Sci. northeastern Brazilian Atlantic Forest. Palaeogeogr. Palaeoclimatol. Palaeoecol. 527,
23 (3), 203–212. 67–76.
Nobre, I., Nobre, C., 2020. Amazon 4.0. Futuribles (1),, 95–108. Song, X.P., Hansen, M.C., Stehman, S.V., Potapov, P.V., Tyukavina, A., Vermote, E.F.,
Nobre, C.A., Sellers, P.J., Shukla, J., 1991. Amazonian deforestation and regional climate Townshend, J.R., 2018. Global land change from 1982 to 2016. Nature 560 (7720),
change. J. Clim. 4 (10), 957–988. 639–643.
Nolan, C., Overpeck, J.T., Allen, J.R., Anderson, P.M., Betancourt, J.L., Binney, H.A., Staal, A., Tuinenburg, O.A., Bosmans, J.H., Holmgren, M., van Nes, E.H., Scheffer, M.,
Brewer, S., Bush, M.B., Chase, B.M., Cheddadi, R., Djamali, M., 2018. Past and future Zemp, D.C., Dekker, S.C., 2018. Forest-rainfall cascades buffer against drought
global transformation of terrestrial ecosystems under climate change. Science 361 across the Amazon. Nature Climate Change 8 (6), 539–543.
(6405), 920–923. Staver, A.C., Archibald, S., Levin, S.A., 2011. The global extent and determinants of
Novello, V.F., Cruz, F.W., Vuille, M., Stríkis, N.M., Edwards, R.L., Cheng, H., Emerick, S., savanna and forest as alternative biome states. Science 334 (6053), 230–232.
De Paula, M.S., Li, X., Barreto, E.D.S., Karmann, I., 2017. A high-resolution history of Stute, M., Forster, M., Frischkorn, H., Serejo, A., Clark, J.F., Schlosser, P., Broecker, W.S.,
the South American Monsoon from Last Glacial Maximum to the Holocene. Sci. Rep. Bonani, G., 1995. Cooling of tropical Brazil (5 C) during the last glacial maximum.
7 (1), 1–8. Science 269 (5222), 379–383.
Novello, V.F., Cruz, F.W., McGlue, M.M., Wong, C.I., Ward, B.M., Vuille, M., Santos, R.A., Taylor E, K, Stouffer J, R, Meehl A, G, 2012. An overview of CMIP5 and the experiment
Jaqueto, P., Pessenda, L.C., Atorre, T., Ribeiro, L.M., 2019. Vegetation and design. Bulletin of the American meteorological Society 93 (4), 485–498.
environmental changes in tropical South America from the last glacial to the Tebaldi, C., Knutti, R., 2007. The use of the multi-model ensemble in probabilistic
Holocene documented by multiple cave sediment proxies. Earth Planet. Sci. Lett. climate projections. Philos. Trans. R. Soc. A Math. Phys. Eng. Sci. 365 (1857),
524, 115717. 2053–2075.
Olson, D.M., Dinerstein, E., Wikramanayake, E.D., Burgess, N.D., Powell, G.V., Ter Steege, H., Prado, P.I., de Lima, R.A., Pos, E., de Souza Coelho, L., de Andrade Lima
Underwood, E.C., D’amico, J.A., Itoua, I., Strand, H.E., Morrison, J.C., Loucks, C.J., Filho, D., Salomão, R.P., Amaral, I.L., de Almeida Matos, F.D., Castilho, C.V.,
2001. Terrestrial ecoregions of the world: a new map of life on Earth. A new global Phillips, O.L., 2020. Biased-corrected richness estimates for the Amazonian tree
map of terrestrial ecoregions provides an innovative tool for conserving biodiversity. flora. Sci. Rep. 10 (1), 1–13.
BioScience 51 (11), 933–938. Terrer, C., Vicca, S., Hungate, B.A., Phillips, R.P., Prentice, I.C., 2016. Mycorrhizal
Oyama, M.D., Nobre, C.A., 2003. A new climate-vegetation equilibrium state for tropical association as a primary control of the CO2 fertilization effect. Science 353 (6294),
South America. Geophys. Res. Lett. 30 (23). 72–74.
Peñuelas, J., Canadell, J.G., Ogaya, R., 2011. Increased water-use efficiency during the Terrer, C., Jackson, R.B., Prentice, I.C., Keenan, T.F., Kaiser, C., Vicca, S., Fisher, J.B.,
20th century did not translate into enhanced tree growth. Glob. Ecol. Biogeogr. 20 Reich, P.B., Stocker, B.D., Hungate, B.A., Peñuelas, J., 2019. Nitrogen and
(4), 597–608. phosphorus constrain the CO 2 fertilization of global plant biomass. Nat. Clim.
Perugini, L., Caporaso, L., Marconi, S., Cescatti, A., Quesada, B., de Noblet-Ducoudre, N., Chang. 9 (9), 684–689.
House, J.I., Arneth, A., 2017. Biophysical effects on temperature and precipitation Tierney, J., Zhu, J., King, J., Malevich, S.B., Hakim, G., Poulsen, C., 2019. Glacial Cooling
due to land cover change. Environ. Res. Lett. 12 (5), 053002. and Climate Sensitivity Revisited.
Piao, S., Wang, X., Park, T., Chen, C., Lian, X.U., He, Y., Bjerke, J.W., Chen, A., Ciais, P., Trisos, C.H., Merow, C., Pigot, A.L., 2020. The projected timing of abrupt ecological
Tømmervik, H., Nemani, R.R., 2020. Characteristics, drivers and feedbacks of global disruption from climate change. Nature 580 (7804), 496–501.
greening. Nat. Rev. Earth Environ. 1 (1), 14–27. Urrego, D.H., Silman, M.R., Bush, M.B., 2005. The Last Glacial Maximum: stability and
Pinaya, J.L., Cruz, F.W., Ceccantini, G.C., Corrêa, P.L., Pitman, N., Vemado, F., Maria del change in a western Amazonian cloud forest. J. Quat. Sci. 20 (7–8), 693–701.
Carmen, S.L., Pereira Filho, A.J., Grohmann, C.H., Chiessi, C.M., Stríkis, N.M., 2019. Urrego, D.H., Bush, M.B., Silman, M.R., 2010. A long history of cloud and forest
Brazilian montane rainforest expansion induced by Heinrich Stadial 1 event. Sci. migration from Lake Consuelo, Peru. Quat. Res. 73 (2), 364–373.
Rep. 9 (1), 1–14. Utida, G., Cruz, F.W., Santos, R.V., Sawakuchi, A.O., Wang, H., Pessenda, L.C.,
Prado, L., Wainer, I., Chiessi, C., Ledru, M.P., Turcq, B., 2013. A Mid-Holocene Climate Novello, V.F., Vuille, M., Strauss, A.M., Borella, A.C., Stríkis, N.M., 2020. Climate
Reconstruction for Eastern South America. changes in Northeastern Brazil from deglacial to Meghalayan periods and related
Prentice, I.C., Harrison, S.P., Bartlein, P.J., 2011. Global vegetation and terrestrial environmental impacts. Quat. Sci. Rev. 250, 106655.
carbon cycle changes after the last ice age. New Phytol. 189 (4), 988–998. Van Der Hammen, T., Hooghiemstra, H., 2000. Neogene and Quaternary history of
Quesada, C.A., Lloyd, J., Schwarz, M., Patiño, S., Baker, T.R., Czimczik, C., Fyllas, N.M., vegetation, climate, and plant diversity in Amazonia. Quat. Sci. Rev. 19 (8),
Martinelli, L., Nardoto, G.B., Schmerler, J., Santos, A.J.B., 2010. Variations in 725–742.
chemical and physical properties of Amazon forest soils in relation to their genesis. Van Der Sleen, P., Groenendijk, P., Vlam, M., Anten, N.P., Boom, A., Bongers, F., Pons, T.
Biogeosciences 7 (5), 1515–1541. L., Terburg, G., Zuidema, P.A., 2015. No growth stimulation of tropical trees by 150
Quesada, C.A., Phillips, O.L., Schwarz, M., Czimczik, C.I., Baker, T.R., Patiño, S., years of CO 2 fertilization but water-use efficiency increased. Nat. Geosci. 8 (1),
Fyllas, N.M., Hodnett, M.G., Herrera, R., Almeida, S., Dávila, E.A., 2012. Basin-wide 24–28.
variations in Amazon forest structure and function are mediated by both soils and VanWey, L.K., Spera, S., de Sa, R., Mahr, D., Mustard, J.F., 2013. Socioeconomic
climate. Biogeosciences 9 (6), 2203–2246. development and agricultural intensification in Mato Grosso. Philos. Trans. R. Soc. B
Root, T.L., Price, J.T., Hall, K.R., Schneider, S.H., Rosenzweig, C., Pounds, J.A., 2003. 368 (1619), 20120168.
Fingerprints of global warming on wild animals and plants. Nature 421 (6918), Vautard, R., Gobiet, A., Sobolowski, S., Kjellström, E., Stegehuis, A., Watkiss, P.,
57–60. Mendlik, T., Landgren, O., Nikulin, G., Teichmann, C., Jacob, D., 2014. The
Rosan, T.M., Aragão, L.E., Oliveras, I., Phillips, O.L., Malhi, Y., Gloor, E., Wagner, F.H., European climate under a 2 C global warming. Environ. Res. Lett. 9 (3), 034006.
2019. Extensive 21st-century woody encroachment in South America’s savanna. Walker, A.P., De Kauwe, M.G., Bastos, A., Belmecheri, S., Georgiou, K., Keeling, R.,
Geophys. Res. Lett. 46 (12), 6594–6603. McMahon, S.M., Medlyn, B.E., Moore, D.J., Norby, R.J., Zaehle, S., 2020. Integrating
the evidence for a terrestrial carbon sink caused by increasing atmospheric CO2.
New Phytol. 229 (5), 2413–2445.
11
Wang, X, Auler, A, Edwards, R, et al., 2004. Wet periods in northeastern Brazil over the Zalles, V., Hansen, M.C., Potapov, P.V., Stehman, S.V., Tyukavina, A., Pickens, A.,
past 210 kyr linked to distant climate anomalies. Nature 432, 740–743. https://doi. Song, X.P., Adusei, B., Okpa, C., Aguilar, R., John, N., 2019. Near doubling of
org/10.1038/nature03067. Brazil’s intensive row crop area since 2000. Proc. Natl. Acad. Sci. 116 (2), 428–435.
Wang, C., Zhang, L., Lee, S.K., Wu, L., Mechoso, C.R., 2014. A global perspective on Zhang, Y., Zhang, X., Chiessi, C.M., Mulitza, S., Zhang, X., Lohmann, G., Prange, M.,
CMIP5 climate model biases. Nat. Clim. Chang. 4 (3), 201–205. Behling, H., Zabel, M., Govin, A., Sawakuchi, A.O., 2016. Equatorial Pacific forcing
Werneck, Fernanda P., Costa, Gabriel C., Colli, Guarino R., Prado, Darién E., of western Amazonian precipitation during Heinrich Stadial 1. Sci. Rep. 6 (1), 1–7.
Sites Jr., Jack W., 2011. Revisiting the historical distribution of Seasonally Dry Zhu, Z., Piao, S., Myneni, R.B., Huang, M., Zeng, Z., Canadell, J.G., Ciais, P., Sitch, S.,
Tropical Forests: new insights based on palaeodistribution modelling and Friedlingstein, P., Arneth, A., Cao, C., 2016. Greening of the Earth and its drivers.
palynological evidencegeb. Global Ecol. Biogeogr. 20 (2), 272–288. Nat. Clim. Chang. 6 (8), 791–795.
Whitney, Bronwen S., et al., 2011. A 45 kyr palaeoclimate record from the lowland Heinrich, V.H.A., Dalagnol, R., Cassol, H.L.G., Rosan, T.M., de Almeida, C.T., Silva
interior of tropical South America. Palaeogeogr. Palaeoclimatol. Palaeoecol. 307 Junior, C.H.L., Campanharo, W.A., House, J.I., Sitch, S., Hales, T.C., Adami, M.,
(1–4), 177–192. Anderson, L.O., Aragão, L.E.O.C., 2021. Large carbon sink potential of secondary
Wille, M., Hooghiemstra, H., Behling, H., van der Borg, K., Negret, A.J., 2001. forests in the Brazilian Amazon to mitigate climate change. Nat. Commun. 12 (1),
Environmental change in the Colombian subandean forest belt from 8 pollen records: 1785. https://doi.org/10.1038/s41467-021-22050-1. PMID: 33741981; PMCID:
the last 50 kyr. Veg. Hist. Archaeobotany 10 (2), 61–77. PMC7979697.
Willmott, C.J., Matsuura, K., 2001. Terrestrial Air Temperature and Precipitation:
Monthly and Annual Time Series (1950–1999) Version 1.02. Center for Climatic
Research, University of Delaware, Newark.
12

Mak Sic 2022

Uploaded by

Copyright:

Available Formats

Mak Sic 2022

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Mak Sic 2022

Uploaded by

Copyright:

Available Formats

Palaeogeography, Palaeoclimatology, Palaeoecology 585 (2022) 110717

Contents lists available at ScienceDirect

Palaeogeography, Palaeoclimatology, Palaeoecology

Brazilian biomes distribution: Past and future

1. Introduction particularly vulnerable biome to climate change (Seddon et al., 2016). In

substantially expand their occurrence with elevated CO2. Future pre

4.1. Biome simulations: from the − 4 ◦ C scenario to the present

Prior to discussing the biome simulations, it is worth noting that

Fig. 4. Sensitivity simulations for biome distri

You might also like

Mak Sic 2022

Uploaded by

Copyright:

Available Formats

Mak Sic 2022

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Mak Sic 2022

Uploaded by

Copyright:

Available Formats

Palaeogeography, Palaeoclimatology, Palaeoecology 585 (2022) 110717

Contents lists available at ScienceDirect

Palaeogeography, Palaeoclimatology, Palaeoecology

Brazilian biomes distribution: Past and future

1. Introduction particularly vulnerable biome to climate change (Seddon et al., 2016). In

substantially expand their occurrence with elevated CO2. Future pre­

4.1. Biome simulations: from the − 4 ◦ C scenario to the present

Prior to discussing the biome simulations, it is worth noting that

Fig. 4. Sensitivity simulations for biome distri­

You might also like

substantially expand their occurrence with elevated CO2. Future pre

Fig. 4. Sensitivity simulations for biome distri