CoChat
Sign Up
Cited by 5Plant Water Relations and Carbon DynamicsClimate variability and modelsEcosystem dynamics and resilience

Amazon dieback beyond the 21st century under high-emission scenarios by Earth System models

Open access
Melnikova, Irina, Hajima, Tomohiro, Shiogama, Hideo, Hayashi, Michiya, Ito, Akihiko, Nishina, Kazuya, Tachiiri, Kaoru, Yokohata, Tokuta·2025·Communications Earth & Environment
DOI 10.1038/s43247-025-02606-5
Open this paper in CoChat — free accountOpen PDF

Abstract

The Amazon rainforest, the world’s largest tropical forest, plays a crucial role in climate regulation but faces risks of dieback under climate change. Here, we examine the Amazon dieback projections from state-of-the-art Earth System Models under high-emission scenarios extending into the 23rd century. Unlike previous studies focused on rapid changes, we investigate both rapid and gradual dieback, capturing drastic productivity declines (>80%). Nine of twelve models project Amazon dieback with variations in timing and spatial extent. The dieback onset is projected within the 21st century under global warming levels ranging from 1.5 to 10.2 °C, triggered by local surface air temperatures above 32.2 ± 4.8 °C and precipitation below 1394.3 ± 306.0 mm year−1, exacerbated by land-use change. We identify large-scale atmospheric circulation changes causing regional drying and warming and examine ecosystem processes contributing to dieback. These findings refine our understanding of Amazon dieback thresholds and mechanisms, highlighting areas for model improvement. Future Amazon dieback could occur during the 21st century, triggered by local annual surface air temperatures above 32.2 °C and precipitation below 1394.3 mm per year, and exacerbated by land-use change, according to an analysis of Earth System Model projections under different levels of warming

AI snapshot

Written from the paper’s full text.

TL;DR

Nine of twelve CMIP5/CMIP6 Earth System Models project Amazon dieback under RCP8.5/SSP5-8.5 by the end of the 23rd century, with onset often in the 21st century and completion around 2195. Dieback is associated with local surface air temperatures above 32.2 ± 4.8 °C, precipitation below 1394.3 ± 306.0 mm year⁻¹, and land-use change.

Key findings

  • Nine of twelve ESMs project substantial GPP reductions and identified dieback; largest dieback areas reach up to two million km² (HadGEM2-ES, CanESM5, IPSL-CM6A-LR, UKESM1-0-LL), while CESM2-WACCM projects less than 0.1 million km². Mean dieback area across models is 1.0 ± 0.8 million km².
  • On average, dieback onset is projected around year 2087 ± 49 and completion around 2195 ± 25; six of nine dieback models project onset within the 21st century. GSAT at onset ranges from 1.5 °C to 10.2 °C, with an inter-model mean of 5.0 ± 2.9 °C.
  • Local thresholds at dieback onset average 32.2 ± 4.8 °C surface air temperature and 1394.3 ± 306.0 mm year⁻¹ precipitation. Temperature is the dominant driver in most models, though precipitation is strong in some CMIP5 models.
  • The dieback region shifts from a carbon sink to a temporary carbon source of up to 0.5 GtC year⁻¹ during the 22nd century, returning to near-neutral or weak source conditions by the 23rd century.

Methods

  • Analyzed concentration-driven historical and RCP8.5 (CMIP5) or SSP5-8.5 (CMIP6) ESM outputs with carbon-cycle variables extending beyond the 21st century, including GPP, temperature, precipitation, LAI, NPP, respiration, vegetation and soil carbon, and land-use fluxes.
  • Defined dieback in grid cells with >80% decrease in mean GPP from 2259–2299 compared to 1859–1899 and historical GPP >2 kgC m⁻² year⁻¹. Onset was identified from a Savitzky-Golay–smoothed GPP derivative falling below −0.05; completion when smoothed GPP dropped to 20 ± 1% of maximum.
  • Used z-scored logistic regression to assess temperature versus precipitation influence, Köppen-Geiger classification for climate shifts, and diagnostics for AMOC and ENSO-like SST gradients.

Limitations

  • Stated: coarse spatial resolution can produce adjacent dieback and non-dieback grid cells; ESMs typically lack lateral ecological interactions, potentially overestimating Amazon stability; fire modules are limited and inconsistent, so fire impacts are likely underestimated; nutrient limitation is poorly represented, with all non-dieback models lacking a nitrogen cycle; tree mortality thresholds are highly uncertain; Köppen-Geiger categories may miss ecological thresholds; piControl drift was not corrected.
  • Stated: models with high equilibrium climate sensitivity overestimate present-day surface air temperature in dieback regions, reducing confidence in their dieback timing; many models do not account for acclimation processes that could mitigate dieback.
  • Evident from the text: the study is restricted to high-emission RCP8.5/SSP5-8.5 scenarios, land cover is assumed constant after 2100, and the 80% GPP-loss definition is a study-specific threshold that may not capture all gradual or abrupt ecosystem transitions.

Model-written summary. Check it against the paper before relying on it.

Integrity signals

Registry

Crossref · 10.1038/s43247-025-02606-5

Cites

87 of 87 references verified0 citation errors · 0 not checkable

Statistics

Nothing to recompute.

Claims the paper makes

15 of 25 cited sources checked · 8 supported · 0 flagged · accuracy not yet measured

  • Premisecitation

    Observational studies suggest that local temperature and precipitation thresholds for dieback vary substantially depending on ecological and species-specific factors.

    “Moreover, observational studies suggest that local temperature and precipitation thresholds for dieback vary substantially depending on ecological and species-specific factors 56,57 .”
    Quote in the paperSubstantial claim2 of 2 sources checked
    • cites [56] Powers, J. S. et al. A catastrophic tropical drought kills hydraulically vulnerable tree species. Glob. Chang. Biol. 26, 3122–3133 (2020). source

      The Results (page unavailable) report that mortality during the 2015 drought varied greatly among species, and a further results passage links mortality to hydraulic traits and rainfall seasonality. The paper does not establish varying local temperature or precipitation thresholds for dieback; its thresholds concern plant hydraulic function, not climate conditions. (LLM, conf=0.86)

      Experimental — accuracy not yet measured
    • cites [57] McDowell, N. et al. Drivers and mechanisms of tree mortality in moist tropical forests. N. Phytol. 219, 851–869 (2018). source

      The ESM discussion (pp. 19, 23) says climate-envelope models miss variation across climates, species, and topography, and that trees do not all share the same drought-response threshold. But the paper does not report observationally established local temperature and precipitation thresholds for dieback; it calls for experiments to investigate threshold responses (p. 22). (LLM, conf=0.82)

      Experimental — accuracy not yet measured
  • Findingown data

    In some regions, model-derived GPP exceeds observational products by more than 1 kgC m−2 year−1, indicating substantial biases in magnitude and spatial patterns.

    “In some regions, model-derived GPP exceeds observational products by more than 1 kgC m -2 year -1 , highlighting substantial biases in both magnitude and spatial patterns.”
    Quote in the paperNot a substantial claim
  • Findingown data

    On average, models that sustain higher Amazon GPP under moderate warming tend to experience more abrupt GPP losses once higher GSAT levels are reached.

    “On average, models that sustain higher Amazon GPP under moderate warming tend to experience more abrupt GPP losses once higher GSAT levels are reached.”
    Quote in the paperSubstantial claim
  • Findingcitation

    The western-high/eastern-low GPP pattern is not supported by FLUXCOM-X or MUSES.

    “This pattern is not supported by the machine learning-based FLUXCOM-X product, which is trained on eddy covariance flux tower data 31 , nor by the satellite-based MUSES product 32 (Fig. S1).”
    Quote in the paperNot a substantial claim
  • Findingown data

    HadGEM2-ES, CanESM5, IPSL-CM6A-LR and UKESM1-0-LL project the largest dieback areas, covering up to two million km2.

    “HadGEM2-ES, CanESM5, IPSL-CM6A-LR and UKESM1-0-LL project the largest dieback areas, covering up to two million km 2 .”
    Quote in the paperSubstantial claim
  • May combine two claims. 31
  • Classified as probably not a substantial claim. 10
  • The quote only partly states this claim. 1

  • May combine two claims.

    Premisecitation

    Fire-vegetation feedbacks can reduce ecosystem resilience and promote persistent forest degradation.

    “Recent studies highlight that fire-vegetation feedbacks can reduce ecosystem resilience and promote persistent forest degradation 26,59,60 .”
    Quote in the paperSubstantial claim2 of 3 sources checked
    • cites [26] Cano, I. M. et al. Abrupt loss and uncertain recovery from fires of Amazon forests under low climate mitigation scenarios. Proc. Natl. Acad. Sci. USA 119, e2203200119 (2022). source
    • cites [59] Drüke, M. et al. Fire may prevent future Amazon forest recovery after large-scale deforestation. Commun. Earth Environ. 4, 248 (2023). source
    • cites [60] Flores, B. M. & Holmgren, M. White-sand savannas expand at the core of the Amazon after forest wildfires. Ecosystems 24, 1624–1637 (2021). source

  • May combine two claims.

    Premisecitation

    Internal feedbacks sustain alternative ecological states, consistent with tipping point theory.

    “These dynamics align with tipping point theory, where internal feedbacks sustain alternative ecological states 61 .”
    Quote in the paperNot a substantial claim1 of 1 sources checked
    • cites [61] van Nes, E. H. et al. Causal feedbacks in climate change. Nat. Clim. Chang. 5, 445–448 (2015). source

      The available indexed passages contain bibliographic metadata and keywords, not the paper’s substantive discussion (the records labelled Abstract have no abstract text). Searches for internal feedbacks sustaining alternative ecological states and for alternative stable states and tipping points found no passage addressing the claim. (LLM, conf=0.70)

      Experimental — accuracy not yet measured

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    ENSO is an important driver of climate variability in the Amazon basin, influencing temperature, precipitation and ecosystem dynamics.

    “The ENSO is an important driver of climate variability in the Amazon basin, influencing temperature, precipitation and ecosystem dynamics [66][67][68] .”
    Quote in the paperNot a substantial claim1 of 3 sources checked
    • cites [66] Mastropierro, M., Peano, D. & Zanchettin, D. Uncertainty in Amazon vegetation productivity in CMIP6 projections driven by surface energy fluxes. EGUsphere 2024, 1–26 (2024). source

      Checked before — its earlier verdict is held back until a new check runs.

    • cites [67] Geng, T., Cai, W., Jia, F. & Wu, L. Decreased ENSO post-2100 in response to formation of a permanent El Niño-like state under greenhouse warming. Nat. Commun. 15, 5810 (2024). source

      The Introduction calls ENSO a major mode of climate variability, and the Results section on a permanent El Niño-like state describes rainfall anomalies over eastern South America. The paper does not identify the Amazon basin or address ENSO’s effects on its temperature or ecosystem dynamics (searched: Amazon, ecosystem, temperature, South America). (LLM, conf=0.84)

      Experimental — accuracy not yet measured
    • cites [68] Peng, Q., Xie, S.-P. & Deser, C. Collapsed upwelling projected to weaken ENSO under sustained warming beyond the twenty-first century. Nat. Clim. Chang. 14, 815–822 (2024). source

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Premisecitation

    SSP5-8.5 projects a fossil-fuel-intensive future with high greenhouse gas emissions and minimal policy interventions, reflecting radiative forcing levels similar to RCP8.5.

    “Its successor, SSP5-8.5 projects a fossil-fuel-intensive future with high greenhouse gas emissions and minimal policy interventions, reflecting similar radiative forcing levels.”
    Quote in the paperNot a substantial claim1 of 1 sources checked
    • cites [63] Kriegler, E. et al. Fossil-fueled development (SSP5): an energy and resource intensive scenario for the 21st century. Glob. Environ. Chang. 42, 297–315 (2017). source

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Study design and conductown data

    Present-day GPP estimates from the ESMs were evaluated against the FLUXCOM-X-BASE and MUSES GPP products.

    “We evaluate the present-day GPP estimates against FLUXCOM-X-BASE GPP product 31 , which uses machine learning trained on eddy covariance data to generate global estimates, and MUltiscale Satellite remotE Sensing (MUSES) GPP product 32 .”
    Quote in the paperNot a substantial claim1 of 2 sources checked
    • cites [31] Nelson, J. A. et al. X-BASE: the first terrestrial carbon and water flux products from an extended data-driven scaling framework, FLUXCOM-X. Biogeosciences 21, 5079–5115 (2024). source
    • cites [32] Wang, J. et al. New global MuSyQ GPP/NPP remote sensing products from 1981 to 2018. IEEE J. Sel. Top. Appl. Earth Obs. Remote Sens. 14, 5596–5612 (2021). source

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Premisecitation

    RCP8.5 assumes a continued rise in greenhouse gas emissions without mitigation, leading to a radiative forcing of 8.5 W m-2 by 2100.

    “RCP8.5 assumes a continued rise in greenhouse gas emissions without mitigation, leading to a radiative forcing of 8.5 W m -2 by 2100.”
    Quote in the paperNot a substantial claim1 of 1 sources checked
    • cites [62] Riahi, K. et al. RCP 8.5—a scenario of comparatively high greenhouse gas emissions. Clim. Chang. 109, 33 (2011). source

  • Classified as probably not a substantial claim. May combine two claims.

    Study design and conductown data

    ERA5 reanalysis data were used to estimate the present-day climate in the dieback grid cells and to evaluate ESM performance for current precipitation and temperature.

    “We use ERA5 monthly averaged data 64 to estimate the present-day climate in the dieback grid cells and evaluate the performance of the ESM for current precipitation and temperature.”
    Quote in the paperSubstantial claim1 of 1 sources checked
    • cites [64] Hersbach, H. et al. ERA5 monthly averaged data on single levels from 1940 to present. Copernicus Climate Change Service (C3S) Climate Data Store (CDS) https://doi.org/10.24381/cds.f17050d7 (2023) (Accessed on 27 July 2025). source

  • May combine two claims.

    Premisecitation

    Both RCP8.5 and SSP5-8.5 include substantial land-use change emissions, which are largely offset by climate-induced forest regrowth.

    “While both scenarios include substantial LUC emissions, these are largely offset by climate-induced forest regrowth 43 .”
    Quote in the paperNot a substantial claim1 of 1 sources checked
    • cites [43] Davies-Barnard, T., Valdes, P. J., Singarayer, J. S., Wiltshire, A. J. & Jones, C. D. Quantifying the relative importance of land cover change from climate and land use in the representative concentration pathways. Glob. Biogeochem. Cycles 29, 842–853 (2015). source

      In the Discussion and Conclusions (lines 379–381), the paper says climate-induced forest increase partially offsets the reduced carbon sink from land-use change; it also reports that climate-induced land-cover change offsets 91% of anthropogenic deforestation in RCP8.5 (lines 398–401). These findings support an offset in RCP8.5, but do not establish that land-use-change emissions themselves are largely offset. The paper does not address SSP5-8.5, so it cannot support the claim for both scenarios. (LLM, conf=0.86)

      Experimental — accuracy not yet measured

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Premisecitation

    Recent studies suggest potential changes in ENSO variability and an increased occurrence of El Niño-like warming under sustained warming.

    “Recent studies on ENSO's long-term behaviour under sustained warming suggest potential changes in its variability and the increased occurrence of El Niño-like warming 67,68 .”
    Quote in the paperNot a substantial claim1 of 2 sources checked
    • cites [67] Geng, T., Cai, W., Jia, F. & Wu, L. Decreased ENSO post-2100 in response to formation of a permanent El Niño-like state under greenhouse warming. Nat. Commun. 15, 5810 (2024). source
    • cites [68] Peng, Q., Xie, S.-P. & Deser, C. Collapsed upwelling projected to weaken ENSO under sustained warming beyond the twenty-first century. Nat. Clim. Chang. 14, 815–822 (2024). source

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    Deforestation peaks in the first half of the 21st century.

    “Besides, the deforestation peaks in the first half of the 21st century 41 , with land cover assumed to remain constant beyond 2100.”
    Quote in the paperNot a substantial claim1 of 1 sources checked
    • cites [41] Hurtt, G. C. et al. Harmonization of land-use scenarios for the period 1500–2100: 600 years of global gridded annual land-use transitions, wood harvest, and resulting secondary lands. Clim. Chang. 109, 117 (2011). source

      The RCP8.5 scenario description (section 2.3 Biomass density; page unavailable) projects forest-cover loss of 300 million hectares from 2000 to 2050, versus 150 million hectares from 2050 to 2100. That supports faster loss in the first half of the century for one scenario, but does not establish a deforestation peak or the unqualified claim across scenarios. (LLM, conf=0.78)

      Experimental — accuracy not yet measured

  • May combine two claims.

    Premisecitation

    Repeated fires can lead to grass invasion, increased flammability and suppressed canopy recovery, particularly at the forest-savanna transition.

    “Repeated fires can lead to grass invasion, increased flammability and suppressed canopy recovery, particularly at the forest-savanna transition.”
    Quote in the paperSubstantial claim2 of 3 sources checked
    • cites [26] Cano, I. M. et al. Abrupt loss and uncertain recovery from fires of Amazon forests under low climate mitigation scenarios. Proc. Natl. Acad. Sci. USA 119, e2203200119 (2022). source
    • cites [59] Drüke, M. et al. Fire may prevent future Amazon forest recovery after large-scale deforestation. Commun. Earth Environ. 4, 248 (2023). source
    • cites [60] Flores, B. M. & Holmgren, M. White-sand savannas expand at the core of the Amazon after forest wildfires. Ecosystems 24, 1624–1637 (2021). source

      The Introduction and Soil Shift sections (pages unavailable) say recurrent fires can keep burnt forests open, while the Vegetation Shift section reports persistent herbaceous cover after fire. But the paper reports native herbs rather than grass invasion in its study sites, explicitly notes that non-native invasive grasses were absent, and studies floodplain forests at the Amazon’s core rather than establishing the claim particularly for forest–savanna transitions. (LLM, conf=0.86)

      Experimental — accuracy not yet measured

  • May combine two claims.

    Findingown data

    ESMs indicate high past GPP levels in the Amazon basin, with a wide range of GPP estimates in the historical period.

    “ESMs indicate high past GPP levels in the Amazon basin, which is the main driver of gross carbon flux into the land ecosystem and an indicator of vegetation activity, with a wide range of GPP estimates in the historical period (Figs. 1 and S1).”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Findingown data

    Amazon dieback is projected to begin around year 2087 ± 49 and complete by year 2195 ± 25 in the considered grid cells under RCP8.5 and SSP5-8.5.

    “On average, Amazon dieback is projected to begin around year 2087 ± 49 and complete by year 2195 ± 25 in the considered grid cells under the RCP8.5 and SSP5-8.5 scenarios.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    MIROC-ES2L projects the latest onset of dieback, beginning in the second half of the 22nd century, with the most abrupt GPP decline.

    “In contrast, MIROC-ES2L projects the latest onset of dieback, beginning in the second half of the 22nd century, with the most abrupt GPP decline.”
    Quote in the paperNot a substantial claim

  • The quote only partly states this claim.

    Findingown data

    On average, annual local surface air temperature above 32.2 ± 4.8 °C is associated with the onset of Amazon dieback.

    “On average, annual local surface air temperature above 32.2 ± 4.8 °C and precipitation below 1394.3 ± 306.0 mm year -1 are associated with the onset of the Amazon dieback.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingreasoning

    Models with coarser spatial resolution show large spatial variations in GPP changes by the end of the 23rd century, with dieback grid cells sometimes adjacent to grid cells of GPP increase.

    “Models with coarser spatial resolution, such as IPSL-CM5A-LR and MIROC-ES2L, show large spatial variations in GPP changes by the end of the 23rd century, with dieback grid cells sometimes adjacent to the grid cells of GPP increase.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    Global warming levels at Amazon dieback onset range from 1.5 °C in CESM2-WACCM to 10.2 °C in CanESM5, with an inter-model mean of 5.0 ± 2.9 °C.

    “with a range from 1.5 °C in CESM2-WACCM to 10.2 °C in CanESM5, and the inter-model mean value of 5.0° ± 2.9 °C (Table 2, Fig. 2b).”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    GISS-E2-R and HadGEM2-ES estimate the highest historical GPP among all ESMs, higher than machine-learning and satellite-based products.

    “GISS-E2-R and HadGEM2-ES estimate the highest historical GPP among all ESMs, and higher estimates compared to machine-learning and satellite-based products.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Findingown data

    The timing of dieback onset shows a strong linear relationship with GSAT, with models that delay dieback exhibiting higher GSAT thresholds.

    “The timing of dieback onset shows a strong linear relationship with GSAT, with models that delay dieback exhibiting higher GSAT thresholds (Fig. 2e and Table 2).”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    There is less inter-model variability in the timing of dieback completion than in onset, with a strong negative linear relationship between onset year and dieback duration.

    “Notably, there is less inter-model variability in the timing of dieback completion than in onset, with a strong negative linear relationship between onset year and dieback duration (see Table 2).”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Findingown data

    Temperature is the dominant driver of dieback in most models, although precipitation exerts a strong influence in some, particularly among the CMIP5 ensemble.

    “The results show that temperature is the dominant driver of dieback in most models, although precipitation exerts a strong influence in some, particularly among the CMIP5 ensemble.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    CanESM5 and MIROC-ES2L show a stronger sensitivity of GPP to rising temperatures, while other models show a dependence on the combined effects of rising temperatures and decreasing precipitation.

    “CanESM5 and MIROC-ES2L show a stronger sensitivity of GPP to rising temperatures, while other models show a dependence on the combined effects of rising temperatures and decreasing precipitation.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    The models without dieback predict future GSAT changes ranging from the lowest values to near the inter-model mean, with moderate to negligible reductions in annual precipitation over the Amazon.

    “These models predict future GSAT changes ranging from the lowest values to near the inter-model mean, with moderate to negligible reductions in annual precipitation over the Amazon (Fig. S3).”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    MPI-ESM-LR shows a moderate GPP decline despite an increased temperature (>40 °C) and reduced precipitation (<700 mm year-1), but this does not meet the dieback criteria.

    “MPI-ESM-LR shows a moderate GPP decline despite an increased temperature (>40 °C) and reduced precipitation (<700 mm year -1 ), but this does not meet the dieback criteria.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingreasoning

    BCC-CSM1-1 and MPI-ESM-LR do not exhibit dieback despite having global and Amazon temperature trends similar to models that do.

    “However, two other models (BCC-CSM1-1 and MPI-ESM-LR) also do not exhibit dieback, despite having global and Amazon temperature trends similar to models that do.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    GISS-E2-R has the smallest GSAT increase by the late 21st century, alongside slight precipitation increases and a substantial increase in mean Amazon GPP.

    “GISS-E2-R has the smallest GSAT increase by the late 21st century, alongside slight precipitation increases and a substantial increase in mean Amazon GPP.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Findingown data

    Among the models with dynamic vegetation components, UKESM1-0-LL and HadGEM2-ES both show extensive Amazon dieback.

    “Of these, UKESM1-0-LL, which projects substantial temperature increases and decreases in precipitation, and HadGEM2-ES, with a large decrease in precipitation, both show extensive dieback.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    By the end of the 21st century, CanESM5 simulates the highest global land temperatures, leading to a large Amazon dieback area due to surface temperature increases.

    “By the end of the 21st century, CanESM5 simulates the highest global land temperatures, leading to a large Amazon dieback area due to surface temperature increases (Fig. S3a).”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Findingown data

    Land-use change occurs primarily on the periphery of the Amazon basin.

    “Further analysis of the two LUH phases data (Fig. S8) indicates that LUC occurs primarily on the periphery of the Amazon basin.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Findingown data

    MPI-ESM-LR, a model with a dynamic vegetation component, does not predict Amazon dieback but simulates a GPP decrease.

    “Another model with a dynamic vegetation component, MPI-ESM-LR, does not predict dieback but simulates GPP decrease (Fig. 1).”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Findingreasoning

    MIROC-ES2L projects localised dieback associated with mean annual surface air temperatures reaching 36.8 ± 0.6 °C, occurring later due to its lower global temperature trajectory.

    “Similarly, MIROC-ES2L projects localised dieback associated with mean annual surface air temperatures reaching 36.8 ± 0.6 °C, but this occurs later due to its lower global temperature trajectory.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Interpretationreasoning

    Models with more resilient vegetation or slower carbon cycle feedbacks may delay dieback initiation, leading to a more sudden loss of productivity once thresholds are crossed.

    “This suggests that models with more resilient vegetation or slower carbon cycle feedbacks may delay dieback initiation, leading to a more sudden loss of productivity once thresholds are crossed.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Interpretationreasoning

    Models with more resilient Amazon vegetation maintain GPP for longer but experience more abrupt losses at higher warming levels.

    “This suggests that models with more resilient Amazon vegetation maintain GPP for longer but experience more abrupt losses at higher warming levels.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Premisenone given

    Coupled ESMs' ecosystem components typically do not represent lateral ecological interactions between grid cells, so dieback in one location does not directly influence neighbouring cells through ecological interactions.

    “Furthermore, although coupled ESMs capture some cross-grid feedbacks through the atmosphere, their ecosystem components typically do not represent lateral ecological interactions between grid cells, meaning that dieback in one location does not directly influence neighbouring cells through ecological interactions.”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    Current models may underrepresent fire as a key accelerator of Amazon dieback.

    “The Global Tipping Points Report (2023) further emphasises that current models may underrepresent fire as a key accelerator of Amazon dieback 6 .”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [6] McKay, D. I. A. et al. Section 1. Earth system tipping points. Chapter 1.3 Tipping points in the biosphere. in The Global Tipping Points Report 2023 (University of Exeter, 2023).

      Couldn’t be matched to a paper in Crossref, so it can’t be checked.

  • May combine two claims.

    Interpretationown data

    The increased precipitation projected by BCC-CSM1-1 in these grid cells is likely due to a northward ITCZ shift, creating more humid conditions that prevent dieback.

    “In contrast, BCC-CSM1-1 projects increased precipitation in these grid cells, likely due to a northward ITCZ shift (Figs.S3a, S7), creating more humid conditions that prevent dieback.”
    Quote in the paperSubstantial claim

  • Classified as probably not a substantial claim.

    Study design and conductcitation

    A Savitzky-Golay filter with a 10-year window and a second-order polynomial was applied to the GPP series to reduce the impact of interannual variability.

    “To reduce the impact of interannual variability, which often masks long-term vegetation trends, we applied a Savitzky-Golay filter 65 with a 10-year window and a second-order polynomial.”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites [65] Gallagher, N. B. Savitzky-golay smoothing and differentiation filter. Eigenvector Research Incorporated, Manson, WA, USA, 2020 [Online]. Available: https://eigenvector.com/wp-content/uploads/2020/01/SavitzkyGolay.pdf (Aaccessed on 27 July 2025).

      Not checked in the free run.

  • May combine two claims.

    Interpretationreasoning

    Dynamic vegetation enables species adaptation to warmer, drier conditions, but under severe climate change plant functional types may not adjust quickly enough to prevent large-scale dieback.

    “Dynamic vegetation enables species adaptation to warmer, drier conditions, but under severe climate change, plant functional types may not adjust quickly enough to prevent large-scale dieback.”
    Quote in the paperSubstantial claim
38 of 38

  • The authors state this as an assumption.

    Land cover remains constant beyond 2100 under the RCP8.5 and SSP5-8.5 scenarios.

    “land cover is assumed to remain constant beyond 2100 41,42”
1 of 1

Self-check

Percentages, sums and reported differences re-computed from the paper’s own numbers.

Each contradiction between two passages listed here was confirmed by an independent check.

  • Figure 3 and Methods use different GPP baselines for dieback thresholds

    Major

    The Figure 3 caption says its climatic thresholds are measured when GPP reaches 20% of its maximum, whereas the Methods say the corresponding climatic conditions are extracted when GPP reaches 20% of its pre-industrial mean. Those baselines need not coincide.

  • Figure 4 period labels disagree with the stated classification periods

    Minor

    The Figure 4 caption describes a 50-year period around 2000, but the Methods list 1850-1899 as the first classification period and no period around 2000.

  • Conclusion reports a different uncertainty for the warming threshold

    Minor

    The conclusion gives the mean warming at dieback onset as 5.0 ± 1.9 °C, while the results and Table 2 give 5.0 ± 2.9 °C for the same quantity.

  • Dieback onset is not confined to the 21st century

    Minor

    The abstract places dieback onset within the 21st century across its stated warming range, but the results identify a model whose onset begins in the second half of the 22nd century.

4 of 4

Cites

Flagged first, then by citations

Research works printed but not linked: 2

  1. McKay, D. I. A. et al. Section 1. Earth system tipping points. Chapter 1.3 Tipping points in the biosphere. in The Global Tipping Points Report 2023 (University of Exeter, 2023).
  2. Eva, H. D. et al. A Proposal for Defining the Geographical Boundaries of Amazonia; Synthesis of the Results from an Expert Consultation Workshop Organized by the European Commission in Collaboration with the Amazon Cooperation Treaty Organization-JRC Ispra, 7-8 June 2005 (European Commission, 2005).

Cited by

Sorted by citations

Related

Authors
Melnikova, I., Hajima, T., Shiogama, H., Hayashi, M., Ito, A., Nishina, K., Tachiiri, K., & Yokohata, T.
Affiliations
  • National Institute for Environmental Studies (JP)
  • Japan Agency for Marine-Earth Science and Technology (JP)
  • The University of Tokyo (JP)
Corresponding author
Melnikova, Irina
ORCID
Venue
Communications Earth & Environment
Publisher
Springer Science and Business Media LLC
Published
2025-08-20
Volume
6
Issue
1
ISSN
2662-4435 (ISSN-L)
DOI
10.1038/s43247-025-02606-5
OpenAlex
W4413337866
Semantic Scholar
58875bc2692c3df559159e27438608c37d491294
Cited by
5
References
74
Citations by year
2026: 5
FWCI
2.06
Type
journal article
Language
en
Open access
Licence
Locations
  • Communications Earth & Environment (publishedVersion, cc-by-nc-nd)
  • DOAJ (DOAJ: Directory of Open Access Journals) (submittedVersion, cc-by-sa)
Funders
  • Environmental Restoration and Conservation Agency (JP-MEERF20242001)
  • Ministry of Education, Culture, Sports, Science and Technology (JPMXD0722681344, JP24K20979)
  • Ministry of Environment
  • Japan Society for the Promotion of Science (24K20979, JPMXD0722681344)
Topics
  • Physical Sciences › Environmental Science › Global and Planetary Change › Plant Water Relations and Carbon Dynamics
  • Physical Sciences › Environmental Science › Global and Planetary Change › Climate variability and models
  • Physical Sciences › Environmental Science › Global and Planetary Change › Ecosystem dynamics and resilience
Keywords
climate change projections, Earth system models, high-emission scenario, precipitation thresholds, land use change, atmospheric circulation changes

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

Claims read by deepseek-v4.1-flash under definition version 3 (extract-v3) on 2026-11-03.

Classified by jev-1.13-20260917 (gold-guideline-v0.2) on 2026-11-03.

Cited sources are checked by gpt-6-sol.

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review