CoChat
Sign Up
Cited by 9Oceanographic and Atmospheric ProcessesClimate variability and modelsMeteorological Phenomena and Simulations

Examining the ability of CMIP6 models to reproduce the upwelling SST imprint in the Eastern Boundary Upwelling Systems

Open access
Varela, Rubén, DeCastro, Maite, Rodriguez-Diaz, Laura, Dias, João Miguel, Gómez-Gesteira, Moncho·2022·Journal of Marine Science and Engineering
DOI 10.3390/jmse10121970
Open this paper in CoChat — free accountOpen PDF

Abstract

Knowing future changes in the sea surface temperature (SST) is of vital importance since they can affect marine ecosystems, especially in areas of high productivity such as the Eastern Boundary Upwelling Systems (EBUS). In this sense, it is key to have fine resolution models to study the SST patterns as close as possible to the coast where the upwelling influence is greater. Thus, the main objective of the present work is to assess the ability of 23 General Circulation Models (GCMs) from phase six of the Coupled Model Intercomparison Project (CMIP6) in reproducing the upwelling SST imprint in the EBUS through a comparison with the Optimum Interpolation of Sea Surface Temperature (OISST ¼) database of the National Oceanic and Atmospheric Administration for the common period of 1982–2014. The results have shown that most of the CMIP6 GCMs overestimate nearshore SST for all the EBUS with the exception of Canary. Overall, the models with better resolution showed lower Normalized Root Mean Squared Error (NRMSE) and Normalized Bias (NBias), although the ability of the models is dependent on the study area. Thus, the most suitable models for each EBUS are the CNRM-HR, GFDL-CM4, HadGEM-MM, CMCC-VHR4, and EC-Earth3P for Canary; CESM1-HR, CMCC-VHR4, ECMWF-HR, and HadGEM-HM for Humboldt; and HadGEM-HH and HadGEM-HM for California. In the case of Benguela, no model adequately reproduces the SST imprint under the conditions established in the present study.

AI snapshot

Written from the paper’s full text.

TL;DR

This study compares 23 CMIP6 GCMs with NOAA OISST 1/4 to assess how well they reproduce the sea surface temperature imprint of coastal upwelling in the four main Eastern Boundary Upwelling Systems over 1982–2014. Most models overestimate nearshore SST except in the Canary system, and higher-resolution models generally perform better, but skill is strongly region-dependent and no model adequately reproduces the Benguela imprint under the study’s criteria.

Key findings

  • Most CMIP6 GCMs overestimate nearshore SST for all EBUS except Canary; some models exceed OISST 1/4 by more than 4 °C. For Canary, practically all models show similar SST to OISST 1/4, with slight underestimation for most models.
  • Coastal NRMSE and NBias are clearly higher than oceanic values for each upwelling system. Benguela shows the greatest difficulties (most models >10% and even >20%); Canary shows the smallest errors (almost all models within ±10%, with the best within ±5%); Humboldt shows moderate errors (<10% for hist-1950 models, while historical models tend to exceed that); California behaves similarly to Benguela, with NRMSE and NBias >20%.
  • Higher-resolution models (hist-1950 experiment) generally show lower NRMSE and NBias than historical models, except in the Canary system, where almost all models adequately reproduce the upwelling influence on SST regardless of resolution.
  • Most suitable models by system: Canary—CNRM-HR (3, 13), GFDL-CM4 (4), HadGEM-MM (6), CMCC-VHR4 (12), EC-Earth3P (14); Humboldt—CESM1-HR (10), CMCC-VHR4 (12), ECMWF-HR (16), HadGEM-HM (20); California—HadGEM-HH and HadGEM-HM (19, 20); Benguela—no model adequately reproduces the SST imprint under the conditions established.

Methods

  • Monthly SST from 23 CMIP6 GCMs (Historical and Hist-1950 experiments, including HighResMIP; NICAM16-8S as a control) was compared with NOAA OISST 1/4 (0.25°) for the common period 1982–2014. Data were bilinearly interpolated to a common 0.25° × 0.25° grid.
  • Coastal points were the grid points closest to the coast; oceanic points were 3° offshore in the direction perpendicular to the coast. Satellite and model data were averaged over 1982–2014. Validation used normalized root mean square error (NRMSE) and normalized bias (NBias), with smaller values indicating better reproduction. Seasonal analysis is provided in the Supplementary Material.

Limitations

  • Stated: OISST 1/4 may overestimate coastal SST and therefore underestimate the thermal imprint of upwelling; because CMIP6 SSTs are even higher than OISST 1/4, most CMIP6 GCMs greatly overestimate coastal SST.
  • Stated: most CMIP6 GCMs have difficulties reproducing the nearshore upwelling SST imprint, with coastal NRMSE/NBias higher than oceanic values; model skill is strongly dependent on the study area, and no Benguela model met the criteria.
  • Evident: no dedicated limitations section is provided; the analysis uses monthly/annual averages over a common 1982–2014 period and relies on OISST 1/4 as the reference, so results are conditioned on that reference and period.

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

Integrity signals

Registry

Crossref · 10.3390/jmse10121970

Cites

67 of 67 references verified0 citation errors · 0 not checkable

Statistics

Nothing to recompute.

Claims the paper makes

7 of 17 cited sources checked · 2 supported · 1 flagged · accuracy not yet measured

  • Findingown data

    Contrary to Wang et al. (2022), we did not find important differences considering the seasonal performance of the CMIP6 GCMs under study.

    “However, contrary to Wang et al. (2022) [57], we did not find important differences considering the seasonal performance of the CMIP6 GCMs under study.”
    Quote in the paperNot a substantial claim1 of 1 sources checked
    • cites Wang et al. (2022) [57] source
  • Findingcitation

    The CAUS OISST 1/4 maps show lower SST along the coast than in the open ocean.

    “The CAUS OISST 1 /4 maps show lower SST along the coast than in the open ocean (Figure 8) as previously observed by different authors [8,46].”
    Quote in the paperSubstantial claim2 of 2 sources checked
    • cites [8] Seabra, 2019, Reduced Nearshore Warming Associated With Eastern Boundary Upwelling Systems, Front. Mar. Sci. source

      The Introduction says that upwelled cold water shapes the thermal environment of upwelling regions, and Materials and Methods uses quarter-degree OISST data. But the paper’s coastal-to-offshore comparisons concern SST warming rates, not the absolute SST difference shown in the citing document’s CAUS maps. It therefore supports the general context, but not that specific map observation. (LLM, conf=0.78)

      Experimental — accuracy not yet measured
    • cites [46] Pardo, 2011, Evolution of upwelling systems coupled to the long-term variability in sea surface temperature and Ekman transport, Clim. Res. source

      The Introduction (page unavailable) explains that coastal upwelling brings cooler water to a narrow region near the coast and identifies the California Current as an upwelling system. That supports the general observation of cool coastal water, but the available text does not explicitly compare California coastal SST with open-ocean SST. The citing paper’s OISST maps are its own, not a result attributed to this source. (LLM, conf=0.76)

      Experimental — accuracy not yet measured
  • Premisecitation

    Oceanic points situated 3 degrees away from the coast are far enough to not be affected by upwelling.

    “Thus, oceanic points are situated far enough from the coast to not be affected by upwelling [7,15].”
    Quote in the paperSubstantial claim1 of 2 sources checked
    • cites [7] Varela, 2018, Coastal warming and wind-driven upwelling: A global analysis, Sci. Total. Environ. source
    • cites [15] Varela, 2021, Influence of Canary upwelling system on coastal SST warming along the 21st century using CMIP6 GCMs, Glob. Planet. Chang. source
  • Premisecitation

    The greatest influence of upwelling on SST patterns in the EBUS occurs in the first 100 km from the coast.

    “the greatest influence of upwelling on the SST patterns in the EBUS occurs in the first 100 km [8].”
    • Also stated in: § 1. Introduction · plain
      “which usually takes place in the first 100 km from the coast [8].”
    Quote in the paperSubstantial claim1 of 1 sources checked
    • cites [8] Seabra, 2019, Reduced Nearshore Warming Associated With Eastern Boundary Upwelling Systems, Front. Mar. Sci. source

      The Results and Figure 2 caption report that SST warming was slowest nearest shore in EBUS, supporting a stronger nearshore thermal signal. They do not identify the first 100 km as the zone of greatest upwelling influence; the analysis covers a 500-km coastal band and describes warming-rate patterns rather than SST patterns generally. Page numbers are unavailable in our copy. (LLM, conf=0.84)

      Experimental — accuracy not yet measured
  • Findingown data

    In the CAUS, oceanic values remain within acceptable margins compared to satellite data.

    “However, oceanic values remain within acceptable margins compared to satellite data (Figure 9b).”
    Quote in the paperNot a substantial claim
  • May combine two claims. 15
  • The quote only partly states this claim. 1
  • Classified as probably not a substantial claim. 5

  • May combine two claims.

    Findingcitation

    This improvement is visible for all EBUS with the exception of CUS, where most models adequately reproduce the influence of upwelling on SST patterns for both experiments independently of resolution.

    “In fact, this improvement is visible for all EBUS with the exception of CUS, where most of models adequately reproduce the influence of upwelling on the SST patterns for both experiments independently of the resolution as Sylla et al. (2022) [53] found in their study.”
    Quote in the paperSubstantial claim1 of 1 sources checked
    • cites Sylla et al. (2022) [53] source

      Section 4.1 (page not available) finds that higher-resolution models represent the SST-based upwelling index more realistically in the Moroccan subregions of the CUS; lower-resolution models markedly underestimate its amplitude in the southern Moroccan subregion. The Summary and conclusions likewise report clear improvement in the southern CUS, contrary to the claim that CUS is an exception whose representation is adequate independently of resolution. The study evaluates CUS, not improvements across all EBUS. (LLM, conf=0.88)

      Checked by a reasoning model — may be wrong.

      Experimental — accuracy not yet measured

  • May combine two claims.

    Premisecitation

    Only Varela et al. (2022) studied coastal SST warming for the Canary Upwelling System using 6 CMIP6 GCMs, leaving the rest of the main EBUS unexplored.

    “Only Varela et al. (2022) [15] studied coastal SST warming for the Canary Upwelling System using 6 CMIP6 GCMs.”
    • Also stated in: § 1. Introduction · plain
      “However, the work was only focused on the Canary Upwelling System, leaving the rest of the main EBUS unexplored.”
    Quote in the paperSubstantial claim1 of 1 sources checked
    • cites Varela et al. (2022) [15] source

      Section 2.1 confirms that the paper used SST data from six CMIP6 models for the Canary upwelling system. The Introduction discusses earlier work on coastal SST and other upwelling systems, but does not establish that this was the only such study or that the other main EBUS were unexplored. Page numbers are not available in our copy. (LLM, conf=0.87)

      Experimental — accuracy not yet measured

  • May combine two claims.

    Findingown data

    In the CAUS, almost all CMIP6 GCMs overestimate coastal SST, with half of them exceeding the error range of OISST 1/4.

    “Regarding the performance of the CMIP6 GCMs to reproduce the coastal and oceanic values provided by OISST 1 /4, almost all models overestimate coastal SST with half of them exceeding the error range of OISST 1 /4 (Figure 9a).”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    In the CUS, practically all CMIP6 GCMs show similar SST values to OISST 1/4 both for the coast and the ocean, displaying a slight underestimation for most models.

    “Regarding the ability of each CMIP6 GCM to reproduce the SST values of OISST 1 /4, practically all the CMIP6 GCMs show similar SST values to OISST 1 /4 both for the coast (Figure 5a) and the ocean (Figure 5b), displaying a slight underestimation for most models.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    The smallest NRMSE and NBias are observed for the CUS, with almost all models within the interval ±10%.

    “On the other hand, the smallest NRMSE and NBias are observed for the CUS with almost all models within the interval ±10%.”
    Quote in the paperSubstantial claim
    • Rests on an assumption: The NRMSE and NBias thresholds used to classify model adequacy (e.g., ±5%, ±10%, >20%) are appropriate for judging whether a CMIP6 GCM adequately reproduces OISST SST.

  • May combine two claims.

    Findingown data

    In the BUS, most CMIP6 GCMs show the SST imprint of upwelling, although in many cases the cold coastal area is smaller than observed for OISST 1/4.

    “On the one hand, most models show the SST imprint of upwelling, although in many cases the cold coastal area is smaller than observed for OISST 1 /4.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    The HUS OISST 1/4 map shows the influence of upwelling on SST in the southernmost part of the domain (~38°S), around 24°S in the area of Chile, and in Peru between 17°S and 13°S.

    “The HUS OISST ¼ map shows the influence of upwelling on SST in the southernmost part of the domain (~38° S), around 24° S in the area of Chile, and in Peru between 17° S and 13° S (Figure 6).”
    • Also stated in: § Supplementary Materials: · plain
      “The HUS OISST 1 /4 map shows the influence of upwelling on SST in the southernmost part of the domain (~38 • S), around 24 • S in the area of Chile, and in Peru between 17 • S and 13 • S (Figure 6).”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Findingown data

    The OISST 1/4 map for the BUS shows a clear gradient between coastal and oceanic temperature, with lower coastal SST (14-16°C) than in the open ocean (over 18°C), and a clear upwelling imprint from 35°S to 18°S.

    “The OISST 1 /4 map shows a clear gradient between coastal and oceanic temperature, with a lower coastal SST (between 14-16 • C) than in the open ocean (over 18 • C).Thus, a clear upwelling imprint is evident from 35 • S to 18 • S.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    In the CAUS, most CMIP6 GCMs display lower SST values along the coast than in the ocean, occupying a less extensive region than provided by OISST 1/4.

    “Most of them display lower SST values along the coast than in the ocean, occupying a less extensive region than provided by OISST 1 /4.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    For the CUS, the SST pattern for OISST 1/4 and most CMIP6 GCMs is very similar, with the influence of upwelling evident along coastal zones and causing lower temperatures from 20°N to 34°N.

    “The SST pattern for OISST 1 /4 and most of the CMIP6 GCMs is very similar for the CUS (Figure 4).In both cases, the influence of the upwelling on SST patterns is evident along the coastal zones, causing lower temperatures from 20 • N to 34 • N.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    In the HUS, moderate NRMSE and NBias, smaller than 10%, are obtained for the hist-1950 models while the historical models tend to exceed that value.

    “Regarding the HUS, moderate NRMSE and NBias, smaller than 10%, are obtained for the hist-1950 models while the historical models tend to exceed that value.”
    Quote in the paperNot a substantial claim
    • Rests on an assumption: The NRMSE and NBias thresholds used to classify model adequacy (e.g., ±5%, ±10%, >20%) are appropriate for judging whether a CMIP6 GCM adequately reproduces OISST SST.

  • The quote only partly states this claim.

    Findingown data

    In the HUS, only a few CMIP6 GCMs (CMCC-HR4 and HadGEM-MM) do not show the upwelling imprint in the area.

    “Only a few CMIP6 GCMs (CMCC-HR4, historical and hist-1950 (2 and 11), and HadGEM-MM (6)) do not show the upwelling imprint in the area.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Findingcitation

    Both at an annual and seasonal scale, CMIP6 GCMs tend to overestimate coastal and oceanic SST with respect to OISST 1/4, although this overestimation is much more pronounced nearshore, especially in the BUS and CAUS.

    “Both at an annual and seasonal scale, CMIP6 GCMs tend to overestimate coastal and oceanic SST with respect to OISST 1 /4, although this overestimation is much more pronounced nearshore, especially in the BUS and CAUS, as pointed out by previous authors [56].”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites [56] Liu, 2022, An ocean perspective on CMIP6 climate model evaluations, Deep. Sea Res. Part II Top. Stud. Oceanogr. source

  • Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based).

    Interpretationreasoning

    The lower NRMSE and NBias of hist-1950 models are probably due to resolution improvement.

    “Moreover, it is also evident that those models belonging to the hist-1950 experiment show lower NRMSE and NBias than those from the historical experiment, probably due to the resolution improvement (Table 1).”
    Quote in the paperNot a substantial claim

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

    Premisecitation

    Dufois et al. (2012) found warm bias in the monthly Pathfinder data during summer in the EBUS.

    “Moreover, Dufois et al. (2012) [10] found warm bias in the monthly Pathfinder data during summer in the Eastern Boundary Upwelling Systems (EBUS), where high SST gradients exist.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites Dufois et al. (2012) [10] source

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

    Study design and conductcitation

    The selection of coastal and oceanic points was made following previous studies (Varela et al., 2018).

    “The selection of both coastal and oceanic points was made following previous studies (Varela et al., 2018 [7], and the references therein).”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites Varela et al., 2018 [7] source

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

    Premisecitation

    Gutiérrez et al. (2011) obtained similar results for the summer SST on the Peruvian coast from 1985 to 2005 using Pathfinder High Resolution data (~4 km).

    “Gutiérrez et al. (2011) [45] obtained similar results for the summer SST on the Peruvian coast from 1985 to 2005 using Pathfinder High Resolution data (~4 km).”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites Gutiérrez et al. (2011) [45] source

  • May combine two claims.

    Premisecitation

    CMIP5 exhibits warm SST biases in EBUS, making it difficult to realistically reproduce the influence of upwelling on SST.

    “Second, CMIP5 exhibits warm SST biases in EBUS [23][24][25], making it difficult to realistically reproduce the influence of upwelling on SST.”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites [23][24][25]

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

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

    Premisecitation

    EBUS can act as thermal refugia, reducing warming and even causing cooling trends in the areas affected by this mechanism.

    “It is a known fact that EBUS can act as thermal refugia, reducing warming and even causing cooling trends in the areas affected by this mechanism [7].”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [7] Varela, 2018, Coastal warming and wind-driven upwelling: A global analysis, Sci. Total. Environ. source

  • May combine two claims.

    Premisecitation

    Meneghesso et al. (2020) found that Level 4 High Resolution Sea Surface Temperature (GHRSST) products can cause an overestimation of coastal SST and an underestimation of the thermal imprint of upwelling.

    “Meneghesso et al. (2020) [9] found that Level 4 High Resolution Sea Surface Temperature (GHRSST) products can cause an overestimation of coastal SST and, thus, an underestimation of the thermal imprint of upwelling.”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites Meneghesso et al. (2020) [9] source
20 of 20

  • The paper relies on this without support.

    The NRMSE and NBias thresholds used to classify model adequacy (e.g., ±5%, ±10%, >20%) are appropriate for judging whether a CMIP6 GCM adequately reproduces OISST SST.

    “with NRMSE and NBias values within the interval ±5%.”
    “almost all models within the interval ±10%.”
    “with NRMSE and NBias values greater than 20%.”
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.

  • Hist-1950 ocean resolutions are not uniformly 0.25 degrees

    Minor

    The methods describe Hist-1950 simulations as having an ocean resolution of 0.25 degrees, but Table 1 lists several other resolutions, including 1 degree for EC-Earth3P.

  • Six models are listed after a claim of seven

    Minor

    The Benguela results say seven GCMs adequately reproduce nearshore SST, but the accompanying list names only six: two ECMWF models, FGOALS, two HadGEM models, and MPI-XR.

2 of 2

Cites

Flagged first, then by citations

Cited by

Sorted by citations

Related

Authors
Varela, R., DeCastro, M., Rodriguez-Diaz, L., Dias, J. M., & Gómez-Gesteira, M.
Affiliations
  • CESAM—Centre for Environmental and Marine Studies, Department of Physics, University of Aveiro, 3810-193 Aveiro, Portugal
  • EphysLab—Environmental Physics Laboratory, CIM-UVIGO, Universidade de Vigo, Edificio Campus da Auga, 32004 Ourense, Spain
ORCID
Venue
Journal of Marine Science and Engineering
Publisher
MDPI AG
Published
2022-12-11
Volume
10
Issue
12
Pages
1970
ISSN
2077-1312 (ISSN-L)
DOI
10.3390/jmse10121970
OpenAlex
W4312126131
Semantic Scholar
a400f05fcf5e701e993a031de0dffa0be95c7248
Cited by
9
References
67
Citations by year
2025: 3 · 2024: 5 · 2023: 1
FWCI
1.10
Type
journal article
Language
en
Open access
Licence
Locations
  • Journal of Marine Science and Engineering (publishedVersion, cc-by)
  • DOAJ (DOAJ: Directory of Open Access Journals) (submittedVersion, cc-by-sa)
  • Investigo Institutional repository of UVigo (Universidade de Vigo) (submittedVersion, cc-by)
Funders
  • Programa de Consolidación e Estructuración de Unidades de Investigación Competitivas (ED431C 2021/44, UIDP/50017/2020+UIDB/50017/2020+LA/P/0094/2020, MAR-02.01.01-FEAMP-0022, PRTR-C17.I1, ED481B-2021-108)
  • FCT/MCTES (ED431C 2021/44, UIDP/50017/2020+UIDB/50017/2020+LA/P/0094/2020, MAR-02.01.01-FEAMP-0022, PRTR-C17.I1, ED481B-2021-108)
  • AquiMap (ED431C 2021/44, UIDP/50017/2020+UIDB/50017/2020+LA/P/0094/2020, MAR-02.01.01-FEAMP-0022, PRTR-C17.I1, ED481B-2021-108)
  • MAR2020 Program (ED431C 2021/44, UIDP/50017/2020+UIDB/50017/2020+LA/P/0094/2020, MAR-02.01.01-FEAMP-0022, PRTR-C17.I1, ED481B-2021-108)
  • Portugal 2020 (ED431C 2021/44, UIDP/50017/2020+UIDB/50017/2020+LA/P/0094/2020, MAR-02.01.01-FEAMP-0022, PRTR-C17.I1, ED481B-2021-108)
  • European Union (ED431C 2021/44, UIDP/50017/2020+UIDB/50017/2020+LA/P/0094/2020, MAR-02.01.01-FEAMP-0022, PRTR-C17.I1, ED481B-2021-108)
  • European Union NextGenerationEU (ED431C 2021/44, UIDP/50017/2020+UIDB/50017/2020+LA/P/0094/2020, MAR-02.01.01-FEAMP-0022, PRTR-C17.I1, ED481B-2021-108)
  • European Maritime and Fisheries (ED431C 2021/44, UIDP/50017/2020+UIDB/50017/2020+LA/P/0094/2020, MAR-02.01.01-FEAMP-0022, PRTR-C17.I1, ED481B-2021-108)
  • Xunta de Galicia (ED431C 2021/44, UIDP/50017/2020+UIDB/50017/2020+LA/P/0094/2020, MAR-02.01.01-FEAMP-0022, PRTR-C17.I1, ED481B-2021-108)
Topics
  • Physical Sciences › Earth and Planetary Sciences › Oceanography › Oceanographic and Atmospheric Processes
  • Physical Sciences › Environmental Science › Global and Planetary Change › Climate variability and models
  • Physical Sciences › Earth and Planetary Sciences › Atmospheric Science › Meteorological Phenomena and Simulations
Keywords
eastern boundary upwelling systems, sea surface temperature, upwelling, CMIP6, global climate models, model evaluation, OISST, model resolution, Benguela upwelling, California Current

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 2026-10-16.

Classified by jev-1.13-20260917 (gold-guideline-v0.2) on 2026-10-16.

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