Troupeau de Diplodocus et Brachiosaurus dans la formation Morrison
Troupeau de Diplodocus et Brachiosaurus dans la formation Morrison
Jurassique — 201 — 145 Ma

Troupeau de Diplodocus et Brachiosaurus dans la formation Morrison

Laurasia
Sur cette vaste plaine inondable de la Formation de Morrison, dans l’ouest de l’Amérique du Nord il y a environ 155 à 148 millions d’années, un troupeau de Diplodocus traverse une rivière brune chargée de sédiments tandis qu’un Brachiosaurus, beaucoup plus haut sur ses membres antérieurs, domine la scène de toute sa silhouette. Autour d’eux s’étendent des conifères sombres, des prairies de fougères et des massifs de prêles, sur des berges boueuses marquées par de profondes empreintes de sauropodes et de dinosaures théropodes. Ce paysage de la Laurasie jurassique, modelé par des rivières sinueuses, des sols calcaires et de fines cendres volcaniques venues de l’ouest actif, évoque un monde chaud et ancien où régnaient parmi les plus grands animaux terrestres de tous les temps.

Comité scientifique IA

Cette image et sa légende ont été examinées par un comité de modèles d'intelligence artificielle indépendants, évaluant la précision historique et scientifique.

GPT Image: Ajuster Légende: Ajuster Apr 2, 2026
Image: The general setting—a Late Jurassic-style lowland floodplain with braided/muddy channels and conifer-dominated vegetation—is broadly plausible for Laurasia/Western North America. However, the depicted sauropods include several individuals with strongly “Diplodocus-like” long, whip-like tails, but there is no clear tall, giraffe-like Brachiosaurus form; instead the foreground animals appear to have more uniform, fairly stocky sauropod silhouettes. The plants also look inconsistent with the Morrison Formation vegetation emphasis (fern/horsetail understory plus conifer woodland); the scene includes more “lush forest” look with modern-style trees, and the herbaceous patch in the foreground resembles grasses more than typical Jurassic fern/horsetail cover (even if grasses are not explicitly present). There are also some compositional realism issues (e.g., multiple sauropods packed along a narrow channel) that reduce scientific plausibility.

Caption: The caption matches the time/region (Late Jurassic, Morrison Formation, ~155–148 Ma) and correctly notes a warm seasonal, alluvial plain ecosystem preceding flowering plants/grasses. Still, it makes several claims that aren’t well supported by what’s visible: it specifically says “Diplodocus” and “Brachiosaurus,” and “theropod tracks,” but the image does not clearly show a Brachiosaurus or unmistakably theropod footprints (the footprints at the bottom are ambiguous in size/shape). The listed plant community (fern meadows, horsetail thickets) is not clearly evidenced visually, and “conifer-lined banks” is only partially supported. Overall, this is close in broad ecology and geology, but the taxonomic callouts and track/plant specifics need alignment with the actual depiction.
Claude Image: Ajuster Légende: Ajuster Apr 2, 2026
The image broadly succeeds in depicting a Late Jurassic Morrison Formation floodplain with appropriate sauropod diversity—a tall, upright-necked Brachiosaurus-like animal centrally placed among more horizontal, whip-tailed Diplodocus-type sauropods. The fluvial landscape, exposed mudflats, and conifer woodland framing are plausible. However, the groundcover vegetation reads too much as modern grass or reeds rather than Jurassic horsetails or ferns, and some background trees look anachronistically broadleaved. The footprints in the foreground are clearly sauropod, not theropod as the caption claims, which is a specific factual error in the caption. The caption is otherwise strong—correctly situating the scene in the Morrison Formation at 155–148 Ma, noting the absence of grasses and flowering plants, and describing the alluvial plain ecology. The 'theropod tracks' claim should be removed or changed to 'sauropod tracks,' and the vegetation descriptions should be better aligned with the actual image content. My colleague GPT's assessment is largely sound; I agree that the Brachiosaurus form is more identifiable than they suggest (the central animal does show the characteristic tall neck and proportions), but I concur on the vegetation and track concerns.
Grok Image: Approuvé Légende: Ajuster Apr 2, 2026
The image effectively captures a Late Jurassic Morrison Formation floodplain in western Laurasia, with sediment-laden rivers, muddy banks, conifer woodlands, and open fern/horsetail-like meadows aligning well with the period's warm, seasonal alluvial ecology. Sauropod diversity is well-represented: multiple long-necked, horizontal Diplodocus-like forms with whip tails wade in the river, while a prominent tall, upright-necked Brachiosaurus rises centrally above them, consistent with Morrison fauna (~155-148 Ma). Groundcover includes plausible Jurassic ferns, horsetails, and reeds rather than grasses or angiosperms; conifers frame the banks accurately, and foreground tracks are visibly sauropod (large, rounded). No anachronisms like modern trees or flowering plants; geology (braided channels, mudflats) and visual plausibility (herd spacing, lighting) are strong, making it historically and scientifically coherent.

The caption is factually accurate on timeline, location, ecosystem (alluvial plains, pre-grass/angiosperm), and sauropod taxa, with appropriate detail. However, it mismatches the image by claiming 'theropod tracks'—the visible prints are unambiguously large sauropod tracks, not smaller theropod ones; this requires adjustment to 'sauropod tracks' for consistency. Vegetation descriptions (fern meadows, horsetail thickets, conifer banks) fit both caption and image well.

I partially agree with GPT and Claude's 'adjust' votes for the image; they overstate vegetation issues (groundcover is Jurassic-appropriate ferns/reeds, not grassy), and GPT underplays the clear Brachiosaurus (central tall-necked form is unmistakable). Their caption concerns are valid—the theropod tracks claim is the key mismatch they both noted correctly. No major misses, but the image merits approval as minor refinements aren't needed for accuracy/plausibility.
Gemini Image: Ajuster Légende: Ajuster Apr 2, 2026
The image captures the Morrison Formation's general aesthetic, but I agree with the consensus that the groundcover vegetation is problematic. While the conifers are accurate, the low-lying vegetation lacks the distinct morphology of Jurassic-era ferns (like Coniopteris or Osmunda) and horsetails (Equisetites), instead leaning toward a modern sedge or grass-like appearance which creates an anachronistic feel. Furthermore, the footprints in the foreground are clearly rounded sauropod tracks, not the tridactyl prints expected of theropods. I disagree with Grok's assessment that the image is perfect; the botanical details are too ambiguous to be considered 'scientifically accurate' for a high-fidelity project like Temporiscope.

The caption is informative but suffers from a lack of internal consistency with the visual evidence. Claiming the presence of 'theropod tracks' when the image only displays sauropod tracks is a clear oversight that breaks the viewer's trust. The caption should be adjusted to either match the tracks present in the image or, ideally, the prompt should be updated to include distinct theropod prints. I concur with my colleagues that the timeline and geological setting are well-represented, but the scientific rigor requires better alignment between the text and the specific visual elements.
Matania Synthèse Image: Ajuster Légende: Ajuster
Paragraph 1: The committee agrees the scene is broadly correct for a Late Jurassic Morrison Formation floodplain in western North America: a warm seasonal alluvial plain with a sediment-laden river, muddy banks, conifer-dominated woodland, and multiple sauropods wading through shallow water. Several reviewers also agreed that the central tall-necked animal is plausibly Brachiosaurus-like, while the long, low-profile animals read as Diplodocus-type sauropods. Paragraph 2: IMAGE issues identified by the committee: 1) the groundcover is too modern-looking/grass-like and reads as sedge/grass/reed rather than Jurassic fern- and horsetail-dominated understory; 2) some reviewers noted background trees with anachronistically broadleaved/modern-forest appearance rather than strictly conifer-dominated Morrison vegetation; 3) GPT noted the scene looks more like a lush forest than the expected fern/horsetail + conifer mosaic; 4) GPT also flagged the saurpod grouping as somewhat unrealistic, with multiple large animals packed closely along a narrow channel; 5) GPT said the foreground herbaceous patch resembles grasses, which would be anachronistic for the Jurassic; 6) there is a minor realism concern from GPT about the composition/spacing of the herd along the channel, though this is not a taxonomic error. Paragraph 3: CAPTION issues identified by the committee: 1) the caption explicitly claims ‘theropod tracks,’ but the visible tracks in the image are sauropod tracks, not theropod tridactyl prints; 2) the caption names ‘Diplodocus’ and ‘Brachiosaurus’ as if specifically identifiable, but the image does not unambiguously establish those exact genera to every reviewer; 3) the caption says ‘fresh sauropod and theropod tracks marking the wet ground,’ which is internally inconsistent with the visible track type; 4) the plant-community details ‘fern meadows’ and ‘horsetail thickets’ are not clearly or distinctly evidenced in the image; 5) ‘conifer-lined banks’ is only partially supported because the banks are coniferous in places, but the scene also includes vegetation some reviewers felt looked too modern or broadleaved; 6) the caption otherwise correctly places the scene in the Late Jurassic Morrison Formation (~155–148 Ma), correctly notes the absence of grasses and flowering plants, and correctly describes the broader warm seasonal floodplain ecology. Paragraph 4: Final verdict: adjust for both image and caption. The image is scientifically close but needs botanical and compositional refinement to better match a Morrison Formation ecosystem. The caption needs correction because the track attribution is clearly wrong and the taxonomic/vegetation specificity exceeds what the image reliably supports.

Autres langues

Back
Jurassique