Fire-scarred fossil tree from the Late Triassic shows a pre-fire drought signal

Evidence from fossil charcoal and pyrolytic polycyclic aromatic hydrocarbons indicates that wildfires are an ancient ecological phenomenon, with studies suggesting increasing wildfire activity through the Late Triassic and across the Triassic–Jurassic boundary. Proposed drivers include increasing aridity, climate warming, and elevated atmospheric CO2, potentially shifting vegetation toward more flammable taxa. However, such proxies do not directly resolve ancient fire regimes (frequency, severity, size, seasonality, spatial patterns). Fire-adapted traits evolve under specific fire regimes, with different trait suites (e.g., serotiny vs thick bark and self-pruning) associated with crown- vs surface-fire regimes. Analyses of fire scars in modern trees provide diagnostic external and internal features (including bands of compressed/distorted tracheids) that can reconstruct fire history. Modern conifers often show an immediate post-fire reduction in tracheid size and subsequent growth release via increased ring width and tracheid dimensions. Drought and fire are linked in modern systems; trees physiologically respond to water stress by producing smaller, thicker-walled tracheids that increase cavitation resistance. This study analyzes the microscopic wood anatomy of a Late Triassic fossil trunk with clear external fire-scar features from Petrified Forest National Park to infer environmental conditions immediately preceding the paleowildfire. The aims were to test whether drought conditions preceded the fire and whether the tree’s anatomical response to fire matches modern analogs. The hypotheses were: (1) pre-scarring tracheids would exhibit a drought signal; (2) a scar-associated band of compressed/distorted tracheids would be present; (3) immediate post-scarring tracheids would be reduced in size; and (4) tracheid size would increase after recovery.

Multiple lines of paleoenvironmental evidence (charcoal, PAHs) document Triassic wildfires, with several basins showing increased fire activity near the Triassic–Jurassic transition. Paleosol geochemistry and palynology indicate growing aridity and floral turnover in the Chinle Formation during the Late Triassic. In modern systems, fire-scar analyses elucidate fire regimes and reveal characteristic internal anatomical responses, including compressed/distorted tracheids at the wound margin and temporary post-fire reductions in tracheid size, followed by growth release. Growth releases after fire in gymnosperms arise from increased tracheid size and/or number, reflecting reduced competition and nutrient pulses. Drought–fire linkages are well-established, and tree-ring and tracheid features are sensitive proxies for water stress; smaller, thicker-walled tracheids enhance cavitation resistance under drought. Evolutionary and phylogenetic work indicates long-term associations between fire and plant traits, with different strategies (thick bark, self-pruning, serotiny, epicormic sprouting) arising under distinct fire regimes, though direct Triassic fossil evidence for such traits remains scarce.

Field reconnaissance (October 2013, April 2014) across five major fossil log exposures in Petrified Forest National Park identified 13 fossil log segments with external morphology resembling modern fire scars (Black Forest: 3; Rainbow Forest: 4; Crystal Forest: 4; Jasper Forest: 2; Blue Mesa: 0). The analyzed specimen (Agathoxylon arizonicum; PEFO 40757) was collected from the Late Triassic Black Forest Bed. Preparatory work included cutting and polishing large cross-sections; a 2-cm-thick cross-section containing woundwood overgrowth and the scar-associated band was prepared for microscopy. Photomicrographs were acquired along nine 22-mm transects perpendicular to the scar-associated band, spanning pre- and post-scarring wood. Three transects were positioned in each of three tangential regions relative to the wound margin: close (~1 cm), middle (~3 cm), and far (~5 cm). A total of 12,410 tracheids were measured using the tgram R package, recording radial tracheid diameter (TD), radial lumen diameter (LD), and cell wall thickness (CWT), and indexing each tracheid’s relative position from the edge of the scar-associated band: pre-scarring positions −140 to −1 and post-scarring positions 1 to 140 (position 0 at band edge). Linear mixed models tested: (i) a pre-fire drought signal by comparing immediate pre-scarring (−20 to −1) vs early pre-scarring (−40 to −21) and average/normal pre-scarring (−140 to −61); and (ii) post-fire responses by comparing immediate post-scarring (1 to 40) and later post-scarring (61 to 140) vs average/normal pre-scarring (−140 to −61), stratified by tangential region (close, middle, far). Permits, curation, and imaging were conducted under PEFO-2014-SCI-0002; specimen curated at Petrified Forest National Park (PEFO 40757).

• External morphology: The specimen exhibits classic fire-scar features (basal location, woundwood overgrowth, “cat face” arches). Nearby logs showed similar features, including apparent re-burning of exposed scar tissue.
• Scar-associated band: A distinct light-colored band at the scar radius comprises compressed and distorted tracheids, matching modern fire-wounding responses.
• Pre-fire drought signal: Tracheids formed immediately before scarring (positions −20 to −1) show 6–12% reductions in tracheid and lumen diameters relative to both earlier pre-scarring intervals (−40 to −21) and average pre-scarring conditions (−140 to −61), consistent with water stress.
• Immediate post-fire response: Tracheids in positions 1 to 40 exhibit a 5–18% reduction in tracheid and lumen diameters compared to average pre-scarring conditions, indicating a recovery period following wounding.
• Post-recovery growth increase: After recovery, larger tracheid (≈6%) and lumen (≈7%) diameters were detected in the far tangential region, indicating localized growth release analogous to modern conifers after fire.
• Context: The combined features suggest a low-severity surface-fire regime in the Late Triassic Black Forest landscape, with drought preceding fire events.

The fossil trunk’s external and internal anatomical features parallel those of modern fire-scarred conifers: a scar-associated band of compressed/distorted tracheids at the wound radius, an immediate post-fire reduction in tracheid dimensions during recovery, and subsequent localized growth release reflected in increased tracheid sizes. Crucially, the immediate pre-scarring reduction in tracheid and lumen diameters indicates drought stress preceding the fire, aligning with modern observations that drought often precedes and promotes low-severity fires. The presence of multiple nearby specimens with fire-scar-like morphology further supports the occurrence of surface-fire regimes in this Late Triassic forest. While regional paleoenvironmental records indicate broader trends toward increased aridity and fire activity in the Late Triassic, the specific drought–fire sequence observed here occurs on shorter timescales (years) and does not require direct linkage to long-term climatic trends. The results imply that environmental conditions favoring the evolution of fire-adapted traits existed around 210 Ma. Direct fossil evidence for such traits (e.g., thick bark, self-pruning, serotiny, epicormic buds) in this flora is limited and often inconclusive, although some indications (bark fragments, potential epicormic structures) exist. Integrating fossil searches with phylogenetic reconstructions may further clarify the timing and emergence of fire adaptations.

This study provides direct anatomical evidence from a Late Triassic fossil tree that drought conditions preceded a fire event, as indicated by reduced tracheid and lumen diameters immediately before scarring, a scar-associated band of compressed/distorted tracheids at the fire event, an immediate post-fire reduction in tracheid size during recovery, and a subsequent localized growth release. These findings mirror responses in modern conifers and suggest that low-severity surface fires and associated selective pressures were present ca. 210 Ma, potentially contributing to the early evolution of fire-adapted traits. Future work should include targeted searches for direct fossil evidence of fire adaptations in Triassic floras, expanded anatomical surveys across additional specimens and sites, and integration with high-resolution paleoenvironmental proxies and phylogenetic analyses to refine the timing and pathways of fire-adaptation evolution.

• The fossil wood is homoxylous without visible growth rings, limiting precise annual resolution.
• Mineralization heterogeneity prevented continuous measurements in parts of the middle region beyond position 60.
• Inference is based on a single well-characterized specimen (with contextual observations from nearby logs), which may limit generalizability.
• The study does not establish a direct causal link to broader long-term climatic trends, focusing instead on short-term drought–fire sequences.
• Direct fossil evidence for specific fire-adapted traits in the local flora remains scarce and inconclusive.