60 million years of glaciation in the Transantarctic Mountains

During the Late Cretaceous, global climates were considerably warmer than present, allowing dense subtropical forests to occupy Antarctica despite its polar location. Through the Cenozoic, declining atmospheric CO2 drove a long-term shift from greenhouse to icehouse conditions, with dramatic cooling at the Eocene–Oligocene (EO) boundary (~34 Ma) and the development of Antarctic ice sheets. While the evolution of Antarctic ice sheets since the EO boundary has been studied extensively, far less is known about the timing and nature of the earliest mountain glaciers that predate or accompany ice-sheet inception. Key questions include: when did glaciation begin in Antarctica, were glaciers present under full greenhouse conditions, and how long did temperate (warm-based) mountain glaciers persist? Extensive ice-sheet cover and erosion obscure onshore records, so distal marine proxies have dominated prior reconstructions but are less sensitive to small ice masses. The authors propose using glacial cirques—ubiquitous armchair-shaped erosional landforms formed at glaciation onset—to infer past glacier equilibrium line altitudes (ELAs) and, via modern temperate-glacier analogues, reconstruct palaeotemperatures to constrain the timing and style of early Antarctic mountain glaciation.

Previous long-term reconstructions of Antarctic glaciation have relied chiefly on distal records such as offshore sedimentation, marine isotope data, and eustatic sea-level indicators, which are not well suited to detecting smaller mountain glaciers present at glaciation onset. Geological evidence indicates widespread Late Eocene glaciation in parts of Antarctica (e.g., ice-rafted debris near the Weddell Sea by ~36.6 Ma). Thermochronologic studies in the Transantarctic Mountains (TAM) also suggest enhanced exhumation during the Late Eocene–Early Oligocene, interpreted as increased erosion by temperate glaciers. Biological proxies indicate high warm-month temperatures during the Late Paleocene–Early Eocene, decreasing through the Middle to Late Eocene, with associated declines in precipitation through the Eocene into the Oligocene and Early Miocene. These prior studies frame the context for evaluating the presence and thermal regime (temperate versus cold-based) of early Antarctic mountain glaciers using cirque metrics.

• Study region and mapping: The authors mapped glacial cirques across the Transantarctic Mountains (TAM) using the Reference Elevation Model of Antarctica (REMA) digital surface model at 8 m resolution (vertical error <1 m). Cirques were identified as depressions bounded upslope by arcuate headwalls and open down-valley. Headwalls and, where visible, cirque thresholds were mapped following established procedures. Where thresholds were indistinct, lower limits were approximated by straight lines between headwall/lateral spur termini. This produced a comprehensive TAM cirque database, cross-validated where prior local inventories existed.
• Glacier cover classification: To focus on reconstructing early temperate mountain glaciation, only glacier-free cirques were analysed, as these preserve landform altitudes without uncertainty from unknown ice thicknesses. Glacier cover within each mapped cirque was assessed using a continent-wide rock outcrop dataset derived from Landsat 8 imagery. Glacier-occupied cirques were excluded from ELA calculations.
• ELA and palaeotopography adjustment: Past ELAs associated with glacier-free cirques were estimated from cirque morphometry and altitudes. ELAs derived in the modern topography were adjusted to reconstructed palaeotopography to account for post-formational topographic change.
• Palaeoclimate inference: Assuming cirques formed under temperate (warm-based) glaciers, the mean summertime air temperature (MSAT) at the ELA was taken as 3.6 ± 2.5 °C, based on modern temperate-glacier analogues. ELAs and local lapse-rate corrections were used to convert to sea-level (SL) MSAT estimates. These MSAT reconstructions were compared to independent published Antarctic palaeotemperature and precipitation proxies to infer the timing and spatial extent of temperate versus cold-based mountain glaciation and transitions to ice-sheet cover throughout the Cenozoic.

• Inventory: 14,060 cirques were mapped across the TAM; 1,292 are glacier-free and suitable for analysis.
• ELA and temperature ranges: Glacier ELAs (after palaeotopography adjustment) range from 211 to 3,888 m. Assuming MSAT at ELAs of 3.6 ± 2.5 °C implies MSAT increases relative to present of ~10.0 ± 2.5 to 40.8 ± 2.5 °C, depending on location/altitude. Sea-level (SL) MSATs required for temperate glaciation range from 5.0 ± 2.5 to 28.9 ± 2.5 °C.
• Onset and prevalence by period (percentages refer to the total glacier-free cirque population, n=1,292): • Late Paleocene (~60–56 Ma): Limited temperate mountain glaciation likely at the highest elevations (~3% temperate; <1% cold-based), consistent with high warm-month temperatures (~25.7 ± 2.7 °C) but high precipitation (~2110 mm mean annual). • Early Eocene (~56–48 Ma): Similar to Late Paleocene—few high-altitude temperate glaciers possible. • Middle Eocene (~48–40 Ma): Expansion to ~8% temperate mountain glacier occupancy (precipitation reduced to ~1534 mm mean annual). • Late Eocene (~40–34 Ma): Peak temperate mountain glaciation—about 76% of cirques occupied by temperate glaciers; ~9% glacier-free; remainder likely occupied by larger, probably cold-based ice masses. Independent evidence (e.g., ice-rafted debris by 36.6 Ma) corroborates widespread glaciation. • Oligocene (~34–23 Ma): With MSAT 4–12 °C and 500–800 mm mean annual precipitation, all cirques likely ice-covered; ~15% remained temperate, the majority cold-based. • Early Miocene (~23–15 Ma): Likely complete submergence of cirques beneath ice sheets and/or predominance of cold-based glaciers—temperate mountain glaciation effectively ceased. • Mid-Miocene Climatic Optimum (~15 Ma): Warming and increased precipitation allowed a brief return of temperate mountain glaciers (~11% of cirques). By ~13.96 Ma, cooling and aridification (e.g., MSAT ~−1.7 °C; ~150 mm MAP in Dry Valleys) drove a switch back to cold-based conditions.
• Overall: Mountain glaciers likely existed in the TAM during greenhouse intervals (Late Paleocene, Middle Eocene), became widespread and warm-based in the Late Eocene, transitioned to dominantly cold-based/ice-sheet conditions at and after the EO boundary, with a short-lived temperate resurgence during the MCO before persistent cold-based glaciation to present.

The results indicate that Antarctic mountain glaciation in the TAM began earlier than the formation of continent-scale ice sheets and even during greenhouse conditions. The ELA-derived MSATs, when matched with independent biological proxies, constrain the timing of temperate glacier presence and show spatial/altitudinal sequencing of cirque occupation through progressive Cenozoic cooling and aridification. The Late Eocene emerges as a key interval of widespread temperate mountain glaciation, followed by a shift to colder and drier conditions at the EO boundary when glaciers coalesced into ice sheets and became predominantly cold-based. The brief Mid-Miocene warming allowed limited temperate glaciation to reappear before a sustained return to cold-based regimes. These findings address the research questions by establishing an earlier onset of glaciation, demonstrating glacier presence under greenhouse climates, and quantifying the persistence and extent of temperate glaciation. They also corroborate independent geological and thermochronologic evidence for enhanced erosion and glacial activity during the Late Eocene–Early Oligocene, underscoring the utility of cirque analyses as terrestrial high-latitude climate archives.

This study compiles the first comprehensive inventory of glacier-free cirques across the Transantarctic Mountains and uses their altitudes to infer past ELAs and palaeotemperatures. The analysis demonstrates that: (1) mountain glaciers likely initiated by the Late Paleocene and expanded during the Middle Eocene; (2) temperate glaciers were widespread in the Late Eocene; (3) a major transition to dominantly cold-based, ice-sheet conditions occurred from the EO boundary through the Oligocene and Early Miocene; and (4) temperate mountain glaciers briefly returned during the Mid-Miocene Climatic Optimum before a persistent cold-based regime to present. These results extend the known longevity of Antarctic glaciation into greenhouse intervals and provide quantitative climatic constraints on early glaciation in the TAM.

• The approach assumes cirques formed under temperate (warm-based) glacier conditions; deviations from this could bias ELA–temperature inferences.
• Only glacier-free cirques were analysed to avoid uncertainties from unknown ice thickness; this excludes potentially informative glacier-occupied cirques.
• Conversion from cirque altitudes to ELAs and to sea-level MSATs carries uncertainties (e.g., MSAT at ELA assumed 3.6 ± 2.5 °C; lapse-rate and palaeotopography corrections), and results vary spatially with cirque altitude and location.
• Precipitation changes through time influence glacier mass balance; proxy-based precipitation reconstructions are limited and introduce additional uncertainty in interpreting glacier thermal regimes and extent.
• Extensive ice-sheet cover and erosion may have modified or obscured some landforms, and some cirques may have been intermittently covered by minimally erosive cold-based ice, complicating precise occupation histories.