Tropical land surface air temperatures cooled by around 5.4°C (or more in some reconstructions) during the Last Glacial Maximum (LGM), pointing to substantial global cooling far from the ice sheets.

The Last Glacial Maximum (LGM) was the most recent period when Earth’s ice sheets reached their greatest extent during the last glacial cycle (the Late Pleistocene).

Most definitions place the LGM roughly between 26,500 and 19,000–20,000 years ago, with peak ice volume and coldest conditions around 21,000–20,000 years ago.

Ice sheets grew toward their maxima from about 33,000–26,500 years ago in response to reduced Northern Hemisphere summer insolation, lower atmospheric CO₂, and cooler tropical Pacific sea-surface temperatures. Deglaciation began around 19–20 ka in the Northern Hemisphere.

It corresponds mainly to Marine Isotope Stage 2 (MIS 2).

Massive ice sheets covered large parts of North America (Laurentide Ice Sheet), northern Europe and Asia (Fennoscandian/Scandinavian Ice Sheet), and expanded ice in other high-latitude and mountain regions.

Permanent ice covered roughly 8% of Earth’s surface and about 25% of the land area.

Global sea level was approximately 120–130 m lower than today because so much water was locked up in ice, exposing continental shelves and creating land bridges (e.g., Beringia between Asia and North America).

The planet was substantially colder and generally drier in many regions:

• Global mean surface temperature estimates typically range from about 4–6+ °C colder than pre-industrial or present conditions. Recent data-assimilation and multiproxy studies often cluster around −5 to −6 °C (e.g., −6.1 °C in one major reconstruction; −5.4 °C or −5.9 °C in the 2026 emergent-constraint study discussed earlier). nature.com
• Cooling was strongly polar-amplified (much larger at high latitudes). Tropical land surfaces cooled by roughly 4–6 °C (noble-gas groundwater records indicate ~5.8 °C for low-to-mid latitudes; multiproxy tropical land estimates around −4.2 to −4.8 °C in recent work). Tropical sea-surface temperatures cooled less, typically 2–3+ °C.
• Atmospheric CO₂ was lower (~180–190 ppm vs. ~280 ppm pre-industrial). Dust levels were higher, vegetation was reduced in many areas, and atmospheric circulation patterns shifted.
• Many tropical and subtropical regions that are now lush were cooler and often more arid or had different vegetation zones; snowlines on tropical mountains were substantially lower.

Tropical regions were cooler and often more arid or seasonal. Vegetation belts shifted:

• Tropical rainforests (e.g., Amazon) contracted or became less dense; savanna and grassland expanded at margins.
• Montane taxa moved downslope, creating non-analogue forest communities.
• Snowlines on tropical mountains lowered by hundreds of meters.
• Overall, cooler, drier conditions and lower CO₂ reduced forest biomass and canopy density in many areas.

Primary drivers of the LGM cold state:

• Reduced Northern Hemisphere summer insolation (orbital/Milankovitch forcing).
• Lower greenhouse-gas concentrations (mainly CO₂, also CH₄ and N₂O).
• Ice-sheet albedo and topography feedbacks.
• Increased dust and vegetation changes that amplified cooling.

The LGM is one of the best-documented past climate states and serves as a key benchmark for testing climate models and estimating climate sensitivity (how much global temperature responds to changes in greenhouse gases and other forcings). Stronger reconstructed cooling generally supports moderate-to-higher climate sensitivity and helps rule out very low sensitivity values.

In summary, the LGM was a colder, drier, lower-CO₂, ice-dominated world with dramatically altered coastlines, ecosystems, and atmospheric/oceanic circulation — a key window into how Earth’s climate system responds to large forcings.

_____________________________________________________________________________________

Reassessing Last Glacial Maximum Cooling With Proxy-Constrained Emergent Relationships

This paper (first published online around 22 July 2026) reassesses Last Glacial Maximum (LGM) cooling using a new multiproxy tropical land-temperature dataset combined with an emergent-constraint approach based on paleoclimate model simulations.

They compile a new multiproxy dataset of tropical land temperature reconstructions for the LGM and apply statistical (“emergent”) relationships derived from climate model ensembles. These relationships link the better-constrained terrestrial temperatures to harder-to-observe variables such as tropical sea-surface temperature (SST) and global mean surface temperature (GMST).

Reported central estimates (with 95% confidence intervals; slight numerical differences appear between preprint and published versions):

• Tropical terrestrial cooling: roughly −4.2°C to −4.8°C (e.g., −4.2°C [−5.0 to −3.5°C] or −4.8°C [−5.6 to −4.0°C]).
• Tropical SST cooling: roughly −2.6°C to −2.9°C (e.g., −2.6°C [−3.3 to −2.0°C] or −2.9°C [−3.6 to −2.3°C]).
• Global mean surface temperature (GMST) change: −5.4°C (−6.5 to −4.4°C) in one version, or −5.9°C (−7.3 to −4.7°C) in another.

These results indicate a colder LGM than reconstructions that rely heavily on the MARGO SST compilation. The colder estimate implies that very low equilibrium climate sensitivity (ECS ≲ ~3°C) is unlikely.

The core innovation is an emergent-constraint framework: robust statistical relationships that appear across model simulations are combined with actual proxy data to “constrain” variables that are sparsely observed.

The approach is presented as flexible and adaptable to other time periods or climate variables. It builds on the long-standing difficulty of reconstructing LGM temperatures (especially in the tropics and for the global mean) from incomplete and heterogeneous proxies, a theme connected to earlier tropical-cooling studies (noble-gas groundwater records, pollen, snowline depression, etc.).

In short, the work supports substantial LGM cooling (GMST around −5.4°C in the version that matches the figure highlighted in related discussion), driven in part by stronger tropical land cooling, and uses model–proxy relationships to place tighter bounds on global and ocean temperatures. This has direct implications for paleoclimate reconstructions and for estimating climate sensitivity.

Published: Geophysical Research Letters (2026)

DOI: DOI: 10.1029/2026gl123083

Authors: A. Hou, T. Caley, T. Extier, M. Chevalier, M. Brugère, D. Swingedouw, M. Kageyama, and P. Braconnot (primarily affiliated with institutions in Bordeaux, Bonn, CNRS, and LSCE-IPSL).

Abstract

Reconstructing global temperatures from sparse proxy data remains a key uncertainty in paleoclimate science. Here, we apply an emergent constraint framework that combines statistical relationships from paleoclimate model simulations with proxy-based estimates to infer unknown climate variables. We compiled a new multiproxy data set of tropical land temperature reconstructions for the Last Glacial Maximum (LGM), which implies tropical terrestrial cooling of −4.2°C (−5.0 to −3.5°C, 95% Confidence Interval, CI). Applying our emergent constraint framework to this terrestrial temperature reconstruction, we derived updated estimates of tropical sea surface temperature (−2.6°C; −3.3 to −2.0°C, 95% CI) and global mean surface temperature (−5.4°C; −6.5 to −4.4°C, 95% CI). These observation-based estimates support LGM cooling toward the colder end of published ranges, implying that very low equilibrium climate sensitivity (ECS < ∼3°C) is unlikely. Our method is adaptable for different time periods and climate variables, providing a flexible framework for integrating proxy and model information.