New research from White Sands National Park in New Mexico shows that early humans were present during the Last Glacial Maximum (roughly 23,500- 21,700 years ago) and lived through repeated, rapid climate shifts, including drought and major landscape changes.

The findings come from an October 2026 Science Advances paper (and a related popular article by co-authors Matthew Robert Bennett and Sally Christine Reynolds in The Conversation). Expanded excavations identified 20 distinct sediment layers (horizons) containing human footprints, spanning nearly 2,000 years at the height of the last ice age. This builds on earlier (2021 and later) dating of the famous White Sands footprints that placed people in North America thousands of years earlier than many traditional models allowed.

Environmental reconstruction

Sediments, pollen, and plant- leaf waxes (biomarkers) reveal a dynamic setting quite unlike today’s desert:

• Early in the record, a large perennial lake (Lake Otero) filled much of the basin. Nearby mountains held conifer forests, and average temperatures were about 4°C colder than today.
• From around 23,600 years ago the lake began shrinking as temperatures rose. Conifers declined and were replaced by oak and sagebrush steppe; the basin grew drier due to higher evaporation.
• Around 23,200 years ago a short but severe drought struck: rainfall dropped, the lake became more saline, and drought- tolerant plants expanded. This local aridification matched a rapid Northern Hemisphere climate shift (Dansgaard- Oeschger event 2) recorded in Greenland ice cores.

People were present during these transitions, the earliest expanded footprint horizons appear around the drought/climate shift, and tracks continue at intervals as the lake retreated.

Interpreting the footprint record

The authors caution against over- interpreting the pattern.

An agent- based computer model of people moving around a fluctuating lake showed that preservation is highly selective: footprints survive mainly when mud is exposed (e.g., during drying) and then quickly buried by rising water or sediment.

As a result:

• Footprint- rich surfaces tend to cluster around climatic and hydrological transitions (exactly the pattern seen at White Sands).
• Gaps between layers do not prove people were absent, and dense tracks do not necessarily mean a population boom, a few people crossing soft mud repeatedly can leave many prints.
• The record therefore interleaves a human history (repeated use of the landscape amid changing conditions) with a geological filter that preserved only certain moments.⁠

In short, the study strengthens evidence that humans occupied the southern United States during the peak of the ice age and demonstrates they experienced abrupt, noticeable environmental change (lake recession, vegetation turnover, centennial- scale wet, dry swings) over roughly two millennia. It does not claim people only appeared during droughts, but shows they were resilient in a rapidly shifting world.

Dansgaard- Oeschger (D-O) events are abrupt, large- scale climate oscillations that occurred repeatedly during the last glacial period (roughly the past 120,000- 11,000 years). They are named after Danish geophysicist Willi Dansgaard and Swiss geophysicist Hans Oeschger, who identified them in Greenland ice-core records.

General characteristics

Each D- O event typically features:

• A very rapid warming phase (over decades, sometimes just a few years to decades) in the North Atlantic region, especially Greenland. Temperatures there could rise by 5- 15 °C (9- 27 °F), roughly half the full glacial- to- interglacial difference.
• This is followed by a more gradual cooling over centuries to millennia, often ending with a sharper drop back into cold (stadial) conditions.

These produce a characteristic “sawtooth” pattern in temperature proxies (especially oxygen isotope ratios, δ¹⁸O) from Greenland ice cores. About 25 such events are recognized over the last glacial cycle; roughly 13 well- dated ones fall between ~45,000 and 11,600 years ago, with an average spacing near 1,470 years (though exact regularity is debated).

They are most strongly expressed around the North Atlantic but have global or hemispheric effects, appearing in records from Europe, Asia, the Americas, and the Southern Hemisphere (often with some lag or opposite polarity via ocean and atmospheric teleconnections).

Causes

The precise triggers remain uncertain, but the leading explanation involves shifts in the Atlantic Meridional Overturning Circulation (AMOC), the large- scale ocean current system that transports heat northward.

Weakened AMOC (linked to expanded sea ice, freshwater input from melting ice, or other factors) produces cold stadials; a sudden reinvigoration produces the abrupt warmings. Sea- ice extent in the North Atlantic, atmospheric circulation changes, and possible external pacing (e.g., solar or orbital influences) are also implicated. Exact mechanisms are still actively researched.

Dansgaard- Oeschger event 2 specifically

D -O event 2 (also called Greenland Interstadial 2 or GI- 2) is one of the later events in the sequence. It began around 23.3- 23.2 thousand years ago (ka), during the Last Glacial Maximum (the coldest phase of the last ice age, ~26.5- 19/20 ka).

In Greenland ice cores (on the GICC05 timescale), it marks a relatively short interstadial warming. Recent high-resolution work, including from White Sands National Park in New Mexico, places its peak or expression at approximately 23.2 ± 0.3 ka.

At White Sands, this interval coincided with a short but severe regional drought: reduced rainfall, increased lake salinity in ancient Lake Otero, expansion of drought- tolerant vegetation, and shrinkage of the lake that exposed shoreline areas. This local aridification is interpreted as a teleconnected response to the broader Northern Hemisphere climate shift associated with D- O 2.

D- O 2 is notable because it occurred deep within the Last Glacial Maximum, when large ice sheets still dominated the Northern Hemisphere, yet it still produced detectable, rapid environmental changes far from Greenland, including in the American Southwest.

In short, D- O events illustrate that even during full glacial conditions the climate system could flip rapidly between cold and warmer states on human- relevant timescales, driven largely by ocean- atmosphere interactions centered on the North Atlantic. Event 2 is a specific instance of this behavior dated to ~23.2 ka.

Dansgaard- Oeschger (D- O) events, including event 2 (~23.2 ka), produced rapid, widespread climate impacts far beyond Greenland and the North Atlantic, though the strongest signals occurred in the Northern Hemisphere.

These effects arose mainly from changes in the Atlantic Meridional Overturning Circulation (AMOC), shifts in atmospheric circulation (including the Intertropical Convergence Zone, or ITCZ, and storm tracks), and interhemispheric heat redistribution via the bipolar seesaw.

Northern Hemisphere

• Europe and western Mediterranean: Interstadial (warm) phases brought warmer and wetter conditions. Vegetation expanded (more temperate forests or woodland), and lake levels or precipitation proxies often rose. Stadials were colder and drier. Temperature swings were smaller than in Greenland but still significant (several degrees).
• Eurasia (including central and eastern Europe, western Asia): Largest land- based warming during interstadials occurred here according to pollen- based reconstructions. Winters warmed more than summers, reducing seasonality. Moisture responses varied but often included wetter conditions in many areas.
• East Asia and Asian monsoon regions: Northern Hemisphere summer monsoons typically strengthened during interstadials (more rainfall in China and surrounding areas, recorded in speleothem δ¹⁸O). Southern Hemisphere monsoons weakened in a seesaw pattern.
• North America: Patterns were spatially complex. Eastern or more continental areas often warmed and sometimes became wetter during interstadials. Western North America (including the Southwest) frequently showed cooling or increased aridity in some reconstructions. For D- O event 2 specifically, multiproxy evidence from White Sands (New Mexico) and other southwestern U.S. sites records a short but severe aridification: reduced rainfall, higher lake salinity, recession of Lake Otero, and a shift toward drought- tolerant vegetation (conifer forest giving way to sagebrush/oak steppe). This matched the Greenland timing of D- O 2 and allowed humans and megafauna to traverse newly exposed lake margins.⁠

Tropics and low latitudes

The ITCZ shifted northward during Northern Hemisphere interstadials. This increased rainfall in northern tropical regions (e.g., Caribbean, northern South America, parts of Africa and Asia) and decreased it in southern tropical areas. Speleothem and lake records across the tropics consistently show these hydroclimate flips, often with little lag relative to Greenland.

Southern Hemisphere

• Antarctica and Southern Ocean: Temperature changes were typically out of phase with Greenland (the bipolar seesaw). Antarctic warming (Antarctic Isotope Maximum events) often led or lagged Greenland warmings by centuries, driven mainly by slower oceanic heat transport changes. Atmospheric teleconnections were faster: mid- latitude storm tracks and westerly winds shifted within decades of North Atlantic changes, altering moisture sources for Antarctic precipitation (seen in deuterium- excess records).
• Southern extratropics more broadly often cooled or showed muted/opposite responses relative to northern warming.

Spatial and temporal patterns

Impacts diminished with distance from the North Atlantic but remained detectable globally. Warming was strongest in northern extratropics (especially Eurasia), with antiphasing between northern and southern extratropics.

In North America a west- east contrast (cooling and aridity in the west, warming farther east) appears in some pollen syntheses. Moisture (precipitation) changes did not always mirror temperature changes and were regionally specific. Most teleconnections outside the high southern latitudes operated rapidly (decades or less) via the atmosphere; oceanic links to Antarctica were slower (centuries).

These patterns hold across multiple D- O events, including weaker ones like event 2 during the Last Glacial Maximum, although signal strength and coherence are greatest for the larger events. Proxy records (ice cores, speleothems, pollen, marine sediments, lake levels) and climate models consistently support this global reorganization of atmospheric and oceanic circulation.

Humans experienced abrupt environmental changes at White Sands during the Last Glacial Maximum

Key findings

Expanded excavations at White Sands National Park (WHSA) Locality 2 (Tularosa Basin, New Mexico) uncovered 20 stratigraphic horizons containing in situ human footprints, megafaunal trackways (mammoth/Proboscidea, ground sloth, canids, possible ursids, rare birds), and linear features interpreted as travois (drag marks from poles used to transport resources).

Refined chronology (using improved radiocarbon dating of conifer pollen and aquatic Ruppia seeds, plus supporting OSL ages):

• Footprint- bearing sequence spans 23.5 to 21.7 thousand years ago (ka).
• Firmly within the Last Glacial Maximum (LGM).

Paleoenvironmental reconstruction (pollen, leaf- wax biomarkers and hydrogen isotopes, stratigraphy):

• The setting began as a shallow perennial lake (pluvial Lake Otero) with surrounding conifer forests and cooler temperatures.
• It underwent a complete environmental overturn to a lake- margin and wet- playa system.
• A major aridification episode at ~23.2 ka coincided with Dansgaard- Oeschger event 2 (a rapid Northern Hemisphere climate oscillation recorded in Greenland ice cores). Rainfall declined, the lake shrank and became more saline, and vegetation shifted toward drought- tolerant sagebrush and oak steppe.
• Centennial- scale swings in hydroclimate continued afterward.
• Lake recession during the aridification exposed shoreline areas that humans and megafauna repeatedly traversed for nearly two millennia.

This provides a high-resolution terrestrial record showing that early humans in North America experienced marked, abrupt environmental changes (including severe drought) during the height of the last ice age. The earliest track horizons occur around the D- O 2 transition.

The paper strengthens the LGM age of the White Sands footprints (previously debated) and demonstrates human presence and activity (including possible resource transport with travois) amid a dynamic glacial climate.

Journal: Science Advances, 14 October 2026, Vol. 12, Issue 42, eaeh0707

DOI: 10.1126/sciadv.aeh0707

PMC full text: PMC13666055

Provided: The Conversation

Authors: Kathleen B. Springer, Jeffrey S. Pigati, Jessica E. Tierney, et al. (including Sally C. Reynolds and Matthew R. Bennett)

Abstract

Expanded excavations at White Sands National Park have unearthed 20 stratigraphic horizons containing in situ human footprints, megafaunal trackways, and linear features created by travois. Advances in radiocarbon dating of conifer pollen, combined with ages of aquatic seeds, refines the chronology of the footprint-bearing sequence to between 23.5 and 21.7 thousand years ago (ka). Biomarker, pollen, and stratigraphic evidence indicate a complete overturning of the environment from a perennial lake to a playa and capture a large aridification episode coincident with Dansgaard-Oeschger event 2 at 23.2 ka, which shrank pluvial Lake Otero and allowed people to traverse the former shoreline area for nearly two millennia. This well-dated and highly resolved paleoclimate record provides compelling evidence that humans in North America experienced marked environmental changes during the Last Glacial Maximum.