Salt may have amplified Earth’s descent into a “Snowball Earth” state around 700 million years ago through a salt- albedo feedback, according to research published in Climate of the Past.

During the Cryogenian Period (roughly 720- 635 million years ago), Earth experienced extreme glaciations in which ice advanced from the poles to the tropics, covering oceans and continents. Geological evidence includes glacial deposits at low latitudes.

The classic ice- albedo feedback is central to explanations: snow and ice reflect most incoming sunlight (high albedo), reducing absorption of solar energy, which cools the planet further and allows more ice to form.

Researchers from UiT, The Arctic University of Norway (Aksel Samuelsberg, Per Kristen Jakobsen, and Martin Rypdal) proposed an additional mechanism.

As sea ice forms, most salt is expelled, but some remains trapped in brine pockets. Under very cold, dry conditions (relevant to a Snowball Earth), salts crystallize, e.g., mirabilite near –8°C and hydrohalite (a hydrated sodium chloride) around –23°C or lower.

As bare sea ice sublimates (turns directly from solid to vapor, common in dry tropical regions with limited snowfall), these salts are left behind as a lag deposit of bright white crystals.

These salt crystals have extremely high albedo, models used values around 93% reflectance for a hydrohalite- rich surface, compared with about 55% for bare sea ice and 80% for snow/ice.

This creates a positive feedback: more reflection → less warming → further cooling and more ice/salt.

In a simple climate (energy- balance) model that incorporated this salt- albedo feedback, the process intensified cooling once activated during the early stages of global glaciation. It acted like an accelerator, helping push the planet into a deeper frozen state than ice- albedo alone, and made the frozen state more resistant to melting (requiring substantially more greenhouse warming to exit).

The study notes limitations: it is an initial investigation using a simplified model.

It is uncertain whether large, persistent salt deposits would have formed and lasted; more complex models including clouds, winds, ice dynamics, dust, and other processes are needed.

Over long timescales, ice flow from higher latitudes could also dilute tropical salt. The salt feedback complements rather than replaces the ice- albedo effect and may help explain aspects of how Earth entered (and stayed in) these extreme states.

Snowball Earth events are thought to have occurred more than once in the Neoproterozoic and may have influenced early life evolution. This work highlights how surface physical processes can strongly affect planetary climate under conditions very different from today’s.

_____________________________________________________________________________________

Cryogenian climate models investigate the extreme “Snowball Earth” glaciations of the Cryogenian Period (roughly 720- 635 million years ago), particularly the long Sturtian (~717- 660 Ma) and Marinoan (~645- 635 Ma) events. These models explore how Earth could enter a near- global ice-covered state, the climate dynamics once frozen, and how it could escape that state.

The central mechanism is the ice- albedo feedback. Snow and ice have high reflectivity (albedo), so expanding ice reflects more sunlight, cools the surface further, and promotes more ice growth.

This creates bistability (hysteresis) in climate models:

• A warm, partially ice- covered or ice-free state can tip into a fully glaciated “Snowball” state under reduced solar luminosity (Neoproterozoic Sun was ~6% weaker), low CO₂, continental configurations favoring high albedo, or other triggers.
• Once in the Snowball state, escape typically requires very high atmospheric CO₂ (from volcanic outgassing, as silicate weathering is suppressed under ice) to overcome the high planetary albedo.

Simple one- dimensional energy- balance models (EBMs, e.g., Budyko- Sellers type) capture this hysteresis clearly, showing stable warm, intermediate, and fully glaciated branches, with an unstable regime between them. More complex models produce qualitatively similar behavior.

“Hard Snowball” (near- total ice cover) versus “slushball” or “Waterbelt” (narrow equatorial open- water band) remains debated. Geological evidence (equatorial glacial deposits, cap carbonates) supports extensive ice, but models and some proxy data allow limited open water or thin ice in places.

Types of Models Used

Energy- balance models (EBMs): Idealized, low-dimensional (often 1D meridional). They efficiently explore bifurcation points, albedo parameterizations, and feedback. Recent work incorporates a salt- albedo feedback: in net-ablation (sublimation- dominated) tropical zones, cold bare sea ice can leave highly reflective salt crystal lag deposits (e.g., hydrohalite, albedo ~0.93). This intensifies cooling once initiated and makes the frozen state more resistant to melting.

General circulation models (GCMs) / Atmosphere-Ocean GCMs (AOGCMs): 3D models (e.g., CAM, ECHAM, FOAM, GENESIS, LMDz, MIROC, MPI-ESM) simulate atmospheric circulation, clouds, hydrology, ocean dynamics, and sometimes coupled ice sheets. They resolve Hadley cells, polar vortices, and regional effects under prescribed or evolving ice cover.

Intermediate- complexity models (e.g., CLIMBER) and specialized setups: Used for sensitivity studies on paleogeography, orbital parameters, and volcanism.

Coupled climate- ice- sheet and biogeochemical models: Explore long- term ice dynamics, sea- glacier flow, and carbon- cycle interactions (e.g., CO₂ buildup for deglaciation).

Boundary conditions typically include reduced solar constant (~94% of modern), Cryogenian continental configurations (often tropical land bias after Rodinia breakup), variable CO₂, and sometimes large igneous province sulfur aerosols or orbital variations.

Main Findings from Modeling

Initiation:

• Ice- albedo feedback can drive rapid global glaciation (years to decades in some simulations) once a critical ice- line latitude is crossed.
• Sensitivity studies show CO₂ thresholds for inception vary widely (roughly 10- 250 ppm or more) depending on continental layout, orbital geometry, and volcanic aerosols. Tropical continents raise planetary albedo and can favor colder climates. Asteroid or large volcanic sulfur injections can tip already- cool states into Snowball conditions.

Snowball climate dynamics:

• Atmosphere becomes very dry and cold. Tropical circulation can reverse due to low thermal inertia of an ice- covered surface, creating an equatorial desert with net sublimation and accumulation of snow/frost at higher latitudes.
• Intermodel comparisons of fully ice- covered states show substantial differences in surface temperatures and ice thickness (one outlier model, FOAM, is notably colder partly due to limited clouds).
• Clouds generally provide a net warming effect and can lower the CO₂ needed for deglaciation; high-resolution simulations reveal complex, meandering cloud structures.
• Limited open water (~15% or less) can allow climate oscillations (seasonal, interannual) similar to modern ones, consistent with some sedimentary records.
• Ocean circulation changes include strong salinity stratification during sea- ice expansion.

Termination:

• Requires high CO₂ (hundreds of times pre- industrial in some older GCM results; lower in models with stronger cloud or dust effects). Dust accumulation, snow aging, or mid- latitude processes can aid melting in some simulations.
• Deglaciation is often abrupt once melting begins, followed by extreme greenhouse conditions and deposition of distinctive “cap carbonates.”

Additional feedbacks and processes:

• Salt precipitation from sea ice (as noted above) can amplify early- stage cooling.
• Sea- glacier flow, cryoconite holes, meltwater ponds, and subglacial oceans are explored for their roles in ice dynamics and potential refugia for life.
• Paleogeography, reduced degassing rates, and weathering of volcanic provinces have been linked to initiation in combined tectonic– climate models.

Challenges and Uncertainties

Models disagree on exact CO₂ thresholds, the persistence of open water, cloud radiative effects, and the relative importance of various feedbacks.

Surface albedo schemes (snow vs. bare ice vs. salt vs. dust) strongly influence results.

Many simulations idealize geography or omit full ice dynamics, clouds, or dust. Geological constraints (duration of glaciations, proxy temperatures, paleogeography) continue to refine model setups.

Overall, models of increasing complexity confirm that ice -albedo feedback makes global glaciation a robust possibility under Neoproterozoic conditions, while highlighting the sensitivity of the climate system to surface processes, clouds, and boundary conditions.

Ongoing work integrates better paleogeography, biogeochemistry, and high- resolution physics to better match the geological record of these extreme events.

_____________________________________________________________________________________

Salt- albedo feedback is positive climate feedback in which the precipitation of salts from sea ice increases surface reflectivity (albedo), leading to greater reflection of sunlight, further cooling, and more ice/salt formation. It is particularly relevant to extreme cold climates such as Neoproterozoic Snowball Earth events (~720- 635 million years ago).

When seawater freezes, most salt is rejected into the surrounding water or concentrated in brine pockets within the ice (brine rejection).

Under sufficiently low temperatures, salts crystallize out of these brine inclusions:

• Different salts precipitate at different temperatures. Mirabilite (Na₂SO₄·10H₂O) can form around –8°C. Hydrohalite (NaCl·2H₂O), the most abundant, begins precipitating significantly below about –23°C. The full seawater eutectic (complete solidification of remaining brine) occurs near –36°C to –37°C.
• Precipitation shows hysteresis: crystals form at lower temperatures than those at which they dissolve upon warming.

In a net- ablation (sublimation- dominated) environment, expected in the dry tropics of a Snowball Earth, ice sublimates (solid → vapor), leaving behind a lag deposit of bright salt crystals on the surface. These crystals are highly scattering and reflective, especially in the near infrared where pure ice absorbs more strongly.

Laboratory measurements show broadband albedos for hydrohalite-rich crusts reaching ~0.93, compared with ~0.5- 0.55 for bare cold sea ice, ~0.8 for snow, and lower values for melting or ponded ice.

This creates a positive feedback loop:

1. Cooling allows salt precipitation and/or crust formation.
2. Higher albedo reflects more solar radiation.
3. Less energy is absorbed → further cooling → more precipitation/ice.

There are related sub- processes: a milder “precipitation feedback” (salts inside ice raise albedo slightly) and stronger “crust- dissolution feedback” (warming dissolves the lag deposit, sharply lowering albedo and accelerating melt).

Observational and Experimental Evidence:

• Field measurements (e.g., McMurdo Sound, Antarctica): Bare cold sea ice albedo increases gradually as temperatures drop below the hydrohalite precipitation threshold, consistent with salt crystallization.
• Laboratory experiments: Controlled growth of NaCl ice, sublimation to form hydrohalite lag deposits, and spectral albedo measurements confirm very high reflectivity (~0.93 all- wave). Warming causes rapid dissolution, ponding, and albedo collapse (sometimes below 0.4). Optical modeling derives inherent optical properties of hydrohalite and provides parameterizations for climate models.

On the modern Earth the effect is negligible because temperatures cold enough for significant precipitation occur mainly in polar winter under snow cover with limited sunlight. It becomes climatically important only under prolonged, low- latitude, cold, dry conditions with exposed sea ice.

Most Snowball Earth simulations have used simplified albedo schemes (snow- covered, bare cold ice, melting ice, ponds). The salt- albedo effect was hypothesized earlier (around 2009- 2016) but only recently incorporated into climate models.

In a 2026 study, researchers added a salt-albedo parameterization to a one- dimensional energy-balance model (EBM). Key features included:

• Temperature- dependent switch to high salt albedo (0.93) below a eutectic threshold in the net-ablation zone.
• Resulting bifurcation diagrams showed two coexisting stable Snowball states: one with salt deposits (significantly colder) and one without.
• Once activated, the feedback intensified early- stage cooling (acting as an “accelerator”) and raised the radiative forcing (e.g., CO₂) needed for deglaciation, making the frozen state more resistant to melting.

Albedo values used in related modeling include ~0.55 for bare sea ice, ~0.80 for snow/ice, and ~0.93 for hydrohalite- rich surfaces.

The feedback has also been explored for exoplanets (e.g., M- dwarf habitable- zone worlds), where hydrohalite’s high near- infrared reflectivity interacts strongly with the host star’s spectrum and can cool planets further.

In Snowball Earth contexts, the feedback helps explain how an already cooling planet could plunge deeper into a hard- frozen state, especially in tropical ablation zones where snow cover was limited. It adds nuance to the classic ice- albedo feedback and hysteresis loop.

In Snowball Earth contexts, the feedback helps explain how an already cooling planet could plunge deeper into a hard- frozen state, especially in tropical ablation zones where snow cover was limited. It adds nuance to the classic ice- albedo feedback and hysteresis loop.

Limitations and uncertainties include:

• Duration and persistence of salt crusts (ice flow from higher latitudes could supply fresher ice; winds, dust, or clouds could modify the effect).
• Exact temperatures and rates under realistic multi-salt seawater chemistry and dynamics.
• Need for fuller 3D GCMs that include clouds, winds, ice dynamics, and dust to quantify the net strength.
• Potential competing effects during deglaciation (dissolution of salt could lower albedo and aid melting once temperatures rise).

Overall, salt-albedo feedback illustrates how micro-scale surface processes (brine chemistry and crystal optics) can influence planetary- scale climate under extreme conditions. It remains an active area of research, with the recent EBM results providing a foundation for more sophisticated modeling.

_____________________________________________________________________________________

Amplified cooling of Snowball Earth from a salt– albedo feedback

On a Snowball Earth, atmospheric circulation creates a net ablation (sublimation- dominated) zone at lower latitudes. Bare sea ice is exposed for long periods under extreme cold. At sufficiently low temperatures, salts precipitate from brine pockets in the ice. As the ice sublimates, these salts form a lag deposit of highly reflective crystals (primarily hydrohalite, NaCl·2H₂O).

This raises surface albedo and creates positive salt- albedo feedback that intensifies cooling. The effect had not previously been included in Snowball Earth climate models.

The authors implemented a simplified salt-albedo parameterization in a classic one- dimensional (meridional) diffusive energy- balance model (EBM).

The model equation balances heat capacity, meridional diffusion of heat, absorbed solar radiation (modified by albedo), and outgoing longwave radiation (Planck response, parameterized as

$A + BT$A + BT).

Key albedo features:

• Standard ice- albedo feedback (higher albedo below an ice- line temperature threshold).
• Lower bare- sea-i ce albedo in the net ablation zone (Jormungand/Waterbelt- style setup).
• Salt albedo ($\alpha_4 = 0.93$) Activated below a eutectic temperature ($T_e \approx -36^\circ$C) in the ablation zone, representing hydrohalite lag deposits.

Equilibrium solutions were found analytically via boundary integrals (locating both an ice line and a “salt line”). Bifurcation diagrams were constructed using outgoing longwave parameter (A) as the control (proxy for greenhouse forcing such as CO₂). Stability was assessed with the slope- stability theorem and finite-difference simulations.

Main Results

• The salt- albedo feedback produces two coexisting stable fully glaciated (Snowball) states: one with salt deposits on equatorial bare ice (significantly colder) and one without.
• Once the feedback activates, it amplifies cooling during the early phase of global glaciation, acting like an accelerator that pushes the climate deeper into a frozen state than ice- albedo feedback alone would achieve.
• The salt- covered state is more resistant to deglaciation: substantially stronger warming (higher greenhouse forcing) is required to exit it compared with the salt- free Snowball state.
• For the model parameters examined, the salt-deposit state is the one preferentially reached when cooling from a warm climate, consistent with the classic Snowball Earth hysteresis (warm state → Snowball → eventual escape under high CO₂).

The authors note that multiple unstable states with partial salt deposits also appear but are not the focus.

Salt precipitation may have played a role in shaping the early climate of Snowball Earth by rapidly intensifying cooling via the high albedo of salt crystals.

The work is described as an initial investigation.

Limitations include the simplicity of the EBM (no clouds, winds, full ice dynamics, dust, or multi- component ocean chemistry) and uncertainty about how long salt deposits would persist (ice flow from higher latitudes could dilute them).

More sophisticated 3D models are needed to quantify the effect under realistic conditions.

This study builds directly on earlier laboratory and field work (Light, Carns, and colleagues, ~2009- 2016) that measured the high albedo of hydrohalite crusts and hypothesized the feedback, providing the first climate- model demonstration of its potential climatic impact.

Published: Climate of the Past

DOI: 10.5194/egusphere-2026-679 preprint on EGUsphere

DOI: final version: Clim. Past, 22, 1499–1506, 2026

Authors: Aksel Samuelsberg, Per Kristen Jakobsen, and Martin Rypdal (UiT—The Arctic University of Norway)

Abstract. 

It is believed that the atmospheric circulation on Snowball Earth produced a net ablation zone exposing bare sea ice. Under sufficiently low temperatures, salt begins to precipitate out of sea ice, forming a lag deposit of crystals with high albedo as the ice sublimates. This could have resulted in a salt- albedo feedback that has not previously been included in modeling studies of Snowball Earth. We implement salt-albedo feedback in a simple climate model and show that, once initiated, this mechanism could have intensified global cooling in the initial phase of Snowball Earth. Our results suggest that salt precipitation may have played a role in shaping the early climate of Snowball Earth.