Glacial Dynamics and Ice Rheology
|
|
|
Polar ice plays a critical but poorly understood role in global climate. From driving sea-level rise to altering Earth’s reflectivity, glaciers and sea ice influence the planet’s energy balance in complex, nonlinear ways. Yet these systems are rapidly changing; Antarctica and Greenland’s vast ice sheets are thinning and flowing toward the ocean faster than ever, with retreating glaciers, collapsing ice shelves, and even glacial earthquakes.
Traditionally viewed as slow, steady flows, ice has shown it can fail catastrophically; ice shelves have collapsed in weeks, and iceberg capsizes can release energy on the scale of kilotons of TNT. Our lab recreates these events at laboratory scale, modeling iceberg capsizes and measuring the forces that could trigger glacial earthquakes. We also study ice mélange, a dense mixture of icebergs, sea ice, and fragments that can stabilize calving fronts. Through experiments and modeling, we apply nonlocal granular fluidity rheology to quantify mélange motion and its capacity to buttress glacier termini.
Building on these experimental approaches, we employ physics-informed neural networks to determine the rheology of non-Newtonian fluids, such as ice shelves, using a laboratory-scale system that recreates glacier–ocean interface conditions. By training on flow velocity and shelf thickness from controlled experiments, these models can recover complex, nonlinear constitutive laws that are difficult to measure using traditional methods.
Traditionally viewed as slow, steady flows, ice has shown it can fail catastrophically; ice shelves have collapsed in weeks, and iceberg capsizes can release energy on the scale of kilotons of TNT. Our lab recreates these events at laboratory scale, modeling iceberg capsizes and measuring the forces that could trigger glacial earthquakes. We also study ice mélange, a dense mixture of icebergs, sea ice, and fragments that can stabilize calving fronts. Through experiments and modeling, we apply nonlocal granular fluidity rheology to quantify mélange motion and its capacity to buttress glacier termini.
Building on these experimental approaches, we employ physics-informed neural networks to determine the rheology of non-Newtonian fluids, such as ice shelves, using a laboratory-scale system that recreates glacier–ocean interface conditions. By training on flow velocity and shelf thickness from controlled experiments, these models can recover complex, nonlinear constitutive laws that are difficult to measure using traditional methods.
Extending from our laboratory and modeling work, we have developed a constitutive theoretical model for the time-dependent deformation of ice that captures both the rapid, transient creep seen immediately after loading and the slower, steady-state flow that follows. The model combines Andrade-like creep, which represents collective dislocation activity, with the evolution of an internal structure parameter that controls long-term viscosity. This formulation naturally reproduces features observed in laboratory experiments and field measurements, such as the strain-rate minimum at ~1% strain and the stress dependence of the transient phase. By bridging short-term and long-term deformation, the model provides predictive capability for ice flow across timescales from minutes to centuries, offering a unified framework for interpreting both experimental and glaciological data.
Relevant Publications:
- Laboratory Investigations of Iceberg-Capsize dynamics, Energy Dissipation and Tsunamigenesis
- Seasonal Changes of Mélange Thickness Coincide With Greenland Calving Dynamics
- A quasi-one-dimensional ice mélange flow model based on continuum descriptions of granular materials
- Experimental Investigations of Ice Mélange and the Flow of Floating Granular Materials
- Soft matter physics of the ground beneath our feet
- Physics of the cryosphere
- Improved Estimation of Glacial-Earthquake Size Through New Modeling of the Seismic Source
- Granular decoherence precedes ice mélange failure and glacier calving at Jakobshavn Isbræ
Particles: Flowing and Creeping
Another recent project has employed diffusing-wave spectroscopy, an imaging technique that allows us to probe the localized strain in a granular system. A gas laser and series of optical filters creates a speckled interference pattern across the plane of interest, and by correlating the pixel intensities of different metapixels, we can see how granular clusters shift over time. The video above captures this interference pattern produced while pouring fine sand into our imaging cell; note the different pixel brightness across the sand pile, as well as how the grains shift and form avalanches at the top. While this video shows a fast flow, the effects of this internal strain can be seen in a slow creep continuing hours after the system appears to be at rest. Importantly, though, better understanding of these kinds of granular systems could have applications in the real-world geophysics of landslides and avalanches.
The flow time measured by traditional hourglasses or sand timers is often determined by the quantity of sand they contain, as well as how large the sand grains are relative to the timer's neck width. However, there are a number of other factors that can influence the granular flow in the system, and thus the timer duration. When the timers are opened to air with a small vent or hole, the number of holes, pressure, ambient gas, and even contact materials in the assembly all play interesting roles in altering the flow time.
The Secret Stickiness of Sand
- Ever wondered how a sand castle holds its shape? We know the water between the grains acts as a temporary glue, but a fascinating puzzle emerges long after it has dried. Even when a cluster of sand is fully dry, it can remain rigid. This phenomenon is everywhere, from a crumbly sand cluster on a beach to the soil on other planets. Our research explores the fundamental physics behind these persistent forces that hold grains together long after the water is gone.
To investigate why dry particles stick together, we bring the phenomenon into the lab. We recreate these rigid clusters using spherical glass particles as a clean, controlled model system. The process involves carefully wetting the particles, drying them, and then measuring the cluster's strength.
Wetting and Drying: We immerse the glass particles in various liquids including different polarities, water, ethanol, hexane and chloroform. The samples are then dried by evaporating the liquid in an oven.
We also treat a different group of particles with high-pressure steam in an autoclave, these samples go then into the oven to be dried.
Measuring Stickiness: After drying all samples, we use a robotic arm equipped with a force sensor to perform an intrusion test. By measuring the force required to break the cluster apart, we can quantify the strength of the adhesion between the particles.
We also treat a different group of particles with high-pressure steam in an autoclave, these samples go then into the oven to be dried.
Measuring Stickiness: After drying all samples, we use a robotic arm equipped with a force sensor to perform an intrusion test. By measuring the force required to break the cluster apart, we can quantify the strength of the adhesion between the particles.
|
|
|
|
|
High-rressure steam creates significanttly strong bonds on sand. Using an autoclave or "Instant Pot" to apply high-pressure, steam creates the strongest clusters by far. This shows that for typical samples, humidity combined with pressure and temperature is effective at creating powerful adhesion between the glass beeds.
But cleaning the sand reveals a deeper truth. We found that as sand is cleaned multiple times, the strengthening effect of pressure becomes less important. For sand cleaned four times, the force measured on Instant Pot or Autoclave samples (with pressure) was only slightly higher than for loose sand (with no pressure). However, the adhesion from simply wetting with pure water and drying remained significant throughout. Our central question is, then: why the residual stickiness?
But cleaning the sand reveals a deeper truth. We found that as sand is cleaned multiple times, the strengthening effect of pressure becomes less important. For sand cleaned four times, the force measured on Instant Pot or Autoclave samples (with pressure) was only slightly higher than for loose sand (with no pressure). However, the adhesion from simply wetting with pure water and drying remained significant throughout. Our central question is, then: why the residual stickiness?
Since the adhesive effect persists even after extensive cleaning and without pressure, the liquid itself must be causing a fundamental change to the particle surfaces. Could this be due to surface chemistry modification driven by a liquid's specific properties, like its polarity?We tested different liquids and found a pattern: Polar liquids like water created significantly stronger clusters than non-polar liquids like hexane or chloroform. Thissuggests that the polarity of the liquid could be a primary driver of the residual adhesion we observe.
But this may not be the only mechanism at play. We are actively investigating other physical phenomena that could contribute to this residual stickiness. Could tiny, stable capillary bridges of liquid remain between the grains, even after the sample appears fully dry? Or could electrostatic charges be contributing to these forces? Hopefully, we will know soon.
But this may not be the only mechanism at play. We are actively investigating other physical phenomena that could contribute to this residual stickiness. Could tiny, stable capillary bridges of liquid remain between the grains, even after the sample appears fully dry? Or could electrostatic charges be contributing to these forces? Hopefully, we will know soon.