Research
Continuum Damage Mechanics
Half of the water coming from the Antarctica comes off the ice by breaking off into icebergs, which float away and melt later, but the processes describing ice fracture and continental ice flow span almost ten orders of magnitude in length- and time-scales. To make predictions of large time and space scales, though, maybe we don’t need to get all the atomic details right - similar to how you don’t need to know the position and spin of every atom in a bar magnet to say something about how it interacts with a piece of iron. This project uses continuum mechanics to devlop and explore effective models of calving, continuum damage mechanics, to predict what the ice sheets will do in a changing climate.
Here’s a talk I gave on Continuum Damage Mechanics at the West Antarctic Ice Sheet Initiative 2021 Workshop:
Solid-Earth Feedbacks on Ice Dynamics
In this project we look at what’s going on deep beneath the ice, at the earth’s mantle, and what it does in response to all the ice moving around on the Earth’s surface, and then what that can mean for the ice’s flow. We’re finding that mantle motion can affect our forecasts of ice mass loss and sea level change even on the relatively short time-scales of decades and centuries, but many questions remain. We’re investigating how this might significantly affect our projections of sea-level in the future and quantifying what controls the magnitude of this feedback, in particular by looking into observations of ice sheets over the last 20 thousand years.
Here’s a talk I gave introducing these concepts and applying them to simulations of ice loss scenarios at Pine Island Glacier in West Antarctica:
Stochastic Calving Models
Iceberg calving events come in all shapes and sizes, and occur somewhat unpredictably, but appear to follow a common distribution that also explains, among other things, the earthquake Gutenberg-Richter distribution. Statistical physics suggests this may be due to a tendency for complex systems to evolve to a state of self-organized criticality, where each part of the system is in communication with all others at all length scales. Can we use this connection to postulate a better long-term, average calving law?
I’ve been advising UC Davis Undergraduate, Paige Brady, in this work. This plot going into her thesis shows how, with a simple, 1D ice model and a way of giving ice random defects (see below) can generate the flavor of calving events - frequency and size - observed at the Erebus Ice Tongue in East Antarctica:
Global Deformation and Sea Level Modeling
The dynamics of ice and the solid earth are drivers of sea level change and its unequal distribution will disproportionately affect already underserved communities. When ice flows or melts into the ocean, the effect on sea level isn’t the same everywhere, like it is when slowly filling up a bathtup. Instead, the whole globe changes shape (deforms), and gravity shifts around, making sea levels rise much faster in some places (like Florida and the Marshall Islands) than others (like Northern Canada and Scandinavia). We are developing open-source tools for computing global deformation and the effect on sea levels, with the aim of including these processes into Earth Systems Models, like the E3SM.
History and Ontology of Computational Science
The sciences in general, and climate science in particular, increasingly rely on computer simulations to explore complicated theories and datasets. As models attain a complexity commensurate with the complexity of nature, particularly salient with the iminent rollout of exascale computing facilities and ever-expanding machine learning, researchers have no choice but to fall back on their empirical intuitions and methods to understand them. What are the consequences for observations and simulations looking more alike than ever? Can it give us insights into how we establish uncertainties in our models and projections? I combine the historical record of model predictions - how they and their assessed uncertainties change - with philosophy of science to unpack these questions.
Also see my cv or my github profile for currently active projects..