EXPLAINABLE MACHINE LEARNING
Machine learning, a branch of artificial intelligence, has allowed climate scientists to better understand nonlinearities in the Earth system. By opening the black box of machine learning to explain how and what the model learned, we are able to build trust in our models and learn new science! My research applies XAI (explainable AI) techniques to quantify predictability on subseasonal to decadal timescales and identify forecasts of opportunity.
​
Click here to access a comprehensive tutorial created by the Barnes Group on applying machine learning to climate science.
analyticsvidhya.com
CLIMATE VARIABILITY AND PREDICTABILITY
My research focuses on states of the climate system, known as forecast windows of opportunity, that are more predictable and lead to higher forecast skill. Sources of predictability typically come from modes of climate variability (e.g. MJO, ENSO, AMO) and I research sources of climate predictability from subseasonal
(2 weeks - 2 months) through decadal timescales.
TROPICAL - EXTRATROPICAL TELECONNECTIONS
Teleconnections are large-scale pressure and circulation patterns internal to the climate system which connect the globe through atmospheric and oceanic waves. My research primarily analyzes waves that travel from the tropics to the midlatitudes, and how their signals can affect conditions in North America, such as precipitation and temperature.
CLIMATE MODELING
Climate models run on computers using differential equations that describe physical laws to simulate the climate system. They use data to describe processes in the atmosphere, ocean, land, ice, etc. as well as interactions between Earth system components. Climate models range from simple conceptual models to complex, intricate general circulation models. My research incorporates a hierarchy of climate models to better understand climate dynamics.
str.llnl.gov
climatecentral.org
SEA LEVEL RISE AND COASTAL FLOODING
Under a continuously warming climate, sea level rise is increasingly threatening coastal communities, with many coastal and island communities already feeling the threats of sea level rise. My research looks at remote influences on U.S. coastal sea levels as well as future projections of localized flooding scenarios.
Our approach uses common impact flooding thresholds which allows the public access to information that is digestible and pertinent to their local communities.
