Research Interests

My research interests are centered around using statistical, data-driven, and machine learning models to study and model a variety of phenomena in weather and climate. I enjoy finding new signals in big, complex, noisy geospatial datasets, and I love thinking about how we as Earth scientists use these tools as much as I love applying them to exciting and relevant problems in weather and climate. Below, you can see some of the recurring themes that have turned up in my work throughout the years.
AI and Tropical Meteorology

Much of my work has been focused on developing machine learning models for various tropical cyclone forecasting tasks. Many of these models are developed with the end goal of producting operational forecast products for use at the National Hurricane Center or elsewhere at NOAA. Some of my work in this area includes:

  • Tropical Cyclone Artificial Neural-network Error Model: A neural network model that predicts tropical cyclone forecast errors for track and intensity; currently in the research to operations pipeline at the National Hurricane Center. See Mark DeMaria's conference talk, Martin Fernandez's paper in Artificial Intelligence in the Earth Systems, and the real-time forecasts at CIRA.
  • TC Structure in Synthetic Microwave Imagery: Using synthetic passive microwave data to study tropical cyclone evolution and structure at high time resolutions that aren’t available with observed satellite imagery. See my 2025 AMS talk; paper to be submitted soon!
  • TCBench: Effort led by the Universite de Lausanne to create a benchmarking dataset for tropical cyclone predictions in AI weather models; see preprint at arXiv
  • Tropical Cyclone Formation Probability: Using AI models to create extended global tropical cyclogenesis forecasts. See 2022 Hollings Scholar Marshall Baldwin's talk and the real-time TCFP page.
  • Causal discovery for hurricane forecasting: Dr. Saranya Ganesh Sudheesh developed a causal discovery model that identified potential new predictors for statistical hurricane intensity forecasts. Her work identified 6 potential new predictors that are currently being tested in existing operational tropical cyclone intensity forecast models; see preprint at arXiv and forthcoming manuscript in Weather and Forecasting.

    • AI Models in Weather and Climate Science

    I use AI and data-driven models to investigate a variety of problems in weather and climate science. In addition to science tasks, I am also interested in improving the AI models themselves. Some of my work in this area includes:

  • Machine learning and climate modeling: I have been part of a working group that identifies challenges and opportunities in machine learning and climate modeling. This group's work is ongoing as part of a series of Gordon Research Conferences.
  • Developing benchmark datasets for AI: graduate student Milton Gomez spearheaded an effort to develop a benchmark dataset for AI tropical cyclone forecast models.
  • Bias in AI: As part of the NSF Institute for Trustworthy AI in Weather, Climate, and Coastal Oceanography, we put forth guidelines and warnings about forms of bias in AI models for weather and climate tasks.
  • Uncertainty quantification for neural networks: Along with my colleagues in the CIRA-ML group, we evaluated a variety of uncertainty quantification techniques for neural network tasks in the environmental sciences.

    • Climate Variability and Change

    I have studied many aspects of climate variability and change using reanalysis products, simple atmospheric models, fully-coupled global climate models, and targeted climate modeling simulations. Some of the topics I have studied include:

  • Jet stream variability in a changing climate: I have studied how the jet stream moves and changes in response to a changing climate using idealized models, model hierarchies, and climate model simulations. I have also used causal inference modeling to understand the observed variability of the jet streams.
  • Polar climate variability: At the University of Washington, I studied the predictability of extreme sea ice loss in subseasonal to seasonal prediction models, as well as the impact of Arctic cylones on extreme sea ice loss. I have also modeled the climate-driven variability of Arctic moisture transport, and used causal inference modeling to better model Arctic - midlatitude atmospheric teleconnections.
  • Tropical climate variability: Before I was a tropical cyclone researcher, I joined Dr. Alex Gonzalez on an exploration of tropical climate variability by studying the dynamics of the springtime Intertropical Convergence Zone (ITCZ).

    • Climate Risk

    At Jupiter, my research has focused on quantifying physical and socioeconomic climate risk. Some of my work in this space includes:

  • Climate Tipping Points: We study the non-linear shock-like impacts of climate tipping points. We recently released an ebook detailing our findings on the coastal flood risk in the aftermath of an Atlantic Meridonal Overturning Circulation collapse. Spoiler: the economic impact is massive!.
  • Climate Risk Assessments: I have produced climate risk assessments for private sector customers and non-profits, such as the United Nations Development Program.
  • Climate Adaption and Climate Change: We model the impacts of climate change and climate adaptation on current events, such as the Winter Olympics. We also study the impacts of historic events such as Hurricane Katrina, and quantify what has--and hasn't--changed after 20 years of adaptation.

    • Causality, Interpretability, and Uncertainty in Climate Science

    I have been interested in causal inference modeling and climate science since my graduate student days. I have also given many talks on this topic (see my Talks page!). Some of the projects I have worked on include:

  • Interpeting AI-Generated Satellite Data for Science Discovery: I led an effort to evaluate the interpretability of AI-generated satellite data by using it to understand tropical cyclone dynamics and evolution (paper coming soon!).
  • Causal inference in hurricane forecasting: Dr. Saranya Ganesh Sudheesh led a study that used causal inference modeling to improve hurricane intensity forecasts.
  • Uncertainty quantification for neural networks: Along with my colleagues in the CIRA-ML group, we evaluated a variety of uncertainty quantification techniques for neural network tasks in the environmental sciences.

    • Extreme Event Analysis

    In the course of my research, I have worked on several types of extreme event modeling and analysis, including extreme sea ice loss, and tropical cyclone rapid intensification. These events are rare, but high consequence, meaning they are difficult (but important) to study. You can also check out my blogs at Jupiter Research to see discussion of current extreme events, such as atmospheric rivers, , extreme floods and extreme heat.

    AI and Geoscience Education

    I have worked on several initiatives to improve AI education for geoscientists, and to help AI practitioners understand Earth sciences. Some of my efforts on this front include developing Learning Journeys for the NOAA Center for Artificial Intelligence, and participating as a mentor for the ClimateChangeAI Summer School, where my team was awarded a Spotlight Presentation at the NeurIPS Tackling Climate Change with Machine Learning Workshop in 2022.