SURF 2026

I have a possible project for a SURF student in the area of data-driven methods for application in the earth system. If interested, please contact me indicating which project(s) you would want to work on. See here for more information about SURF.

Analyzing analogue forecasting methods for climate attribution:

Analogue forecasting tools — which make predictions (or forecasts) based on previous observations of similar states (analogues) — are used in data-driven climate attribution tasks as well as the prediction tasks for which they are named. Despite the growing usage of these methods, there is not yet robust theory or guidelines for when analogue forecasting methods will give accurate measurements of the effects of changing forcing or parameters. As such, it is not clear how accruate these tools can be for measuring the effect of climate change on phenomena like wildfires.

This project will analyze and compare various forms of analogue forecasting methods (including linear, kernel, and deep-learning based methods) for use in the climate attribution problem, asking when these methods can confidently be applied and where they differ. The student working on these problem will implement these methods in both simple systems (where the answer is known) and in climate data (where the true answer is not). Additionally, if interested, the student working on this problem will have the opportunity to develop theory which characterizes the abilities of these methods to solve attribution problems.

Relevant papers

• https://www.pnas.org/doi/10.1073/pnas.2111875118
• https://doi.org/10.5194/wcd-6-817-2025
• https://doi.org/10.1029/2019MS001958

Student Qualifications

Students will need to have a background in linear algebra, (ordinary) differential equations, and coding proficiency — either in Python or Julia. It will be necessary to have an interest in the project, and it will be helpful (but not necessary) to have coursework in dynamical systems, optimization, climate dynamics, and data-driven methods.