Kaitlyn Wen ’26, Computer Science

Kaitlyn Wen ’26

Major

Computer Science

Project Title

Quantifying Spread in Radiative Feedbacks Due to Differences in Kernels

Minor(s): Cognitive Science, Quantitative Economics

Equilibrium climate sensitivity (ECS), which describes how much the Earth will warm if atmospheric CO₂ concentrations double from levels in the year 1850, is one of the most important metrics in climate science. Estimates of ECS have a wide uncertainty range that is commonly attributed to individual radiative feedbacks. Scientists typically compute feedbacks with the help of climate models using the radiative “kernel” technique. These “kernels” are pre-calculated quantities commonly shared among research groups. However, the existing literature focuses on the spread of feedbacks across different models and the potential impact of kernel choice remains to be explored. I examined the influence of 11 different radiative kernels on the spread across feedback estimates from a set of 49 climate models. With the help of the Python package ClimKern, I performed feedback calculations for each model using different kernels and compared the cross-kernel and cross-model spread of estimates. I found that up to 50% of the intermodel spread across globally averaged individual feedbacks can be reproduced by using different kernels. Working on this project was incredibly enjoyable; it allowed me to explore the positive impacts of coding in a new context, which reinforced my passion for studying computer science.