The distribution of radiant energy throughout the atmosphere and at the surface drives every wind that blows and leads to every drop of rain that falls. We work to make state-of-the-art understanding available in climate, weather, and research models of the earth system through a combination of physical modeling and machine learning.
The flow of radiation through the earth system is among the best-understood processes in atmospheric physics. Representing those flows accurately, though, presents a host of interesting conceptual and computational challenges.
With a growing network of collaborators around the world we develop RTE+RRTMGP, a software toolbox for computing the flow of energy in planetary atmospheres. By encapsulating conceptually-distinct tasks the toolbox can be used to solve a wide range of problems on a wide range of computing architectures.
We organize a international consortium in the US and Germany working to streamline access to cutting-edge information about how gases interact with radiation, and to summarize ("parameterize") that information to best match the cost and accuracy for the user's problem. One outcome has been a new method for spectral integration that's simpler and more transparent than k-distributions:
- Paulina Czarnecki, Lorenzo M. Polvani, and Robert Pincus, 2023. Sparse, empirically optimized quadrature for broadband spectral integration. J. Adv. Modeling Earth Syst., 15, e2023MS003819. doi:10.1029/2023MS003819
Recent work carefully linking the two-stream equations for angularly-integrated flux to the radiative transfer equation for directional intensity highlighted the impact of assumptions made for more than a century, leading to more accurate and more physically-consistent calculations.
-
Ho, D. J. X. and R. Pincus, 2024. Two-streams revisited: General equations, exact coefficients, and optimized closures. Submitted to J. Adv. Model. Earth Syst., Jun 2024. Preprint at doi:10.22541/essoar.171867251.13739862/v1