GF4ACE -- Data from: Reanalysis-based global radiative response to sea surface temperature patterns: Evaluating the Ai2 climate emulator
Data files
Mar 04, 2025 version files 46.20 MB
-
GF4ACE_data.zip
46.19 MB
-
README.md
2.16 KB
Abstract
The sensitivity of the radiative flux at the top of the atmosphere to surface temperature perturbations cannot be directly observed. The relationship between sea surface temperature and top-of-atmosphere radiation can be estimated with Green's function simulations by locally perturbing the sea surface temperature boundary conditions in atmospheric climate models. We perform such simulations with the Ai2 Climate Emulator (ACE), a machine learning-based emulator trained on ERA5 reanalysis data (ACE2-ERA5). This produces a sensitivity map of the top-of-atmosphere radiative response to surface warming that aligns with our physical understanding of radiative feedback. However, ACE2-ERA5 likely underestimates the radiative response to historical warming. We argue that Green's function experiments can be used to evaluate the performance and limitations of machine learning-based climate emulators by examining if causal physical relationships are correctly represented and testing their capability for out-of-distribution predictions.
Senne Van Loon, Maria Rugenstein, & Elizabeth A. Barnes (2025), Reanalysis-based Global Radiative Response to Sea Surface Temperature Patterns: Evaluating the Ai2 Climate Emulator, arXiv:2502.10893
Data & Software:
https://doi.org/10.5061/dryad.d2547d8cf
https://doi.org/10.5281/zenodo.14963836
Data for GF4ACE
This repository contains initial conditions, patch perturbations and global-mean output data for:
Senne Van Loon, Maria Rugenstein, & Elizabeth A. Barnes (2025), Reanalysis-based Global Radiative Response to Sea Surface Temperature Patterns: Evaluating the Ai2 Climate Emulator, arXiv:2502.10893
Files and variables
File: GF4ACE_data.zip
Description: initial conditions, forcing, and output data
The zip file contains sample data to be used in combination with the software available on Zenodo, and is ordered in a file structure consistent with the code. In particular:
forcing
contains patch forcing files for each model used, for different patch amplitudes- Each folder in
forcing
denotes the version of ACE used (‘eam’, ‘era’, or ‘fv3’) patches_*.nc
contains SST patches of amplitude2K
,-2K
,4K
, or-4K
.- These patches were created with the
create_patches
function in the software available on on Zenodo
- Each folder in
initialization
contains initial conditions for the control run of each model- Each folder in
initialization
denotes the version of ACE used (‘eam’, ‘era’, or ‘fv3’) 20191231_1800.nc
contains the initial conditions for the control run
- Each folder in
output
contains output files of the Green’s function experiments- Each folder in
output
denotes the version of ACE used (‘eam’, ‘era’, or ‘fv3’) - pre-processed global-mean radiation from all patch simulations in
R_gm_monthly_*.nc
- Ocean masks for each version in
ocean_mask.nc
- Incoming solar radiation for each version in
rsdt.nc
- Each folder in