The Fast and Fourier: Using neural operators to speed up radiative modeling of protoplanetary disks
Year: 2026
Type: article
Abstract: Edge-on protoplanetary disks provide a spatially resolved view of the vertical distribution of icy planetary building blocks, but inferring their relative abundances requires radiative-transfer (RT) modeling with multiple anisotropic scattering. Each such model costs 10^1-10^3 CPU hours, which precludes principled Bayesian inference and homogeneous, population-scale analyses. We develop and validate a fast and reusable surrogate for Bayesian inference with ice-bearing edge-on disk models. We trained a Fourier neural operator on a grid of spectra. The network maps disk parameters and analytic physical input fields to the spectrum, using mixed-resolution spectra. We used the surrogate in an explicit-likelihood nested sampler. Then we validated it on held-out simulations and synthetic recovery tests with Test of Accuracy with Random Points (TARP) coverage tests. Finally, we used it to retrieve the disk structure and ice abundances of ESO-Hα 569 from its spectral energy distribution (SED). RADMC-3D The surrogate successfully reproduces the spectra with a median error of fwdPerSpecQFifty%. We recover synthetic parameters without detectable bias and with nominal coverage. For ESO-Hα 569, the posterior favors a massive (M_ RADMC-3D ̊m gas = nearly edge-on (i= . Its ice abundances and cavity size remain weakly constrained by broadband data alone. The marginal spectrum-evaluation speed-up is approximately 10^9--10^10. disk consistent with the image-plus-SED model of wolff2017hubblespacetelescope Neural-operator emulation makes full posterior inference feasible within the adopted model and training prior. Scientifically important posterior points should still be checked with high-fidelity RT models. By dramatically decreasing the evaluation time, for the first time this approach enables homogeneous population studies of ices in disks, which can be capitalized on for new JWST surveys to map the range of compositions within planetary nurseries.
Source: Astronomy and Astrophysics
Authors: L. Welzel, M. K. McClure
Language: English
Subfield: Astronomy and Astrophysics
Field: Physics and Astronomy
Domain: Physical Sciences
Open Access status: gold
Estimated APC paid: $1,865
Awards: G3260597334, G9001172342
OpenAlex ID: W7221625611