Latent Twins: A Framework for Scene Recognition and Fast Radiative Transfer Inversion in FORUM All-Sky Observations

TL;DR

This paper introduces a physics-aware, data-driven inversion framework using latent twins for rapid, accurate scene recognition and atmospheric variable retrieval from all-sky infrared observations, enhancing real-time climate monitoring capabilities.

Contribution

The work presents a novel latent twin autoencoder approach for fast, physically consistent atmospheric and cloud variable retrievals from FORUM all-sky data, improving upon traditional methods.

Findings

Demonstrates accurate retrievals on synthetic data

Enables near-instantaneous scene classification

Provides physically plausible cloud reconstructions

Abstract

The FORUM (Far-infrared Outgoing Radiation Understanding and Monitoring) mission will provide, for the first time, systematic far-infrared spectral measurements of Earth's outgoing radiation, enabling improved understanding of atmospheric processes and the radiation budget. Retrieving atmospheric states from these observations constitutes a high-dimensional, ill-posed inverse problem, particularly under cloudy-sky conditions where multiple-scattering effects are present. In this work, we develop a data-driven, physics-aware inversion framework for FORUM all-sky retrievals based on latent twins: coupled autoencoders for atmospheric states and spectra, combined with bidirectional latent-space mappings. A lightweight model-consistency correction ensures physically plausible cloud variable reconstructions. The resulting framework demonstrates potential for retrievals of atmospheric, cloud and surface variables, providing information that can serve as a prior, initial guess, or surrogate for computationally expensive full-physics inversion methods. It also enables robust scene classification and near-instantaneous inference, making it suitable for operational near-real-time applications. We demonstrate its performance on synthetic FORUM-like data and discuss implications for future data assimilation and climate studies.