Work

Refined Reconstruction of Multimodal Mesospheric Temperature Fields Based on Diffusion Model

Year: 2025

Type: article

Abstract: Mesosphere is the core region of atmospheric wave activities, where fine temperature data are essential for analyzing its dynamic processes. We use an interpretable Data Diffusion Fusion Model to fuse sparse Sounding of the Atmosphere using Broadband Emission Radiometry (SABER) observations with coarse‐resolution Modern‐Era Retrospective Analysis for Research and Applications V2. By combining spatiotemporal properties, F10.7 and Kp, the mesospheric temperature is reconstructed with high‐precision, increasing resolution by 5 times. Results show that Denoising Diffusion Fusion Model outputs incorporates more information from SABER data, improving the Peak Signal‐to‐Noise Ratio by 5.02 dB. Our reconstruction model demonstrates excellent global stability ( R 2 > 0.94). In the January 2019 Sudden Stratospheric Warming event, the model not only accurately reproduces the small‐scale features, but also significantly reduces the deviation from 21.7% to 14.9% with SABER data. This outcome fully proves its superior performance in handling extreme events. Importantly, even in the absence of observations, our model can still deliver high‐precision data set to support scientific investigations.

Source: Geophysical Research Letters

DOI: 10.1029/2025gl118581

Authors: Huanwei Zhang, Yang He, Zheng Sheng, Haoyang Sui

Institution: National University of Defense Technology

Language: English

Topic: Ionosphere and magnetosphere dynamics

Subfield: Astronomy and Astrophysics

Field: Physics and Astronomy

Domain: Physical Sciences

Open Access status: gold

Estimated APC paid: $3,190

Funders: Strong, National Natural Science Foundation of China, Chinese Academy of Sciences, National University of Defense Technology