Generative Atmospheric Super-Resolution from Heterogeneous In Situ Observations through Composable Interfaces
Abstract
Atmospheric observations are sparse, heterogeneous, and unevenly distributed, whereas many generative atmospheric models learn distributions over regularly gridded multivariate states. Once pretrained, diffusion models can supply atmospheric priors that can be combined with observation-derived likelihood factors in a Bayesian formulation. However, these observation sources differ substantially in geometry and sampling density, complicating the consistent use of their observations within a common inference framework. Here, we formulate this reconstruction problem as generative atmospheric super-resolution and introduce composable observation interfaces for conditioning a single pretrained 13-variable atmospheric diffusion model. The interfaces convert sparse radiosonde (R), clustered aircraft (A), and dense irregular surface-station (S) observations into source-specific likelihood factors that specify where observations constrain the gridded state, how residuals are counted under uneven sampling, and how strongly each source guides posterior sampling. We developed the aircraft and surface observation interfaces using 2019 observations and evaluated the selected interfaces throughout 2020 without further tuning. Compared with reconstructions conditioned only on radiosonde observations, the composed R+A+S interface reduces RMSE evaluated against ERA5 by 9.24% across all 13 state variables over the CONUS domain. The aircraft and surface factors provide complementary improvements in upper-air and surface variables. The R+A+S combination also lowers the Continuous Ranked Probability Score (CRPS), while evaluations at held-out aircraft and surface-station observations show reduced prediction errors. Together, these results demonstrate a modular route for conditioning a pretrained atmospheric generative prior on heterogeneous in situ observations without retraining the underlying model.
Related
- NeuPlume: generative inversion of atmospheric point-source emissions from sparse observations
- 4D sensitivity kernels for infrasound atmospheric propagation: considering the need and impact of time dependent observation operators for acoustic wave observations in data assimilation methods
- Hydrometeorological Consistency and Structural Differences in High-Resolution Atmospheric Forcing Products over the Gulf Coast
- Recurrent High-Resolution Satellite Observations Quantify Facility-Resolved Annual Large-Emitter Methane Emissions in the Permian Basin
- The Harmony Scientific Workbench-Part 1: Modeling High-Resolution Multistatic Radar Observations of the Ocean
BibTeX
@misc{xu2026generative,
title = {{Generative Atmospheric Super-Resolution from Heterogeneous In Situ Observations through Composable Interfaces}},
author = {Yang Xu and Dibyajyoti Chakraborty and Haiwen Guan and Sen Wang and Romit Maulik},
howpublished = {arXiv},
year = {2026},
month = sep,
doi = {10.48550/arxiv.2609.29027},
eprint = {2609.29027},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2609.29027},
note = {CC BY 4.0},
}
RIS
TY - UNPB
TI - Generative Atmospheric Super-Resolution from Heterogeneous In Situ Observations through Composable Interfaces
AU - Yang Xu
AU - Dibyajyoti Chakraborty
AU - Haiwen Guan
AU - Sen Wang
AU - Romit Maulik
PY - 2026
DA - 2026/09/29
PB - arXiv
DO - 10.48550/arxiv.2609.29027
UR - https://arxiv.org/abs/2609.29027
AB - Atmospheric observations are sparse, heterogeneous, and unevenly distributed, whereas many generative atmospheric models learn distributions over regularly gridded multivariate states. Once pretrained, diffusion models can supply atmospheric priors that can be combined with observation-derived likelihood factors in a Bayesian formulation. However, these observation sources differ substantially in geometry and sampling density, complicating the consistent use of their observations within a common inference framework. Here, we formulate this reconstruction problem as generative atmospheric super-resolution and introduce composable observation interfaces for conditioning a single pretrained 13-variable atmospheric diffusion model. The interfaces convert sparse radiosonde (R), clustered aircraft (A), and dense irregular surface-station (S) observations into source-specific likelihood factors that specify where observations constrain the gridded state, how residuals are counted under uneven sampling, and how strongly each source guides posterior sampling. We developed the aircraft and surface observation interfaces using 2019 observations and evaluated the selected interfaces throughout 2020 without further tuning. Compared with reconstructions conditioned only on radiosonde observations, the composed R+A+S interface reduces RMSE evaluated against ERA5 by 9.24% across all 13 state variables over the CONUS domain. The aircraft and surface factors provide complementary improvements in upper-air and surface variables. The R+A+S combination also lowers the Continuous Ranked Probability Score (CRPS), while evaluations at held-out aircraft and surface-station observations show reduced prediction errors. Together, these results demonstrate a modular route for conditioning a pretrained atmospheric generative prior on heterogeneous in situ observations without retraining the underlying model.
ER -