Generative model improves atmospheric data from sparse observations
Generative Atmospheric Super-Resolution from Heterogeneous In Situ Observations through Composable Interfaces
Machine Learning
Summary
Atmospheric data is often sparse and collected from different sources like weather balloons, airplanes, and surface stations. The authors designed a system that uses a pretrained model combined with special methods to include these different types of observations together. This produces better weather data with more detail across many atmospheric variables. Their approach works without retraining the whole model and improves accuracy compared to using only one type of observation.
What this means in practice
- •For weather forecasters: Enhance weather prediction by integrating diverse observational data sources using one pretrained atmospheric generative model.
- •For climate monitoring teams: Generate high-resolution atmospheric states from mixed types of in situ observations for improved climate data products.
Authors
Yang Xu, Dibyajyoti Chakraborty, Haiwen Guan, Sen Wang, Romit Maulik
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.