← Back to all news
arXiv Machine Learning By Zhao Meng, Yinan Cai, Siru Zhong, Juepeng Zheng, Haohuan Fu

GeoPrior-Mamba: Structured Process Priors with Mamba for Fine-Resolution XCO2 Reconstruction

Read the original on arXiv Machine Learning →

GeoPrior-Mamba is a new multi-directional Mamba framework that reconstructs fine‑resolution XCO2 fields from sparse satellite data by incorporating structured process priors derived from a language model. These priors encode biospheric uptake, ecosystem respiration, and anthropogenic emissions into deterministic tables, which are spatially instantiated with geographic, ecological, emission, and seasonal information and injected into the model via a lightweight knowledge adapter. On OCO‑2 data from 2018‑2020, the method achieves an RMSE of 0.81 ppm and an R² of 0.93, outperforming CAMS background interpolation and Trans‑XCO2, and shows faster convergence and consistency with ground‑based TCCON measurements.

Machine-generated by The Flow from the publisher's headline and feed description — not written or checked by a human. The full article lives at arXiv Machine Learning.

Related stories

arXiv Machine Learning
Aug 18

Earth Observation Foundation Models for Terrestrial Ecohydrology: From Representation Learning to Process Inference

arXiv:2608. 15282v1 Announce Type: new Abstract: Earth observation foundation models (EOFMs) are emerging as reusable representation frameworks for data-driven retrieval, prediction and process modelling within ecohydrology, which integrate EO, meteorological forcing and process models to characterise coupled water, energy and carbon dynamics in vegetation and soil across scales.

By Yi Yu, Jian Peng, Yucheng Lin, Trevor F. Keenan, Thomas F. A. Bishop
benchmarks
More like this →
arXiv Machine Learning
Sep 25

Generative Atmospheric Super-Resolution from Heterogeneous In Situ Observations through Composable Interfaces

The paper presents a method for improving atmospheric state reconstructions by conditioning a pretrained diffusion model on heterogeneous in‑situ observations. It introduces composable interfaces that transform sparse radiosonde, clustered aircraft, and dense surface‑station data into likelihood factors, allowing these diverse sources to guide posterior sampling consistently. Using 2019 data to build the interfaces and evaluating them in 2020, the combined R+A+S approach reduces RMSE by 9.24% and improves CRPS compared to using radiosondes alone, demonstrating a modular way to incorporate varied observations without retraining the model.

By Yang Xu, Dibyajyoti Chakraborty, Haiwen Guan, Sen Wang, Romit Maulik
diffusion
More like this →