SENDAI: A Hierarchical Sparse-measurement, EfficieNt Data AssImilation Framework

Xingyue Zhang, Yuxuan Bao, Mars Liyao Gao, J. Nathan Kutz

Keywords: data assimilation, remote sensing, sim2real, multi-scale modeling, sparse observation

TL;DR: We present SENDAI, a hierarchical Sparse-measurement, EfficieNt Data AssImilation Framework that reconstructs full spatial states from hyper sparse sensor observations by combining simulation-derived priors with learned discrepancy corrections.

Abstract: Bridging the gap between data-rich training regimes and observation-sparse deployment conditions remains a central challenge in spatiotemporal field reconstruction, particularly when target domains exhibit distributional shifts, heterogeneous structure, and multi-scale dynamics absent from available training data. We present SENDAI, a hierarchical $\textbf{S}$parse-measurement, $\textbf{E}$fficie$\textbf{N}$t $\textbf{D}$ata $\textbf{A}$ss$\textbf{I}$milation Framework that reconstructs full spatial states from hyper-sparse sensor observations by combining simulation-derived priors with learned discrepancy corrections. We demonstrate performance on satellite remote sensing, reconstructing MODIS-derived vegetation index fields across globally distributed sites. Using seasonal periods as a proxy for domain shift, the framework consistently outperforms established baselines that require substantially denser observations - SENDAI achieves substantial SSIM improvements over traditional baselines and recent high-frequency-based methods. These gains are particularly pronounced for landscapes with sharp boundaries and sub-seasonal dynamics; more importantly, the framework effectively preserves diagnostically relevant structures - such as field topologies, land cover discontinuities, and spatial gradients. The results highlight a lightweight and operationally viable framework for sparse-measurement reconstruction applicable to physically grounded inference, resource-limited deployment, and real-time monitoring and control.

Submission Number: 65