Frequency-Enhanced Hilbert Scanning Mamba for Short-Term Arctic Sea Ice Concentration Prediction

TL;DR

This paper introduces FH-Mamba, a novel framework combining Hilbert scanning, wavelet transform, and attention mechanisms to improve short-term Arctic sea ice concentration prediction by capturing spatial, temporal, and frequency features.

Contribution

The paper proposes a Frequency-enhanced Hilbert scanning Mamba framework that enhances sequence modeling for Arctic SIC prediction through a 3D Hilbert scan, wavelet transform, and hybrid attention.

Findings

Outperforms state-of-the-art baselines on OSI-450a1 and AMSR2 datasets

Improves temporal consistency and edge reconstruction in SIC forecasting

Demonstrates the effectiveness of Hilbert scanning and frequency-aware attention mechanisms

Abstract

While Mamba models offer efficient sequence modeling, vanilla versions struggle with temporal correlations and boundary details in Arctic sea ice concentration (SIC) prediction. To address these limitations, we propose Frequency-enhanced Hilbert scanning Mamba Framework (FH-Mamba) for short-term Arctic SIC prediction. Specifically, we introduce a 3D Hilbert scan mechanism that traverses the 3D spatiotemporal grid along a locality-preserving path, ensuring that adjacent indices in the flattened sequence correspond to neighboring voxels in both spatial and temporal dimensions. Additionally, we incorporate wavelet transform to amplify high-frequency details and we also design a Hybrid Shuffle Attention module to adaptively aggregate sequence and frequency features. Experiments conducted on the OSI-450a1 and AMSR2 datasets demonstrate that our FH-Mamba achieves superior prediction performance compared with state-of-the-art baselines. The results confirm the effectiveness of Hilbert scanning and frequency-aware attention in improving both temporal consistency and edge reconstruction for Arctic SIC forecasting. Our codes are publicly available at https://github.com/oucailab/FH-Mamba.