Source-linked AI summary
SeisMamba: Low-Latency Single-Station Seismic Magnitude Estimation for Spatially Distributed Earthquake Early Warning
Quenton Yeo, Zhaoge Bi, Linghan Huang, Luke Stephen Higgins, Flora Salim, Huaming Chen
TL;DR
Earthquake early warning often relies on dense regional networks, while low-latency single-station magnitude estimation under regional deployment shift remains underexplored. SeisMamba addresses this gap with selective state-space waveform modelling, achieving the best MSE, RMSE, and R^2 among tested models while preserving low inference latency.
Problem
Low-latency single-station magnitude estimation under regional deployment shift remains underexplored despite sparse regions’ need for low-cost earthquake early warning.
Method
SeisMamba combines hierarchical convolutional encoding, sparse Mamba blocks, multi-scale feature fusion, and auxiliary temporal supervision for single-station waveform magnitude estimation.
Results
SeisMamba achieves the best MSE, RMSE, and R^2 among tested models on STEAD, with 0.55 ms inference time per batch.
Takeaways & Limitations
Mamba-based waveform models are promising components for scalable, low-cost, spatially distributed earthquake early warning.
Takeaways & Limitations
Transfer to unseen seismic regions remains affected by attenuation, site effects, source properties, and station conditions.
Abstract
from arXiv · showhide
Rapid earthquake magnitude estimation is central to earthquake early warning, yet many operational systems depend on dense regional seismic networks and region-specific calibration. This creates a spatial coverage barrier for high-risk areas with sparse sensing infrastructure. Single-station learning offers a lower-cost alternative, but existing models often face an accuracy--latency trade-off and may degrade under regional distribution shift. We present SeisMamba, a lightweight Mamba-based architecture for low-latency magnitude estimation from minimally processed three-component seismic waveforms recorded at a single station. SeisMamba combines hierarchical convolutional encoding, sparse selective state-space modelling, multi-scale feature fusion, and an auxiliary temporal prediction head to support efficient long-sequence waveform analysis. On the STEAD benchmark, SeisMamba achieves the best MSE, RMSE, and $R^2$ among tested baselines while requiring only 0.55 ms for a batch of 32 waveforms on an NVIDIA T4 GPU, making it about three times faster than transformer-based baselines. We further conduct a Chile--Taiwan regional hold-out experiment as a diagnostic test of cross-region deployment, where SeisMamba retains useful performance on geographically unseen seismic regions. These results suggest that selective state-space waveform modelling provides a promising accuracy--latency backbone for spatially distributed, low-cost earthquake early warning.
1 Introduction
Single-station earthquake early warning can reduce the spatial coverage barrier created by dense-network dependence, but requires low-latency inference and robustness to regional distribution shift. SeisMamba is proposed as a compact Mamba-based architecture designed to balance efficient local inference, accuracy, and cross-region deployment evaluation.
- Motivation: Dense regional networks, historical catalogues, and region-specific calibration create a spatial coverage barrier for high-risk regions with sparse monitoring infrastructure.This motivates low-cost magnitude estimation from three-component waveforms at a single station.
- Challenges: Single-station EEW requires spatially distributed sensing, low-latency local inference, and robustness to regional distribution shift.Models must remain useful when deployed in regions with waveform characteristics differing from training data.
- Research gap: The potential of low-latency single-station magnitude estimation under regional deployment shift remains underexplored.
- Contribution: SeisMamba is proposed as a compact Mamba-based architecture for low-cost single-station earthquake magnitude estimation.Its design targets spatially distributed EEW requirements rather than treating magnitude estimation as a standalone waveform regression task.
- Contribution: SeisMamba combines convolutional waveform encoding with sparsely placed selective state-space blocks and multi-scale feature fusion.
2 Related Work
Prior work reduces earthquake early warning’s dependence on dense regional networks through single-station waveform-based magnitude estimation. Efficient sequence modelling has progressed from quadratic-complexity Transformers toward linear-time selective state-space models such as Mamba.
- Single-station seismic magnitude estimation: Single-station EEW estimates earthquake properties from local waveform observations, reducing reliance on dense regional networks.Early learning-based methods used engineered P-wave features, transfer learning, or auxiliary source information.
- Single-station seismic magnitude estimation: MagNet demonstrated that neural models can estimate earthquake magnitude directly from raw waveform data.This followed earlier approaches based on engineered features, transfer learning, or auxiliary source information.
- Efficient sequence modelling for spatially distributed seismic sensing: Transformers model long-range sequences effectively, but attention has quadratic complexity with sequence length.This complexity motivates alternatives for efficient processing of long seismic waveforms.
- Efficient sequence modelling for spatially distributed seismic sensing: Mamba offers a selective state-space alternative with linear-time scaling and input-dependent state transitions.In seismology, PhaseMamba indicates that Mamba-style models can capture useful waveform structure for phase picking.
3 Method
SeisMamba addresses single-station magnitude estimation from minimally processed three-component waveforms under requirements for accurate regression, low latency, and geographic robustness. Its architecture combines hierarchical convolutions, sparse Mamba sequence modelling, multi-scale fusion, and scalar and temporal prediction heads.
- Task formulation: Given a 30-second ENZ waveform sampled at 100 Hz, SeisMamba predicts a scalar earthquake magnitude from a single station.The deployment setting assumes inference near one sensing device.
- Design requirements: The model targets accurate magnitude regression, low inference latency, and robustness across geographically different seismic regions.STEAD measures average predictive performance, while regional hold-out testing evaluates representation usefulness in unseen regions.
- Architecture: Its encoder uses hierarchical one-dimensional convolutions, sparsely placed Mamba blocks, multi-scale feature fusion, and two prediction heads.Convolutions extract local waveform features, while sparse Mamba blocks model longer temporal context.
- Efficient sequence modelling: Mamba blocks operate only at selected encoder resolutions after convolutions shorten the sequences, with bypass paths avoiding unnecessary long-context processing.This sparse placement is motivated by efficiency and preserves direct propagation of local convolutional features where long-context modelling is unnecessary.
- Prediction heads: The scalar head concatenates pooled descriptors across encoder depths, while an auxiliary temporal head predicts a P-wave-aligned magnitude trajectory.The scalar output is used for benchmark comparison; temporal targets are zero before P-wave arrival and equal to event magnitude afterwards.
4 Experiments
Experiments on STEAD evaluate SeisMamba under minimally processed single-station waveform inputs against representative seismic models, measuring accuracy and latency. A geographically unseen Chile–Taiwan hold-out and ablations further examine deployment robustness and architectural contributions.
- STEAD benchmark: On STEAD, SeisMamba achieves the best MSE, RMSE, and R^2 among tested models while remaining close to the best MAE.The benchmark uses 30-second three-component waveforms with only a 1–40 Hz fourth-order zerophase Butterworth band-pass filter and no hand-crafted P-wave features, source-location inputs, or station-specific calibration.
- STEAD benchmark: 0.55 ms per batch is SeisMamba’s STEAD inference time on an NVIDIA T4 GPU, slightly slower than PhaseNet but substantially faster than EQTransformer, AMAG, and U-Mamba.The comparison includes MagNet, PhaseNet, EQTransformer, AMAG, and U-Mamba; PhaseNet and EQTransformer use adapted regression heads.
- Regional hold-out: 5,478 recordings are excluded from training in the Chile–Taiwan regional hold-out, creating testing on geographically unseen seismic regions.The held-out split contains a broad magnitude range, including a substantial proportion of medium-to-large events.
- Regional hold-out: R^2 = 0.8518 and MAE = 0.2808 are achieved on the harder Chile–Taiwan hold-out, with latency remaining nearly unchanged at 0.58 ms.The observed performance drop from the standard benchmark is attributed to regional variation in attenuation, site effects, source properties, and station conditions.
- Ablation study: Ablations show that deeper encoders increase latency, dense Mamba placement performs worse than sparse placement, and removing multi-scale fusion degrades accuracy.These results indicate that state-space modelling is most useful after local waveform compression and that shallow and deep features provide complementary magnitude evidence.
5 Discussion
The discussion identifies selective state-space modelling as a promising low-latency backbone for single-station earthquake early warning, while emphasizing unresolved regional transfer limitations. It also frames the auxiliary temporal head as an observability mechanism rather than a complete uncertainty estimator.
- Selective state-space modelling can improve single-station magnitude accuracy while preserving low inference latency for spatially distributed earthquake early warning.This supports deploying models near individual stations.
- Regional waveform shift remains unresolved, as attenuation, site effects, source properties, and station conditions affect transfer to unseen seismic regions.The Chile–Taiwan hold-out revealed a performance drop under regional distribution shift.
- The auxiliary temporal head provides time-resolved magnitude supervision and shows how estimates evolve as waveform evidence accumulates, but it is not a complete uncertainty estimator.Operational alerting would still require calibrated uncertainty estimates and decision rules.
6 Conclusion
The paper presents SeisMamba as a lightweight selective state-space architecture for single-station earthquake magnitude estimation. Its design combines local encoding, sparse Mamba modelling, multi-scale fusion, and auxiliary temporal supervision to improve the accuracy–latency trade-off while retaining useful regional hold-out performance.
- SeisMamba is a lightweight selective state-space architecture for single-station earthquake magnitude estimation.
- The architecture combines local waveform encoding, sparse Mamba sequence modelling, multi-scale fusion, and auxiliary temporal supervision.
- SeisMamba improves the accuracy–latency trade-off on a global seismic benchmark while retaining useful performance under a regional hold-out setting.