.png){kind=icon}
.png){kind=icon}
Return
Physics-Informed Ensemble Learning for Detection of Malfunctioning Ocean Bottom Pressure Gauges: A Case Study on the S-Net Network
.png){kind=icon}
PRE
AI
O
N
S
A
DOI：10.1109/joe.2026.3686725.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
In this study, we analyze bottom pressure time series from 139 operational stations of the S-net network to detect potential anomalies. As of 12 March 2025, 15 stations were identified as malfunctioning: S1N09, S1N13, S1N21, S2N06, S2N18, S2N20, S2N25, S3N02, S3N04, S3N22, S4N10, S4N24, S5N10, S5N12, and S5N16. Notably, the failure of station S2N25 was captured in real time during observations. The analysis revealed that increased depth is statistically associated with a higher risk of sensor malfunction (<inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><tex-math notation="LaTeX">$p < 0.05$</tex-math></inline-formula>). To enable automatic identification of malfunctioning pressure gauges, we introduce an ensemble learning model that combines three complementary reconstruction techniques—Seasonal Auto Regressive Integrated Moving Average forecasting, physics-based tidal prediction using the TPXO model, and interstation similarity analysis—evaluated with two error metrics (mean squared error and DTW). Their outputs are integrated by a lightweight random-forest meta-classifier, yielding an interpretable and data-efficient hybrid model. The model aggregates prior knowledge of the physical characteristics of the signal with machine learning to classify stations as either operational or malfunctioning. The average MCC-score on the test set reached 0.90 with a standard deviation of 0.06. Because the proposed model is fully interpretable and computationally inexpensive, it is suitable for real-time deployment within existing tsunami warning infrastructures.
Keywords:
Anomaly detection
ensemble classifier
pressure gauge (PG)
random forest
S-Net
seafloor observatory
sensor testing
Journal
IF:
5.3
Papers:
2.6K
Citations:
7.4K
Organization
Cited Papers