In this work, we exploited the ubiquitous seismic noise generated by energy transfer from the sea to the solid Earth (called microseism) to infer the significant wave height data, with the aim of developing a microseismbased monitoring system of the Sicily Channel. We used a combined approach based on statistical analysis and machine learning by using seismic and sea state data (provided by the hindcast maps), recorded between 2018 and 2021.Through spectral and amplitude analysis, we observed that microseism was influenced by the conditions of the seas surrounding Sicily. Correlation analysis demonstrates that microseism mostly originates from sources located up to 400 km from the coastlines. Moreover, employing machine learning algorithms, we successfully reconstruct spatial and temporal sea wave distributions using microseism data. Among the tested methods, the Random Forest algorithm yields the best results, with an R2 value of 0.89 and a mean prediction error of about 0.21 m.
Towards a monitoring system of the sea state based on microseism and machine learning
Minio V.
Primo
;Borzi A. M.;Saitta S.;Cannata A.;
2023-01-01
Abstract
In this work, we exploited the ubiquitous seismic noise generated by energy transfer from the sea to the solid Earth (called microseism) to infer the significant wave height data, with the aim of developing a microseismbased monitoring system of the Sicily Channel. We used a combined approach based on statistical analysis and machine learning by using seismic and sea state data (provided by the hindcast maps), recorded between 2018 and 2021.Through spectral and amplitude analysis, we observed that microseism was influenced by the conditions of the seas surrounding Sicily. Correlation analysis demonstrates that microseism mostly originates from sources located up to 400 km from the coastlines. Moreover, employing machine learning algorithms, we successfully reconstruct spatial and temporal sea wave distributions using microseism data. Among the tested methods, the Random Forest algorithm yields the best results, with an R2 value of 0.89 and a mean prediction error of about 0.21 m.File | Dimensione | Formato | |
---|---|---|---|
85_Minio+et+al.,+2023+EMS_compressed.pdf
accesso aperto
Tipologia:
Versione Editoriale (PDF)
Licenza:
Creative commons
Dimensione
1.06 MB
Formato
Adobe PDF
|
1.06 MB | Adobe PDF | Visualizza/Apri |
I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.