Unravelling the Relationship Between Microseisms and Spatial Distribution of Sea Wave Height by Statistical and Machine Learning Approaches

被引：15

作者：

Cannata, Andrea ^{[1
,2
]}

Cannavo, Flavio ^{[2
]}

Moschella, Salvatore ^{[1
]}

Di Grazia, Giuseppe ^{[2
]}

Nardone, Gabriele ^{[3
]}

Orasi, Arianna ^{[3
]}

Picone, Marco ^{[3
]}

Ferla, Maurizio ^{[3
]}

Gresta, Stefano ^{[1
]}

机构：

[1] Univ Catania, Dipartimento Sci Biol Geol & Ambientali, Sez Sci Terra, Corso Italia 57, I-95129 Catania, Italy

[2] Ist Nazl Geofis & Vulcanol, Osservatorio Etneo Sez Catania, Piazza Roma 2, I-95125 Catania, Italy

[3] Ist Super Protez & Ric Ambientale, CN COS, Via Vitaliano Brancati 48, I-00144 Rome, Italy

来源：

REMOTE SENSING | 2020年 / 12卷 / 05期

关键词：

microseism; significant wave height; machine learning; correlation coefficient; HF RADAR; GENERATION;

D O I：

10.3390/rs12050761

中图分类号：

X [环境科学、安全科学];

学科分类号：

08 ; 0830 ;

摘要：

Global warming is making extreme wave events more intense and frequent. Hence, the importance of monitoring the sea state for marine risk assessment and mitigation is increasing day-by-day. In this work, we exploit the ubiquitous seismic noise generated by energy transfer from the ocean to the solid earth (called microseisms) to infer the sea wave height data provided by hindcast maps. To this aim, we use a combined approach based on statistical analysis and machine learning. In particular, a random forest model shows very promising results in the spatial and temporal reconstruction of sea wave height by microseisms. The observed dependence of input importance from the distance sea grid cell-seismic station suggests how the reliable monitoring of the sea state in a wide area by microseisms needs data recorded by dense networks, comprising stations evenly distributed along the coastlines.

引用

页数：15

共 47 条

[31]

Johnson K., 2013, Applied Predictive Modeling

[32] Application of random forest regression to the calculation of gas-phase chemistry within the GEOS-Chem chemistry model v10 [J].