Detecting and characterizing upwelling filaments in a numerical ocean model

Osvaldo Artal; Héctor H. Sepúlveda; Domingo Mery; Christian Pieringer

doi:10.1016/j.cageo.2018.10.005

Detecting and characterizing upwelling filaments in a numerical ocean model

Osvaldo Artal, Héctor H. Sepúlveda, Domingo Mery, Christian Pieringer

Research output: Contribution to journal › Article › peer-review

11 Scopus citations

Abstract

Upwelling filaments are long (≈ 100's km) narrow (O ≈ 10 km) structures in the coastal ocean. They export nutrients and prevent the movement of larvae along the coast. Filaments can be observed in satellite images and in numerical models, but their manual identification and characterization is complex and time consuming. Here we present a Matlab code for a manual method to assist experts in this task, and a code for an automatic filament detection method (AFD) based on image processing and pattern recognition to identify and extract features in output files from a numerical ocean model. AFD was tested with a simulation of northern Chile. AFD had a similar performance in filament detection to that of human experts. AFD provides substantial time savings when analyzing a large number of images from a numerical ocean model. AFD is open source and freely available.

Original language	English
Pages (from-to)	25-34
Number of pages	10
Journal	Computers and Geosciences
Volume	122
DOIs	https://doi.org/10.1016/j.cageo.2018.10.005
State	Published - Jan 2019
Externally published	Yes

Keywords

Chile
Coastal ocean
Image processing
Numerical models
Upwelling filaments

Access to Document

10.1016/j.cageo.2018.10.005

Cite this

@article{bef99fbcb55040959eee5866cb63050f,

title = "Detecting and characterizing upwelling filaments in a numerical ocean model",

abstract = "Upwelling filaments are long (≈ 100's km) narrow (O ≈ 10 km) structures in the coastal ocean. They export nutrients and prevent the movement of larvae along the coast. Filaments can be observed in satellite images and in numerical models, but their manual identification and characterization is complex and time consuming. Here we present a Matlab code for a manual method to assist experts in this task, and a code for an automatic filament detection method (AFD) based on image processing and pattern recognition to identify and extract features in output files from a numerical ocean model. AFD was tested with a simulation of northern Chile. AFD had a similar performance in filament detection to that of human experts. AFD provides substantial time savings when analyzing a large number of images from a numerical ocean model. AFD is open source and freely available.",

keywords = "Chile, Coastal ocean, Image processing, Numerical models, Upwelling filaments",

author = "Osvaldo Artal and Sep{\'u}lveda, {H{\'e}ctor H.} and Domingo Mery and Christian Pieringer",

note = "Publisher Copyright: {\textcopyright} 2018 Elsevier Ltd",

year = "2019",

month = jan,

doi = "10.1016/j.cageo.2018.10.005",

language = "English",

volume = "122",

pages = "25--34",

journal = "Computers and Geosciences",

issn = "0098-3004",

}

TY - JOUR

T1 - Detecting and characterizing upwelling filaments in a numerical ocean model

AU - Artal, Osvaldo

AU - Sepúlveda, Héctor H.

AU - Mery, Domingo

AU - Pieringer, Christian

PY - 2019/1

Y1 - 2019/1

N2 - Upwelling filaments are long (≈ 100's km) narrow (O ≈ 10 km) structures in the coastal ocean. They export nutrients and prevent the movement of larvae along the coast. Filaments can be observed in satellite images and in numerical models, but their manual identification and characterization is complex and time consuming. Here we present a Matlab code for a manual method to assist experts in this task, and a code for an automatic filament detection method (AFD) based on image processing and pattern recognition to identify and extract features in output files from a numerical ocean model. AFD was tested with a simulation of northern Chile. AFD had a similar performance in filament detection to that of human experts. AFD provides substantial time savings when analyzing a large number of images from a numerical ocean model. AFD is open source and freely available.

AB - Upwelling filaments are long (≈ 100's km) narrow (O ≈ 10 km) structures in the coastal ocean. They export nutrients and prevent the movement of larvae along the coast. Filaments can be observed in satellite images and in numerical models, but their manual identification and characterization is complex and time consuming. Here we present a Matlab code for a manual method to assist experts in this task, and a code for an automatic filament detection method (AFD) based on image processing and pattern recognition to identify and extract features in output files from a numerical ocean model. AFD was tested with a simulation of northern Chile. AFD had a similar performance in filament detection to that of human experts. AFD provides substantial time savings when analyzing a large number of images from a numerical ocean model. AFD is open source and freely available.

KW - Chile

KW - Coastal ocean

KW - Image processing

KW - Numerical models

KW - Upwelling filaments

UR - http://www.scopus.com/inward/record.url?scp=85054868217&partnerID=8YFLogxK

U2 - 10.1016/j.cageo.2018.10.005

DO - 10.1016/j.cageo.2018.10.005

M3 - Article

AN - SCOPUS:85054868217

SN - 0098-3004

VL - 122

SP - 25

EP - 34

JO - Computers and Geosciences

JF - Computers and Geosciences

ER -

Detecting and characterizing upwelling filaments in a numerical ocean model

Abstract

Keywords

Access to Document

Other files and links

Fingerprint

Cite this