Analysis of backscatter properties and application of classification procedures for the identification of small pelagic fish species in the Central Mediterranean

M. D'Elia*, B. Patti, A. Bonanno, I. Fontana, G. Giacalone, G. Basilone, P. G. Fernandes

*Corresponding author for this work

Research output: Contribution to journalArticle

12 Citations (Scopus)

Abstract

The pelagic realm of the Central Mediterranean Sea is populated by four main species of fish: sardine, anchovy, horse mackerel and a mix of other pelagic fish species. In this study we employed a multifrequency acoustics approach to detect and classify fish schools of these groups. Monospecific trawl catches were selected from eight acoustic surveys and examined in relation to the coincident acoustic data. The backscattering properties of the three main species were determined using the decibel difference (Sv(120) - Sv(38)) and the frequency response (NASC(120)/NASC(38)). The results indicate that schools of these species cannot be distinguished on the basis of energetic properties alone, because they are very similar in physiology and scattering is dominated by the swimbladder, which is similar in shape and size. However, the use of classification models (classification tree, random forest), using energetic features, as well as bathymetric and morphometric parameters, allowed for some discrimination among the groups. According to the classification tree, school depth was found to play an important role in the identification of these fish groups, especially for anchovy and horse mackerel, for which the contribution to the overall performance of the tree was about 20%. The tree models, with only energetic or morphometric parameters, were able to classify sardine schools reasonably well, but not so well for anchovy and horse mackerel. Using a random forest method, which accounted for the variability in the learning sample, an accuracy of 85% in the overall classification rate was reached with a greater power of discrimination for sardine and anchovy schools. Published by Elsevier B.V.

Original languageEnglish
Pages (from-to)33-42
Number of pages10
JournalFisheries Research
Volume149
Early online date14 Oct 2013
DOIs
Publication statusPublished - Jan 2014

Keywords

  • multifrequency
  • pelagic schools
  • species identification
  • backscattering property
  • classification tree methods
  • anchovy engraulis-encrasicolus
  • African Continental-Shelf
  • herring Clupea-Harengus
  • feeding-behavior
  • horse mackerel
  • random forests
  • acoustic data
  • fisheries acoustics
  • zooplankton groups
  • sound-scattering

Cite this

Analysis of backscatter properties and application of classification procedures for the identification of small pelagic fish species in the Central Mediterranean. / D'Elia, M.; Patti, B.; Bonanno, A.; Fontana, I.; Giacalone, G.; Basilone, G.; Fernandes, P. G.

In: Fisheries Research, Vol. 149, 01.2014, p. 33-42.

Research output: Contribution to journalArticle

@article{8193827cef334683b50352df00483c83,
title = "Analysis of backscatter properties and application of classification procedures for the identification of small pelagic fish species in the Central Mediterranean",
abstract = "The pelagic realm of the Central Mediterranean Sea is populated by four main species of fish: sardine, anchovy, horse mackerel and a mix of other pelagic fish species. In this study we employed a multifrequency acoustics approach to detect and classify fish schools of these groups. Monospecific trawl catches were selected from eight acoustic surveys and examined in relation to the coincident acoustic data. The backscattering properties of the three main species were determined using the decibel difference (Sv(120) - Sv(38)) and the frequency response (NASC(120)/NASC(38)). The results indicate that schools of these species cannot be distinguished on the basis of energetic properties alone, because they are very similar in physiology and scattering is dominated by the swimbladder, which is similar in shape and size. However, the use of classification models (classification tree, random forest), using energetic features, as well as bathymetric and morphometric parameters, allowed for some discrimination among the groups. According to the classification tree, school depth was found to play an important role in the identification of these fish groups, especially for anchovy and horse mackerel, for which the contribution to the overall performance of the tree was about 20{\%}. The tree models, with only energetic or morphometric parameters, were able to classify sardine schools reasonably well, but not so well for anchovy and horse mackerel. Using a random forest method, which accounted for the variability in the learning sample, an accuracy of 85{\%} in the overall classification rate was reached with a greater power of discrimination for sardine and anchovy schools. Published by Elsevier B.V.",
keywords = "multifrequency, pelagic schools, species identification, backscattering property, classification tree methods, anchovy engraulis-encrasicolus, African Continental-Shelf, herring Clupea-Harengus, feeding-behavior, horse mackerel, random forests, acoustic data, fisheries acoustics, zooplankton groups, sound-scattering",
author = "M. D'Elia and B. Patti and A. Bonanno and I. Fontana and G. Giacalone and G. Basilone and Fernandes, {P. G.}",
note = "Received 6 March 2013, Revised 8 August 2013, Accepted 11 August 2013, Available online 14 October 2013 Acknowledgements We are grateful to all the Captains and the crew of R/V “G. Dallaporta” for their help in collecting acoustic and trawl data from 2002 to 2011, and to the colleagues of the Marine Laboratory, in Aberdeen, for their help and assistance at various stages of the work.",
year = "2014",
month = "1",
doi = "10.1016/j.fishres.2013.08.006",
language = "English",
volume = "149",
pages = "33--42",
journal = "Fisheries Research",
issn = "0165-7836",
publisher = "Elsevier Science B. V.",

}

TY - JOUR

T1 - Analysis of backscatter properties and application of classification procedures for the identification of small pelagic fish species in the Central Mediterranean

AU - D'Elia, M.

AU - Patti, B.

AU - Bonanno, A.

AU - Fontana, I.

AU - Giacalone, G.

AU - Basilone, G.

AU - Fernandes, P. G.

N1 - Received 6 March 2013, Revised 8 August 2013, Accepted 11 August 2013, Available online 14 October 2013 Acknowledgements We are grateful to all the Captains and the crew of R/V “G. Dallaporta” for their help in collecting acoustic and trawl data from 2002 to 2011, and to the colleagues of the Marine Laboratory, in Aberdeen, for their help and assistance at various stages of the work.

PY - 2014/1

Y1 - 2014/1

N2 - The pelagic realm of the Central Mediterranean Sea is populated by four main species of fish: sardine, anchovy, horse mackerel and a mix of other pelagic fish species. In this study we employed a multifrequency acoustics approach to detect and classify fish schools of these groups. Monospecific trawl catches were selected from eight acoustic surveys and examined in relation to the coincident acoustic data. The backscattering properties of the three main species were determined using the decibel difference (Sv(120) - Sv(38)) and the frequency response (NASC(120)/NASC(38)). The results indicate that schools of these species cannot be distinguished on the basis of energetic properties alone, because they are very similar in physiology and scattering is dominated by the swimbladder, which is similar in shape and size. However, the use of classification models (classification tree, random forest), using energetic features, as well as bathymetric and morphometric parameters, allowed for some discrimination among the groups. According to the classification tree, school depth was found to play an important role in the identification of these fish groups, especially for anchovy and horse mackerel, for which the contribution to the overall performance of the tree was about 20%. The tree models, with only energetic or morphometric parameters, were able to classify sardine schools reasonably well, but not so well for anchovy and horse mackerel. Using a random forest method, which accounted for the variability in the learning sample, an accuracy of 85% in the overall classification rate was reached with a greater power of discrimination for sardine and anchovy schools. Published by Elsevier B.V.

AB - The pelagic realm of the Central Mediterranean Sea is populated by four main species of fish: sardine, anchovy, horse mackerel and a mix of other pelagic fish species. In this study we employed a multifrequency acoustics approach to detect and classify fish schools of these groups. Monospecific trawl catches were selected from eight acoustic surveys and examined in relation to the coincident acoustic data. The backscattering properties of the three main species were determined using the decibel difference (Sv(120) - Sv(38)) and the frequency response (NASC(120)/NASC(38)). The results indicate that schools of these species cannot be distinguished on the basis of energetic properties alone, because they are very similar in physiology and scattering is dominated by the swimbladder, which is similar in shape and size. However, the use of classification models (classification tree, random forest), using energetic features, as well as bathymetric and morphometric parameters, allowed for some discrimination among the groups. According to the classification tree, school depth was found to play an important role in the identification of these fish groups, especially for anchovy and horse mackerel, for which the contribution to the overall performance of the tree was about 20%. The tree models, with only energetic or morphometric parameters, were able to classify sardine schools reasonably well, but not so well for anchovy and horse mackerel. Using a random forest method, which accounted for the variability in the learning sample, an accuracy of 85% in the overall classification rate was reached with a greater power of discrimination for sardine and anchovy schools. Published by Elsevier B.V.

KW - multifrequency

KW - pelagic schools

KW - species identification

KW - backscattering property

KW - classification tree methods

KW - anchovy engraulis-encrasicolus

KW - African Continental-Shelf

KW - herring Clupea-Harengus

KW - feeding-behavior

KW - horse mackerel

KW - random forests

KW - acoustic data

KW - fisheries acoustics

KW - zooplankton groups

KW - sound-scattering

U2 - 10.1016/j.fishres.2013.08.006

DO - 10.1016/j.fishres.2013.08.006

M3 - Article

VL - 149

SP - 33

EP - 42

JO - Fisheries Research

JF - Fisheries Research

SN - 0165-7836

ER -