Mesopelagic fish belong to the mid-trophic level functional group in
the marine ecosystem, and they are a key component of oceanic
food web, linking primary and tertiary consumers. Despite their
ecological relevance, little is known about the distribution patterns
of mesopelagic fishes (Young et al. 2001), and there is a strong
demand for more research for ocean-biogeochemical-population
(Lehodey 2004) and ecosystem (Kloser et al. 2009, Handegard et al.
2012) models. Fisheries acoustics provides a suitable tool for
monitoring mesopelagic fish abundance and distribution (Simmonds
& MacLennan 2005). Using acoustics from vessels of opportunity
may provide the best chance of achieving the ocean basin scale
spatial coverage required to inform ecosystem models.
Introduction
1. Identify large-scale patterns of pelagic fish abundance using area
backscattering coefficient (sa in m2 m-2) at 38 kHz as a measure
of biological abundance across the South Pacific Ocean.
2. Determine the vertical distribution of sa.
3. Explore the relationship between distribution of sa and
environmental and geographic variables to determine the major
factors driving pelagic fish distribution.
Objectives
Two return trips between New Zealand and Chile in Oct-Dec 2007
and 2010 from RV Kaharoa deploying ARGO buoys.
Data collected using a Simrad ES60 echosounder with hull-
mounted 38 kHz transducer (2000 W power, 1.024 ms pulse).
Analysis software Sonardata Echoview 4.90 and 5.0 (Myriax).
Acoustic records integrated in 10 n miles elementary distance
sampling units (EDSU) along the vessel track, and in 50 m depth
bins. Maximum data range up to 600 m depth.
Boosted regression trees (BRT) were fitted in R (version 2.13.2),
using GMB package version 1.6–3.1 plus custom code available
online (Elith et al. 2008).
Methods
Results
Conclusions The key finding was that there was a strong correlation between chl a and distribution of acoustic backscatter
as shown by BRT and spatial overlap (Fig. 2). The process underlying this correlation is likely food availability.
Major vertical patterns of pelagic fish distribution were diel migration and layering.
Major horizontal patterns were coastal intensification and oceanic patchiness, with abundance decreasing
from the coasts towards the core of the south tropical gyre.
Different mesopelagic species may have different acoustic properties, so correlation between acoustic backscatter
and fish abundance may not be linear. Nevertheless, acoustic information based on total backscatter still provided
further valuable knowledge about patterns of distribution and productivity.
There is a growing interest for developing programs focused on global ecosystem models for studying large-scale
patterns of distribution and abundance of biota and their changes over long time scales. This study provides data to
validate model predictions of mid-trophic functional groups.
Reference Escobar-Flores, P., O’Driscoll, R., & Montgomery, J. (2013).
Acoustic characterization of pelagic fish distribution across the
South Pacific Ocean. Marine Ecology Progress Series, 490, 169–
183. doi:10.3354/meps10435
* Email: [email protected]
Acknowledgments and contact Funding for this study was provided by the Advanced Human Capital Programme, National Commission of Scientific and
Technological Research CONICYT CHILE, through a Bicentennial Becas-Chile Master scholarship.
1School of Biological Sciences, University of Auckland, Private Bag 92019, Auckland 1142, New Zealand
2National Institute of Water and Atmospheric Research, Private Bag 14-901, Kilbirnie, Wellington 6241, New Zealand
Pablo Escobar-Flores1*, Richard L. O’Driscoll2, John C. Montgomery1
Acoustic characterization of pelagic fish
distribution across the South Pacific Ocean
-chlorophyll (chl a in mg m-3)
-sea surface temperature
(SST in degrees Celsius)
-sea surface height (SSH in m)
-minimum distance
from main land
masses (km)
-depth (m)
GEOGRAPHIC
PARAMETERS
- time since
sunrise (hours)
TEMPORAL
VARIABLE
1 EDSU = 10 n miles
10 n miles 20 n miles
ENVIRONMENTAL
VARIABLES
Fig. 2. Vertically summed acoustic backscatter coefficient (sa in m2 km−2) per 10 n miles elementary distance sampling
unit along each transect overlying maps of chl a (mg m−3) for the area and period of study: (a) T1-2007, (b) T2-2007,
(c) T1-2010, and (d) T2-2010. Maximum bar size = 140 m2 km−2. Grey bars: daytime; black bars: nighttime.
Table 1: Table 1. Comparison of total acoustic area backscattering coefficient (sa) along the 4 transects. EDSU:
elementary distance sampling unit size (10 n miles).
Fig. 3. Proportion of total acoustic backscatter coefficient (sa in m2 km−2) summarised in 50 m depth bins for each
individual transect: (a) T1-2007, (b) T2-2007, (c) T1-2010, and (d) T2-2010.
Fig. 4. Longitudinal section along transect T1-2010. (a) Vertically summed acoustic backscatter coefficient (sa in m2 km−2) in 10 n
miles elementary distance sampling units along the transect. (b) Daytime and nighttime. (c) Mean volume backscattering strength
(Sv in decibels) in each cell (red-brown= strong, blue-grey= weak). Each cell in the colour central panel is 10 n miles long by 50 m
deep. (d) Bottom depth.
Table 2. Summary of boosted regression trees model parameters (lr : learning rate; tc: tree complexity; nt: number of trees) and
performance. Model parameters were optimised by k-fold cross-validation (CV); SE: standard error.
Table 3. Summary of relative contributions (%) of the predictor variables for boosted regression trees Models 2 and 3, with cross-
validation data on 2462 locations. See Table 2 for model descriptions and parameters.
References
Elith J, Leathwick JR, Hastie T (2008) A working guide to boosted regression trees. J Anim Ecol 77: 802−813
Handegard NO, du Buisson L, Brehmer P, Chalmers SJ and others (2012) Towards an acoustic-based coupled observation and
modelling system for monitoring and predicting ecosystem dynamics of the open ocean. Fish Fish doi:10.1111/j.1467-
2979.2012.00480.x
Kloser RJ, Ryan TE, Young JW, Lewis ME (2009) Acoustic observations of micronekton fish on the scale of an ocean basin:
potential and challenges. ICES J Mar Sci 66: 998−1006
Lehodey P (2004) A spatial ecosystem and populations dynamics model (SEAPODYM) for tuna and associated oceanic top-predator
species. II. Tuna populations and fisheries. In: 17th Meeting of the Standing Committee on Tuna and Billfish, Majuro, Marshall
Islands, 5−18 August, 2004. SCTB17, Working Paper ECO-2, July 2004. Oceanic Fisheries Programme, Secretariat of the Pacific
Community, Noumea
Simmonds EJ, MacLennan DN (2005) Fisheries acoustics: theory and practice, 2nd edn. Blackwell Science, Oxford
Young JW, Lamb TD, Bradford R, Clementson L, Kloser R, Galea H (2001) Yellowfin tuna (Thunnus albacares) aggregations along
the shelf break of southeastern Australia: links between inshore and offshore processes. Mar Freshw Res 52: 463−474
Fig. 1. Different species of mesopelagic fish caught with mid-water trawls over
the Chatam Rise, western South Pacific Ocean.