Acoustic biomass estimation of mesopelagic fish: backscattering from individuals, populations, and communities

ICES Journal of Marine Science ICES Journal of Marine Science (2015), 72(5), 1413– 1424. doi:10.1093/icesjms/fsv023 Original Article Acoustic biomass estimation of mesopelagic fish: backscattering from individuals, populations, and communities Peter C. Davison 1*, J. Anthony Koslow2, and Rudy J. Kloser 3 1 *Corresponding author: tel/fax: +1 707 981 8033; e-mail: Davison, P. C., Koslow, J. A., and Kloser, R. J. Acoustic biomass estimation of mesopelagic fish: backscattering from individuals, populations, and communities. – ICES Journal of Marine Science, 72: 1413 – 1424. Received 7 November 2014; revised 26 January 2015; accepted 28 January 2015; advance access publication 19 February 2015. Acoustic survey methods are useful to estimate the distribution, abundance, and biomass of mesopelagic fish, a key component of open ocean ecosystems. However, mesopelagic fish pose several challenges for acoustic biomass estimation based on their small size, wide depth range, mixed aggregations, and length-dependent acoustic reflectance, which differentiate them from the larger epipelagic and neritic fish for which these methods were developed. Foremost, there is a strong effect of depth on swimbladder resonance, so acoustic surveys of mesopelagic fish must incorporate depth-stratification. Additionally, the 1 – 3 cm juveniles of many species are not only more abundant, but can also be stronger acoustic backscatterers than the larger adults that comprise most of the biomass. The dominant species in terms of biomass may thus be weak acoustic backscatters. Failure to properly incorporate depth, the full size distribution, and certain less-abundant species into mesopelagic acoustic analyses could lead to errors in estimated biomass of up to three orders of magnitude. Thus, thorough validation, or “ground-truthing ”, of the species composition, depth structure, population size distribution, capture efficiency of the sampling device, and acoustic properties of the fish present is critical for credible acoustic estimates of mesopelagic fish biomass. This is not insurmountable, but requires more ancillary data than is usually collected. Keywords: acoustic backscatter, biomass assessment, mesopelagic fish, resonance, swimbladder, target strength. Introduction Mesopelagic fish are virtually ubiquitous in the world’s oceans, being found in all oceans except the Arctic (Gjosaeter and Kawaguchi, 1980). In most of the ocean, they appear the dominant zooplankton consumers, so they play a key role in marine foodwebs (Mann, 1984; Pakhomov et al., 1996; Gartner et al., 1997). They reside at mesopelagic depths (200 –1000 m) in daylight, where they help form the acoustic deep scattering layer (DSL). Many ascend to feed in near-surface waters at night, and return to depth before dawn in a diel vertical migration (DVM). They thus link epipelagic and deep-water food chains and by transporting considerable zooplankton production to the deep ocean, may play a key role in biogeochemical cycling and carbon sequestration (Davison et al., 2013; Irigoien et al., 2014). Their biomass is therefore a critical parameter for global and regional models of marine ecosystems and biogeochemistry. However, at present, there remain approximately order of magnitude uncertainties with regard to their biomass both globally and regionally (Irigoien et al., 2014). Gjosaeter and Kawaguchi (GK; 1980) estimated the global biomass of mesopelagic fish was 1 billion t, based on a review of global mesopelagic sampling programmes to date. They relied on micronekton net sampling for their biomass estimates, which generally ranged between 1 and 5 g m22. Since publication of GK’s landmark study, there is increased evidence from studies based on combining acoustics with trawling that escapement and avoidance from pelagic trawls leads to underestimation by trawls of a factor of 7 or more (Koslow et al., 1997; Kloser et al., 2009; Yasuma and Yamamura, 2010; Davison, 2011b). Substantial avoidance of pelagic trawls has also been directly observed (Kaartvedt et al., 2012). Thus, acoustic estimates of biomass generally, but not # International Council for the Exploration of the Sea 2015. All rights reserved. For Permissions, please email: Farallon Institute for Advanced Ecosystem Research, Petaluma, CA 94952, USA Scripps Institution of Oceanography, University of California at San Diego, 9500 Gilman Drive, La Jolla, CA 92093-0218, USA 3 CSIRO Oceans and Atmosphere Flagship, GPO Box 1538, Hobart, TAS 7001, Australia 2 1414 Methods Fish collection A total of 125 oblique trawls to depths .400 m were made over the course of nine cruises (CCE-P0704, CCE-P0810, ORCAWALE 2008, SEAPLEX, and CalCOFI cruises 1008, 1011, 1108, 1110, and 1202) in the CCE and NPSG from 2007 to 2012 using a 5-m2 Matsuda-Oozeki-Hu trawl (MOHT; Oozeki et al., 2004) or 3-m2 Isaacs–Kidd midwater trawl (IKMT; Isaacs and Kidd, 1953) fitted with a constant 1.7 mm mesh (Figure 1). Water flow through the trawl mouth was measured with a TSK flowmeter for the MOHT and calculated from ship speed for the IKMT. Standard length (Ls) was measured to the nearest millimetre and blotted wet weight (Ww) measured to a precision of 0.01 g or estimated from Ww(Ls) regressions (Figure 2). Abundance and biomass were estimated by dividing the number and weight (respectively) of captured fish by the volume of water filtered, and then multiplying by the depth of the trawl. Three model species of mesopelagic fish in the CCE were chosen for study of acoustic backscattering. Diaphus theta, Stenobrachius leucopsarus, and Leuroglossus stilbius are the most abundant vertically migratory species in our collections from the CCE with gas-filled swimbladders, regressed swimbladders as adults, and no swimbladder, respectively (Davison, 2011a). All individuals of these species collected from 122 deep trawls in the CCE (Figure 1) were pooled to estimate the length and weight distributions of their populations. Catches were pooled because the distribution of mesopelagic fish is patchy, catches of individual species from single trawls are often low, and the larger size classes are relatively rare. Both IKMT and MOHT trawls were used to collect fish of the model species, and abundance was corrected for gear differences (MOHT catch is 2.1 times that of the IKMT; Yamamura et al., 2010; Davison et al., 2013) before pooling. For community-level comparisons, the entire normalized catch from a station in the CCE (four trawls from CCE-P0810 cruise, “Cycle 3”) was compared with that from a station in the NPSG (three trawls from SEAPLEX cruise, “Station 2”; Figure 1). Only MOHT trawls were used at the CCE and NPSG stations. We use the word “community” here and hereafter in the sense of the species present at our two stations, and their size distribution. These stations may or may not be representative of broader time or space scales. Figure 1. Trawl locations. Individuals of D. theta, S. leucopsarus, and L. stilbius we (...truncated)