Progress Reported in Identifying Deep-7 Hawaii Bottomfish from Acoustic Data

Presently, estimates of biomass used in stock assessment of "Deep-7" Hawaii bottomfish (6 snapper species and an endemic grouper) are derived only by fisheries-dependent methods. To improve the quality of biomass estimates, three fisheries-independent approaches — acoustic, optical, and experimental fishing — are being studied and compared to arrive at a single approach or a combination of methods that provides the most accurate abundance estimates. As part of this research, Staff of the PIFSC Ecosystems and Oceanography Division and Fisheries Research and Monitoring Division are developing acoustic descriptors to distinguish Deep-7 from other demersal and semi-demersal species and, if possible, to identify Deep-7 bottomfish to the species level.

The application of acoustics methods to assess these fish is challenging due to the high rugosity of bottomfish habitats and the frequent intermixing and association of Deep-7 bottomfish with each other and other species. The task involves improving our capability to separate echoes generated by the target species from those of other species, and, using this capability, estimating biomass and abundance of Deep-7 bottomfish. To fulfill the above objectives, acoustic data were collected on board the NOAA Ship Oscar Elton Sette in waters of the "Maui Complex" in the main Hawaiian Islands. The acoustic data were analyzed and compared to species and size information obtained simultaneously from experimental bottomfish fishing operations and stereo-video cameras attached to various sampling platforms (e.g., Botcam, Toad, CTD frame, AUV, and ROV). Of the camera platforms used, the AUV seemed to provide the most useful data. The slowly-moving AUV could be kept within the transducers' beams far more frequently than Botcam or ROV, giving it a higher potential for effectiveness in future data collection. However, so far, most of the data points for validating the acoustic method have come from synchronous fishing operations, which have high uncertainty due to difficulties in accurately estimating locations of fish catch from time of catch, depth of the hook, and drift of the vessel. More observations from simultaneous acoustics and optical sampling, ideally with reliable cameras mounted on an AUV, are needed to improve the accuracy of acoustic descriptors developed thus far for Deep-7 species. In addition to optical data, the experience of fishermen using acoustic "fish finders" was a valuable source of information for distinguishing fish on echograms.

From all data available, the separation of Deep-7 species has been partially achieved. The accompanying graphic shows examples of species as they appear on echograms at the shallow (echogram a), mid-depth (b), and deep (c) ranges of Deep-7 habitat. In the shallower habitat (~100 m), smaller opakapaka and kalekale form thick aggregations on the ocean floor and they can be intermixed with each other and with taape and/or wekeula. Lehi and larger opakapaka swim close to the bottom as non-aggregated individuals. Opelu typically form vertically elongated or "V" shaped aggregations and are located farther up in the water column (echograms a and d). However, smaller opakapaka and kalekale can also migrate up in the water column and are hard to separate from opelu and taape. There are some indications that dB differencing (comparing acoustic signals between different frequencies) could be used to separate taape and opelu from Deep-7 species, but more in situ data are needed to verify this hypothesis.

Echograms depicting examples of aggregated and loosely-associated fish identified during simultaneous observation using acoustics 
        and alternative methods, such as video cameras or hook-and line fishing.  Examples of backscatter from Deep-7 bottomfish and other 
        fish species intermixed with them are shown at shallow (a), mid-depth (b), and deep (c) ranges of Deep-7 
        bottomfish habitat.  The middle-top echogram (d) shows an example of a near-surface school of opelu.
Echograms depicting examples of aggregated and loosely-associated fish identified during simultaneous observation using acoustics and alternative methods, such as video cameras or hook-and line fishing. Examples of backscatter from Deep-7 bottomfish and other fish species intermixed with them are shown at shallow (a), mid-depth (b), and deep (c) ranges of Deep-7 bottomfish habitat. The middle-top echogram (d) shows an example of a near-surface school of opelu.

In mid-depth range (echogram b), smaller kalekale and opakapaka form thick aggregations on the bottom with shapes that taper off with decreasing depth. Larger opakapaka and lehi swim individually. At the shallower and mid-ranges, these Deep-7 species mix with kahala, another species that cannot be distinguished from Deep-7 at this time. In the deeper ranges of Deep-7 habitat (below 200 m, echogram c), onaga and ehu swim close to the bottom and do not form aggregations. They do not seem to intermix with other species that cannot be separated on the echogram. In addition to dB differencing methods, non-Deep-7 species might be separated using their size range if more data were available. Species that could not be separated from Deep-7 at this point are either at the lowest (e.g., taape) or highest (e.g., kahala) end of the Deep-7 size range.