NOAA Protocols for Fisheries Acoustics Surveys and Related ...

elation also assumes that backscattering cross section is proportional to the square of the fish
length (Foote, 1987), which may not necessarily be a viable assumption (McClatchie et al.,
1996). The latter feature of the TS-length model has implications for the accuracy to which the
relation can predict TS, especially beyond the narrow size range of hake used in the Traynor
(1996) equation.
Another consideration regarding bias in the derived TS from fish size distribution is the
assumption of representativeness across all length classes for sampled Pacific hake. Net
selectivity is typically asymptotic, with smaller fish proportionately less represented in the trawl
catches. If the younger fish are indeed a significant proportion of the backscatter, but are not
represented in the catch, appropriate compensation by weighting in the size distributions will be
needed. There is evidence of variable catchability of Pacific hake acoustic survey (Helser et al.,
2004), but this pattern incorporates other features of the survey (e.g., availability, sampling bias)
beyond simple net selectivity.
Considerations
Remediation – In the event that the currently accepted TS-fish length relation for Pacific hake
is deemed incorrect or not as accurate as a successor, an analysis will be undertaken to
determine the effects of the past practices on Pacific hake population estimates.
Improvements – A combination of in situ, ex situ, and modeling experiments are currently
underway and are designed to investigate and compare measured and predicted target strength
measurements from a wide range of sizes of Pacific hake. The results of this work will shed
additional light on the reliability of the currently accepted TS-length relation, including hake
target strength variation as a function of tilt. If needed, the problem of remotely determining
the in situ orientation distribution of fish may be assessed by an inferential method (Foote and
Traynor, 1988). This method, which couples an understanding of swimbladder morphology
and fish TS values measured at multiple frequencies, may provide a general method for
determining the parameters of the tilt angle distribution in situ. Key to advancing this research
is the capability to place transducers of different frequencies closer to the hake, either through
drop transducer systems or autonomous underwater vehicles.
Data Collection
Not Applicable
Detection Probability
Not Applicable
Classification
See Protocol 2
Performance Degradation
See Protocol 2
120