 |
Alternative Observer Sampling within the Kodiak Trawl Fishery |
North Pacific Research Board Project 1017 |
|
Welcome to the NPRB 1017 project web page!
NPRB Project 1017 is a North Pacific Research Board (NPRB) funded cooperative research project that was conducted
during 2011-12 by members of the National Marine Fisheries Service (NMFS), Pacific States Marine Fisheries Commission,
and Alaska Groundfish Data Bank Incorporated. Together with the Kodiak commercial trawl fleet, this project aimed to
evaluate observer sampling methods that might improve estimates of catch used to manage Alaska's groundfish fisheries.
Our study had two components: at-sea sampling of discards and shore-side sampling of landed catch. These two components
correspond to the two portions of catch used for quota management by NMFS for Alaskan groundfish. The at-sea portion of
the project relied on the cooperation of vessel operators to facilitate our study, and the cooperation of the Kodiak processing plants for full access to the landed catch.
| |
| AT SEA SAMPLING | On trawl catcher vessels, catches are difficult to sample. Observers work on the trawl
deck where sampling space and access to catch is limited. Due to these sampling challenges, the haul-specific catch estimates
may be highly variable. Many factors contribute to the variance in estimates of species-specific catches and at-sea discards,
such as the difficulty in 1) sampling (that may result in only a small portion of the catch being observed, 2) estimating the
size (weight) of the total catches, and 3) determining the percentage of each species discarded at-sea.
Alternative methods were designed that should decrease variance in observer catch and estimates of at-sea discards. The alternative
methods sample the portion of the catch that would be discarded at sea while the standard observer methods sample from the entire
catch. By focusing sampling on the at-sea discards and relying on shoreside reports of retained catch, the sampling efficiency of
observers and precision of catch estimates may be improved.
Two observers were deployed in three fisheries (on 12 trips) and simultaneously sampled catches using the two sampling
methodologies. The alternative methodology was successfully implemented in two of the three fisheries; however, logistical
constraints decreased sampling effectiveness in the third. Observers were unable to collect multiple samples under both
methodologies, preventing variance estimation, although this occurred more often for the observer using standard methods.
Although catch estimates from the two methods were not found to be significantly different, catch estimates under the alternative
methodology had smaller variances than standard methodology estimates for 64% of hauls where variances could be computed.
|
| DOCKSIDE MONITORING | While the at-sea sampling portion of this project focused on observer sampling of the
bycatch to be discarded at sea, the second portion of the project was concerned with the landed portion of the catch. Since
species-specific weights are provided by industry in landing reports, it is not standard practice for the Observer Program to
sample for species composition from the delivered catch. The accurate identification of all landed species is critical to accurate
fisheries management. Species identification trainings are held in processing facilities on occasion, however, the impact of such
activities diminishes with time. We examined the effectiveness of using observer sampling techniques to collect data that could be
used to corroborate the recorded catch reported on landing reports. The overall goal of this exercise was to test the feasibility of
using this type of comparison as a tool towards improving the accuracy of species identification on industry reports of retained
catch that are used in quota management.
We sampled 13 deliveries to different processing plants in Kodiak. Delivered fish could be divided into two portions: fish that were
pumped by vacuum directly into the plant, and fish that were sorted on deck and delivered by crane in containers or totes.
These two populations of delivered catch were monitored independently by two observers for each offload. Since both portions of the
delivered fish are being offloaded at the same time, it would not have been possible to track the offload process with a single
observer.
Observers independently verified the weights of each tote for the deck sorted fish. In this study skates were the only group of
fishes that were treated in this way. Based on differences between the skate weights recorded by the observers and the weights of
skates recorded on landing reports, it appears that some confusion between Big and longnose skates occurred during the study.
Plant staff underestimated Big skate and overestimated Longnose skate.
Observer data that resulted from sampling proved difficult to compare against landing reports because an observer may have missed
a species due to a small sample size relative to the entire offload. Still, observers were able to obtain up to 24 samples each
weighing between 300 and 1,000 kgs from a single offload. Our analysis show that for relatively common species, say over 10% of
the delivery, observer samples were able to adequately identify species. While there was some indication that species
identifications by plant staff could be improved for flatfish, there was not enough data to say for certain this was a chronic
problem. For comparison in the rockfish fishery, where Catch Monitoring and Control Plans are in place, we found that industry
reports were much better at identifying rare (less than 1% of the delivery) species than observers. Thus while the use of
observers to improve landing reports has utility, this does not represent the only option available to fishery managers and
policy makers to improve data quality.
|
| Additional Information and Published Results | Additional Information and Published Results http://www.alaskamarinescience.org/documents/AMSS_2013_J.Cahalan.pdf http://www.alaskamarinescience.org/CraigFaunce_000.mp3 |
| Acknowledgements |
This work would not have been possible without the dedicated professional fisheries observers who served as the project's at-sea scientists or without the voluntary participation of the owners, operators, and crew of the fishing vessels that allowed us onboard and the processors who allowed us into their plants.
Specifically we would like to thank observers: Robert Davis, Laura Galoti, Kenny Hardin, Carla Hilts, Mike Levine, Lauren Loe, Tom Mauer, Dave Pearmain, and Rory Ricks; the vessels F/V Caravelle, F/V Laura, F/V Marathon, F/V Mar Del Norte, F/V Michelle Renee, F/V Pacific Star, F/V Sea Mac, and F/V Topaz; and the Kodiak processing plants APS, ISA, Ocean Beauty, Trident Seafoods.
Katy McGauley (Alaska Groundfish Data Bank) provided invaluable at-sea support and technical expertise to the project. The programmatic and logistical support on which this project depended was provided by the training, data processing, and contract management staff of the Fisheries Monitoring and Analysis Division of the Alaska Fisheries Science Center, in particular Brian Mason, Mike Vector, Bob Maier and Heather Weikart. Saltwater, Inc provided observers while under contract to NOAA/NMFS.
Funding for this project was provided by the North Pacific Research Board, NOAA/NMFS Alaska Observer Program, and the Pacific States Marine Fisheries Commission.
|
| CONTACT INFORMATION |
To find out more or to participate contact any of the following persons:
Craig H. Faunce
Jennifer Cahalan
Julie Bonney
|
| FURTHER READING |
For more information on the history of the NPGOP, see
For more information on how the NPGOP samples, see:
For more information on the way catch is estimated by NMFS in Alaska, see:
|
    |