Link to NMFS Homepage Link to NOAA Homepage Keyboard Navigation Alaska Fisheries Science Center Home banner

The Fishery Interaction Team:  Investigating the Potential
Impacts of Commercial Fishing On the Foraging Success
of Endangered Steller Sea Lions  (cont.)

Estimating Movement and Abundance of Atka Mackerel With Tag and Release Data

map of Sequam Pass 2000 Atka mackerel tagging study site
Figure 6.  Seguam Pass - the site of the 2000 Atka mackerel tagging study. Shown are areas inside (Area 1) and outside (Area 2) the 20 nm Trawl Exclusion Zone as well as the transects along which tagged Atka mackerel were released

Atka mackerel have become the focus of increased attention as one of the major prey items of the endangered Steller sea lion. They represent one of the largest groundfish stocks in the Aleutian archipelago, with current estimates of exploitable biomass in excess of 500,000 metric tons (t). Atka mackerel occur in dense aggregations in the Aleutian Islands passes and prefer habitat with strong currents. Atka mackerel have been observed at great densities and are extremely patchily distributed in space and time.

In order to prevent possible prey shortage for the Steller sea lions, the National Marine Fisheries Service (NMFS) established 20-nautical mile (nmi) no-trawl zones around six Steller sea lion rookeries in the eastern Aleutian Islands in 1991. In 1998, further measures were taken to spread out fishing effort temporarily as well as spatially. Instead of one fishing season per year, an A (winter) and B(summer) season were mandated, and more quota was assigned to be caught outside of critical habitat. These restrictions affected the fishery by shifting effort away from traditional fishing grounds, and the question arose as to the efficacy of the no-trawl zones and the impact of the fishery on Atka mackerel local abundance and movement. Estimates of Atka mackerel abundance on small time- and space-scales were needed to identify potential changes in behavior or movement of Atka mackerel that could be attributed to the fishery. Mark recapture estimation methods were used to estimate local abundance and small-scale movement of Atka mackerel around Steller sea lion rookeries and to examine potential fishery effects on Atka mackerel movement and abundance.

During August 1999, NMFS scientists in cooperation with the School of Fisheries and Aquatic Sciences, University of Washington, conducted a tagging feasibility study as part of a trawl survey in Seguam Pass in the Aleutian Islands. During the study a mortality experiment was carried out on board with four live tanks, which held fish for 12 days. Two other live tanks were used to hold fish for opportunistic release; 936 tagged fish were released in the area open to the fishery. The results of the mortality experiment and recoveries by the fishery showed that the tagged fish survived well and that the fishery was able to identify and report tagged fish.

In July-August 2000 a full-scale tagging study was conducted in Seguam Pass. Fish were caught and released in two dedicated areas: inside (Area 1) and outside (Area 2) the Trawl Exclusion Zone (Figure 6 above). Atka mackerel were caught with trawl gear, transferred into live tanks, tagged, and released in a continuous transect pattern (Figure 6 above). The tag loss rate was estimated by double tagging approximately 20% of all fish released. Total number of fish released was 6,096 inside the Trawl Exclusion Zone and 2,677 fish outside the Trawl Exclusion Zone.

Tagged fish were recovered by the fishing fleet during its regular fishing procedures with the help of fishery observers in the area open to the fishery (Area 2). In the area closed to the fishery (Area 1), the fishing vessel Seafisher, a 220-foot factory trawler, was chartered by NMFS to recover tagged Atka mackerel. Approximately 34,891 fish were examined for tags in Area 1 by the charter vessel, and 3,960,000 fish were examined for tags in Area 2 by the commercial fleet and the charter vessel combined. The tag reporting rate was estimated by randomly seeding ten dummy tags per haul into the catch that was examined for tags. This was done for every haul on the charter vessel and for all hauls sampled by the observers on the commercial vessels.

Table 1.  Tag recoveries by the two fishery events
and by the NMFS charter vessel.
  Recovery of tags by the commercial fishery
  Recovered Area 1 Recovered Area 2
Tagged Area 1 0   0  
Tagged Area 2 0   77  
Recovery of tags by the NMFS charter vessel
  Recovered Area 1 Recovered Area 2
Tagged Area 1 11   1  
Tagged Area 2

Results of the tag recoveries are summarized in Table 1 (right). The commercial fleet recovered 77 tags in Area 2, all of which had also been released in Area 2; therefore, no movement was reported. The charter vessel recovered 14 tags in Area 1, out of which 3 tags had moved from Area 2; the charter vessel recovered 13 tags in Area 2, out of which 1 tag had moved from Area 1. The results from all recoveries show that only 1 tag out of 90 tags recovered in Area 2 had moved and 3 tags out of 14 tags in Area 1 had moved.

The tagging model used was an integrated model using maximum likelihood to estimate all parameters simultaneously. There were four components of the model including expected tag recoveries, tag loss, tag survival, and tag reporting. Parameter estimates with 95% confidence intervals are summarized in Table 2. Numbers of fish per area were converted into metric tons by calculating the average weight of Atka mackerel at the time of recovery. Biomass estimates from the tagging model ranged from 117,900 t in Area 1 to 82,057 t in Area 2 (Table 2 below).

table 2

Table 3.  Movement rates per area over time.
  Time event Days since
tagging (dk)
Value Lower C.I. Upper C.I.
Movement rate p12 Tag release 1   0.0001 0.0000 0.0002
(from inside area Fishery 37   0.0019 0.0000 0.0068
to outside area) Charter 59   0.0031 0.0000 0.0109
  Fishery (CDQ) 107   0.0056 0.0000 0.0197

Movement rate p21

Tag release



(from outside area Fishery 37   0.4376 0.0000 1.0000
to inside area) Charter 59   0.6006 0.0000 1.00002

Movement rate was calculated as instantaneous or daily movement rate and was then extrapolated to a corresponding movement rate for days in the water since release (Table 3 above). Estimated movement rate from inside (Area 1) to outside (Area 2) the Trawl Exclusion Zone after 59 days (the time of the recovery charter) was less than 1% of the population (Figure 7a below). Estimated movement rate was much greater for fish moving from the open area to the closed area at 60% of the population (Figure 7b below). However, the recovery effort inside the closed area was much smaller, so there is a high degree of uncertainty around the estimate of movement rate into the closed area the 95% confidence bounds included zero and 100% of movement.

graph of movement probability from Area 1 to 2
Figure 7a.  Probability of movement from Area 1 to 2. Instantaneous movement rate estimated for both sexes together and extrapolated to movement rate for days out in the water since release.  (Error bars are the 95% C.I.)

graph of movement probability from Area 2 to 1
Figure 7b.  Probability of movement from Area 2 to 1. Instantaneous movement rate estimated for both sexes together and extrapolated to movement rate for days out in the water since release.  (Error bars are the 95% C.I.)

In addition to examining local abundance and movement rates, sex ratio at time of tagging and at time of tag recovery was examined and compared. Sex ratio for all fish examined for tags at the various recovery events was calculated using sexed length frequency data that were taken for each haul during the recovery charter and for all hauls sampled by observers during the commercial fishery tag recovery. It appeared that sex ratio in the population was not the same in both areas and changed from time of release in July to time of recovery in September and November. In both areas there was a much greater percentage of males present at time of recovery than at time of tagging. This suggests that male and female populations might have different movement rates; therefore, the model was run separately for both sexes.

table 4


  graph of sex ratio of the female and male biomass
         Figure 8.   Sex ratio of the female and male biomass
         in metric tons per area.

Total population sizes were approximately the same as the estimates for both sexes combined (Table 4 above). However, the model estimated the biomass in Area 1 to consist mostly of females; with 100,197 t females and only 18,456 t males. In Area 2 the sexes were more evenly distributed with 47,302 t of females and 29,793 t of males (Figure 8 right) Movement rates also differed dramatically, as it appeared that females were not showing much movement at all while the males were responsible for the large point estimate of the movement rate from Area 2 to Area 1 (Figures 9b, 9b below). This movement rate, however, was associated with large confidence bounds, and not much inference can be drawn from this result. It makes sense, however, that the model associated a high rate of movement for males from Area 2 to Area 1: the sex ratios in both areas were quite different between tagging and recovery periods, and it appeared that an influx of males might cause the shift in sex ratios. It is noteworthy that this model does not allow for immigration or emigration of fish.
graphs of movement rates from Area 1 to 2
Figures 9a, 9b.  Instantaneous movement rate estimated for each sex separately and extrapolated to movement rate for days out in the water since release. Open diamonds represent data for females, solid squares represent data for males (error bars are the 95% C.I.).  Top figure shows the probability of movement from Area 1 to 2 and the bottom figure shows the probability of movement from Area 2 to 1.

graphs of movement rates from Area 2 to 1

Results from the Atka mackerel tag and release study suggest that there is relatively little movement of Atka mackerel from inside to outside the Trawl Exclusion Zone, indicating that the Trawl Exclusion Zones are effective at protecting Atka mackerel near Steller sea lion rookeries around Seguam Pass. Caution should be used in applying these results to other areas, with resident Atka mackerel populations and fisheries of different size and distribution. To examine geographical variation in movement and local abundance, FIT scientists will begin a parallel tag release-recovery study in the Tanaga Pass area (west of Seguam Pass) in 2002. The data described above provide information on local abundance and movement rate of a fished population of Atka mackerel. To increase the precision of the movement rate estimate from outside to inside the Trawl Exclusion Zones, we will increase the number of tagged Atka mackerel from approximately 8,800 to 25,000 tags released in Seguam Pass and will release 12,000 tags in Tanaga Pass. To address the question of whether and how fisheries impact Atka mackerel, FIT scientists will conduct tag recovery surveys both before and after the September 2002 Atka mackerel fishery in Seguam Pass. The specific goals of this experiment are to determine whether localized depletion occurs in a fished area (outside the Trawl Exclusion Zone) and whether movement rates into and out of the Trawl Exclusion Zone are influenced by fishing.

Future Research

In addition to those mentioned above, FIT researchers are planning additional studies. For example, FIT plans to implement an analysis of the potential effects of winter Bering Sea trawl fisheries for Pacific cod on the local distribution and abundance of Pacific cod in December 2002. This experiment will be conducted using pot gear. The design of this experiment calls comparison of the ratios of CPUE (catch-per-unit-effort) before and after fishing in fished and un-fished sites. FIT researchers in collaboration with RACE scientists hope to expand their studies on pollock to additional regions and other seasons. The experiments described here and those planned for the near future focus on identifying commercial fishing effects on the distribution and abundance of Steller sea lion prey. If an effect is detected, research will focus on detecting what levels of fishing effort illicit marked changes in prey distribution and abundance. These experiments will play an integral role in the design and evaluation of management strategies for commercial fisheries in Federal waters. FIT also anticipates that members of this research team will provide information that is directly relevant to the development of biological opinions and stock assessment advice.


Quarterly April-June 2002 sidebar Contents


Auke Bay Lab

National Marine
Mammal Lab

RACE Division

REFM Division

Quarterly Index

Quarterly Home