Comparison of stock assessment and trawl survey estimates of total halibut abundance at length

Assessment and trawl survey estimates of total halibut abundance by 10-cm length interval in Areas 3A and 3B combined agree quite well. This agreement indicates that the assessment estimates are not grossly inaccurate. A similar comparison for Area 2C suggests that the NMFS trawl survey underestimates halibut abundance in southeast Alaska. Background The large mark-recapture experiment begun by the Commission in 2003 shows a much larger abundance of halibut than the annual stock assessment. In Area 3A, the mark-recapture results indicate about four times the abundance estimated in the assessment, and in Area 3B about ten times (Webster 2008). A third estimate of abundance is available from the NMFS trawl survey. The swept area estimates are accurate if survey catchability is equal to one for all sizes. There are some unpublished video observations suggesting it is near one for fi sh 50-60 cm in length (the mode of trawl length frequencies). It is clearly less than one for very small halibut, and there is some suspicion that it may also be less than one for very large halibut, which may be able to outswim the trawl. This paper reports the trawl survey estimates of abundance by length interval and compares them to the stock assessment estimates. Data and estimation procedures The annual stock assessment is done by fi tting an age- and sex-structured model to coastwide commercial and survey setline data for the years beginning with 1996. The estimate of abundance in each IPHC regulatory area is calculated by applying setline survey estimates of relative abundance to the coastwide total estimated in the assessment (Clark and Hare 2008). No trawl survey data are used to estimate abundance in the assessment, but trawl survey data on the distribution of length at age by sex are used to calculate the bycatch at length, so it is a simple matter to calculate estimates of total abundance at length from the assessment estimates of total abundance at age by sex. These totals were grouped by 10-cm length interval (60-69 cm, 7079, …, 110-119, 120+) and distributed among regulatory areas using the setline survey estimate of the distribution of each length group in each year. In recent years, the NMFS trawl survey of the Gulf of Alaska was conducted in 1996, 1999, 2001, 2003, 2005, and 2007. The survey extends from Dixon Entrance (133oW) to the Islands of Four Mountains (170oW), but both Area 2C in the east and Area 4A in the west contain a lot of rough ground, so the trawl survey is most reliable in Areas 3A and 3B. Swept-area estimates of abundance at length in Areas 3A and 3B (actually by INPFC area) are directly available from NMFS. For the summaries computed herein, Area 3A comprises INPFC regions Yakutat and Kodiak. Area 3B comprises Kodiak and 70% of Shumagin. The value of 70% is the 3B proportion