Data Compilation Workshop on Northeast Arctic Greenland Halibut and Assessment Methods (DCWKNGHD)

According to ToRs four main issues regarding data on NEA Greenland halibut were addressed; commercial cpue series, survey data, new age readings and work on analytical models, with main emphasis on cpue and survey data. Updated data from Russia and Norway were exchanged, and it was agreed to develop joint data exchange formats for Greenland halibut.
Cpue series that were used as tuning series for a surplus production model presented at WKBUT 2013 were examined. Cpue series from the time when fisheries were unreg-ulated (before 1991) are considered to have potential to reflect stock dynamics for that period, and are most likely useful as model tuning series. Exploratory cpue series from the period after regulations were implemented on the fisheries fluctuate and may not reflect stock dynamics, and thus should be carefully scrutinised before considered useful as tuning series. Further work is ongoing to standardise Russian cpue series.
It was concluded that fisheries-independent data stand out as the most promising to illuminate stock dynamics, especially in the time period after 1992. Four surveys in the Barents Sea were studied. These surveys cover different parts of the slope and shelf areas, and consequently different components of NEA stock. In recent trend based as-sessment at ICES only data from the Norwegian Autumn Slope Survey and the Russian Autumn Survey have been used, being the surveys that include deep adult areas at the slope (>400 m), but the indices have shown considerable discrepancy in later years. At the workshop results from the area at the slope that these surveys have in common were studied, and interesting findings need further examination. Especially drop in the numbers of females in 2011 and 2013 in the Norwegian survey, and inverse fluctuations of mature female biomass at slope and shelf areas in the Russian survey. These and other issues need to be addressed in relation to the distribution of stock components, reflected in length distributions, including information from the Joint Ecosystem survey and Joint Winter Survey. Based on these studies the goal is to decide on indices that are the most representative to use as tuning series.

Update on routine age readings with new age-reading method in Norway was pre-sented. More age readings on the smallest individuals are needed to construct better growth curves.
Models were not the main issue at this point, but current stage in model work was presented. Based on scrutiny of the cpue series it was recommended to examine runs with the surplus production model for the period 1964–1991 and 1964–2005, in addi-tion to runs for the whole 1964–2013 period. It is possible to run the GADGET model with different growth parameters, with direct age–length data, or with only length data. This latter option would remove the need for age estimation and estimate growth directly from the length data. Experiments have suggested this is feasible, but comes at the expense of creating unrealistic recruitment patterns, as the model does not know (from the length information alone) which year to put recruitment in. This could po-tentially be remedied with a recruitment index, or by including age–length data on the younger individuals (where all age-reading methods are in agreement) as an indication of which year the fish had entered the population. The GADGET model is in a state where it is ready to be run once the suite of possible tuning series is available, but may be best utilized in conjunction with the production model. In this case Gadget could give the annual stock estimate, while the production model would help define the ref-erence points.