NORTHEAST ARCTIC SAITHE (SUB-AREAS I AND II) .1 Status of the Fishery

5.1.1 Historical development of the fisheries (Tables 5.1–5.2)

Since the early 1960s the fishery has been dominated by purse seine and trawl fisheries, usually accounting for about 75% of the landings (Table 5.2). A traditional gill net fishery for spawning saithe accounts for about 15%. The remaining catches are by-catches from mixed fisheries. Catches declined sharply after 1976 (Table 5.1). This was partly caused by the introduction of national economic zones in 1977. The stock was accepted as exclusively Norwegian and quota restrictions were put on fishing by other countries, while the Norwegian fishery for some years remained unrestricted. In recent years the purse seine and trawl fisheries have been regulated by quotas where account has been taken of expected landings from other gears. Quotas can be transferred between purse seine and trawl fisheries if the quota allocated to one of the gears will not be taken. The target set for the total landings has generally been consistent with the scientific recommendations. Norway presently accounts for about 95% of the landings.

The purse seine fishery concentrates on schools of immature saithe in coastal areas and fjords. The trawlers operate on the coastal banks and catch both immature and mature fish. Over the years purse seiners and trawlers have taken roughly equal shares of the catches. In the first half of the 1990s trawlers took the biggest share while purse seine landings declined. Thus, the purse seine landings were only about 20% of the total in 1992–1995, whereas trawl landings accounted for more than half of the total. The decline in purse seine landings appeared to have been caused predominantly by changing market conditions. However, purse seine landings have about doubled since 1995 and now make up almost 30% of the total, while trawl landings have had a corresponding decline and have made up 30 – 40% of the total in the last three years.

1 March 1999 the minimum landing size was increased from 35–40 cm to 45 cm for trawl and conventional gears, and to 40–42 cm for purse seine, with an exception for the first 3000 t purse seine catch between 62° N and 65° 30 N, where the minimum landing size still is 35 cm.

5.1.2 Landings prior to 2000 (Table 5.1, Figure 5.1A, 5.2A)

Landings of saithe were highest in 1970–1976 with an average of 238,000 t and a maximum of 265,000 t in 1970. This period was followed by a sharp decline to a level of about 160,000 t in the years 1978–1984. Another decline followed, and from 1985 to 1991 the landings ranged from 67,000–122,000 t (Table 5.1). An increasing trend was seen after 1990 to 171,498 t in 1996. The stock predictions, however, again showed a decreasing trend. In order to reduce the exploitation rate to a sustainable level the TAC for 1997 was set at 125,000 t. Due to overfishing the total catch in 1997 was 143,760 t. The TAC for 1998 was also first set at 125,000 t. After a request from the Norwegian Ministry of Fisheries, the Institute of Marine Research (IMR) conducted an inter-sessional stock assessment on Northeast Arctic saithe in April 1998 (Anon., WD 1998). The reason behind this request was reports from Norwegian fishermen of high saithe abundance with extremely good catchabilities. This suggested that the previous assessment underestimated the stock and that the quota had been set too low. Based on this assessment IMR advised that the catch for 1998 should not exceed 150,000 t. Norwegian authorities increased the TAC for 1998 to 145,000 t, but due to overfishing the total catch was 153,822 t.

Based on the stock assessment made in August 1998 and a precautionary fishing mortality (F_pa) of 0.26 ACFM advised a TAC for 1999 of less than 87,000 t. Due to continuing reports from Norwegian fishermen of high saithe abundance and increased abundance indices of adult fish in the most recent acoustic saithe survey, Norwegian authorities asked for an updated advice. IMR therefore again conducted an inter-sessional stock assessment on Northeast Arctic saithe in December 1998 (Mehl, WD 22 1999). In this assessment preliminary catch at age data for 1998 were applied together with acoustic abundance indices from the latest survey and updated fishing effort and CPUE data from the purse seine and trawl fisheries in the tuning. In 1998 there was, however, a large inconsistency between the CPUE in the commercial fishery and the survey data (Table 5.5). The catch rates showed a considerable reduction in 1998 while the survey, as well as reports from the fishery, showed an opposite trend. It was therefore decided to run two analyses: one standard with tuning data from both the commercial fishery (Option 1) and the acoustic survey, and one with tuning data from the acoustic survey only (Option 2). Based on this assessment IMR advised that, depending on how quickly the fishing mortality could be reduced to the Fpa level, the TAC for 1999 ought to be between 120,000 t and 145,000 t (1998 TAC). Norwegian authorities set the TAC for 1999 to 144,000 t, while provisional reports of landings in 1999 indicate overfishing to a total of 150,272 t.

O:\ACFM\WGREPS\AFWG\REPORTS\2001\AFWG01.Doc 190

5.1.3 Expected landings in 2000

In August 1999 both options of the December 1998 assessment were revised with updated catch-at-age data and commercial CPUE data (ICES 2000/ACFM:3). The results of the assessment were similar to those obtained in December 1998. Also the management option tables (catch options) for 1999, which were the basis for the IMR advice to the Norwegian authorities, were updated and presented in the report. However, taking into account the probability of a new assessment at IMR in November 1999, including both data from the acoustic survey in October and commercial catch data from 1999, no short-term predictions (catch options for 2000) or medium-term forecasts were given.

ACFM, however, chose to give an advice based on the August 1999 assessment, and recommended a TAC for 2000 of less than 89,000 t (F_pa-level). ACFM also commented upon the inconsistency between the CPUE in the commercial fishery and the survey. ACFM requested that all series should be included in the tuning, but recommended an evaluation of the data before the next assessment. Norwegian authorities again asked IMR for their advice for year 2000 catch options, and IMR made a new standard assessment to base their advice upon. The assessment showed an increase in Fs from 1998 to 1999 to levels well above F_pa and a decline in SSB towards B_pa. A catch in 2000 corresponding to F_pa would be 95,000 t, with about 25% increase in SSB at the beginning of 2001, while a F₉₉(0.44) catch of 145,000 t would reduce the SSB further. A precautionary management requires a plan for bringing the biological reference points within the limit values. To meet these requirements IMR advised a gradual reduction of F towards Fpa, and suggested that a F of 0.35 in 2000 could be an appropriate first step. The corresponding catch would be 125,000 t, and Norwegian authorities set the TAC for 2000 at this level. Landings in 2000 are expected to be around the TAC of 125,000 t.

5.2 Status of Research

5.2.1 Fishing effort and catch-per-unit-effort (Tables C1–C3)

Traditionally indices of fishing effort in the purse seine fishery have been based on the number of vessels of 20–24.9 m length (Table C1a). This category has in recent years accounted for about 35–45% of the purse seine landings, and constitutes most of the specialised saithe purse seiners. The effort (number of vessels) of this length category is raised by the catches to represent the total purse seine effort. From 1994 to 1997 this fishing effort increased by nearly 40%

and from 1997 to 1998 by over 70%. The number of vessels taking part in the fishery almost doubled from 1997 to 1998, but due to regulations the catches were almost the same as in 1997. In such a situation the total number of vessels participating in a fishery is perhaps not a good measure of effort. Many of the vessels which have taken part in the fishery the last decade have accounted for only a small fraction of the purse seine catches (Table C1b). Roughly half of the vessels have caught less than 100 tonnes per year, and the sum of these catches are only about 5–10% of the total purse seine catch. Therefore, the number of vessels catching more than 100 tonnes annually seems to be a more representative and more stable measure of effort in the purse seine fishery. These numbers have been raised to the total purse seine catch (Table C1b), and the new effort series show a smaller decrease in later years than the old one (40 and 70%, respectively). Exploratory XSA runs showed higher scaled weights for the CPUE-series based on the new effort data (Mehl, WD 20 2000). The WG decided to use the new data in the assessment.

Table C2 gives catch, effort, and catch per unit effort for Norwegian trawlers since 1976. This summarises hauls where the effort has almost certainly been directed towards saithe, i.e., days with more than 50% saithe and only on trips with more than 50% saithe in the catch. The effort estimated for the directed fishery was raised by the catches to give total effort of Norwegian trawlers (Table C3). The index more than doubled from 1991 to maximum recorded level in 1995, and then decreased by more than 60% towards 1997 to the lowest effort during the last ten years. From 1997 to 1998 the effort increased by more than 50%, but due to regulations the catches were slightly lower in 1998 and the CPUE decreased by almost 40% from 1997 to 1998 (Table C2). This may at least partly be explained by the increasing problem with bycatch of saithe in the declining cod fishery. It is uncertain whether hauls from days with more than 50%

saithe and trips with more than 50% saithe in the catch were directed only towards saithe. Work is in progress to estimates new CPUE indices based on the logbook database having daily resolution (Mehl, WD 20 2000).

The combined effort index should reflect the main trends in total effort (Table C3). The total effort increased by almost 50% from 1997 to 1998 and continued to increase somewhat in 1999. As mentioned above, this increase was not accompanied by higher catches.

5.2.2 Survey results (Tables C4-C5)

Since 1985 a Norwegian acoustic survey specially designed for saithe has been conducted annually in October-November. The survey covers the near coastal banks from the Varangerfjord close to the Russian border and southwards to 62° N. The whole area has been covered since 1992, and the major parts since 1988. The aim of conducting an acoustic survey targeting Northeast Arctic saithe has been to support the stock assessment with

fishery-independent data of the abundance of the youngest saithe. The survey mainly covers the grounds where the trawl fishery takes place, normally dominated by 3–5 year-old-fish (Table C4). 2-year-old saithe, mainly inhabiting the fjords and more coastal areas, are also represented in the survey, although highly variable from year to year. In 1997 and 1998 there was a large increase in the abundance of age 5 and older saithe, confirming reports from the fishery. In 1999 the abundance of these age groups decreased somewhat, but was still at a high level compared to years before 1997 (Mehl 2000). Abundance indices for ages 2–5 from 1988 and onwards have traditionally been used for tuning, but including older ages as a 6+ group in the tuning series improved the scaled weights a little (Mehl, WD 20 2000). The WG decided to apply this series in the assessment.

Since 1995 a Norwegian acoustic survey for coastal cod has been conducted along the coast and in the fjords from Varanger to Stad in September (Berg and Albert, WD 33 2000), just prior to the saithe survey described above. This survey covers coastal areas not included in the regular saithe survey. Because saithe is also acoustically registered, this survey may provide supplementary information, especially about 2- and 3-year-old saithe which have not yet migrated out to the banks. Results from the coastal cod survey from the areas not overlapping with the saithe survey are shown in Table C5. Combining these indices with indices from the regular saithe survey in the tuning series did, however, not influence the assessment much. The WG therefore decided to only apply indices from the regular saithe survey in the assessment since this series is longer.

5.3 Data Used in the Assessment 5.3.1 Catch numbers-at-age (Table 5.3)

The age composition of Norwegian landings in 1998 was revised, resulting in only minor changes. Age composition data for 1999 was available from Norway and Germany, accounting for 96% of the landings. A Russian length composition was also available, and was applied on the Russian landings together with an age-length-key from the Norwegian trawl landings. Other countries were assumed to have the same age composition as Norwegian trawlers.

5.3.2 Weight-at-age (Table 5.4)

Constant weight at age values were used for the period 1960–1979. For subsequent years, annual estimates of weight at age in the catches were used. Weight at age in the stock was assumed to be the same as weight at age in the catch.

5.3.3 Natural mortality

A fixed natural mortality of 0.2 was used both in the assessment and the forecast.

5.3.4 Maturity-at-age (Table 5.14)

Traditionally, knife-edge maturity at age 6 has been used for this stock. In 1995, data on spawning zones recorded in otoliths in Norway were investigated. There was no evidence of change in maturation rates over the period in the assessment, and it was decided to use the same ogive for all years. This ogive is based on the distribution of age at first spawning among 8 year and older fish. It represents an approximation of the data from 1973 to 1994, with most weight given to recent observations.

5.3.5 Tuning data (Table 5.5)

The tuning is based on three data series: indices from the Norwegian acoustic survey on saithe, data from the new purse seine tuning series, and CPUE from the trawl fisheries. Due to low catches of age 2 saithe in the purse seine fishery in later years, this age group was excluded from the tuning series.

5.3.6 Recruitment indices

Reliable recruitment indices are crucial for the predictions. Attempts at establishing year class strength at age 0 or 1 have so far failed. An observer program aiming at establishing a 0-group index series has just started. In a few years (1992–1993) acoustic survey data showed promise for improving the estimate of year class strength at age 2. It may, however, vary from year to year to what extent the 2 year old saithe have migrated out from the near coast areas and are available for the acoustic saithe survey on the banks. In recent years there have been conflicting results between the catch and survey data, especially for the 1993 year class.

O:\ACFM\WGREPS\AFWG\REPORTS\2001\AFWG01.Doc 192

5.3.7 Prediction data (Table 5.14)

The input data for the predictions based on results from the XSA-analysis are given in Table 5.14. The stock number at age in 2000 was taken from the XSA for age 5 and older. The recruitment at age 2 and 3 in 1999 (1996 and 1997 year classes) was estimated using RCT3 (Section 5.5.2). The corresponding numbers at age 3 and 4 in 2000 was calculated applying a natural mortality of 0.2 and fishing mortalities according to the catches taken of these year classes. The long-term geometric mean recruitment (1960–1996) of 210 million was used for the 1998 and subsequent year classes. The natural mortality and the maturity ogive are the same as used in the assessment. For the exploitation pattern the average of 1997–1999 has been used, scaled to the 1999 level. For weight-at-age in the catch and stock, the average weight at age for the last three years in the VPA has been used.

5.4 Methods Used in the Assessment

5.4.1 VPA and tuning (Tables 5.6, Figures 5.2A–C)

Extended Survivors Analysis (XSA) was used for the assessment with the same settings as last year in the analyses.

Catchability was assumed to be independent of stock size for all ages. The tuning diagnostics are given in Table 5.6.

Figures 5.2A-C show plots of the tuning indices versus stock numbers from the VPA.

5.4.2 Recruitment (Tables 5.12–5.13, Figures 5.3 B–C)

Estimates of the recruiting year classes up to the 1995 year class from the XSA were accepted. Catches of age group 2 have declined to very low levels in recent years (Table 5.3), and retrospective analysis shows that estimates of recruitment at age 2 in the last VPA year have been unreliable (Figures 5.3B). Estimates of recruitment at age 3 have been somewhat more precise (Figure 5.3C). However, in the last two years catches of 3-year-olds have also been low.

The low catch in 1998 may be due to a weak 1995 year class, while the relatively low catch in 1999 was probably caused by the new minimum landing size introduced in 1999. RCT3-runs were therefore conducted to estimate both the 1996 and 1997 year classes, with 2 and 3 year olds from the survey as input for the estimation.

5.5 Results of the Assessment

5.5.1 Fishing mortalities and VPA (Tables 5.7–5.11, Figures 5.1A-B, 5.3A-C)

The fishing mortality (F_3–6) in 1998 was 0.29, which is well below the value of 0.39 from last year’s assessment (Figure 5.3A). Using the RCT3 estimation of the 1996 year class gives a fishing mortality (F_3–6) in 1999 of 0.37.

The XSA-estimates of the 1996–1997 year classes are not considered to be valid and these estimates are therefore put in brackets (Tables 5.8–5.9). In Table 5.11 the long-term average recruitment and recalculated total biomass are presented.

The 1990–1991 year classes are still abundant, and the 1992 year class is well represented in the catches. The 1993 year class is weak, and the 1994 year class also seems to be below average. In the acoustic survey, however, the 1993 and 1994 year classes come out as above average. The 1995 year class is poor both in the XSA-estimate and in the survey, while the 1996 year class so far is below average in the XSA-estimate but one of the strongest in the survey.

The SOP corrected stock and spawning stock biomass tables are included (Tables 5.9–5.11). The total biomass was at a maximum in 1993 and the SSB in 1998. Since then both has decreased somewhat. There are considerable SOP discrepancies in the early part of the time series which are caused by the fixed weights in the database prior to 1980.

SOP correction should therefore give better estimates of biomass, but it is not advisable to recalculate the weights on this basis because they could be interpreted as observed values.

5.5.2 Recruitment (Tables 5.11–5.13)

The RCT3 estimates (with 2-year-olds as input, Table 5.12) of the 1997 year class is 213 million individuals, while the RCT3 estimate (with 3-year-olds as input and back calculating the strength as 2 year olds, Table 5.13) of the 1996 year class gives 267 million individuals. Since the latter year class comes out as one of the strongest in the survey it was decided to use the RCT3 estimates for both the 1996 and 1997 year class and the long-term geometric mean of 210 million individuals for the 1998 and subsequent year classes in the predictions.

5.6 Reference Points

5.6.1 Biomass reference points

In 1995 MBAL for Northeast Arctic saithe was set at 170,000 t. (ICES 1996/Assess:4). This was also proposed as a suitable level for Bpa by the Study Group on the Precautionary Approach to Fisheries Management (SGPAFM, ICES 1998/ACFM:10). Based on an examination of the stock-recruitment plot ACFM reduced the B_pa to 150,000 t (ICES 1998A).

5.6.2 Fishing mortality reference points (Table 5.15, Figures 5.1C, 5.4)

Yield and SSB per recruit were based on the parameters in Table 5.14 and are presented in Table 5.15. F_0.1 and F_max were estimated to be 0.09 and 0.17, respectively, which is slightly above the values of 0.08 and 0.14 obtained last year.

The plot of SSB versus recruitment is shown in Figure 5.4. F_low, F_med, and F_high were not recalculated. The values obtained in 1998 were 0.17, 0.32, and 0.58, respectively. ACFM estimated F_pa using the formula F_pa = F_lim e^–1.645σ, with σ = 0.3 giving a Fpa = 0.26 based on an estimated F_lim = 0.45 (ICES 1998b).

5.7 Catch Options for 2001 (Short-Term Predictions) (Tables 5.16)

The management option table (Table 5.16) shows that the expected catch of 125,000 t in 2000 will reduce fishing mortality from F₉₉ (status quo) of 0.37 to 0.31. The status quo catch in 2001 is 154,000 t compared to a catch at F_pa of

The management option table (Table 5.16) shows that the expected catch of 125,000 t in 2000 will reduce fishing mortality from F₉₉ (status quo) of 0.37 to 0.31. The status quo catch in 2001 is 154,000 t compared to a catch at F_pa of