This page will be permanently under construction: Information is added on occasion. If you would like to contribute, please contact the webmaster.
Currently, only the horse mackerel stocks fished in the North-East Atlantic are assessed by a scientific working group. Assessments for stocks in the Mediterranean and off the African coast are so far not available.
| Trachurus trachurus in the North-East Atlantic |
Allocation of catches to stocks | Stock size |
North Sea horse mackerel |
Western horse mackerel [ ICES advice | Fishery independent information | State of the stock | Reference points | Harvest control laws | Management considerations] |
Southern horse mackerel [ ICES advice | Fishery independent information | Recruitment forecasting | State of the stock | Reference points | Harvest control laws | Management considerations] |
References
More information on these topics, especially figures and tables, will be added later. In the meantime, they can be found in the latest report of the ICES Working Group for the Assessment of Mackerel, Horse Mackerel, Sardine and Anchovy.
| Allocation of catches to stocks |
The last 10 years the Working Group has considered the horse mackerel in the north east Atlantic as separated into three management stocks: the North Sea, The Southern and the Western stocks (ICES 1990/Assess: 24, ICES 1991/Assess: 22). Since little information from research surveys is available, this separation is based on the observed egg distributions and the temporal and spatial distribution of the fishery. Western horse mackerel are thought to have similar migration patterns as Western mackerel. As for mackerel, the egg surveys have demonstrated that it is difficult to determine a realistic border between a western and southern spawning area. In later years some horse mackerel have been tagged in Portuguese and Spanish waters, but so far no tags have been recovered.
Based on spatial and temporal distribution of the horse mackerel fishery the catches were as in previous years allocated to the three management stocks as follows:
Western stock: Divisions IIa, IIIa (western part), Vb, IVa, VIa, VIIa–c,e–k and VIIIa,b,d,e. It seems strange that only catches from western part of Division IIIa are allocated to this stock. The reason for this is that the catches in the western part of this Division taken in the fourth quarter usually are in fact taken in Division IVa. In 1999 there were no information about where and when the Swedish catches were taken in Division IIIa (1957 t). The Working group therefore decided as in most years to allocate the total catches of Division IIIa to the western stock. At present there is only set a TAC for the western stock in EU waters. The present management area for this stock is therefore restricted to Divisions VIa, VIIa–c,e–k and VIIIa,b,d,e and western part of Division IVa, which do not cover the total distribution area. If TACs are set by stocks, they should apply to all areas where the different stocks are distributed.
North Sea stock: Divisions IIIa (eastern part), IVb,c and VIId. All catches in Division IIIa in 1999 (2,095 t) were allocated to the western stock.
Southern stock: Divisions VIIIc and IXa. All catches from these aeas are allocated to the southern stock.
Figure 1: ICES areas and Exclusive Economic Zones (EEZs) in the North-east Atlantic
| Stock Size |
Southern and western horse mackerel stocks are assessed usually by VPA (Adapt or other variations) (e.g. ICES, 1998) and tuned with biomass estimates from egg surveys (e.g. ICES, 1996 and 1997). The stock size depends on (a) the stock definition, (b) recruitment and (c) the rate of exploitation. Nevertheless, with all caution it is possible to provide estimates of the order of magnitude of the stock or biomass size of an exploited but not severely overfished stock. The spawning stock (?) biomass in Iberian waters, particular the southern continental shelf parts of the Bay of Biscay (Southern stock) is estimated by analytical methods (ICES, 1998) and by means of the egg production method (e.g. Walsh et al., 1990, Franco et al., 1993) and ranged recently between 200 and 265 thousand tonnes (ICES, 1998). Based on egg surveys, the size of the stock in the North Sea was estimated to be in the range of 220 - 250 thousand tonnes (Eltink, 1990, 1991 and 1992). However, undoubtedly the largest part of the spawning stock biomass is attributed to the Western stock. The stock size has been fluctuating throughout the past decades due to very unstable recruitment and changing levels of exploitation. The stock reached its maximum in 1988 with nearly 6.5 million tonnes and is now in the range of 1 to 1.5 million tonnes. Over the last 15 years, the stock was dominated by the extraordinarily strong 1982 year class. The 1982 cohort can be traced until it disappears in the 15+group (see Figure below). Accordingly, the fishing pressure on this stock increased drastically, with raising F’s since the mid 80’ies. In spite of having an assessment which is methodologically weak, it becomes clear that the stock is in a state of decline, which is indicated by the different methods.
Figure 2: Age composition in commercial catches of WESTERN horse mackerel (1982-1999)
| North Sea horse mackerel |
ACFM advice Applicable to 1998 and 1999
As usual no TAC advice was given by the ACFM. ACFM suggested that due to the age composition of the relatively small catches and past biomass estimates from egg-surveys, 1988-1991, the exploitation rate might have been low. From 1997 to 1999 ICES recommended that consistent with a precautionary approach a management plan including monitoring of the development of the stock and fishery with corresponding regulations should be developed and implemented.
EU has since 1987 set a TAC for EU waters in Division IIa and Sub-area IV which is a wider area than the North Sea stock is distributed in. This TAC has since 1993 been fixed at 60,000 t.
Fishery-independent Information from Egg Surveys
No egg surveys for horse mackerel have been carried out in the North since 1991. Such surveys were carried out during the period 1988-1991 and the SSB was estimated between 217 and 255 thousand tonnes the last three survey years (Eltink, 1992)
State of the Stock
It was not possible to do any analytical assessment. Estimates of total age composition are available since 1995 mainly based on Dutch samples. Estimates of age composition prior to 1995 are considered unreliable, that is, not representative for the entire fishery, and should not be used for analytical assessment. During the period the catches were relatively low with an average of 18,000 t. The catch, however, has gone up considerably in recent years, and the state of the stock is unknown. The egg surveys in later years for mackerel in the North Sea do not cover the spawning area of horse mackerel. In 1999 the catch level increased by 92% compared to the average long-term catch level, and the 1999-catch of 37224 tons is the highest on record. The present stock level is uncertain since the last SSB estimate was made in 1991. Since allocation of catches to the stock is based on the temporal and spatial distribution of the fishery it is important that catches are reported by ICES rectangle and quarters. Since there are no information of the SSB since 1991 it is not known if this stock is still exploited moderately. The Working Group therefore recommends that a new egg survey should be carried out and collection of age distribution data is improved.
Reference Points for Management Purposes
At present there is not sufficient information to estimate appropriate reference points.
Harvest Control Rules
No harvest control rules were considered since no assessment was carried out.
Management Measures and Considerations
EU has since 1987 set a TAC for EU waters in Division IIa and Sub-area IV. This TAC has been 60,000 t since 1993. However, this TAC is set for a wider area than the North Sea horse mackerel is distributed in. This TAC area also covers parts of the distribution area of western horse mackerel in EU waters of Divisions IVa and IIa. No forecast for the North Sea stock has been made for 2001. The data were insufficient to define a management plan for this stock.
The Working Group recommends that if a TAC is set for this stock, it should apply to those areas where the North Sea horse mackerel are fished, i.e. Divisions IVb,c, VIId and eastern part of Division IIIa.
| Western Horse mackerel |
ICES Advice Applicable to 1999 and 2000
For 1999 ICES advised that the catches should be effectively limited to no more than 200,000 t. This was aimed at maintaining the SSB above that which produced the 1982 year class. This advice was repeated for 2000. In addition ICES advised to close the directed trawl fishery for horse mackerel and the industrial fisheries in Divisions VIIe,f due to relatively large catches of juvenile horse mackerel. EU has set TACs for horse mackerel since 1989 in Division Vb (EU waters only), Sub areas VI and VII, Divisions VIIIa,b,d,e. These areas do not correspond to the total distribution area of western horse mackerel. The TAC should apply to all areas where western horse mackerel are fished. During the period 1994-1997 the TAC set by EU was 300,000 t, 320,000 t in 1998 and 265,000 t in 1999 and 240,000 t in 2000.
In 1998 and 1999 the catches of western horse mackerel were respectively 100% and 37% above the recommended TACs by ACFM.
Fishery Independent Information
Egg surveys
In 1998 the level of atresia observed in the western spawning area was very low (ICES, 1999/G:5). However, the fecundity estimate in 1998 was very low, possibly because of very early spawning. To clarify this the Netherlands sampled ovaries in January-April 2000. However, the problem is still not solved and there are indications that horse mackerel might be an indeterminate spawner (Eltink, WD 2000) . According Eltink (WD 2000) the historic fecundity of 1557 eggs/g female does not appear to be a serious underestimate of the potential fecundity. A revised fecundity (1481 eggs/g) and atresia (15eggs/g) estimate by Portugal for southern horse mackerel collected in 1998 (Costa, WD 2000) suggests also that the historic fecundity of 1557 eggs/g female might be valid. Furthermore, the new assessment on western horse mackerel shows that the biomass estimates from egg survey match quite well with the spawning stock biomass estimates. Therefore the WG decided to continue to use the historic fecundity estimate of 1557 eggs/g female and therefore also to use the biomass estimates from egg surveys for tuning the assessment. The working group considers the SSB estimate based on the 1998 egg surveys of 1.4 mill t (ICES 1999/G:5) still to be valid.
Environmental effects
The Norwegian fishery for horse mackerel is unregulated and is carried out by purse seiners mainly in the Norwegian economical zone in the North Sea in October. This fishery is therefore reflecting the availability of horse mackerel in these areas. There is good correlation between modelled inflow of Atlantic water the first quarter of a year and the Norwegian horse mackerel catches later that year (Iversen et al. 1998). This relation has been used to predict the catches in 1997, 1998 and 1999. The predicted catches during 1997-1999 have reflected the trend in the actual catches very well. The modelled inflow of Atlantic water the first quarter of 2000 was 2.4 Sverdrup (Iversen et.al., WD 2000) corresponding to a predicted catch of 60,000 t in 2000.
As during last year’s WG, data exploration, preliminary modelling and preliminary catch predictions were conducted by the ‘ADAPT’-type method (Gavaris, 1988) in which an arbitrary choice of selection pattern is made. This method was used at earlier Working Group meetings (1994 - 1998) to estimate the size of this stock and associated mortality rates. Since 1998, it has been used for comparability with a Baysian VPA - based assessment. The Bayesian model structure has shown extreme sensitivity of the results to inherent structural deficiencies; therefore, this year, the Working Group decided to examine the use of alternative models for the assessment of this stock. Two models were constructed which were based on an assumption of the separability of fishing mortality. The Instantaneous Separable VPA model (Kizner and Vasilyev 1997) was applied to the catch at age matrix and used to estimate time series of population abundance and fishing mortality. In addition a new model was constructed using a combination of the Pope and Shepherd (1982) separable VPA algorithm for the most recent three years of the time series and an ADAPT type structure for the earlier years. Details about the different methods used can be found in the respective Working Group reports (e.g. ICES CM 2001/ACFM:06).
The accepted SAD assessment model is fitted to the catch data for the years 1982 - 1999. The years 1997 - 1999 are modelled within the Separable VPA with a reference age for unit selection of 7 and a terminal selection of 1.2. The ADAPT VPA is applied to the years 1982 - 1996. Apart from 1992, fishing mortality at the oldest age is estimated as a scaling of the fishing mortality at ages 7 - 9 in the same year. The scaling factor is estimated as a parameter within the minimisation. After scaling, the fishing mortality at the oldest age is also used to estimate the population abundance of the plus group. The value of fishing mortality at age 10 in 1992, the oldest age of the 1982 year class (and also that of the plus group), is estimated as a parameter. At the current stage of development no estimates of the uncertainty in the point estimates is calculated.
Figure 3 (below) shows the most important stock parameters over the recent years. SSB is estimated by the model to have increased to a peak value of 2,850,000t in 1988 following the recruitment of the 1982 year class. With the lack of recruitments of equivalent magnitude, SSB declined has declined steadily until 1999. The 1999 estimate of SSB, at 1,424,000t, estimated to be above the historic low that gave rise to the 1982 year class. F is estimated by the model to have remained relatively stable within the range 0.1 - 0.25 throughout the history of the fishery. Apart from the strong 1982 year class, recruitment to the stock showed an increasing trend between 1991 and 1994 and is then estimated to have declined. However, the age of full recruitment to the fishery is 5 and catch at age data at the youngest ages is subject to higher relative errors. Given the additional sensitivity of the estimated recruitment to the value selection at the oldest age, recent recruitment trends should be treated with caution.
Reliability of the Assessment
The SAD model is at an early stage of development. The current specification of the separable model structure does not allow estimation of the selectivity at the oldest age and a formulation using similar constraints to those used in ISVPA should be considered in future developments. With the gradual reduction in the size of the 1982 year class and a consequent improvement in the assumption of the separability of fishing mortality, the assessment of this stock should become more stable. Future work should examine the sensitivity of the model to extension to the period of separability, especially back to the 1995 egg survey estimate. Estimates of uncertainty of the point estimates are not calculated, therefore the reliability of the assessment cannot be determined statistically. However, the minimisation is extremely stable, re-starts over a wide range of values converge to one solution in relatively few iteration. This gives confidence that there are no local minima and that the solution surface has a well defined global minimum.
Figure 3: Western Horse Mackerel: Results of the ADAPT assessment. a. Total landings; b. Spawning stock biomass (median, 5th, 25th, 50th, 75th and 95th percentiles of the expected SSB fitted SSB estimates from egg surveys), compared to SSB values estimated from egg surveys (as circles) and the minimum biological acceptable level (MBAL); c. Recruitment at age; d, Fishing mortality (median, 5th, 25th, 50th, 75th and 95th percentiles), mean for age groups 5-14 (unweighted and weighted by stock numbers), mean F on ages 2-4 and mean F resulting from an exploratory run with reduced natural mortality (M=0.05 instead of 0.15).
Reference points for management purposes
Biomass reference points
This stock is characterised by infrequent, extremely large recruitments. As only a short time series of data are available, it is not possible to quantify stock-recruit relationships, but one may make the precautionary assumption that the likelihood of a strong year class appearing would decline if stock size were to fall lower than the stock size at which the only such event has been observed. The basis for the level of Bpa is the stock size in 1983 (as estimated by an egg survey and the assessment), which is used as a proxy for the stock size present in 1982; that which produced the strong 1982 year class.
The egg survey biomass estimate was 530,000 t, the ISVPA version 2 model estimate of the SSB in 1982 is 930,000t and the SAD assessment estimate is 500,000. The assessment of uncertainty in the population model estimates is incomplete, and therefore it is proposed to retain the use of the egg survey biomass estimate as the reference value for Bpa. Conventionally this has been rounded to 500,000 t. The Study Group on the Precautionary Approach to Fisheries Management has accepted this Working Groups recommendation that these values should be used as Bpa.
Fishing mortality reference points
Model development for the assessment of this stock is incomplete. Two fishing mortality reference points have been calculated from the current implementation, they are F0.1 (0.15) and F35%SPR (0.15).
The stock is at present in a transition from harvesting the large 1982 year class to the fishing of younger ages. Given the early stage in the development cycle of the SAD model it was considered that the definition of Harvest control rules would, currently, be inappropriate. Further development work for the estimation of uncertainty and on the sensitivity of the model to the imposed structural constraints, will allow an evaluation of Harvest control rules in the near future.
This stock has been dependent on the abundant 1982 year class for many years and there were no significant recruitments. Recently however fisheries in Divisions VIId and VIIe,f have taken large catches of mainly juvenile horse mackerel from both the North Sea and western stocks. For example in 1998 about 13,400 t and in 1999 about 27,500 t were taken in the third and fourth quarter from Division VIId. In 1998 about 22% and in 1999 about 36% of the catches in numbers were between 1-4 years old. Similarly in Divisions VIIe-f over 42,600 t of horse mackerel were taken in the third and fourth quarter in 1998 and about 32,000 in 1999 of which 86% and 53% of the catches in numbers in the respective years were between 1-4 years old. A clear change in the age-structure of the catches from older to younger fish since 1996 is visible.
The Working Group expresses concern about this high exploitation rate of juvenile fish at a time when the TAC is considered too high for the long-term exploitation of the stock. Juvenile fisheries are common in many pelagic stocks and harvesting strategies have been developed that allow a balance of competing market demands (Herring WG 1999). In general the TAC for fisheries which heavily exploit juveniles, is lower than an adult fishery, to account for the inherent variability in the targeted year classes and the loss of potential yield. If the current increase in targeted juvenile mortality continues, landings will have to be reduced at a faster rate than that for an adult fishery. The Working Group recommends that a management strategy which allows regulation of the conflicting exploitation patterns be devised and evaluated.
If the fishing mortality in 2000 is the same as in 1999 the catch will be 280,000 t, it is predicted that continued fishing at that level will result in a catch of 260,000t in 2001. SSB will continue to decline at these catch levels from the 2000 estimate of 1322,000t to 1098,000 in 2001 and 900,000t in 2002.
The TAC has been overshot considerably since 1988 (ICES 1997/Assess:3). However, the TAC has only been given for parts of the distribution and fishing areas (EU waters). The Working Group advises that if a TAC is set for this stock, it should apply to all areas where western horse mackerel are caught, i.e. Divisions IIa, IIIa (western part), IVa, Vb, VIa, VIIa–c, VIIe–k and VIIIa,b,d,e.
| Southern horse mackerel |
ICES Advice Applicable to 1999 and 2000
ICES in 1999 stated that there are no explicit management objectives for this stock. However, for any management objectives to meet precautionary criteria, their aim should be to reduce or maintain F below Fpa and to increase or maintain spawning stock biomass above Bpa. The stock is considered to be harvested outside safe biological limits although the spawning stock is estimated above the proposed Bpa. ICES stated that fishing mortality should be reduced to below Fpa, corresponding to landings less than 59,000 t in 2000. ICES proposes that Bpa be set at 205,000 t and Fpa be established at 0.17, which is considered to provide approximately 95% probability of avoiding Flim. A total catch of 61,000 t in 2000 corresponding to Fstatus quo (F = 0.18), was considered inconsistent with the precautionary approach. ICES recommended that the TAC for this stock should only apply to Trachurus trachurus and that other species of horse mackerel be excluded. The TAC up to 1997 (73,000 t) included catches of other species of horse mackerel.
Fishery Independent Information
Trawl surveys
There are three survey series: The Portuguese July survey, the Portuguese October survey and the Spanish October survey. The two October surveys covered Sub-divisions VIIIc East, VIIIc West, IXa North (Spain) from 20–500 m depth and Sub-divisions IXa Central North, Central South and South, in Portugal, from 20–750 m depth. The same sampling methodology was used in both surveys but there were differences in the gear design, as described in ICES (1991/G:13). The Portuguese October and July survey indices and the Spanish September/October survey indices are estimated by strata for the range of distribution of horse mackerel in the area, which has been consistently sampled over the years. This corresponds to the 20–500 m strata boundaries. It was demonstrated that horse mackerel off the Portuguese shelf are stratified by length according to the depth and spawning time (ICES 1993/Assess:19). This explains the special characteristics of the composition of the catches, the lower availability of fish after first maturing which creates a peculiar selection pattern.
The catch rates from research vessel surveys in kg per towcan be compared to the total biomass trend. In 1999 the two Portuguese surveys (July and October surveys) were carried out by the research vessel "Capricornio” which is very different from the one previously used, both in terms of the vessel basic performance and gear type used. There is no estimation of the calibration factor to compare the Portuguese indices obtained in 1999 from "Capricornio”, with the rest of the series and then the 1999 data were not used for the assessment.
Portuguese surveys show similar catch rates and variability in the data, showing the following mean and standard deviation in the time series: 24.3 (±19.7) and 21.6 (±17) for July and October surveys respectively. Both surveys present similar trends for the 1995-1998 period. The Spanish October survey biomass index shows a decrease of 24% compared with the index obtained in 1998 but this is still a high value compared with the rest of the series. This series has less variability than the observed in the Portuguese series, giving a mean yield of 21 (±11.5). Spanish surveys shows a closer agreement in yields trends with the Portuguese July surveys, excepting in the 1995-1998 period.
Numbers at age are obtained from the Spanish and Portuguese bottom trawl fleets in the October surveys and from the Portuguese July survey. Age disaggregated data is only available from 1985. The Spanish September/October survey and the Portuguese October survey are carried out during the fourth quarter when the recruits have entered the area. As it was explained above, in 1999 the indices obtained from the Portuguese surveys are not comparable with the rest of the series. In the Spanish area, in 1999, the index at age 0 from the October survey shows a slight increase compared with the 1998 index, but it is still continuing the low levels obtained since 1995. In the Portuguese October survey the recruitment (age 0) observed in 1998 was one of the lowest value in the series contrasting with the extremely high value reached in 1997. It seems that there exists no good agreement in trends between these surveys in the abundance index for the 0 group. In the Spanish October survey in 1999 the yields in the range of ages from 4 to 9 years old were noticeable, as they were in 1998, changing the pattern observed in 1997. In the Portuguese July survey there is a strong fall in the observed 1995 abundance indices, except for ages 0 and 1, that it is continuing in 1997 and 1998.
Egg surveys
Some problems have been detected in the research work related with egg surveys which are important SSB index for tuning the assessment of the stock. As it is stated in ICES (2000/G:01 Ref:D, 2000/ACFM:5) more research work is needed for the adult parameters estimation (fecundity, determinate spawning, atresia and maturity ) and egg identification. The MHMEGG WG (ICES 2000/G:01 Ref:D) provided a revised estimate of the 1998 egg production using mean values instead the unusual high egg density values for two rectangles described above. Then the annual stage I egg production estimate was 17.85 x 1013 eggs (CV=42.2%). As only about 30% of the fecundity data were available from the area between Cadiz and Finisterra (IXa ICES Division), it was not possible to have an estimation of the SSB. These data have been presented to the Working Group (WD, Costa, 2000). Then the Working Group recommends to combine these data with those already presented previously for the Division VIIIc to obtain, as soon as possible, an estimation of the SSB from 1998 egg survey.
Effort and Catch per Unit Effort
The evolution of the commercial standardized effort series from the Spanish trawl fleets fishing in Sub-division VIIIc West (A Coruña) and in Sub-division VIIIc East (Avilés) from 1984 to 1999 is used for this data set. A Coruña bottom trawl fleet in 1999 reached the lowest level of effort in the series, continuing with the decreasing trend that started in 1996. In 1998 there was no reliable estimation on the A Coruña bottom trawl fleet effort. The effort in Avilés bottom trawl fleet has increased by 35.7%, comparing with the 1998 observed effort, anyway, it is maintained below the mean effort level from the total series. There is no estimation of effort from the purse seine fleets.
Further used are the commercial catch rates from the trawl fleet fishing in Sub-divisions IXa Central North, IXa Central South and South (Portugal) from 1979 to 1990 and trawl fleets from Spain fishing in Sub-division VIIIc West (A Coruña) and in Sub-division VIIIc East (Avilés) from 1983 to 1999. In 1999 both fleets show minor changes in catch rates comparing with previous years. The Avilés trawl fleet show a slight decrease in 1999, reaching a catch rate similar to the relatively low levels obtained in 1997 and 1998. For A Coruña trawl fleet a slight increase, comparing with the 1997 catch rate, is observed, but it still is at a lower level than the CPUE mean of the series (147 ± 25). In 1998 there was no effort estimation from A Coruña bottom trawl fleet. Horse mackerel trawl catch rates from the Portuguese trawl fleet fishing in Division IXa are yet not available since 1991, and the whole series needs to be revised.
Catch per unit effort at age
CPUE at age from the Galician (A Coruña) bottom trawl fleet (Sub-division VIIIc West) and from the Cantabrian (Avilés) trawl fleet fishing in Sub-division VIIIc East are available from 1984 to 1999 . As it has been observed in 1997, the catch rates of juveniles (up to age 3) from the Galician trawl fleet has been mantained in 1999. Also in 1999, there was an increase of the intermediate ages (4 – 9) for the same fleet. A similar pattern is obtained with the Aviles trawl fleet during the period 1997- 99: poor representation of the younger ages and a noticeable catch rate on intermediate ages (4 – 9). There is no estimation of effort in 1998 for A Coruña bottom trawl fleet.
In 1999 the index of the 0 group from the Spanish survey carried out in the recruitment season (October) was 30.74 fish/1/2h. The Portuguese October survey was not used in this year's assessment because was carried out with a different vessel and fishing gear from the rest of the series, and to date there is no conversion factor between vessels and gears. Figure 7.6.1 shows the evolution of these indices from 1985 to 1999. Both surveys present a high variability, especially in recent years. The variability in the Portuguese survey is higher than in the Spanish one, and no clear trends are evident over the whole Portuguese survey series. The abundance indices of the Spanish survey present a slight decreasing trend over the years. From 1989 to 1994 these surveys gave different estimates, but in 1995 both surveys indicated a low level of 0 group abundance which is in agreement with the VPA estimate. In 1996 and 1997 the recruitment indices from the Portuguese survey were much higher than the ones from the Spanish one.
Data exploration and preliminary modelling
All available data were used in the preliminary assessment of this stock. Given the high coherence of the time series and of the previous assessments carried out using Extended Survivors Analysis (XSA), no alternative methods were considered to be used with this stock. However, a production model was used, as a preliminary attempt to assess this stock with this alternative method (WD Abaunza et al, 2000). This model gave a similar perception of the stock as that from the VPA-based model: stability in the SSB estimates and the level of fishing mortality is slightly higher than the Fpa and FMSY The use of such model is not intended as a replacement to age-structured models, but as a way to corroborate the coherence of the assessment.
As in last year's assessment, XSA parameters were set at catchability independent of age for ages equal or greater than 9 years old, and the plus group at 12. The strength of shrinkage has a significant decreasing effect on the standard errors of the log catchability (Anon. 1995/Assess:2). In order to compare the independent information provided by the different fleets, XSA was firstly run with each fleet in separate, without shrinkage.
The external information used in the tuning was:
Fleet 1: Catch per unit of effort of the trawl fleet from A Coruña (VIIIc West - North Galicia)
Fleet 2: Catch per unit of effort of the trawl fleet from Avilé s (VIIIc East - Cantabrian Sea)
Fleet 3: Portuguese October Trawl Survey during the recruitment season (Division IXa)
Fleet 4: Portuguese July Trawl Survey end of spawning season in Division IXa
Fleet 5: Spanish October trawl Survey during the recruitment season (Sub-division IXa North and Division VIIIc)
In 1999 the July and October Portuguese bottom-trawl surveys were carried out in a different vessel and with a different gear. Given that a conversion factor between gears and vessels is not available, these CPUE indices for 1999 were not used in the assessment. The October Portuguese survey has been a very influential index in previous assessments, therefore a comparison was done between this year's assessment with and without the 1999 estimates of this survey. The result suggests that the inclusion or not of the last year of the October Portuguese survey may slightly change the perception of the state of the stock, not in terms of trends but of the biomass estimates over time.
The slopes of the linear regressions between log-catchability and log-population were analysed for the ages with catchability dependent on year class strength: fleets 1 and 4 presented a negative slope at age 0 with a low coefficient of determination, as did fleet 3 at ages 0 and 1 with a zero R-square. Therefore those ages were not included in the tuning, because they were not providing any information.
The SSBs estimated for 1997, 1998 and 1999 are compared by tuning fleet without shrinkage. The lowest SSB values were estimated from fleet 1 - A Coruña (VIIIc west) and the highest ones correspond to the estimates provided by fleet 3 (October Portuguese bottom-trawl survey). The 1997 and 1998 SSB estimates from the 1999 assessment agree closely with those given by fleets 1, 4 and 5. Fleets 2 and 3 provided higher values of SSB. In 1998 there was no estimate of fishing effort for fleet 1, hence that year was removed from the CPUE series. The options for the final assessment were taken in accordance with this exploratory analysis, and keeping consistency with last year's assessment.
Stock assessment
The final stock assessment was performed following the conclusions of the preliminary modelling. These assessment's SSB estimates are compared with those from the last assessment and from the preliminary assessments with each fleet at a time. Results show coherence among assessments, except those made only with fleets 2 and 3.
F estimates from this year and last year's assessment included all fleets with an F shrinkage of 1.00. It is clear that for the reference Fbar (1-11) the estimates show an extremely close agreement. Given the pattern of exploitation this stock is under a higher fishing mortality in the younger and older ages with a more reduced mortality at 4-6 years old. The estimates of Fbar (0-3) and Fbar (7-11) also show a close agreement with the assessment of last year. The retrospective SSB estimates performed by the final VPA, and the 1995 egg survey estimate, indicate a very good agreement among them.
Reliability of the assessment and uncertainty estimation
This assessment is relatively consistent with the assessments performed in previous years. The spawning stock biomass estimated from the 1995 egg surveys is in close agreement with the 1995 SSB level estimated using the two October surveys, the July survey information and the two commercial fleets. Thus this assessment seems to be reliable, with a relatively low level of uncertainty.
Reference points for management purposes
The range of SSBs is quite narrow, and no stock-dependent trend in the recruitment can be inferred from these observations. The very strong 1982 year class has contributed substantially to the SSB during the whole period 1985-1999. The lowest biomass attained during the period was 132,000 t in 1985, which originated a medium recruitment.
In 1998 ACFM defined Blim as Bloss, and Bpa was defined as Bloss x 1.5 that corresponded to 205000 t. In the past this Working Group proposed Fmax as Fpa. This was further supported by the Study Group on the Precautionary Approach to Fisheries Management (ICES 1998/ACFM:10). ACFM established Fpa as Flim * 0.63 = 0.17, which is close to the current Fmax (0.1755). Flim was considered equal to Floss. This working group considers that there are not reasons to change these reference points.
No harvest control rules were proposed neither by the Study Group on the Precautionary Approach to Fisheries Management (ICES 1998/ACFM:10) nor by this Working Group.
In the year 2000 the TAC was revised to 68000 t, which is in close agreement with last year recommendation from this working group. In 1999, F attained the same value as Fpa (F99 = 0.17). Table 7.12.1 summarises 2 management options: F status-quo and Fpa.
| References |
... will be added soon...
(C) EU-project HOMSIR 2001 | HOMSIR home | mail