No new study on sardinella stock identity has been carried out since the meeting of the Working Group in March 2000. As in the previous Working Groups agreement was reached on the existence of a single stock of both species of sardinella (FAO 2001). For further information on stock identity refer to the report of the 2001 Working Group (FAO 2001).
Total catches
Sardinella aurita
Total catches of Sardinella aurita for the whole region are presented in Table 3.2.1a and Figure 3.2.1a. It can be seen that the gradual decrease in catch of S. aurita continued in 2002, with catches declining from 295 000 tonnes in 2001 to 283 000 tonnes in 2002. Catches are now below the high level of the period 1996-2000, but still above the average level of the early 1990s.
The decline in catches over recent years was caused mainly by the reduction of catches in Zone C, north of Cape Blanc. The Russian fleet stopped operating in this area in 1999 and the Ukrainian fleet in 2001. In 2002, there was no fishery by non-Moroccan fleets at all in this zone. Since a main part of the stock is found to the north of Cape Blanc towards the end of the year (see Section 3.3.2), it was not accessible to the fishery during part of the year.
Catches in the Mauritanian zone remained at the same level as in the previous three years. A decline in catch per unit of effort in the fleet of the European Union (Section 3.3.1) was compensated for by an increase in fishing effort in this fleet. During the last quarter of the year, the abundance of S. aurita sharply declined in Mauritanian waters. Most of the EU-fleet returned to Europe, and the three remaining vessels concentrated on sardine, which appeared in very large quantities in the northern part of the area.
In Senegal, preliminary catch figures show a slight decrease in catches. This decrease, however, has to be confirmed by the final figures for the catch in 2002.
Figures for the Gambia show a huge increase in total landings of sardinellas. The species composition (S. aurita or S. maderensis) of these catches was calculated using the Senegalese artisanal catches together with that of R/V DR. FRIDTJOF NANSEN. It would appear that the increase in catches was due to an increase in availability of sardinella in Gambian waters, and not to an increase in fishing effort.
Sardinella maderensis
In 2002, the total catch of S. maderensis was 126 000 tonnes, which is close to the average of 124 000 tonnes of the previous five years (Table 3.2.1b and Figure 3.2.1b). As usual, the bulk of the catch was taken by the Senegalese artisanal fishery.
Total effort
Sardinellas are fished in four major fisheries in the subregion: the industrial fishery in Morocco, the EU-industrial fishery in Mauritania, the non-EU fishery in Mauritania, and the artisanal fishery in Senegal. Effort data for each zone are presented in Table 3.2.2. and Figures 3.2.2.a, b and c. No attempt has been made to separate effort on sardinella from effort directed at other species, and the data presented here represent the total effort exerted by the industrial fleets.
Effort in the industrial fishery in Morocco declined to zero in 2002 following the expiry of rental agreements with Ukraine and other countries. The Russian fishery in the area had already been stopped by the end of 1999 following the termination of the fisheries agreement between Morocco and the Russian Federation.
In Mauritania, the total nominal effort in the non-EU fleet has remained very stable during the last five years at an average level of 8 000 fishing days per year. This is the same level as before the temporary drop in effort in 1994-95. This fleet concentrates primarily on horse mackerel, but it will also target sardinella if horse mackerel is scarce in the area.
The EU fleet in Mauritania targets primarily sardinella. This fleet started operating in Mauritania in 1996, and it has gradually increased its effort in subsequent years. Because the vessels that have joined the fleet in recent years were larger and more powerful than the first vessels, the effort exerted by the EU vessels has to be corrected for fishing power. This has been done by assuming that fishing power is directly proportional to total engine power (main engine plus auxiliary engines), and by expressing fishing effort in "standard fishing days". The standardized fishing effort for each vessel is calculated according to the following formula:
The standardized effort calculated this way shows a strong increase in the two most recent years.
In Senegal, fishing effort in the artisanal fishery has remained fairly constant during the last five years. The effort in the industrial fleet remained low as in previous years.
Recent developments by country
Recent changes in the various economic zones are described below.
Morocco
In 2002, no fishing activity for sardinella was recorded in the zone north of Cape Blanc.
Mauritania
Sardinella (Sardinella aurita and Sardinella maderensis) is exploited by the industrial and artisanal fisheries.
The industrial fishery is carried out by foreign trawlers, most notably from the European Union, operating mainly under fishing agreements.
As far as the industrial fishery is concerned, sardinella (both S. aurita and S. maderensis) catch has trebled, increasing from 82 000 tonnes to 247 000 tonnes between 1995 and 1996. Over the last three years the catch has dramatically fallen to around 182 000 tonnes. The round sardinella is the main target species and represents over 86 percent.
As for the artisanal fishery, sardinella fishing is carried out all year round and particularly between March and July. The fleet is principally based in Nouakchott and uses purse seines. About 67 Senegalese type Mauritanian canoes are involved. Since 1999, under a fishing agreement between Mauritania and Senegal, 250 Senegalese artisanal canoes fish in Mauritanian waters subject to the same conditions as the national vessels. This fleet lands at Saint-Louis in Senegal.
The catch of pelagic species by the national artisanal fleet can be estimated at 8 300 tonnes (Mahfoudh 2002), whereas that caught by the Senegalese fleet in the Mauritanian EEZ is between 30 000 and 50 000 tonnes a year (Fifth IMROP Working Group 2002). This fishery mainly targets the round sardinella and uses purse seines at depths of between 20 and 30 m.
Senegal
In Senegal, unlike in Mauritania, the artisanal fishery (see table below) dominates the exploitation of coastal pelagic resources. The main fishing gears used are the purse seine, the ringnet and the beach seine.
|
Canoes/period |
September 1993 |
October 1995 |
September 1997 |
October 2001 |
October 2002 |
|
Purse seine |
344 |
294 |
394 |
476 |
393 |
|
Ringnet |
72 |
89 |
184 |
101 |
67 |
|
Beach seine |
91 |
95 |
177 |
85 |
85 |
The industrial fleet is composed of small local purse seiners of low tonnage ("sardiniers dakarois") and powerful foreign vessels. The number of these "sardiniers dakarois" is constantly decreasing. In 2001 and 2002 they numbered five.
The Gambia
Although sardinellas are not target species for the artisanal and industrial fisheries, landings of these species are becoming noticeable/abundant in recent years. In spite of the fact that artisanal fishers are equipped to catch Sardinella aurita and S. maderensis they target Ethmalosa fimbriata (bonga) an estuarine species as a matter of preference. Industrial fisheries for the small pelagics including sardinella are yet to be developed despite the Government's effort to encourage its development.
As already mentioned above, an increase in landings of sardinellas (Sardinella aurita and S. maderensis) by the artisanal and industrial fisheries has been observed even though they are not target species. No specific effort could be attributed to this fishery as they are considered to be a by-catch.
Mauritania
For Mauritania, two series of catch per unit of effort have been calculated; one for the EU fleet and one for the non-EU fleet. Both series refer to the combined catch of S. aurita and S. maderensis, as no distinction can be made between effort directed at either of the two species individually.
The CPUE for the EU fleet provides an index for the abundance of sardinella in Mauritanian waters. This fleet targets primarily sardinella, and switches to other species only when sardinella is not available in the area. The CPUE series for the non-EU fleet is a less reliable index for the abundance of sardinella since this fleet targets sardinella only part of the time. Splitting the effort of this fleet into a sardinella component and a horse mackerel component requires a number of arbitrary assumptions, and no attempt has been made to do so.
The two series are presented in Figure 3.3.1.1. The CPUE for the EU fleet shows a continuous decline since 1998. This decline in abundance is confirmed by vessel captains, who state that sardinellas are presently scarcer than during the early years of the fishery. The decreasing trend in abundance can also be seen in the acoustic estimates for the Mauritanian zone (Section 3.3.2).
Senegal
The CPUE series for S. aurita and S. maderensis are only given for the artisanal fishery of Senegal (Figure 3.3.1.2). These series are expressed in tonnes per trip.
The CPUE series of Sardinella aurita for the artisanal fishery show a decrease over the period 1996-1999, followed by a small increasing trend from 1999 to 2001 and a slight decline in 2002.
For Sardinella maderensis, the yearly trends are more fluctuating with the highest peaks observed in 1996 and 1999. Relative stability can be seen over the last three years.
Morocco
Figure 3.3.2.1 shows sardinella biomass in Zone C, north of Cape Blanc. Strong fluctuations from 1995 to 2002 are clearly evident. The year 1996 and subsequently 1999 and 2001 show an increase in biomass as compared to the year before. In 1997 and 1998 however, sardinella is almost absent and a drastic decrease in biomass was registered.
The round sardinella (S. aurita) predominate with over 65 percent of the total estimated sardinella biomass. In 2000 and 2002, the biomass was essentially made up of round sardinella with the notable absence in 2000 of the flat sardinella (S. maderensis). It should be noted that the total biomass of sardinella has increased considerably north of Cape Blanc since 1998.
Mauritania
The sardinella biomass estimates for the Mauritanian EEZ are shown in Figure 3.3.2.2. The flat sardinella biomass represents about 54 percent of the total sardinella biomass (except for 1999 where the percentage was only 24 percent). The overall trend to note is the decline in sardinella biomass between 1995 and 2002.
Senegal - the Gambia
Figure 3.3.2.3 shows sardinella biomass in the Senegalese-Gambian zone. Strong fluctuations of biomass can be observed with a low level from 1996 to 1998 followed by a remarkable high in 1999. From 1999 to 2000 a decrease is observed followed by an increase in 2001 and 2002. As in Mauritania, the flat sardinella is the more abundant of the two species, with about 64 percent of the total sardinella biomass. In the Senegalese-Gambian region this species was the most dominant throughout the whole period of observation.
Total subregion (Morocco, Mauritania, Senegal and the Gambia)
Figure 3.3.2.4 shows the annual sardinella biomass estimated for the subregion. Considering the total biomass in the subregion, the stability at around 3 million tonnes from 1995-1996 is followed by a reduction of around 40 percent to 1.5 million tonnes in 1997 and 1.8 million tonnes in 1998. A strong increase in biomass of the two sardinella species to over 3.6 million tonnes can be observed in 1999. In 2000, this biomass is reduced to 2.6 million tonnes before increasing again in 2001 to 3.15 million tonnes. A lower biomass of 2.1 million tonnes was registered in 2002.
Conclusion
During 1995-2002, for the whole subregion, round sardinella (S. aurita) is relatively more abundant than flat sardinella (S. maderensis). It appears that over the period covered, most of the round sardinella biomass was found north of Cape Blanc and the flat sardinella stock more to the south.
Sampling effort for S. aurita and S. maderensis in 2002 is presented in Tables 3.4.1 and 3.4.2.
In Mauritania, sampling effort was relatively high in the Russian and EU fleet, due to the presence of scientific observers on board some of the trawlers. The average number of samples per 1 000 tonnes of fish was 3.7 in the EU fleet (applies to samples for length measurements only), which is above the requirements specified in Section 1.5. For the Russian fleet, no exact number of samples per 1 000 tonnes could be calculated. No sampling was conducted of catches taken by other fleets. For the time being it is assumed that species and length composition of the Russian fleet is representative of all non-EU catches.
The figures for sampling effort in Senegal reported in Tables 3.4.1 and 3.4.2 are provisional and will be updated. As mentioned in Section 1.5, sampling in the Senegalese fishery has been low in recent years.
For the Gambia, no sampling data were available, despite an increase in sardinella catches in 2002.
Much research is presently directed at studies on growth rate and ageing of sardinella. The results of a workshop on age reading of sardinella were already presented in Section 1.6.
The last year research on growth of sardinella is concentrated mainly on the study of the formation of rings on otoliths. However, information on the growth rate of sardinella may also be obtained by following length cohorts in surveys or catch data.
The increase in number of acoustic surveys in the area may present new opportunities to study the progression of length modes over periods shorter than one year. The surveys by R/V DR. FRIDTJOF NANSEN for instance are now conducted twice a year (starting in 2001), and there may be a possibility to follow cohorts of fish from one survey to the next. Since these surveys also cover parts of the coastal areas, they may pick up cohorts of juvenile fish at an early stage and give an indication of the relative recruitment strength.
As a first attempt to follow this approach, an exploratory analysis of length data from R/V DR. FRIDTJOF NANSEN was made for the two surveys conducted in 2002 (in May-July and in November-December). Total length composition of the stock estimated in each of the economic zones is presented in Figures 3.5.1a,b,c,d,e and f. It can be seen that the spring survey found a distinct cohort of small fish with a modal length of 13-15 cm in Mauritania in May-July, whereas none of these fish were found in November-December. However, a distinct cohort of fish with a modal length of 22-25 cm was encountered in the Moroccan zone in November-December.
If the results of R/V DR. FRIDTJOF NANSEN are considered to represent the abundance and length composition of fish in the total area, one can assume that the distinct cohort of sardinella found in Morocco in November-December 2002 corresponds to the cohort of juvenile fish in Mauritania in June. That could suggest that these fish have migrated from Mauritania to Morocco, and also that they may have grown an average of 9 cm during a five month period. Assuming that the cohort found in Mauritania was born in the summer of 2001, it may mean that the fish reach a length of about 24 cm in one and a half year.
The cohorts of small sardinella found by R/V DR. FRIDTJOF NANSEN in Mauritania were also encountered in catches by the Dutch trawlers in Mauritania. Length frequency distributions of these catches (including discards) show a cohort with a modal length of 16 cm in May, and a cohort with a modal length of 22 cm in November. Taking into account the spatial distribution of the EU fishery, these cohorts may correspond to the ones observed in the acoustic surveys.
The study of length cohorts in acoustic surveys and commercial catches thus seems to offer the possibility of obtaining additional information on growth and age of sardinella. The WG therefore recommends that further studies on length distributions of sardinella from acoustic surveys and commercial catches should be undertaken in order to validate current age readings of the fish. The results of these studies should be presented in a working document to next year's meeting.
Age reading of sardinellas is still on the exploratory stage, and the age reading workshop on sardinella indicated that it may be possible to do age reading on S. aurita in the future. However, it will still take some time before the method is applicable and useful assessment time series can be produced. Therefore, it was decided to employ a production model using the BIODYN software package (Punt and Hilborn 1996), taking into account the fact that the model does not account for recruitment variability and other biological parameters.
Methods
The BIODYN software (Punt and Hilborn 1996) was used for the assessment. The "OBSTWO" spreadsheet which adapts a discrete Schaefer model to the catch and CPUE data using dynamic observation - error analysis and a maximum likelihood estimate was used (Pella and Tomlinson 1969; Butterworth and Andrew 1984; Ludwig and Walters 1985; Walters 1986). It does this by projecting the simulated CPUEs forward in time from the beginning of the catch time series and estimating parameters using the observed and estimated CPUEs of the Senegalese artisanal fishery (1990-2001).
Application of the model requires a long time series of total catch of the stock and an abundance index of the stock for the same time period.
Input data
The available catch data for the two sardinella species (S. aurita and S. maderensis) were obtained for the different countries in the study zone. The analysis covered two periods 1982-2002 and 1990-2002. During the same period the corresponding efforts for the different fisheries have also been obtained.
The CPUEs, expressed in tonnes per trip for the Senegalese artisanal fishery, were used in the application of the model. This fishery targets sardinella and is available for a long period. Even though the Dutch fleet operating in the Mauritanian EEZ mainly targets sardinella, the CPUE series only covers the years 1996 to 2002. The other fisheries, in particular Russian Federation and Mauritania, mainly target horse mackerel.
The abundance indices from the R/V DR. FRIDTJOF NANSEN surveys from 1995 to 2002 were also used.
It is important to remember that the use of dynamic models requires the estimation of supplementary parameters, which are: r (intrinsic growth rate), K (mean equilibrium biomass of the unexploited population) and Binit (biomass before the start of the first recorded catch). If the data are available, it is reasonable to assume that Binit is equal to K. The ratio between CPUE and biomass, q (catchability coefficient) was also estimated.
Results and discussion
Application of the production model shows that, for the data used, the MSY values decrease when the time series is longer. In the case of S. aurita the MSY increases from 378 000 tonnes using the 1982-2002 time series to 403 000 tonnes for 1990-2002. This agrees with the results obtained during the fifth IMROP Working Group in 2002.
In accordance with the analyses carried out in previous Working Groups, application of the BIODYN model using the Senegalese artisanal fishery CPUE indices (1990-2002) and the Nansen indices for 1995-2002 gives the following results, showing the best fit for sardinella:
Sardinella aurita (CPUE index of the Senegalese artisanal fishery) 1990-2002
|
r |
= |
1.23 |
|
K |
= |
1 309 295 tonnes |
|
Binit |
= |
740 390 tonnes |
|
MSY |
= |
403 000 tonnes |
|
SSQ |
= |
0.17 |
|
q |
= |
0.0000016 |
Sardinella maderensis (Nansen abundance index) 1995-2002
|
r |
= |
1.10 |
|
K |
= |
500 000 tonnes |
|
Binit |
= |
500 000 tonnes |
|
MSY |
= |
138 000 tonnes |
|
SSQ |
= |
0.36 |
|
q |
= |
0.0000033 |
Sardinella spp (CPUE index of the Senegalese artisanal fishery) 1990-2002
|
r |
= |
2.07 |
|
K |
= |
1 500 000 tonnes |
|
Binit |
= |
1 000 000 tonnes |
|
MSY |
= |
775 000 tonnes |
|
SSQ |
= |
0.26 |
|
q |
= |
0.0000021 |
where SSQ = the sum of the differences between the observed and estimated CPUE logarithms.
It should be noted that the results obtained for the two sardinella species separately are close to those obtained during the last Working Group. For the flat sardinella the best fit was obtained using the Nansen indices whereas for the round sardinella the best fit was found using the CPUEs from the Senegalese artisanal fishery.
Overall, for all the runs of the model, the best fit was obtained grouping together the two sardinella species.
Figure 3.6.1 shows a good fit of the model to the observed data for S. aurita for the last five years of the series. For S. maderensis the model fit is not as good (Figure 3.6.2). The model for the two species combined shows a similar fit as for S. aurita (Figure 3.6.3).
Until the problems of age reading in sardinella have been resolved, the WG cannot apply age-structured models, nor make short-term catch predictions. For the time being, the assessment is based on production models that can only give estimates of maximum sustainable yield and the corresponding effort. As mentioned above, the values for MSY strongly depend on the time period used. If a long series is used (starting from 1982), the estimated MSY is considerably lower than in case of a shorter time series (starting from 1990). This is due probably to long-term natural variations in stock abundance.
It can be noted that total catches of S. aurita have declined in recent years. Although some of this decline may be due to a reduced effort in Zone C, north of Cape Blanc, it should be noted that the total acoustic abundance of S. aurita has also declined in recent years. Another decline can be seen in the CPUE of the EU fishery in Mauritania, a fishery that is directed specifically at sardinella.
The occurrence of a distinct cohort of young S. aurita in Mauritania and Morocco in 2002 might be an indication of good recruitment in 2003. However, the contribution of this cohort to the adult stock cannot be quantified at present, and it is therefore recommended to maintain a precautionary approach. The Working Group recommends that the combined catch of both species of sardinella in 2003 should not exceed the mean level of the preceding three years, which is 420 000 tonnes.
