3.1 Stock identity
3.2 The fisheries
3.3 Catch and effort data
3.4 Biological data
3.5 Abundance indices
3.6 Assessment
3.7 Management recommendations
3.8 Future research
Sardinellas caught in the Northwest African zone are composed of two species, the round sardinella (Sardinella aurita) and the flat sardinella (Sardinella maderensis). The numerous surveys carried out show that the two species are to be found in a vast area stretching from the southern Moroccan zone to the south of Senegal. The migratory patterns of the two species strictly follow the thermal front.
The round sardinella population in the eastern-central Atlantic is composed of three stocks: (a) the Saharan sedentary stock (30°-22° lat. N); (b) the Senegalese-Mauritanian stock (approximately 22°-12° lat. N); (c) the sherbo stock (approximately 12°-7° lat. N). (Garcia 1982; Maxim and Maxim, 1988).
Young round sardinellas are found along the coast of Northwest Africa both during the cold season and during the hot season. These observations confirm the presence of nurseries to the west of the Bank of Arguin and in southern Senegal. The adults of this species migrate on a seasonal basis.
The flat sardinella is also a tropical species. It can be found from the southern Mediterranean to Angola. The Senegalese-Mauritanian stock of Sardinella maderensis is distributed along the coast (26°N-10°N).
There are two nurseries, located more or less in the same geographic zone as those of the round sardinella, for flat sardinella. The adults, which migrate seasonally, are to be found in varying quantities in all areas of the zone.
For the moment, the Working Group has agreed on the existence of one stock for these two species in the study zone (FAO/CECAF, 1994).
Gambia
Since 1991, the absence of industrial pelagic fishing vessels can be noticed. The Gambia does not have a national industrial fleet. The artisanal fishery which does not target sardinella, has not registered any major variation in landings, which are very low in any case. Sardinellas are in actual fact just a by-catch as the main target species is the bonga shad (Ethmalosa spp.).
Senegal
The fishery targets sardinella with an emphasis on the round sardinella. The industrial pelagic fishery has been dominated these past few years by the presence of chartered, Russian purse seiners and pelagic trawlers. Large tonnage purse seiners have disappeared since 1994 and only the pelagic trawlers are still in operation. The number of these has gone from fifteen in 1992 to eleven in 1999. Russian trawlers ceased operations in the Senegalese EEZ in June 1999.
The largest landings of coastal pelagic fish, mainly sardinellas, come from the artisanal fishery. The purse seiners target the round sardinella whereas the encircling gill nets mainly catch the flat sardinella.
Mauritania
This fishery is described in detail in the section on horse mackerel (Section 4.2).
The average number of vessels operating in the area is about sixty each year. During the period 1991-1999 the number of vessels deployed by the Soviet fleet (at the beginning of the period) and by the Russian Federation fleet after that was by far the largest, followed by vessels flying the Ukrainian, Dutch and Lithuanian flags. Fishing by vessels of other nationalities is of an irregular nature with a presence which ranged from just a few months to one, two or three years during the period analysed.
Morocco
Fishing for sardinellas and other pelagic species in the Moroccan EEZ began in 1993 with the decline of the sardine stock in zone C. The setting in 1996 of fishing quotas for this species at about 5% of total catch in the zone in the fishing agreements between Morocco and Russia and pelagic trawlers chartered by Moroccan professionals has increased the effort directed towards sardinella.
The fleets that fish for small pelagic fish are made up of long range trawlers which are capable of following the concentrations of fish and of processing the catch at sea (freezing, canning, fish-meal). Their number varied between ten and fifteen in 1999.
The EU fleet with two or three pelagic trawlers fishes the sardinella stocks to the south of Morocco (south 26°N) during the hot season, thereby profiting from the seasonal migration.
Catches for the period 1990-1999 are shown in Tables 3.3.1a (S. aurita) and 3.3.1b (S. maderensis). Tables 3.3.2a and 3.3.2b show the historic catches for S. aurita and S. maderensis respectively. The efforts are shown in table 3.3.3.
Catch
Gambia
The Gambia presented data on the annual catch of the two sardinella species for both industrial and artisanal fishery for the years 1981-1999. It is important to mention however, that since 1991 sardinellas have ceased to be targeted.
Senegal
Senegal has data on annual catch for the artisanal fishery for 1974-1999 and for the industrial fishery for 1966-1999.
The data collected onboard the Russian vessels is not split into Sardinella aurita and Sardinella maderensis. The two species are reported together as “sardinellas”. To obtain an estimate of the catches for each of the species, the total catch was split based on the percentages obtained from the R/V Dr. Fridtjof Nansen surveys since 1997. Prior to this the breakdowns were carried out by scientists aboard the vessels themselves.
Mauritania
Mauritania has annual catch data available for the two sardinella species for 1979-1999.
In terms of weight, sardinella catches since 1996 in the Mauritanian EEZ are the most important. This change is due in part to the entry of a new pelagic fleet from the European Union that targets the round sardinella (Sardinella aurita) in particular. The sampling on board these vessels over the past two years (1999-2000) has shown an annual percentage of sardinella catch of about 84% of the total catch.
In the log-books the catches are grouped. They have been broken down into the two species (Sardinella aurita and Sardinella maderensis) on the basis of the following sampling keys: 1990-1992 Sampling CNROP/AtlantNiro; 1993-1994 Average annual percentage of the samplings from 1990-1992 (CNROP/AtlantNiro); 1995-1998 Average annual value of the Russian samplings for 1995-1999 (AtlantNiro); Sampling by CNROP on board the European Union pelagic fishing fleet for 1999; Annual average for the two years of sampling (1999-2000) on board the European Union pelagic vessels. This average was applied to European Union catches for 1996-1998.
Morocco - Zone C
Monthly sardinella catch data are available to the group for the Russian and “Ukrainian and others” fleets from 1993. The breakdown by species (Sardinella aurita and Sardinella maderensis) for the catch was done based on samplings carried out on board the vessels during 1995-1999 to the north of Cape Blanc.
Effort
The Gambia does not possess any data on effort as its fishery does not target sardinella.
Senegal provided effort data for the artisanal fishery on a monthly and annual basis from 1982 and for the industrial fishery on a monthly basis from 1976 and on an annual basis from 1962 to 1999 (fishing days).
In Mauritania the different data that are available on effort are the effort in fishing days for the Dutch fleet for 1996-1999 and the effort in fishing days for the whole pelagic fleet for 1991-1999.
Morocco, INRH provided data on effort (fishing days) for 1993-1999 for the Russian and “Ukrainian and others” fleets in zone C. As regards the Moroccan fleet, fishing effort is concentrated on sardine. Consequently the group did not consider it.
Senegal
For each of the two species of sardinella, CRODT provided data on length frequency distributions on a monthly basis for 1986-1998 (except for 1992) for the artisanal fishery, and for the industrial fishery for 1986-1993 for the round sardinella (FAO Fisheries Report No. 657 - Supplement). These length frequency distributions are from the artisanal fishery samplings carried out on catches caught by purse seiners and encircling gill nets.
For the industrial fishery, the length measurements were carried out on species caught by sardiniers Dakarois. The data were weighted to total catch.
The other biological information available is that of the FAO Working Group of 1993 on sardinella (FAO, 1994)
Mauritania
For the European Union catches, length frequency distributions were available for the years 1999-2000.
For the Russian catches in CECAF Zones 34.1.3 and 34.3.1, a large amount of data had been received by FAO and CNROP just prior to the meeting. These data included length and age distributions from 94-99 for Sardinella aurita. Also a large amount of biological data including growth parameters was available for earlier years starting in 1973. Due to lack of time these data could not be fully analysed during the present meeting. A summary of the Russian data are presented in Appendix II.
Morocco
Morocco INRH provided the length composition for Sardinella aurita for the years 1996-1998 in zone C. These data were obtained from samplings carried out by Russian observers onboard Russian vessels. The weighting was done based on the total catch of the sample boat (FAO Fisheries Report No. 657 - Supplement).
3.5.1 Catch per unit of effort
3.5.2 Acoustic surveys
General remarks concerning sardinella CPUE series
CPUE series for sardinella were available for fisheries in Senegal, Mauritania and Morocco.
In Senegal, CPUE series for both S. aurita and S. maderensis were available for artisanal and industrial fisheries from 1980. The artisanal fishery in Senegal targets sardinella; in particular S. aurita. The CPUE for the artisanal fishery is expressed in tonnes per trip. The industrial fishery in Senegal is aimed primarily at sardinella and horse mackerel. The CPUE, expressed in tonnes per day fishing, is not converted for unities of horsepower.
For the Mauritanian area, several CPUE series were available. The last FAO Working Group that made an assessment of sardinella (Nouadhibou, 1993) used a CPUE series for 1979 - 1991 using Romanian trawlers as standard. Due to the disappearance of this type of trawler after 1991, this historic series could not be extended. The working group meeting in Dakar in 1993 used a series based on Russian effort data for 1979 - 1991.
The CNROP database presently contains catch and effort data for all vessels in Mauritania since 1991. From the material in this database, class 10 Russian trawlers (2000-4000 GRT) appeared to be the most suitable standard vessel for the construction of a new CPUE series. Vessels of this type have worked in Mauritanian waters throughout 1991 - 1999, and their characteristics (particularly their mean GRT) have been practically constant during this period. It should be noted that most of these vessels target horse mackerel and take sardinella only as a by-catch. The CPUE of these vessels, therefore, is not an ideal index of sardinella abundance. Some Russian vessels do target sardinella, but the number of vessels in this category (defined by catches containing more than 20% sardinella) is too low to construct a CPUE series for the entire period 1991-1999. Catches by the Russian fleet are reported for the two species of sardinella combined. On average, S. aurita constitutes the bulk of the catch, about 80-90% in 1999.
A shorter series of CPUE data was available for the EU trawlers in Mauritania. This fishery started in 1996 and is aimed specifically at sardinella. Catches are reported for the two species of sardinella combined. On average, S. aurita constitutes the bulk of the catch. From a preliminary analysis it appeared that the catch per day of these vessels is linearly related to total horsepower. The CPUE data were therefore converted into values for a vessel of 10,000 HP which is the average HP of vessels in this fleet.
In Mauritania and Morocco, the industrial vessels seem to fish mainly the older component of the population. The CPUE in this fishery may therefore reflect the abundance of older age groups in the stock, rather than the total biomass. In Senegal, the mean length in the catch is smaller than in Mauritania and Morocco (ref. tables in survey reports). This indicates that the fishery in Senegal is also exploiting younger age groups. Because the Senegalese fishery is exploiting a wider range of age groups in the stock, the CPUE in this fishery was assumed to represent a better index of total stock abundance than the CPUE for other areas.
CPUE series Sardinella aurita 1991 - 1999
CPUE series for S. aurita during the period 1991 - 1999 are presented in Figure 3.5.1. In addition to the two series that specifically represent S. aurita (artisanal and the industrial fishery in Senegal), the series for total sardinella of the industrial fisheries in Mauritania have also been included. The catches in these fisheries consist mainly of S. aurita, and their CPUE, therefore, is assumed to reflect the abundance of S. aurita rather than that of S. maderensis. It can be seen that the CPUE series for Mauritania and Senegal the industrial fishery generally show a decreasing trend since 1996-97.
CPUE series Sardinella maderensis 1991 - 1999
CPUE series for S. maderensis are only available for Senegal (industrial and artisanal). Both series show a decline since 1996, except for the artisanal fishery in 1999 (Figure 3.5.2).
CPUE series for S. aurita and S. maderensis combined for 1991-1999
CPUE series for both species combined are presented in Figure 3.5.3. For the more recent years a peak is noticed in 1995 and a decline in subsequent years for the industrial fishery whereas the artisanal fishery shows a peak in 1996, followed by a decline until 1998. From 1998-1999, a small increase can be observed.
General remarks concerning sardinella acoustic surveys
Series of acoustic surveys are available from surveys by the Norwegian R/V Dr. Fridtjof Nansen and the Russian R/V AtlantNiro and R/V Atlantida. Since 1995, the Norwegian surveys have always been conducted in November-December. This is the time of the year when sardinella are migrating south. Although the distribution of the stock during the survey may change from year to year (depending on the timing of the southward migration), the surveys are supposed to provide reliable indices for the total stock in the area. So far, the results of the surveys have been expressed in biomass and in numbers per length group. No age compositions are yet available for the surveys. The total biomass in some years includes a juvenile component (FAO Fisheries Report No. 657 - Supplement).
The Russian surveys were conducted at various times of the year and they covered different areas. Since they never extended into Senegal, they will only have covered part of the stock.
Results of acoustic surveys
The indices from surveys of the R/V Dr. Fridtjof Nansen (Figure 3.5.4) show the same trend for both species of sardinella, i.e. a decline from 1995 to 1997, and a strong recovery in 1999. In the last year (2000), the indices for both species have decreased again. The strong increase in biomass in 1999 was only partly due to a good recruitment in that year. Also the number of larger (and older) fish recorded by the survey increased from 1998 to 1999. The proportion of S. aurita and S. maderensis in the acoustic estimates was about equal in the years 1995-1998. Starting from 1999, the proportion of S. maderensis in the biomass estimates was less than that of S. aurita.
The surveys by the Russian vessels are presented in Figure 3.5.5. In years with 2 surveys (1995 and 1998) only the value for the second survey (July-September) has been plotted. Compared to the Norwegian surveys, the Russian surveys provided much lower estimates of stock size. This is most likely due to incomplete coverage of the stock.
The group was able to compile the available information on catch and effort. The annual catch series for Sardinella aurita, Sardinella maderensis and for the two species grouped together from 1980 to 1999 were compiled for the whole zone. For the same period the group was able to obtain the corresponding effort for the artisanal fishery (number of trips) and the industrial fishery (fishing days) for Senegal. The Mauritanian effort series starts in 1991 but that of the Dutch fleet is only available for the years 1996-2000. The abundance indices from the surveys conducted by R/V Dr. Fridtjof Nansen for the years 1992 and 1995-2000 were also used.
In the absence of more recent data on sardinella age composition, the possibility of applying a biomass dynamic model using the BIODYN software (Punt and Hilborn, 1996) was explored, bearing in mind the fact that the model does not incorporate recruitment variability nor other biological parameters.
Methods
The BIODYN package (Punt and Hilborn, 1996) was used for the assessment. Best-fit parameters were estimated for both species using the “OBSTWO” spreadsheet, which fits a discrete Schaefer model to the catch and catch per unit of effort data using a dynamic observation-error analysis and a maximum likelihood estimator (Pella and Tomlinson, 1969; Butterworth and Andrew, 1984; Ludwig and Walters, 1985; Walters, 1986). It does this through projecting the simulated CPUEs forward from the start of the catch time-series and estimating parameters using observed and estimated CPUEs of the artisanal fishery in Senegal (1990-1999).
Application of the model requires a reasonably long time series of total catch from the stock and an index of abundance of the stock during this time period. For the model estimates to be acceptably accurate and precise, the time-series should include periods of contrasting biomass, including a period of depletion under heavy catches followed by some recovery under reduced catches.
Data input
For both Sardinella aurita and S. maderensis there were reasonable data on the total catches, but very little reliable data on the effort in the fishery for all fleets, so the catch per unit of effort (CPUE) from the artisanal fishery of Senegal was used as an index of abundance. As an exercise, the biomass estimates from the Nansen surveys were used as well. The catches from Table 3.3.1ab were used.
When using dynamic models for fitting a global model, additional parameters that need to be estimated are: r (the intrinsic growth rate), K (the average unexploited equilibrium biomass) and Binit (the biomass prior to the first recorded catch).
If data are available from the start of the fishery, then it is reasonable to assume that Binit was equal to K (the average unexploited equilibrium biomass). The estimate of the constant of proportionality between the CPUE and biomass, q (catchability coefficient) was also assumed.
Results and discussion
A reasonable fit to the Sardinella aurita and S. maderensis data series was obtained for the Schaefer model (Figure 3.6.1a and b). Estimates of the key parameters were:
Sardinella aurita
|
r |
= |
1.42 |
|
K |
= |
1 000 000t |
|
Binit |
= |
1 000 000t |
|
MSY |
= |
354 570t |
|
SSQ |
= |
0.62 q = 0.0000018 |
|
r |
= |
0.63 |
|
K |
= |
500 000t |
|
Binit |
= |
500 000t |
|
MSY |
= |
78 713t |
|
SSQ |
= |
1.17 |
|
q |
= |
0.000003 |
For S. aurita the model estimates suggested a stable CPUE with some fluctuation from 1990 to 1996 and a decrease thereafter (Figure 3.6.1a).
For S. maderensis the model estimates suggested a decreasing CPUE with some fluctuation from 1990 to 1999 except for the years 1996 and 1999 (Figure 3.6.1a).
The data series used in this model included several gaps at the level of CPUE series and several important sources of potential error. In order to overcome these problems, a number of important assumptions have been made, and hence the results need to be interpreted with caution. However, the results do provide some useful information on the dynamics and probable status of the stocks, therefore the Working Group proposed a precautionary approximation of 500 000 tonnes for both species at the maximum level observed in the catches in the series analysed (1990-1999).
Due to the uncertainty of the data, it would be better to avoid any great increase in fishing effort on sardinella in the zone. The pressing necessity of a concerted effort in the management of this resource can be easily understood when considering the characteristics of a single stock widely distributed between several countries.
1. The continuation of the R/V Dr. Fridtjof Nansen surveys2. Joint surveys between vessels of the sub-region
3. Organise workshops on age reading
4. In the intervals between the meetings of the working groups, undertake to resolve the problems encountered, notably with regard to fishing effort and the unreliable results obtained with the global models
5. Utilise the Russian data in analytical models
