COMMERCIAL TRAWLING FOR FISH AND DEEP SEA LOBSTER (Puerulus spp.) OFF SOMALIA

June 1985
RAF/79/065/WP/13/84

Torbjorn Johnsen

The designations employed and the presentation of material in this publication do not imply the expression of any opinion whatsoever by the Food and Agriculture Organization of the United Nations concerning the legal status of any country, territory, city or area or of its authorities, or concerning the delimitation of its frontiers or boundaries. The views expressed are those of the authors.

1. INTRODUCTION.

The Somitfish Company was an Italian aid/joint-venture project with the Somalian Government, which has now ceased operations. Italians acted as management consultants and officers on the company vessels. The Company operated three stern trawlers in a trawl fishery along the coasts of Somalia. Both demersal fish and deep-sea lobsters (Puerulus spp.) were caught, the emphasis, however, being on the latter species. The catch was frozen for export to Italy. On board the vessels, the crew was mostly Somalian, but the officers were Italian, and were expected to train the Somalian crew in vessel maintenance, seamanship and fish processing.

This report is a summary of information collected on a 40 day cruise on the Somitfish Company trawler "Cusmaan Geedi Raage", from 5 March to 8 April 1983. It includes vessel and gear specifications and charts showing all the grounds utilized for trawling in Somalian waters. A brief description of the areas and the species/genera/families dominating the catches from the different areas is included. An estimate of the standing stocks of two species of deep sea spiny lobsters, Puerulus carinatus and Puerulus sewellii, is also presented.

Thirty days during this cruise were spent on trawling for deep sea spiny lobster and five on trawling for fish.

2. THE BOATS AND FISHING GEAR

The vessels were equipped with radar, gyrocompass, echo sounder, sonar and satellite navigator. Two types of trawls were in use, respectively for fish and lobsters. Polyvalent type trawl doors were used, suitable both for bottom and pelagic trawling. The three vessels belonging to the company were all built in Italy.

Table 1 presents specifications the company's three sisterships, "Cusmaan Geedi Raage", "21 Octoobar" and "Faarax Oomar". Table 2 presents specifications of the two trawl types in the Somitfish vessels.

Table 2 gives specifications of the trawls.

3. FISHING STRATEGY, DOMINANT SPECIES, AND PRICE STRUCTURE

3.1. The trawling grounds and dominant commercial species.

Table 3 shows the dominant species in each of the areas exploited by the Somitfish Company. This information was recorded mainly by interviewing experienced crew members. Detailed charts with depth soundings and radar bearings for each area, available at the SWIO project headqarters, are also referred to. The area and the depth range of each fishing ground were derived from the charts and from recordings made on board. Figure 1 shows the approximate location of each fishing area.

3.2. FOB prices obtained on export to Italy.

Table 4 shows the "Free on board"(FOB) price in US dollars per kilogram of each commercial category caught by trawl when exported to Italy. The categories present in the catches made during the cruise are identified by fami ly, genus or species (Fischer, W. & G. Bianchi (eds), 1984; Holthuis, L.B., 1980 and Roper, C.F.E., M.J. Sweeney & C.E. Nauen, 1984).

3.3. Limiting factors for the fishery.

Crew members with long experience fishing in Somalian waters identified the main factors limiting fishing activity as wind strength and current, rather than rough seas. The current velocity poses a problem from Brava to Ras Kiamboni, limiting trawling to the up-current direction, the speed of the vessel going down-current being such that control of the trawl was difficult. The wind seems to pose problems for fishing along the Southern Somalian coast all year long.

Only 5 of the 19 trawling areas detected up to now were usable for fishing all year. Four of these areas are on the Northern coast, and one is a sheltered shrimp trawling area near Chisimayo (Section 3.1). Fishing activity is limited to the period of the North East monsoon in the remaining areas. Since other vessels have previously been operating along the coast all the year around, this limitation may have been due to inadequate vessel design.

3.4. General strategy for locating concentrations of deep sea spiny lobsters.

The officers on board the trawler reported that the concentration of P. carinatus and P. sewellii vary with depth within a trawling ground.

The decision at which depth to start trawling within an area depended on the dominant species of spiny lobster expected. Log book records of previous operations in the area were consulted and trawling proceeded at pre-determined depths until a good catch was obtained. The area giving this good catch was then swept at other depths to find further dense concentrations of spiny lobster. When a good catch was obtained, the next haul was taken on a reciprocal track. The same ground was sometimes covered three times in this fashion.

3.5. Observed fishing activity of other vessels.

No vessels other than "Cusmaan Geedi Raage" were known to be fishing in area 20 during this cruise. In area 8 another Somitfish Co. vessel, "21 Oktobaar" was observed fishing at the same depths as "Cusmaan-Geedi Raage" in the time period from 2-11 April.

In area 11, the vessels "21 Oktobaar" and "Gabriella", a stern trawler belonging to the Seafish Co., a Sicilian fishing company licenced to fish with two vessels on the Somalian coast, were observed fishing at the same depths as "Cusmaan Geedi Raage" in the time period from 15-28 March.

No other vessels were seen from the "Cusmaan Geedi Raage" in area 14 or while this vessel was trawling for fish. Through radar observations and radio contact, however, it was known that "Faarax Omar" and "Maria Antoinette", a sistership of "Gabriella", were trawling for fish South of area 8.

Figure 1. Approximative location of the different trawling grounds on the Somalian coast.

4. RECORDS TAKEN ON BOARD

Due to the fact that virtually no information was available on fishing strategy or on working routines before the start of the cruise, a design for data recording had to be made on board (Forms 1 to 4, Appendix 1).

4.1. Catch records.

I. Lobster and fish trawling on the cruise.

Lobster trawling on the Somalian coast was in general performed at depths between 240 and 400 metres. The gear used was of the type described in Table 2.

Fish trawling in Somalia in general covered a depth range from 20 to 70 metres. The gear used was of the type described in Table 2.

Table 5 shows the coverage in time of the different areas, the depth intervals covered and the number of hauls completed within the periods of exploitation in each area.

II. Lobster trawling.

The catches are reported in Appendix 2. Each species of lobster was sorted into three categories, large (total length 19-24 cm), medium (total length 16-19 cm) and small (total length 10-16 cm), packed and frozen in 12 kg boxes. Lobsters smaller than 10 cm were discarded. The total catch of lobsters was recorded in the logbook in terms of the number of boxes processed. Box counts established the relative proportions of the two species of lobster for each depth fished.

Other crustaceans, cephalopods and gurnard were processed in the same manner, and are recorded by genus in the catch records.

III. Fish trawling.

Fish in the catch were sorted into commercial categories, packed and frozen in 20 Kg. packets (Appendix 3). Individuals weighing more than about 1.5 Kg. were gutted, headed and tailed before packing. No transformation to total weight was made in the total catch records for fish processed in this way. 'Dentice verde' (Lethrinus nebulosus), which composed 40-60% of the catch when trawling for fish, was not reported in the catch records, although they were stored. During the short period of fish trawling on this cruise the main activity of the observer involved taking photographs of the commercial species in the catches. The catch records for this period are a transcription of the logbook, and no additional records were taken.

4.2. Initial information collected on distribution of spiny lobsters by depth and location.

In all areas trawled South of Ras Hafun (area 8, 11 and 20), lobster concentrations appeared to exist within narrow depth ranges and the size and sex composition vary with depth. In area 14, however, P. sewellii is evenly distributed with depth, and no difference in size and sex composition has been observed. It was reported that all the areas North of Ras Hafun have a distribution of lobsters similar to area 14.

4.3. Length/weight relationship.

Total length and weight were recorded for P. carinatus. These recordings were not done by sex, owing to the inaccuracy of the scales available. A kitchen scale (measurement accuracy plus or minus 2.5 gm) was used for weighing. The derivation of length/weight relations for P. sewellii was done by using data recorded in Yemen (Sanders 1981). The data from this report were not processed by sex.

4.4. Length frequencies.

Total length, taken from eyestalk to end of tail, was recorded from random samples from catches at different depths. Individual lengths were grouped in 1 cm intervals. Lengths of P. carinatus were recorded in area 8 and 11. These two areas were not separated when length frequencies were recorded. For P. sewellii records were taken also in area 14 which was, however, recorded separately. All length frequencies were recorded by sex and the females were sorted into categories carrying and not carrying fertilized eggs.

4.5. Catch records.

The vessel, when trawling both for fish and for lobster, was normally able to complete 10 hauls per day. The time spent on each haul usually ranged from 1.5-2 hours (Appendix 3).

The catches of the two species of lobster (Appendix 2) usually ranged from 100-400 Kg per haul.

When trawling for fish (Appendix 3) catches usually ranged from 500-1000 Kg.

4.6. Some additional observations made on board.

I. Tuna.

The Northern waters, from Eyl (area 8) and Northwards seem to be well populated with tuna, especially skipjack tuna (Katsuwonus pelamis). Big schools of tuna were observed near the surface in an area where the depth ranged from 50 to 400 metres. Schools of tuna were following the vessel while she was trawling for deep sea lobster, feeding on the small benthic species of fish by-catch being thrown overboard.

Some attempts were made to catch the tuna using handlines made from nylon netting twine and hooks made from bent nails. One of the fish species (Chlorophthalmus bicornis) always numerous in the lobster catches turned out to be an excellent bait. With this primitive fishing gear, six of the crew members were catching 400 Kg. of tuna, for half an hour's fishing.

Since the time spent on sorting and freezing lobster and other commercial species seldom exceeds an hour, most of the sailors have more than one hour off before the next catch comes in. While fishing in one area, ten hauls were taken per day. During daylight hours, therefore, the crew could have fished for tuna for 5 one-hour periods.

II. Sharks.

Schools of big sharks were nearly always seen when trawling for deep sea lobsters. Some of these sharks were caught in the trawl as it was hauled, the largest number in one catch being 10 individuals. These sharks were 2-2.5 metres long, and such a catch represents a substantial weight. The fins and tails of the sharks were cut off and prepared for drying on board by the Somalian crew, but the rest was thrown overboard. Facilities for processing shark fillets/steaks were available on board - there should be a market somewhere for this meat.

5. ANALYTICAL METHODS.

The amount of information recorded on this cruise is limited as there was only one observer on board and fishing continued 24 hours per day.

The methods in this section represent an attempt to optimize the output derived from data collected on a commercial cruise, where the fishing strategy is motivated by profit and not research.

The catch records from lobster trawling provide enough information to enable the derivation of catch per nautical mile square (Section 5.3) for each species processed (Appendix 2).

Since the species accounting for the largest part of the catch when trawling for fish was not recorded by weight in the catch records, only total catch per nautical mile square was derived for fish (Appendix 3).

5.1. The strategy of grouping catches by coast line and depth interval.

Each area trawled for deep sea spiny lobsters can be described as a matrix, defined by depth intervals and coastline intervals. The depth interval used in the matrix has to be small because large variations in density of lobsters could be found over small depth ranges. The depth interval used here was five metres.

The appropriate coast line interval, or for the grounds South of the Horn of Africa, the latitude interval, depends on how hauls were clustered within an area. The standard coastline interval in this matrix corresponds to the distance covered per haul, which was usually between 6 and 9 miles. In areas with few trawl hauls, a longer coastline interval was taken. This was done in order to increase the precision of the estimate by having a larger number of samples within the interval and only in cases where catch rates were similar within given depth ranges. In cases where a haul was taken covering more than one coastline interval, it was ascribed to the interval where most of the haul was taken.

When trawling for fish, the mean depth of each haul was recorded. The separation into coastline intervals was performed in the same way as when trawling for lobsters.

5.2. The derivation of the area of the different trawling grounds.

For the trawling grounds being swept for lobsters during this cruise, the areas within each coast line interval (TA) were estimated from the length of the coastline interval and the mean slope of the continental shelf within the depth range trawled (S/100 metres), measured by means of echo-sounder records and position of the vessel with respect to the coast:

(3) TA = ((100/S) x I x N)/1852,

where I is the depth interval chosen (here 5 metres) and N is the number of minutes (nautical miles) in the coastline interval.

For all the fish areas and the lobster areas which were not swept during the cruise, the estimates of area are based on charts available at SWIO project headquarters.

5.3. Catch per unit area.

The catch per unit area (Appendix 2 and 3) was derived from the swept area method. The vessel was, if possible, kept at a speed of 4 knots (S) when trawling. From the trawling time (T) and the horizontal opening of the trawl (M), the swept area per haul in nautical miles square (A) can be derived by:

(1) A = S x T x (M/1852)

Given the total catch (W), the catch per nautical mile square (C) is:

(2) C = W/A

5.4. The retention coefficient of the trawl.

The retention factor used for the trawl in the estimates of standing stock is 1.0. This value corresponds to the figure most frequently used by fishery biologists (Sanders 1981, Ulltang 1980).

5.5. Standing stocks.

From the previous set of equations (1-3) the expected biomass per nautical mile square of one species of lobster (NNn) can be derived for each latitude/depth area:

(4) Nnn = (C/R) x P/100

where C is the catch per nautical mile square, R is the retention coefficient and P is the relative weight of the target species of lobster given as a percentage of the total catch of lobster (Appendix 2).

The concentrations of lobsters were observed to vary within each coast interval at different depths, with some specific depth ranges providing significantly higher catches. Although the absolute catch values varied, there appears to be a relationship between concentrations at given depths from one coastline interval to another.

Because of this depth-dependent distribution, a graphical representation of lobster concentration by depth range was generated for a coastline interval within each area which had been fished at all or most depth ranges. This 'standard curve' could then be applied to areas with less coverage in order to provide estimates of lobster density for depth ranges which had not been fished. The 'raising factor' for fitting the standard curve to the average density of any given coastline interval was obtained from trawl data in the depth ranges where trawl hauls were made.

If any depth interval had not been swept in the coastline interval chosen for standard curve derivation, the concentration derived from a catch at this depth in another coast line interval within the same area was transferred to the chosen interval. The coast line interval presenting the smallest standard deviation from the mean raising factor (Ri) was used for transfer (priority 1 in Table 6).

If the depth interval had not been swept within the same area, the concentration at this depth in a coast line interval within another area was selected in the same way as for a transfer within an area (priority 2 in Table 6).

If the depth interval had not been swept within any area or if the raising factor was derived from only one pair of hauls (i.e. R2 derived from AA2 and AB2 in Table 6), the biomass per mile square was derived through interpolation between the concentrations in adjacent shallower and deeper depth intervals (priority 3 in Table 6).

One of the coast line intervals in area 8 (7 32'N-7 41'N) turned out to be the pilot interval on this cruise. This interval was swept at many depths within the whole depth range exploited for lobster on most of the trawling grounds. This coastline interval was therefore selected for standard curve derivation in area 8.

The same proceedure was followed on selecting the coast line interval for standard curve derivation for areas 11 and 20.

When a standard curve is derived for an area, the estimated concentrations at each depth interval were transformed to biomass first for the interval in which the standard curve was derived (Table 6):

(5) B1 = AA1 x MAA1

and then for each depth interval within the other coast line intervals (Table 6):

(6) B2 = AA1 x (1/Ri) x MAB1

where MAA1 is the area of the coastline/depth interval having a lobster concentration of AA1, Ri is the raising factor and MAB1 is the area of the coastline/depth interval having a concentration of lobster of AB1.

The sum of the biomass in the coastline/depth matrix within an area is defined as standing stocks for P. carinatus and P. sewellii. (N.B. lobsters of less than 10 cm were caught in number in only one area, and are thought to be found in depths greater than those normally exploited).

TABLE 6 Standard curve derivation for one species of spiny lobster within an area. Concentrations (AA1-NNn) at different depth intervals within coastline intervals.

The standing stock within area 20 is used for prediction of standing stocks of P. carinatus in each of the other areas further South (Snn):

(7) Snn = (S20/A20) x Ann

where S20 is the standing stock of area 20 within the trawlable depth interval of the other area, A20 is the number of square miles within this depth interval part of area 20, Ann is the size of the other area.

The standing stock estimate from area 14 is used for prediction of standing stocks of P. sewellii within all the areas North of Ras Hafun. The minimum depth of occurence of P. sewellii in areas trawlable for both shrimp and lobster is assumed to be 270 metres as for the areas covered on this cruise (Appendix 2).

5.6. Length/weight relationships.

Linear regression of: W = a x L^b

was performed for both species in a 1n-1n transformation where W is total weight and L is total length, 'a' is the condition factor and 'b' indicates whether growth is allometric or isometric.

5.7. Length frequencies.

In order to obtain a realistic picture of the depth distribution of the different length groups within the population and of the length at which recruitment is initialized to the fishery for areas 8,11 and 20, the length frequency subsamples taken for each depth range were raised to reflect the biomass distribution. Bias from the depth-dependent size and frequency patterns noted in Section 4.2 was thus avoided.

Biomass for each depth interval was therefore converted to numbers of individuals using the length-weight relationship estimated for each species (Section 5.6).

6. RESULTS

Two species of deep sea spiny lobster, P. carinatus and P. sewellii were present in the catches within the depth range exploited. The Northern limit for occurence of P. carinatus appears to be around area 11, and the Southern limit for occurence of P. sewellii is around area 8. Cephalopods (Sepia spp. and Loligo spp.) do not seem to be important commercial categories (Appendix 2). Andaman lobster (Metanephrops andamanicus) is present in the catches in small amounts at depths above 350 metres in areas 8 and 11 and 320 metres in area 14. One gurnard species (Triglidae) is present in small amounts in the catches at depths shallower than 270 metres in area 20. Deep sea shrimp (Heterocarpus spp.) were a fairly important by catch in area 14.

Trawling for fish was performed within a rather restricted area. Few species are identified in catch records (Appendix 3). Records on dominant species in each fish trawling area are reported in Table 3 and Table 4.

6.1. Catch per unit area.

During lobster trawling (Appendix 2) the catch per unit area varied greatly with depth in the areas 8, 11 and 20, while that of area 14 seemed to be quite constant. In general, P. carinatus dominates the depth range from 240 to 300 metres and P. sewellii dominates from 300 to 400 metres.

The catch rate during fish trawling seems to be at an optimal level in the depth range from 34 to 40 metres, without any large variation between different areas of the coast.

6.2. Standing stocks of Puerulus spp.

Area 8

Figure 2 shows the standard curve of distribution for P. carinatus expressed in biomass. Standard deviation is shown for each depth interval. Figure 2 show the standard curve raised to fit the state of the P. carinatus stock in the other coast line intervals. The points on Figure 2 are those where catch data were available.

Figure 3 refer to P. sewellii. The biomass matrix and the estimated standing stock for both P. carinatus (161.29 tons) and P. sewellii (97.46 tons) in this area are shown in Table 7.

Area 11

Figure 4 shows the standard biomass curve for P. carinatus. Standard deviation is shown for each depth interval. Figure 4 show the standard curve raised to fit the state of the P. carinatus stock in the other coast line intervals. The points on Figure 4 are those where catch data were available.

Figure 5 refer to P. sewellii. The biomass matrix and the estimated standing stock for both P. carinatus (180.10 tons) and P. sewellii (413.55 tons) in this area are shown in Table 8.

Area 20

Figure 6 shows the standard biomass curve for P. carinatus. Standard deviation is shown for each depth interval. Figure 6 show the standard curve raised to fit the state of the P. carinatus stock in the other coast line intervals. The points on Figure 6 are those where catch data were available. The biomass matrix and the estimated standing stock for P. carinatus (74.71 tons) in this area are shown in Table 9.

Area 14

Table 10 shows for P. sewellii the mean biomass per square mile and standard deviation. The estimated standing stock within this area is 109 tons.

The whole Somalian coast.

P. carinatus is the dominant species South of area 20. The standing stock in area 20 is therefore used to estimate the stocks of this species in all areas to the South (Table 10). Similarly, the standing stock of P. sewellii in area 14 is used to estimate the stocks of this species in all areas North of area 14 (Table 10). The estimated standing stocks for the whole coast for P. carinatus is 639 tons and 1094 tons for P. sewellii (Table 10).

6.3. Spawning seasons.

During the cruise nearly all female lobsters (97-100%) with a total length more than 16 cm, both of P. carinatus and P. sewellii , were carrying fertilized eggs. The total length of 16 cm appears to be the lower limit of spawning for both species.

Figure 2. Area 8. a. Biomass within different depth intervals for the standard curve interval for P. carinatus. Standard deviation for each transfered biomass is given. b-c. Standard curve raised to fit the state of the P. carinatus stock in the other intervals in the area. The points are biomass estimates derived from catches within each of the latitude intervals.

Figure 3. Area 8. a. Biomass within different depth intervals for the standard curve interval for P. sewellii. Standard deviation for each transfered biomass is given. b-c. Standard curve raised to fit the state of the P. sewellii stock in the other intervals in the area. The points are biomass estimates derived from catches within each of the latitude intervals.

Figure 4. Area 11. a. Biomass within the different depth intervals for the standard curve interval for P. carinatus. Standard deviation for each transfered biomass is given. b-e. Standard curve raised to fit the state of the P. carinatus stock in the other intervals in the area. The points are biomass estimates derived from catches within each of the coastline intervals.

Figure 5. Area 11. a. Biomass within the different depth intervals for the standard curve interval for P. sewellii. Standard deviation for each transfered biomass is given. b-e. Standard curve raised to fit the state of the P. sewellii stock in the other intervals in the area. The points are biomass estimates derived from catches within each of the coastline intervals.

Figure 6. Area 20. a. Biomass within the different depth intervals for the standard curve interval for P. carinatus. Standard deviation for each transfered biomass is given. b-h. Standard curve raised to fit the state of the P. carinatus stock in the other intervals in the area. The points are biomass estimates derived from catches within each of the coastline intervals.

Figure 7. From area 8. The sum of the number of individuals per length group of P. carinatus, males and females, depth range 300-400 metres.

6.4. Length/weight relationship.

The linear regression of weight versus length in 1n-1n transformation gave:

6.5. Length frequencies.

Area 8.

The two equations in section 6.4 were used to transform biomass to number of individuals in the depth intervals where length frequencies were recorded. This was done in area 8 for both species (Table 11). Large individuals of P.sewelli and small individuals of P. carinatus were both found in the 330-400 metres depth range.

The number of individuals per length group for males and females of P. carinatus for area 8, derived from the depth intervals where length frequencies were recorded, is presented in Figure 7. Within the depth intervals trawled, there appear to be two exploited age groups composed of both males and females. There is a significant decrease in mean length for both males and females by depth (Figure 10).

The number of individuals per length group for males and females of P. sewellii for area 8, derived from the depth intervals where length frequencies were recorded, is presented in Figure 8. For this species there seems to be mainly one age group exploited, both for males and females, within the depth range trawled. There is also for this species a clear tendency for decreasing mean length for both males and females by depth (Figure 11). The smallest age groups are most probably deeper than the maximum depth trawled (400 metres).

Area 14.

The number of individuals per length group for males and females of P. sewellii for area 14, derived from the depth intervals where length frequencies were recorded, is presented in Figure 9 and Table 12. There appear to be two age groups for both males and females of this species exploited in this area. The length distribution is thus quite different from area 8, and the depth range narrower, from 315-335 metres.

The overall data recorded on this cruise were not sufficient to do any growth or to complete stock assessment studies.

Figure 9. The sum of individuals per length group of P. sewellii, males and females from area 14.

Figure 10. Mean length and standard deviation from the mean by depth for P. carinatus, males and females in area 8.

Figure 11. Mean length and standard deviation from the mean by depth for P. sewellii, males and females in area 8.

TABLE 11. Biomass as number of individuals from each length group per depth interval where length frequencies were recorded in area 8.

7. DISCUSSION

7.1. Standing stocks and the relative abundance of different age groups

The standing stocks derived for P. carinatus and P. sewellii are quite low. For the areas South of area 14 the estimates are derived at the end of the fishing season but for the areas from area 14 and Northwards one can expect the fishery to continue during the South East monsoon.

The strategy of exploitation of an area where only one of the deep sea spiny lobster species is expected is different from that of an area where both species coexist. In area 20, where only P. carintus is found, the depth range fished is limited to the one where one should expect to find large individuals of this species. In area 8, where the two species coexist in dense concentrations the depth range hauled is 240-400 metres. Within this range one can find the older age group of P. carinatus in the same depths as in area 20, and the younger age group together with the older age group of P. sewellii in the depth range 330-400 metres (Table 11).

In area 11, however, which is mainly inhabitated by P. sewellii , the depth range hauled is limited to the one where one would expect to find the older age group of this species.

In area 14 the same exploitation pattern seems to occur as in area 11, but in this area small individuals of P. carinatus found in the depth range 330-400 metres in area 8 and 11 .are replaced by small individuals of P. sewellii (Table 12).

In areas where there is coexistence between the two species (Figure 10 and Figure 11), the younger age groups of P. sewellii are displaced into deeper water (more than 400 metres). In areas inhabited by P. sewellii only, all age groups are found in the same depths, and there is no tendency towards a decrease in size with depth (Table 12). The depth preference of the different age groups of P. carinatus seems to be independent of the presence of P. sewellii : the mean size decreases with depth (Figure 10).

The variance between the standard curve and the trawl catch values in a given coastline interval increased from area 20 (Figure 6) in the South through area 8 (Figure 2 and Figure 3) to area 11 (Figure 4 and Figure 5) in the North. This might be due to the fact that "Cusmaan Geedi Raage" was the only vessel fishing in area 20, while there were 2 vessels in area 8 and 3 vessels in area 11 fishing at the same time. The standard curve was based on initial catches from "Cusmaan Geedi Raage" and one could not know whether some depths were already hauled by one of the other vessels. This increase in variance from one area to another might be due to an increase in undetected effort.

7.2. Spawning seasons

According to Sanders 1981, P. sewellii is most probably spawning twice a year in the Gulf of Aden. No data on spawning seasons is available for any of the two species in Somalian waters. The only indication available is verbal statements from the Russians during their period of stay in Somalia to the effect that there are two spawning seasons.

Since no data could be found on spawning cycles for P. carinatus , it is difficult to Judge whether this statement was meant for only P. sewellii or for both species.

7.3. Comments on vessels and performance

The Somitfish Company trawlers appear uneconomic for fishing along the Somalian ( coast, being too expensive both at purchase and in operational costs.

The design of the vessels could be improved upon: their susceptibility to wind drift limits their use to the Northern coast for half of the year. In addition, the diesel storage limits the time interval between each bunker to around 35 days. Of this time, 30 may be spent on fishing and the rest for cruising to and from the fishing grounds. To fill up the holds these vessels would have to obtain a mean catch per day of around 11 tons. This corresponds to about the maximum ever caught in one day by a trawler in Somalian waters. Usually while trawling for fish, the catch rates were 5-7 tons and while trawling for lobster about 1.5-2.5 tons per days fishing. It seems obvious that a freezing capacity of 30 tons per day and storage capacity of 330 tons were excessive.

A freezing capacity of 15 tons per day and total storage of 200 tons would seem convenient. The reduction of the storage capacity offers possibilities for reducing the size of the vessel, and thereby reducing the power of the engine required.

The ideal industrial trawler for the coast of Somalia may be of a smaller kind than the ones of the Somitfish Company. It must, however, be able to operate in waters with a current velocity of 2-2.5 knots, a wind of 10-12 metres/second and in waves of 2-3 metres. In addition, the construction should make pole-and-line fishing for tuna possible.

8. RECOMMENDATIONS

The logbook records on board the Somitfish Company vessels were not sufficient for management purposes. A more detailed recording system has been proposed to provide the information needed for management from both the biological and the economic points of view.

8.1. Information needed for future management purposes

Forms 1 and 2 (Appendix 1) are catch forms, respectively for lobsters and fish, and have been duplicated in sufficient number to cover more than one year of catch records from each of the Somitfish Co. vessels. These sets of forms have been given to the Somitfish Co. and were to be delivered to the captain of each vessel. The forms are written in Italian. Translation to English of each item in the forms is shown in Table 1 in Appendix 1.

I. Lobsters

Form 1 (Appendix 1) provides enough data for catch/effort studies and for the derivation of standing stocks by depth range and area. Additional information on growth through a time series of length frequency recordings and knowledge of the time interval between spawning seasons would be needed for stock assessment purposes. Form 3 (Appendix 1) should, if filled in on 3 or 4 cruises per year, provide enough information to make such a study possible. The Southern coast where P. carinatus can be caught is little fished during the South East monsoon, however, and this might leave a gap in coverage.

Form 3 is to be used for length frequency recording and should ideally be filled in for Samples from all depths trawled.

Form 4 (Appendix 1) is designed to obtain data for the transformation from biomass to number of individuals. Since the length-weight relationship for the categories specified in the form probably varies little throughout the year, only one recording should be needed, preferably at a time when both berried and non-berried females are found within the same length groups.

The samples for length frequencies are to be collected randomly, while the samples for analysis of length-weight relationship should be collected in a deterministic way, covering as wide a length range as possible for each category.

II. Fish

Due to the fact that form 2 (Appendix 1) is based on only 5 days of observations from a rather restricted area, this form might require future modification, particularly for the coloumn "cernia" which covers all species within the family Serranidae. Since the economic dominant species and the importance of this category are not known, it is at this time difficult to sub-divide the coloumns. Form 2 provides enough data for catch/effort studies and for the derivation of catch per area from each commercial category of fish specified in the form.

Only after Form 1 records covering some time are available will it be possible to decide which species have to be sampled for length-frequency. Species representing more than 10% of either catch volume or economic value should be sampled. If the price range for the categories caught in this fishery remains as small as the one shown in Table 4, there should be no need to include any economic factors. If it is decided to record length frequencies for a species, records should be taken whenever individuals from the species are present in a catch.

III. Recording of the data.

The catch record forms (Form 1 and Form 2) should be filled in by the officer in charge.

The length frequency form (Form 3) and the form for recording of length-weight relationships (Form 4) should be filled in on board by observers trained in biological sampling methods. Since fishing continues 24 hours per day, two observers should be placed on board, keeping 6 hour shifts. These observers could take part in 3 or 4 lobster cruises per year.

When more information is available from fish trawling, the same procedure should be followed.

IV. The Seafish Co.

Several attempts were made to establish contact with the branch in Mogadishu of this company, but these remained fruitless. Efforts should be made to distribute the forms on the vessels of this company operating in Somalia.

APPENDIX 1

Form 1. Catch record form to be filled in when trawling for deep sea spiny lobsters and shrimps.

Form 2. Catch record to be filled in when trawling for fish.

Form 3. Form for recording of length frequencies.

Form 4. Form for recording of individual length and weight.

APPENDIX 2

Catch records from the deep sea lobster trading.

APPENDIX 3

Catch records from the fish trawling.