Indicators of performance in the fishery for shrimp Palaemon serratus (pennant) in Irish Coastal Waters.

ABSTRACT A commercial fishery for shrimp Palaemon serratus occurs on the south and west coasts of Ireland. Two temporal phases were identified: expansion from the mid 1970s to the late 1980s when maximum national landings were less than 150 t annually; and a second phase since 1990 when most years yielded more than 200 t. Maximum landings of 548 tons were reached in 1999. Three indicators of fishery performance are examined: total annual landings, weight landed per vessel per day and the average weights of individual shrimp in October and December, reconstructed from processors' records of graded shrimp. Total landings and daily landings weights correlated positively, whereas both correlated negatively with average individual shrimp weight in October and December over a 13-year period. Periodic and possibly cyclical recruitment events are identified as a contributor to improved yield.

KEY WORDS: Palaemon serratus, Palaemon elegans, Ireland, inshore, shrimp fishery 

INTRODUCTION

A trap (creel) fishery for shrimp commenced in the vicinity of Baltimore, southwest Ireland in the mid 1970s (McPadden 1979). It subsequently expanded eastwards along the south and northwards along the west coasts. The species making up the majority of the landings was Palaemon serratus (Pennant), the smaller P. elegans accounting for less than 1% of landings by number (Fahy et al. 1998a). Cole (1958) and Forster (1951) described P. serratus as having a pronounced southern distribution in Britain and the majority of landings in Ireland came from the south west (Fahy & Gleeson 1996) (Fig. 1).

[FIGURE 1 OMITTED]

Shrimp are fished in plastic Chinese-hat-ended traps baited with whitefish or pelagic fish offal. Berried females fetch the best prices and are targeted by the fishery. These are most accessible and concentrated inshore during the autumn when growth is fast and an individual could double its weight in four to five months (Fahy & Gleeson 1996). However, shrimp are poorly selected by the square mesh used in creels (Fahy et al. 1998b); all sizes above 55 mm total length (which was the threshold at which recruitment commenced) are generally retained.

The growth and longevity of P. serratus have been investigated by a number of workers: Sollaud (1916) believed the species lived for five or six years; Cole (1958) thought it probable it survived for four but Forster (1951, 1959) favored two and Figueras (1986) described the expected life span as three years.

Significant elements in the life cycle reconstructed from data in Fahy and Gleeson (1996) and corroborated in Fahy et al. 1998b are presented in Figure 2 on which the traditional fishing season (August to January) is marked with rectangles whose shading indicates its relative importance to the industry (0-group shrimp are harvested and sold but they realize low prices, shrimp in their second autumn are the principal target and are most valuable, whereas female shrimp that might be in their third autumn are infrequent). Shrimp hatch in May and first recruit to the creels the following August and September. Discarding was an exceptional practice during the period under review (the majority of fishers sold the unsorted contents of their creels), hence 0-group animals were retained in the landings. Currently fishing also takes place up to 6 nm offshore when the females migrate into deeper water during the colder months after the end of the year.

[FIGURE 2 OMITTED]

MATERIALS AND METHODS

Total landings of shrimp were collected by the government department responsible for fisheries (currently the Department of Communications, Marine and Natural Resources) on a national scale (Fig. 3).

[FIGURE 3 OMITTED]

Details of shrimp purchases and grading were obtained from five processors, labeled A to E. These processors purchased shrimp from a wide geographic range and several contemporaneously bought landings from the same area. Processors mechanically sorted shrimp using vibrating screens into four or five size grades at least five 1-kg samples of each of which were obtained and analyzed in the Marine Institute.

In the laboratory samples were disaggregated by sex; females have a faster growth rate than males. Individual measurements of total length from the tip of the rostrum to the end of the telson were noted (0.1 mm), the shrimp having been straightened against a rigid surface and individual weights (0.1 g) were recorded. Length frequencies separated by sex were predominantly bimodal with the two modes interpreted as the 0 and 1-y groups. However some females exceeding 24 mo in age may also have been present but not identifiable in the length frequencies. The bimodal length frequencies were sliced (separated at the lowest length frequency) to divide year classes. The consequence for the average shrimp weights of altering the proportions of the two age groups was modeled.

Definitions of grades varied from one buyer to another and the average weights of corresponding grades among processors differed; to maintain consistency of product, a processor maintained his sorting screen characteristics from one year to another. Table 1 contains the average weights of shrimp per grade analyzed by the Marine Institute and used throughout the period of the investigation to estimate the average individual weight of shrimp landed.

Processors provided details of the weight composition of graded shrimp in the landings they purchased (the actual weights of grades 1-4 or 5 after screening) together with dates and places of purchase, and from these and the average weight of shrimp within a grade the average weights of shrimp captured from an area in a specific time period was calculated according to the formula:

[n.summation over (i=1)] (GrWt_I * AvWt_i)

where GrWt_i is the proportion of stock at this grade and AvWt_i is its average individual weight, 1 was the largest and n the smallest grade.

The time period to which average weights referred was dependent on factory procedure: landings might be sorted daily or weekly and some archived data were for landings bought-in over a month.

The average individual weight in October and December was calculated using all consignments in these months. Grading was not always undertaken by a processor and such data as existed were not always made available. In 1991 grading details of 3% of total landings were obtained; the following year this increased to 98%. Thereafter it fluctuated between 20% (in 1993 and 2001) and 50% (in 1995), averaging at 35% over the period. In all 162 estimates of October and December weights were made and these were distributed among 92 combinations of processor, area and year from 1991 to 2003 (Table 2).

The majority of shrimp were fished mainly by small inshore vessels (<10 m) crewed by two persons fishing 300-500 creels. Logbooks were not maintained in this fishery. Records of daily landings by a single vessel, which were then consigned to a processor, were obtained from the same geographical range and from the same companies by scrutiny of payment records; individual boats were not identified. An estimated 3% of all landings were reported as daily landings in 1991 and 25% the following year. An average of 13% of total landings were reported as daily landings overall. Daily consignment numbers averaged 1,076 annually and ranged from 347 in 1991 to 1,906 in 1994.

The areas from which shrimp were purchased (numbered 1-8) are shown in Figure 1. Area 1 is in County Galway, area 2 is County Kerry, areas 3-7 are in County Cork, and area 8 is County Water ford.

The statistical package used in the work was SPSS 11.0.

RESULTS

The Palaemon shrimp fishery in Ireland is conducted close inshore in the autumn and it is concentrated on the south west of the country. Few shrimp are harvested in the Irish Sea (Fig. 1). A fishery exists in the northwest where shrimp accumulated in storage boxes are periodically collected for processing, which is also carried out mainly in the southwest. Between 1995 and 2003 the counties that provided the data on which this investigation is based recorded 93% of the total landings of shrimp (Galway, 20%, Kerry, 23%, Cork 45% and Waterford, 6%).

The fishery targeted 1-group shrimp (Fig. 2), especially females, which are twice as heavy as males during their second autumn. The fishery was mainly conducted from August to January when 1-group females moved close inshore. Females outnumber males in landings, the majority of which are harvested within 1 n m of the coast. Not all females have reached their asymptotic length at this age however and they continue their growth in deeper water during the colder months. Very few females survive to contribute to the next fishing season. Investigations on shrimp in southwest Ireland in the mid1990s supported Forster's interpretation of life expectancy for practical assessment purposes although the longevity of some few females might exceed 24 mo (Fahy & Gleeson 1996).

Annual landings (Fig. 3) had 2 phases: a period of low catches up to 1989 (142 t in 1982 was exceptional) and a subsequent period of expanded catches; other than in 1992 when they totaled 164 and in 1993 269 t, landings of shrimp exceeded 300 t per year during the latter period. In 1999 landings peaked at 548 t after which they fell to 115 t in 2002, recovering again to 413 t in 2004.

Mean daily landings per boat over the period were 34 kg (ranging from 23 kg in 1992 to 45 kg in 1998) and analysis of weight frequency over the period recorded 90% was less than 75 kg (n = 12,669, between 1990 and 2003). Landings of as low as 2 kg were recorded and the upper end of the range was at 250 kg (Fig. 3). Daily average consignment weights correlated significantly with total landings ([r.sup.2] = 0.353, n = 14, P = 0.0251).

All processor/area/year combinations of data covering the period September to December were considered in this analysis and for these the average individual shrimp weights in October and December were calculated. Selections of these data sets are shown in Figure 4 to illustrate the nature of the data. In the majority of panels there was some increase in weight in the autumn months and the average individual weight frequently declined coinciding with the move of females into deeper water in the colder months.

[FIGURE 4 OMITTED]

Forster (1951), Cole (1958), Sollaud (1916) and Desbrosses (1951) identified a number of variations in growth and biology of Palaemonid shrimp, which might be attributed to location (sex ratio favors females in rocky areas and males where the substratum provides less cover, for example) so the average individual weights were examined for bias, which might be linked to processor or area. Average individual weights were higher in areas 1 (4.95 g) and 8 (4.92 g) than elsewhere (3.9-4.2 g). In the rocky environment of Connemara (area 1) shrimp are known to be of better commercial quality because of the higher incidence of berried females. All records from area 1 were provided by processor D. Two processors, A and B, handled shrimp from area 8 and both provided a similar weight range. Area 8 includes the only area in which fishers were known to have graded their catches. An analysis of variance found a-significant difference between processor and average shrimp weight (df = 3, F = 3.420, P = 0.019) and between area and average shrimp weight (df = 6, F = 2.740, P = 0.015) but average shrimp weights in samples from the same area handled by different processors did not differ significantly (df = 4, F = 1.685, P = 0.157).

Areas 1 and 8 had higher average shrimp weights, but data from these areas were more plentiful in some years than others hence, when they did occur they might have distorted the results in those years, so samples from areas 1 and 8 were excluded from all further analyses, and 147 average individual shrimp weights were used in the remaining calculations.

Average individual shrimp weights displayed periodic trends during the study period: between 1991 and 1995 average individual weights in October and December rose (x-var = 0.365; n = 29, [r.sup.2] = 0.0384, P = 0.0003), from 1995 to 1998 they declined (x-var = -0.189, n = 30, [r.sup.2] = 0.152, P = 0.033) and from 1998 until 2001 they increased again (x-var = 0.315, n = 52, [r.sup.2] = 0.270, P < 0.0001) (Fig. 5). A trough occurred in 1998 and the average weights tended downward again, though not significantly, between 2001 and 2003.

[FIGURE 5 OMITTED]

Average individual shrimp weight correlated negatively with daily consignment weight (n = 147, [r.sup.2] = 0.0816, P = 0.0005) and national landings (n = 147, [r.sup.2] = 0.0711, P = 0.001).

A correlation of average shrimp weight with time, superimposed on the data in Figure 6, suggests an increase in average shrimp weight over the period of the study.

[FIGURE 6 OMITTED]

The relationship between the proportion 0-group and average individual shrimp weight was modeled on the composition of shrimp samples provided in October and November by processor A. Disaggregated into males and females the length frequency compositions were sliced into 0 and 1-groups. In the model the proportion of 0-y shrimp was increased and the consequent average individual shrimp weight was noted (Fig. 7). An inverse relationship between average shrimp weight and the proportion of 0-group shrimp also suggested that the recruitment of 0-group shrimp influenced yield.

[FIGURE 7 OMITTED]

DISCUSSION

Expanding from the mid1970s, the inshore fishery for shrimp reached its maximum geographical extent in the early 1990s. The fishery is unregulated.

The maximum average individual weight in a shrimp catch from the inshore fishery is recorded in the late autumn, between October and December. Average individual weight is influenced by the sex ratio within the catch, whether it has been sorted to remove 0-group individuals and the relative proportions of 0 and 1-group animals represented (Fahy & Gleeson 1996, Fahy et al. 1998b).

Annual shrimp landings correlated positively with daily landings suggesting that the latter reflected changes in shrimp abundance. Greater abundance of a target species attracts additional fishing effort and that, in turn, increases yield. While logbooks were not maintained in this fishery, it is possible to obtain an estimate of fishing effort from the average daily and total annual landings. Using these data fishing days annually would have averaged 9,266 between 1991 and 2003 and ranged from 3,263 in 2002 to 13,650 in 1994. Fishing effort is strongly correlated with total landings (n = 13, [r.sup.2] = 0.6919, P = 0.0004).

Successful fishing could be based on a strong 1-y class of higher first sale value or an abundant 0-y class contributing greater volume but less unit value. In the analysis annual landings and daily consignment size both correlated negatively with average shrimp weight in October and December suggesting that increased landings were caused by greater proportions of 0-group shrimp and increased recruitment.

Figure 7 illustrates how a decreasing average shrimp weight signals this recruitment but is not a sensitive indicator. Modeled average shrimp weights were 5.45 g when they consisted of only 1-group animals; demonstrable change was brought about by the addition of an equal weight of 0-group shrimp, which reduced the average weight to approximately half (2.85 g). Doubling the addition of 0-group shrimp further reduced the average weight to 2.42 g (by a further 0.44 g) and tripling it reduced the average weight to 2.27 (by an additional 0.15 g). In terms of numbers, working from the data shown in Figure 2, the addition of 6:1 0-group to 1-group could increase the total weight of a consignment in the autumn fishery by 300%.

An investigation of the commercial shrimp fishery (Fahy & Gleeson 1996) provided some analyses of the age and sex composition of graded shrimp but had limited access to quantitative factory grading data. Without disaggregating for sex, they sliced the length frequency distributions at three points (65, 70 and 75 mm) to separate the 0 and 1-groups. At the time (early 1990s) there would have been a low incidence of 0-group shrimp in the landings anyway, but the slice point was also too far to the left, 85 mm would have been more appropriate. Fahy and Gleeson's review of the fishery in the 1990s considered it well managed. However, they noted the tendency for the autumn fishery to extend over a longer period, a trend that has since accentuated. The pursuit of shrimp offshore during the colder months of the year has also intensified as stock depletion of other inshore species and the capitalization of the inshore sector have progressed.

Fahy and Gleeson (1996) reported that shrimp are berried for much of the year: 1-group females become ovigerous in the autumn but 0-group come into berry the following spring. Berried females are the most commercially valuable category of shrimp and confining the fishery to the autumn allowed the surviving 0-group cohort to reproduce. They expressed concern that a fishery extending over much of the year could deplete shrimp spawning stock biomass.

The gradual upward trend in average weights since the fishery completed its full expansion in the early 1990s (Fig. 6) may indicate a trend towards lower recruitments. Episodic larger recruitments are associated with shrimp of small average size and low value; larger shrimp are more valuable and their exploitation offshore when stocks are at a low level combined with their exploitation during the spring reproductive phase could culminate in growth overfishing, and this may now be a reality. A precautionary approach to prevent growth overfishing is required and an obvious initiative would be the restoration of the traditional fishery confining exploitation to inshore waters in the autumn months.

ACKNOWLEDGMENTS

The following processors provided data and samples for these investigations: Carr & Sons, Tallow, Co Waterford, Glenmar Shellfish, Skibbereen, Co Cork, Eiranova, Castletownbere, Co Cork, Bereci Shellfish, Middleton, Co Cork and Porcupine Fisheries, Cleggan, Co Galway. Eadaoin Healy of RPS consultants, Letterkenny, mapped inshore shrimp fisheries presented in Fig 1. Trevor Alcorn, of the Marine Institute, provided advice on GPS and Jonathan White and Neil O'Maoileidigh commented on an earlier draft of the MS.

LITERATURE CITED

Cole, H. A. 1958. Notes on the biology of the common prawn. Fish Invest. London, Ser. 22:22.

Desbrosses, P. 1951. Les crevettes roses. Office Sci. et Tech. Peches. Maritimes, Bull. Inform. Et Documentation, Ser. B, N.S. No.39. Paris.

Fahy, E. & P. Gleeson. 1996. The commercial exploitation of shrimp Palaemon serratus (Pennant) in Ireland. Ir. Fish. Invests. new series. 1:1-28.

Fahy, E., N. Forrest & P. Gleeson. 1998a. Estimates of the contribution of Palaemon elegans Rathke to commercial shrimp landings in Ireland and observations on its biology. Ir. Nat. J. 26(3/4):93-98.

Fahy, E., N. Forrest & L. Oakey. 1998b. Catch analysis of shrimp Palaemon serratus (Pennant) taken by different mesh sizes. Fisheries Bulletin No, 16. 11 pp.

Figueras, A. J. 1986. Crecimiento de Palaemon adspersus (Rathke, 1837) y P. serratus (Pennant, 1777) (Decapoda) en la Ria de Vigo (SO de Espana) Scientia Marina 50:117-128.

Forster, G. R. 1951. The biology of the common prawn Leander serratus Pennant. J. Mar. Biol. Assoc. U.K. 30:333-360.

Forster, G. R. 1959. The biology of the prawn, Palaemon (= Leander) serratus (Pennant). J. Mar. Biol. Assoc. U.K. 38:622-627.

McPadden, C. 1979 Exploratory and experimental fishing for Palaemon serratus (Pennant). M.Sc. thesis. National University of Ireland, Galway.

Sollaud, E. 1916. Recherches sur la bionomie des "Palemonides" des cotes de France. Rec. Fonds. Bonaparte 1:69-71.

EDWARD FAHY, * NIAMH FORREST, MARGARET O'TOOLE, RUTH MORTIMER AND JIM CARROLL Fisheries Science Services, Marine Institute, Snugboro Road, Dublin 15, Ireland

* Corresponding author. E-mail: [e-mail]

TABLE 1.
Average individual weights (g) of shrimp in grades sorted by five
processors. The averages are of five or more samples of each grade.

Grades/Processors    A     B     C     D     E

        1           7.4   9.3   6.2   8.3   11.5
        2           4.7   7.0   4.5   5.9    7.2
        3           3.3   2.4   3.0   3.7    4.4
        4           2.2   1.8   1.5   2.9    2.0
        5                             2.2

TABLE 2.
The origin of average individual shrimp weights in October and
December.

Processors/Areas   1     2    3   4   5   6    7    8   Totals

       A                 7    2   9   5   4    2    7     36
       B                                       7    9     16
       C                                  2    2           4
       D           11                                     11
       E                13    8       4                   25
       E                13    8       4                   25
     Totals        11   20   10   9   9   6   11   16     92