RESEARCH FOR MANAGEMENT OF THE ORNATE TROPICAL ROCK LOBSTER, PANULIRUS ORNATUS, FISHERY IN TORRES STRAIT:

Transcription

1 MARCH 1997 RESEARCH FOR MANAGEMENT OF THE ORNATE TROPICAL ROCK LOBSTER, PANULIRUS ORNATUS, FISHERY IN TORRES STRAIT: FINAL REPORT ON CSIRO RESEARCH, C.R. PITCHER, T.D. SKEWES, and D.M. DENNIS CSIRO DIVISION OF MARINE RESEARCH C S I R O AUSTRALIA MARINE RESEARCH

2 TORRES STRAIT TROPICAL ROCK LOBSTER i Research for Management of the Ornate Tropical Rock Lobster, Panulirus ornatus, Fishery in Torres Strait: Final Report on CSIRO Research, C.R. Pitcher T.D. Skewes D.M. Dennis CSIRO Division of Marine Research Marine Laboratories PO Box 12 Cleveland, Queensland 4163 This report summarises research carried out by CSIRO on the ornate rock lobster, Panulirus ornatus, fishery in Torres Strait during the triennium. The program investigated: the annual relative abundance of the rock lobster stock in Torres Strait and the strength of the recruiting year-class; the catch and effort of the Islander diver fishery; aspects of the biology of the breeding population in the far northern Great Barrier Reef; and made annual assessments of the potential yield.

3 TORRES STRAIT TROPICAL ROCK LOBSTER ii Table of Contents CSIRO Tropical Rock Lobster Research, Acknowledgements 1 Executive Summary 2 The fishery and past research 2 Annual population surveys 2 Catch monitoring 2 Local depletion/replenishment 3 Breeding grounds survey 3 Stock assessment 3 Future research priorities 4 Introduction 6 The fishery 6 Lobster life history 8 Summary of research 9 Objectives: Section 1: Annual population surveys 14 Introduction 14 Annual survey methods 14 Size and age composition 15 Recruitment to the fishery 16 Fishable stock 16 Stock forecasting 17 Spatial variability in abundance and size 18 Conclusions 2 Section 2: Catch monitoring 21 Introduction 21 Interannual CPUE comparisons 21 Comparison with annual abundance survey 22 Conclusions 23 Section 3: Local depletion/replenishment 24 Introduction 24 Study area 24 Methods 26

4 TORRES STRAIT TROPICAL ROCK LOBSTER iii Results 26 Discussion 31 Conclusions 32 Section 4: Breeding population studies 33 Introduction 33 ROV and diver survey 33 Survey results 34 Conclusions 36 Section 5: Stock assessment 37 Introduction 37 Previous Assessment 37 Developments 39 Data sources 39 Modelling methods 39 Modelling Results 41 Indicators of current stock status 42 Uncertainties in assessment 43 Implications for Management 44 Future data needs 44 Conclusions 45 Section 6: CSIRO Funded Research 46 Distribution of lobster larvae in the NW Coral Sea 46 Aquarium growth experiments 48 Future Research Priorities 52 High priority 52 Medium Priority 53 Low Priority 53 References and Publications from the Project 56

5 TORRES STRAIT TROPICAL ROCK LOBSTER 1 Acknowledgments This research was funded by the Commonwealth Department of Primary Industries and Energy, through the Australian Fisheries Management Authority as part of its Torres Strait Protected Zone fisheries research program. The Torres Strait Fisheries Scientific Advisory Committee was responsible for vetting the research proposal and reviewing progress.

6 TORRES STRAIT TROPICAL ROCK LOBSTER 2 Executive Summary The fishery and past research The commercial fishery for the Torres Strait ornate tropical rock lobster began in the late 196's and has become a major income earner for Torres Strait Islanders. Effort in the fishery has increased substantially since the 196's and the research that CSIRO has undertaken proved valuable to managers because it showed that, in the face of increasing effort and catch levels, recruitment is stable, indicating that the fishery is sustainable. Past research has provided basic information on the life history of the ornate tropical rock lobster, particularly the existence of an annual breeding migration out of Torres Strait to the north east as far as Yule Island at the eastern Gulf of Papua. The migration causes substantial changes in the Torres Strait population and fishery. Lobsters at the breeding grounds are in very poor physiological condition and by the end of a short breeding season virtually all lobsters die. Subsequently, however, other breeding grounds were discovered in Australian waters of eastern Torres Strait, but the extent and mortality rate of this population remains to be determined. More recently, research has been directed at assessment of the fishery for management. In 1989, the abundance of the lobster stock in Torres Strait was estimated, thus enabling the first estimates of the exploitation rate and potential yield. The stock has been surveyed annually since then and annual assessments have been done using an escapement model. Details of these and other research activities during are summarised below. Annual population surveys The annual population surveys have provided fishery independent information on the relative strengths of the recruiting and exploited year-classes. This information has proved valuable to managers because it shows that although mid-year abundance of 2+ lobsters has declined, midyear abundance of recruiting 1+ lobsters has been near or above the 1989 level; indicating that at current effort level the fishery is sustainable. The surveys also provided annual estimates of growth and mortality rates, which are incorporated into the fishery model for stock assessments, and the fishery independent abundance estimates allowed evaluation of commercial catch and effort data. The information on relative strengths of recruiting year-classes, combined with the 1989 absolute abundance estimate, has proved valuable in assessing stock status by allowing forewarning of future fishable stock sizes. These forecasts were, however, unreliable for the 1993 and 1995 fishable stocks and highlighted the spatial and temporal variability in natural mortality rates. The high natural mortality in the 1993 fishable stock was attributed mainly to loss of lobster habitat in the NW quadrant; subsequently identified as a seagrass dieback event. Using long-term average mortality, and given the below average 1996 recruiting year-class, it is expected that the 1997 fishable stock would also be below average. Catch monitoring The catch and effort and size of lobsters caught by the island-based fishery at Badu/Mabuiag Islands has been monitored during June each year from 1988 to This information proved valuable to managers because it showed that catch rates, in this vulnerable sector of the fishery, have been slightly lower, but stable, since 199. Catch per unit effort of the island-based fishery generally compared well with lobster abundance measures from the annual lobster surveys. The recent divergence between the two measures of abundance was likely due to high lobster

7 TORRES STRAIT TROPICAL ROCK LOBSTER 3 abundance in the north-west compared to levels in other regions of Torres Strait. Size distribution information proved useful as it showed that in years where 2+ abundance was very low, fishers maintained catch rates by taking greater proportions of 1+ lobsters. Hookah divers continued to have higher catch rates than free divers, but the gap between these methods is not widening, and therefore, it is unlikely hookah diving is having a deleterious effect on free diver catches. Local depletion/replenishment Replenishment rates to the depleted shallow reef habitat fished by local Islanders was investigated by conducting population surveys at the start (March) and at the end (August) of the main fishing season. We found that 1+ lobsters live only in inter-reefal areas, 2+ lobsters live in inter-reefal areas and on shallow reefs, and 3+ male lobsters live mainly on shallow reefs. Replenishment rates were very low to zero during the main fishing season (winter) and we conclude that recruitment occurs as a movement of 1+ lobsters onto the shallow reef after the annual breeding migration in September/October. Fishing in the deeper inter-reefal habitat adjacent to the shallow reef would therefore not have an immediate effect on the catch on the shallow reef, although it may reduce the number of potential recruits that are available to replenish the shallow reef after the breeding migration. Breeding grounds survey Previous research on the breeding population showed that there was a significant, though patchy, lobster population actively breeding in shallow and deep water in eastern Torres Strait. The size-frequency distribution of this population indicated that the lobsters may not suffer catastrophic mortality after breeding as occurs in the breeding population of the eastern Gulf of Papua. To assess the extent and persistence of the eastern Torres Strait breeding population the eastern breeding grounds were surveyed during February 1996, by divers on the shallow reefs and using a remotely operated vehicle (ROV) in deep water. Only 19 lobsters were counted by divers at 119 shallow reef sites. Further, the 1996 survey of the small relic reefs on the narrow shelf edge outside the far northern Great Barrier did not find a persistent breeding population in relatively high densities, as had been found previously. The explanation for the low number of breeding lobsters observed during the 1996 survey is unknown, but possibilities include: the destination of the breeding migration is variable among years; and/or the mortality of eastern breeding lobsters is severe, as it is at Yule Island; and/or environmental fluctuations stimulated the eastern breeding lobsters to move to other areas or habitats. Stock assessment To provide stock-assessment advice for management, data collected in the field continued to be integrated into a stock-assessment model of the fishery. Analyses of yield-per-recruit indicated that growth overfishing would only become an issue at levels of fishing mortality much higher than current levels and the minimum size of 1 mm tail length does not improve yields. Stockassessment was based on consideration of the level of escapement the proportion of the population that escapes fishing to emigrate and breed. Initial assessment showed the fishery was lightly exploited and a conservative target that would allow rational development was set. The target, on which potential yields were estimated, was fishing mortality (F) could be increased to F=.4, and yet leave a conservative escapement of about 74%. It was advised that increased

8 TORRES STRAIT TROPICAL ROCK LOBSTER 4 effort should be encouraged in the diver fishery. Since then, annual fishery independent surveys have provided data on stock abundance, recruitment, growth and mortality. The updated parameters enabled the model to be refined and now, based on the average level of recruitment, the average long term potential yield estimate is ~275 t, at F=.4. There are uncertainties in the assessment largely due to variability in growth and natural mortality rates and the impact of environmental disturbances. Estimates of other parameters (e.g. F) may be tuned by more refined analysis of data from existing sources and ongoing monitoring. The recruitment data for one year have been used to forecast the potential yield (at F=.4) for the following year. The lower level of recruitment experienced in the 199s may be more typical, and effort and fishing mortality are increasing and approaching the uncertainty range for the assessments. The implications for management are that new management measures may soon need to be considered to address the uncapped effort potential in the fishery. Future research priorities The data and research needs for the Torres Strait lobster fishery and stock were discussed during a series of priority setting meetings of the Torres Strait Fisheries Scientific Advisory Committee (TSFSAC) during October 95 January 96. The high priority data and research needs for lobster, according to the TSFSAC outcomes, were: Commercial catch data. Commercial catch data (from processor landings) provides the most basic information about the status of the fishery, showing monthly and annual trends in the lobster catch. The data are simple and inexpensive to obtain and data collection should continue. Freezer boat CPUE logbooks. Although not all freezer vessel participate in the voluntary logbook program, the data have been collected for 1 years and over this period, surveys have shown the stock size to vary considerably; thus there should be significant contrast in the data from year to year. The PNG catch information should also be collated and included in assessments of the Torres Strait lobster stock. Islander CPUE analysis. There are differences between the Islander based sector and the freezer boat sector of the fishery. The Islander based sector comprises a significant free-diving component and it is susceptible to local depletion as it lacks the mobility of the freezer vessel fleet. The Islander CPUE data collected by CSIRO is important because it is collected by independent observers with a high resolution, and concurrent length frequency data is also collected. This enables calculation of separate CPUE measures for the recruiting (1+) and fished (2+) components of the catch and measures of proportional fishing mortality by size-class which are important parameters for stock assessment. Annual lobster population surveys Annual surveys of the lobster population provide fishery independent indices of the relative abundance of the recruiting (1+) and fished (2+) year-classes. These indices are compared with abundances measured during annual Islander CPUE monitoring programs. The population surveys also provide estimates of growth, mortality rate and exploitation rate in each year. Also, as lobsters are surveyed throughout all areas of the fishery, spatial differences in these parameters are interpreted. Annual population surveys currently provide the underlying data and are required for annual stock assessments.

9 TORRES STRAIT TROPICAL ROCK LOBSTER 5 Stock assessment & modelling Annual stock assessments and status reporting are required for the Torres Strait lobster fishery. The present computer model, on which assessments are based, should be further developed to include spatial and temporal patterns in the fishery and stock, the AFMA logbook data; fishery data from the PNG side of Torres Strait; spatial and temporal effort, CPUE and length frequency data. Several other variables that should be programmed include variability in settlement timing, growth & mortality rates; distribution of survey sites relative to variability in lobster movement/distribution; variability in the size-selection (recruitment) curve dependent on relative abundance of year-classes, and differences in fishing behaviour between Islander freedivers and commercial hookah divers (requires size-distribution data from freezer vessel divers). Currently, the spatial distribution of effort is known only anecdotally this needs to be documented quantitatively for a spatial modelling approach.

10 TORRES STRAIT TROPICAL ROCK LOBSTER 6 Introduction The fishery In Australia, most commercial fishing for the ornate tropical rock lobster, Panulirus ornatus (Fabricius), occurs in Torres Strait, with some activity along the far north-east coast of Queensland (Fig. I-1). Historically, lobsters have been fished by the traditional inhabitants of Torres Strait probably for several centuries before commercial fishing began in the late 196s. These lobsters will not enter pots, so they are speared by divers fishing from small dinghies. Initially, there were about 3-4 small (~1 m) freezer vessels that processed catch from a total of about 1 dinghies. The fishery gradually developed through the 197s and during the 198s there were about 15 small freezer boats active throughout the area and each processed catch from 1-6 dinghies (Channells et al 1987). Independent Islander involvement in the fishery also increased during the 198s and it became a major source of income for Torres Strait Islanders. The annual catch of lobster tails from Torres Strait in the 197s ranged from 68 to 124 tonnes (t). Catches during the 198s were variable, averaging almost 2 t, with a peak catch of 35 t in 1986 (Channells et al 1987; AFMA unpubl. data). In the early 199s, catches have been similarly variable, averaging about 2 t. The first Australian management measure was introduced in 1981 a ban on daytime trawling for migrating lobsters. Subsequently, a total ban on trawling for lobsters was legislated in Since then, the fishery has been managed under Article 22 of the Torres Strait Treaty ratified between Australia and Papua New Guinea (PNG) in 1985 (Haines 1986). The Treaty established the Torres Strait Protected Zone, in which the traditional way of life and livelihood of the inhabitants is to be protected. The main management measure for ornate rock lobsters is strict entry criteria intended to prevent any increase in non-islander involvement in the fishery and encourage Islander participation and restricting participation by non-islanders (Channells et al 1987). The treaty also provides for catch-sharing arrangements between Australia and PNG and for this purpose, the allowable catch of the fishery has, since 1987, been set as the catch of the diver fishery. Currently, divers work from 4-6 m dinghies powered with ~4 hp outboard motors, and either use hookah compressors or free dive. Between 3-4 dinghies operate from island communities and most divers return their catch to island based processors or semi-permanently moored mother vessels. There are also about 2 dozen licenses for small fishing vessels (~8-18 m), although only about half are active regularly throughout the area, each freezing the catch from 1-6 dinghies. Free divers generally take lobsters from coral outcrops in waters 1-4 m deep while hookah divers generally take lobsters from rocky holes and crevices in the deeper grounds (4-15 m) between reefs. Although lobsters are found on most reefs in Torres Strait and the north-eastern Queensland coast, the principal fishing grounds are around Thursday Island, Orman and Warrior Reefs in Torres Strait. Divers report that it is now necessary to fish additional grounds in deeper waters to maintain catch rates at levels experienced over past years. Fishing occurs throughout the year, with lower activity during October-December and peak catches during March-August. Most boats operate during neap tides when currents are slower and the water is clearer.

11 TORRES STRAIT TROPICAL ROCK LOBSTER PNG 15 E Port Moresby Torres Strait 1 Coral Sea 15 S Australia Fig. I-1. Map of Torres Strait and Gulf of Papua showing the main fishing grounds around the central and western reefs and islands (///), the migration pathways (broad arrows), breeding grounds ( ) and current patterns (arrows) in the Coral Sea. Fishing effort (in terms of number of participants, boats, days worked per year, or hours worked per day) has increased substantially since the fishery began, though total effort is unknown. At the same time, the catch per hour has decreased to roughly one-third of that 15-2 years ago hence, there is a continuing need for quality assessment and monitoring of the lobster stocks to provide advice to managers. Recent management measures include: a minimum size limit of 1 mm tail length, introduced in 1988; the catch sharing arrangements required under the treaty with PNG were implemented in 199 with a specified number of PNG dinghies permitted to fish in the Australian side of the protected zone; a two month ban (October/November) on hookah gear was introduced in 1993 in response to Islander concerns of over-fishing in traditional reef fishing areas; the TAC of the fishery is still set as the catch of the diver fishery. There have been no formal studies of the economic status of the fishery; however, the average annual catch of the fishery is about 2 t tail weight and prices paid to fishers vary between $15-$35 per kg. Thus the first-purchase value of the fishery is ~A$5 million and it is a major source of income for Torres Strait Islanders. The lobsters are sold on both domestic and overseas (mainly USA) markets. Individual incomes would be highly variable, nevertheless, daily catch rates mostly range between 2-4 kg per diver/dinghy and professional divers may work about 9 days per year. Costs are relatively low: capital input for a dinghy based fisher is about $6K for an aluminium dinghy, outboard motor and basic dive gear; hookah equipment would cost an additional ~$2.5K (the replacement periods would be about 2-3 years) operating costs comprise mainly petrol, oil and maintenance totalling perhaps $4-$8 per day; dinghy driver's are often paid a percentage of the catch. Costs of operating freezer vessels would be substantially higher: capital investment could range between $3K to over $1K with daily running costs being similarly variable most freezer vessels are supported by several dinghies which would help cover the increased running costs.

12 TORRES STRAIT TROPICAL ROCK LOBSTER 8 Lobster life history Past research has provided fundamental information on the life history of the ornate rock lobster Panulirus ornatus. Hatching and release of larvae (phyllosomes) occurs between November and March each year. The planktonic phyllosomes develop and are transported via ocean currents in the Coral Sea during the ensueing six months. The final planktonic stages (pueruli) settle during winter, into small holes in the seabed in the nursery grounds of central and western Torres Strait and grow rapidly, recruiting into the fishery about one year later at ~1 mm tail length. These juvenile and sub-adult lobsters are fished until they are just over 2½ years old. At this time, in spring each year, most emigrate from Torres Strait to breed and catch rates decline markedly. Tagging studies have shown that some of the emigrating lobsters moved north east into the Gulf of Papua (Fig. I-1), undergoing reproductive development at the same time (Moore & MacFarlane 1984). The tagging studies also showed that lobsters on reefs off the north east Queensland coast do not participate in the migration across the Gulf of Papua but in general tended to move to the south east there is little or no movement of individuals from the northeast coast of Queensland into Torres Strait (Bell et al 1987). Nevertheless, the populations in the two areas are indistinguishable genetically (Salini et al 1986). It is highly likely that rock lobsters from the east coast of Queensland come from the same breeding stock as the resource in Torres Strait, given the genetic similarities of the populations and the ocean current patterns in the north-west Coral Sea (Fig. I-1) that could disperse larvae to both areas from breeding grounds off the southern coast of PNG and northern Great Barrier Reef (Pitcher et al 1992b). Prawn trawlers used to target the migrating lobsters and catches up to ~2 t were recorded, but this activity was banned in 1984 (Williams 1986). Some of the migrating lobsters move as far as the coastal reefs of the eastern Gulf of Papua where there is a breeding ground (MacFarlane & Moore 1986). This breeding population forms the basis of a seasonal artisinal fishery around Yule Island which lasts only a few months during the summer. This fishery existed in traditional form prior to written history but the origin of the lobsters became known only in the early 198s. The lobsters on these Papuan coastal reefs are in very poor physiological condition (Trendall & Prescott 1989) the muscles are wasted and the blood is very watery and virtually all lobsters die after the breeding season and this is why the Papuan artisinal fishery lasts only 2 4 months (Dennis et al 1992). Such catastrophic mortality is very unusual for lobsters, as most species can live and breed for more than 1 years. Until recently, the coastal reefs of the eastern Gulf of Papua were the only significant known breeding grounds. Yet perhaps only a minority of the several million lobsters which emigrate from Torres Strait actually migrate across the Gulf of Papua each year and only a fraction of those arrive at Yule Island. It was suspected that the remainder migrate to other, largely unknown, breeding grounds and in the 1989/9 summer, a small research submarine was used to survey the Gulf of Papua and far northern Great Barrier Reef for other lobster breeding grounds in waters as deep as 4 m (Prescott & Pitcher 1991). Breeding lobsters were seen at several sites off the far northern Great Barrier Reef but very few were found in the Gulf of Papua. The far northern Great Barrier Reef may support a significant breeding population but this has yet to be confirmed; further, it is not known whether the mortality rate of these lobsters is catastrophic, as occurs at Yule Island.

13 TORRES STRAIT TROPICAL ROCK LOBSTER 9 Summary of research The previous two triennium's research on the ornate tropical rock lobster began, in mid 1987, with a basic understanding of stock structure and movement patterns, particularly the existence of an annual breeding migration north east out of Torres Strait into the Gulf of Papua as far east as breeding grounds around Yule Island. Lobsters at these breeding grounds were in very poor physiological condition and by the end of the short breeding season they had all disappeared. Research during the triennium was directed at determining: the causes of this disappearance; the ocean current patterns that may provide recruitment to Torres Strait; the existence of settlement grounds in far western Torres Strait; the impact of the migration and fishing activities on the Torres Strait stock; and whether other breeding populations existed in the Gulf of Papua or northern Great Barrier Reef. Also, attention was turned to stock assessment for management, primarily by making an estimate of the actual abundance of lobsters in Torres Strait and initiating annual fishery independent surveys of the stock. These activities were detailed in Pitcher et al (1992, 1994) but are summarised below. Western settlement grounds The extensive seagrass beds of far western Torres Strait were surveyed in October 1987 for the occurrence of juvenile lobster nursery grounds that could supply recruits to the fishery farther east. Only 18 lobsters were seen at a rate of.35 lobsters per kilometre of bottom surveyed and none of these were newly settled juveniles. The survey indicated that the area was not a significant settlement ground for lobsters nor did it appear to support a significant adult population. Other work showed that the major settlement occurs in central Torres Strait in the same area as the fishing activity. Breeding lobster mortality The cause of the annual decline of the breeding lobster population on the coastal reefs near Yule Island, PNG, was investigated in early 1989, using several complementary methods. The Yule Island fishery followed a typical pattern of punctuated rise and then a rapid decline that was not associated with any observed movement of lobsters off the coastal reefs into deeper water. Tag returns indicated that fishing pressure was responsible for >3% of the decline, but the natural mortality rate was even higher (>1-fold greater than in Torres Strait), probably due to stress of migrating and breeding. The important implication was that if all breeding populations suffer similar catastrophic mortality then settlement into the fishing grounds each year may depend entirely on the breeding success of the preceding years emigration which should therefore be conserved. Sources of recruitment A series of satellite tracked buoys were released in deep water off Yule Island to test whether lobster larvae hatched at Yule Island could potentially be transported to Torres Strait. The buoys moved rapidly eastward to the end of the PNG mainland and then turned south and later south west toward the north Queensland coast. This pattern of movement indicated the presence of a clockwise gyre in the northern Coral Sea which has the potential to carry larval lobsters from PNG waters back to Torres Strait. Population changes due to migration Marked changes occur in the lobster population around August each year as a result of the annual breeding emigration. The movement of larger lobsters out of the fishery begins between early August and early September and lasts for 4 to 8 weeks; after which mostly only smaller lobsters remain in Torres Strait. More female than male lobsters emigrate, causing the sex ratio to be biased towards males after the emigration. Lobsters undergo a maturation moult before

14 TORRES STRAIT TROPICAL ROCK LOBSTER 1 they emigrate; this was evident as one or more large peaks in moulting activity that coincided with the third lunar quarter. Size-structure and growth The growth rate of the ornate lobster have been investigated by a variety of methods. The growth curve was re-estimated from tag-recapture, aquarium growth, settlement and hatching data. Size distributions from the Torres Strait population, the catch and the emigratory population showed that the population in Torres Strait comprised mainly 1+ and 2+ lobsters, that the catch consisted mostly of 2+ lobsters with smaller and variable numbers of 1+ lobsters, and that the emigratory population consisted mainly of 2+ lobsters with some 3+ males. The ornate rock lobster in Torres Strait is very fast growing, perhaps the fastest growing of any palinurid studied. Growth varies on both spatial and temporal scales: spatial variability may the result of environmental factors such as temperature and/or food availability; temporal variability may also be a result of environmental factors and/or density dependant effects. The fishery exploits primarily the 2+year-class for less than 1 year, thus the catch was susceptible to variations in the modal size of this year-class as well as its abundance. Survey of lobster abundance The lobster stock in a 25, km² area of Torres Strait was estimated in June 1989, by making visual counts of the number of lobsters in strip transects, after pilot studies in 1988 confirmed the feasibility of a full-scale survey. The main survey took seven weeks and analysis of the transect data provided an estimate of abundance between million lobsters. The surveyed population was sampled concurrently to determine its size structure and provided an estimate of the stock size for the fishery between 2,2-3,35 t tail weight, which was roughly 1-fold greater than the 1989 catch of about 24 t. This indicated that the level of exploitation was low and that the fishery may even support greater effort. The survey also provided information on the benthic habitat of central and western Torres Strait. Catch monitoring Catch and effort of Torres Strait lobster fishermen were monitored as part of assessing the impact of the fishery on the stock. The CPUE of the Islander catch reflected changes in lobster abundance, as expected from the results of the annual relative abundance surveys. Nevertheless, it appeared that catchability may increase when lobster abundance is low, thus countering decreases in CPUE to some extent. The difference in catch rate of free-divers and hookah-divers reduced over the period of research and there is no evidence that continued use of hookah was detrimentally affecting the catch-rate of free divers. The magnitude and inter-annual variation of the CPUE measures from the Island based fishery and the freezer boat sectors were different and there were indications that Islander CPUE may reflect changes in abundance more closely than freezer boat CPUE. Annual population surveys The annual population surveys have provided fishery independent information on the relative strengths of the exploited and pre-recruit year-classes that is invaluable for assessing the status of the stock. They also provided annual estimates of growth and mortality rates and evaluation of commercial catch and effort data. The surveys have also forewarned of future fishable stock sizes, and provided insight into the causes of variability in natural mortality for example, the possible impact of the environmental changes on settlement grounds during , which appeared to cause low recruitment and high mortalities.

15 TORRES STRAIT TROPICAL ROCK LOBSTER 11 Puerulus collectors Puerulus collectors were trialed as a method of providing annual settlement indices, forecasting recruitment into the fishery, detecting of the effect of any significant changes in the fishery or disturbances in the environment and estimating compensatory processes occurring in the first year of benthic life. Unfortunately the trials were not particularly successful; as a consequence, expansion of the sub-project to full-scale sampling was not undertaken. It is possible that the poor catches of pueruli in the collectors reflected very low numbers of pueruli in the water column and/or that collectors could not compete with abundant natural settlement habitat. It is also possible that the large distance between the settlement grounds and the larval retention area in the north western Coral Sea massive depletion of pueruli from the water column as they move west from the edge of the Great Barrier Reef to the central and western Torres Strait. Juvenile ecology Small juvenile lobsters were found throughout central Torres Strait, mainly coinciding with the broad range of habitats occupied by adults with no apparent depth preferences; most juveniles were found between 7-9 m. The juveniles grew an average 6 mm CL per month in the first year of post-settlement life which is the fastest growth rate documented for a palinurid lobster. Postpueruli and juveniles sheltered in holes in rock substrata that matched closely their body diameter and length. There was a shift in habitat with growth; juveniles <4 mm CL preferred holes in rock covered by macro-algae (eg. Sargassum sp.) whereas larger juveniles preferred bare holes and crevices. The pueruli settling into Torres Strait measured ~6 mm carapace length and settlement occurred during winter, after an oceanic larval phase of about 4-6 months. Breeding grounds surveys The Gulf of Papua and far northern Great Barrier Reef were surveyed for the existence of other lobster breeding grounds in depths from 3 to 2 m using a small research submarine. High densities were seen on a few deep (3 1 m) reef habitats on the edge of the shelf of the far northern GBR; but very few lobsters were found in the Gulf of Papua or in deep water adjacent to the coastal reefs of the eastern Gulf. It was thought possible that the far northern GBR supported an important breeding population, but it was unknown whether these lobsters suffer catastrophic mortality after breeding as occurs in the breeding population at Yule Island. Two pilot surveys were conducted to develop methods for using a remote vehicle and divers to research the breeding population. These showed that there was a significant, though patchy, lobster population actively breeding in shallow water in eastern Torres Strait. The sizefrequency distribution of this population hinted that these lobsters may not suffer catastrophic mortality after breeding as occurs in the breeding population of the eastern Gulf of Papua. The shelf edge outside the far northern Great Barrier Reef was narrow with small relic reefs but these appeared to support a persistent breeding population in relatively high densities. Methods were successfully developed for a full-scale survey to confirm the distribution, abundance and mortality rate of the breeding population in the far northern Great Barrier. Stock assessment To provide stock-assessment advice for management, data collected in the field continued to be integrated into a stock-assessment model of the fishery. The first assessment, in 199 included an analysis of yield-per-recruit that indicated that the minimum size would only improve yield at levels of fishing mortality much higher than current levels. The model was extended to consider the proportion of the population that escapes fishing to emigrate and breed and, based on 1989 data, it was estimated that fishing mortality (F) could be increased 4-fold (to F=.4), giving an average potential yield estimate of >8 t, and yet leave a conservative escapement of about 74%. It was advised that increased effort should be encouraged in the diver fishery. Since

16 TORRES STRAIT TROPICAL ROCK LOBSTER 12 then, annual fishery independent surveys provided updated parameters that enabled the model to be refined and, based on average levels of recruitment up to 1994, the potential long term average yield was estimated at ~39 t. The recruitment data for one year can be used to forecast the potential yield (at F=.4) for the following year. There were uncertainties in the assessment due to variability in parameters but these, and a variety of other uncertainties, may be reduced by continued data collection. Objectives: In the triennium , research continued to focus on the measurement of parameters such as lobster abundance, fishing and natural mortality, recruitment, and the breeding grounds. This information is essential for sustainable development of the fishery through sound stock assessment. The overall objectives for the project during the triennium were to accurately quantify fishery dynamics parameters for input into models that provide potential catch levels as output, and develop monitoring systems to assess the impact of implementation of the catch levels. Such an empirical approach with a feedback loop was necessary as many variables were and are still unknown. The specific objectives for the period July 1993 to June 1996 were: Annual population surveys to obtain an unbiased annual index of the relative abundance of all year-classes of lobsters (including 1+) in Torres Strait this will provide information on the relative abundance of the stock and recruits, growth and mortality rates, for stock status assessments. Islander catch monitoring to obtain length-frequency distributions and underlying catch-effort data from the fishery this, in conjunction with the annual population surveys, would provide information on total mortality, growth, fishing pressure and size selection (bias) by fishermen, as well as CPUE of Islander fishermen which is not recorded by logbooks. Local depletion to determine the rate of replenishment of lobsters from deep water to a depleted shallow water population this objective originated from concerns raised by some Island community representatives that commercial lobster fishing by hookah divers in deeper water adjacent to traditional "home-reefs" would deplete replenishment to areas fished by Islander free-divers. Breeding population studies to survey the size and extent of the breeding lobster population in eastern Torres Strait other important questions included estimation of the mortality rate of the breeding lobsters. Stock assessment to provide outputs for the management of the lobster fishery based on input fisheries dynamics parameters obtained from field research sub-projects. Detailed models enable assessment of the fishery's sensitivity to parameter estimates and can provide possible outcomes of different management strategies (eg. changing level of effort, minimum size etc). This report is organised into sections that correspond to each of the above objectives. Also reported are CSIRO funded objectives including, the distribution of lobster larvae in the northwest Coral Sea and aquarium experiments on factors affecting lobster growth.

17 TORRES STRAIT TROPICAL ROCK LOBSTER 13

18 TORRES STRAIT TROPICAL ROCK LOBSTER 14 Section 1: Annual population surveys Introduction In May/June 1989, CSIRO carried out a major survey of the ornate tropical rock lobster Panulirus ornatus stock in central and western Torres Strait (Pitcher et al., 1992a). Data from this survey, which involved quantitative sampling of measured (4 5 m) transects at 572 sites, provided information on the absolute abundance of lobsters and the relative proportions of commercially fished and pre-recruit year-classes. Subsequent to the major stock survey in 1989, CSIRO has undertaken annual surveys of relative abundance at the 1 of the original 6 sites that accounted for 85% of the lobsters observed at the time. These surveys provided an annual index of the relative abundance of the two year-classes (1+ and 2+) in the Torres Strait population, including the strength of the recruiting year-class (1+), and updated estimates of growth and mortality, as a basis for annual stock assessments. The 1989 survey data was adjusted to match subsequent surveys due to the necessary change in methods between the two types of surveys. The annual survey data is also compared with data from the Islander-catch monitoring sub-project (see Section 2) to allow assessment of changes in commercial catch corresponding with changes in abundance of different year classes. Note that the 1996 survey does not actually fall within the reporting period, but is included in any case. Annual survey methods Of the original 572 sites surveyed in 1989, a sub-set of 1 sites were chosen with highest density of lobsters in Australian waters of Torres Strait. These sites were expected to provide the highest precision of inter-annual relative abundance estimates; there was little cost-benefit in continuing to sample low/zero density sites. The sub-set of 1 sites accounted for 86% of lobsters observed in Australian waters during the 1989 survey and we assumed that the sizedistribution of lobsters sampled in all sites did not differ from that of the sub-set. The distribution of the sub-set of 1 sites is shown in Fig The annual surveys were conducted in June/July each year, during the middle of the fishing season when the 1+ cohort were large enough to be conspicuous, but before they recruited into the legal-sized stock. The durations of the surveys were about 2-3 weeks, with a single team of three divers operating from a dinghy. The previously determined starting points of each transect were accurately re-located each year with GPS-navigation; the coordinates have remained unchanged since Paired divers recorded the number and sampled lobsters, during unidirected (with the current) 2 minute bottom-time dives. Consequently, survey effort was constant among years. Any lobsters not successfully sampled were noted for size (1+ or 2+). The distances surveyed during each 2 minute dive were recorded using GPS. The tail-width, sex and moult stage of each lobster were measured and recorded. The inter-reefal benthic habitat at each site (substratum type, seagrass and epibiota, and other inter-reefal resources including pearl shell) was recorded using a standardised description (Pitcher et al. 1992b) to examine relationships between lobsters and habitat, and to monitor any changes in habitat that may impact the lobster population.

19 TORRES STRAIT TROPICAL ROCK LOBSTER 15 P.N.G E Parama Reef Warrior Reefs Buru Is 9 3 Mabuiag Is Badu Is 1 S Commercial catch monitoring site Thursday Is Torres Strait 1 3 Australia Annual Survey Sites Fig Map of Torres Strait showing sites (2 dives are done at each site) surveyed between and the arbitrary division of the study area into 4 quadrants to allow assessment of spatial differences in abundance and size of lobsters. The location of the land-based processor where commercial catch and effort is monitored is also shown. In 1996, the number of sites was reduced to 82 because of reduced funding the sites omitted accounted for ~7% of the lobsters observed during previous surveys. For the purpose of comparison in this report, the 1996 total counts were re-scaled to estimate counts/1 sites based on regression between counts for the 82 and 1 sites for the years (r²=.991). Size and age composition The tail-width size-distribution data were analysed with Mix (MacDonald and Pitcher 1979) to estimate the proportion and mean size of the recruiting (1+) and fished (2+) year-classes. When these proportions were corrected by the noted miss-bias, they provided a standardised index of the relative abundance of the 1+ and 2+ year-class that could be compared among years and contrasted with measures of fishery catch per unit effort. The relative abundance data were also used to estimate growth and mortality (by comparing 2+ abundance in any year with 1+ abundance in the previous year) parameters for input into fishery dynamics models to provide assessments of the stock. The surveyed population comprised mainly two size-modes in all years, representing 1+ and 2+ year-class lobsters, with a very small 3+ year-class component (Fig. 1-2). The average size of the 1+ year-class lobsters was estimated at ~4 mm tail width, while 2+ year-class lobsters averaged ~7 mm tail width. The commercial catches were also comprised of these age classes, however, the 1+ year-class components were truncated due to the imposed minimum size limit of 1 mm tail length ( 52 mm tail width, Fig. 1-2). The 1+ year-class is the prerecruit/recruiting year-class and the 2+ year-class is the exploited year-class or the "fishable stock". In all years, except 1989, the 1+ year-class was relatively more abundant than the 2+ year-class.

20 TORRES STRAIT TROPICAL ROCK LOBSTER 16 Recruitment to the fishery The relative abundance of the recruiting year-class (1+ year olds) increased between 1989 and 1992 (Fig. 1-3), but was variable thereafter. The 1991, 92 and 93 recruiting year-classes were above average and the 1993, 95 and 96 recruiting year-classes were below average. Larger recruitments probably occurred before the surveys began, i.e. the 1985 and 1988 recruiting yearclasses may have been the largest and second largest, as they gave rise to the largest and second largest catches. The 1993 recruiting year-class was the smallest since the surveys began (Fig. 1-3); only ~55% of that in 1992, which may have been a result of major habitat changes in the nursery grounds (see below) Catch 9 Survey Number of lobsters/1 hours Tail Width (mm) Fig The size-frequency distributions of the islander catch landed at Mabuiag and/or Badu Islands in June from 1989 to 1996 and the size-frequency distribution of the Torres Strait lobster population surveyed by research divers in June/July in the same years. The Islander catch and survey population histograms are scaled by CPUE of the islander fishery and survey abundance respectively. Number of lobsters Fishable stock The fishable stock, in any one year, comprised mostly the 2+ year-class and a proportion of the 1+ year-class that was greater than the minimum legal size of 1 mm tail length ( 52 mm tail width). The relative abundance of 2+ year-class lobsters in 199 and 1992 were similarly above average, but less than half as abundant as in The 1991 and lobsters were in similarly below average abundance, about only half that in 199 and 1992 (Fig. 1-3). The year-class was the smallest since the surveys began (Fig. 1-3). These inter-annual variations in relative abundance of the 2+ year-class are reflected in interannual variations in CPUE of 2+ lobsters in the Islander catch the relationship between these

21 TORRES STRAIT TROPICAL ROCK LOBSTER 17 two indices of 2+ abundance is relatively close (see Fig. 2-4, Section 2). Total annual catches are only loosely related to inter-annual variations in 2+ abundance (Section 5), presumably because total effort, and fishing mortality, also vary inter-annually. Relative lobster abundance years old 1+ years old Year Fig Relative abundance of 1+ year-class and 2+ year-class lobsters in central and western Torres Strait, estimated from annual surveys between 1989 and Stock forecasting It was anticipated that the 2+ stock in a future year could be forecast from the relative abundance of the previous year's recruiting year-class, by applying mortality rates estimated by tracking the relative abundance of 1+ in one year to the 2+ in the following year, given some consistency in mortality rates among years. Noting the change in survey methods between 1989 and 199, the data from subsequent surveys (199 to 1996) should be the most reliable for estimating mortality and making forecasts. However, the total mortality appeared to vary considerably among these years and possible causes of this variablity are discussed below. Overall, the ratio of 1+:2+ in the 199 to 1996 surveys is 1:.3, which corresponds to an average survival of only ~3% and indicates a total mortality rate of Z 1.19 (Section 5). By applying this average mortality rate to the below average 1996 recruiting year-class (Fig. 1-3), it is expected that the 1997 fishable stock would also be below average and similar to the 1996 level. Relative lobster abundance or catch/potential yield (t) Potential yield Catch Year Fig Relationship between relative abundance of 2+ year-class lobsters ( ) in year (x) and relative abundance of 1+ year-class lobsters ( ) in the previous year (x-1) estimated from annual surveys between 1989 and 1997 and annual catch and estimated potential yield for the same years.

22 TORRES STRAIT TROPICAL ROCK LOBSTER 18 The sizes of fishable stocks were not correlated (r 2 =.23,p>.1) with sizes of the recruiting stocks in respective previous years (Fig. 1-4). However, this result was mainly due to data from two years, 1993 and In these years the fishable stocks, which resulted from large recruiting year-classes (Fig. 1-4), were much smaller than forecasted. However, these discrepencies were not due to survey methods, since in both of these years the estimated islander CPUE for 2+ lobsters corroborated the low 2+ abundance (Fig. 2-4). In 1993, the low 2+ abundance was likely due to low survival of recruits as a result of loss of lobster habitat and food in the NW quadrant (caused by a major seagrass dieback, reported at TSFSAC #2), and 2+ abundance in the NW quadrant was the lowest ever recorded (Fig. 1-5). Anecdotal evidence showed that in response, the freezer vessel fleet directed more effort into the NE quadrant of the fishery to maintain catch rates and catch in this year was higher than in 1991,92. In 1995, the low 2+ abundance was mainly due to low survival of recruits in the SW quadrant of the fishery (Fig. 1-5). In fact, 1+ abundance in the SW quadrant in 1994 was the highest ever recorded in any quadrant and subsequent low survival may have been due to density dependant effects. However, as in 1993 the low survival of this recruiting year class was likely due to deterioration of lobster habitat. This was highlighted by the fact that in 1995 the abundance of 1+ lobsters in the SW quadrant was the lowest ever recorded. Because of these changes in the rate of natural mortality between years (as a result of density dependant effects or environmentally induced density independent effects) or changes in the distribution of lobsters between years (as a result of habitat modification or shelter or food distribution), the relationship between the relative abundance of 1+ in one year to the 2+ in the following year (Fig. 1-4) was not as close as expected. These changes also affected the relationship between potential yield and actual catch in some years. Further, because of unquantifiable changes in total effort and changes in distribution of effort (particularly for the mobile freezer vessel fleet) the relationship between fishable stock (2+ abundance) or potential yield and total catch was not consistent among years. It is also possible that changes in the way fishers exploit the 1+ and 2+ year classes in different years (e.g. Fig. 2-2, Section 2, which shows that 1+ lobsters have become a significant part of the total catch in recent years) could alter the relationship between 1+ abundance in one year and 2+ abundance in the subsequent year. In summary, large fishable stocks forecasted for the years 1993 and 1995 were not realised due to spatial differences in mortality, likely as a result of deterioration in lobster habitats. Stock forecasts in other years between 199 and 1996 were more reliable. Annual catches, between 1989 and 1996, were not well correlated with size of the fishable stock, likely as a result of increased total effort in recent years and changes in the distribution of effort by fishers to maintain catch rates. Commercial catches in recent years were only 1-15% smaller than the 1989 catch, whereas the fishable stocks in these years were 8-85% smaller than in As a result, CPUE of fishers; particularly in the mobile freezer vessel sector, may not be a reliable measure of overall lobster abundance. This highlights the need for fishery independent surveys and the value of information from these surveys for management Spatial variability in abundance and size The relative abundances, among 3 quadrants of Torres Strait, of 1+ and 2+ lobsters were comparable, and the inter-annual trends were similar though differing in detail (Fig. 1-5). This variability may be caused by a combination of patchy delivery of settling pueruli, habitat