PART 2: TECHNICAL

7. SUMMARY

It is necessary to define management objectives so that if these objectives are met, the CITES goal will automatically follow.

The CITES objective will be to control exports so that the stock remains in a state sufficiently close to its unexploited state so that it will not become endangered. Whether this objective is met, should be verifiable through various government reports and export monitoring.

CITES objectives should be met through applying fisheries management objectives.

Fisheries management objectives should be to:

• maintain the conch population above the overfished level;

• maintain fishing effort below level that would lead to overfishing;

• maintain the reproductive capacity of the stock;

• maintain socio-economic returns as close to the theoretical optimum as possible.

Note that these objectives are not limited to exports, but may encompass wider controls and concerns of which CITES objectives are a part.

These objectives are met by achieving management and scientific outputs. Management is required to set and enforce the controls that change the state of the fishery so that the:

• current stock state is above its lower limit;

• current mortality rate is below its upper limit;

• reproductive state is above its limit;

• socio-economic returns are closer to the optimum.

The outputs from the scientific authority include indicators and reference points for the current stock state, fishing mortality, reproductive capacity and socioeconomic performance. These are use as feedback to the management controls.

This manual provides advice on how management and scientific authorities might achieve these outputs. This is largely through implementing a data collection program and then basing advice upon the data collected.

Data are converted to simplified indicators and reference points which measure the state of the fishery and the stock. While difficult to estimate statistically, chosen indicators should be simple to understand and capture all relevant important information about the fishery.

Definitions of such states as "overfished" should be agreed based on indicator values. For example, if the catch per boat day is an indicator, overfishing could be defined as a point when this indicator is too low. Management actions already agreed among stakeholders should be associated with all such points. In this example, a rebuilding plan would automatically be applied when the indicator showed that the stock is overfished.

It is important that the decision-making is transparent and based upon objective assessments of the fishery. It will be necessary for all stakeholders to identify the current state of the fishery and verify that agreed measures have been implemented.

8. SPECIFIC RECOMMENDATIONS

The management plan should include documentation of:

• policy objectives and the resulting principles on which decisions should be based;

• the indicators and reference points that will be used;

• controls which will be applied to alter the stock or fishery as shown by the indicators;

• plan of actions when indicators pass reference points;

• technical information on the assessments conducted and the default operational model on which management decisions are taken.

8.1 Management objectives

8.1.1 Stock state

The state of the stock should be maintained above the overfishing point. Standardized catch-per-unit-effort (CPUE) as a proportion of the estimated unexploited CPUE should be used as the biomass indicator. There are two reference points:

• fifty percent unexploited CPUE precautionary: If the indicator falls below this point a rebuilding programme is implemented (CPUE should be monitored during rebuilding or a set recovery period applied);

• thirty percent unexploited CPUE limit: If a fishery falls below this point, the fishery should close;

8.1.2 Exploitation rate

Maintain the current fishing mortality and potential fishing mortality below that level which could drive the fishery into an overfished state.

• FMSY limit: Effort should be reduced to be below this level at which yield is maximized. FMSY (effort at the maximum sustainable yield) is the fishing effort which theoretically maximizes the total catch from the fishery.

• f_TEY (effort at the target economic yield) target: This is the fishing effort required to move the stock to some optimum level which can be defined using socio-economic indicators. The f_TEY should be adjusted to approach the target level as quickly as possible. A socio-economic target should be developed immediately and can be based on average costs, conch price and/or on fisher interviews. It can be refined through further socio-economic research.

It cannot be emphasized enough that most other controls are secondary compared to controlling fishing mortality and fishing capacity (including processing capacity). If fishing mortality is controlled to the appropriate level, other requirements are very likely to be met. Fleet capacity should be reduced to be commensurate with FMSY.

8.1.3 Size composition

A simple empirical indicator should be applied using observed catch size composition and mean meat weight to assess whether growth overfishing is occurring and whether the spawning stock (i.e the mature fraction of the total stock) is large enough (see Appendix II).

8.2 Initial data collection

All fishing vessels should be registered. A survey may need to be conducted to initiate the register.

If catch is not available, a survey should be undertaken to estimate current total catch. The survey would cover vessel activity and catch per unit effort.

Size composition sampling should be carried out on the landed catch. Catches should be chosen at random and the size recorded for all samples (e.g. meat weight). A sub-sample should be selected ensuring it covers the widest range of catch size to estimate conversion parameters between various measures, such as processed and unprocessed meat weight.

Fisher interviews can be conducted to obtain fishers’ views on the productivity of the stock. This is not only useful in estimating the current state of the stock, but also is a useful move towards co-management, which should improve enforcement.

Data can and should be shared between countries to enable those with little information to develop a monitoring programme immediately.

Other data, such as biomass estimates from surveys, are very useful if they can be conducted, but are not necessary for initializing controls.

8.3 Long-term data collection and monitoring

The following should be implemented over a reasonable, planned time frame:

• all fishing vessels should be required to register. An existing register needs to be maintained and kept up-to-date;

• all commercial catches should be recorded through purchase receipts and processor records;

• other catch and effort data should be estimated through sampling programmes;

• catch size composition should be obtained through random sampling of all catches;

• logbooks should be required for commercial vessels at sea more than one day.

8.4 Assessments linking data to indicators

8.4.1 Primary assessment model

A biomass dynamics model should be used to set limit, precautionary and target reference points for total catch and effort. The default population model should be the logistic biomass dynamic model unless another can be shown to fit the available data better. The logistic model has maximum sustainable yield occurring when the biomass is at half of its unexploited size (see Figure 3).

8.4.2 Secondary assessment model

A size- and age-based model will be required to interpret size composition, ensure adequate replenishment of the spawning stock and optimize yield in relation to growth. Indicators and reference points should be based upon robust decision rules rather than population models.

8.5 Initiating management advice

Robust decision rules must be developed in relation to indicators. In particular, it is necessary to decide upon a rebuilding plan and at what stock level it would be necessary to apply that plan to the fishery. A plan compatible with CITES goal should be agreed with the fishing industry.

8.6 Management control

A vessel register and licensing scheme should be applied to maintain or reduce fishing capacity to the appropriate level.

Controls should be chosen so that they are enforceable and can be shown to have the desired effect on the fishery. An effort control should be applied to cap the maximum effort which can be applied in any year. A quota control should ideally be set on total catch rather than on exports. Additional controls, such as marine reserves, can be used to reduce the risk of overfishing.

8.7 Monitoring

Regular assessments of the data should be carried out. It is not necessary, or even desirable, that assessments be carried out every year. It is however necessary that monitoring of indicators be continuous. A fall in indicators of biomass should initiate an assessment.

It will be necessary to demonstrate that the chosen management controls can affect the fishery. For example, reducing quotas should result in increases in the CPUE index. This can and should be verified by management.

8.8 Long-term research

Longer term research should be co-ordinated in the region. Models for growth and mortality are a priority. Not only do estimates for parameters need to be refined, but explicit measures of uncertainty are also required. Ideally, a default growth and mortality model paradigm should be agreed amongst scientists.

9. GENERAL APPROACH

9.1 Stock identification and management units

As queen conch has a pelagic larval stage, conch stocks may recruit individuals from many surrounding populations. In the strict sense of stocks being completely isolated, it is quite possible that several fisheries will share a single conch population. Although this should lead to calls for greater cooperation, it should not lead to inaction on the part of individual fisheries for two reasons.

Firstly, joint actions are likely to be close to the sum of individual fishery actions taken as though they are exploiting isolated stocks. Therefore it does not take cooperation to know what to do. An individual fishery should undertake action immediately and then seek adjoining fisheries to do likewise.

Secondly, recruitment links between populations may not be strong. Conch has only a short three week pelagic larval stage. It is therefore unlikely that the larvae travel far and the majority of recruits are probably derived locally. In this case, running a country’s fishery as a separate management unit is appropriate.

9.2 From policy to data collection

To apply policy in fisheries, it is necessary to obtain information on the state of the fishery. This is achieved by routine data collection which allows scientists to estimate variables which measure different aspects of the state of the fishery (FAO, 1999).

Fisheries are generally by too complex to manage without some simplification. This can be achieved by reducing the description of change in the fishery to a small set of indicators representing the fishery’s performance.

Interpretation of indicators may be difficult. It is achieved by decision rules usually defined by boundaries, called reference points, at which points the fishery is deemed to change from one state to another. Each policy needs to be represented by a set of indicators and reference points.

For example, overfishing is often talked about but rarely defined. Application of the policy to avoid overfishing needs a precise definition of what overfishing is. A common interpretation is the point where biomass has fallen below 50 percent of its unexploited state.

Policy usually requires two types of reference points:

• limit and precautionary reference points define when the fishery might close and when rebuilding programmes might be instituted;

• target reference points define the optimum point towards which managers are trying to move the fishery.

It is important that both types of reference points are defined for each fishery.

As well as providing an objective way to summarize the state of a fishery, indicators and reference points allow easier communication between scientists and managers and others within the fishery. Although there is considerably more to managing a fishery than simply defining indicators and reference points, they can be used to plan for important changes in management actions and allow easier communication between stakeholders.

Data collection must provide information for the estimation of indicators and reference points. Regular data collection forms the foundation for providing independent advice and arbitration between different interest groups.

Controls that can have a demonstrable effect on the indicators need to be defined for the fishery. Controls on catch and effort should allow managers to bring about changes in the stock and the indicators, demonstrably improving the performance of the fishery.

Indicators must not only accurately represent the state of the fishery, they must be independent of the controls. A control that affects an indicator directly makes the indicator invalid. For example, if CPUE is being used as an indicator, a control affecting the gear efficiency, for example prohibiting the use of SCUBA gear, when it was previously permitted, may change CPUE even if biomass does not change. If this is the case, an alternative, possibly more costly indicator, may have to be developed to bridge this change.

All these issues must be addressed in the management plan. The management plan should encapsulate not just the general policy, but the framework through which it will be applied.

9.3 Logical framework

It is necessary to define a logical system that will achieve CITES overall goal: to avoid endangering Strombus gigas. This is achieved by preventing the various stocks of conch from becoming endangered, which in turn is achieved by meeting various fisheries management objectives (Table 1).

Objectives should be defined not only for the direct avoidance of particular fishery states, such as overfishing, but to achieve socio-economic conditions which will encourage a sustainable fishery. In particular, avoiding over-capacity is necessary to achieve a sustainable fishery.

Consultation and joint decision-making are essential in determining the objectives. The objectives should reflect the reasonable desires of interest groups within the constraints imposed by ecological limits and overriding objectives of national planning. Identifying the various interest groups and encouraging them into partnership with the management authority will often form an initial objective in itself.

The system can be defined by a hierarchical set of objectives. Meeting the lowest set of objectives should automatically achieve the highest goal. Each fishery should set out such a plan. This need not necessarily follow the specific recommendations in this manual, but must apply some logic that guarantees CITES objectives are met.

There is an international standard against which a fishery can be measured represented by the FAO Code of Conduct. There is no perfect management regime. Management is always limited by resources and various social constraints. A checklist is given in Appendix I which can be used to assess a fishery’s management regime.

Figure 1: An illustration of the relationship between the policy and a data collection programme. In this case, catch and effort data are collected to estimate the state of the stock and the exploitation rate. It will be necessary to check that the data collection programme gives reasonably precise estimates of the indicators and that the indicators are truly related to the state of the fishery. For example, applying a reduced catch quota should eventually lead to a demonstrable increase in the CPUE and immediately reduce effort indicators (Adapted from FAO, 1999).

Table 1: Logical framework for conch fishery management. Management should be focused on the means of fishery control and the levels of control to achieve objectives. Objectives are translated into suitable indicators and reference points. If these objectives are met, the CITES goal will automatically follow.

9.4 Standard indicators and reference points

The fundamental indicators are the proportion of the stock that is being removed by fishing (fishing mortality) and the remaining biomass of mature females upon which recruitment ultimately depends.

These variables can be estimated directly or monitored through the use of proxies (Table 2). Proxy variables ideally should be approximately proportional to, rather than equal to, the variable of interest. Even if a stock assessment model cannot be fitted, these variables can be plotted to monitor trends in the fishery.

Table 2: Proxy variables should be related to standard fisheries indicators. The relationship should ideally be verified. The proxies can be used to monitor the fishery without carrying out a full stock assessment.

Particular types of data need to be collected to estimate indicators and reference points (Table 3). Interpretation of indicators may require research (Table 4) or at the very least a decision on the default growth and mortality models to apply. These types of data either are used for monitoring (collected continuously) or as single research programmes. Information from research programmes can be shared within the region, thereby reducing costs.

Table 3: The main data types and the indicators or models with which the data may be used for estimation of parameters and to study the dynamics of the stock and fishery.

Indicators are statistics calculated from the available data. Like any statistic, they are associated with some error. For example, the indicator could be an estimate of the biomass. The biomass estimate will be uncertain, and that uncertainty will need to be taken into account in decision-making.

Table 4: Priority research to enable interpretation of routinely collected data

Reference points require an application of some principle (Table 5). Reference points should define boundaries where fisheries management takes actions or changes a control following some predefined agreement (Figure 2). These points need to be decided by management in consultation with stakeholders.

Table 5: Basic reference points and their principles

Other measures of the fisheries state are also required, particularly fleet capacity and vessel fishing power. This can be achieved through a vessel register. Most countries already operate vessel registers and a licensing system.

Vessel fishing power is of particular concern in many fisheries as on-going improvements in fishing power of vessels may invalidate CPUE indicators. Fortunately in conch fisheries, fishing power is likely to remain stable and easy to monitor if it does change. Critical change in fishing power would be between free diving and use of SCUBA or hookah allowing deeper populations to be exploited. Otherwise vessel length and engine power could indicate fishing power and be monitored through a vessel registration system. These vessel characteristics will not only affect the numbers of divers a vessel might carry, but also will affect the fishing grounds used.

Controlling fleet capacity so that it is appropriate to the potential yield of the fishery is an important aim of management. Excessive fleet capacity is one of the biggest underlying causes for overfishing. For commercial conch fishing, a vessel register is necessary for all vessels exploiting conch. A register should allow managers to control the number and type of vessels which have access to the resource. This in turn limits potential fishing mortality and reduces risks of overfishing.

Figure 2: Example fishery-state diagram of the two main indicator variables. The diagram shows the effort in this hypothetical example that could be allowed at different observed catches per unit effort. The reference points are marked as boundaries. The best estimate for the fisheries state (measured as CPUE) is plotted for each year (1999-2002). In this example, there has been excessive effort and the fishery has entered its rebuilding programme. The main current aim would be to reduce fishing effort below 2 500 boat days until CPUE has risen to above 200 kg day^-1.

9.5 Empirical or non-parametric indicators

While the focus has been on interpreting standard indicators such as spawning stock biomass or fishing mortality, in many cases indicators can be developed which are less dependent on model assumptions, but try to apply controls through decision rules and empirical measures of performance.

While such measures would be more robust where information is limited, they will still be founded upon some underlying models of the fishery or stock. These models still need to be researched, verified and improved. However, day to day management can be made a great deal easier through development of simple rules using indicators which are easy for non-technical staff to understand.

9.6 Probabilistic indicators

Indicators based on probability explicitly take into account risk. That is, the reference points are defined to take account of the risk decision-makers are prepared to accept and the uncertainties in the observed indicators. Decision analysis can be used to define target reference points. Probabilities defining acceptable risk can be used to define limit reference points.

For example, instead of defining a reference point as the biomass being 50 percent of the unexploited biomass, it can be defined as a 10 percent probability that the current biomass is below 50 percent of the unexploited biomass. This is more realistic than fixed points as the current state of the stock can never be defined precisely and this approach allows the uncertainties, for example as shown by the standard error, to be considered as well.

Approaches based on probability are more sophisticated than classical stock assessment and require some expertise particularly where software is unavailable. Methods are generally based on decision analysis which provides powerful methods for assessments.

Probabilistic approaches are most useful where data are poor and uncertainty is a significant component of decision-making. Under these circumstances, reducing risks of overfishing may form the central policy. If data collection is expensive, probabilistic approaches allow a balance between the cost of overfishing and costs of increased data collection.

A policy is required to define how much risk to take. The risk in this case would be defined as the probability of the stock falling below the limit reference point for stock size, or fishing effort being above its limit reference point. While the "precautionary approach" is often put forward as the desirable risk-averse option, it would have to be quantified as specific probabilities.

10. DATA COLLECTION

10.1 Introduction

Data collection forms the foundation for all management decision-making. Data provide independent assessment on the real state of the fishery (Table 6). It is important that data collection is founded on good statistical principles. However, for data which will be used for long-term monitoring, the collection system must be sustainable and therefore not be too logistically or financially demanding.

It is also important to note that conch data will probably be collected alongside catch statistics for other species.

Indicators are calculated from data on the basis of scientific research into the appropriate models and supporting information required. These scientific activities are usually reported through a stock assessment which should be carried out regularly. In particular, stock assessment is required to estimate current biomass and unexploited biomass of the stock (see Section 11).

Table 6: The main indicators and the data and underlying research required for them.

10.2 Catch and effort

10.2.1 Overview

The fishery will commonly consist of two parts: commercial and non-commercial. The commercial fishery, supplying conch for export and local tourist consumption through restaurants, should be able to supply total catches with high accuracy, and effort for the majority of the fishery. The non-commercial fishery, including traditional subsistence, is more difficult to monitor and it would be advisable to use methods based on sampling.

10.2.2 Commercial landings monitoring

Processors should report all the catches they purchase from fishers. This information should be maintained for their own financial records anyway. The requirement is only to copy this information to the management authority.

Receipts should be obtained for all transactions. This is particularly important if the fishery is controlled through catch quotas. If quotas are operating, reported catches will need to be verified by inspection.

Data should be kept as simple as possible. Given a vessel register, only an identifier for the vessel is required (e.g. licence number or captain’s name). For daily trips, only the catch landed on each day is required. If possible the time of departure and return is useful, but not necessary. The fishing times are valuable only where there is significant variation in the number of hours in a trip. In many fisheries the trip time does not vary much.

It is not reasonable to expect a commercial company or cooperative to supply much more data than they record anyway as part of their normal operations. Clearly, the quantity of catch purchased falls into this category, but much beyond this would require another approach.

10.2.3 Commercial logbooks

Commercial vessels going out for more than one day, perhaps carrying out processing on board, should be required to complete logbooks. For larger vessels, this is not particularly onerous as such vessels keep their own private fishing logs.

Logbooks should record on each day the date, location and amount of fishing (some measure of effort depending on gear - in this case hours fishing multiplied by the number of fishers), and catch weight.

10.2.4 Exports

Commercial catches for export can be obtained from two main sources. Actual exports should be reported to the Government Customs Department. As exports require CITES certification by the scientific authority, it should be relatively straightforward to measure all exported landings.

Processed weight needs to be linked to landings through a statistical model. This requires good estimates of the conversion parameter(s) based upon a large sample from the landings.

10.2.5 Commercial purchase receipts

All local restaurant purchases could be monitored through purchase receipts. Restaurants should supply copies each month to the management authority of purchase receipts with the date, fisher’s identity, and amount bought and value of the conch. Receipts should be simple to complete. It may be necessary to carry out spot checks on quantities in storage to verify receipts, particularly if such purchases cover a significant proportion of the catches.

With growth in tourism, restaurant purchase may be one of the largest changes in local consumption. It is important to monitor changes in catches as these have a direct impact on the exploitation level. Therefore, if tourism is changing or fluctuates, monitoring tourist consumption will be an important component of the data collection system.

10.2.6 Obtaining non-commercial catch and effort

Stamatopoulos (2002) describes the sample-based fishery survey method in detail. This method is appropriate to non-commercial conch fishing where there is no easy way to institute or enforce fishers to keep records. Sample-based fishery monitoring requires three components.

Firstly, a sampling frame is required. This usually consists of a complete enumeration of all landing sites and the vessels which will land there. Up-to-date information should available through a vessel register. If no register exists, a frame survey will be required every few years until a register is developed.

Secondly, fisheries staff must visit landing sites and measure the vessel activity. Sites can be chosen at random if there are more sites than staff available. If staff know how many vessels could land at a site from the vessel register, they can check which of these vessels have left to fish and are expected back. They need to visit as frequently as possible in a month. If visits must be limited, they should be made at random times during a month.

Finally vessel trips need to be sampled to obtain estimates of mean catch-per-unit effort. This will require that staff choose vessels randomly when they land if staff cannot measure all vessel catches. Vessels must let them measure their landings. At this point, size frequencies and other data may also be collected through a trip interview.

These data can be combined to estimate total fishing effort and total catch, which can be added to the commercial fishery. Estimation methods also allow the statistical uncertainty to be estimated and included in assessments. This makes the approach potentially very efficient in collecting information cheaply.

10.2.7 Consumption interview data

Where a substantial subsistence fishery exists with no set landing sites, consumption surveys may give some estimate of the quantity of catch taken. This would sample the local population and ask how much and how often people eat conch. Given a total population census, it is possible to raise this estimate to obtain the amount of the total catch that was consumed by local inhabitants which will need to be added into any assessment if not already included elsewhere.

10.3 Catch size frequency sampling

Catches can be sampled at landing sites. Possible measurements will depend on what is being landed. It is recommended that the following are obtained through sampling landings:

• shell length and lip thickness (where the shell is landed)

• unprocessed meat weight

• sex

• maturity

• processed meat weight

For size catch frequencies, a random sample is required. For conversion relationships, a wide range of weights is required. Therefore not all measurements are required for all animals. It is recommended that unprocessed meat weight, sex and maturity be measured for a random sample of landings (around 300 conches should be adequate in the first instance). Of these, a smaller selection taken from the largest, smallest and medium size individuals can be weighed before and after processing. The same process would apply to shell length, lip thickness, (un)processed meat weight, maturity and sex. Any control would have to be consistent so that fishers could avoid discarding dead conches and instead apply selection towards flared lipped conches. This may involve employing fishers to land shells to allow sampling.

There are no commercial size categories that can be used. However, if a minimum weight limit is applied, it is possible animals below this limit will be sold locally. This may still achieve the conservation objective if the price is reduced, but would have to be monitored and assessed.

10.4 Fishery independent surveys

Stock surveys can give immediate estimates of biomass and are particularly recommended where no fishery has previously existed. However, they only give an estimate of current biomass, may be expensive and are not absolutely necessary. They are currently used by Jamaica to monitor biomass.

Likewise fishing experiments may indicate biomass in particular areas and can also be used to estimate fishing mortality. Fishing experiments use intensive fishing in a monitored area, effectively closed to immigration and emigration of conch, to relate the decline in density to the catch. Coupled with a survey, they would be particularly useful and provide rapid information, but again may be prohibitively expensive for some fisheries.

Survey design is a specialist activity requiring particular expertise. However, the basic method will follow common design needs (See Thompson, 1992).

It is necessary to have a map of the survey area on which the sample design is based. The map might be stratified by habitat type. The simplest strata might be by depth.

It is necessary to choose random locations across the sample area within each stratum. The number of transects should be chosen on the basis of the required precision: in general, the more transects that are included, the more precise the estimate will be (i.e. the narrower the confidence intervals on the estimate). All transects must either have their length recorded or must be of equal length.

Transects can be conducted in two ways. The simplest is to apply an absolute width where the surveyor counts all conch within a set distance of the transect line and ignores all others. The alternative is to count all conch that are seen, but measure their perpendicular distance to the transect line. This last allows a detection function to be estimated.

Where conches are common, a fixed width transect is probably the best alternative. However, where they are rare, a method based on perpendicular distance may prove a better technique. It would be possible to mix the two methods, with separate methods applied in different strata for example, although subsequent estimates would be more difficult to obtain.

Juveniles can be separated from mature by their flared lip. Finer divisions based on size and how worn the shell is can be developed with training.

More efficient designs than random sampling may be useful where conches are heavily aggregated. Cluster sampling and adaptive cluster sampling allows sampling effort to focus where conches have been found. This is a more sophisticated approach to sampling, but under some circumstances, particularly where densities are low, may be useful.

10.5 Tagging

Tagging could form an important source of information on conch. Many of the problems associated with tagging other species do not occur with conch. The tag is placed on the shell, so it is fast, relatively cheap and tag loss is negligible. The main problem is ensuring tagged conches are returned by fishers to the scientists.

Conch can be tagged by wire and numbered plastic disks. The sex, maturity, shell length, weight, flared lip thickness are recorded. These data, with meat weight, should be recorded again when the tagged conch is returned. This allows growth and mortality to be estimated. Where possible, tagging should concentrate on smaller, younger conch.

The main problem is getting fishers to return tagged conch to the scientists. Even where fishers agree to cooperate, the tags will be overgrown by algae over time and therefore difficult to spot. Ensuring fishers check for tags during the experiment would be important.

As tagging can provide considerable information quickly on the current state of the resource and fishery, it is always worth reviewing whether a tagging experiment would be feasible in each fishery.

10.6 Other information

Beyond researching specific parameters, biological research can be carried out on life history, ecology and habitat use. This information is valuable in management as well as for developing new improved models for assessments. Such research is not a priority however, but should follow basic stock assessment activities and implementation of appropriate fishery controls. For most scientific authorities, monitoring the scientific literature and attending appropriate workshops and conferences will be adequate to obtain and use this information.

11. ASSESSMENTS LINKING DATA TO INDICATORS

11.1 Introduction

In order to assess any fishery, an operational model of the fishery is required. The aim is to estimate the parameters for the model chosen. The model parameters will then define the appropriate reference points for the indicator. Various assessments are possible, but two approaches are highlighted here.

11.2 Biomass dynamics assessment

11.2.1 Indicators and reference points

An assessment of biomass should be able to use all indicators and estimates of biomass, including surveys and CPUE. The assessment aims to construct a history of biomass change, and the potential and safe yields from the population. Indicators can be simple (e.g. CPUE, catch, effort) or based on model variables (e.g. biomass, spawning stock biomass or fishing mortality). The former has the advantage that it is more easily understood by managers.

The key reference points are the biomass and fishing mortality at maximum sustainable yields (MSY). If biomass falls below the biomass at MSY, a rebuilding programme should be applied. Target reference points can be developed based on economics and fisher’s requirements.

Figure 3: The diagram illustrates stock biomass growth described by the logistic form of a biomass dynamic model. The stock produces no surplus production (that is, it does not change its size) when it is unexploited and when it is extinct (population is zero). When unexploited, the stock reaches its carrying capacity for the environment and only reproduces enough young to replace those which die. When the population is reduced in size, it will produce more biomass than is lost through natural causes, so that the population grows. This excess production is shown on the y-axis of the curve. Production is zero at extinction and at the unexploited size, but there is some maximum in between, labelled as the maximum sustainable yield (MSY) and often used as a reference point. Maximum sustainable yield is too risky for management to aim for as it is right on the edge of overfishing. Alternative target reference points have been proposed to reduce the chance of overfishing. F_0.1 (illustrated) identifies a point before the maximum yield is reached, but still allows most economic benefits to be obtained from the fishery. Similar definitions can be made for reference points in other models.

11.2.2 Data sources

The primary data source for estimating biomass will be catch and effort. These data are important not only to estimate biomass, but to estimate appropriate catch and effort controls.

Survey density can also be used to estimate biomass either as an index or in absolute terms. An index would require less data to get a reasonable result and hence is less costly. In order to ensure comparability in estimates, the same survey must be repeated annually or every few years and may be included in the stock assessment. If an index of abundance is used, then the stock assessment will need to estimate a scaling parameter that will raise the index to an estimate of absolute abundance. Absolute abundance estimates require more sampling, but do not require, in theory, an extra scaling parameter in the stock assessment model and therefore the first estimate can be used immediately. A compromise for monitoring could be to apply many transects to obtain an absolute abundance estimate at the start of the time series and thereafter to revisit a smaller proportion of the original transects, perhaps stratified^[6] by the observed conch density.

Short-term research may provide estimates of potential biomass growth. For example, monitoring recovery within closed areas may indicate the potential productivity rate. Such experiments could be designed and implemented with fisher cooperation.

11.2.3 Assessment model

The simplest and most appropriate operational model would be a biomass dynamics model. The logistic model is used in the Turks and Caicos Islands (Medley and Ninnes, 1999). Other models might be used, but the logistic population model seems most applicable without evidence for alternatives.

The model can be fitted to catch and effort time series as well as any other appropriate index of population size, such as those based on fishery-independent surveys. With enough contrast in the catches, it can be fitted in a spreadsheet (e.g. Punt and Hilborn, 1996, 2001).

11.2.4 Getting started

If a fishery has no time series of catch and effort, or the time series does not include both depletion and recovery periods, it will not be possible to apply classical stock assessment methods. Alternative methods must be used.

If survey estimates of biomass are available, potential yield can be estimated assuming that the survey estimates the unexploited stock. If fishing is already occurring, such potential yield calculations are likely to underestimate the actual yield possible from the stock.

The growth rate parameter estimates (K) for conch vary from 0.2 to 0.7 year^-1 (CFMC, 1999), with the most reliable estimates for recruited conch being towards the lower end of this range. Assuming recruitment at around two years old, and post-recruitment natural mortality and growth rates being about 0.3 and 0.2 year^-1 respectively, the MSY would be around eight percent of the unexploited biomass (Beddington and Cooke, 1983). Sustainable catches therefore need to be maintained at somewhat less than this amount.

If biomass is estimated from a survey with a mature fishery already operating, the survey will estimate current biomass not unexploited biomass. The MSY estimated from Beddington and Cooke in such cases would therefore be lower than the true MSY. The estimate can still be used in a biomass dynamics model or to estimate fishing mortality directly (fishing mortality is approximately the catch as a proportion of the stock biomass).

A more general approach which can make use of all information (including a survey if available) is to apply decision analysis. The method uses available information more efficiently and explicitly takes into account risks. For this reason, this approach is recommended.

The biomass dynamics stock assessment model is relatively simple, requiring only a few parameters. This makes it particularly appropriate when time series data are lacking. Decision analysis needs or may use the following information:

• Size of fishery: An estimate of the current effort and catch in the fishery is required to provide the relevant scale.

• Any catch-effort data series: Even a short series without contrast can provide useful information.

• Fisher interviews: fishers can be interviewed to obtain their view on the productivity of the resource. While fishers may tend to be optimistic, they may be the only source of time series data and in particular the only source on information on what the unexploited state of the resource was (see Appendix III for an example of using only these data).

• Current biomass from a resource survey: The current biomass can be incorporated with its uncertainty.

• Other data from other fisheries: traditionally parameter point values from other fisheries are used in stock assessments, but this will underestimate the uncertainty in the results. Decision analysis allows information to be shared as probability distributions which include uncertainty.

• Measures of cost of actions and outcomes. It is often the case that a review of the potential costs and benefits will allow a decision to be made even if information of the likelihood of outcomes is lacking.

Bayesian statistical software exists allowing various numerical techniques to be applied to model probability distributions, such as Bayes SA (Punt and Hilborn, 2001) which applies a Bayesian approach to fitting stock assessment models. These require some expertise in using and developing models and data sources.

One way that these data and model results can be combined is to use the new software, PFSA (2003; see Appendix III). The software requires that parameter estimates be represented by frequencies, which can be derived from various sources and techniques. The parameter frequencies are used to model the parameter probability distributions which can then be used in decision analysis. Data types that the software can use include interviews and catch and effort data, for example, as well as output from other software (e.g. Bayes SA) which can draw parameter frequencies from likelihoods.

Decision analysis will initiate the management process with clearly defined reference points from the beginning. It will not be necessary to wait for scientific research to be complete and the information can be updated smoothly as more data from adaptive management become available. However, the Bayesian approach is potentially sophisticated and may require some expert help.

11.2.5 Recommendations

The requirement for only a few parameters makes the biomass dynamics model the recommended approach. It has been shown to fit long term time series data from at least one conch fishery (Medley and Ninnes, 1999). The assessment model, in the absence of proving a better alternative, should be the logistic population model. Even where no historical information exists, sharing information among countries may allow an initial stock assessment to be conducted and then updated as information becomes available.

Where a suitable time series of catch and effort is not available, the Beddington and Cooke approach combined with an estimate of current biomass and, if available, an estimate of unexploited or lightly exploited biomass can provide useful information on potential yield until sufficient data exist for more reliable analyses.

It is recommended that decision analysis is used. This not only allows assessment when data are lacking, but is transparent and allows explicit application of the precautionary approach.

11.3 Per-recruit assessment

11.3.1 Indicators and reference points

Per-recruit assessments focus on fishing mortality as the main indicator and control variable. Spawners-per-recruit measures the rate at which the spawning stock is replenished at different levels of fishing mortality. In yield-per-recruit, the aim is to find a fishing mortality level to achieve a particular level of yield for each conch recruited to the fishery. The yield can be adapted to convert to processed meat yield or value. Per-recruit measures can also be used to monitor and interpret mean meat weight (see Appendix II).

The method will also allow size selectivity to be addressed. For example, the impact of a minimum size or flared lip only control can be assessed using yield-per-recruit.

In general, estimates of current fishing mortality and the size at first capture (or a full selectivity function) are required. Initial size is easy to obtain, but generally fishing mortality and selectivity can be difficult. Fishing mortality is usually related to fishing effort.

Reference points are heavily dependent on the growth model and natural mortality estimates. Work has been conducted on this, but wider agreement is required on standard default models and parameters to use. If agreement could be reached by scientists on these models, the way would be paved for regional agreement on harmonized controls such as minimum size.

A limit reference would be F_MSY, although this does not always exist in yield-per-recruit models which may estimate that yield will continue to increase without limit, for increasing fishing mortality. Yield begins to fall after this point, so there is absolutely no reason for fishing mortality to exceed it. Although F_MSY is often at a higher level than desirable for economic benefits or to protect the spawning stock, it may still be used as a limit reference point. F_0.1 always exists and, in the absence of economic information, can be used as a target reference point. It has often been found to be close to the economic optimum when more data becomes available.

It will be necessary for a spawners-per-recruit analysis to be conducted as well as yield-per-recruit to manage the fisheries properly. Fishing mortality must be maintained at a level that does not reduce spawner biomass to too low a level. There is currently no standard spawners-per-recruit reference point for conch. One will need to be developed based on the conch life history and biology, and based upon experience in other fisheries (see Mace and Sissenwine, 1993).

11.3.2 Data sources

The main data source for conch would be catch-at-size composition which can be collected rapidly. Rapid collection is an important advantage of this assessment approach. To estimate mortality however, size must be converted to age. As there is no way to age conch directly, it will be important to have a good growth model.

Catch size composition will depend on two factors:

• Size composition of the population will depend on age composition and growth variation. In its turn, age composition will depend on the recruitment history of the stock and history of mortality.

• Selectivity of the fishing method will decide which sized conches are more likely to be caught. As collection methods are basically the same, selectivity will depend on the distribution of the population by size and the distribution of fishing effort.

Supporting scientific research should produce information important to interpreting size frequency data. Research is needed to estimate growth model parameters and selectivity and to link yield-per-recruit fishing mortality to the chosen indicator (e.g. effort). Data may include more complex expensive methods, such as tagging, in the short term, as well as standard monitoring data.

An important but relatively simple requirement in fisheries is the ability to convert between different measures. For example, conversions will be required between shell length and weight, lip thickness, sex, maturity, processed and unprocessed meat weights. Conversion parameters can be estimated using generalized linear models.

11.3.3 Estimation of indicators and reference points

Reference points require growth and natural mortality models. Either previous published models or parameters can be used (CFMC, 1999) or special scientific research can be conducted to estimate them for each fishery. Also, these two sources of information could be combined, using decision analysis.

Fishing mortality must be estimated for each fishery. Proxy variables, such as fishing effort, can be used but will need to be converted to fishing mortality for interpretation in a yield-per-recruit context. Inaccurate selectivity models are perhaps the greatest weakness of the classical method.

An alternative, still based on yield-per-recruit, would depend on developing appropriate statistics for applying decision rules (e.g. see Appendix II). These approaches can by-pass critical dubious assumptions, but ultimately still depend on growth and mortality models.

11.3.4 Getting started

Length and weight compositions can be obtained immediately. As long as growth and mortality models are available, these can be interpreted.

Although they can be obtained immediately, mortality estimates from length or weight converted catch curves are probably unreliable. They do not take into account growth variability and require most of the catch to be young animals. It is also necessary to know the size/age selectivity of the method used in the collection of the sample (e.g. the fishery) in order to be able to estimate the true length or weight distribution from the sample.

The decision rule described in Appendix II based on an observed catch composition could be used to provide immediate management advice. Clearly, immediate monitoring of the proportion of mature animals being landed could also be used to apply a decision rule. Such rules would at least indicate the appropriate level of control to apply initially while monitoring continues.

11.3.5 Recommendations

A good growth model is required for yield-per-recruit assessment. A number of growth models and parameter estimates exist. These need to be assessed and a final acceptable method for modelling meat weight growth proposed. Results are sensitive to parameter values, so precision and uncertainty would have to be considered.

Estimation is a significant problem for yield-per-recruit approaches. Conversion of size to age is problematic and may prove impossible. As conch cannot be aged directly, it is necessary to rely on conversion using a growth model. Such models are not necessarily reliable for individual animals as variation in individual growth may be high and size does not increase perceptibly with age beyond maturity.

Use of meat weight should revolve around a decision rule rather than absolute reference points (for example, see Appendix II). This allows greater flexibility as well as dealing effectively with uncertainty. Decision rules could be developed for rejecting or allowing exports based on samples, for example, as well as applying management controls. While developing a decision rule may be sophisticated, applying it should be very straightforward.

Yield-per-recruit and spawner-biomass-per-recruit require more information than a biomass dynamics model. However, their ability to use size frequencies which can be collected quickly makes the per-recruit methods useful at least as a comparison to the results from the biomass dynamics model. Ultimately they may replace biomass models as the primary source of management advice as they allow assessment of selectivity which may prove important.

11.4 Other assessments

The natural extension to these assessment approaches would combine the dynamic aspects of the biomass dynamics model with the growth and mortality models. Such age structured dynamic models allow recruitment to be monitored and ultimately modelled. However, they have a very high demand for data, including a time series of catches, effort and size frequencies as well as reliable growth to age conversion. It is not possible to apply this approach at this time.

12. LONG-TERM POPULATION MONITORING

12.1 Population assessments

Long-term monitoring would depend on being able to generate annual catch-per-unit effort (CPUE) by fishing ground. CPUE is the indicator of choice to monitor biomass as it is usually the cheapest method and can be maintained over a long time period. Short breaks in the series do not create a problem, but total catch time series must be complete. Total catches measure the impact of the fishery on the stock and must be recorded and be available.

It is important that the catch and effort monitoring be robust to changes in available capacity. Such monitoring should form a core activity of the management and scientific authorities. Much of the monitoring of commercial catches should use records completed by processors and other purchasers. The scientific authority needs the ability to collect, to check and to store these data.

If possible, stock surveys can be conducted to provide a fishery independent assessment of biomass. Such surveys are valuable as they provide a check on CPUE as well as providing an independent assessment on the scale of the fishery. Such surveys are often used in fisheries to improve assessments. If CPUE is being monitored well, however, they need not be frequent.

The main concern with using CPUE is that there will be changes in catchability. Catchability is a measure of the portion of the biomass caught by a unit of effort and is therefore the scaling parameter between biomass and the CPUE variable. It can change if vessels become more efficient, for example, or if management controls change the way vessels fish (e.g. introducing a minimum size control). Stock surveys can be used to bridge such changes. Vessel register data can be used to standardize CPUE for many changes in fishing power.

Stock assessments require contrasts in the data in terms of population depletion and growth to allow accurate estimates of appropriate controls. Periods of depletion, in particular, may not be considered desirable. However, if it is to be verified that chronic overfishing is not taking place, a period of reduced fishing mortality after monitoring is in place should be applied to see whether the population increases in response to the reduce fishing, and at what rate.

12.2 Standardized minimum meat weight

Preliminary analysis suggests a decision rule could be developed for interpreting meat weights. Two rules are possible:

• A rule based on the optimal size to prevent growth overfishing (see Appendix II). It would apply to controlling gross landings, fishing effort and other fishing mortality related controls. The rule would also allow management to evaluate controls in relation to selectivity (see Appendix II for an example of a simple decision rule).

• A rule based on only landing conch with flared lip. The rule would be applied to enforce selectivity by fishers to take only mature conch. It needs to be verified that a meat weight control can be used to enforce this.

It should be noted that prescriptive regional controls could have a detrimental effect on some fisheries. In particular, different selectivity among fisheries remains a problem. Where free diving is the sole gear, mature conches that have escaped the fishery can be found in deeper water in abundance. A policy shifting emphasis to large conch could encourage exploitation of the spawning stock which is probably best left alone in these fisheries. There should therefore be scope for some flexibility in controls at the national level and possibly even at the local level.

13. CONTROLS AND MONITORING

13.1 Introduction

Stock assessments are only useful where they lead to some control on the level of fishing. Management authorities must be prepared to limit and reduce fishing activity, where required, to protect the resource.

Any control aiming to improve the stock state must reduce catches at least in the short term. If they do not reduce catches, they are not being effective. For example, putting marine reserves where fishers do not fish will have no positive effect (although it may reduce problems later). Reducing catches can make controls unpopular with fishers, but it is up to management to minimize problems through consultation and joint decision-making, and to demonstrate the advantages of the control to fishers through emphasis on longer term benefits.

Experience has shown that it is always necessary to consider the socio-economic effects of controls. Enforcement will, at best, be difficult if the needs of industry and fisher communities are not taken into account. Every effort should be made to apply controls which not only achieve biological objectives, but socio-economic ones as well. As overfishing is the worst socio-economic outcome for a conch fishery, it should be possible to find some level of control that satisfies most realistic objectives. In cases where socio-economic dependence on a conch stock exceeds, the sustained production potential of the stock, the socio-economic needs can only be met by external intervention, either by providing short-term support during restructuring or by assisting fishers to find alternative livelihoods.

13.2 Fleet capacity and effort

All commercial vessels should only be allowed to fish if they are registered and licensed by the management authority. The number of licences issued for conch should be commensurate with the number of vessels needed to harvest the resource.

If fleet capacity is much greater than the allowable effort, there is likely to be tremendous economic and political pressure to allow greater fishing activity. Fleet capacity in commercial fisheries must therefore be controlled. Both numbers and sizes (fishing power) of vessels can be controlled by limiting registration.

Effort and fishing mortality may be controlled by limiting the number of days that the registered, and hence legal, vessels spend at sea. Closed seasons may allow this, but specific effort controls are often difficult to administer. Enforced closure of processing plants may also allow some control.

13.3 Catch quotas

Catch quotas will control fishing and at least exports can be enforced. However, they require careful monitoring of the stock and of the catches and can easily allow overfishing. For example, setting the quota to the maximum sustainable yield is inherently unstable and will always lead to overfishing the resource. It is recommended catch quotas are used alongside other controls.

To be applied, it will be necessary to be able to convert catch quotas to export quotas accurately. All exports need to have a CITES certification so quotas could be enforced at this point if all or nearly all of the commercial catch is exported. Conversion needs to be monitored, as meat yield may change with the size composition of the catch.

Fishing mortality can be related to catch in the form of a feedback system. Lowering catch quotas will lower F, but setting the precise quota to achieve a particular fishing mortality will be difficult. Using biomass dynamics models will allow direct estimation of appropriate quotas.

13.4 Minimum size and maturity

Minimum size should be related to growth and size at maturity. In general, unless the shell is landed, minimum size is difficult to apply. The meat size composition may indicate violations of a flared lip rule (only mature conch allowed), but only gross violations may be detectable. Hence, minimum size may prove less useful for direct enforcement, but remains a useful indicator of the performance of other controls.

A control requiring fishers to only take conch with a flared lip would require cooperation from fishers. It would be possible to prevent immature conch from being landed, and this may be a useful control in the non-commercial sector. It would require a fairly sophisticated education awareness campaign, however.

13.5 Closed area

There may be effective closed areas already present in some countries through gear controls. Fisheries which only allow free diving (e.g. Turks and Caicos Islands) have de facto protected deep water populations. Otherwise, properly designed closed areas could be useful for reducing risks of overfishing by protecting a portion of the stock.

13.6 Closed season

In general, closed seasons can be used for four purposes.

• To protect the stock during critical periods, such as spawning.

• To gain from growth by delaying recruitment to the fishery. This would require that the species has a discrete recruitment period.

• To reduce effective fishing mortality by reducing the number of days in which fishing can occur. This would be the most likely use of a closed season as a biological control.

• To achieve socio-economic alternatives. For example, a closed season may allow a quota to be spread more evenly through the year.

Whether closed seasons are used for any of the above purposes will be up to managers, stakeholders and scientists in local fisheries.

13.7 Taxation

Taxation on exports might be used to raise revenue to pay for fisheries management. Management costs governments money, yet fishers should gain significant benefits from good management. Export taxes are one means to help pay for this management.

Importantly, taxation also discourages overfishing by lowering the effective price paid for conch. Regional cooperation on setting taxation levels would reduce problems in terms of competitiveness and thereby help to protect the regional resource.

14. OTHER ISSUES

Unlike other fisheries, conch fishing is not associated with habitat damage or significant by-catch. It is possible conch fishers will take other species opportunistically, but in general conch does not form part of a multispecies commercial fishery.

15. ADAPTIVE MANAGEMENT

15.1 Management cycle

Management should follow a basic cycle. Once the policy and objectives for the fishery have been developed and interpreted in the form of indicators and reference points, the monitoring and assessment cycle can start.

1. Data should be collected for estimating indicators and monitoring the fishery.
2. Indicators should be updated and stock assessments conducted, if necessary.
3. The effectiveness of previous fishing controls should be evaluated.
4. Based on these results, scientists should provide advice to the managers.
5. Based on the scientific advice, fishing controls should be adapted where necessary and applied.
6. Policy and objectives should be reviewed and revised where necessary and desirable.

This cycle should be repeated indefinitely, although not all steps may be conducted in every year. There should also be close consultation with stakeholders during this cycle, for example at steps 3, 5 and 6.

The stock assessment should be undertaken to update indicators and reference points with new data, as appropriate. All technical aspects of the stock assessment should be fully documented, including a description of the data used, method applied and assumptions made. Models and data should be stored for future reference. This makes updating the assessment in the future much more rapid.

The assessment should report separately to managers the stock status and the management advice. Full technical information on the stock assessment will not be required in this report and a non-technical summary should be submitted to decision-makers. This should make clear the uncertainties in the assessments.

For discrete changes of state, for example from fully exploited to stock rebuilding, it is recommended that decision rules are used to specify the management actions (e.g. see Figure 2). Decision rules should already be agreed by decision-makers, so the scientist’s job should be to evaluate the rule and report back the evaluation. This should make communication of results more straightforward.

It is recommended that the rules applied should follow a system similar to that shown in Figure 2. If the stock is already low or the stock size falls below a particular, pre-specified level, a rebuilding plan should apply. This will lead to setting lower effort (or quota) than the sustainable yield, allowing the stock to grow. A principle is needed, such as the time to rebuild, to set a particular level of fishing. Hence the exact plan would have to be developed through consultation between scientists, managers and stakeholders.

Adaptive management can actively apply management measures to gain information about the resource. The main problem with many stock assessments is that they require contrast in the catch time series. Ideally, the time series should include periods of both depletion and recovery. If this information is not available, it is quite possible for a fishery to remain in a chronically overfished (or underutilized) state indefinitely without the potential for recovery and higher yields being known. Carefully reducing and increasing controls within the overfishing limits is quite legitimate and may provide the final test of other scientific results. Other management actions, such as setting up temporary closed areas, may also form part of these adaptive management actions.

15.2 Verification and transparency

The management plan and assessments should be subject to periodic review. This will ensure that the assumptions and advice are reasonable and based on the available evidence. The reviewer can be internal (although using an external reviewer could lead to greater objectivity) but should not have been involved in preparing the plan or in the assessments so as to ensure objectivity. Comments of the reviewer should be addressed, and suggested modifications incorporated in the plan where appropriate.

A checklist for testing a management regime has been proposed for fishery certification (Appendix I). This can be applied to conch fisheries in exactly the same way to provide an independent assessment of how well management is doing against the best practice. The tests go well beyond CITES requirements, but recognize the wider benefits of good management as contained in the FAO Code of Conduct, of which the CITES requirements represent a part.

16. POTENTIAL REGIONAL MANAGEMENT REGIME

Cooperation between countries on many issues will be difficult. Data and information sharing is one area which should not be contentious and is in the best interests of all. Scientific workshops focused on conducting assessments using shared research and data could produce significant insight into how to manage this resource.

Further cooperation would have to be founded on developing commonly accepted models for fisheries and conch biology. Common biological models will form the basis for choosing common reference points. Common indicator and reference points would then allow different countries to apply the same management rules. Regionally accepted and applied growth and mortality rates would be necessary to identify a single optimum minimum meat weight for example.

Although common indicators, reference points and controls should make enforcement easier, there may well be a cost to individual fisheries. A common minimum size policy will assume a particular growth model applies to all stocks. Conches which grow at a rate higher or lower than this model implies will result in an inefficient choice of size, either ineffective or too restrictive.

The most likely areas of immediate concern across the region should probably be a minimum mean meat weight, closed season and export tax level. It is suggested that these issues be considered by scientists first in assessments and then advice passed on to managers to see whether and how much cooperation is advisable.

17. REFERENCES

Beddington, J.R. & Cook, J.G. 1983. The potential yield of fish stocks. FAO Fisheries Technical Paper (242): 47p.

Berkes, F., Mahon, R., McConney, P., Pollnac, R. & Pomeroy, R. 2001. Managing small-scale fisheries. Alternative directions and methods. IDRC, Canada. 320p.

CFMC. 1999. Report on the Queen Conch Stock Assessment and Management Workshop Belize City, Belize, 15-22 March 1999.

Cochrane, K.L. (ed.). 2002. A fishery manager’s guidebook. Management measures and their application. FAO Fisheries Technical Paper. No. 424. Rome, FAO. 231p.

FAO. 1995. Code of Conduct for Responsible Fisheries. Rome, FAO. 41p.

FAO. 1997. Fisheries management. FAO Technical Guidelines for Responsible Fisheries. No. 4. Rome, FAO. 82pp.

FAO. 1999. Guidelines for the routine collection of capture fishery data. FAO Fisheries Technical Paper 382. Rome, FAO. 113p.

FAO. 2003. The ecosystem approach to fisheries. FAO Technical Guidelines for Responsible Fisheries. No. 4, Suppl. 2. Rome. FAO. 112p.

Mace, P.M. & Sissenwine, M.P. 1993. How much spawning per recruit is enough? pp. 101-118. In: Smith, S.J., J.J. Hunt and D. Rivard (ed.). Risk evaluation and biological reference points for fisheries management. Can. Spec. Publ. Fish. Aquat. Sci. 120.

Medley, P.A.H.& Ninnes, C.H. 1999. A stock assessment for conch (Strombus gigas L.) fishery in the Turks and Caicos Islands. Bulletin of Marine Science 64(3): 399-406.

PFSA. 2003. Participatory Fisheries Stock Assessment. Technical Report, Software and Help File. MRAG. (contact: [email protected]).

Punt, A. & Hilborn, R. 1996. BIODYN Biomass dynamic models. User’s manual. FAO Computerized Information Series (Fisheries). No. 10. Rome, FAO. 62p.

Punt, A. & Hilborn, R. 2001. BAYES-SA Bayesian stock assessment methods in fisheries. User’s manual. FAO Computerized Information Series (Fisheries) No. 12. Rome, FAO. 56p.

Silverman, B.W. 1986. Density Estimation for Statistics and Data Analysis. Monographs on Statistics and Applied Probability 26. Chapman and Hall, London.

Stamatopoulos, C. 2002. Sample-based fishery surveys. A technical handbook. FAO Fisheries Technical Paper. No. 425. Rome, FAO. 132p.

Thompson, S.K. 1992. Sampling. John Wiley and Sons, New York.