3. INDICATORS AND DATA VARIABLES

3.1 FISHERIES ACTIVITY DATA

This chapter introduces the basic concepts and terminology, which is required to define a fisheries data collection programme. The definitions of concepts are illustrated by examples.

3.1.1 Catch, Landings and Discards

By “Catch”, we mean the biomass caught by a gear (a trawl, a hook, a purse seine etc.). A part of the catch is brought on to the deck or into the hold of the vessel and a part may be “slipped”, that is released from the gear without being taken onboard the vessel. Of the catch onboard the vessel, a proportion is landed and the rest, if any, is discarded (returned to the sea). It is traditional in fisheries biology to use the word “Catch” for the number of individuals and the word “Yield” for the weight of the biomass caught. However, in this manual, there is no reason to make a special distinction between numbers and weight, so “Catch” can mean either weight or number depending on the context.

“Landings” means the part of the catch that is actually brought on land. In the present context, we shall include the slipped catch into the group “Discards”. Thus, by “Discards” we mean the part of the catch that encountered the gear, but was not landed.

As for “Catch”, “Landings” and “Discards” may refer either to “number of individuals” or “weight”, and the reader must judge from the context which interpretation is applicable.

The division of the landings into commercial groups and the recording of the landings will be discussed in Section 5.3. Discards are discussed in Section 3.4.

3.1.2 Fishing Effort and Activity

One of the main targets of fisheries investigations is to link the mortality of fish with fishing effort. That is, for example, to find the link between the number of vessel-days of each fleet and the proportion of the stock that was harvested. The proportion of deaths caused by fishing is indicated by the “instantaneous rate of fishing mortality”, usually denoted by “F”. The relationship between F and effort is often assumed to be linear, but this may not always be true. To find the relationship between F and effort you must know the historical development in the fishery. That is, it must be known how many effort units each fishing fleet has exerted during some period; for example, the number of fishing days per year by each fishing fleet.

For a bio-economic analysis, effort is the link between the biological and economic models. It is related to production through fishing mortality, as well as to variable and fixed costs. Effort expressed as fishing days or days away from port is the most important variable for the bio-economic assessment of fisheries, as the number of active days is often assumed to be linearly related to the variable costs of fishing.

Fishing effort can be measured in many different ways. The effort measurement may be selected to fit a specific type of vessel and gear. For example, for a trawler you might use the number of trawling hours, for longlines the “number of hooks per line” and for gillnets the “number of gillnets set per night”. What can actually be used as measure for fishing effort, of course, depends on which data are available from the fishing operations.

Subsequently, the number of fishing effort units of each fleet must be compared to the number of fish that died due to fishing by that fleet. To estimate the proportion of fish which died due to fishing is a main objective of fish stock assessment (see Section 3.2.2).

Table 3.1.1 lists possible measures of effort by gear categories. The table is divided into five “priorities”, where the effort measures, which are most likely to achieve the best relationship between effort and fishing mortality, are the “first priority”.

Table 3.1.1 List of effort measures, in order of priority according to the ability of measure to provide a relationship between fishing effort and fishing mortality).

The concept of “Fleet” will be introduced in Section 3.1.3. Here we shall only note that within a “fleet” of fishing vessels, it is assumed that one day of fishing (or one unit of effort) by any member of the fleet generates the same fishing mortality.

The average number of effort units exerted per time unit per vessel (for example, the number of fishing days per year) is called the “activity level”. The activity in general is the “Number of effort units exerted per time unit”, where “Effort unit” can be one of the measures suggested in Table 3.1.1. Often it is not possible to make a complete enumeration of all effort units, but it is possible to estimate the activity level from samples.

The sampling of activity observations may be done with trip interviews (see Chapter 5), but it may also be collected through a frame survey (see Section 5.2.2). From the estimate of the average activity level and the number of vessels, the total number of effort units is estimated as the product:

(Total effort per time unit) = (Number of vessels) × (Average activity level)

3.1.3 The Fleet Concept

A “fleet” is a group of uniform vessels, which have approximately the same size and the same construction. The vessels should use the same type of gear and fishing techniques and most often, they share fishing grounds.

The fleet definitions may change during the year. A vessel may, for example, do pair trawling for fish during one season and do single trawling for shrimps during another season. Some vessels use a combination of gears during a fishing trip, which may complicate the allocation of vessel to fleets. Sometimes a fleet defined as a “gear combination fleet” can solve the problem.

Fleets may be defined by a combination of gear, engine horsepower (size of vessel), type of construction and fishing grounds. Horsepower, tonnage and length of vessel are usually correlated within a group of vessels of the same basic construction type.

One practical problem is that the sampling programme must adequately cover every fleet. The number of samples from each fleet should have a size, which makes the estimation of the mean catch per unit of effort reasonably accurate. The table below contains an example of categories of fishing vessels according to horsepower class, gear and fishing grounds:

Although this division does not appear very narrow, it nevertheless results in potentially 6^*7^*3 = 126 combinations of categories or different fleets. Many of the 126 combinations, however, may be empty. For example, there may not be a lift-net with an engine over 75 HP. The example above suggests a low upper limit on the level of detail, which it is possible to account for in practice.

When the fleets have been defined, we shall in the following assume (as an approximation to reality) that all vessels in a fleet are exactly equal and behave in exactly the same way.

Within the sampling context, the fleet concept is needed for two main objectives:

• To raise catch effort samples to total landings (See Section 4.4). Catch and effort data are combined with activity data and frame survey data (or vessel register) to estimate the total landings by raising sample estimates.

• To assess the competition between fleets (See Section 3.2.3).

There are additional reasons for using the fleet concept. For example, the fleet concept is needed to define suitable measures for CPUE (Catch Per Unit of Effort), which is an important issue in many different types of assessment of fisheries and resources.

The two main objectives may not always lead to the same definition of fleets. However, often the objective of estimating total landings leads to more fleets than necessary for assessing competition between vessels.

The simplest type of “raising of samples” is achieved when it is assumed that all fishing vessels within a fleet have the same “fishing power”. In its simplest form, raising means that if one vessel is “sampled” and it was observed that the catch was 50 kg/day, then we can “raise” this to the fleet, by multiplication with the number of members of the fleet that were active. In other words, when we “raise” samples we can use the same “raising factor” for all vessels. This is a reasonable approach if all members of a fleet are similar.

Two fishing vessels are said to have the same “fishing power” if they can catch the same amounts and types of fish under similar conditions. For example, two trawlers fishing on the same fishing grounds at the same time must catch the same amounts of fish in species, numbers and sizes to have the same fishing power. One may simplify the concepts of fishing power by making it species-specific. Thus if two trawlers, A and B, catch the same amounts of, for example, threadfin breams under similar conditions, then they have the same fishing power relative to threadfin breams. If vessel A catches the double amount of thread-fin breams as vessel B, then vessel A has two times the fishing power of vessel B. This means that one vessel of type A “counts” the same as two vessels of type B.

In practice, this ideal definition can rarely be shown to hold. Instead, if the two trawlers catch the same amount of “demersal fish” during a fishing operation on average, they have the same fishing power, and if one vessel catches X % more on average than the other vessel it has X % more fishing power.

A concept closely linked to fishing power is that of a “standard vessel”. It is often desirable to express the fishing power relative to some selected vessel type. Usually the most common vessel type is selected as “standard vessel”. That may for example be the trawlers of length 15 m with an engine of 60 HP and perhaps some more specific characteristics. Other types of vessels are then expressed in units of standard vessels. If a vessel has 80% of the fishing power of a standard vessel, it counts as a “0.8” standard vessel.

This approach is intuitive, as the purpose is to convert fishing effort into fishing mortality. Two vessels, which remove the same percentage of fish per day from the stock, create the same fishing mortality according to the definition of fishing mortality. If they fish at the same time, then the same percentage will result in the same numbers caught, thus the same fishing power.

Some of the factors that determine fishing power are rather complex. One factor, which is difficult to quantify, is the skill of the skipper. Other factors determining the fishing power are easier to measure, such as engine power, vessel length, year of construction, fitted electronic equipment, mesh size, rigging of the trawl and other physical features of the vessel and gear.

3.2 BIOLOGICAL INFORMATION

Full biological information should cover the composition of the catch, including discards. However, as information about landings usually is easier to obtain than information about the catch, biological information is often based on landings only.

3.2.1 Species Composition of Commercial Groups

Biologists use the fishery data for fish stock assessment. A fish stock is a sub-set of a species, so before stocks (or management units) can be identified, the species must be identified. The landings are often sorted into commercial groups, so the starting point for the sorting into species is usually the commercial group. If a commercial group is only one species, the task should be easy. The enumerator, should however, check that the commercial group is only one single species, as two similar species may be merged in a commercial group. Two similar species will be separated only if the buyers of the landings and their customers appreciate the two species as different products. Other commercial groups, mainly of medium value, contain three to five different, but usually similar, species. For low value products, the number of species in a commercial group may be very large. A commercial group like “Mixed low value small fish” may also contain small specimens of the high value species as well. In any case, the enumerators must take samples from some trips (not necessarily all trips) to record the species composition of the commercial groups.

Most tropical fisheries involve a large number of species. There is no way available manpower and funds will allow for all species to be recorded. The sampling programme must prepare a list of, say, the 400 commercially most important species. Other species found in the species composition samples would then appear as “Other”, although recording the number of species (or “recognisable taxonomic units”) in a sample may prove useful in future assessments. The number “400 commercial important species” is not universal, and the designer of the sampling programme will have to decide how many species the funding and manpower allows for. The more species are recorded, the more training of enumerators in species identification is needed, and the more time will be spent sorting out species.

Theoretically the species should subsequently be split into stocks, but usually the stock separation in tropical waters has to be based on a set of rather crude assumptions, such as all specimens landed in a set of landing places originate from the same stock. There are sophisticated methods that can be used to separate stocks, but for those the reader is referred to specialised literature.

3.2.2 Fish Stock Assessment, Virtual Population Analysis

Although the methodology of fish stock assessment is outside the scope of this manual, this subsection briefly discusses it, with the main purpose of introducing the type of data needed. It is often easier to understand which data to collect if you also understand how the data are to be used. For further details on fish stock assessment, the reader should consult other manuals and textbooks (see, for example Sparre and Venema 1998 and Lasswen and Medley 2000).

As an example, we shall use the input data needed for Virtual Population Analysis (VPA). The principal input data for VPA is the number of fish caught by size group (length or weight group), which must ultimately be converted into age group. These numbers must be the total catches of all fleets exploiting the stock in question.

It is possible to read the age of a fish from the ring structure (reflecting the seasons of the year) in certain hard parts of the body. However, for tropical fish, where differences between seasons are not so great the annual rings are not easy to identify. It is here assumed that no attempt to read ages of fish or squid is made. Instead, we shall use the relationship between body length and age. There is usually a relationship between age and body length for young fish, squid or shrimp.

Figure 3.2.1 An example of a length frequency sample.

The method is called “model progression analysis”. Length frequency data (number of fish in the landings in each length-group) can be collected for each month of the year. A length frequency sample is depicted as in Figure 3.2.1, showing a sample of 5390 specimens distributed over 30 length groups of 1 cm. The length frequency samples should be collected during a longer period, preferably not less than a year. One single length frequency sample gives only a “snapshot” of the situation, but to observe changes over time, you must collect length frequency samples must be collected frequently, every month or every quarter of the year.

The reason why the modal progression analysis is possible, is that recruitment to the fish stocks does not occur at the same rate all year round (Figure 3.2.2). The picture shown in Figure 3.2.2 is typical for the seasonality of recruitment of many fish, cephalopod and shrimp stocks. There are usually two spawning seasons during the year. One major spawning followed by a second smaller season.

Figure 3.2.2 Seasonality of recruitment (hypothetical example).

Figure 3.2.3 Simulated input data for Modal Progression Analysis. The figure illustrates the history of a stock during a year from samples taken each quarter.

These peak recruitments can be traced through the size frequencies. To do the “modal progression analysis”, we put length frequencies together into a larger graphical picture, and try to find the cohorts by splitting the sample into cohort components (Figure 3.2.3).

There are many theories on the explanation for the observed seasonality of recruitment, which is outside the scope of this manual. We shall point here only that the monsoon and the seasonality of oceanographic features are factors, which influence the conditions for survival of fish in the early stages of life and thus also influence the recruitment to the exploited stock.

Figure 3.2.3 is a hypothetical example constructed to show the principles of modal progression analysis. It is rather easy to separate the 1-group of fish in all samples. In the January sample, it has mean length of about 8 cm and extends from about 4 cm to about 12 cm. In the April sample the 1-group is less in number but they have grown to a mean length of about 10 cm and now with a larger variation between individuals. In April, however, a new cohort, the 0-group, is entering the stock. It has a mean length of about 2.5 cm and it extends from 2 cm to 3 cm. In July, the 0-group has grown to about 4.5 cm length. In this hypothetical example, the youngest cohorts gain approximately 2 cm in the first three months.

Now suppose that the graphs are not just samples, but that they represent the length distribution of the entire catch. In that case, we can estimate the fractions of fish, which survive each quarter of the year. If we count the number of survivors in the two first cohorts, we get (approximately):

The 0-group is increasing in numbers, whereas the 1-group is decreasing in numbers. Taking into account that the fishery starts to catch most fish only after they have reached a certain size this is the expected result. The 0-group is not yet fully recruited to the fishery, whereas the 1-group is fully exposed to fishery.

The decrease in numbers in the 1-group can be used to compute the rate of exploitation (how quickly the fishery removes fish from the fish stock). This assumes the same length distribution in both the catch and stock, so the percentage reduction is the same in each.

If we can do the modal progression analysis, we can also use VPA methods. However, sometimes the modal progression analysis is very difficult. That happens in particular when the length frequency samples are not very big, and in that case we use a more robust method, a method which provides a better fit to the data, but which gives less detailed output. This method is called “Length-based cohort analysis”. The length-based cohort analysis is based on a number of assumptions about the fish stock dynamics, which make the computations easier.

The “price” we have to pay for an easier method is that the results become less precise. The assumptions we make are never strictly met in reality, they always represent an approximation. In the length-based cohort analysis, we do not try to separate the cohorts, but we sum the length frequencies over the year. The most critical assumption we make in length-based cohort analysis, is that of constant recruitment. Unfortunately, recruitment is usually highly variable, therefore the assumption of constant recruitment is an approximation, which sometimes may be far from reality. However, the average recruitment for a number of years will be less variable, and if several year classes are considered, the assumption may be less critical. The assumption of constant recruitment should be made only when no other alternative is available.

The sum of the length frequencies of Figure 3.2.3 is shown in Figure 3.2.4. The next step in the length-based cohort analysis is to study the slope of the summed length frequencies and to use the results to estimate the number of recruits to the population and the fishing mortality rates for each size group.

Figure 3.2.4 Input data for length-based cohort analysis (the sum of data of Figure 3.2.3).

3.2.3 Fish Stock Assessment Forecast

VPA and cohort analysis are retrospective, they indicate what happened to a fish stock in the past. The main purpose of a fish stock assessment is to predict what will happen in the future. We may say that the ultimate objective of fish stock assessment is to answer “What-if-questions”, like “what will happen to the stock if the fleet size increases by 10%”. Where the VPA focused on the “stock”, the forecast focuses on the “fleets”.

Another important element of fisheries assessment is to analyse the competition and interaction between different fleets. To illustrate the nature of the problem, consider a country with only two different types of fishing vessels, say “fishing fleet 1” and “fishing fleet 2”. In 1996 fleet 1 consisted of 100 large industrial vessels and fleet 2 of 2500 artisanal vessels (Figure 3.2.5).

The total value of the landings is $10 million for fleet 1, which is $0.1 million per vessel per year. The figures for the artisanal vessels are $12.5 million for the total and $0.005 million per vessel per year.

We can then ask the question “What happens if the number of industrial vessels are raised from 100 to 500. We could, for example, answer the question by computing (or “by predicting”) the landings two years later, that is in year 1998.

In the example given here, the total value of the catch goes down from $10 million to $5 million for the industrial vessels and from $12.5 million to $5 million for the artisanal vessels. A total decline in value from $22.5 million to $10 million.

Figure 3.2.5 Illustration of a “what-if” question regarding interaction between artisanal and industrial fleets.

3.2.4 Data Requirements for Fish Stock Assessment

As discussed above, to run the retrospective analysis and the forecast, two principal types of data are needed for each fish stock considered:

1. Total catches by species and by size group for each period (quarter or month) caught by each fleet.

2. Total effort by fleet and by period (for example number of fishing days/month).

The total catch data (by fleet) is derived from two types of basic data:

1. Total weight of catch by species (or species group).

2. Size frequencies for selected species.

Figure 3.2.6 The principal input data for length-based cohort analysis by fleet. Species 2 grows to a larger size than species 1 and there are some differences between the sizes selected by the two fleets. Effort is also useful for fitting models to the catches. As species 2 is caught by the same fleets as species 1, the corresponding fishing effort is the same for the two species.

Figure 3.2.7 The principal input data for length-based cohort analysis summed over fleets.

Figure 3.2.6 illustrates the dimensions of the data in a given period. It shows an example with length frequencies for two fleets catching two species. Species 1 gets up to 12 cm long and species 2 gets up to 20 cm long in this hypothetical example. The sampling programme should produce a graph like that of Figure 3.2.6 for each time period of the year, containing all species and all fleets.

As input to the length-based cohort analysis we will use the frequencies (for each species) summed over fleets. Figure 3.2.7 presents the input for two independent cohort analyses, as the cohort analysis is carried out separately for each species.

Usually, it is not a great problem to collect samples of length frequencies. The funds for buying the samples must be available (although the fish can be sold after measurement), or the fishers/buyers may allow measurements without charge. The necessary scientific and technical manpower to collect and analyse the samples must also be available. Furthermore, you must have the co-operation of the fishers and/or buyers to take the samples.

Although it may not always be possible to get certain important types of data, some data should always be obtained for monitoring purposes. If a data collection programme cannot get total catches, the analysis should be based on relative figures. For example, relative figures of length frequencies can be derived from samples. If length frequencies cannot be obtained, other types of data, such as catch (in weight) per day, may still be applied for fisheries assessment.

3.3 BIO-ECONOMIC INFORMATION

Bio-economics is a combination of resource evaluation (i.e. fish stock assessment) and a cost and earnings analysis of the harvesting sector. Ideally, the management and development of fisheries should be based on data and analyses representing all major aspects of the fisheries sector. Thus, biological, economic and sociological information should be collected and processed so the combined output can be used for rational decision making (Figure 3.3.1). Sometimes the processing sector is also included in the analysis of fisheries bio-economics. Here we shall consider only the combination of resource evaluation and an economic analysis of the harvesting sector. The bio-economic methodology is structured by a biological/technical sub-model and by an economic sub-model (Sparre and Willmann 198? And Seijo 1999).

The economic sub-model introduces prices, costs and a range of economic performance criteria. The biological/technical sub-model establishes a physical relationship between fishing effort and fish production (landings). The optimisation of the fishery based on physical quantities (e.g. total catch in tonnes) would result in maximising the sustainable physical yield. As the exertion of fishing effort incurs costs, and as different species and sizes of fish realise different prices in the market, maximum sustainable yield is not, in most instances, a desirable objective of fishery management from an economic point of view.

There should not be two independent data collection programmes, one for the biological data and one for economic data. The biological, technical and economic information are most often collected from the same sources, and are used in combination in analyses. Therefore, the collection of biological/technical data and economic data should be combined in one integrated programme. This approach is cost-effective, and will ensure that the data collected are compatible. Hence, the aim here is not to give a complete list of all possible fisheries economics data, but rather to mention some of the key information, and to discuss how these data can be collected together with the data for resource evaluation.

Figure 3.3.1 Illustration of an approach to bio-economic assessment of a fishery.

3.3.1 Fish Prices

Bio-economics operates with two principal types of prices, ex-vessel prices and wholesale prices. Ex-vessel prices are those received by the fisher/vessel-owner at the landing site and those are the only prices considered in this manual.

Different species and sizes of fish generally fetch different prices in the market. In addition, fish of the same species and size may realise different prices because of differences in product qualities, often related to the handling and storage facilities on board of the vessels.

Wholesale prices are those received by the first hand buyer when selling fish to either the domestic or export market. Fish sold for export will earn foreign exchange, which may have additional economic benefit.

Prices are always given by commercial group, and the commercial group is therefore a very important element linking biology to economics. The value of the catch is a useful quantity. When combining catches of different species groups, it may not make sense in a bio-economic context to add together the biomass of species groups, if the price per kilogram differs substantially between species groups. By converting biomass into value, all quantities will be given in the same unit. For example, adding biomass of shrimps, cephalopods and fish often makes little sense, as this sum is rarely part of an explanation for the behaviour of the fishers or the resources.

3.3.2 Costs of Fish Harvesting

There are three different types of harvesting costs, namely

1. Costs depending on the number of fishing effort units;

2. Costs depending on the number of fishing vessels;

3. Costs depending on the value of the landings (i.e. ex-vessel value).

The first category of costs is often assumed to be linearly related to fishing effort. If fishing effort goes up by a factor of two, these costs also go up by a factor of two. This category of costs comprises elements such as fuel and oil, repairs and maintenance, ice, crew wages (independent of yield in value), etc. and is also referred to as “variable costs”. For a particular fleet, the total of this cost category is calculated by multiplying the total number of units of effort (e.g. number of fishing days) expended per year with the total costs per unit of effort (e.g. costs per fishing day).

The costs that depend on the number of vessels refer to costs that arise even if a vessel does not go out fishing. They are also referred to as “fixed costs” and comprise elements such as capital depreciation, insurance, interest, mooring fees, refit etc. These costs are usually given on an annual basis. The total costs of this category for a certain fleet would be given by multiplying the number of vessels by the average total fixed costs per vessel per year.

The third category consists of those costs that depend on the ex-vessel value of the landings. They comprise two elements, namely:

1. The crew share, and

2. Fish auctioning/marketing fees.

In fishing, crews are often paid a share of the value of the landings. Although the details of the share system vary from place to place, usually some variable (effort dependent) costs, such as fuel, are deducted from the ex-vessel value prior to sharing the proceeds between the crew and the owner of the fishing vessel. A good share system provides an incentive for the crew to catch more as well as to reduce operation costs. On the other hand, auction fees are determined as a share of the ex-vessel value of the landings without prior deduction of costs.

The total harvest costs of a particular fleet is obtained as the sum over all three cost categories. Clearly, total harvest costs will not increase linearly with fishing effort because some costs depend on the value of the landings produced by that effort, which changes with the intensity of exploiting the fishery resource. The total costs of the entire harvesting sector are given as the sum of the costs of each fleet.

The costs, in particular the costs depending on the effort, are another link between the biological and economic models. The variable costs are linearly related to effort, which in turn is related to fishing mortality and thus the quantities removed from the living resources.

3.3.3 Investments, Foreign Exchange Costs, Taxes and Subsidies

The profit from fishing (value of catch minus costs of fishing) relative to the investment is an important measure of the performance of the fishing sector. The capital investments are mostly the cost of the long term assets, such as hull, engine, gear, electronic equipment, safety equipment etc.

A quantity closely related to investment is the rate of interest to be paid for the loans used to fund the investment. Another important variable is the depreciation rate, for example the percentage depreciated per year for different types of investments.

Data collection of costs will need to reflect the currencies concerned. For a developing country, each of the cost elements and items presented above may entirely or partially be incurred in foreign currency. Fuel, engines, hulls, echosounders, for example, often need to be imported and paid for in foreign exchange. Thus, parts or all depreciation allowances need to be accounted for in foreign exchange. In joint-venture fisheries, crew remuneration may occur in foreign currency because the vessels may be manned by foreign fishers. Also, interest and insurance payments may need to be settled in a foreign currency.

Taxes are payments from private investors/parties to the government. Subsidies are negative taxes, i.e. payments from government to private individuals or companies. Taxes and subsidies apply mainly to fuel and lubricants, repair and maintenance. Taxes will increase the costs of these inputs to the entrepreneur while subsidies will reduce them. In both cases, the real cost to the economy of using these inputs is distorted and needs to be adjusted for determining the optimum management regime.

3.4 AN EXAMPLE FROM A TROPICAL FISHERY

Below follows a brief description of the fisheries sector in a “representative tropical country” assumed in the following chapters, which introduces the methodology of data collection. In particular Chapter 5, which introduces a set of data collection forms, is consistent only in the context of this fishery.

The example is constructed to reflect the situation in a tropical country with a large artisanal marine fisheries sector and some industrial or semi-industrial fleets. Logbooks and complete enumeration through the vessel register are only available for the industrial fishers.

Most of the concepts introduced below have already been introduced, so the description given here has the form of a summary. The only new aspect introduced in this chapter is the methodology suggested for estimation of discards (i.e. catch minus landings) from limited information

3.4.1 Data Types Collected

Trip-Interview: Catch and effort data (and other trip-related data) are collected by enumerators on a “per fishing trip” basis, usually at the time of landing (interview), although the actual collection of data also involves biological measurements by the enumerator.

Vessel Registration: Vessels are assumed to be registered in a “Home port”, which may or may not be the same as the “Fishing base port”, which may or may not be the landing place. The registration details are stored in a “vessel register”. The vessel register is intended to be a complete enumeration of all vessels, although gaps and errors may occur.

Landings: Landings (=Catches - discards) may be occur in one or more landing places. For example, shrimps may be landed in place A, whereas fish are landed in place B. Landings may also be transferred to collector vessels or other fishing vessels at sea. The fishing vessels may fish in groups on distant fishing grounds. They may fish in groups for security and fish finding purposes, but also to reduce the cost of landings. One or two vessels may land the catch of all vessels, while the remaining vessels stay at the fishing grounds.

Discards: The fishers do not generally record discards, but they may be able to give some indication of the discard amounts and species composition. For example, a trawler may keep the total catch of the last haul, in which case it is possible to estimate the entire discard, on the assumption the last haul is a representative sample of all hauls made. During the interview the fishers will be requested to give the number of gear operations (say, trawl hauls) with discarding and the number without discarding.

Commercial Groups: Landings are recorded by weight of commercial groups. A “commercial group” may comprise several species, be a single species or a size group within a species. The basic raw data are the weight and the price per kilogram of the commercial group. Examples of commercial groups are: “Mixed small demersal fish”, “Mixed threadfin breams”, “Middle-sized groupers (usually only one species)”, 10–15 peeled shrimp tails per lb.”, “Middle-sized dried Loligo chinensis”. The definition of commercial groups are not under the control of the data managers, they are defined by the demands of the market.

Species Composition: Samples may be taken from commercial groups with more than one species for the estimation of species composition. Sub-samples may furthermore be taken from a species for estimation of, for example, length frequency. In the case of a typical tropical country, the number of species observed may be large. However, only the important species will be recorded, and the non-important species are lumped into “other species”.

“Biological Data”: These may include several types of data, such as species composition, maturity stages, meristic characters, age distribution, sex distribution, length composition data etc. In the present example, only four types of biological data are assumed to be collected. The data in question are:

• Species composition of commercial groups;

• Length frequency data;

• Body weight by length group for selected species;

• Sex distribution (notably of length frequency data of shrimps).

The number of commercially important species is assumed to be so large that resources are only available to collect biological data for a set of selected “representative” species (see Section 4.3.2.7).

Fishing Activities: Fishing activities, in the form of effort per unit time (e.g. number of fishing days per month) is assumed to be sampled. These data may be collected as a part of the interview, but other authorities than those responsible for the catch and effort data may also collect activity data (e.g. a coast guard under the Ministry for Defence). Recording of fishing activity may or may not be done by complete enumeration.

Migrations of Fishing Vessels: Migration of fishing vessels refers to vessels that do not use the homeport (the port of registration) as the base port of fishing. A port authority or a coast guard may record vessel migrations. Migration data may also be obtained from the interviews together with the catch and effort data. Recording of migration of fishing vessels may or may not be done by complete enumeration.

Costs and Earnings: During the interview, the enumerator also collects costs and earnings data. The data will either (1) be recorded during the weighing of landings by commercial group or (2) be copied from the record-book of the skipper or vessel owner. The enumerator may also collect data already available from the vessel register for the purpose of validation.

Province, District, Village and Site: It is assumed that the country is divided into a number of administrative units (“Province”, “District”, “Village” and “Site”).

Fishing Fleets: The fishing vessels are grouped into “fleets”.

“Fish Stocks” and “Management Units”: It is not assumed that conventional fish stocks have been identified. Species (and sometimes sex) represents the lowest level of division of the living resources. The definition of stocks in the case of tropical systems is usually rather problematic, and some division of the resources based on fishing grounds is usually the only practical option. Thus, fish (shrimps, cephalopods etc.) caught at one or more pre-defined fishing grounds are assumed to belong to a “stock”. This definition of stocks, will not live up to the more strict definitions and therefore is referred to under the broader term “management unit” (Section 2.3).

Spatial Information: When interviewed, fishers will usually be able to identify the fishing grounds. This may, for example, be “East of island X” or other name of the fishing ground. They may sometimes be able to give a more exact position in terms of latitudes and longitudes. Fishing grounds can be defined by statistical rectangles (say, 30 by 30 nautical miles) or finer divisions.

Frame Survey: The frame survey is an inventory list of fishing vessels structured by homeport (port of registration). The frame survey provides the number of vessels in each fleet of each homeport. The frame survey is not required if the vessel register and the data on vessel migration are reliable. However, in the present example, these data are assumed to be questionable, so the frame survey is therefore also used for validation of the vessel register.

3.4.2 Methodology for Estimation of Total Catches

Total landings at one place over some period (e.g. one month) can be estimated by:

C = TNV × FV × AC × CPUE

where

The procedure of estimating total landings will be further elaborated in Section 4.4

3.4.3 Estimation of Discards from Limited Data

The most accurate, but also most costly, method for estimating discards is to place observers (enumerators) onboard a representative selection of fishing vessels, and then let the observers record the total catch (landings and discards) and let them take samples from the discards. The use of observers may have many more objectives than collecting discard data, and in general, they would be required to monitor the fishing operations. Observers onboard the vessel during the fishing trip is recommended if the budget and personnel allows for this extra activity.

Using an experimental fishery to estimate discards can be recommended as a sound approach to collecting samples of detailed data. To what degree experimental fishing should be used is very dependent on the resources available to the data collection programme. Setting up an experimental fishery, however, is outside the scope of the present manual.

The table illustrates a third simpler and less dependable (amongst several alternatives) method to estimate LPUE and DPUE. The estimates are based on a sub-sample of hauls of a trip where the discards are retained for examination. These samples allow CPUE, including discards, to be estimated.

Table 3.4.1 An example taken from a trawl fishery of a method to estimate LPUE and DPUE. The table assumes that catch can be divided into two groups: high-value species (or size) and low-value species (or size). The high-value species are never discarded, whereas the low-value species are discarded except for the last two trawl hauls. In case the low-value species are always discarded, an alternative approach must be applied, for example, by placing observers onboard the vessel, or by making the fishers collect samples of the discards.