8. FRAME SURVEYS

8.1 OBJECTIVES
8.2 SITES AND BOAT/GEAR CLASSIFICATIONS
8.3 SEASONAL/SEQUENTIAL VARIATION OF FISHING GEAR
8.4 CONCURRENT USE OF FISHING GEAR
8.5 FORMS FOR DATA COLLECTION
8.6 BRIEFING OF DATA COLLECTORS
8.7 SURVEY TESTING AND IMPLEMENTATION
8.8 FRAME SURVEY DATA SUMMARIES
8.9 SUMMARIES WITH GROUPED HOMEPORTS
8.10 ABSOLUTE AND RELATIVE ACCURACY

This section completes the description of Frame Surveys by reviewing the objectives of frame surveys and discussing survey preparation, implementation and application in the process of estimating fishing effort.

8.1 OBJECTIVES

A Frame Survey is a census-based approach in which data is collected on all fishing vessels and gear (at all homeports/fishing sites), which could be potentially operating within the estimation context or stratum,

Usually, Frame Surveys also provide the opportunity for recording supplementary information useful for planning and implementation purposes, such as fishing trip patterns and seasonal use of fishing gear. They can also be used to provide information on the socio-economics and demography of fishing communities.

8.2 SITES AND BOAT/GEAR CLASSIFICATIONS

Prior to implementing a Frame Survey users must decide on a general framework regarding the homeports (hence the geographical strata) and the boat/gear types to be covered. Therefore, a list of known homeports and a first attempt to set-up a standard boat/gear classification must precede data collection.

During, or as a result of, a Frame Survey, information may be collected that will demonstrate a need to include more homeports than previous surveys, or to exclude homeports that are no longer relevant. Similarly, fleet and operational changes since a previous survey will indicate a need to include new boat/gear categories or to group some categories together.

The immediate task after a frame survey has been completed is to finalize the list of homeports and boat/gear types as survey standards, which should then be used as the basis for the conduct of other surveys, including Boat Activity surveys, Active Days Surveys and Landings Surveys.

8.3 SEASONAL/SEQUENTIAL VARIATION OF FISHING GEAR

8.3.1 Multiple gears used sequentially

In determining total numbers of fishing craft at a homeport, it is often observed that fishers use one type of fishing gear during one fishing season and a different one during another season. Generally, multiple use of fishing gear may not depend on the season but are employed according to circumstances, that is one day fishing with gear A, next with gear B, etc., but not simultaneously. Such multiple use of gear is conventionally described as sequential or seasonal, meaning that although the same fishing unit uses different gear types such use is strictly not concurrent.

8.3.2 Recording of boats with gears used sequentially

In such cases the boats ought to be recorded as many times as the number of the different gears used sequentially. This will not result in double counting because each estimation process operates within a fixed context of a minor stratum, a month and a specific boat/gear type.

Example: Assume that at a homeport there exist 20 boats operating gillnets and 10 that operate traps. Of the 20 gillnetters, 5 also operate traps but never together with gillnets.

The record of the total number of boats at this homeport would thus be:

Boats operating gillnets:	20
Boats operating traps:	10 + 5 = 15

8.4 CONCURRENT USE OF FISHING GEAR

Sometimes fishing boats use two or more gears simultaneously. In such cases it is usually not possible to estimate the proportion of catch that has resulted from each gear separately, unless the different gears are targeting completely different species. For example, one boat might gillnet for small sardines while also setting lines for tunas thus enabling a statistical separation of catch and effort for these species/gear combinations. However, in another case, a boat might carry both trammel nets and traps for lobster and crab.

The problem of concurrent use of fishing gear (for similar species or species groups) cannot be solved statistically in a satisfactory manner. In practice the following methods of recording might be used:

• The predominant gear is used to describe the boat/gear type.
• Setting-up a new combination boat/gear category describing the combined use of different gears (i.e. Gillnet + Hook & Line).
• Describing all boats with such use of gear as “other”.

8.5 FORMS FOR DATA COLLECTION

There are several ways for recording frame survey data, depending on the data required for the census. For numbers of boats, separate forms (one per homeport) are used containing the following information:

• Name of homeport.
• Date of recording.
• Name of data collector.
• As many records as the boats found at the homeport, with an indication of all gears used concurrently. For sequential use of gears boat records are repeated.
• Remarks concerning fishing trip patterns and other information useful in the subsequent planning of sampling operations.

The following model form may be used as an example.

Table 8.5 Example of a form for the recording of frame survey data

Statistical monitoring of small-scale marine fisheries - Frame survey
Date: 05/03/2001 Homeport: Old Harbour (SW Coast) Recorder: John Ovusu
Fishing Unit	Gillnet	Hook & Line	Traps	Castnets	Other
A	X
B	X
C	X	X
D	X	X
E	X
E			X
E				X
F	X
F			X
F				X
G			X
H			X
I			X
J				X
K				X
L				X
M		X
TOTALS	4	1	5	5
Gillnet + Hook & Line	2
Remarks: All boats except those using traps land between 08:00 and 11:00 Boats with traps land between 14:00 and 16:00

Notes on form:

• Boats C and D make concurrent use of Gillnets and Hook & Line. A new boat/gear type describing this combined use of gear is created since such cases are common for several homeports.
• The total number of boats using Gillnets is 4 and not 6, since 2 units were already included in the combined boat/gear type. Likewise the Hook & Line boats will be 1 and not 3.
• Boats E and F were repeated to show sequential use of three different gears. In total there are 5 boats using Traps and 5 using Castnets.
• An ‘Other’ column is reserved for unforeseen boat/gear types.
• Remarks were included indicating normal landing times.
• Forms can be configured to show only those boat/gear types that are relevant to minor or major strata, thus simplifying the use of forms by the data collectors.

8.6 BRIEFING OF DATA COLLECTORS

By definition frame surveys are census-based and may often demand a large number of data collectors, who may be employed on a temporary basis to supplement the activities of regular staff. Providing data collectors with precise and unambiguous instructions is fundamental for the reliability of the data obtained through Frame Surveys.

The following points are important in briefing data collectors:

• Explaining in full detail the recording forms, their use in data collection and their purpose and utility in the overall sampling programme.
• Clarifying the nature of boat/gear types, sequential and concurrent gear use, and whether new boat/gear types are significant or can be ignored, etc.
• Planning the visits to homeports.
• Methods for approaching and interviewing fishers and village authorities to obtain complete and reliable information on numbers of boat/gear types operating from homeports.
• Ways of cross-checking the obtained information and the appropriate course of action in cases of serious discrepancies.

8.7 SURVEY TESTING AND IMPLEMENTATION

Implementation of successful Frame Surveys needs careful planning because they are costly and will form the basis (possibly for several years) for many statistical estimations and for planning other surveys. The process of planning includes:

Pilot phase: small scale testing (e.g. within one or two homeports) to identify possible design and operational drawbacks and to assess likely timeframes for full-scale operations.

Test the entire sampling programme: combine the Frame Survey (including revisions) with pilot implementation of the entire sampling programme (normally 6-12 months) to ensure both meet their requirements.

Evaluation and revision phase: forms and classifications are revised and new instructions issued to data collectors.

8.8 FRAME SURVEY DATA SUMMARIES

Frame Survey results are summarized prior to their use in supporting surveys, as in table 8.8 below.

Table 8.8 Frame Survey -summaries by homeport and boat/gear type

Minor Stratum	Home port	Gill net	Beach Seine	Hand line	Trap	GN+ HL	TP+ HL
		Single gear				Multigear
SW Coast	Old Harbour	14	3	-	-	2	5
	Montagu	6	-	3	-	1	2
	Long Beach	10	-	4	-	-	3
	Pirates’ Hide	5	2	6	-	-	5
	Fishbone	10	-	2	-	-	10
	West Arm	30	-	-	-	8	-
	Mousetrap	15	-	-	-	1	-
Sub-total		90	5	15	-	12	25
SE Coast	New Harbour	-	5	-	20	-	-
	Airport	-	10	-	10	-	-
	Blue Village	-	4	-	30	-	-
	Windy Beach	-	6	-	40	-	-
	White Sands	-	-	-	15	-	-
	Coral	-	-	-	5	-	-
	Paradise	-	-	-	25	-	-
	Cactus	-	-	-	5	-	-
	Joseph’s Cave	-	-	-	30	-	-
Sub-total		-	25	-	180	-	-
TOTALS		90	30	15	180	12	25

Notes on the summary form:

• The summary form contains all boat/gear types found in the entire statistical area (all strata) and at all homeports.
• Each sub-total at minor stratum level provides the raising factor F for estimating total fishing effort within the context of that minor stratum, calendar month and boat/gear type.
• Frame survey results are “static”, that is they refer only to the period when the frame survey was conducted.
• Totals that refer to the entire statistical area are for information purposes only, not for the estimation processes resulting from other surveys.

8.9 SUMMARIES WITH GROUPED HOMEPORTS

Table 8.8 provides individual homeport information so that a sampling scheme (e.g. Boat Activity Survey) may change sampling locations and still have ‘frame’ information to do the appropriate raising at each site. However, sometimes operational and logistics constraints may mean that surveys need to be conducted at fixed, pre-selected homeports and used as representative of a minor stratum. This is undertaken by grouping homeport information and producing a frame summary that represents pre-selected homeports as in Table 8.9.

Table 8.9 Frame survey - summaries with grouped homeports

Minor Stratum	Home port	Gill net	Beach Seine	Hand line	Trap	GN+ HL	TP+ HL
		Single gear				Multigear
SW Coast	Old Harbour	14	3	-	-	2	5
	Pirates’ Hide	5	2	6	-	-	5
	UNSAMPLED	71	-	9	-	10	15
Sub-total		90	5	15	-	12	25
SE Coast	Airport	-	10	-	10	-	-
	Blue Village	-	4	-	30	-	-
	Windy Beach	-	6	-	40	-	-
	UNSAMPLED	-	5	-	100	-	-
Sub-total		-	25	-	180	-	-
TOTALS		90	30	15	180	12	25

Notes on form:

• Sampling sites are fixed at SW Coast (Old Harbour and Pirates’ Cave) and SE Coast (Airport, Blue Village and Windy Beach).
• In each minor stratum homeports that are not used in sampling are grouped under the name “UNSAMPLED” so that sub-totals at the minor stratum level are maintained.
• This approach is useful when there are long lists of homeports and only a few are used for sampling on a fixed-site basis.

8.10 ABSOLUTE AND RELATIVE ACCURACY

The weakness of frame surveys is that they provide “static” information that it is valid only at the time of the survey - a “snapshot” of the fishery. Any significant changes to the fisheries, such as increases or decreases of fishing boats, introduction of new fishing gear, etc will affect the use of frame survey results as raising factors for estimating total fishing effort.

Since,

Effort = BAC × F × A

Where: BAC is the Boat Activity Coefficient, F the total number of boats in the minor stratum provided by a frame survey, and A is a time raising factor,

if F is outdated and there are more boats operating, fishing effort will be underestimated, despite the accuracy with which BAC and A have been formulated. Similarly, if there has been a decrease in the number of boats since the last frame survey, total effort will be overestimated.

Some of these problems can be overcome by introducing calculations that measure the “relative’ accuracy of frame surveys. When the Boat Activity survey is conducted in such a manner that homeports are sampled with the same frequency, then the impact of outdated frame survey data is much reduced.

This is best illustrated with a numerical example.

Assume a minor stratum with two sampling homeports A and B. The last frame survey reported that:

• Number of trawlers in homeport A: 10.
• Number of trawlers in homeport B: 20.
• Total number of trawlers in minor stratum: 300

BAC was formulated during three selected sampling days as follows:

• Days 8, 13, 22 for homeport A
• Days 7, 11, 29 for homeport B

The method used was to observe all boats that have been active (= fishing) on the selected days and compare these numbers to those assumed by the Frame Survey.

Following is the summary of the results:

Table 8.10 Comparison of Results of Frame and Boat Activity Surveys

Sampling days	7	8	11	13	22	29
Homeport A
Frame data		10		10	10
Active		7		3	5
Homeport B
Frame data	20		20			20
Active	4		6			20

Based on the total number of active boats and the number of boats assumed to be present at each site, the standard approach for formulating BAC is as follows:

BAC =[(7+3+5)+(4+6+20)]/[(10+10+10)+(20+20+20)] = 45/90 = 0.5.

BAC = [(7+3+5)+(4+6+20)]/(3 × 30),

Thus, fishing effort will be estimated as:

Effort = 45/3 × (F/30) × A = 45/3 × (300/30) × A.

In other words the reliability of the estimate now depends on the ratio 300/30. Compared to the absolute accuracy of F discussed earlier, this ratio expresses the relative accuracy of a frame survey, and is likely to be more resistant to overall increases or decreases in boat numbers.

This approach requires that:

• All active boats must be counted.
• Active boats must be compared to their total number assumed by the frame survey.
• Homeports must be visited with the same frequency (same number of days).

SUMMARY In this section frame surveys were reviewed, including: Objectives of frame surveys. Lists of homeports and boat/gear classifications. Methods for the recording of boats that use multiple gears. Examples of forms for the recording of primary data and for preparing frame survey summaries. Implementation aspects of frame surveys. Simplified frame survey summaries with grouped sites. Absolute and relative accuracy in frame surveys.