Data Availability and
Analytic Workshop for Trawl Survey Experiment 28 October - 6 November 2002


At and after the trawl warp workshop, ad hoc groups of commercial fishermen and industry representatives designated lists of potential factors which they felt could affect survey catchability, beyond just the change in trawl warp offset. There was interest in evaluating a combination of effects simultaneously to estimate an "upper bound" to the possible extent of changes due to gear configurations in the most recent years, and the potential effect of that suite of changes on survey catch rates.

An experiment was undertaken 28 October - 6 November 2002, which was designed to compare catch rates of the Albatross IV under two different gear configurations. The control configuration consisted of newly constructed nets inspected by industry, doors inspected by industry, even trawl warps, and backstraps fitted with swivels. The treatment configuration consisted of a recently-used net with a recently-implemented change in footrope seizing, a pair of doors felt by industry to have performed poorly in a previous cruise (on which they participated), trawl warps offset to match documented error, and a backstrap with no swivels intentionally twisted. The experiment was undertaken as a randomized block design, with one randomly chosen gear configuration used per 24 hour period, after which gear configurations were switched and tows repeated in order. Three different areas were selected for study by industry representatives. The F/V Sea Breeze towed beside the Albatross IV, using commercial gear. Catch per tow data were collected from the F/V Sea Breeze to index of fish availability from day to day and within day, as a covariate. A preliminary version of the data has been provided to the participating captain, and has been circulated throughout the commercial fishing community; and a preliminary report of raw data has been made at a New England Fishery Management Council meeting.

The cruise report from that experiment is available at AL0211rpt.pdf .

Availability of the Experimental Survey Data

We would like to make the analysis process as accessible and transparent as possible. To do that, we would like to make the data available to anyone who is interested in analyzing the data independently. We will then hold an analytic workshop, where all the results generated by NEFSC and any other interested individuals will be presented and discussed. We will summarize all results in a workshop report, which will then be forwarded to a group of independent peer reviewers who will be evaluating scientific advice provided to the New England Fishery Management Council in support of Amendment 13 to the Multispecies Fishery Management Plan.

Data availability: The data collected from the experiment have now been audited by hand and electronically. Data collected from the F/V Seabreeze are available at SBtrawldata.xls as a spreadsheet. Descriptions of the variables in that spreadsheet are listed in SBvariabledefinitions.pdf. Data on length composition of F/V Seabreeze catches are available at SBlengthdata.xls. All the variables from the R/V Albatross IV data set are listed in SVDBSvariabledefinitions.pdf.

Two minor errors were identified and corrected in the Sea Breeze data file on 8 Jan 2003. They were: 1) begin and end tow times were reversed on tow #32, and 2) end time on tow #52 was corrected from 1715hrs to 1615hrs.

Because of data security issues, we plan to make a separate, standard audited data set available to meet the needs of interested external analysts. To do that, we need input from the analysts in terms of which variables are of interest, and what format would be most suitable. Please respond to Frank Almeida ( , 508-495-2308) as soon as possible with contact information, the variables you are interested in, and the format that would be most accessible for you. We will then collate those variables into a single data set and forward a copy of the data to you. Alternatively, we have a spreadsheet version containing data on the following variables for each tow: Area, Block, Sub-Block, Gear Configuration, Date, Time (GMT), Time (EST), Average Depth (m), Surface Temp, Sea State, Begin Latitude, Begin Longitude, End Latitude, End Longitude, COG, Pitch, RPM, Tide, and catch in weight and number by species. The variable definitions are available in the two preceding variable description files. If this spreadsheet version will meet your needs, it is available here ALBtrawldata.xls.

The ALBtrawldata.xls file was updated on 22 January 2003 to correct the start time, begin latitude, and begin longitude of tow #12. This tow actually started 6 minutes prior to the time on the original data file.

The Analytic Workshop and Peer Review

An Analytic Workshop will be held to evaluate the significance of the changes in experimental survey trawl configurations on survey trawl catch rates 14 - 15 January 2003. A draft agenda is attached. The NEFSC plans to undertake an analysis of covariance, to determine if there is a statistically significant effect of survey trawl configuration on survey trawl catch rates, and will determine the statistical power of the test (i.e., how big an effect could exist but be undetected in the analysis). Other analyses addressing the significance of the changes in survey trawl configuration on survey trawl catch rates are also welcomed. Presentations, discussions and report contents will be limited to analyses which directly address the workshop's objective. (A planned regional advisory board consisting of industry, academic and agency representatives may indicate directions of additional experimental and analytic work in support of the initial trawl warp workshop recommendations.)

A report of the workshop will include the results of the analyses presented and discussed, a summary of consensus conclusions, and, if necessary, separate sections for different analytic approaches and their associated conclusions. Analytic presenters should forward copies of material to be circulated during the workshop by 10 January 2002 for duplication for the meeting or bring copies to the meeting (contact Wendy Gabriel 508-495-2213 or for number of copies). Analytic presenters should also bring or forward electronic files detailing their methodologies, results, supporting tables and figures, and conclusions to the workshop, so that the material can be directly incorporated into the report during the workshop.

The combined report will be forwarded to the group of independent peer reviewers. Peer reviewers from the group may also participate in the workshop itself, as schedules permit.

Draft Agenda

Workshop on Analysis of Trawl Survey Experiment
14-15 January 2003
Northeast Fisheries Science Center
Woods Hole, MA
NEFSC Aquarium Conference Room
9:00 a.m - 4:30 p.m

Objective: To analyze results of trawl survey experiment conducted 28 October - 6 November 2002 in order to determine the potential effect of a suite of changes in gear configurations (including unequal trawl warp lengths) on trawl survey catchability; to summarize the results of those analyses for peer review.

  1. Introduction

    Introduce participants

    Adopt agenda

  2. Background (Working Paper Section 1: Gabriel and Almeida; Fogarty)

    Review history of the design of the experiment: factors to be evaluated, statistical design

    Summarize field and sampling methods

    Summarize types and amount of data collected

  3. Analysis of covariance (Working Paper Section 2: Fogarty)

    Model structure

    Analytic results:

    Statistical significance of gear configuration effect

    Power of test: how large an effect could exist but be undetected given the variability in this data

    Analytic considerations: a balanced design with fewer observations or an unbalanced design with more observations? More observations give us more statistical power, but unbalanced designs are harder to analyze and interpret.

  4. Other analyses (Working Papers to be contributed by external participants)

  5. Discussion of results (Working Paper section to be drafted during meeting)

    Consensus on statistical conclusions, if possible

    Recommendations for future experiments (hypotheses, design)

Back to main page
Link disclaimer | Email webmaster | Privacy policy |     File Modified Nov 24, 2004