+ - 0:00:00
Notes for current slide
Notes for next slide

mscatch:
An R package for
calculating catch-at-length and age

10 October 2022

Andy Beet
Northeast Fisheries Science Center

Contributors:
Sarah Gaichas, Sean Lucey, Kiersten Curti
EDAB, numerous Population Dynamics Scientists

1 / 16

Background: Data

  • Single species catch at age models require a lot of data. In the Northeast region the assessment scientist is responsible for obtaining and wrangling the data for use in their model.

    • This is time consuming
    • Many decisions made in the data wrangling process are not uniform across species (data availability, life history)

  • Multispecies models (age and length based models) require a similar effort but for many more species.

  • Interest centers on the decision making process for wrangling catch data

Goal: Facilitate a reproducible and transparent approach for wrangling multiple species catch data while incorporating many species specific decisions

2 / 16

Background: Data wrangling process

  • Pull landings data and sampled fish length data

  • Aggregate landings, by time, gear type, market code (based on availability of sampled fish)

  • Fit length-weight relationships

  • Expand the catch to length compositions

  • Create age length key

  • Calculate numbers at age and length

3 / 16

Species specific considerations: Rules

Need to recognize differences among species.

  • How are market codes aggregated?

  • How are landing aggregated temporally? QTR, SEMESTER, ANNUAL?

  • What are the rules implemented when length samples are missing?

    • Are there a minimum number of samples required?
    • 200 mt landings for every 100 fish lengths measured?
    • Borrow length sample from which time period?
    • Nearest neighbor? previous year? previous year same semester? within last 5 years? (based on life history?)

  • Is species transient, resident, or resident in portion of the area of interest?

  • Time scales for length-weight relationships and age-length keys

4 / 16

Species specific considerations: Data

  • Sources for length-weight data/parameters?

    • From survey and/or commercial?
    • Fit own relationship or use parameters pulled from internal database.
    • How many length-weight relationships are used? QTR, SEMESTER, sex, gear type?

  • Sources for discard data?

    • Are length samples available?
    • Add to totals prior to length expansion?

  • Sources for age data?

    • From survey and/or commercial?

  • Sources of landings and length data?

    • From internal NMFS commercial fisheries database, state data, survey data

  • Sources for foreign data?

  • Stock area definition? Statistical areas?

5 / 16

mscatch as an R package

  • Address the decision making process

  • Define species specific rules

    • How to aggregate catch based on availability of length samples by time (Quarterly, Semi annually, Yearly)
    • Associated gear types with fleets
    • Market code relabeling
    • Data sources
    • Define length-weight relationship
    • Define age-length key

  • Reproducible method

https://noaa-edab.github.io/mscatch/index.html

6 / 16

A case study: Mackerel

  • All gears combined into a single gear type

  • Temporal aggregation: semesters (Jan-Jun, Jul-Dec)

  • Market codes:

    • SQ, ES, SV combined to SMALL
    • MD
    • XG, JB, LG combined into LARGE

  • Missing samples: Borrow sample from previous semester within the same market category

    • 5 year average over both semesters
    • market category time series average

  • Length-weight relationships: 6 (3 time intervals per semester)

    • Fitted to bottom trawl data (after QA/QC for anomalous values)

  • Age-length data: from Bottom trawl survey and Commercial data. Combine

  • Age-length key: By Year

  • Start date: 1992

7 / 16

Species Rules

https://noaa-edab.github.io/ms-keyrun/articles/GBLandingsByLength.html

8 / 16

Sample output: Gear

9 / 16

Sample output: Gear

9 / 16

Sample output: Gear

9 / 16

Sample output: Market codes

10 / 16

Sample output: Market codes

10 / 16

Sample output: Market codes

10 / 16

Sample output: Temporal

11 / 16

Sample output: Temporal

11 / 16

Sample output: Temporal

11 / 16

Sample output: Length-Weight relationships

12 / 16

To Do

  • Data pull from comlandr for both landings (domestic + foreign), discards, lengths

  • QA/QC code to detect and remove anomalous values

  • More flexibility in species options

  • Additional diagnostics and figures

13 / 16
14 / 16

Walkthrough: Aggregating data

15 / 16
16 / 16

Background: Data

  • Single species catch at age models require a lot of data. In the Northeast region the assessment scientist is responsible for obtaining and wrangling the data for use in their model.

    • This is time consuming
    • Many decisions made in the data wrangling process are not uniform across species (data availability, life history)

  • Multispecies models (age and length based models) require a similar effort but for many more species.

  • Interest centers on the decision making process for wrangling catch data

Goal: Facilitate a reproducible and transparent approach for wrangling multiple species catch data while incorporating many species specific decisions

2 / 16
Paused

Help

Keyboard shortcuts

↑, ←, Pg Up, k Go to previous slide
↓, →, Pg Dn, Space, j Go to next slide
Home Go to first slide
End Go to last slide
Number + Return Go to specific slide
b / m / f Toggle blackout / mirrored / fullscreen mode
c Clone slideshow
p Toggle presenter mode
t Restart the presentation timer
?, h Toggle this help
Esc Back to slideshow