際際滷

際際滷Share a Scribd company logo

Presentation at IFOMC 2018

Edvin Fuglebakk
Edvin Fuglebakk

Testing fisheries data sets from port-intercept sampling and reference fleet for approximate representativeness using model validation tools. Presented at IFOMC 2018 in Vigo, panel 5: Assessing bias from monitoring programs.

1 of 10
Download to read offline
Representativeness in Norwegian reference fleet and intercept sampling Assessing representativeness and bias in vessel selection in commercial catch sampling programs conducted in collaboration with the industry Edvin Fuglebakk, Gjert Dings淡r, Tom Clegg, H奪kon Otter奪, Jon Helge V淡lstad
Key catch-sampling programs Reference fleet (High-Seas) Intercept sampling Sampling frame Catches, self-sampled Vessel > 15m Landings Fresh fish North of 64属 N Selection 14 vessels Tender 4-6y contracts Commercial fishing trips Systematic day visits to landing sites Intercepts vessels at port arrival Gears Trawl Longline Gillnet D. seine Longline Gillnet Jig Strata Gear Gear - Season (High/Low) Hierarchy Area-Quarter - Vessel - Catch Site-day - landing
Reference fleet coverage (cod) area / quarter w 2 vessels
Reference fleet temporal and spatial Clusters: area / quarter
Validation Model validation approach Generic estimators - parameter xp can be exchanged. Estimate of parameters recorded in census: xp = catch weight xp = position xp = time Errors quantified in terms of estimated precision (confidence intervals)
Reference fleet - validation
Reference fleet - validation
Intercept sampling - validation
Conclusions Classical model validation supports use of key sampling programs. The reference fleet sample well important parameters: catch- weight, position, time an depth (within clusters). Gear not adressed Clustering trade-off with coverage. Total coverage 77% for cod.
Acknowledgement Co-authors: Gjert Dings淡r, Tom Clegg, H奪kon Otter奪, Jon Helge V淡lstad Intercept sampling crew Reference fleet

More Related Content

Presentation at IFOMC 2018

  • 1. Representativeness in Norwegian reference fleet and intercept sampling Assessing representativeness and bias in vessel selection in commercial catch sampling programs conducted in collaboration with the industry Edvin Fuglebakk, Gjert Dings淡r, Tom Clegg, H奪kon Otter奪, Jon Helge V淡lstad
  • 2. Key catch-sampling programs Reference fleet (High-Seas) Intercept sampling Sampling frame Catches, self-sampled Vessel > 15m Landings Fresh fish North of 64属 N Selection 14 vessels Tender 4-6y contracts Commercial fishing trips Systematic day visits to landing sites Intercepts vessels at port arrival Gears Trawl Longline Gillnet D. seine Longline Gillnet Jig Strata Gear Gear - Season (High/Low) Hierarchy Area-Quarter - Vessel - Catch Site-day - landing
  • 3. Reference fleet coverage (cod) area / quarter w 2 vessels
  • 4. Reference fleet temporal and spatial Clusters: area / quarter
  • 5. Validation Model validation approach Generic estimators - parameter xp can be exchanged. Estimate of parameters recorded in census: xp = catch weight xp = position xp = time Errors quantified in terms of estimated precision (confidence intervals)
  • 6. Reference fleet - validation
  • 7. Reference fleet - validation
  • 8. Intercept sampling - validation
  • 9. Conclusions Classical model validation supports use of key sampling programs. The reference fleet sample well important parameters: catch- weight, position, time an depth (within clusters). Gear not adressed Clustering trade-off with coverage. Total coverage 77% for cod.
  • 10. Acknowledgement Co-authors: Gjert Dings淡r, Tom Clegg, H奪kon Otter奪, Jon Helge V淡lstad Intercept sampling crew Reference fleet