
Add swept-area estimates following the FishGlob workflow
Source:R/fishglob.R
add_swept_area_fishglob.RdCompute haul-level swept-area estimates for a datras_raw / DATRASraw
object using the FishGlob methodology.
Value
A datras_raw object with two additional columns in HH:
SweptArea: swept area in \(km^2\),DoorsArea: door-swept area in \(km^2\).
Details
The function adds haul-level swept area (SweptArea) and door-swept area
(DoorsArea) to the HH table. Missing WingSpread, DoorSpread, and
Distance values are reconstructed using survey-specific prediction models
and hierarchical fallback rules.
The implementation follows the methodology used in the FishGlob workflow.
Missing WingSpread and DoorSpread values are predicted from internally
stored survey-specific models. When observed values are missing or judged
implausible, the function applies survey-specific corrections and fallback
prediction rules.
Missing towing distance is reconstructed from GroundSpeed and HaulDur.
Missing GroundSpeed values are imputed hierarchically using averages at
the following levels:
survey-year-ship,
survey-year-country,
survey-country,
global mean.
Swept area is then computed as: $$ SweptArea = Distance \times WingSpread \times 10^{-6} $$
Door-swept area is computed analogously as: $$ DoorsArea = Distance \times DoorSpread \times 10^{-6} $$
Both quantities are returned in \(km^2\), assuming Distance,
WingSpread, and DoorSpread are recorded in metres.
Some survey- and gear-specific filtering and manual corrections are applied to match the FishGlob workflow.
The original spread-estimation logic was developed by Aurore Maureaud and P. Daniël van Denderen and is available from the FishGlob code repository.
References
FishGlob workflow code: https://github.com/fishglob/FishGlob_data/blob/main/cleaning_codes/source_DATRAS_wing_doorspread.R