In noaa-garfo/discaRd: Cochran bycatch estimation

\newpage

Discard Rate Current Method Summary (J. Michael Lanning summary)

1) Rates determine by observer reported values (gear, area, etc) 2) Incomplete observed trips have missing 'hauls' prorated by observed information from that trip 3) Trips with observer get reported/calculated observed discards of that specific trip 4) Unobserved trips get discards from the rate calculated from 1) 5) QM is only interested in the summary total of discards for each trip, not subtrips. Often the interested number is a summary of trips, ie. the herring total of bycatch for an area/season or a sector's season's total of GB Cod.

6) Other others are driven by regs. Here I would place transition rates and EM methods.

I recommend separating out issues such as mismatching gear and area for the future QA data system.

And of course, any change management must be worked through proper and transparent interaction with all end users and clients including council and SFD.

Tables created and steps to date

Rates determine by observer reported values (gear, area, etc) Observed discards

make_obdbs_table_cams_v2.sql

created:

apsd.bg_obdbs_cams_mock2018
apsd.bg_obdbs_cams_mock2019
apsd.bg_obdbs_cams_mock2020

Incomplete observed trips have missing 'hauls' prorated by observed information from that trip

Prorate observed discards on unobserved hauls within a subtrip. This is done by applying a ratio of kept all on the entire trip to kept all on the unobserved hauls only

$$d_{total} = d_{observedhauls}*(1+KALL_{unobserved hauls}/KALL_{subtrip})$$

make_obdbs_prorate.sql

created:

apsd.obs_cams_prorate
- this table was made using apsd.bg_obdbs_cams_mock2018 and apsd.bg_obdbs_cams_mock2019

Trips with observer get reported/calculated observed discards of that specific trip

Match observed hauls to subtrips

explore_link3_mesh_match.sql

This step matches on AREA, GEAR and MESHGROUP (sm, lg, xlg). This is a hard match and will go awry if there is a mismatch in the data.

created:

apsd.bg_cams_catch_mock
- follows the steps layed out for mid-Atlantic discard estimation. Gear, mesh and area CASE statements should be replaced at some point with table driven code.
- Utilizes the current apportionment table: apsd.cams_apport_20201230
apsd.bg_obs_cams_tmp1
- links to dmis.d_match_obs_link and apsd.bg_cams_catch_mock
to date (r lubridate::today()), this is only using trips that have multiple subtrips. These are the only cases where the pro-ration step matters.
apsd.bg_obs_cams_tmp2 is used in the squid example and include all trips.