Readability code with age comps

[jysullivan]

Currently use all ages, regardless of readability code.

[kwmcneel]

Readability codes are currently single digits between 0 and 9. Region 1 has historically only uploaded readability codes 1 through 3, but codes 4-9 should be excluded from modeling. Attached is a memo explaining the definitions of the readability codes and references of the historical codes. Readability Memo 1.6.2019.docx

[jysullivan]

Notes for future ageing error matrix: there is both training (i.e. employee's data who are in training) and production data. Contact K. McNeel for more information of reading flags from the ageing database.

[jysullivan]

@kwmcneel It appears that age readability codes 04 and 05 are being uploaded into the database. Maybe I'm using the wrong view?

Here's the query for the NSEI fishery age comps:

query <-
  paste0(" select  year, project_code, trip_no, adfg_no, vessel_name, sell_date, g_stat_area,
          g_management_area_code, sample_type, species_code, length_type_code, 
          length_type, length_millimeters / 10 as length, weight_kilograms,
          age, age_readability_code, age_readability, sex_code, 
          maturity_code, maturity, gear_code, gear

  from    out_g_bio_age_sex_size

  where   species_code = '710' and
          project_code in ('02', '17') and
          g_management_area_code = 'NSEI' and year = ", YEAR)

And here's a table of the numbers of readability codes for 2019:

Age readability codes	2019
01	25
02	622
03	587
04	55
05	17

[jysullivan]

@kwmcneel Once I filtered out the age readability codes > 3 here's the data I have for the NSEI fishery age comps:

Age readability code	2001	2002	2003	2004	2005	2006	2007	2008	2009	2010	2011	2012	2013	2014	2015	2016	2017	2018	2019
1	0	7	9	55	34	9	43	0	197	211	15	96	133	103	29	111	58	8	25
2	0	450	353	729	664	370	952	631	894	701	1049	850	692	912	858	834	906	651	622
3	1	1736	1357	1512	1037	1302	504	905	469	621	406	608	471	304	315	482	430	708	587

[jysullivan]

@kwmcneel These are the sample sizes I have for age readability codes > 3 for the NSEI longline survey:

Age readability code	1988	1989	1990	1991	1992	1993	1994	1995	1996	1997	1998	1999	2000	2001	2002	2003	2004	2005	2006	2007	2008	2009	2010	2011	2012	2013	2014	2015	2016	2017	2018	2019
1	19	102	9	16	37	31	22	43	17	84	35	3	0	1	3	0	0	23	29	0	1	8	16	49	20	40	24	9	109	44	21	19
2	241	99	198	262	160	374	279	188	258	440	278	103	94	47	74	10	76	264	576	215	234	200	239	427	350	236	448	352	296	436	339	268
3	22	15	11	8	11	28	33	13	32	48	42	228	311	648	581	774	666	589	133	560	397	411	350	200	356	260	90	136	156	117	212	182

The associated query updated to exclude bad age readability codes:

query <-
  paste0(" select  year, project_code, trip_no, target_species_code, adfg_no, vessel_name, 
          time_first_buoy_onboard, number_of_stations, hooks_per_set, hook_size, 
          hook_spacing_inches, sample_freq, last_skate_sampled, effort_no, station_no, species_code, 
          g_stat_area as stat, start_latitude_decimal_degrees as start_lat,
          start_longitude_decimal_degree as start_lon, end_latitude_decimal_degrees as end_lat,
          end_longitude_decimal_degrees as end_lon, avg_depth_fathoms * 1.8288 as depth_meters, 
          length_millimeters / 10 as length, weight_kilograms as weight, 
          age, age_type_code, age_readability_code, sex_code, maturity_code, otolith_condition_code

  from    output.out_g_sur_longline_specimen

  where   species_code = '710' and
          age_readability_code in ('01', '02', '03') and
          project_code in ('603', '03') ") 

[jysullivan]

Hi Jane,

These numbers closely match our database. I think Mike used to go through and personally delete ages with readabilities >3, but 4s and 5s have always existed. I can work with James and try to get the recent higher codes erased from your database if you’d like.

Thanks, Kevin Jan 24, 2020

[jysullivan]

@kwmcneel Excellent! Let's hold off on deleting stuff from the database for now. These high codes are now filtered out in the query, I've added significant documentation, and I plan on including your memo as an appendix in our assessment this year. Hopefully this will sufficiently ward off future mistakes. Thanks again!