stock synthesis tutorial - capamresearch.org
TRANSCRIPT
![Page 1: Stock Synthesis Tutorial - capamresearch.org](https://reader034.vdocuments.us/reader034/viewer/2022050214/626e71a357a4b362912a390a/html5/thumbnails/1.jpg)
Stock Synthesis Tutorial
CAPAM Spatial Assessment Models Workshop Oct. 01, 2018, La Jolla, CA
![Page 2: Stock Synthesis Tutorial - capamresearch.org](https://reader034.vdocuments.us/reader034/viewer/2022050214/626e71a357a4b362912a390a/html5/thumbnails/2.jpg)
From: Mark Maunder To: Juan Valero Subject: SS spatial models Juan. I have tried everyone to teach a SS course on the Monday morning of the CAPAM workshop. Spatial model, movement, and integrated tagging. But nobody can do it. Do you have experience and are you willing to do it? Mark IATTC
![Page 3: Stock Synthesis Tutorial - capamresearch.org](https://reader034.vdocuments.us/reader034/viewer/2022050214/626e71a357a4b362912a390a/html5/thumbnails/3.jpg)
From: Juan Valero To: Mark Maunder Subject: Re: SS spatial models
Mark,
Yes, it should work. I have some materials and experience on spatial models and movement, less so re: use of integrated tagging in stock assessment models ... We should be further ahead on the spatial BET by then so that would be useful too. Juan
![Page 4: Stock Synthesis Tutorial - capamresearch.org](https://reader034.vdocuments.us/reader034/viewer/2022050214/626e71a357a4b362912a390a/html5/thumbnails/4.jpg)
Stock Synthesis spatial options
• Any number of areas (limited by memory, speed, data…)
• Fleets are all area-specific
• Design matrix for movement between areas for each season, growth pattern
• Global recruitment partitioned to areas
• Uses generalized framework for parameters in SS (priors, time-varying options, etc.)
![Page 5: Stock Synthesis Tutorial - capamresearch.org](https://reader034.vdocuments.us/reader034/viewer/2022050214/626e71a357a4b362912a390a/html5/thumbnails/5.jpg)
Movement in Stock Synthesis
• Box transfer with no explicit adjacency of areas, so fish can move from any area to any other area in each time step.
Area 1 Area 2 Area 3 Area 4 Time =1
Area 1 Area 2 Area 3 Area 4 Time =2
![Page 6: Stock Synthesis Tutorial - capamresearch.org](https://reader034.vdocuments.us/reader034/viewer/2022050214/626e71a357a4b362912a390a/html5/thumbnails/6.jpg)
• Two parameters per movement definition to allow separate rates for young (A) and old (B) fish, with ramp in between (linear in log space)
• For each source area the implicit movement parameter value is 0.0
• Exponentiated and scaled to sum to 1.0:
1
rate ji
Npp
i
j
e e 1
1
ratelog if 0
ratei
ip p
Movement in Stock Synthesis
![Page 7: Stock Synthesis Tutorial - capamresearch.org](https://reader034.vdocuments.us/reader034/viewer/2022050214/626e71a357a4b362912a390a/html5/thumbnails/7.jpg)
• Two parameters per movement definition to allow separate rates for young (A) and old (B) fish, with ramp in between (linear in log space)
Movement in Stock Synthesis
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
0 5 10 15 20 25 30 35 40
Age
Mo
ve
me
nt
rate
area 1 to area 1
area 1 to area 2
area 1 to area 3
area 1 to area 4
B = 25 A = 3
A
B
![Page 8: Stock Synthesis Tutorial - capamresearch.org](https://reader034.vdocuments.us/reader034/viewer/2022050214/626e71a357a4b362912a390a/html5/thumbnails/8.jpg)
Movement and Areas in SS
• A fleet operates in only one area – but a fleet in one area can mirror a fleet in another area
• Movement at end of each season – fraction of fish in area A that move to area B
• The equilibrium calculator starts by distributing the total recruits to areas, settlement events, and morphs
![Page 9: Stock Synthesis Tutorial - capamresearch.org](https://reader034.vdocuments.us/reader034/viewer/2022050214/626e71a357a4b362912a390a/html5/thumbnails/9.jpg)
Movement and Areas in SS (biology)
• Biology is assigned to morphs, not areas. So when fish move they retain same biological characteristics
– it would be feasible to have area-specific M, but might be challenging to retrofit now
– it would be feasible to have area-specific growth such that the mean size of fish in an area are weighted mean of the fish that stay in that area and the size of the fish entering that area. Lots of coding to do this, but seems feasible.
![Page 10: Stock Synthesis Tutorial - capamresearch.org](https://reader034.vdocuments.us/reader034/viewer/2022050214/626e71a357a4b362912a390a/html5/thumbnails/10.jpg)
Movement and Areas in SS (tags)
• When a tag group is released in an area, p, at time, t, at age, a, they are distributed proportionally among all biology morphs according to the current distribution of morphs at that p, g, a,
• Then each p.g component of that tag group moves to other areas according to the movement parameters for that p,g; and in an area they experience F*selex and M along with their p,g group.
![Page 11: Stock Synthesis Tutorial - capamresearch.org](https://reader034.vdocuments.us/reader034/viewer/2022050214/626e71a357a4b362912a390a/html5/thumbnails/11.jpg)
Recruitment
• Total global recruitment calculated from all areas, seasons, and growth patterns
• Recruits assigned to these divisions according to apportionment weights (estimated or fixed, constant or time-varying)
• Tranformation of apportionment weight parameters to fractions is equivalent to calculation used for movement rates.
![Page 12: Stock Synthesis Tutorial - capamresearch.org](https://reader034.vdocuments.us/reader034/viewer/2022050214/626e71a357a4b362912a390a/html5/thumbnails/12.jpg)
Stock Synthesis DATA file setup
# TOP of DATA file
#Model Dimensions . . . #Fleet/Survey Setup . . . #Catch and Abundance Indices . . . #Discards . . . #Lenght Compositions . . . #Age Compositions
How many areas?
Area of each fleet?
![Page 13: Stock Synthesis Tutorial - capamresearch.org](https://reader034.vdocuments.us/reader034/viewer/2022050214/626e71a357a4b362912a390a/html5/thumbnails/13.jpg)
Stock Synthesis DATA file setup
# TOP OF DATA file 1971 #_StartYr 2001 #_EndYr 1 #_Nseas 12 #_months/season 2 #_Nsubseasons (even number, minimum is 2) 1 #_spawn_month 2 #_Ngenders 40 #_Nages=accumulator age 4 #_Nareas . . .
How many areas
![Page 14: Stock Synthesis Tutorial - capamresearch.org](https://reader034.vdocuments.us/reader034/viewer/2022050214/626e71a357a4b362912a390a/html5/thumbnails/14.jpg)
.
.
.
5 #_Nfleets (including surveys)
# fleet.type fishery.time area catch.units catch_mult fleetname
1 -1 1 1 0 FISHERY1 # 1
1 -1 2 1 0 FISHERY2 # 2
1 -1 3 1 0 FISHERY3 # 3
3 1 1 2 0 SURVEY1 # 4
3 1 2 2 0 SURVEY2 # 5 . . .
Stock Synthesis DATA file setup
Area of each fleet
![Page 15: Stock Synthesis Tutorial - capamresearch.org](https://reader034.vdocuments.us/reader034/viewer/2022050214/626e71a357a4b362912a390a/html5/thumbnails/15.jpg)
Stock Synthesis CONTROL file setup
# TOP of CONTROL file
#General configuration . . . #Biology options . . . #Biology parameters . . . #Productivity inputs . . . #Fishery inputs (F, q, selectivity) . . . #Data weighting
Recruitment distribution setup
Movement setup
Recruitment distribution parameters
Movement parameters
![Page 16: Stock Synthesis Tutorial - capamresearch.org](https://reader034.vdocuments.us/reader034/viewer/2022050214/626e71a357a4b362912a390a/html5/thumbnails/16.jpg)
.
.
. 3 # recr_dist_method for parameters: 2=main effects for GP, Settle timing, Area; 3=each Settle entity; 4=none, only when N_GP*Nsettle*pop==1 1 # not yet implemented; Future : Spawner-Recruitment: 1=global; 2=by area 4 # number of recruitment settlement assignments 0 # unused option #GPattern month area age (for each settlement assignment) 1 1 1 0 1 1 2 0 1 1 3 0 1 1 4 0 . . .
Stock Synthesis CONTROL file setup Recruitment distribution setup
![Page 17: Stock Synthesis Tutorial - capamresearch.org](https://reader034.vdocuments.us/reader034/viewer/2022050214/626e71a357a4b362912a390a/html5/thumbnails/17.jpg)
.
.
.
2 #_N_movement_definitions
0 # first age that moves (real age at begin of season, not integer)
# seas,GP,source_area,dest_area,minage,maxage
1 1 1 2 14 15
1 1 3 4 14 15
.
.
.
Stock Synthesis CONTROL file setup Movement setup
![Page 18: Stock Synthesis Tutorial - capamresearch.org](https://reader034.vdocuments.us/reader034/viewer/2022050214/626e71a357a4b362912a390a/html5/thumbnails/18.jpg)
.
. # Bottom of Biology parameters
# Recruitment Distribution LO, HI, INIT, PR, PR.SD, PR.type PHASE env dev mnyr mxyr devPH Block Blk_Fxn
-5 5 0.5 0.2 0.5 1 -3 0 0 0 0 0 0 0 # RecrDist_GP_1_area_1_month_1
-5 5 0.2 0.2 0.5 1 3 0 1 1971 2001 4 0 0 # RecrDist_GP_1_area_2_month_1
-5 5 0.2 0.2 0.5 1 3 0 1 1971 2001 4 0 0 # RecrDist_GP_1_area_3_month_1
-5 5 0.2 0.2 0.5 1 3 0 1 1971 2001 4 0 0 # RecrDist_GP_1_area_4_month_1 . .
Stock Synthesis CONTROL file setup Recruitment distribution parameters
![Page 19: Stock Synthesis Tutorial - capamresearch.org](https://reader034.vdocuments.us/reader034/viewer/2022050214/626e71a357a4b362912a390a/html5/thumbnails/19.jpg)
.
. # Bottom of Biology parameters
# Recruitment Distribution LO, HI, INIT, PR, PR.SD, PR.type PHASE env dev mnyr mxyr devPH Block Blk_Fxn
-5 4 -1.8 0 0.05 1 -3 0 0 0 0 0.5 0 0 # MoveParm_A_seas_1_GP_1from_1to_2
-5 4 -7.0 0 0.05 1 -3 0 0 0 0 0.5 0 0 # MoveParm_B_seas_1_GP_1from_1to_2
-5 4 -1.8 0 0.05 1 -3 0 0 0 0 0.5 0 0 # MoveParm_A_seas_1_GP_1from_3to_4
-5 4 -7.0 0 0.05 1 -3 0 0 0 0 0.5 0 0 # MoveParm_B_seas_1_GP_1from_3to_4
.
.
.
Stock Synthesis CONTROL file setup Movement parameters
![Page 20: Stock Synthesis Tutorial - capamresearch.org](https://reader034.vdocuments.us/reader034/viewer/2022050214/626e71a357a4b362912a390a/html5/thumbnails/20.jpg)
Simulation testing example making use of spatial options in Stock Synthesis
From Ian Taylor and Rick Methot
![Page 21: Stock Synthesis Tutorial - capamresearch.org](https://reader034.vdocuments.us/reader034/viewer/2022050214/626e71a357a4b362912a390a/html5/thumbnails/21.jpg)
• Simulation study to test ability of single area model with separate fleets to approximate spatial shift
• Simulate using three spatial areas (shallow, middle, deep)
• Recruitment is to shallow area
• Age-based movement to deeper zones
• Apply single area models with different selectivities for each fleet
• Compare estimated parameters and quantities of interest to simulated values
•Example: ontogenetic movement
![Page 22: Stock Synthesis Tutorial - capamresearch.org](https://reader034.vdocuments.us/reader034/viewer/2022050214/626e71a357a4b362912a390a/html5/thumbnails/22.jpg)
Simulation model configuration movement rates equilibrium age comp
catch time series bootstrap data
shallow area
middle area
deep area
![Page 23: Stock Synthesis Tutorial - capamresearch.org](https://reader034.vdocuments.us/reader034/viewer/2022050214/626e71a357a4b362912a390a/html5/thumbnails/23.jpg)
Estimation model configuration
In all cases estimated parameters are
selectivity for each fleet,
initial biomass,
recruitment deviations from 1970-2007
Fitting 3 area operating model to itself with logistic selectivity as a reference group
Fitting 1 area model with three selectivity arrangements:
Length-based (double-normal)
Age-based (double-normal)
Both length- and age-based (all double-normal)
![Page 24: Stock Synthesis Tutorial - capamresearch.org](https://reader034.vdocuments.us/reader034/viewer/2022050214/626e71a357a4b362912a390a/html5/thumbnails/24.jpg)
Example biomass trajectory
shallow area
middle area
deep area
catch in middle and deep areas
![Page 25: Stock Synthesis Tutorial - capamresearch.org](https://reader034.vdocuments.us/reader034/viewer/2022050214/626e71a357a4b362912a390a/html5/thumbnails/25.jpg)
Example simulation: numbers at age
shallow area middle area deep area
![Page 26: Stock Synthesis Tutorial - capamresearch.org](https://reader034.vdocuments.us/reader034/viewer/2022050214/626e71a357a4b362912a390a/html5/thumbnails/26.jpg)
Estimated selectivities
shallow survey length
middle survey length
deep survey length
shallow survey age
middle survey age
deep survey age
![Page 27: Stock Synthesis Tutorial - capamresearch.org](https://reader034.vdocuments.us/reader034/viewer/2022050214/626e71a357a4b362912a390a/html5/thumbnails/27.jpg)
Biomass estimates
![Page 28: Stock Synthesis Tutorial - capamresearch.org](https://reader034.vdocuments.us/reader034/viewer/2022050214/626e71a357a4b362912a390a/html5/thumbnails/28.jpg)
Age distribution revisited
Age distribution by area from unfished equilibrium population in 3-area simulation
Median estimated age-based selectivity from 1-area model with age- and length- based selectivity
shallow area
middle area
deep area
shallow survey
middle survey
deep survey
![Page 29: Stock Synthesis Tutorial - capamresearch.org](https://reader034.vdocuments.us/reader034/viewer/2022050214/626e71a357a4b362912a390a/html5/thumbnails/29.jpg)
Simulation model conclusions
• Single area model with only length- or age-based selectivity does not perform well.
• Single area model with selectivity as dome-shaped functions of both length and age shows less bias, but high variability.
• Spatial model which matches generating model performs well.