-
Notifications
You must be signed in to change notification settings - Fork 0
/
run_GOApollock.R
227 lines (200 loc) · 11.4 KB
/
run_GOApollock.R
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
# Install WHAM package (growth branch):
# remotes::install_github(repo = 'GiancarloMCorrea/wham', ref='growth', INSTALL_opts = c("--no-docs", "--no-multiarch", "--no-demo"))
# Set you WD:
setwd("~/GitHub/AKWHAM")
rm(list = ls())
require(dplyr)
require(ggplot2)
library(wham)
source('functions_pollock.R')
source('aux_fun.R')
# Load some files:
load('aux_data/input.RData') # input data as in original WHAM model (Cole's)
load('aux_data/cv_survey.RData') # CV for observed WAA (survey)
load('aux_data/asdrep.RData') # CV for observed WAA (survey)
# see here: https://github.com/afsc-assessments/GOApollock/blob/dev/pkwham/pkwham_test.R
# Some model parameters:
LAA_jan1 = c(8.8, 21.1, 30.4, 37.4, 42.7, 46.8, 49.8, 52.1, 53.9, 55.2) # obtained from survey data
LW_pars = c(2.847e-06, 3.261) # obtained from survey data
WAA_jan1 = round(LW_pars[1]*LAA_jan1^LW_pars[2], digits = 3)
# -------------------------------------------------------------------------
# Prepare input data for WHAM:
wham_data = list()
wham_data$ages = 1:input$data$n_ages
wham_data$lengths = seq(from = 2, to = 80, by = 2)
wham_data$years = as.integer(input$years)
wham_data$n_fleets = input$data$n_fleets
wham_data$agg_catch = input$data$agg_catch
wham_data$use_agg_catch = input$data$use_agg_catch
wham_data$catch_cv = input$data$agg_catch_sigma
# Age comps fleet:
wham_data$catch_paa = input$data$catch_paa
wham_data$catch_Neff = input$data$catch_Neff
wham_data$use_catch_paa = input$data$use_catch_paa
wham_data$selblock_pointer_fleets = input$data$selblock_pointer_fleets
wham_data$F = matrix(0.2, ncol = 1, nrow = input$data$n_years_catch)
wham_data$n_indices = input$data$n_indices
wham_data$agg_indices = input$data$agg_indices
# Age comps survey
wham_data$index_paa = input$data$index_paa
wham_data$index_Neff = input$data$index_Neff
wham_data$use_index_paa = input$data$use_index_paa
wham_data$units_indices = input$data$units_indices
wham_data$use_indices = input$data$use_indices
wham_data$units_index_paa = input$data$units_index_paa
wham_data$selblock_pointer_indices = input$data$selblock_pointer_indices
wham_data$fracyr_indices = input$data$fracyr_indices
wham_data$waa = input$data$waa
wham_data$waa_pointer_indices = input$data$waa_pointer_indices
wham_data$waa_pointer_fleets = input$data$waa_pointer_fleets
wham_data$waa_pointer_totcatch = input$data$waa_pointer_totcatch
wham_data$waa_pointer_ssb = input$data$waa_pointer_ssb
wham_data$waa_pointer_jan1 = input$data$waa_pointer_jan1
wham_data$maturity = input$data$mature
wham_data$fracyr_SSB = input$data$fracyr_SSB
wham_data$Fbar_ages = input$data$Fbar_ages
wham_data$percentSPR = input$data$percentSPR
wham_data$percentFXSPR = 100
wham_data$percentFMSY = 100
wham_data$XSPR_R_avg_yrs = 1:length(input$years)
wham_data$XSPR_R_opt = input$data$XSPR_R_opt
wham_data$simulate_period = input$data$simulate_period
wham_data$bias_correct_process = 1
wham_data$bias_correct_observation = 1
# -------------------------------------------------------------------------
# Empirical WAA
input_a = prepare_wham_input(model_name="pollock_a",
selectivity=list(model = c('double-logistic', 'age-specific', 'double-logistic',
'double-logistic', 'age-specific', 'age-specific',
'double-logistic'),
re = c('iid', 'none', 'none', 'none', 'none', 'none', 'none'),
initial_pars=list(c(4, 1, 20, 0.36), rep(1, 10),
c(4, 1, 20, 0.36),
c(4, 1, 20, 0.36), rep(1, 10), rep(1, 10),
c(4, 1, 20, 0.36)),
fix_pars = list(NULL, 1:2, 3:4, 3:4, 1:10, 1:10, 1:4),
n_selblocks = 7),
M = list(model = 'age-specific', re = 'none',
initial_means = exp(input$par$M_a)),
NAA_re = list(sigma="rec", cor = 'iid', N1_model = 1,
recruit_model = 2,
N1_pars = exp(input$par$log_N1_pars),
recruit_pars = exp(input$par$mean_rec_pars)),
catchability = list(re = c('ar1', 'none', 'ar1', 'none', 'none', 'none'),
initial_q = rep(1, 6), q_lower = rep(0, 6),
q_upper = rep(1000, 6), prior_sd = rep(NA, 6)), #DO NOT CHANGE UPPER Q!
basic_info = wham_data)
# update some inputs as base WHAM model:
input_a = post_input_pollock(input_a, input)
# no random effects:
input_a$random <- c("log_NAA", "q_re")
# Fix survey selex for age 3,4, and 8 to reach convergence:
tmp = matrix(input_a$map$logit_selpars, nrow = 7)
tmp[2,c(3:4,8)] = NA
input_a$map$logit_selpars = factor(tmp)
#Run model:
fit_a = fit_wham(input_a, do.osa = FALSE, do.fit = TRUE, do.retro = FALSE)
check_convergence(fit_a)
save(fit_a, file = 'GOA_pollock/fit_a.RData')
# Make plots
dir.create(path = 'GOA_pollock/fit_a')
plot_wham_output(mod = fit_a, dir.main = 'GOA_pollock/fit_a', out.type = 'pdf')
# Make projections:
proj_a = project_wham(model = fit_a, MakeADFun.silent = TRUE)
save(proj_a, file = 'GOA_pollock/proj_a.RData')
# -------------------------------------------------------------------------
# iid WAA
# Add CV and turnon likelihood for waa:
cvsurvey = as.matrix(out_data)
cvsurvey[is.na(cvsurvey)] = 0
wham_data$waa_cv = array(0, dim = dim(wham_data$waa))
wham_data$waa_cv[2,,] = cvsurvey
wham_data$use_index_waa = matrix(0L, ncol = wham_data$n_indices, nrow = length(wham_data$years))
wham_data$use_index_waa[17:52,1] = 1L
input_b = prepare_wham_input(model_name = "pollock_b",
selectivity = list(model = c('double-logistic', 'age-specific', 'double-logistic',
'double-logistic', 'age-specific', 'age-specific',
'double-logistic'),
re = c('iid', 'none', 'none', 'none', 'none', 'none', 'none'),
initial_pars=list(c(4, 1, 20, 0.36), rep(1, 10),
c(4, 1, 20, 0.36),
c(4, 1, 20, 0.36), rep(1, 10), rep(1, 10),
c(4, 1, 20, 0.36)),
fix_pars = list(NULL, 1:2, 3:4, 3:4, 1:10, 1:10, 1:4),
n_selblocks = 7),
M = list(model = 'age-specific', re = 'none',
initial_means = exp(input$par$M_a)),
NAA_re = list(sigma="rec", cor = 'iid', N1_model = 1,
recruit_model = 2,
N1_pars = exp(input$par$log_N1_pars),
recruit_pars = exp(input$par$mean_rec_pars)),
WAA = list(WAA_vals = WAA_jan1,
re = c('iid'),
est_pars = 1:input$data$n_ages),
catchability = list(re = c('ar1', 'none', 'ar1', 'none', 'none', 'none'),
initial_q = rep(1, 6), q_lower = rep(0, 6),
q_upper = rep(1000, 6), prior_sd = rep(NA, 6)),
basic_info = wham_data)
# update some inputs:
input_b = post_input_pollock(input_b, input)
# random WAA only
input_b$random <- c('WAA_re', "log_NAA", "q_re")
# Fix survey selex for age 3,4, and 8 to reach convergence:
tmp = matrix(input_b$map$logit_selpars, nrow = 7)
tmp[2,c(3:4,8)] = NA
input_b$map$logit_selpars = factor(tmp)
#Run model:
fit_b = fit_wham(input_b, do.osa = FALSE, do.fit = TRUE, do.retro = FALSE)
check_convergence(fit_b)
save(fit_b, file = 'GOA_pollock/fit_b.RData')
# Make plots
dir.create(path = 'GOA_pollock/fit_b')
plot_wham_output(mod = fit_b, dir.main = 'GOA_pollock/fit_b', out.type = 'pdf')
# Make projections:
proj_b = project_wham(model = fit_b, MakeADFun.silent = TRUE)
save(proj_b, file = 'GOA_pollock/proj_b.RData')
# -------------------------------------------------------------------------
# 2DAR1 WAA
# use same input data created in iid WAA
input_c = prepare_wham_input(model_name = "pollock_c",
selectivity = list(model = c('double-logistic', 'age-specific', 'double-logistic',
'double-logistic', 'age-specific', 'age-specific',
'double-logistic'),
re = c('iid', 'none', 'none', 'none', 'none', 'none', 'none'),
initial_pars=list(c(4, 1, 20, 0.36), rep(1, 10),
c(4, 1, 20, 0.36),
c(4, 1, 20, 0.36), rep(1, 10), rep(1, 10),
c(4, 1, 20, 0.36)),
fix_pars = list(NULL, 1:2, 3:4, 3:4, 1:10, 1:10, 1:4),
n_selblocks = 7),
M = list(model = 'age-specific', re = 'none',
initial_means = exp(input$par$M_a)),
NAA_re = list(sigma="rec", cor = 'iid', N1_model = 1,
recruit_model = 2,
N1_pars = exp(input$par$log_N1_pars),
recruit_pars = exp(input$par$mean_rec_pars)),
WAA = list(WAA_vals = WAA_jan1,
re = c('2dar1'),
est_pars = 1:input$data$n_ages),
catchability = list(re = c('ar1', 'none', 'ar1', 'none', 'none', 'none'),
initial_q = rep(1, 6), q_lower = rep(0, 6),
q_upper = rep(1000, 6), prior_sd = rep(NA, 6)),
basic_info = wham_data)
# update some inputs:
input_c = post_input_pollock(input_c, input)
# random WAA only
input_c$random <- c('WAA_re', "log_NAA", "q_re")
# Fix survey selex for age 3,4, and 8 to reach convergence:
tmp = matrix(input_c$map$logit_selpars, nrow = 7)
tmp[2,c(3:4,8)] = NA
input_c$map$logit_selpars = factor(tmp)
#Run model:
fit_c = fit_wham(input_c, do.osa = FALSE, do.fit = TRUE, do.retro = FALSE)
check_convergence(fit_c)
save(fit_c, file = 'GOA_pollock/fit_c.RData')
# Make plots
dir.create(path = 'GOA_pollock/fit_c')
plot_wham_output(mod = fit_c, dir.main = 'GOA_pollock/fit_c', out.type = 'pdf')
# Make projections:
proj_c = project_wham(model = fit_c, MakeADFun.silent = TRUE)
save(proj_c, file = 'GOA_pollock/proj_c.RData')