From 358ac983b2db83a40e093acc1eb2c2e6ebf3ea74 Mon Sep 17 00:00:00 2001 From: brianlangseth-NOAA Date: Tue, 5 Mar 2024 11:56:51 -0800 Subject: [PATCH] Add exploration runs for CA rebuilder questions *For questions from Corey leading up to March 2024 Council meeting --- code/_ca_quillback_runs.R | 73 +++++++++++++++++++++++++++++++++++++++ 1 file changed, 73 insertions(+) diff --git a/code/_ca_quillback_runs.R b/code/_ca_quillback_runs.R index cef30d7..2c0ff18 100644 --- a/code/_ca_quillback_runs.R +++ b/code/_ca_quillback_runs.R @@ -2407,3 +2407,76 @@ reb[loc+5] <- paste(tmp_val, collapse = " ") writeLines(reb, file.path(wd, "rebuilder", new_dir, "rebuild.dat")) + +#--------------------------------------------------------------------------------------------# + +## +#The following three test runs were done to explore things brought up before 2024 March Council meeting +## + +#Test model run to figure out why time to recovery differs so much between relative biomass (SB/SB40) +#and probability for the 0.5 and 0.55 SPR policies. +#1113b_NoStateNature_no_abc_max_relF: Reset states of nature lines +new_dir <- file.path(wd, "rebuilder", "1113b_NoStateNature_no_abc_max_relF") +dir.create(new_dir) + +file.copy(from = file.path(wd, "rebuilder", "1100_2023", "rebuild.exe"), + to = file.path(new_dir, "rebuild.exe")) + +reb <- readLines(file.path(wd, "rebuilder", "1111b_no_abc_max_relF", "rebuild.dat"),n=-1) +loc <- grep("# Conduct projections", reb) +reb[loc+1] <- 0 +loc <- grep("# Number of parameter vectors", reb) +reb[loc+1] <- 1000 +writeLines(reb, file.path(new_dir, "rebuild.dat")) +#Difference in time to rebuilding remains. However, values for SB manually divided by SB40 +#align with rebuilder output for SB/SB40 when have no alt states. With alt states of nature, +#these dont quite align + + +#Test model run to figure out whether the alternative definition for prob rebuilding makes +#a difference in timing of probability of recovery and relative biomass for 0.5 and 0.55 policies. +#1113c_NoStateNature_no_abc_max_relF_altProb: Reset states of nature lines +new_dir <- file.path(wd, "rebuilder", "1113c_NoStateNature_no_abc_max_relF_altProb") +dir.create(new_dir) + +file.copy(from = file.path(wd, "rebuilder", "1100_2023", "rebuild.exe"), + to = file.path(new_dir, "rebuild.exe")) + +reb <- readLines(file.path(wd, "rebuilder", "1113b_NoStateNature_no_abc_max_relF", "rebuild.dat"),n=-1) +loc <- grep("# Defintion of recovery", reb) +reb[loc+1] <- 1 +writeLines(reb, file.path(new_dir, "rebuild.dat")) +#It does. For this run, the years for rebuilding based on SB/SB40 and probability were the same + + +#Same as above but now base on model with alternative states of nature. +#Test model run to figure out whether the alternative definition for prob rebuilding makes +#a difference in timing of probability of recovery and relative biomass for 0.5 and 0.55 policies. +#1111c_no_abc_max_relF_altProb: Reset states of nature lines +new_dir <- file.path(wd, "rebuilder", "1111c_no_abc_max_relF_altProb") +dir.create(new_dir) + +file.copy(from = file.path(wd, "rebuilder", "1100_2023", "rebuild.exe"), + to = file.path(new_dir, "rebuild.exe")) +file.copy(from = file.path(wd, "rebuilder", "rebuild_m_2023.SSO"), + to = file.path(new_dir, "rebuild_m_2023.SSO")) + +reb <- readLines(file.path(wd, "rebuilder", "1111b_no_abc_max_relF", "rebuild.dat"),n=-1) +loc <- grep("# Defintion of recovery", reb) +reb[loc+1] <- 1 +writeLines(reb, file.path(new_dir, "rebuild.dat")) +#Years for rebuilding based on SB/SB40 and probability are the same. States of nature dont affect this. + +#During these explorations I noticed that the histogram of B0 values in the output of Res.csv, +#which are then plotted, does not capture the range of B0 properly. The upper limit seems to be +#capped at a specific range from the midpoint, so values beyond that are plotted into the last +#category instead of at their actual location. This makes it appear a run has fewer modes for B0 +#than is actually the case. This occurred for my base runs, when each state of nature is applied +#once with equal probability, and when only the high and low states of nature are applied. +#The runs are correct though. I say this because the historgram of Tmin-yinit shows three +#distinct modes and these modes dissappear when rebuilder is rerun without the specific states of +#nature (whereas the x-values of the histogram of B0 stay the same in some cases), and that the +#B0 histogram of a run with just the low and high state of nature doesn't show the correct +#upper value of B0 (i.e. the B0 for the low state of nature). +#--------------------------------------------------------------------------------------------#