CCRFP_available_data_for_assessments.Rmd

---
title: "California Collaborative Fisheries Research Program"
subtitle: "Data availability for stock assessments"
author: "Compiled by Melissa H. Monk (SWFSC)"
date: "`r format(Sys.time(), '%B %d, %Y')`"
output:
  bookdown::pdf_document2: 
    keep_tex: true
    keep_md: true
toc: false
header-includes:
  - \usepackage{booktabs}
  - \usepackage{longtable}
  - \usepackage{array}
  - \usepackage{multirow}
  - \usepackage{wrapfig}
  - \usepackage{float}
  - \usepackage{colortbl}
  - \usepackage{pdflscape}
  - \usepackage{tabu}
  - \usepackage{threeparttable}
  - \usepackage[normalem]{ulem}
  - \usepackage{makecell}
  - \usepackage{xcolor}
  - \usepackage{placeins}
always_allow_html: true
---

\newcommand\CapeM{$40^\circ 10^\prime N. lat.$}

```{r load-packages, include = FALSE}
knitr::opts_chunk$set(echo = FALSE)
library(devtools)
library(dplyr)
library(ggplot2)
library(ggridges)
library(stringr)
library(tidyr)
library(purrr)
library(unikn)
library(viridis)
library(kableExtra)
library(knitr)
library(forcats)
```

```{r setup, echo = FALSE, include = FALSE, warning = FALSE, message = FALSE, cache = TRUE}
knitr::opts_chunk$set(echo = FALSE)

setwd("C:/GitHub/CCFRP")

#-------------------------------------------------------------------------------
# Read in data and basic cleanup
load("CCFRP_cleanedup.RData")
#-------------------------------------------------------------------------------

Assess2023 <- data.frame(
  Species = c(
#    "Black Rockfish",
#    "Copper Rockfish",
    "Quillback Rockfish",
    "Yellowtail Rockfish",
    "Vermilion Rockfish",
    "Gopher Rockfish"
  ),
  AssessGroup = "Assess2023"
)

# AssessTop30 <- data.frame(
#   Species = c(
#     "Brown Rockfish",
#     "Treefish",
#     "Blue Rockfish",
#     "Deacon Rockfish",
#     "Canary Rockfish",
#     "Vermilion Rockfish",
#     "China Rockfish",
#     "Olive Rockfish",
#     "Grass Rockfish"
#   ),
#   AssessGroup = "AssessTop30"
# )
SpeciesOfInterest <- Assess2023 #rbind(Assess2023, AssessTop30)
```

```{r echo = FALSE, include = FALSE, warning = FALSE, message = FALSE, cache = TRUE}
# Reorder university so that the table has years in order - not north to south,
# but this is a kable issue - if you put HSU first, the first year they sampled,
# 2017, is first in the table

#remove tilde from Ano, now causing issues
#dat_final$Name = str_replace(dat_final$Name,"Año Nuevo", "Ano Nuevo")

dat_final <- dat_final %>%
  ungroup() %>%
  dplyr::mutate(Monitoring.Group = factor(Monitoring.Group, levels = c(
    "MLML",
    "Cal Poly",
    "HSU", "BML",
    "UCSB", "SIO"
  ))) %>%
  dplyr::mutate(`Monitoring Group` = Monitoring.Group) %>%
  dplyr::mutate(`Monitoring Group` = recode(`Monitoring Group`,
    "HSU" = "Cal Poly Humboldt",
    "BML" = "Bodega Marine Lab",
    "MLML" = "Moss Landing Marine Lab",
    "Cal Poly" = "Cal Poly SLO",
    "UCSB" = "UC Santa Barbara",
    "SIO" = "Scripps Institute Ocean."
  )) %>%
  dplyr::mutate(Monitoring.Group = recode(Monitoring.Group,
    "HSU" = "Humboldt",
    "BML" = "Bodega",
    "MLML" = "Moss Landing",
    "Cal Poly" = "Cal Poly",
    "UCSB" = "UCSB",
    "SIO" = "Scripps"
  )) %>%
#  dplyr::mutate(Name = factor(Name)) %>%
# dplyr:: mutate(Name = dplyr::recode(Name,
#                       "Año Nuevo" = "Ano Nuevo")) %>%
  dplyr::mutate(Name = factor(Name, levels = c(
    "South Cape Mendocino", "Ten Mile",
    "Stewarts Point", "Bodega Head",
    "Año Nuevo", "Point Lobos",
    "Piedras Blancas", "Point Buchon",
    "Carrington Point", "Anacapa Island",
    "Swamis", "South La Jolla"
  ))) %>%
  dplyr::mutate(Name = recode(Name,
                              "Año Nuevo" = "Ano Nuevo")) %>%
  dplyr::mutate(SITE = factor(SITE))

# Change the levels of monitoring groups from north to south - didn't do before because of weird tables stuff
# Reorder the factor levels north to south
dat_final <- dat_final %>%
  mutate(Monitoring.Group = fct_relevel(Monitoring.Group, "Moss Landing", after = 3)) %>%
  mutate(Monitoring.Group = fct_relevel(Monitoring.Group, "Cal Poly", after = 3))

# Just drift info
drifts_final <- dat_final %>%
  ungroup() %>%
  select(
    Drift.ID, Trip.ID, Grid.Cell.ID, SITE, Name, MPA.Designation, YEAR, Month, 
    Day, Start.Depth..ft., End.Depth..ft., Monitoring.Group,
    StartGISDepth2mRes_ft, StartGISDepth90mRes_ft, EndGISDepth2mRes_ft, 
    EndGISDepth90mRes_ft
  ) %>%
  unique()

# lengths and then join to drift info
species_lengths <- catches %>%
  filter(!is.na(Length..cm.))
species_lengths <- inner_join(species_lengths, drifts_final)

spp_to_plot <- SpeciesOfInterest$Species[SpeciesOfInterest$AssessGroup == "Assess2023"]
```

# Summary
This document summarizes the data from the California Collaborative 
Fisheries Research Program (CCFRP), a survey that monitors groundfish populations 
in California's network of Marine Protected Areas (MPAs) and adjacent reference areas. 
Preliminary analyses suggest the following  data availability to inform stock 
assessments for each species:

**Black rockfish (*Sebastes melanops*)**: There are enough data to consider an 
index of abundance and also include ages in the assessment model north of Point
Conception. An exploration of an index could be considered if California is 
modeled as one area. Black rockfish are present in a high proportion of the fishing drops 
at the one MPA north of the \CapeM management line. However, this area was added in 
2017 and represents a shortened time series. There are likely enough otoliths available 
to include ages in the assessment model if California is modeled as a single area, and to 
inform growth externally north of \CapeM if the assessment is split at the management 
boundary.


**Copper rockfish (*S. caurinus*)**: There are enough data north of Point Conception 
to consider an index of abundance.  The two MPAs south of Point Conception with consistent encounters 
of copper rockfish cover a very small fraction of the available habitat and were not 
included in the original sampling design.  There are enough age data available north 
of Point Conception for inclusion in an assessment model. Otoliths will be 
collected south of Point Conception in 2022.  

**Quillback rockfish (*S. maliger*)**: There is only one MPA monitored north of the \CapeM management 
boundary in California that was not part of the original sampling design and hence 
does not have a long time series.  It is not likely that robust index can be developed
from the sparse data separately north and south of \CapeM. An exploration of an index 
could be considered if California is modeled as one area. Age data are available 
that could inform growth external to the assessment model.    

**Yellowtail rockfish  (*S. flavidus*)**: There are enough data to consider an index of abundance north of 
Point Conception for yellowtail rockfish as well as include age data from the otoliths
within the assessment model.



# Survey Background
The 1999 Marine Life Protection Act resulted in the creation of a network of 
Marine Protected Areas (MPAs) along California's coast. The state of California
designated both State Marine Reserves (SMRs) and State Marine Recreational 
Management areas (SMCAs). The SMRs prohibit all recreational and commercial take 
and SMCAs allow some recreational and/or commercial take that varies by SMCA. 
A number of MPAs consist of an SMR adjacent to an SMCA, of which the SMR is closer to shore.
The California Collaborative Fisheries Research Program, 
[CCFRP](https://www.mlml.calstate.edu/ccfrp/), 
is a fishery-independent hook-and-line survey designed to monitor nearshore fish 
populations at a series of sampling locations both inside and adjacent to 
California's network of MPAs. 

The CCFRP survey began in 2007 with  California Polytechnic University 
San Luis Obispo (Cal Poly) and San Jose State University Moss Landing Marine Lab 
(Moss Landing) in collaboration with NMFS scientists and the fishing community.  The core 
area of the survey includes A&ntilde;o Nuevo SMR and Point Lobos SMR sampled by 
Moss Landing, and Point Buchon SMR and Piedras Blancas SMR sampled by Cal Poly (Figure \@ref(fig:fig-mpa-map)). In 2017, CCFRP expanded within 
California to include four additional partners, Cal Poly Humboldt (formerly Humboldt State 
University), University of California Davis' Bodega Marine Lab, University 
of California Santa Barbara, and the Scripps Institute of Oceanography. 
The CCFRP now monitors 12 MPA and reference area pairs (Table \@ref(tab:monitoring)). 
Cal Poly Humboldt samples the furthest north sites, which are south of Cape 
Mendocino, but north of the management line at \CapeM. The COVID-19 pandemic also affected the 
survey effort, but all partners were able to conduct sampling in 2020 and 2021.

The CCFRP survey design is consistent across all partners. Each MPA and reference 
area consists of a number of 500 x 500 m cells that were selected because the contained 
appropriate rockfish habitat. The survey was designed as a capture 
and release survey, with a sub-study tag/recapture program. Therefore, CCFRP restricts 
sampling areas to depths shallower than approximately 120 feet in order to reduce 
potential effects of barotrauma.  On any given survey 
day site cells are randomly selected within a stratum (MPA and/or reference cells).
Commercial passenger fishing vessels (CPFVs) are chartered for the survey and 
the captain is allowed to search within the cell for a fishing location.
Due to the nature of the fishery in northern California, Cal Poly Humboldt conducts 
sampling aboard 6-pack vessels, and therefore fishes for fewer total angler hours per 
year compared to the other partners (Tables \@ref(tab:anghrs) and \@ref(tab:fishingdrops)). During 
a sampling event, each cell is fished for a total of 30-45 minutes by volunteer 
anglers. Volunteer anglers are allowed to reel up their lines at any time during a 
fishing drop if they think they've hooked fish. Anglers can then re-bait and 
continue fishing until the the drop is complete. Each fish encountered 
can be linked back to an angler. Each anglers fishes one line, with two hooks. 
The jig and bait is assigned to each angler, but an angler may fish with a personal 
fishing rod.

All fish encountered are measured to the nearest centimeter (fork length), and the 
majority of fish are released or descended to depth. 
A total of `r dim(catches %>% filter(str_length(Tag.ID)>0))[1]` fish were tagged 
since 2007, and recapture data are available from each partner. Starting in 2017, 
at the request of Melissa Monk (NMFS SWFSC), a fraction of the fish encountered 
in the reference cells have been retained to collect otoliths and fin clips that 
provide needed biological information for nearshore species.  In 2022, the goal 
is to increase biological collections for commonly encountered species for use 
in the 2023 stock assessments.

 
# Available Data for Indices

From 2007-2021 a total of `r length(unique(drifts_all$Trip.ID))` fishing trips 
were taken, consisting of `r dim(drifts_all)[1]` fishing drops. When the CCFRP 
expanded in 2017, some MPAs/sites were fished in only one or two years during an 
exploratory phase.  These included Laguna Beach, the southeast Farallon Islands, 
Point Conception and Trinidad, which were excluded from this summary since we 
would not include them in a stock assessment. 
Fishing drops that drifted outside a cell were also excluded.  These site filter 
result in an available `r length(unique(dat$Drift.ID))`. The final filter removed 
drifts within a cell that were not fished for at least ten minutes within a sampling 
occasion, resulting in a total of `r length(unique(dat_final$Drift.ID))` fishing 
drops available for analyses for stock assessments. The total number of fish 
encountered by CCFRP partners and the percent of positive drops by species and MPA
can be found in Tables \@ref(tab:totalfishbygroup) - \@ref(tab:percentpos).



# Available Lengths and Otoliths

The CCFRP measures every fish to the nearest centimeter. Distributions of fish 
lengths inside and outside the MPAs are in Figures 
\@ref(fig:lengths-1) - \@ref(fig:lengths-4).  Length data were 
filtered to the drifts that would be used to develop indices of abaundacne. 
Any species and site (MPA or reference) combination with fewer than 20 observed 
fish over the entirety of the program were not plotted.

The total number of fish retained by university partner can be found in Table 
\@ref(tab:otoliths).  This represents the maximum number of available otoliths, 
which will be verified once the stock assessments for 2023 are selected.  The 
rule of thumb for including conditional age-at-length samples is a minimum of 30 
available fish in a year/fleet stratum.  Given this, there are likely not enough 
fish from Bodega or Cal Poly to support conditional age-at-lengths for any species. 

\FloatBarrier
\newpage

# Tables

\FloatBarrier

```{r monitoring, echo = FALSE, warning = FALSE, message = FALSE}
knitr::kable(
  booktabs = T,
  caption = "Monitoring groups and the associated MPAs they sample. The abbreviated names will be 
  used throughout most of the tables in this document",
  dat_final %>%
    select(`Monitoring Group`, Monitoring.Group, Name) %>%
    unique() %>%
    arrange(Monitoring.Group, Name) %>%
    rename('Abbreviated Name' = Monitoring.Group,
           'MPA' = Name)
) %>%
  kable_styling(latex_options = "striped")
```



```{r anghrs, echo = FALSE, warning = FALSE, message = FALSE}
# Total effort by year among the programs
knitr::kable(
  booktabs = T,
  caption = "Total angler hours by institution summed across all active years.",
  dat_final %>%
    ungroup() %>%
    select(Drift.ID, YEAR, Effort, Total...Anglers.Fishing, Monitoring.Group, Name) %>%
    unique() %>%
    select(Monitoring.Group, YEAR, Effort) %>%
    group_by(Monitoring.Group, YEAR) %>%
    summarise(Total_AnglerHours = round(sum(Effort), 0)) %>%
    pivot_wider(id_cols = YEAR, names_from = Monitoring.Group, values_from = Total_AnglerHours, values_fill = 0) %>%
    arrange(YEAR)
) %>%
  kable_styling(latex_options = "striped") %>%
  column_spec(2:7, width = "6em")
```



```{r fishingdrops, echo = FALSE, warning = FALSE, message = FALSE}
# Number of drift within/outside each mpa by year
knitr::kable(
  booktabs = T,
  caption = "Total number of fishing drops by year at each monitored site in the reference areas and inside the MPAs, in parentheses.",
  dat_final %>%
    select(Drift.ID, YEAR, SITE, Name) %>%
    unique() %>%
    group_by(Name, SITE, YEAR) %>%
    tally() %>%
    pivot_wider(id_cols = c(YEAR, Name), names_from = SITE, values_from = n) %>%
    mutate(Samples = paste0(REF, "(", MPA, ")")) %>%
    select(YEAR, Name, Samples) %>%
    pivot_wider(id_cols = YEAR, names_from = Name, values_from = Samples, values_fill = '-') %>%
    arrange(YEAR)
) %>%
  add_header_above(c(" ", "Cal Poly Humboldt" = 2, "Bodega Marine Lab" = 2, 
                   "Moss Landing" = 2, "Cal Poly SLO" = 2, 
                   "UC Santa Barbara" = 2, "Scripps" = 2))  %>%
  kable_styling(latex_options = c("striped", "scale_down")) %>%
  column_spec(2, width = "1.7cm") %>%
  column_spec(3:12, width = "1.5cm") %>%
   landscape()
```

\FloatBarrier
\newpage


\FloatBarrier



```{r totalfishbygroup, echo = FALSE, warning = FALSE, message = FALSE}
# Total number of fish encountered by monitoring group
knitr::kable(
  booktabs = T,
  caption = "Total number of fish encountered by each monitoring group.",
  dat_final %>%
    filter(Common.Name %in% SpeciesOfInterest$Species) %>%
    group_by(Common.Name, Monitoring.Group) %>%
    summarise(total = sum(n)) %>%
    pivot_wider(id_cols = Common.Name, names_from = Monitoring.Group, values_from = total, values_fill = 0)
) %>%
  kable_styling(latex_options = "striped")
```  

```{r echo = FALSE, include = FALSE, warning = FALSE, message = FALSE}

# Percent positive drifts at each MPA by year
drifts_by_MPA <- dat_final %>%
  select(Drift.ID, YEAR, Monitoring.Group, Name) %>%
  unique() %>%
  group_by(Name, YEAR) %>%
  summarise(tot_drifts = n())
```

```{r percentpos, results = 'asis',warning = FALSE, message = FALSE}
# calculate the number of positive drifts by each of the 2023 assess species
for (i in 1:length(SpeciesOfInterest$Species[SpeciesOfInterest$AssessGroup == "Assess2023"])) { # nolint
  pos_drifts <- dat_final %>%
    filter(Common.Name %in% SpeciesOfInterest$Species[SpeciesOfInterest$AssessGroup == "Assess2023"][i]) %>%
    select(Drift.ID, YEAR, Monitoring.Group, Name) %>%
    unique() %>%
    group_by(Name, YEAR) %>%
    summarise(tot_pos_drifts = n())

  print(knitr::kable(
    booktabs = T,
    caption = paste0(
      "Percent of drifts with encounters of ",
      SpeciesOfInterest$Species[SpeciesOfInterest$AssessGroup == "Assess2023"][i],
      " at each monitoring location (inside and outside araes combined) and year."
    ),
    left_join(drifts_by_MPA, pos_drifts) %>%
      mutate(percent_pos = scales::percent(tot_pos_drifts / tot_drifts, accuracy = 2)) %>%
      pivot_wider(id_cols = YEAR, names_from = Name, values_from = percent_pos, values_fill = "-") %>%
      replace(is.na(.), "-") %>%
      arrange(YEAR)
  ) %>%
    kable_styling(latex_options = "striped") %>%
    column_spec(2, width = "1.7cm") %>%
    column_spec(3:12, width = "1.2cm") %>%
    column_spec(10, width = "1.7cm") %>%
    landscape()
  )
  cat("\n\n<!-- -->\n\n")
}

```


\FloatBarrier


```{r otoliths, results = 'asis',warning = FALSE, message = FALSE}
knitr::kable(
  booktabs = T,
  caption = "Total number of fish retained by monitoring group over the duration of the program.",
   species_lengths %>%
   filter(Retained == 'TRUE',
          Common.Name %in% spp_to_plot) %>%
   left_join(., drifts_final) %>%
    group_by(Common.Name, Monitoring.Group) %>%
    tally() %>%
    pivot_wider(id_cols = Common.Name, names_from = Monitoring.Group, values_from = n, values_fill = 0) %>%
     rename(`Common Name` = Common.Name)
) %>%
  kable_styling(latex_options = "striped")

```

\FloatBarrier

# Figures

```{r, fig-mpa-map, echo = FALSE, fig.cap = "Map of the State Marine Reserves (SMRs) monitored by the CCFRP program."}
knitr::include_graphics('MPA_map.pdf')
```


```{r lengthplot, warning = FALSE, message = FALSE, echo=FALSE}
#"Frequency of fork length (cm) for inside each MPA (MPA) and outside at reference areas (REF).",

#create plot list
plot_list = vector('list', length(spp_to_plot))


for(i in 1:length(spp_to_plot)){ # nolint
 #subset to the species
  spp_length_plot_dat = species_lengths %>%
     dplyr::filter(Common.Name == spp_to_plot[i])

#tally the sample sizes and create the label
 length_counts <- species_lengths %>%
    filter(Common.Name == spp_to_plot[i]) %>%
  group_by(Common.Name, Name, SITE) %>%
  summarise(n = n()) %>%
   filter(n>19) %>%
   pivot_wider(id_cols = Name, names_from = SITE, values_from = n, values_fill = NA) %>%
   mutate(Label = paste("MPA: N = ", MPA, "\nREF: N = ", REF)) %>%
   rename(MPA1 = MPA,
          REF1 = REF)
 

 #join the labels to the dataframe and remove the rows with <20 in a site
 spp_length_plot_dat <- left_join(spp_length_plot_dat, length_counts)
 spp_length_plot_dat <- spp_length_plot_dat %>%
    filter((!is.na(MPA1) & SITE=='MPA') | (!is.na(REF1) & SITE=='REF')) %>%
   droplevels()
   
 
 #set x and y coordinates for the labels
 # and number of columns
# x_pos = c(15, 15, 5, 5)
 y_pos = c(0.12, 0.11, 0.1, 0.1, 0.1, 0.1)
 ncols = c(2, 3, 2, 2, 2, 2)
 
 
 
  plot_list[[i]]  <- local({
    i <- i
  p1 <- ggplot(aes(x = Length..cm., fill = SITE), data = spp_length_plot_dat) +
  geom_density(alpha=0.5) + #inherit.aes = FALSE, , trim = TRUE
    #scale_y_continuous(labels = scales::percent_format()) +
     facet_wrap(~Name, ncol = ncols[i], dir = "v") + #, scale = "free_y", space = "free"
     scale_color_brewer(palette = "Dark2") +
    labs(
      x = "Fork length (cm)",
      y = "Density"
  ) +
    scale_x_continuous(limit = c(0,60)) +
    theme_bw() +
    theme(axis.line = element_line(color='black'),
    plot.background = element_blank(),
    panel.grid.major = element_blank(),
    panel.grid.minor = element_blank(),
    legend.title = element_blank()) +
   geom_text(aes(x = 1, y = y_pos[i], label = Label), 
             size = 2.5, family = "sans", #inherit.aes = FALSE, 
             data = spp_length_plot_dat, check_overlap = TRUE,
             hjust = 0)
        # print(p1)
})
}


names(plot_list) = spp_to_plot
 
```


```{r lengths, fig.cap = paste("Density plot of",names(plot_list), "fork lengths form fish encountered inside each MPA and outside at reference areas (REF) over all years of the program. A sample size of NA indicates fewer than 20 fish were encountered in that MPA stratum and were not plotted. Areas north of Ano Nuevo and south of Point Buchon were sampled beginning in 2017."), echo=FALSE, warning= FALSE, results = "asis"}

for(i in 1:length(plot_list)) { # nolint
     print(plot_list[[i]])
    cat('\n\n')
}
#fig.cap = "Percent of fish in 2 cm fork length bins inside each MPA and outside at reference areas (REF)."
#eval.after = "fig.cap",warning = FALSE, message = FALSE, echo = FALSE,
```



\FloatBarrier