index.qmd

---
title: "Movement Ecology and Behavior of Ribbon and Spotted Seals in the Bering Sea, 2024 "
execute:
  echo: false
  warning: false
  message: false
format: 
  html:
    fig-width: 8
    fig-height: 6
search: false

author:
  - id: jml
    name:
      literal: Josh M. London
    email: josh.london@noaa.gov
    orcid: 0000-0002-3647-5046
    attributes:
      corresponding: true
    affiliations: 
      - ref: mml
  - id: hlz
    name:
      literal: Heather L. Ziel
    email: heather.ziel@noaa.gov
    affiliations: 
      - ref: mml
  - id: skh
    name:
      literal: Stacie M. Koslovsky
    email: stacie.koslovsky@noaa.gov
    affiliations: 
      - ref: mml
  - id: mfc
    name:
      literal: Michael F. Cameron
    email: michael.cameron@noaa.gov
    affiliations: 
      - ref: mml
affiliations:
  - id: mml
    name: AFSC Marine Mammal Laboratory, NOAA Fisheries
    address: 7600 Sand Point Way NE
    city: Seattle
    region: Washington
    country: United States
    postal-code: 98115
    url: https://www.fisheries.noaa.gov/about/marine-mammal-laboratory
---

```{r}
#| include: false
library(dplyr)
library(purrr)
library(fs)
library(tidyr)
library(lubridate)
library(stringr)
library(ggplot2)
library(janitor)
library(readr)
library(googlesheets4)
library(wcUtils)
library(scam)
library(sf)
library(googlesheets4)
library(rdeck)
library(aniMotum)
library(rcartocolor)
library(gt)
library(ggdist)

r_files <- fs::dir_ls(here::here('R'), glob = "*.R")
purrr::map(r_files, source)
deploy_tbl <- readr::read_rds(here::here('data/deploy_tbl.rds'))  |>  
  dplyr::mutate(deploy_end_date_time_gmt = lubridate::as_date('2025-12-31'))

```

```{r}
#| include: false
#data_list <- get_wc_data(deploy_tbl$deployid[!deploy_tbl$deployid %in% c('PL2024_1023_18U1318','HF2024_1021_19U2627','HF2024_1019_19U2623')])

data_list <- get_wc_data(deploy_tbl$deployid) 
```

::: callout-warning
The information on this page is automatically updated every hour and without any
significant quality checks. Errors may be present and are likely. We are
providing access to the information in a public forum in the spirit of open
science and to provide easy access to updated information for the authors and
interested researchers.

**The information provided here should not be cited or referenced in any form**
:::

::: column-margin
Disclaimer

The scientific results and conclusions, as well as any views or opinions
expressed herein, are those of the author(s) and do not necessarily reflect the
views of NMFS, NOAA, or the Department of Commerce.
:::

## Sampling and Deployment Summary

```{r}
#| include: false
summary_tbl <- readr::read_rds(here::here('data/summary_tbl.rds'))
```

```{r}
gt(summary_tbl) |> 
cols_label(
  speno = "SPENO",
  date = "Date",
  species = "Species",
  sex = "Sex",
  age = "Age",
  molt_status = "Molt Status",
  mass = "Mass (kg)",
  splash = "SPLASH",
  spot6 = "SPOT6"
) |> 
tab_spanner(
      label = md('Bio-logger Type & Location'),
      columns = 8:9
    ) |> 
opt_stylize(style = 6, color = 'gray') |> 
tab_options(
  table.font.size = px(12)
)
```



## Predicted Movements

Each seal is released with a bio-logger that provides regular updates on
movement and behavior. For each seal, the observed locations are used to fit a
predicted path. Both the predicted track (line) and the observed bio-logger
locations (points) are shown on the map. The ship track of the _R/V Norseman II_
is shown in black and can be toggled on/off using the layer selector in the
upper left corner.

**Because some of the tracks span the 180 anti-meridian, you may need to re-center
and zoom the map to the Bering Sea region.**



```{r}
#| include: false
map_proj <- "+proj=laea +lat_0=90 +lon_0=180 +x_0=0 +y_0=0 +datum=WGS84 +units=m +no_defs +type=crs"

locs <- sf::st_as_sf(data_list$locs,
                     coords = c("longitude","latitude"),
                     crs = 4326) %>% 
  left_join(deploy_tbl) %>% 
  dplyr::filter(between(date_time,
                        deploy_date_time_gmt,
                        deploy_end_date_time_gmt)) %>% 
  dplyr::rename(datetime=date_time)

locs <- locs %>%
  group_by(speno) %>%
  arrange(datetime, error_radius) %>%
    mutate(
      rank = 1L,
      rank = case_when(duplicated(datetime, fromLast = FALSE) ~
                         lag(rank) + 1L, TRUE ~ rank)) %>%
    dplyr::filter(rank == 1) %>% 
  arrange(speno,datetime) %>% 
  ungroup() 

locs <- locs |> 
  dplyr::mutate(
    quality = case_when(
      type == 'FastGPS' ~ 'G',
      type == 'User' ~ 'G',
      .default = quality
    )
  ) |> 
  dplyr::filter(!quality %in% c('Z')) |> 
  sf::st_transform(map_proj)

locs_fit <- fit_ssm(
  x = locs,
  vmax = 8,
  model = "crw",
  time.step = 0.25,
  id = "speno",
  date = "datetime",
  lc = "quality",
  epar = c(
    "error_semi_major_axis",
    "error_semi_minor_axis",
    "error_ellipse_orientation"
  ),
  tz = "UTC"
)

predict_pts_sf <-
  grab(locs_fit,
       what = "predicted",
       as_sf = TRUE,
       group = TRUE) |> 
  rename(speno = id,
         datetime = date) |> 
  left_join(deploy_tbl)

predict_lines_sf <- predict_pts_sf %>% 
  group_by(speno) %>%
  summarise(do_union = FALSE) %>%
  st_cast("LINESTRING") %>%
  left_join(deploy_tbl)

map_lines <- predict_lines_sf |> 
st_transform(4326)

bbox_center <- st_bbox(predict_lines_sf) %>% 
  st_as_sfc() |> 
st_transform(4326) |> 
st_break_antimeridian(lon_0=180)

map_last_loc <- predict_pts_sf %>% 
  sf::st_transform(4326) %>% 
  group_by(speno) %>% 
  arrange(speno,datetime) %>% 
  group_map(~ tail(.x, 1L), .keep=TRUE) %>% 
  bind_rows()

map_locs <- locs %>% 
  sf::st_transform(4326) %>% 
  group_by(speno) 

map_locs_adult <- map_locs |> 
  dplyr::filter(age == 'adult')
map_locs_subadult <- map_locs |> 
  dplyr::filter(age == 'sub-adult')
map_locs_yoy <- map_locs |> 
  dplyr::filter(age == 'young-of-year')

map_lines_adult <- map_lines |> 
  dplyr::filter(age == 'adult')
map_lines_subadult <- map_lines |> 
  dplyr::filter(age == 'sub-adult')
map_lines_yoy <- map_lines |> 
  dplyr::filter(age == 'young-of-year')

map_last_loc_adult <- map_last_loc |> 
  dplyr::filter(age == 'adult')
map_last_loc_subadult <- map_last_loc |> 
  dplyr::filter(age == 'sub-adult')
map_last_loc_yoy <- map_last_loc |> 
  dplyr::filter(age == 'young-of-year')

ship_track <- readr::read_rds(here::here('data/ship_track.rds'))

mapbox_map <- rdeck(map_style = "mapbox://styles/jmlondon/cl9kktfl2000v16pooqw5p2wz",
                initial_bounds = bbox_center[1],
                theme = "light",
                controller = TRUE) %>%
  add_path_layer(data = ship_track,
                    name = "Ship Track",
                    wrap_longitude = TRUE,
                    position_format = "XY",
                    width_min_pixels = 1.5,
                    get_color = "#000000",
                    opacity = 0.25,
                    get_path = geometry) |> 
  add_path_layer(data = map_lines,
                    name = "Predicted Movements",
                    opacity = 0.8,
                    get_color = scale_color_category(
                      col = speno,
                      palette = rcartocolor::carto_pal(name = "Pastel")
                      ),
                    wrap_longitude = TRUE,
                    position_format = "XY",
                    width_min_pixels = 1.5,
                    get_path = geometry) %>%
    add_scatterplot_layer(data = map_locs,
                    name = "Observed Locations",
                    opacity = 0.35,
                    get_fill_color = scale_color_category(
                      col = speno,
                      palette = rcartocolor::carto_pal(name = "Pastel"),
                      legend = FALSE
                      ),
                    wrap_longitude = TRUE,
                    position_format = "XY",
                    radius_min_pixels = 2,
                    get_position = geometry,
                    pickable = TRUE,
                    tooltip = c(speno,datetime)) %>% 
  add_scatterplot_layer(data = map_last_loc,
                    name = "Latest Location",
                    opacity = 1,
                    get_fill_color = scale_color_category(
                      col = speno,
                      palette = rcartocolor::carto_pal(name = "Pastel"),
                      legend = FALSE
                      ),
                    wrap_longitude = TRUE,
                    position_format = "XY",
                    radius_min_pixels = 4,
                    get_position = geometry,
                    pickable = TRUE,
                    tooltip = c(speno,datetime))

mapbox_map_adult <- rdeck(map_style = "mapbox://styles/jmlondon/cl9kktfl2000v16pooqw5p2wz",
                initial_bounds = bbox_center[1],
                theme = "light",
                controller = TRUE) %>%
  add_path_layer(data = ship_track,
                    name = "Ship Track",
                    wrap_longitude = TRUE,
                    position_format = "XY",
                    width_min_pixels = 1.5,
                    get_color = "#000000",
                    opacity = 0.25,
                    get_path = geometry) |> 
  add_path_layer(data = map_lines_adult,
                    name = "Predicted Movements of Adults",
                    opacity = 0.8,
                    get_color = scale_color_category(
                      col = speno,
                      palette = rcartocolor::carto_pal(name = "Pastel")
                      ),
                    wrap_longitude = TRUE,
                    position_format = "XY",
                    width_min_pixels = 1.5,
                    get_path = geometry) %>%
    add_scatterplot_layer(data = map_locs_adult,
                    name = "Observed Location",
                    opacity = 0.35,
                    get_fill_color = scale_color_category(
                      col = speno,
                      palette = rcartocolor::carto_pal(name = "Pastel"),
                      legend = FALSE
                      ),
                    wrap_longitude = TRUE,
                    position_format = "XY",
                    radius_min_pixels = 2,
                    get_position = geometry,
                    pickable = TRUE,
                    tooltip = c(speno,datetime)) %>% 
  add_scatterplot_layer(data = map_last_loc_adult,
                    name = "Latest Location",
                    opacity = 1,
                    get_fill_color = scale_color_category(
                      col = speno,
                      palette = rcartocolor::carto_pal(name = "Pastel"),
                      legend = FALSE
                      ),
                    wrap_longitude = TRUE,
                    position_format = "XY",
                    radius_min_pixels = 4,
                    get_position = geometry,
                    pickable = TRUE,
                    tooltip = c(speno,datetime))

mapbox_map_subadult <- rdeck(map_style = "mapbox://styles/jmlondon/cl9kktfl2000v16pooqw5p2wz",
                initial_bounds = bbox_center[1],
                theme = "light",
                controller = TRUE) %>%
  add_path_layer(data = ship_track,
                    name = "Ship Track",
                    wrap_longitude = TRUE,
                    position_format = "XY",
                    width_min_pixels = 1.5,
                    get_color = "#000000",
                    opacity = 0.25,
                    get_path = geometry) |> 
  add_path_layer(data = map_lines_subadult,
                    name = "Predicted Movements of Sub-adults",
                    opacity = 0.8,
                    get_color = scale_color_category(
                      col = speno,
                      palette = rcartocolor::carto_pal(name = "Pastel")
                      ),
                    wrap_longitude = TRUE,
                    position_format = "XY",
                    width_min_pixels = 1.5,
                    get_path = geometry) %>%
    add_scatterplot_layer(data = map_locs_subadult,
                    name = "Observed Location",
                    opacity = 0.35,
                    get_fill_color = scale_color_category(
                      col = speno,
                      palette = rcartocolor::carto_pal(name = "Pastel"),
                      legend = FALSE
                      ),
                    wrap_longitude = TRUE,
                    position_format = "XY",
                    radius_min_pixels = 2,
                    get_position = geometry,
                    pickable = TRUE,
                    tooltip = c(speno,datetime)) %>% 
  add_scatterplot_layer(data = map_last_loc_subadult,
                    name = "Latest Location",
                    opacity = 1,
                    get_fill_color = scale_color_category(
                      col = speno,
                      palette = rcartocolor::carto_pal(name = "Pastel"),
                      legend = FALSE
                      ),
                    wrap_longitude = TRUE,
                    position_format = "XY",
                    radius_min_pixels = 4,
                    get_position = geometry,
                    pickable = TRUE,
                    tooltip = c(speno,datetime))

mapbox_map_yoy <- rdeck(map_style = "mapbox://styles/jmlondon/cl9kktfl2000v16pooqw5p2wz",
                initial_bounds = bbox_center[1],
                theme = "light",
                controller = TRUE) %>%
  add_path_layer(data = ship_track,
                    name = "Ship Track",
                    wrap_longitude = TRUE,
                    position_format = "XY",
                    width_min_pixels = 1.5,
                    get_color = "#000000",
                    opacity = 0.25,
                    get_path = geometry) |> 
  add_path_layer(data = map_lines_yoy,
                    name = "Predicted Movements of Young-of-Year",
                    opacity = 0.8,
                    get_color = scale_color_category(
                      col = speno,
                      palette = rcartocolor::carto_pal(name = "Pastel")
                      ),
                    wrap_longitude = TRUE,
                    position_format = "XY",
                    width_min_pixels = 1.5,
                    get_path = geometry) %>%
    add_scatterplot_layer(data = map_locs_yoy,
                    name = "Observed Location",
                    opacity = 0.35,
                    get_fill_color = scale_color_category(
                      col = speno,
                      palette = rcartocolor::carto_pal(name = "Pastel"),
                      legend = FALSE
                      ),
                    wrap_longitude = TRUE,
                    position_format = "XY",
                    radius_min_pixels = 2,
                    get_position = geometry,
                    pickable = TRUE,
                    tooltip = c(speno,datetime)) %>% 
  add_scatterplot_layer(data = map_last_loc_yoy,
                    name = "Latest Location",
                    opacity = 1,
                    get_fill_color = scale_color_category(
                      col = speno,
                      palette = rcartocolor::carto_pal(name = "Pastel"),
                      legend = FALSE
                      ),
                    wrap_longitude = TRUE,
                    position_format = "XY",
                    radius_min_pixels = 4,
                    get_position = geometry,
                    pickable = TRUE,
                    tooltip = c(speno,datetime))
```

::: column-page

::: {.panel-tabset}

### All Age Classes
```{r}
mapbox_map
```

### Adults
```{r}
mapbox_map_adult
```

### Sub-adults
```{r}
mapbox_map_subadult
```

### Young-of-Year
```{r}
mapbox_map_yoy
```

:::

:::


## Haul-out Behavior

Each of the bio-loggers are programmed to report the percentage of time the
device was dry for each hour. For example, a seal that was hauled out for an
entire hour would report as 100% dry. The hours are shown as UTC hours. To
adjust for apparent solar noon, subtract approximately 12 hours.

::: {.panel-tabset}

```{r}
ho_data <- data_list$haul_out |> 
  left_join(deploy_tbl) |> 
  dplyr::filter(between(timeline_start_dt,
                        deploy_date_time_gmt,
                        deploy_end_date_time_gmt))  |>  
  dplyr::rename(haulout_dt=timeline_start_dt) |> 
  dplyr::select(species = common_name, speno, haulout_dt, percent_dry)  |> 
  dplyr::mutate(percent_dry = mean(percent_dry,na.rm=TRUE), 
                                  .by = c(speno, haulout_dt)) |> 
  arrange(speno,haulout_dt) |>  
  mutate(year = lubridate::year(haulout_dt),
         month = lubridate::month(haulout_dt,label=TRUE),
         day = lubridate::day(haulout_dt),
         hour = lubridate::hour(haulout_dt))
```

### Ribbon seals

```{r}
#| fig-asp: 1.3

p <- ggplot(ho_data |> dplyr::filter(species == 'ribbon seal'),aes(day,hour,fill=percent_dry))+
  geom_tile(linewidth = 0.1, color = "grey75") + 
  scico::scale_fill_scico(palette = "nuuk", direction = 1,
  labels = scales::percent_format(scale=1),
                       aesthetics = "fill",
    guide = guide_colorbar(title.position = "top", barwidth = 15,
                           barheight = 0.5, title.hjust = 0)
    ) +
  coord_cartesian(clip = "off")

p <- p + facet_grid(speno~month)
p <- p + scale_x_continuous(breaks = c(5,15,25), expand = c(0,0)) +
  scale_y_continuous(breaks = c(4,12,20), labels = c("04:00","12:00","20:00"))

p <- p + theme_minimal() +
  theme(legend.position = "top", legend.justification = "left") +
  theme(strip.background = element_rect(colour="white")) +
  theme(axis.ticks=element_blank()) +
  xlab("day of month") + ylab("hour (UTC)")
p
```

### Spotted seals

```{r}
#| fig-asp: 1.3

p <- ggplot(ho_data |> dplyr::filter(species == 'spotted seal'),
              aes(day,hour,fill=percent_dry))+
  geom_tile(linewidth = 0.1, color = "grey75") + 
  scico::scale_fill_scico(palette = "nuuk", direction = 1,
  labels = scales::percent_format(scale=1),
                       aesthetics = "fill",
    guide = guide_colorbar(title.position = "top", barwidth = 15,
                           barheight = 0.5, title.hjust = 0)
    ) +
  coord_cartesian(clip = "off")

p <- p + facet_grid(speno~month)
p <- p + scale_x_continuous(breaks = c(5,15,25), expand = c(0,0)) +
  scale_y_continuous(breaks = c(4,12,20), labels = c("04:00","12:00","20:00"))

p <- p + theme_minimal() +
  theme(legend.position = "top", legend.justification = "left") +
  theme(strip.background = element_rect(colour="white")) +
  theme(axis.ticks=element_blank()) +
  xlab("day of month") + ylab("hour (UTC)")
p
```

:::

## Dive Behavior

The bio-loggers that are attached to the hair of the nape or head of seals are
capable of determining depth through a pressure transducer. This allows for a
variety of data products that describe the dive behavior of seals.

### Time At Depth from Cumulative Distribution

Here, dive behavior is described as proportion of time at given depths. This
relatively new data product transmits a cumulative distribution of time at depth
for each 2 hour summary period. The distribution adjusts to accomodate different
maximum depths and darker red areas indicate depths where the seal is choosing
to spend more time. The grey blocks above indicate proportion of time dry within
the summary period.

::: {.panel-tabset}

```{r}
ecd_data <- data_list$ecdf %>% 
  dplyr::rename(deployid = deploy_id) %>% 
  dplyr::arrange(deployid, start) %>% 
  left_join(deploy_tbl) %>% 
  rowwise() %>% 
  dplyr::mutate(spline_interp = list(spline_ecdf(full_ecdf, bin.width=5)),
                depth50pct = x_pct_depth(full_ecdf,pct_tad = 0.5),
                n_dives = if_else(all(is.na(n_dives_shallow),is.na(n_dives_deep)),
                                  NA, mean(c(n_dives_shallow,n_dives_deep), na.rm=TRUE))
  ) %>% 
  dplyr::select(c(species = common_name,speno,deployid,deploy_date_time_gmt,deploy_end_date_time_gmt,start:percent_dry,shallow_ecdf, 
                  deep_ecdf,full_ecdf,spline_interp,depth50pct, n_dives)) %>% 
  unnest(spline_interp, keep_empty = TRUE) %>% 
  group_by(deployid, start, end, kind, percent_dry) %>% 
  dplyr::arrange(deployid, start, depth_break) %>% 
  dplyr::mutate(propTAD = c(0,diff(ecd_prop)),
                minTAD = 120*(1-0.01*percent_dry) * propTAD) %>% 
  # determine the next depth value
  dplyr::mutate(next_depth = dplyr::lead(depth_break)) %>%
  # determine previous depth for plotting
  dplyr::mutate(prev_depth = dplyr::lag(depth_break)) %>%
  dplyr::select(-next_depth) %>%
  # transform depths to negative values
  dplyr::mutate(depth_break = depth_break * -1,
                prev_depth = prev_depth * -1)

ecd_data <- ecd_data %>% 
  group_by(deployid) %>% 
  arrange(start)
```

#### Ribbon seals

```{r}
#| fig-asp: 1.3
ecd_data %>% 
  dplyr::filter(species == 'ribbon seal',
                between(end,
                        deploy_date_time_gmt,
                        deploy_end_date_time_gmt)) %>%
ggplot() +
  geom_rect(aes(xmin = start, xmax = end,
                ymin = 0, ymax = percent_dry/10),
            fill = "seashell3") +
  geom_rect(aes(xmin = start, xmax = end,
                ymin = depth_break, ymax = prev_depth,
                fill = propTAD),
            color = NA) +
  scale_fill_distiller(palette = "Reds",
                   direction = 1, trans = "log10",
                       guide = guide_colorbar(
                         title = 'proportion of time submerged',
                         title.position = 'bottom',
                         title.hjust = 0.5,
                         barwidth = unit(75, units = "mm"),
                         barheight = unit(2, units = "mm"))) +
  scale_x_datetime() +
  facet_wrap(speno ~ ., ncol=1) +
  ylab("depth (meters)") +
  labs(title = "Ribbon Seals (HF), 2024",
       subtitle = "each bar represents 2 hours binned at 5m depth increments",
       caption = stringr::str_wrap("data derived from the Emperical Cumulative 
                                   Distribution (ECD) of time at depth 
                                   transmitted via the Argos satellite network")) +
  theme_minimal() +
  theme(strip.text = element_text(hjust = 1),
        legend.position = "bottom")
```

#### Spotted seals

```{r}
#| fig-asp: 1.3
ecd_data %>% 
  left_join(deploy_tbl) %>% 
  dplyr::filter(species == 'spotted seal',
                between(end,
                        deploy_date_time_gmt,
                        deploy_end_date_time_gmt)) %>%
ggplot() +
  geom_rect(aes(xmin = start, xmax = end,
                ymin = 0, ymax = percent_dry/10),
            fill = "seashell3") +
  geom_rect(aes(xmin = start, xmax = end,
                ymin = depth_break, ymax = prev_depth,
                fill = propTAD),
            color = NA) +
  scale_fill_distiller(palette = "Reds",
                   direction = 1, trans = "log10",
                       guide = guide_colorbar(
                         title = 'proportion of time submerged',
                         title.position = 'bottom',
                         title.hjust = 0.5,
                         barwidth = unit(75, units = "mm"),
                         barheight = unit(2, units = "mm"))) +
  scale_x_datetime() +
  facet_wrap(speno ~ ., ncol=1) +
  ylab("depth (meters)") +
  labs(title = "Spotted Seals (PL), 2024",
       subtitle = "each bar represents 2 hours binned at 5m depth increments",
       caption = stringr::str_wrap("data derived from the Emperical Cumulative 
                                   Distribution (ECD) of time at depth 
                                   transmitted via the Argos satellite network")) +
  theme_minimal() +
  theme(strip.text = element_text(hjust = 1),
        legend.position = "bottom")
```

:::

### Time At Depth from Histogram Bins

Two bio-loggers deployed on sub-adult spotted seals (SPENO: PL2024_1003,
PL2024_1017) were programmed to collect time at depth information based on
pre-defined histogram bins. The depth range encompassed by each bin is courser
than with the above cumulative distribution data.

```{r}
#| fig-asp: 0.618
data_list$tad %>% 
  left_join(deploy_tbl) %>% 
  dplyr::rename(start = tad_start_dt,
                depth_break = bin_upper_limit) |> 
  dplyr::arrange(deployid, start) |> 
  dplyr::mutate(end = start + lubridate::hours(2)) |>
  dplyr::filter(between(end,deploy_date_time_gmt,deploy_end_date_time_gmt)) %>%
  dplyr::filter(bin != 'bin14') |> 
  dplyr::mutate(depth_break = as.numeric(depth_break),
                pct_tad = pct_tad*0.01) |> 
  group_by(deployid, start, end) |> 
  # determine the next depth value
  dplyr::mutate(next_depth = dplyr::lead(depth_break)) %>%
  # determine previous depth for plotting
  dplyr::mutate(prev_depth = dplyr::lag(depth_break)) %>%
  dplyr::select(-next_depth) %>%
  dplyr::mutate(prev_depth = ifelse(is.na(prev_depth),0,prev_depth)) |>
  # transform depths to negative values
  dplyr::mutate(depth_break = depth_break * -1,
                prev_depth = prev_depth * -1) |> 
  dplyr::filter(!is.na(pct_tad) & pct_tad > 0) |> 
                
ggplot() +
  geom_rect(aes(xmin = start, xmax = end,
                ymin = depth_break, ymax = prev_depth,
                fill = pct_tad),
            color = NA) +
  scale_fill_distiller(palette = "Reds",
                   direction = 1, trans = "log10",
                       guide = guide_colorbar(
                         title = 'proportion of time submerged',
                         title.position = 'bottom',
                         title.hjust = 0.5,
                         barwidth = unit(75, units = "mm"),
                         barheight = unit(2, units = "mm"))) +
  scale_x_datetime() +
  facet_wrap(speno ~ ., ncol=1) +
  ylab("depth (meters)") +
  labs(title = "Spotted Seal (PL2024_1003) Time at Depth",
       subtitle = "each bar represents 2 hours binned at set depth increments",
       caption = stringr::str_wrap("histograms of time at depth 
                                   transmitted via the Argos satellite network")) +
  theme_minimal() +
  theme(legend.position = "bottom",
        strip.text = element_text(hjust = 1))
```

### Dive Behavior Records

The two sub-adult spotted seals (SPENO: PL2024_1003, PL2024_1017) were also
programmed to provide more detailed information regarding individual dives. For
each dive, the start time, end time, and maximum depth are recorded. The
subsequent post-dive interval at the surface is also recorded. This is bundled
into 5 consecutive dive + post-dive intervals. In addition to providing 
information on dive depths and dive frequency we can also explore the overall
distribution of dive duration.

```{r}
#| fig-asp: 0.618
data_list$behav |> 
  left_join(deploy_tbl) |> 
ggplot() + 
  geom_rect(aes(xmin=start,
                xmax=end,
                ymin=0,
                ymax=-1*depth_min),
                linetype=0,
                fill = rcartocolor::carto_pal(3,"Bold")[1]) +
  scale_x_datetime() +
  facet_wrap(speno ~ ., ncol=1) +
  ylab("depth (meters)") +
  labs(title = "Spotted Seal Dive Behavior",
       subtitle = "each bar represents a single dive",
  caption = stringr::str_wrap("dive behavior details 
                                   transmitted via the Argos satellite network")) +
  theme_minimal() +
  theme(axis.title = element_text(),
        strip.text = element_text(hjust = 1))
```

```{r}
#| fig-asp: 0.618
data_list$behav |> 
  left_join(deploy_tbl) |> 
  dplyr::filter(what == "Dive") |> 
  dplyr::select(speno,what,depth_max,duration_max) |> 
  ggplot(aes(x = duration_max, y = what)) +
  geom_boxplot(width = 0.05,
              color = rcartocolor::carto_pal(3,"Bold")[1]) +
  geom_point(shape = "|",
             size = 8,
             alpha = 0.15,
             position = position_nudge(y = -0.15),
             color = rcartocolor::carto_pal(3,"Bold")[1]
  ) +
  stat_slab(height = 0.75,
            position = position_nudge(y = 0.1),
            fill = rcartocolor::carto_pal(3,"Bold")[1]
  ) +
  facet_wrap(speno ~ ., ncol=1) +
  xlab("dive duration (seconds)") +
  labs(title = "Spotted Seal Dive Duration",
       subtitle = "rain cloud plot showing distribution of dive durations") +
  theme_minimal() +
  theme(
    strip.text = element_text(hjust = 1),
    axis.title.y = element_blank(),
    axis.text.y = element_blank(),
    axis.ticks.y = element_blank()
  ) +
  theme(axis.title = element_text())
  
```


## Oceanographic Temperature Profiles

Approximately two times per day, the bio-loggers will record a temperature
profile upcast from the deepest dive during a two hour summary period. This
provides unique insight into the oceanography and structure of the water these
seals are foraging in.

::: {.panel-tabset}

```{r}
pdt_data <- data_list$pdt %>% 
  group_by(deploy_id,date) %>% 
  mutate(data_hr = lubridate::round_date(date,unit="hour"),
         x1 = data_hr - minutes(90),
         x2 = data_hr + minutes(90)) %>% 
  dplyr::mutate(next_depth = dplyr::lead(depth)) %>%
  # determine previous depth for plotting
  dplyr::mutate(prev_depth = dplyr::lag(depth)) %>%
  dplyr::select(-next_depth) %>%
  # transform depths to negative values
  dplyr::mutate(depth = depth * -1,
                prev_depth = prev_depth * -1) %>% 
  ungroup() %>% 
  left_join(deploy_tbl, by=c("deploy_id" = "deployid")) %>% 
  dplyr::rename(species = common_name) |> 
  dplyr::filter(between(data_hr,deploy_date_time_gmt,deploy_end_date_time_gmt))
```

### Ribbon seals

```{r}
#| fig-asp: 1.3
pdt_data %>% 
  filter(depth<0,
         species == 'ribbon seal') %>% 
ggplot() +
  geom_rect(aes(xmin = x1, xmax = x2,
                ymin = prev_depth, ymax=depth,
                fill=min_te), colour=NA) +
  scale_fill_distiller(palette = "Spectral",
                   direction = -1,
                       guide = guide_colorbar(
                         title = 'temperature (C)',
                         title.position = 'bottom',
                         title.hjust = 0.5,
                         barwidth = unit(75, units = "mm"),
                         barheight = unit(2, units = "mm"))) +
  scale_x_datetime() +
  facet_wrap(speno ~ ., ncol=1) +
  ylab("depth (meters)") +
  labs(title = "Temperature at Depth (transmitted)",
       subtitle = "each bar depicts data from a single upcast as a 3-hour window") +
  theme_minimal() +
  theme(legend.position = "bottom",
        strip.text = element_text(hjust = 1),)
```

### Spotted seals

```{r}
#| fig-asp: 1.3
pdt_data %>% 
  filter(depth<0,speno != 'PL2024_1017',
         species == 'spotted seal') %>% 
ggplot() +
  geom_rect(aes(xmin = x1, xmax = x2,
                ymin = prev_depth, ymax=depth,
                fill=min_te), colour=NA) +
  scale_fill_distiller(palette = "Spectral",
                   direction = -1,
                       guide = guide_colorbar(
                         title = 'temperature (C)',
                         title.position = 'bottom',
                         title.hjust = 0.5,
                         barwidth = unit(75, units = "mm"),
                         barheight = unit(2, units = "mm"))) +
  scale_x_datetime() +
  facet_wrap(speno ~ .,ncol=1) +
  ylab("depth (meters)") +
  labs(title = "Temperature at Depth (transmitted)",
       subtitle = "each bar depicts data from a single upcast as a 3-hour window") +
  theme_minimal() +
  theme(strip.text = element_text(hjust = 1),
        legend.position = "bottom")
```

:::

## Satellite Transmissions

This figure provides some diagnostic and performance insights to tag deployments
and attachment locations.

::: {.panel-tabset}

### Ribbon seals

```{r}
#| fig.asp: 1.3
cols <- carto_pal(n=4,'Vivid')[1:3]
attachment <- c("Flipper","Head","Nape")

cols <- setNames(cols,attachment)

msg_data <- data_list$messages |> 
  left_join(deploy_tbl) |> 
  dplyr::filter(between(msg_date,
                        deploy_date_time_gmt,
                        deploy_end_date_time_gmt)) |> 
  dplyr::filter(common_name == 'ribbon seal') |> 
  dplyr::select(species=common_name,speno,attachment,msg_date,msg_count=msg) |> 
  dplyr::mutate(msg_day = lubridate::date(msg_date)) |> 
  dplyr::group_by(speno,attachment,msg_day) |> 
  dplyr::summarise(total_count = sum(msg_count))

ggplot(data = msg_data) + 
  geom_col(aes(x=msg_day,y=total_count,fill = attachment)) +
  scale_fill_manual(values = cols) +
  facet_wrap(speno ~ ., ncol = 1) +
  ylab("date") +
  labs(title = "Satellite Messages Recieved by Attachment Type",
       subtitle = "each bar the total number of messages recieved each day",) +
  theme_minimal() +
  theme(strip.text = element_text(hjust = 1),
        legend.position = "top",
        legend.title=element_blank())
```

### Spotted seals

```{r}
#| fig.asp: 1.3

msg_data <- data_list$messages |> 
  left_join(deploy_tbl) |> 
  dplyr::filter(between(msg_date,
                        deploy_date_time_gmt,
                        deploy_end_date_time_gmt)) |> 
  dplyr::filter(common_name == 'spotted seal') |> 
  dplyr::select(species=common_name,speno,attachment,msg_date,msg_count=msg) |> 
  dplyr::mutate(msg_day = lubridate::date(msg_date)) |> 
  dplyr::group_by(speno,attachment,msg_day) |> 
  dplyr::summarise(total_count = sum(msg_count))

ggplot(data = msg_data) + 
  geom_col(aes(x=msg_day,y=total_count,fill = attachment)) +
  scale_fill_manual(values = cols) +
  facet_wrap(speno ~ .,ncol=1) +
  ylab("date") +
  labs(title = "Satellite Messages Recieved by Attachment Type",
       subtitle = "each bar the total number of messages recieved each day",) +
  theme_minimal() +
  theme(strip.text = element_text(hjust = 1),
        legend.position = "top",
        legend.title=element_blank())
```

:::