flow_dash_ncoast.qmd

---
title: "FLOWdash N Coast"
logo: assets/wdfw_logo_stacked_fullcolor.png
format: 
  dashboard:
    orientation: rows
theme: lux
embed-resources: true
---

```{r setup, include=FALSE, warning=FALSE, message=FALSE}
#knitr::opts_chunk$set(echo = FALSE, warning = FALSE, message = FALSE, fig.width = 10, fig.height = 13)

library("tidyverse", quietly = T)
theme_set(theme_minimal()) 

# #workbook of basic site/station metadata: `usgs_sites`
# readxl::read_excel("~/T/DFW-Team WDFW Watershed Synthesis - flow_trees_heat/usgs_sites_dailyQ_focal.xlsx") |>
#   select(site_no, station_nm) |>
#   mutate(station_nm = str_remove(station_nm, ", WA$") |> 
#            str_replace("RIVER", "R") |> 
#            str_replace("NEAR", "Nr") |> 
#            str_replace("HIGHWAY", "HWY") |> 
#            str_to_title()) |>
#   slice(19:16) |> unlist()

sites <- data.frame(
  site_no = c("12043000","12041200","12040500","12039500"),
  station_nm = c("Calawah R Nr Forks","Hoh R At Us Hwy 101 Nr Forks","Queets R Nr Clearwater","Quinault R At Quinault Lake")
)
  
q_obs_range <- c(as.Date("1979-01-01"), Sys.Date())


#build uniform object inserting NAs for missing obs during q_obs_range
#rebuild/overwrite y, m, and yday since `complete` inserts many NAs
#then calc per site per CALENDAR year 
# - cumulative sum of daily mean flow (weird but useful proxy for overall annual volume, 'wet/dry year')
# - 7day moving average of daily mean flow
usgs_q_dv <- list.files("data", pattern = paste0("usgs_dailyQ_", sites$site_no, collapse = "|"), full.names = T) |>
  map_df(~readRDS(.x)) |> 
  #group_by(site_no) |> summarise(dmin = min(date), dmax = max(date))
  tidyr::complete(site_no, date = full_seq(q_obs_range, 1))|> 
  mutate(
    year = year(date),
    month = month(date), yday = yday(date)
    #,q_dv_mean = if_else(q_dv_mean < 0, NA_real_, q_dv_mean)
  ) |> 
  # mutate(
  #   q_dv_mean_sum = cumsum(q_dv_mean),
  #   #,q_dv_mean_7d = slider::slide_dbl(q_dv_mean, ~mean(., na.rm=T), .before = 3, .after = 3),
  #   .by = c(site_no, year)
  # ) |> 
  left_join(sites, by = "site_no") |> 
  select(site_no, station_nm, everything())

##add yday median daily Q across years (all and decadal) 
usgs_q_dv <- bind_rows(
  usgs_q_dv |> mutate(year = as.character(year))
  ,
  map2_df(
    c(1980,1990,2000,2010,1979),
    c(1989,1999,2009,2019,2024),
    ~usgs_q_dv |>
      filter(between(year, .x, .y)) |> 
      summarise(
        year = paste0("median_",.x,"_",.y),
        q_dv_mean = median(q_dv_mean),
        .by = c(site_no, yday)
      )
  )
)

usgs_bfs <- readRDS("data/usgs_bfs_pred.rds") |> 
  filter(site_no %in% sites$site_no)

nwm_fcst <- list.files("data", pattern = paste0("nwm_", sites$site_no, collapse = "|"), full.names = T) |>
  map_df(~readRDS(.x))

nwfsc_st <- list.files("data", pattern = paste0("nwfsc_st_pred_", sites$site_no, collapse = "|"), full.names = T) |>
  map_df(~readRDS(.x)) |> 
  mutate(
    year = as.character(year(date)), month = month(date), yday = yday(date)
  )

crc_sites <- list.files("data", pattern = paste0("wdfw_crc_", sites$site_no, collapse = "|"), full.names = T) |>
  map_df(~readRDS(.x)) |> 
  left_join(sites, by = "site_no") |> 
  select(site_no, station_nm, everything())

  
#rr_wb_coho <- readRDS("data/rr_wb_coho.rds")

rangeslider_thickness <- 0.05
#wacolors::pal_vector("ferries",n=46)
#length(unlist(wacolors::wacolors[1:9])) #50
#wacolors::pal_vector("washington_pass",n=length(2024:1979))

pal <- set_names(
  c(
    rep(c(wacolors::wacolors$washington_pass, wacolors::wacolors$palouse), 
        length.out = length(2024:1979)),
    # #median(s)
    c("#B96000","darkblue","lightgreen","grey80","cyan"),
    #nwm forecast
    alpha("purple",c(0.8,0.4)),
    #usgs baseflow forecast
    "#DF3383", rep("#8A6172",2)
    ),
  c(2024:1979, 
    usgs_q_dv |> distinct(year) |> filter(str_detect(year, "median")) |> pull(year) |> sort(), # 'median',
    'nwm_mr_mean', 'nwm_lr_mean',
    'usgs_bfs','usgs_bfs_05','usgs_bfs_95'
    ))

plot_q_dv <- function(site, log10 = T){

  d <- usgs_q_dv |>
    filter(site_no==site) |>
    select(date, year, yday, q_dv_mean) |> 
    bind_rows(
      nwm_fcst |> 
        filter(site_no == site, str_detect(memb, "mean")) |> 
        select(yday, year = memb, q_dv_mean = cfs)
    )
  
  if(site %in% unique(usgs_bfs$site_no)){
    d <- bind_rows(
      d,
      usgs_bfs |> 
        filter(site_no == site) |> 
        select(yday, starts_with("usgs_bfs")) |> 
        pivot_longer(cols = starts_with("usgs_bfs"), 
                     names_to = 'year',
                     values_to = 'q_dv_mean'
        )
      )
  }
  
  p <- d |> 
    plotly::plot_ly(
      type = 'scatter', mode = 'lines',
      name = ~year, x = ~yday, y = ~q_dv_mean,
      color = ~year, colors = pal,
      hovertext = ~format(date, format = '%b-%d')
    ) |>
    plotly::layout(
      legend = list(traceorder = 'reversed'),
      xaxis = list(
        title = 'Day of year', 
        ticktext = format(seq.Date(as.Date("2024-01-01"),as.Date("2024-12-01"), by = 'month'),  format = '%b-%d'),
        tickvals = yday(seq.Date(as.Date("2024-01-01"),as.Date("2024-12-01"), by = 'month'))
        ),
      yaxis = list(title = 'cfs')
    )

  if (log10){
    p <- p |> plotly::layout(yaxis = list(type = "log"))
  }

  p
}
#plot_q_dv(sites$site_no[1])

plot_q_quant <- function(site){
  usgs_q_dv |> 
    filter(site_no == site, year != "2024") |>
    group_by(year) |> 
    summarise(
      q10 = quantile(q_dv_mean, p = 0.10, na.rm = T),
      q05 = quantile(q_dv_mean, p = 0.05, na.rm = T),
      q01 = quantile(q_dv_mean, p = 0.01, na.rm = T),
      min = min(q_dv_mean),
      .groups = "drop") |>
    pivot_longer(-year, names_to = "quantile", values_to = "cfs") |>
    plotly::plot_ly(
      type = "bar", x = ~year, y = ~cfs,
      color = ~quantile, 
      colors = c(q10 = "darkblue",
                 q05 = "blue",
                 q01 = "lightblue",
                 min = "#016C72"
                 )) |>
    plotly::rangeslider(thickness = rangeslider_thickness) |>
    plotly::layout(
      #barmode = 'overlay',
      xaxis = list(title='')
    )
} 

plot_t_dv <- function(site){

  d <- nwfsc_st |>
      filter(site_no==site) |>
      select(date, year, yday, st_pred)
  
  p <- d |> 
    plotly::plot_ly(
      type = 'scatter', mode = 'lines',
      name = ~year, x = ~yday, y = ~st_pred,
      color = ~year, colors = pal,
      hovertext = ~format(date, format = '%b-%d')
    ) |>
    plotly::layout(
      legend = list(traceorder = 'reversed'),
      shapes = list(
        list(type = "rect", fillcolor = "red", opacity = 0.2, 
             y0 = 20, y1 = 22, x0 = 0, x1 = 366),
        list(type = "rect", fillcolor = "orange", opacity = 0.2, 
             y0 = 18, y1 = 20, x0 = 0, x1 = 366),
        list(type = "rect", fillcolor = "yellow", opacity = 0.2, 
             y0 = 16, y1 = 18, x0 = 0, x1 = 366)
      ),
      xaxis = list(
        title = 'Day of year', 
        ticktext = format(seq.Date(as.Date("2024-01-01"),as.Date("2024-12-01"), by = 'month'),  format = '%b-%d'),
        tickvals = yday(seq.Date(as.Date("2024-01-01"),as.Date("2024-12-01"), by = 'month'))
        ),
      yaxis = list(title = 'degC')
    )

  p
}

plot_crc <- function(site){
  d <- crc_sites |>
    filter(
      site_no == site,
      between(statmonth, 6, 10)
    ) |> 
    arrange(year, statmonth) |> 
    mutate(
      year = as.character(year),
      statmonth = factor(statmonth, levels = 6:10),
      year_species = paste0(year,"_",species),
      est_tot = cumsum(est), .by = c(site_no, station_nm, species, year)
      ) |> 
    select(year, species, year_species, statmonth, est_tot)

  p <- d |> 
    plotly::plot_ly(
      type = 'scatter', mode = 'lines+markers',
      name = ~year_species, x = ~statmonth, y = ~est_tot, symbol = ~species,
      color = ~year, colors = pal
    ) |>
    plotly::layout(
      legend = list(traceorder = 'reversed'),
      xaxis = list(title = 'Month'),
      yaxis = list(title = 'Cumulative catch est.')
    )

  p
}

```

```{r data_rebuild_q_dv, eval=FALSE}
#not `complete()` here since no reason to store potentially lots of NA for long qobsrange
walk(
  sites$site_no
  ,
  ~dataRetrieval::readNWISdv(
    .x, parameterCd = "00060",
    startDate = q_obs_range[1],
    endDate = q_obs_range[2]
    ) |>
    as_tibble() |>
    mutate(year = year(Date), month = month(Date), yday = yday(Date)) |>
    select(site_no, date = Date, year, month, yday, q_dv_mean = X_00060_00003) |>
    saveRDS(paste0("data/usgs_dailyQ_", .x,".rds"))
)
```

```{r data_rebuild_bfs, eval=FALSE}
#add new USGS baseflow predictions...
huc4 <- c("1710","1711")
url <- paste0("https://wa.water.usgs.gov/projects/baseflows/out/bfprj_HUC",huc4,".csv")
map_df(url, ~readr::read_csv(.x)) |>
  mutate(yday = yday(Date)) |>
  select(
    site_no = SiteID, date = Date, yday,
    usgs_bfs = Baseflow.cfs,
    usgs_bfs_05 = StreamflowCB05.cfs,
    usgs_bfs_95 = StreamflowCB95.cfs
    ) |>
    saveRDS("data/usgs_bfs_pred.rds")
```

```{r data_rebuild_nwm, eval=FALSE}
sites$COMID <- map_int(
  sites$site_no,
  ~nhdplusTools::discover_nhdplus_id(
    nldi_feature = list(featureSource = "nwissite",
                        featureID = paste0("USGS-",.x)))
)

get_nwm <- function(comid){
  mr <- httr2::request(paste0("https://api.water.noaa.gov/nwps/v1/reaches/",comid,"/streamflow?series=medium_range")) |> 
    httr2::req_headers(Accept = "application/json") |> 
    httr2::req_perform() |> 
    httr2::resp_body_json() |> 
    purrr::pluck("mediumRange")
  lr <- httr2::request(paste0("https://api.water.noaa.gov/nwps/v1/reaches/",comid,"/streamflow?series=long_range")) |> 
    httr2::req_headers(Accept = "application/json") |> 
    httr2::req_perform() |> 
    httr2::resp_body_json() |> 
    purrr::pluck("longRange")
  #probably a more elegant way to do this...
  nwm <- set_names(
    c(mr, lr),
    c(paste0("mr_",names(mr)),paste0("lr_",names(lr)))
  ) 

  nwm <- map_df(
    names(nwm),
    ~bind_rows(nwm[[.x]]$data) |> mutate(memb = paste0("nwm_",.x))
  ) |>
    mutate(
      yday = lubridate::yday(lubridate::as_datetime(validTime))
    ) |>
    summarise(
      cfs = mean(flow), .by = c(memb,yday)
    )
  
  return(nwm)
}

# get_nwm(sites$COMID[2]) -> nwm
# nwm |> pivot_wider(names_from = memb, values_from = flow) |> print(n=Inf)

nwm <- map2(
  sites$site_no, sites$COMID,
  ~get_nwm(.y) |> 
    mutate(site_no = .x) 
  )

walk(nwm,  ~saveRDS(.x, file = paste0("data/nwm_",.x$site_no[1],".rds")))

```

```{r data_rebuild_crc, eval=FALSE}
mdb_file_path <- "~/T/DFW-Team WDFW Watershed Synthesis - data_common/crc/Sport Harvest Estimates 20230213.mdb"

crc <- inner_join(
  readr::read_csv(I(
    system2(
      "mdb-export",
      args = paste(str_replace_all(mdb_file_path, " ", "\\\\ "),"Area"),
      stdout = T)
  ))
  ,
  readr::read_csv(I(
    system2(
      "mdb-export",
      args = paste(str_replace_all(mdb_file_path, " ", "\\\\ "),"Catch"),
      stdout = T)
  ))
  ,  by = "AreaID"
) |>
  select(
    AreaCode, AreaName, AreaType, AreaWRIA,
    CatchYear, CatchStatMonth, Species, CatchEst #, CatchVariance?
  ) |>
  rename_with(~tolower(.) |> str_remove("catch")) |>
  filter(
    #year >= 2000,
    species %in% c("Coho","Chinook")
  )

crc |>
  filter(str_detect(areaname, "Hoh|Queets|Quinault|Sol|Calawah|Bogach")) |>
  count(areacode, areaname)

#Calawah, combining Bogachiel, Calawah, Sol Duc
#may need to revise?
crc |> 
  filter(areacode %in% c("398","400","406")) |> 
  mutate(site_no = sites$site_no[1]) |> 
  group_by(site_no, year, statmonth, species) |> 
  summarise(est = sum(est), .groups = "drop") |> 
  saveRDS(file = paste0("data/wdfw_crc_",sites$site_no[1],".rds"))

#Hoh, combining above & below Oxbow areas
crc |> 
  filter(areacode %in% c("348","350")) |> 
  mutate(site_no = sites$site_no[2]) |> 
  group_by(site_no, year, statmonth, species) |> 
  summarise(est = sum(est), .groups = "drop") |> 
  saveRDS(file = paste0("data/wdfw_crc_",sites$site_no[2],".rds"))

#Queets
crc |> 
  filter(areacode == "394") |> 
  mutate(site_no = sites$site_no[3]) |> 
  saveRDS(file = paste0("data/wdfw_crc_",sites$site_no[3],".rds"))
#Quinault
crc |> 
  filter(areacode == "410") |> 
  mutate(site_no = sites$site_no[4]) |> 
  saveRDS(file = paste0("data/wdfw_crc_",sites$site_no[4],".rds"))

```

```{r data_rebuild_st_pred, eval=FALSE}
#could also think about adding/integrating NWM via AWS flow_trees_apps.qmd>>nwm_zarr_pull2 

# # #single HUC10 prediction file is n-COMIDs by n-days in 1990-2021
# # #need to know gage HUC10 to figure out model prediction file
# # #then need to know gage COMID since values are by COMID-day
# # #but gage/reach may or may not actually be in the dataset

# #gets HUC10 by service given an sf object
# #this workbook already has HUC8s per gage...

sites <- sites |> 
  left_join(
    readxl::read_excel("~/T/DFW-Team WDFW Watershed Synthesis - flow_trees_heat/usgs_sites_dailyQ_focal.xlsx") |>
      filter(site_no %in% sites$site_no) |> 
      select(site_no, lon = dec_long_va, lat = dec_lat_va) |>
      mutate(
        huc10 = map2_chr(
          lon, lat,
          ~suppressMessages(
            nhdplusTools::get_huc(
              sf::st_as_sf(data.frame(lon = .x, lat = .y), coords = c("lon","lat"), crs = sf::st_crs(4326)),
              type = 'huc10')$huc10)
        )
      )
    , by = "site_no"
  )

sites$COMID <- map_int(
  sites$site_no,
  ~nhdplusTools::discover_nhdplus_id(
    nldi_feature = list(featureSource = "nwissite",
                        featureID = paste0("USGS-",.x)))
)

#big object
stp <- map_df(
  sites$huc10,
  ~read_csv(paste0("~/T/DFW-Team WDFW Watershed Synthesis - data_common/st_pred/st_pred_171001/st_pred_",.x,".csv")) |>
  select(date = tim.date, COMID, st_pred = prd.stream_temp)
  ) |>
  drop_na(st_pred)

# 3 of 4 for N Coast gages
distinct(stp, COMID) |> 
  inner_join(sites, by = "COMID") |> 
  split(~site_no) |> 
  map(
    ~left_join(
      .x |> select(site_no, COMID),
      stp,
      by = "COMID") |> 
      saveRDS(paste0("data/nwfsc_st_pred_",.x$site_no,".rds"))
  )

# #do not want HUC10 'average' steam temp across COMIDs
# #but could also take max daily pred across spatial range
# #would be better to add HUC10 stratification, mutate in col in map_df
# stp |>
#   group_by(date) |>
#   summarise(st_pred_max = max(st_pred, na.rm = T))

```

# About

## Row

::: {.card title="Daily flows - observed & forecast"}
This card displays mean daily streamflow per-day-of-year overlaid by year, with the median per day across years for reference.

In addition, [current predictions](https://wa.water.usgs.gov/projects/baseflows/BFS_downloads_index.html) and confidence intervals from the USGS WA Water Science Center near-term [baseflow forecast](https://www.usgs.gov/tools/baseflow-forecasts-selected-sites-united-states) are shown alongside National Water Model [(NWM)](https://water.noaa.gov/about/nwm) reach streamflow forecasts from NOAA's National Water Prediction Service [NWPS API](https://api.water.noaa.gov/nwps/v1/docs/). 

As for all other cards:

  - click the lower right corner to expand the card
  
  - double-click on any legend entry to highlight it (then single-click others to add individually or double-click again to return all) 
  
  - zoom to any area of interest.
 
built `r Sys.time()`
:::

## Row

::: {.card title="Annual low flows: Q10, Q05, Q01, Qmin"}
This card displays per-year minimum values of daily mean streamflow volume in cfs (cubic feet per second) as measured at the USGS gaging stations (`r paste(unlist(unite(sites, col = "nn", sep = " ")), collapse = " & ")`).

[`r sites$site_no[1]`](`r paste0('https://waterdata.usgs.gov/monitoring-location/',sites$site_no[1],'/#parameterCode=00060&period=P365D&showMedian=true')`)

[`r sites$site_no[2]`](`r paste0('https://waterdata.usgs.gov/monitoring-location/',sites$site_no[2],'/#parameterCode=00060&period=P365D&showMedian=true')`)

[`r sites$site_no[3]`](`r paste0('https://waterdata.usgs.gov/monitoring-location/',sites$site_no[3],'/#parameterCode=00060&period=P365D&showMedian=true')`)

[`r sites$site_no[4]`](`r paste0('https://waterdata.usgs.gov/monitoring-location/',sites$site_no[4],'/#parameterCode=00060&period=P365D&showMedian=true')`)

In addition to the annual minimum, per-year sample quantiles illustrate longer term relative differences at several low flow magnitudes. For example, ~90% of observed daily mean flows in a year were greater than and ~10% were less than the dark blue 'Q10'.
:::

::: {.card title="Annual Catch Record Card estimates"}
This card displays annual estimated Chinook and coho recreational harvest in the [CRC](https://wdfw.wa.gov/licenses/fishing/catch-record-card) data records. 
:::

::: {.card title="Estimated daily stream temperatures, Siegel et al. 2023"}
This card displays the [Siegel et al. 2023](https://journals.plos.org/water/article?id=10.1371/journal.pwat.0000119) estimated daily stream temperature for the medium resolution NHD+ COMID (flowline+local catchment) associated with the displayed USGS streamflow gage.

Fitting to the NorWeST database, "This model reflects mechanistic processes using publicly available climate and landscape covariates in a Generalized Additive Model framework. We allowed covariates to interact while accounting for nonlinear relationships between temporal and spatial covariates to better capture seasonal patterns."

Values displayed here are from the [publicly available datasets of results](https://zenodo.org/records/8174951).
:::


```{r, include=FALSE}
i <- 1
```

# `r unite(sites,"no_nm")$no_nm[i]`

## Row

```{r}
#| title: Flows - observed & forecast
plot_q_dv(site = sites$site_no[i], log10 = T)
```

## Row

```{r}
#| title: Annual low flows
plot_q_quant(sites$site_no[i])
```

```{r}
#| title: Catch (CRC)
plot_crc(sites$site_no[i])
```

```{r}
#| title: Stream temp (Siegel et al. 2023)
plot_t_dv(sites$site_no[i])
```


```{r, include=FALSE}
i <- 2
```

# `r unite(sites,"no_nm")$no_nm[i]`

## Row

```{r}
#| title: Flows - observed & forecast
plot_q_dv(site = sites$site_no[i], log10 = T)
```

## Row

```{r}
#| title: Annual low flows
plot_q_quant(sites$site_no[i])
```

```{r}
#| title: Catch (CRC)
plot_crc(sites$site_no[i])
```

```{r}
#| title: Stream temp (Siegel et al. 2023)
plot_t_dv(sites$site_no[i])
```


```{r, include=FALSE}
i <- 3
```

# `r unite(sites,"no_nm")$no_nm[i]`

## Row

```{r}
#| title: Flows - observed & forecast
plot_q_dv(site = sites$site_no[i], log10 = T)
```

## Row

```{r}
#| title: Annual low flows
plot_q_quant(sites$site_no[i])
```

```{r}
#| title: Catch (CRC)
plot_crc(sites$site_no[i])
```

```{r}
#| title: Stream temp (Siegel et al. 2023)
plot_t_dv(sites$site_no[i])
```


```{r, include=FALSE}
i <- 4
```

# `r unite(sites,"no_nm")$no_nm[i]`

## Row

```{r}
#| title: Flows - observed & forecast
plot_q_dv(site = sites$site_no[i], log10 = T)
```

## Row

```{r}
#| title: Annual low flows
plot_q_quant(sites$site_no[i])
```

```{r}
#| title: Catch (CRC)
plot_crc(sites$site_no[i])
```

```{r}
#| title: Stream temp (Siegel et al. 2023)
plot_t_dv(sites$site_no[i])
```