Point Pelee Frogs and Toads

Water and Life

Water vapour amplifies the effects of the other more toxic Greenhouse Gasses and by diminishing the land we have increased the amount moving through Atmospheric Cycles. Much of our urban and agricultural infrastructure is designed to drain the land quickly which contradicts the earths natural systems.

Essex County (Ontario, Canada) was described as a Eden when it was first settled. Forests were then burned, endless marshlands, estuaries drained and the native wildlife eliminated, poisoned or pushed into very small pockets. Point Pelee is such a place, located on the southern tip of Essex county and is a source of local pride. Before you surround yourself with the beauty of the park; note the canals and pump stations keeping the great marsh in check. Imagine the potential of this marsh protecting small fish, birds, insects and plants. Imagine how healthy the surrounding lake would have been.

Restoration of our natural systems will be a crucial part of land stewardship for a better future.

---

Point Pelee Open Data

I’ve downloaded a data set and corresponding dictionary from open.canada.ca which details the results of a yearly survey of the parks frog and toad populations. I’ve created the following workbook using the RMarkdown package in RStudio. I’ve kept the data raw when importing so that you may download and achieve the same results with the following code chunks.

#load the relevant packages and data from the global environment
knitr::opts_chunk$set(echo = TRUE)
load("myData.RData")
library(tidyverse)
library(naniar)
library(leaflet)
library(htmltools)
library(rbokeh)

Listening Stations

I’ve extracted the coordinates from the dictionary and prepared a Leaflet map showing the listening stations where the surveys were completed through the years.

leaflet(pelee_station_coords, options = leafletOptions(minZoom = 12.3, dragging = TRUE)) %>% 
  addProviderTiles(provider = "Esri.WorldTopoMap") %>% 
  setView(lng = -82.515746, lat = 41.948, zoom = 12.5) %>% 
  addMarkers(~lon, ~lat, popup = ~htmlEscape(pelee_station), 
             options = markerOptions(riseOnHover = TRUE, opacity = 0.4, title = "Pelee Station")) %>% 
  addCircleMarkers(data = pelee_station_coords, radius = 7, label = "pelee_station", color = "Green")

Data Cleaning

Let’s start by renaming the data frames and installing script friendly column names

#rename the dataframes
frogs_toads_data <- Point_Pelee_NP_Frogs_and_Toads_1999_2017_data[c(-1),]
frogs_toads_dict <- Point_Pelee_NP_Frogs_and_Toads_1999_2017_data_dictionary[, c(1, 3, 4)]

#code friendly column_names
colnames(frogs_toads_data) <- c("year", "visit", "date", "start_time", 
                         "cloud_cover_10_ths", "air_temp", "wind_scale_code", 
                         "precipitation_code", "species_code", "pelee_station", 
                         "calling_code", "frog_count", "inside_100m", 
                         "outside_100m", "background_code", "notes")

colnames(frogs_toads_dict) <- c("reference", "key", "data")

Separation

The data set contains references or keys to the corresponding dictionary. In order to combine all of the variables into a single table, we will need to separate the dictionary into smaller tables and use Dplyr to piece the information into the data set.

#separate the dictionary by filtering the first column

#species code references the latin name of the frogs and toads
species_code_dict<- frogs_toads_dict %>% 
  filter(reference == "Species Code") %>% #join the english frognames
  left_join(frog_names_english, by = "key")
species_code_dict <- species_code_dict[, c(2, 4)]
colnames(species_code_dict) <- c("species_code", "frog_species")

#create a dictionary for the calling code (single, chorus)
calling_code_dict <- frogs_toads_dict %>% 
  filter(reference == "Calling Code")
calling_code_dict <- calling_code_dict[, c(2, 3)]
colnames(calling_code_dict) <- c("calling_code", "frog_call")

#Beaufort Windscale is a poetic scientific classification from the 19th century
wind_scale_dict <- frogs_toads_dict %>% 
  filter(reference == "Beaufort Wind Scale")
wind_scale_dict <- wind_scale_dict[, c(2, 3)]
colnames(wind_scale_dict) <- c("wind_scale_code", "wind_scale")

#scale noting traffic or other concerns that may affect the count
background_noise_dict <- frogs_toads_dict %>% 
  filter(reference == "Background Noise Code")
background_noise_dict <- background_noise_dict[, c(2, 3)]
colnames(background_noise_dict) <- c("background_code", "background_noise")

#raining or not very little data in this column anyway
precipitation_dict <- frogs_toads_dict %>% 
  filter(reference == "Precipitation")
precipitation_dict <- precipitation_dict[, c(2, 3)]
colnames(precipitation_dict) <- c("precipitation_code", "precipitation")

#create a dictionary for the stations and separate the Northing and Easting columns
pelee_station_dict <- frogs_toads_dict %>%
  filter(reference == "Pelee Station")  
pelee_station_dict <- separate(pelee_station_dict, data, 
          c("NAD", "UTM", "lng", "lat", "station_change", 
            "change_lng", "change_lat"), sep = " ")
pelee_station_dict_colnames <- c("reference", "pelee_station", "NAD", "UTM",
                              "lng", "lat", "station_change",
                              "change_lng", "change_lat")
colnames(pelee_station_dict) <- pelee_station_dict_colnames
#drop the redundant columns
pelee_station_dict <- pelee_station_dict[, c(2, 5, 6)]

Combination

Combining all of the variables into a single table and coercing column types.

#join the tables
amphibians <- frogs_toads_data %>%
  left_join(pelee_station_dict, by = "pelee_station") %>%
  left_join(species_code_dict, by = "species_code") %>% 
  left_join(calling_code_dict, by = "calling_code") %>% 
  left_join(precipitation_dict, by = "precipitation_code") %>% 
  left_join(wind_scale_dict, by = "wind_scale_code")

#create a date time column by pasting together the date and start_time
amphibians$date_time <- as.POSIXct(paste(amphibians$date,
                                    amphibians$start_time),
                                    format = "%Y-%m-%d %H:%M")
#mutate the columns to the apropriate type
amphibians <- amphibians %>%
  mutate(visit = as.numeric(visit),
         date = as.Date(date, format = "%Y-%m-%d"),
         cloud_cover_10_ths = as.numeric(cloud_cover_10_ths),
         air_temp = as.numeric(air_temp),
         wind_scale_code = as.numeric(wind_scale_code),
         calling_code = as.numeric(calling_code),
         frog_count = as.numeric(frog_count),
         lng = as.numeric(lng),
         lat = as.numeric(lat))

Missingness

I prefer to use the Naniar package to visualize the missingness prior to making any decisions on cleaning or disregarding the data.

#plot call for missing variables
gg_miss_var(amphibians, show_pct = TRUE)

Touching Up the Variables

Replace the NA’s with the appropriate values, fill in the empty variables with booleans or text, impute the air temperature and recode the inside; outside 100 meters.

amphibians <- amphibians %>% 
  replace_na(list(frog_count = 0, species_code = 0, frog_species = "No Count",
                  frog_call = 0, notes = "No Observation", inside_100m = "FALSE",
                  outside_100m = "FALSE", background_code = "No Observation", 
                  precipitation = "No Observation", wind_scale_code = 0,
                  wind_scale = "No Observation", cloud_cover_10_ths = 0)) %>% 
  mutate(air_temp = impute_mean(air_temp)) %>% 
  mutate(inside_100m = recode(inside_100m, "X" = "TRUE")) %>% 
  mutate(outside_100m = recode(outside_100m, "X" = "TRUE")) %>% 
  drop_na(date_time)

Lonely Nights

I’ve created a plot using the tidyverse package to manipulate the data and pipe it into an rbokeh object, detailing the number of nights each year that teams did not hear any frog or toad calls. Just really felt bad for the researchers gearing up to sit in a marsh for nothing.

Please Consider the following plot a tribute to their dedication.

zero_frogs_toads <- amphibians %>% 
  select(year, pelee_station, frog_count, date_time) %>% 
  filter(frog_count == 0) %>% 
  group_by(year, pelee_station) %>% 
  summarise(no_amphibians = n_distinct(pelee_station)) %>% 
  group_by(year) %>% 
  summarize(nights_without_frogs = sum(no_amphibians)) %>% 
    figure(title = "Lonely Nights") %>% 
      ly_bar(x = year, y = nights_without_frogs,
             hover = TRUE) %>% 
      theme_axis("x", major_label_orientation = 90) %>% 
      theme_grid(grid_line_color = "#000000") %>% 
      theme_plot(background_fill_color = "#000000")

zero_frogs_toads

Visualizing the Counts

The final plot details the species counts yearly. The rbokeh object will allow you to zoom in and hover over the object to see the species name and counts.

amphibians_bar <- amphibians %>%
  select(year, date_time, frog_species, frog_count) %>%
  filter(frog_count != 0) %>% 
  group_by(year) %>% 
  mutate(species_totals = sum(frog_count)) %>% 
    figure(title = "Point Pelee Frogs and Toads Survey") %>% 
    ly_bar(x = year, y = frog_count, 
           color = frog_species, 
           hover = TRUE) %>%
    theme_plot(background_fill_color = "#FFFFFF",
               outline_line_color = "white") %>% 
    theme_legend(background_fill_alpha = 0.1, 
                 background_fill_color = "#FFFFFF", 
                 label_text_alpha = 0.5) %>% 
    theme_axis("x", major_label_orientation = 90) %>% 
    set_palette(discrete_color = pal_color(c("#6C51A1", "#86F29C", 
       "#CF93F5","#538399", "#05F656", "#5179A1"))) %>% 
    theme_grid(grid_line_color = "#FFFFFF")
  
amphibians_bar

Final Thoughts

I believe that there are simple solutions to the problems that humanity faces and that there is an abundance of wisdom and wise people lifting up their voices for our collective future.

Increasing habitat potential will increase the potential for insects and smaller fish supporting the frog and toad populations, which in turn provide food for birds and smaller mammals.

If I had billions, I would actively be seeking out land for purchase and re-naturalization. My hope is that private citizens can make small changes to their lifestyles, maybe keeping pesticides out of yards, leaving a few areas with long grass for shelter and possibly a stretch of soft shoreline.

One of the fundamental lessons in nature is balance and humanity has created so many challenges. I’m very optimistic that we can find a way to coexist and thrive alongside natural systems.