Koala Conservation
Jenny Green, Natalie Musick
12/16/2019
Are koalas functionally extinct?
Background
In May 2019, the Australian Koala Foundation stated that there are now less that 80,000 koalas left in Australia, establishing them as ‘functionally extinct.’ Functional extinction is when populations decrease to the point that the species no longer plays a significant role in their ecosystem, making the population no longer viable. It is believed that while some individuals can reproduce, the limited number of koalas makes the long-term viability of the species unlikely and highly susceptible to disease (Nace, 2019).
Koalas have been a well known charismatic species throughout time, and their sharp decline has been noticeable. Their main threats include deforestation and disease spread (such as chlamydia); both of which are expemlified by recent large fires in Australia and climate chagge. Their status has been debated as the population numbers are difficult to solidify. Various sources state conflicting ideas about the status of Koalas. Some articles declare their functional extinction whereas others say this is not actually occuring. Figuring out whether or not they are truly functionally extinct in certain areas of Australia is very important to conservation efforts. Thus, in this report, we will delve into available koala population data and their threats, using spatial configurations, to see how Koala’s should be classified in terms of conservation status.
Install Packages
Bring in Data
redlistData <- sf::read_sf("data/redlist_species_data_9adba45f-de93-488b-8d3f-8e9e32dbe8e6/data_0.shp")
qlForest <- sf::read_sf("data/QSC_Extracted_Data_20191207_053058712000-11408/Defined_forest_area.shp")
koala_locations <- read_csv("data/wildnetkoalalocations-2019-06-19.csv")## Parsed with column specification:
## cols(
## Longitude = col_double(),
## Latitude = col_double(),
## ScientificName = col_character(),
## SiteVisitID = col_double(),
## StartDate = col_character(),
## EndDate = col_character(),
## SiteID = col_double(),
## SiteCode = col_character(),
## LocalityDetails = col_character(),
## LocationPrecision = col_double(),
## ProjectID = col_double(),
## ProjectName = col_character(),
## OrganisationID = col_double(),
## OrganisationName = col_character(),
## OrganisationAcronym = col_character()
## )koalaCount <- read_csv(file = "data/records-2019-12-10/records-2019-12-10.csv")## Parsed with column specification:
## cols(
## .default = col_character(),
## `Catalogue Number` = col_logical(),
## `Vernacular name - original` = col_logical(),
## Subspecies = col_logical(),
## `Institution ID` = col_logical(),
## Institution = col_logical(),
## `Collection ID` = col_logical(),
## Collection = col_logical(),
## `Institution Code` = col_logical(),
## `Collection Code` = col_logical(),
## Locality = col_logical(),
## `Latitude - original` = col_double(),
## `Longitude - original` = col_double(),
## `Geodetic datum - original` = col_logical(),
## Latitude = col_double(),
## Longitude = col_double(),
## `Coordinate Precision` = col_logical(),
## `Coordinate Uncertainty in Metres` = col_logical(),
## `Minimum elevation in meters` = col_logical(),
## `Maximum elevation in meters` = col_logical(),
## `Minimum depth in meters` = col_logical()
## # ... with 10 more columns
## )## See spec(...) for full column specifications.koalaC <- koalaCount %>% dplyr::select("Latitude", "Longitude", "Country - parsed", "State - parsed", "Year", "Event Date - parsed", "Scientific Name - original") %>% drop_na("Latitude", "Longitude")
koalaCsp <- st_as_sf(koalaC, coords = c("Longitude", "Latitude"), crs = 4326)
AustraliaBoundaries <- read_sf("data/Australia_Boundaries/aust_cd66states.shp")
#can filter to only states with koala sightings
AustStates <- AustraliaBoundaries %>%
mutate(state = case_when(STE == 1 ~ "New South Wales",
STE == 2 ~ "Victoria",
STE == 3 ~ "Queensland",
STE == 4 ~ "South Australia",
STE == 5 ~ "Western Australia",
STE == 6 ~ "Tasmania",
STE == 7 ~ "Northern Territory"))
fires <- read_sf("data/MODIS_C6_Australia_and_New_Zealand_7d/MODIS_C6_Australia_and_New_Zealand_7d.shp")#https://www.data.qld.gov.au/dataset/wildnet-koala-locations
startdate <- dmy(koala_locations$StartDate)
enddate <-dmy(koala_locations$EndDate)
end.year <- as.numeric(format(enddate,'%Y'))
start.year <- as.numeric(format(startdate,'%Y'))
koala <- koala_locations %>% mutate(Date_lastSeen = enddate,
year_lastSeen = end.year,
Date_found = startdate,
year_found = start.year) %>%
dplyr::select("Longitude", "Latitude",
"Date_lastSeen", "year_lastSeen",
"Date_found", "year_found")
class(koala$Date_found)## [1] "Date"koalaSf <- st_as_sf(koala, coords = c("Longitude", "Latitude"), crs = 4326)
qKoala <- koalaSf %>% filter(year_found >= 2010 & year_found <= 2020)
#Set CRS AustStates
crs_1 <- "+proj=longlat +ellps=WGS84 +no_defs"
st_crs(AustStates) <- crs_1Given “WildNet” Koala locations provided by the Queensland government, we display an animated version (by time) of the dates and locations in which koalas were sighted during field research in Queensland between 1970 to present. Sightings are presented in longitude and latitude, and are in order of when the studies started and ended.
koala_anim <- ggplot(koala, aes(Longitude, Latitude, group = koala$Date_found)) + geom_point() + transition_time(Date_found) + ggtitle('Wildnet Koala Locations', subtitle = '1970-Present')
koala_anim
To display this in another form, this histogram shows that the most amount of studies and therefore sightings took place around 2000.
hist <- koala %>% filter(Date_found > 1950) %>% ggplot(aes(Date_found)) + geom_histogram() + labs(title = "Queensland Koala Field Surveys",
x = "Year",
y = "Koala Sightings During Research")
hist## `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.
Vegetation Loss Over Time
Currently, the main threat to koalas is habitat loss. Below we display loss of vegetation in New South Wales and Queensland, the states of Australia in which koalas exsist. This data is derived from Global Forest Watch, based on surveys between 2001 and 2018.
#data from Global Forest Watch
Aus_veg <- readxl::read_xlsx("data/Aus1.xlsx") %>% filter(subnational1 == "New South Wales" | subnational1 == "Queensland") %>% filter(threshold == 50) %>% dplyr::select("subnational1", ("tc_loss_ha_2001": "tc_loss_ha_2018")) %>%
rename("2001" ="tc_loss_ha_2001", "2002" = "tc_loss_ha_2002", "2003" = "tc_loss_ha_2003", "2004" = "tc_loss_ha_2004","2005" = "tc_loss_ha_2005","2006" = "tc_loss_ha_2006","2007" = "tc_loss_ha_2007", "2008" = "tc_loss_ha_2008","2009" = "tc_loss_ha_2009", "2010" = "tc_loss_ha_2010", "2011" = "tc_loss_ha_2011", "2012" = "tc_loss_ha_2012", "2013" = "tc_loss_ha_2013", "2014" = "tc_loss_ha_2014", "2015" = "tc_loss_ha_2015", "2016" = "tc_loss_ha_2016", "2017" = "tc_loss_ha_2017", "2018" = "tc_loss_ha_2018", "Nation" = "subnational1")
pivoted <- pivot_longer(Aus_veg, names_to = "Years", cols = ("2001": "2018"))
pivoted <- pivoted %>% mutate(years = as.integer(pivoted$Years))
pivoted1<- pivoted %>% dplyr::select("Nation", "years", "value")
vegetation_graph <- pivoted1 %>% ggplot(aes(x = years, y = value, color = Nation)) + geom_line()+ labs(title = "Vegetation Loss Over Time",
x = "Year",
y = "Hectares Lost")
vegetation_graph
Here we can see that vegetation loss per year has increased since 2001, with significant spikes in losses around 2001, 2007, 2013, and 2016 in New South Wales. Each point represents the amount of vegetation loss per year, which, accumulates with time, meaning that the amount of vegetation loss is accelerating. Additionally, the value of land loss was most likeley not 0 prior to 2001, rather, there was no data collected then. It is important to keep in mind that koalas are specialized to certain species of eucalyptus, which is a fraction of this vegetation loss. There was no provided data of just eucalyptus loss, however, eucalyptus is the most common tree in both New South Wales and Queensland, so this data should be largely representative of the eucalyptus species that koalas inhabit. Furthermore, the data had varying levels of “thresholds,” which represents the percent of tree cover with canopy density. We used the threshold value that best represented preferred koala habitat.
The increasing trend, as well as the various spikes contributes to loss of koalas as they depend on vegetation for their habitat and diet. While this can not signify whether or not koalas are now “functionally extinct,” it helps explain their rapid loss.
Interactive Map Displaying Fires and Koala Locations
allData <- tm_shape(AustStates, name = "Australian States") + tm_polygons(col="state", id="state", legend.show = FALSE) +
tm_shape(redlistData, name = "Acceptable Koala Habitat Area, Redlist", id="Koala Sighting") + tm_polygons(col = "blue") +
tm_shape(fires, name = "Australia Bush Fires, 2019") + tm_dots(col = "red", id="Fire Location") +
tm_shape(qKoala, name = "Koala Counts, Queensland Data") + tm_dots(col = "year_found", style = "cat", title = "Year of Koala Sighting", id="Koala Sighting") +
tm_shape(koalaCsp, name = "Koala Counts, Citizen Science") + tm_dots(col = "Year", style = "cat", legend.show = FALSE, id="Koala Sighting") +
tm_layout(title = "Koala Threats & Counts Map", legend.show = FALSE)
allData %>%
tmap_leaflet(mode = "view", add.titles = TRUE)