02 - Turnover with endemic species

02_A_beta-endemics-countries.qmd

Overview

Here we process the data on species native distribution filtering by species with only one occurrence of native distribution per country. This was done in order to reduce the effects of species with large distribution ranges in the the metrics of native turnover and NBT turnover

Packages

library(readr)        # reading CSV files
library(here)         # constructing file paths
library(dplyr)        # data manipulation
library(tidyr)        # data tidying

Data

# From 01_C_data_preparation.qmd
spp_native_distribution <- readr::read_csv(here::here("data", "raw", "spp_native_distribution.csv")) 

# From 01_C_data_preparation.qmd
spp_type_distribution <- readr::read_csv(here::here("data", "raw", "spp_type_distribution.csv")) 

Data processing

Filtering data to keep only species with one occurrence records in native distribution

spp_endemic_native <- 
  spp_native_distribution |> 
  group_by(species) |>
  add_count(name = "n.species") |> 
  filter(n.species == 1) |> 
  ungroup() |> 
  group_by(country_distribution) |> 
  add_count(name = "n.endemics.country")

df_native_endemic_grid <- 
  spp_endemic_native |> 
  dplyr::select(grids = country_distribution, 
                species) |> 
  tidyr::drop_na(grids)

df_type_grid <- 
  spp_type_distribution |> 
  dplyr::select(grids = country_museum, 
                species) |> 
  tidyr::drop_na(grids) |>
  dplyr::mutate(grids = paste(grids, "type", sep = "_"))
  

# joining data frames
df_all_grid <- rbind(df_native_endemic_grid, df_type_grid) # joining both matrices - 
    #native and types composition


#### Just descriptive quantities
country_native_endemic <- unique(df_native_endemic_grid$grids)
country_type <- gsub(pattern = "_type", 
                     replacement = "",
                     unique(df_type_grid$grids))
country_type_zero <- setdiff(country_native_endemic, country_type) # countries with no type specimen

# transforming into a sparse matrix to speed up calculations
sparse_all <- 
  df_all_grid |> 
  phyloregion::long2sparse(grids = "grids", species = "species") |> 
  phyloregion::sparse2dense()

# Transforming in presence absence matrix
sparse_all_pa <- ifelse(sparse_all >= 1, 1, 0) 

# Binding countries with no types - adding zeroes
country_type_zero_names <- paste(country_type_zero, "_type", sep = "") # this will be used to bind together matrix with types and add the countries with no type
matrix_type_zero <- matrix(0,
         nrow = length(country_type_zero_names),
         ncol = ncol(sparse_all_pa), 
         dimnames = list(country_type_zero_names, 
                         colnames(sparse_all_pa)))

sparse_all_pa2 <- rbind(sparse_all_pa, matrix_type_zero)

 

sparse_all_pa3 <- 
  sparse_all_pa2[, colnames(sparse_all_pa2) %in% spp_endemic_native$species]

source(here::here("R", "functions", "function_beta_types_success_fail.R"))
# names_countries <- unique(df_native_endemic_grid$grids) # country names
names_countries <- 
  unique(gsub(pattern = "_type", 
              replacement = "",
              c(unique(df_native_endemic_grid$grids), df_type_grid$grids)))
df_endemic_beta <- beta_types(presab = sparse_all_pa3, 
                          names.countries = names_countries)


# calculating metrics

readr::write_csv(df_endemic_beta, here::here("data", "processed", "df_endemic_beta.csv"))