R/GRSex.R

In GapAnalysis: Conservation Indicators using Spatial Information

#' @title Geographical representativeness score ex situ
#' @name GRSex
#' @description The GRSex process provides a geographic measurement of the proportion of a species’ range
#'  that can be considered to be conserved in ex situ repositories. The GRSex uses buffers (default 50 km radius)
#'  created around each G coordinate point to estimate geographic areas already well collected within the distribution
#'  models of each taxon, and then calculates the proportion of the distribution model covered by these buffers.
#' @param Occurrence_data A data frame object with the species name, geographical coordinates,
#'  and type of records (G or H) for a given species
#' @param Species_list A vector of characters with the species names to calculate the GRSex metrics.
#' @param Raster_list A list of rasters representing the species distribution models for the species list provided
#'  in \var{Species_list}. The order of rasters in this list must match the same order as \var{Species_list}.
#' @param Buffer_distance Geographical distance used to create circular buffers around germplasm.
#'  Default: 50000 (50 km) around germplasm accessions (CA50)
#' @param Gap_Map logical, if \code{TRUE} the function will calculate gap maps for each species analyzed and
#'  will return a list with two slots GRSex and gap_maps. If any value is provided, the function will assume that
#'  Gap_Map = TRUE
#' @return This function returns a data frame with two columns:
#'
#' \tabular{lcc}{
#' species \tab Species name \cr
#' GRSex \tab GRSex value calculated\cr
#' }
#'
#' @examples
#' ##Obtaining occurrences from example
#' data(CucurbitaData)
#' Cucurbita_splist <- unique(CucurbitaData$species)
#' ## Obtaining rasterList object. ##
#' data(CucurbitaRasters)
#' CucurbitaRasters <- raster::unstack(CucurbitaRasters)
#' #Running GRSex
#' GRSex_df <- GRSex(Species_list = Cucurbita_splist,
#'                     Occurrence_data = CucurbitaData,
#'                     Raster_list = CucurbitaRasters,
#'                     Buffer_distance = 50000,
#'                     Gap_Map = TRUE)
#'
#' @references
#' Ramirez-Villegas et al. (2010) PLOS ONE, 5(10), e13497. doi: 10.1371/journal.pone.0013497
#' Khoury et al. (2019) Ecological Indicators 98:420-429. doi: 10.1016/j.ecolind.2018.11.016
#'
#' @export
#' @importFrom sp coordinates proj4string SpatialPoints over CRS
#' @importFrom stats median
#' @importFrom fasterize fasterize
#' @importFrom raster overlay crop raster extent ncell projection



GRSex <- function(Species_list, Occurrence_data, Raster_list, Buffer_distance=50000, Gap_Map=FALSE) {

  longitude <- NULL
  taxon <- NULL
  type <- NULL
  latitude <-NULL

  #Checking Occurrence_data format
  par_names <- c("species","latitude","longitude","type")

  if(missing(Occurrence_data)){
    stop("Please add a valid data frame with columns: species, latitude, longitude, type")
  }

  if(isFALSE(identical(names(Occurrence_data),par_names))){
    stop("Please format the column names in your dataframe as species, latitude, longitude, type")
  }

  #Checking if Gap_Map option is a boolean or if the parameter is missing left Gap_Map as FALSE
  if(is.null(Gap_Map) | missing(Gap_Map)){ Gap_Map <- FALSE
  } else if(isTRUE(Gap_Map) | isFALSE(Gap_Map)){
    Gap_Map <- Gap_Map
  } else {
    stop("Choose a valid option for GapMap (TRUE or FALSE)")
  }

  #Checking if user is using a raster list or a raster stack
  if (isTRUE("RasterStack" %in% class(Raster_list))) {
    Raster_list <- raster::unstack(Raster_list)
  } else {
    Raster_list <- Raster_list
  }


  # create a dataframe to hold the components
  df <- data.frame(matrix(ncol = 2, nrow = length(Species_list)))
  colnames(df) <- c("species", "GRSex")

  if(isTRUE(Gap_Map)){
    GapMapEx_list <- list()
  }

  for(i in seq_len(length(sort(Species_list)))){

    # select species G occurrences
    OccData  <- Occurrence_data[which(Occurrence_data$species==Species_list[i]),]
    OccData  <- OccData [which(OccData$type == "G" & !is.na(OccData$latitude) & !is.na(OccData$longitude)),]
    OccData  <- OccData [,c("longitude","latitude")]


    # select raster with species name
    for(j in seq_len(length(Raster_list))){
      if(grepl(j, i, ignore.case = TRUE)){
        sdm <- Raster_list[[j]]
      }

      d1 <- Occurrence_data[Occurrence_data$species == Species_list[i],]
      test <- GapAnalysis::ParamTest(d1, sdm)
      if(isTRUE(test[1])){
        stop(paste0("No Occurrence data exists, but and SDM was provide. Please check your occurrence data input for ", Species_list[i]))
      }

    };rm(j)

    if(isFALSE(test[2])){
      df$species[i] <- as.character(Species_list[i])
      df$GRSex[i] <- 0
      warning(paste0("Either no occurrence data or SDM was found for species ", as.character(Species_list[i]),
                     " the conservation metric was automatically assigned 0"))
    } else {

      #
      # sp::coordinates(OccData ) <- ~longitude+latitude

      #Checking raster projection and assuming it for the occurrences dataframe shapefile
      if(is.na(raster::crs(sdm))){
        warning("No coordinate system was provided, assuming  +proj=longlat +datum=WGS84 +no_defs +ellps=WGS84 +towgs84=0,0,0","\n")
        raster::projection(sdm) <- "+proj=longlat +datum=WGS84 +no_defs +ellps=WGS84 +towgs84=0,0,0"
      }
      # suppressWarnings(sp::proj4string(OccData) <- sp::CRS(raster::projection(sdm)))

      # select raster with species name

      # convert SDM from binary to 1-NA for mask and area
      sdmMask <- sdm
      sdmMask[sdmMask[] != 1] <- NA #USING THIS TO AVOID PROBLEMS WITH NA FLOATING VALUES AS -9999 OR 3E8-178
      # buffer G points
      buffer <- GapAnalysis::Gbuffer(xy = OccData ,
                                     dist_m = Buffer_distance,
                                     output = 'sf')
      # rasterizing and making it into a mask
      buffer_rs <- fasterize::fasterize(buffer, sdm)
      buffer_rs[!is.na(buffer_rs[])] <- 1
      buffer_rs <- buffer_rs * sdmMask
      # calculate area of buffer
      cell_size<-raster::area(buffer_rs, na.rm=TRUE, weights=FALSE)
      cell_size<-cell_size[!is.na(cell_size)]
      gBufferRas_area<-length(cell_size)*median(cell_size)

      # calculate area of the threshold model
      cell_size<- raster::area(sdmMask, na.rm=TRUE, weights=FALSE)
      cell_size<- cell_size[!is.na(cell_size)]
      pa_spp_area <- length(cell_size)*median(cell_size)
      # calculate GRSex
      GRSex <- min(c(100, gBufferRas_area/pa_spp_area*100))

      df$species[i] <- as.character(Species_list[i])
      df$GRSex[i] <- GRSex

      #GRSex gap map

      if(isTRUE(Gap_Map)){
        message(paste0("Calculating GRSex gap map for ",as.character(Species_list[i])),"\n")
        bf2 <- buffer_rs
        bf2[is.na(bf2),] <- 0
        gap_map <- sdmMask - bf2
        gap_map[gap_map[] != 1] <- NA
        GapMapEx_list[[i]] <- gap_map
        names(GapMapEx_list[[i]] ) <- Species_list[[i]]
      }
    }
  }
    if(isTRUE(Gap_Map)){
      df <- list(GRSex= df,gap_maps=GapMapEx_list)
    } else {
      df <- df
    }

  return(df)
}