In nicholasjclark/BBS.occurrences: BBS community composition and co-occurrence analysis

knitr::opts_chunk$set(echo = TRUE)
library(knitr)
opts_chunk$set(tidy.opts=list(width.cutoff=50),tidy=TRUE)

This appendix describes how to extract the presence-absence and abundance datasets that can be accessed in the BBS.occurrences library using data('BBS.occurrences') and data('BBS.abundances'). It also demonstrates how to store latitude and longitude data needed for generating landcover and flyway variables for each site (accessed in the BBS.occurrences library using data('Site.descriptors')).

Specify the years to gather North American Breeding Bird Survey (BBS) data. Note, crashes can sometimes occur when using multiple years. We found it is best to run the below functions separately for each year and save the resulting object as an .Rdata file. You can then bring all saved files in as a list and compile them at once (see below for methods to filter and compile)

year <- 2003

Gather the full dataset of BBS observations from USA (840) for the specified years

# install.packages("pullBBS")
species.df <- pullBBS::pullBBS(year = year, country = 840, useCache = T)

Add species names in Birdtree.org format and eliminate records of unrecognised species

# install.packages("devtools")
devtools::install_github("nicholasjclark/BBS.occurrences")
library(BBS.occurrences)
species.df <- validateSpecies(species.df)

Save the dataset to a new folder

dir.create('./BBS.Data.cleaned')
save(species.df, file = "./BBS.Data.cleaned/USA.BBS.2003cleaned.Rdata")

Repeat for the next year, etc. Once all of the data is downloaded, we can read in all the separate yearly BBS datasets and filter them

all.files <- list.files(path="./BBS.Data.cleaned/", pattern="USA.BBS")

mylist<- lapply(all.files, function(x) {
  load(file = paste("./BBS.Data.cleaned/",x,sep=""))
  get(ls()[ls()!= "filename"])
})

names(mylist) <- all.files
species.df <- do.call("rbind", mylist)

\pagebreak

BBS runs are only valid if RunType == 1 and if Run protocol (RPID) == 101. In addition, only RouteTypeDetailID == 1 and RouteTypeDetailID == 2 are considered random. We filter to match these conditions, and add a ydays numeric variable observation for the day

# install.packages("dplyr")
# install.packages("lubridate")
species.df = species.df %>%

  dplyr::filter(RPID == 101 & RunType == 1 & RouteTypeID == 1 
                & RouteTypeDetailId %in% c(1,2)) %>%

  dplyr::mutate(ydays = lubridate::yday(lubridate::ymd(paste(year, Month, Day,
                                                             sep = "-")))) %>%

  dplyr::mutate(start.times = as.numeric(as.difftime(
      lubridate::hm(gsub("^(.)(.*)$", "\\1:\\2", StartTime)))) / 60 / 60)

We can now create site by species abundance and occurrence matrices

route.abundance = species.df %>%
  dplyr::mutate_at(dplyr::vars(dplyr::starts_with("Stop")), dplyr::funs(as.numeric)) %>%
  dplyr::mutate(Sp.abun = rowSums(.[grep("Stop", names(.))], na.rm = TRUE)) %>%
  dplyr::select(RouteDataID, BirdTree.species, Sp.abun)%>%
  dplyr::group_by(RouteDataID, BirdTree.species) %>%
  dplyr::summarise(total = sum(Sp.abun))%>%
  tidyr::spread(BirdTree.species, total, fill = 0)

row.names(route.abundance) <- route.abundance$RouteDataID
route.abundance$RouteDataID <- NULL

route.presence.absence <- route.abundance
route.presence.absence[route.presence.absence > 0] <- 1

Finally, we compress the pullBBS dataset to one row per site, using slice to remove duplicate sites with both faranheit & celsius records (keeping only the farenheit record)

sites.df = species.df %>%
  dplyr::select(matches("Route|Temp|Wind|Sky|Type"), year, 
                countrynum, statenum, RPID, Month,
                Day, ObsN, Stratum, Latitude, Longitude, ydays, 
                start.times, BCR) %>%
  distinct %>%
  group_by(RouteDataID) %>%
  slice(which.max(StartTemp))

Before saving, we order site data to match row orders of site x species matrices

sites.df <- sites.df[match(rownames(route.abundance), 
                           sites.df$RouteDataID),]

Save the matrices, site-level data and a matrix of lats and longs (for downloading MODIS landcover data). Note, these presence-absence and abundance datasets can be directly accessed in the BBS.occurrences library using data('BBS.occurrences') and data('BBS.abundances')

latlon <- cbind(sites.df$Longitude, sites.df$Latitude)
save(
  route.presence.absence = route.presence.absence,
  route.abundance = route.abundance,
  latlon = latlon,
  sites.df = sites.df,
  file = "./BBS.Data.cleaned/BBS.2003-2009.filtered.Rdata"
)

nicholasjclark/BBS.occurrences documentation built on July 19, 2020, 8:31 p.m.