Nothing
R/movegroup.R
In movegroup: Visualizing and Quantifying Space Use Data for Groups of Animals
Defines functions
movegroup
Documented in
movegroup
#' Automates dynamic Brownian bridge movement model construction across individuals
#'
#' Automates dynamic Brownian bridge movement model calculation for utilization distribution (UD)
#' estimation for multiple individuals simultaneously, using functions in the 'move' package. The
#' authors are indebted to the move package authors Bart Kraunstauber, Marco Smolla, and Anne K
#' Scharf, and to Sarah Becker for seed code which inspired the development of the
#' movegroup::movegroup function.
#'
#' Step 1. Filter individuals.
#' Remove those individuals for which there are insufficient data i.e. number of re-locations is
#' smaller than the window size parameter value (default = 31).
#' Step 2. Generate universal raster.
#' Based on all remaining data, a universal raster is generated where the calculated UDs are plotted
#' into.
#'
#' Step 3. Loop through individuals.
#' Individuals are looped through to construct individual-level movement models (on an absolute
#' scale).
#'
#' See www.GitHub.com/SimonDedman/movegroup for issues, feedback, and development
#' suggestions.
#'
#' install_git('https://gitlab.com/bartk/move.git') #Installs 'move' development version
#' @import utils
#' @importFrom lubridate is.POSIXt
#' @importFrom sp CRS SpatialPoints spTransform proj4string
#' @importFrom beepr beep
#' @importFrom dplyr mutate rename group_by summarise filter semi_join distinct ungroup arrange across bind_cols pull
#' @importFrom tidyr drop_na
#' @importFrom raster raster projectExtent res ncell setValues calc values writeRaster
#' @importFrom move move timeLag burst brownian.bridge.dyn brownian.motion.variance.dyn getVolumeUD
#' @importFrom rlang .data
#' @importFrom grDevices graphics.off
#' @importFrom graphics par
#' @importFrom methods new
#' @importFrom stats setNames
#'
#' @export movegroup
#'
#' @param data Data frame object containing the data. Requires columns Lat Lon DateTime ID and
#' potentially a grouping column (not currently implemented, email to request). Column names
#' specified in later parameters.
#' @param ID Name of animal tag ID column in data. "Character".
#' @param Datetime Column name in data that contains date/time stamps for each recorded detection.
#' Must be in POSIXct format. "Character".
#' @param Lat Name of latitude column in data. "Character".
#' @param Lon Name of longitude column in data. "Character".
# #' @param Group Name of grouping column in data. Unused 2023-08-14.
#' @param dat.TZ Timezone of data for as.POSIXct. Default "US/Eastern".
#' @param proj CRS for move function. Default sp::CRS("+proj=longlat +datum=WGS84").
#' @param sensor Sensor for move function. Single character or vector with length of the number of
#' coordinates. Optional. Default "VR2W".
#' @param moveLocError Location error (m) in the 'brownian.bridge.dyn' function in the 'move'
#' package. Numeric. Either single or a vector of length nrow data. If using passive acoustic data
#' this is the detection range of the receiver(s). Default 1. See MoveLocErrorCalc function for
#' satellite data with state space modelled locations with 95% confidence intervals for latlon i.e.
#' lat and lon025 and 975.
#' @param timeDiffLong Single numeric value. Threshold value in timeDiffUnits designating the length
#' of long breaks in re-locations. Used for bursting a movement track into segments, thereby
#' removing long breaks from the movement track. See ?move::bursted for details. Default 2 hours is
#' arbitrary, looping through 18, 24, and 36 hours for satellite data on great hammerhead sharks
#' revealed volume areas for core and general use gradually rise with timeDiffLong increases,
#' multiple small dots of presence get blobbed together, and therefore sometimes this covers land.
#' Ideally one would not discard any data, in which case one should choose a value higher than the
#' largest between-detections gap in their dataset (or just pick a very large number). This
#' parameter is useful when the model would otherwise get stuck trying to calculate a UD for an
#' individual with a very large home range that is inadequately captured by a receiver array.
#' Default 2.
#' @param timeDiffUnits Character. Unit for timeDiffLong. Default "hours".
#' @param center US English alternate to centre. Do you want to center the move object within
#' extent? See spTransform. Default TRUE.
#' @param centre British English alternate to center. Do you want to centre the move object within
#' extent? See spTransform. Default NULL.
#' @param buffpct Buffer extent for raster creation, proportion of 1. Default 0.3, can try e.g. 3
#' for a large buffer to avoid clipping, at the cost of file size, but later cropping in
#' plotraster.R will remove extraneous blank space.
#' @param rasterExtent Extent of raster created around data. If NULL (default), calculated from
#' data, buffpct, rasterResolution. Else length 4 vector, c(xmn, xmx, ymn, ymx) decimal latlon
#' degrees. Don't go to 90 (degrees) north or south for ymax or ymin. Doesn't prevent constraint to
#' data limits (in plot anyway), but prevents raster clipping crash.
#' @param rasterCRS CRS for raster creation. Default sp::CRS("+proj=utm +zone=17 +datum=WGS84").
#' @param rasterResolution Single numeric value to set raster resolution - cell size (width and
#' height) in metres. 111000: 1 degree lat = 111km. Trade-off between small res = big file & processing
#' time. Should be a function of the spatial resolution of your receivers or positioning tags.
#' Higher resolution will lead to more precision in the volume areas calculations. Try using
#' 2*dbblocationerror, if dbblocationerror is a single value. Default 50 = 50m = 50m² = 0.00005 km²
#' (divide by 1000000). Try around the median of your moveLocError.
# #' @param dbblocationerror Location.error param in 'brownian.bridge.dyn' function in the 'move'
# #' package. single numeric value or vector of the length of coordinates that describes the error of
# #' the location (sender/receiver) system in map units. Or a character string with the name of the
# #' column containing the location error can be provided. Default is moveLocError. See
# #' MoveLocErrorCalc function for satellite data with state space modelled locations with 95%
# #' confidence intervals for latlon i.e. lat and lon025 and 975.
#' @param movemargin Margin size for variance calc in move::brownian.motion.variance.dyn and
#' behavioural change point analysis in move::brownian.bridge.dyn. Must be an odd number. Default 11.
#' @param dbbext Ext param in the 'brownian.bridge.dyn' function in the 'move' package. Extends
#' bounding box around track. Numeric single (all edges), double (x & y), or 4 (xmin xmax ymin ymax)
#' . Default 3. Excessive buffering will get cropped automatically.
#' @param dbbwindowsize The window.size param in the 'brownian.bridge.dyn' function
#' in the 'move' package. The size of the moving window along the track. Larger windows provide more
#' stable/accurate estimates of the brownian motion variance but are less well able to capture more
#' frequent changes in behaviour. Number must be odd. Code will not run if total detections of
#' individual < window size (default 23), which must be >= 2*movemargin (default 11).
#' @param writeRasterFormat Character. Output file type for raster::writeRaster param format.
#' Default "ascii". TO DEPRECIATE.
#' @param writeRasterExtension Character. Output file extension for raster::writeRaster param
#' extension. Default ".asc". TO DEPRECIATE.
#' @param writeRasterDatatype Character. Data type for writing values to disk for
#' raster::writeRaster param Datatype. Default "FLT4S". TO DEPRECIATE.
#' @param absVolumeAreaSaveName File name plus extension where UD estimates are saved. Default
#' "VolumeArea_AbsoluteScale.csv".
#' @param savedir Save outputs to a temporary directory (default) else change
#' to desired directory e.g. "/home/me/folder". Do not use getwd() for this.
#' Do NOT include terminal slash. Directory must exist. Default tempdir().
#' @param alerts Audio warning for failures. Default TRUE.
#'
#' @return Individual-level utilization distributions, saved as rasters, as well
#' as calculated volume area estimates for 50 and 95pct contours, saved in a
#' .csv file. Motion variance csvs per individual ("MotionVariance.csv"), see
#' move::brownian.motion.variance.dyn. No processed object is returned, i.e. bad: "objectname <-
#' movegroup()", good: "movegroup()".
#' @details
#' When used together, the order of functions would be: movegroup, scaleraster, alignraster if
#' required, plotraster.
#'
#' ## Errors and their origins:
#'
#' 1. Error in .local(object, raster, location.error = location.error, ext = ext: Higher y grid not
#' large enough, consider extending the raster in that direction or enlarging the ext argument.
#' Increase buffpct, e.g. to 3.
#'
#' 2. Error in .data\[\[dttm\]\]: Must subset the data pronoun with a string, not a <POSIXct/POSIXt>
#' object. Use "ColName" not dataframe$ColName syntax for Datetime, ID, Lat, Lon.
#'
#' 3. Error in splice(dot_call(capture_dots, frame_env = frame_env, named = named,: object
#' 'DateTime' not found. Use "ColName" not ColName syntax for Datetime, ID, Lat, Lon.
#'
#' 4. Error in .local(object, raster, location.error = location.error, ext = ext: Higher x grid not
#' large enough, consider extending the raster in that direction or enlarging the ext argument. Try
#' "buffpct = 1," , then larger e.g. 3, if still getting the error.
#'
#' 5. cannot allocate vector of size (BIG) Gb: Increase rasterResolution value.
#'
#' 6. In min/max: No non-missing arguments to min; returning Inf: likely not enough memory, increase
#' rasterResolution value.
#'
#' 7. Error in tmp\[\[i\]\]: subscript out of bounds. dbbmmwindowsize may be too large relative to nrow
#' of that individual. Try lowering movemargin (default 11, has to be odd) and then lowering
#' dbbmmwindowsize (default 23, has to be >=2*movemargin, has to be odd).
#'
#' @examples
#' \donttest{
#' # load data
#' data("TracksCleaned")
#' # run function
#' movegroup(
#' data = TracksCleaned,
#' ID = "Shark",
#' Datetime = "Datetime",
#' Lat = "Lat",
#' Lon = "Lon",
#' savedir = tempdir())
#' }
#'
#' @author Simon Dedman, \email{simondedman@@gmail.com}
#' @author Maurits van Zinnicq Bergmann, \email{mauritsvzb@@gmail.com}
#'
#' @references Kranstauber, B., Kays, R., LaPoint, S. D., Wikelski, M. and Safi, K. (2012), A
#' dynamic Brownian bridge movement model to estimate utilization distributions for heterogeneous
#' animal movement. Journal of Animal Ecology. doi: 10.1111/j.1365-2656.2012.01955.x
#'
#' Kranstauber, B., M. Smolla & A. K. Scharf. 2019. Move: visualizing and analyzing animal track
#' data. R package version 4.2.4 (at 2023-08-15). https://CRAN.R-project.org/package=move.
movegroup <- function(
data = NULL, # Data frame object containing the data. Requires columns Lat Lon DateTime ID and optionally a grouping column.
ID = NULL, # Name of animal tag ID column in data.
Datetime = NULL, # Column name in data that contains date/time stamps for each recorded detection. Must be in POSIXct format.
Lat = NULL, # name of Lat & Lon columns in data.
Lon = NULL,
# Group = NULL, # name of grouping column in data. CURRENTLY UNUSED; MAKE USER DO THIS?
dat.TZ = "US/Eastern", # timezone for as.POSIXct.
proj = sp::CRS("+proj=longlat +datum=WGS84"), # CRS for move function.
sensor = "VR2W", # sensor for move function. Single character or vector with length of the number of coordinates. Optional.
moveLocError = 1, # location error in metres for move function. Numeric. Either single or a vector of length nrow data.
timeDiffLong = 2, # threshold length of time in timeDiffUnits designating long breaks in relocations.
timeDiffUnits = "hours", # units for time difference for move function.
center = TRUE, # center move object within extent? See spTransform.
centre = NULL, # British English alternate to center.
buffpct = 0.3, # buffer extent for raster creation, proportion of 1.
rasterExtent = NULL, # if NULL, raster extent calculated from data, buffpct, rasterResolution. Else length 4 vector, c(xmn, xmx, ymn, ymx) decimal latlon degrees. Don't go to 90 for ymax
# Doesn't prevent constraint to data limits (in plot anyway), but prevents raster clipping crash
rasterCRS = sp::CRS("+proj=utm +zone=17 +datum=WGS84"), # CRS for raster creation. This is around Bimini, Bahamas.
rasterResolution = 50, # numeric vector of length 1 or 2 to set raster resolution - cell size in metres. 111000: 1 degree lat = 111km
# dbblocationerror = moveLocError, # location.error param in brownian.bridge.dyn. Could use the same as moveLocError?
movemargin = 11, # Margin size for variance calc in move::brownian.motion.variance.dyn and behavioral change point analysis in
# move::brownian.bridge.dyn. Must be an odd number. Default 11.
dbbext = 3, # ext param in brownian.bridge.dyn. Extends bounding box around track. Numeric single (all edges), double (x & y), or 4 (xmin xmax ymin ymax). Default 3.
dbbwindowsize = 23, # window.size param in brownian.bridge.dyn. The size of the moving window along the track. Larger windows provide more
# stable/accurate estimates of the brownian motion variance but are less well able to capture more frequent changes in behavior.
# This number has to be odd. A dBBMM is not run if total detections of individual < window size (default 23). Must be >=2*margin (movemargin) which is 11 so >=22,
# but odd so >=23.
writeRasterFormat = "ascii",
writeRasterExtension = ".asc",
writeRasterDatatype = "FLT4S",
absVolumeAreaSaveName = "VolumeArea_AbsoluteScale.csv",
savedir = tempdir(), # save outputs to a temporary directory (default) else change to current
# directory e.g. "/home/me/folder". Do not use getwd() here.
alerts = TRUE # audio warning for failures
) {
# If savedir has a terminal slash, remove it, it's added later
if (substr(x = savedir, start = nchar(savedir), stop = nchar(savedir)) == "/") savedir = substr(x = savedir, start = 1, stop = nchar(savedir) - 1)
# if savedir doesn't exist, can't save or setwd into it, therefore create it
if (!file.exists(savedir)) dir.create(savedir)
oldpar <- par(no.readonly = TRUE) # defensive block, thanks to Gregor Sayer
oldwd <- getwd()
oldoptions <- options()
on.exit(par(oldpar))
on.exit(setwd(oldwd), add = TRUE)
on.exit(options(oldoptions), add = TRUE)
setwd(savedir)
if (alerts) options(error = function() {
beepr::beep(9)# give warning noise if it fails
graphics.off()# kill all graphics devices
setwd(oldwd) # reinstate original working directory. Probably redundant given on.exit
} # close options subcurly
) # close options
# edit center if centre used
if (!is.null(centre)) center <- centre
# Create writeRasterExtension from writeRasterFormat
if (is.null(writeRasterExtension)) writeRasterExtension <- switch(EXPR = writeRasterFormat,
"ascii" = ".asc",
"raster" = ".grd",
"SAGA" = ".sdat",
"IDRISI" = ".rst",
"CDF" = ".nc",
"GTiff" = ".tif",
"ENVI" = ".envi",
"EHdr" = ".bil",
"HFA" = ".img")
data <- dplyr::rename(.data = data, # Rename user entry to "ID", Ditto Datetime Lat & Lon
Datetime = .data[[Datetime]],
ID = .data[[ID]],
Lat = .data[[Lat]],
Lon = .data[[Lon]]) |>
# prefixes "X" to numerical-named sharks to avoid issues later
dplyr::mutate(ID = make.names(ID))
if (!any(class(data$Datetime) == "POSIXct")) stop("Datetime column must be POSIXct")
# Add movelocationerror and dbblocationerror as columns if they're vectors, else their length
# desyncs from nrow(data) if any IDs < windowsize
if (nrow(data) == length(moveLocError)) data$moveLocError <- moveLocError
# if (nrow(data) == length(dbblocationerror)) data$dbblocationerror <- dbblocationerror
# Convert coordinate sets to SpatialPoints and project
cord.dec <- sp::SpatialPoints(cbind(data$Lon, data$Lat),
proj4string = sp::CRS("+proj=longlat")
)
# Transform to UTM by setting the EPSG to local UTM zone for the data.
cord.UTM <- as.data.frame(sp::spTransform(cord.dec, rasterCRS))
colnames(cord.UTM) <- c("NewEastingUTM", "NewNorthingUTM")
data <- cbind(data, cord.UTM) # 1308 x 7
# Construct movement models per individual.
# Some notes regarding the 'move package and construction of movement models: Several arguments
# need to be used to run the model. 1. Window size: corresponds with number of locations and moves
# along a given trajectory to estimate the MA parameter within defined subsections of the path.
# This increases the ability to detect breakpoints where changes in behaviour occur. The window
# size should relate to what kind of behaviours the model is desired to identify e.g., a window
# size of 23 means the sliding window is moved every 23 locations or every 23 hours (has to do
# with sampling interval) 2. Margin: motion variance based on only the middle section of the
# trajectory; the ends of the movement trajectory where no changes are allowed because at some
# stage you want to have a few locations to base your estimation of the variance on and how many
# locations in either side of the window we use for this, is called the margin. Smaller values for
# window size and margin is expected to give a higher frequency of behavioural changes; make these
# large for looking at migrations. 3. The raster dictates the grid cell size for the UD to be
# calculated per grid cell per individual. Create the raster of certain size that matches with
# coordinates used to make the move object 4. Extent: is incorporated if there are animal
# locations that border the edges of the raster.
# A dBBMM is not run if total detections of individual < window size (default value, 31).
# Below code checks and filters individuals with insufficient data.
# Below we set window size arbitrarily to 23
check1 <- data |>
dplyr::group_by(.data$ID) |>
dplyr::summarise(relocations = length(.data$Datetime))
check2 <- dplyr::filter(check1, .data$relocations >= dbbwindowsize) # filter: removed 2 rows (14%), 12 rows remaining
if (length(check1$ID) != length(check2$ID)) {
data <- dplyr::semi_join(data, check2) # Joining, by = "ID". semi_join: added no columns
check1 <- data |>
dplyr::group_by(.data$ID) |>
dplyr::summarise(relocations = length(.data$Datetime))
check2 <- dplyr::filter(check1, .data$relocations >= dbbwindowsize) # filter: no rows removed
length(check1$ID) == length(check2$ID)
}
# Create per-individual move object, project it, construct a dBBMM, calculate the volume area
# within the 50% and 95% contours, save as ASCII.
bb <- list()
bb.list <- list()
data <- tidyr::drop_na(data = data, ID) |>
dplyr::group_by(ID) |>
dplyr::distinct(Datetime, .keep_all = TRUE) |> # distinct (grouped): removed one row (<1%), 1,286 rows remaining
# prevents duplicate Datetime crash in move() later
dplyr::ungroup() # 1253 x 7 after removing as.numeric above
# Make a raster for the UD to plot into. Start with UTM.
# These coordinates need to be big enough to cover your data.
# May need to expand x and y ranges (i.e. the "e" variable below) if you encounter errors when constructing the DBBMM in the next code chunk.
# NOTE: "xUTM" grid should be larger than the "x" i.e. lower minima and higher maxima.
# The variable 'e' influences the extent [now done with buffpct]
# The resolution determines the grid size.
# Finer resolution/grid size (in m) i.e. lower value means longer computation time. May need/want to play with this value too.
# e <- 80 * 1000 # make e a function of range of extent as %
# Error in .local(object, raster, location.error = location.error, ext = ext, :
# Lower x grid not large enough, consider extending the raster in that direction or enlarging the ext argument
# make buffpct larger
if (!is.null(rasterExtent)) { # if xUTM object exists
xUTM <- sp::SpatialPoints(cbind(c(rasterExtent[1], rasterExtent[2]), c(rasterExtent[3], rasterExtent[4])), proj4string = sp::CRS("+proj=longlat"))
xUTM <- as.data.frame(sp::spTransform(xUTM, rasterCRS)) # Transform to UTM
xUTM <- raster::raster( # create a raster
xmn = xUTM[1,1] - ((xUTM[2,1] - xUTM[1,1]) * buffpct), # with xUTM's values as extents
xmx = xUTM[2,1] + ((xUTM[2,1] - xUTM[1,1]) * buffpct),
ymn = xUTM[1,2] - ((xUTM[2,2] - xUTM[1,2]) * buffpct),
ymx = xUTM[2,2] + ((xUTM[2,2] - xUTM[1,2]) * buffpct),
crs = rasterCRS, # +proj=utm +zone=27 +datum=WGS84 +units=m +no_defs
resolution = rasterResolution # 50
) # close raster
} else { # exists(xUTM), else create from data
xUTM <- raster::raster(
xmn = min(data$NewEastingUTM, na.rm = TRUE) - ((max(data$NewEastingUTM, na.rm = TRUE) - min(data$NewEastingUTM, na.rm = TRUE)) * buffpct),
xmx = max(data$NewEastingUTM, na.rm = TRUE) + ((max(data$NewEastingUTM, na.rm = TRUE) - min(data$NewEastingUTM, na.rm = TRUE)) * buffpct),
ymn = min(data$NewNorthingUTM, na.rm = TRUE) - ((max(data$NewNorthingUTM, na.rm = TRUE) - min(data$NewNorthingUTM, na.rm = TRUE)) * buffpct),
ymx = max(data$NewNorthingUTM, na.rm = TRUE) + ((max(data$NewNorthingUTM, na.rm = TRUE) - min(data$NewNorthingUTM, na.rm = TRUE)) * buffpct),
crs = rasterCRS, # +proj=utm +zone=27 +datum=WGS84 +units=m +no_defs
resolution = rasterResolution # 50
) # close raster
} # close else
# run move on all data together to generate global CRS for raster
alldata <- data |>
dplyr::arrange(Datetime) |>
dplyr::group_by(Datetime) |> # remove duplicates
dplyr::summarise(dplyr::across(where(~ is.numeric(.)), mean, na.rm = TRUE),
dplyr::across(where(~ is.character(.) | lubridate::is.POSIXt(.)), dplyr::first)) |> # library(lubridate)
dplyr::mutate(dplyr::across(where(~ is.numeric(.)), ~ ifelse(is.nan(.), NA, .)), #convert NaN to NA. POSIX needs lubridate
dplyr::across(where(~ is.character(.)), ~ ifelse(is.nan(.), NA, .))) |> # https://community.rstudio.com/t/why-does-tidyrs-fill-work-with-nas-but-not-nans/25506/5
dplyr::ungroup()
alldata <- as.data.frame(alldata)
moveall <- move::move(
x = alldata$Lon, # numeric vector with x coordinates if non-Movebank data are provided (e.g. data$x).
y = alldata$Lat, # numeric vector with y coordinates if non-Movebank data are provided (e.g. data$y).
time = alldata$Datetime, # vector of time stamps with POSIXct conversion if non-Movebank data are provided,
# i.e. as.POSIXct(data$timestamp, format="%Y-%m-%d %H:%M:%S", tz="UTC")
data = alldata, # extra data associated with the relocations
proj = proj, # projection method for non-Movebank data; requires a valid CRS (see CRS-class) object, e.g. CRS("+proj=longlat +ellps=WGS84")
# animal = data$ID,
sensor = sensor # Sensor name(s), either single character or a vector with length of the number of coordinates.
)
# Convert projection to Azimuthal Equi-Distance projection (aeqd)
rall <- sp::spTransform(moveall, center = center)
sp::proj4string(rall) # "+proj=aeqd +lat_0=44.99489997 +lon_0=-17.48575004 +x_0=0 +y_0=0 +datum=WGS84 +units=m +no_defs"
rallCRS <- sp::CRS(sp::proj4string(rall))
class(rallCRS) # CRS
write.csv(sp::proj4string(rall), file.path(savedir, "CRS.csv"), row.names = FALSE)
saveRDS(rallCRS, file = file.path(savedir, "CRS.Rds"))
rm(moveall)
rm(alldata)
# Reproject to AEQD. Make a dummy object to get the correct projection.
x.i <- raster::raster(
xmn = min(data$NewEastingUTM),
xmx = max(data$NewEastingUTM),
ymn = min(data$NewNorthingUTM),
ymx = max(data$NewNorthingUTM),
crs = sp::proj4string(rall) # if rasterCRS then same as above! trying proj4string(r.i) which is AEQD from spTransform
# could add here: resolution = rasterResolution # 50 ####
)
# Reproject UTM to AEQD
# ?projectExtent # project the values of a Raster object to a new Raster object with another projection (CRS), i.e. projectExtent(from, to, ...))
newTemplate <- raster::projectExtent(xUTM, sp::proj4string(x.i))
# change newTemplate to xAEQD. Change xAEQD res so x & y match. Kills values.
raster::res(newTemplate) <- rep(rasterResolution, 2)
# Give newTemplate some values. Make Rep equal to the ncell dimension
ones <- rep(1, raster::ncell(newTemplate))
xAEQD <- raster::setValues(newTemplate, ones)
rm(newTemplate)
rm(ones)
# get resolution from raster in rasterlist, assign it object, squared
rasterres <- (raster::res(xAEQD)[1]) ^ 2
write.csv(x = data.frame(rasterres = rasterres,
rasterResolution = rasterResolution),
file = file.path(savedir, "Resolutions.csv"))
# Loop through all unique tags
counter <- 0
for (i in unique(data$ID)) { # i <- unique(data$ID)[1]
# Print individual
counter <- counter + 1
message(
"processing ", which(unique(data$ID) %in% i),
" of ", length(unique(data$ID))
)
# Filter individual data
data.i <- as.data.frame(data[data$ID == i, ])
# create move object
move.i <- move::move(
x = data.i$Lon,
y = data.i$Lat,
time = data.i$Datetime,
proj = proj,
data = data.i,
animal = data.i$ID,
sensor = sensor
)
# Check the current projection
sp::proj4string(move.i) # "+proj=longlat +datum=WGS84 +no_defs"
# Incorporate uncertainty in the model by including a location error.
if ("moveLocError" %in% colnames(data.i)) {
move.i$LocationError <- data.i$moveLocError
} else {
if (exists("moveLocError")) {
if (length(moveLocError) == 1) {
move.i$LocationError <- moveLocError # single value, replicated down for each relocation
} else { # else if multiple values
if (length(moveLocError) == nrow(data.i)) { # should be same length as full dataset
move.i$LocationError <- data.i |> # take the full dataset,
dplyr::bind_cols(moveLocError = moveLocError) |> # cbind the full movelocerror
dplyr::filter(ID == i) |> # filter for just this ID
dplyr::pull(moveLocError) # and pull just the movelocerror for this ID
} else { # if not length 1 and not length of nrow(data)
stop(message("moveLocError must be either length 1 or length(nrow(data))")) # if not stop and tell user
} # close not length1 not same length as full dataset else
} # close length1 or else
} # close if exists movelocerror
}
# Convert projection to Azimuthal Equi-Distance projection (aeqd)
r.i <- sp::spTransform(move.i, CRSobj = rallCRS, center = center)
# Calculate time lag between consecutive detections.
# Obviously there is no time lag for the first detection, so we need to tell R this.
# Knowing time lags in acoustic telemetry is important, as the motion variance is based on the
# time difference b/w detections at consecutive locations i.e. larger time lag creates a wider
# bridge where the animal could've been and therefore potentially inflates the motion variance.
# Below code deals with this issue by identifying the location and number of detections with
# large time gaps. Below we will use an arbitrary value of 2 h.
TimeDiff <- move::timeLag(r.i, units = timeDiffUnits)
r.i$TimeDiff <- append(0, TimeDiff)
long <- which(r.i$TimeDiff > timeDiffLong)
# if no timediffs are longer than timedifflong, long is integer(0), a potentially dangerous object.
# Below we exclude the data points that are 'long' i.e. create a large time gap. move::burst ?
bursted <- move::burst(
r.i,
c("normal", "long")[1 + (move::timeLag(r.i, units = timeDiffUnits) > timeDiffLong)]
)
rm(r.i)
# There are 2 types of burst: "normal" and "long". You can select for these in the dbbmm by selecting the factor level in the burstType command.
if ("moveLocError" %in% colnames(data.i)) {
moveLocError.i <- data.i$moveLocError
} else {
if (exists("moveLocError")) {
if (length(moveLocError) == 1) {
moveLocError.i <- moveLocError # single value, replicated down for each relocation
# Robs: 1 m: gps loc of shark inferred from boat coord + bearing + distance estimate
} else { # else if multiple values
if (length(moveLocError) == nrow(data)) { # should be same length as full dataset
moveLocError.i <- data |> # take the full dataset,
dplyr::bind_cols(moveLocError = moveLocError) |> # cbind the full movelocerror
dplyr::filter(ID == i) |> # filter for just this ID
dplyr::pull(moveLocError) # and pull just the movelocerror for this ID
} else { # if not length 1 and not length of nrow(data)
stop(message("moveLocError must be either length 1 or length(nrow(data))")) # if not stop and tell user
} # close not length1 not same length as full dataset else
} # close length1 or else
} # close if exists moveLocError
} # close ifelse
# location error needs to be a positive number. Replace zeroes with 0.00001
moveLocError.i[which(moveLocError.i == 0)] <- 0.00001
# Construct the model. The time.step should reflect the ping frequency of the tag (in minutes)
bursted_dbbmm <- move::brownian.bridge.dyn(bursted,
burstType = "normal",
raster = xAEQD, # has to be AEQD for metre-based calculations, presuambly?
# Error:The projection of the raster and the Move object are not equal.
# Need bursted, r.i, to be the same projection as xAEQD
location.error = moveLocError.i,
ext = dbbext, # dbbext
margin = movemargin,
window.size = dbbwindowsize # must be >=2*margin which is 11 so >=22, but odd so >=23
)
# Build dBMvarianceBurst object with brownian.motion.variance.dyn
bursted_motionvar <- move::brownian.motion.variance.dyn(bursted,
location.error = moveLocError.i,
window.size = dbbwindowsize, # must be >=2*margin which is 11 so >=22, but odd so >=23
margin = movemargin
)
mtnvar <- as.data.frame(bursted_motionvar) # extract the data
mtnvar$motionVariance <- move::getMotionVariance(bursted_motionvar) # add motionvariance
write.csv(mtnvar,
file = file.path(savedir, paste0(i, "_MotionVariance.csv")),
row.names = FALSE)
rm(bursted)
rm(mtnvar)
# Re-standardize (Dr. Kranstauber's trouble shooting solution).
# Occurring errors are "due to limits of accuracy during calculations.
# Given the error is really small (<.000001 %) I would not worry to much and re-standardize".
# Then aggregate UD segments.
tmp <- raster::calc(bursted_dbbmm, sum) # raster::calc :Calculate values for a new Raster* object from another Raster* object
# using a formula. returns a RasterLayer if fun returns a single value (e.g. sum)
rm(bursted_dbbmm)
bb <- new(".UD", tmp / sum(raster::values(tmp))) # new() creates object from Class ("UD.")
rm(tmp)
# Calculate volume area (m^2) within 50% (core) and 95% (general use) contours. Note: absolute scale
# Note: Based on 'An introduction to the 'move' package', it's the opposite of what you might expect:
# "A cell with a very high value in the UD raster will have a very low value in the contour raster"
# area.50 <- round(sum(raster::values(move::getVolumeUD(bb) <= .50)), 4)
# area.95 <- round(sum(raster::values(move::getVolumeUD(bb) <= .95)), 4)
# area.50 <- sum(raster::values(move::getVolumeUD(bb) <= .50)) # 2024-01-08 remove rounding here. Seemingly had no effect
# area.95 <- sum(raster::values(move::getVolumeUD(bb) <= .95)) # 2024-01-08 remove rounding here. Seemingly had no effect
# Below calc introduced in 2022-10-08 commit, message =
#"changed code volume area: instead of using getVolumeUD() from the move package,
# we just did the calculations on the UD raster and not using the function.
# these seem to produce very plausible estimates!"
# This change creates unnaturally small areas.
# Changed back to previous calcs 2023-02-24. Added rounding
# area.50.new <- round(sum(raster::values(bb) >= (max(bb@data@values) * 0.5)), 4)
# area.95.new <- round(sum(raster::values(bb) >= (max(bb@data@values) * 0.05)), 4)
area.50.new <- sum(raster::values(bb) >= (max(bb@data@values) * 0.5)) # 2024-01-08 remove rounding here. Seemingly had no effect
area.95.new <- sum(raster::values(bb) >= (max(bb@data@values) * 0.05))# 2024-01-08 remove rounding here. Seemingly had no effect
# Combine in single df
area.ct <- data.frame(
core.use.new = area.50.new,
general.use.new = area.95.new
) # 2024-01-08 add rounding here if necessary. Maybe not.
# Add ID id
area.ct$ID <- i
# Put in list
bb.list[[counter]] <- area.ct
# need to reproject bb back to UTM else the resolution is unequal in x & y
# only an issue for ascii raster format
names(bb) <- i # puts name in slot, means scaleraster can name its output properly with lapply
# Export aggregated individual UDs to as ascii files for further exploration/spatial analysis in GIS software.
# Needed for when the aim is to plot population-level and normalized UDs per species.
raster::writeRaster(bb, # x has unequal horizontal and vertical resolutions. Such data cannot be stored in arc-ascii format
file.path(savedir, paste0(i, writeRasterExtension)),
format = writeRasterFormat, # "ascii"
datatype = writeRasterDatatype, # "FLT4S"
if (writeRasterFormat != "CDF") bylayer = TRUE, # bylayer kills ncdf4
overwrite = TRUE)
gc() # cleanup
} # close for i in unique data$ID
# Put everything in a data.frame
md <- dplyr::bind_rows(bb.list,
.id = "column_label"
) |>
dplyr::select(!"column_label") # remove column_label column
# 2023-08-30 quoted column_label to hopefully address gbm.factorplot: no visible binding for global variable ‘column_label’
md$core.use <- (rasterres * md$core.use.new) / 1000000 # convert from cells/pixels to metres squared area based on cell size, then to kilometres squared area
md$general.use <- (rasterres * md$general.use.new) / 1000000
write.csv(md,
file = file.path(savedir, absVolumeAreaSaveName),
row.names = FALSE
)
# 2023-10-04 Vital memory bug warning
if (all(md$core.use == md$core.use[1])) message("All core UDs identical. Possibly due to insufficient memory for raster calculations. Check rasterResolution")
if (all(md$general.use == md$general.use[1])) message("All general UDs identical. Possibly due to insufficient memory for raster calculations. Check rasterResolution")
if (length(which(md$core.use == max(md$core.use, na.rm = TRUE))) > 1) message("More than 1 individual share exactly the same max value for core use, possibly due to insufficient memory for raster calculations. Check rasterResolution")
if (length(which(md$general.use == max(md$general.use, na.rm = TRUE))) > 1) message("More than 1 individual share exactly the same max value for general use, possibly due to insufficient memory for raster calculations. Check rasterResolution")
if (length(which((md$core.use - md$general.use) == 0)) > 0) message("1 or more individuals have exactly the same value for core and general use, possibly due to insufficient memory for raster calculations. Check rasterResolution")
} # close function
Try the movegroup package in your browser
Any scripts or data that you put into this service are public.
movegroup documentation built on May 29, 2024, 3:36 a.m.
R Package Documentation
Browse R Packages
We want your feedback!
Note that we can't provide technical support on individual packages. You should contact the package authors for that.
Defines functions movegroup
Documented in movegroup
#' Automates dynamic Brownian bridge movement model construction across individuals
#'
#' Automates dynamic Brownian bridge movement model calculation for utilization distribution (UD)
#' estimation for multiple individuals simultaneously, using functions in the 'move' package. The
#' authors are indebted to the move package authors Bart Kraunstauber, Marco Smolla, and Anne K
#' Scharf, and to Sarah Becker for seed code which inspired the development of the
#' movegroup::movegroup function.
#'
#' Step 1. Filter individuals.
#' Remove those individuals for which there are insufficient data i.e. number of re-locations is
#' smaller than the window size parameter value (default = 31).
#' Step 2. Generate universal raster.
#' Based on all remaining data, a universal raster is generated where the calculated UDs are plotted
#' into.
#'
#' Step 3. Loop through individuals.
#' Individuals are looped through to construct individual-level movement models (on an absolute
#' scale).
#'
#' See www.GitHub.com/SimonDedman/movegroup for issues, feedback, and development
#' suggestions.
#'
#' install_git('https://gitlab.com/bartk/move.git') #Installs 'move' development version
#' @import utils
#' @importFrom lubridate is.POSIXt
#' @importFrom sp CRS SpatialPoints spTransform proj4string
#' @importFrom beepr beep
#' @importFrom dplyr mutate rename group_by summarise filter semi_join distinct ungroup arrange across bind_cols pull
#' @importFrom tidyr drop_na
#' @importFrom raster raster projectExtent res ncell setValues calc values writeRaster
#' @importFrom move move timeLag burst brownian.bridge.dyn brownian.motion.variance.dyn getVolumeUD
#' @importFrom rlang .data
#' @importFrom grDevices graphics.off
#' @importFrom graphics par
#' @importFrom methods new
#' @importFrom stats setNames
#'
#' @export movegroup
#'
#' @param data Data frame object containing the data. Requires columns Lat Lon DateTime ID and
#' potentially a grouping column (not currently implemented, email to request). Column names
#' specified in later parameters.
#' @param ID Name of animal tag ID column in data. "Character".
#' @param Datetime Column name in data that contains date/time stamps for each recorded detection.
#' Must be in POSIXct format. "Character".
#' @param Lat Name of latitude column in data. "Character".
#' @param Lon Name of longitude column in data. "Character".
# #' @param Group Name of grouping column in data. Unused 2023-08-14.
#' @param dat.TZ Timezone of data for as.POSIXct. Default "US/Eastern".
#' @param proj CRS for move function. Default sp::CRS("+proj=longlat +datum=WGS84").
#' @param sensor Sensor for move function. Single character or vector with length of the number of
#' coordinates. Optional. Default "VR2W".
#' @param moveLocError Location error (m) in the 'brownian.bridge.dyn' function in the 'move'
#' package. Numeric. Either single or a vector of length nrow data. If using passive acoustic data
#' this is the detection range of the receiver(s). Default 1. See MoveLocErrorCalc function for
#' satellite data with state space modelled locations with 95% confidence intervals for latlon i.e.
#' lat and lon025 and 975.
#' @param timeDiffLong Single numeric value. Threshold value in timeDiffUnits designating the length
#' of long breaks in re-locations. Used for bursting a movement track into segments, thereby
#' removing long breaks from the movement track. See ?move::bursted for details. Default 2 hours is
#' arbitrary, looping through 18, 24, and 36 hours for satellite data on great hammerhead sharks
#' revealed volume areas for core and general use gradually rise with timeDiffLong increases,
#' multiple small dots of presence get blobbed together, and therefore sometimes this covers land.
#' Ideally one would not discard any data, in which case one should choose a value higher than the
#' largest between-detections gap in their dataset (or just pick a very large number). This
#' parameter is useful when the model would otherwise get stuck trying to calculate a UD for an
#' individual with a very large home range that is inadequately captured by a receiver array.
#' Default 2.
#' @param timeDiffUnits Character. Unit for timeDiffLong. Default "hours".
#' @param center US English alternate to centre. Do you want to center the move object within
#' extent? See spTransform. Default TRUE.
#' @param centre British English alternate to center. Do you want to centre the move object within
#' extent? See spTransform. Default NULL.
#' @param buffpct Buffer extent for raster creation, proportion of 1. Default 0.3, can try e.g. 3
#' for a large buffer to avoid clipping, at the cost of file size, but later cropping in
#' plotraster.R will remove extraneous blank space.
#' @param rasterExtent Extent of raster created around data. If NULL (default), calculated from
#' data, buffpct, rasterResolution. Else length 4 vector, c(xmn, xmx, ymn, ymx) decimal latlon
#' degrees. Don't go to 90 (degrees) north or south for ymax or ymin. Doesn't prevent constraint to
#' data limits (in plot anyway), but prevents raster clipping crash.
#' @param rasterCRS CRS for raster creation. Default sp::CRS("+proj=utm +zone=17 +datum=WGS84").
#' @param rasterResolution Single numeric value to set raster resolution - cell size (width and
#' height) in metres. 111000: 1 degree lat = 111km. Trade-off between small res = big file & processing
#' time. Should be a function of the spatial resolution of your receivers or positioning tags.
#' Higher resolution will lead to more precision in the volume areas calculations. Try using
#' 2*dbblocationerror, if dbblocationerror is a single value. Default 50 = 50m = 50m² = 0.00005 km²
#' (divide by 1000000). Try around the median of your moveLocError.
# #' @param dbblocationerror Location.error param in 'brownian.bridge.dyn' function in the 'move'
# #' package. single numeric value or vector of the length of coordinates that describes the error of
# #' the location (sender/receiver) system in map units. Or a character string with the name of the
# #' column containing the location error can be provided. Default is moveLocError. See
# #' MoveLocErrorCalc function for satellite data with state space modelled locations with 95%
# #' confidence intervals for latlon i.e. lat and lon025 and 975.
#' @param movemargin Margin size for variance calc in move::brownian.motion.variance.dyn and
#' behavioural change point analysis in move::brownian.bridge.dyn. Must be an odd number. Default 11.
#' @param dbbext Ext param in the 'brownian.bridge.dyn' function in the 'move' package. Extends
#' bounding box around track. Numeric single (all edges), double (x & y), or 4 (xmin xmax ymin ymax)
#' . Default 3. Excessive buffering will get cropped automatically.
#' @param dbbwindowsize The window.size param in the 'brownian.bridge.dyn' function
#' in the 'move' package. The size of the moving window along the track. Larger windows provide more
#' stable/accurate estimates of the brownian motion variance but are less well able to capture more
#' frequent changes in behaviour. Number must be odd. Code will not run if total detections of
#' individual < window size (default 23), which must be >= 2*movemargin (default 11).
#' @param writeRasterFormat Character. Output file type for raster::writeRaster param format.
#' Default "ascii". TO DEPRECIATE.
#' @param writeRasterExtension Character. Output file extension for raster::writeRaster param
#' extension. Default ".asc". TO DEPRECIATE.
#' @param writeRasterDatatype Character. Data type for writing values to disk for
#' raster::writeRaster param Datatype. Default "FLT4S". TO DEPRECIATE.
#' @param absVolumeAreaSaveName File name plus extension where UD estimates are saved. Default
#' "VolumeArea_AbsoluteScale.csv".
#' @param savedir Save outputs to a temporary directory (default) else change
#' to desired directory e.g. "/home/me/folder". Do not use getwd() for this.
#' Do NOT include terminal slash. Directory must exist. Default tempdir().
#' @param alerts Audio warning for failures. Default TRUE.
#'
#' @return Individual-level utilization distributions, saved as rasters, as well
#' as calculated volume area estimates for 50 and 95pct contours, saved in a
#' .csv file. Motion variance csvs per individual ("MotionVariance.csv"), see
#' move::brownian.motion.variance.dyn. No processed object is returned, i.e. bad: "objectname <-
#' movegroup()", good: "movegroup()".
#' @details
#' When used together, the order of functions would be: movegroup, scaleraster, alignraster if
#' required, plotraster.
#'
#' ## Errors and their origins:
#'
#' 1. Error in .local(object, raster, location.error = location.error, ext = ext: Higher y grid not
#' large enough, consider extending the raster in that direction or enlarging the ext argument.
#' Increase buffpct, e.g. to 3.
#'
#' 2. Error in .data\[\[dttm\]\]: Must subset the data pronoun with a string, not a <POSIXct/POSIXt>
#' object. Use "ColName" not dataframe$ColName syntax for Datetime, ID, Lat, Lon.
#'
#' 3. Error in splice(dot_call(capture_dots, frame_env = frame_env, named = named,: object
#' 'DateTime' not found. Use "ColName" not ColName syntax for Datetime, ID, Lat, Lon.
#'
#' 4. Error in .local(object, raster, location.error = location.error, ext = ext: Higher x grid not
#' large enough, consider extending the raster in that direction or enlarging the ext argument. Try
#' "buffpct = 1," , then larger e.g. 3, if still getting the error.
#'
#' 5. cannot allocate vector of size (BIG) Gb: Increase rasterResolution value.
#'
#' 6. In min/max: No non-missing arguments to min; returning Inf: likely not enough memory, increase
#' rasterResolution value.
#'
#' 7. Error in tmp\[\[i\]\]: subscript out of bounds. dbbmmwindowsize may be too large relative to nrow
#' of that individual. Try lowering movemargin (default 11, has to be odd) and then lowering
#' dbbmmwindowsize (default 23, has to be >=2*movemargin, has to be odd).
#'
#' @examples
#' \donttest{
#' # load data
#' data("TracksCleaned")
#' # run function
#' movegroup(
#' data = TracksCleaned,
#' ID = "Shark",
#' Datetime = "Datetime",
#' Lat = "Lat",
#' Lon = "Lon",
#' savedir = tempdir())
#' }
#'
#' @author Simon Dedman, \email{simondedman@@gmail.com}
#' @author Maurits van Zinnicq Bergmann, \email{mauritsvzb@@gmail.com}
#'
#' @references Kranstauber, B., Kays, R., LaPoint, S. D., Wikelski, M. and Safi, K. (2012), A
#' dynamic Brownian bridge movement model to estimate utilization distributions for heterogeneous
#' animal movement. Journal of Animal Ecology. doi: 10.1111/j.1365-2656.2012.01955.x
#'
#' Kranstauber, B., M. Smolla & A. K. Scharf. 2019. Move: visualizing and analyzing animal track
#' data. R package version 4.2.4 (at 2023-08-15). https://CRAN.R-project.org/package=move.
movegroup <- function(
data = NULL, # Data frame object containing the data. Requires columns Lat Lon DateTime ID and optionally a grouping column.
ID = NULL, # Name of animal tag ID column in data.
Datetime = NULL, # Column name in data that contains date/time stamps for each recorded detection. Must be in POSIXct format.
Lat = NULL, # name of Lat & Lon columns in data.
Lon = NULL,
# Group = NULL, # name of grouping column in data. CURRENTLY UNUSED; MAKE USER DO THIS?
dat.TZ = "US/Eastern", # timezone for as.POSIXct.
proj = sp::CRS("+proj=longlat +datum=WGS84"), # CRS for move function.
sensor = "VR2W", # sensor for move function. Single character or vector with length of the number of coordinates. Optional.
moveLocError = 1, # location error in metres for move function. Numeric. Either single or a vector of length nrow data.
timeDiffLong = 2, # threshold length of time in timeDiffUnits designating long breaks in relocations.
timeDiffUnits = "hours", # units for time difference for move function.
center = TRUE, # center move object within extent? See spTransform.
centre = NULL, # British English alternate to center.
buffpct = 0.3, # buffer extent for raster creation, proportion of 1.
rasterExtent = NULL, # if NULL, raster extent calculated from data, buffpct, rasterResolution. Else length 4 vector, c(xmn, xmx, ymn, ymx) decimal latlon degrees. Don't go to 90 for ymax
# Doesn't prevent constraint to data limits (in plot anyway), but prevents raster clipping crash
rasterCRS = sp::CRS("+proj=utm +zone=17 +datum=WGS84"), # CRS for raster creation. This is around Bimini, Bahamas.
rasterResolution = 50, # numeric vector of length 1 or 2 to set raster resolution - cell size in metres. 111000: 1 degree lat = 111km
# dbblocationerror = moveLocError, # location.error param in brownian.bridge.dyn. Could use the same as moveLocError?
movemargin = 11, # Margin size for variance calc in move::brownian.motion.variance.dyn and behavioral change point analysis in
# move::brownian.bridge.dyn. Must be an odd number. Default 11.
dbbext = 3, # ext param in brownian.bridge.dyn. Extends bounding box around track. Numeric single (all edges), double (x & y), or 4 (xmin xmax ymin ymax). Default 3.
dbbwindowsize = 23, # window.size param in brownian.bridge.dyn. The size of the moving window along the track. Larger windows provide more
# stable/accurate estimates of the brownian motion variance but are less well able to capture more frequent changes in behavior.
# This number has to be odd. A dBBMM is not run if total detections of individual < window size (default 23). Must be >=2*margin (movemargin) which is 11 so >=22,
# but odd so >=23.
writeRasterFormat = "ascii",
writeRasterExtension = ".asc",
writeRasterDatatype = "FLT4S",
absVolumeAreaSaveName = "VolumeArea_AbsoluteScale.csv",
savedir = tempdir(), # save outputs to a temporary directory (default) else change to current
# directory e.g. "/home/me/folder". Do not use getwd() here.
alerts = TRUE # audio warning for failures
) {
# If savedir has a terminal slash, remove it, it's added later
if (substr(x = savedir, start = nchar(savedir), stop = nchar(savedir)) == "/") savedir = substr(x = savedir, start = 1, stop = nchar(savedir) - 1)
# if savedir doesn't exist, can't save or setwd into it, therefore create it
if (!file.exists(savedir)) dir.create(savedir)
oldpar <- par(no.readonly = TRUE) # defensive block, thanks to Gregor Sayer
oldwd <- getwd()
oldoptions <- options()
on.exit(par(oldpar))
on.exit(setwd(oldwd), add = TRUE)
on.exit(options(oldoptions), add = TRUE)
setwd(savedir)
if (alerts) options(error = function() {
beepr::beep(9)# give warning noise if it fails
graphics.off()# kill all graphics devices
setwd(oldwd) # reinstate original working directory. Probably redundant given on.exit
} # close options subcurly
) # close options
# edit center if centre used
if (!is.null(centre)) center <- centre
# Create writeRasterExtension from writeRasterFormat
if (is.null(writeRasterExtension)) writeRasterExtension <- switch(EXPR = writeRasterFormat,
"ascii" = ".asc",
"raster" = ".grd",
"SAGA" = ".sdat",
"IDRISI" = ".rst",
"CDF" = ".nc",
"GTiff" = ".tif",
"ENVI" = ".envi",
"EHdr" = ".bil",
"HFA" = ".img")
data <- dplyr::rename(.data = data, # Rename user entry to "ID", Ditto Datetime Lat & Lon
Datetime = .data[[Datetime]],
ID = .data[[ID]],
Lat = .data[[Lat]],
Lon = .data[[Lon]]) |>
# prefixes "X" to numerical-named sharks to avoid issues later
dplyr::mutate(ID = make.names(ID))
if (!any(class(data$Datetime) == "POSIXct")) stop("Datetime column must be POSIXct")
# Add movelocationerror and dbblocationerror as columns if they're vectors, else their length
# desyncs from nrow(data) if any IDs < windowsize
if (nrow(data) == length(moveLocError)) data$moveLocError <- moveLocError
# if (nrow(data) == length(dbblocationerror)) data$dbblocationerror <- dbblocationerror
# Convert coordinate sets to SpatialPoints and project
cord.dec <- sp::SpatialPoints(cbind(data$Lon, data$Lat),
proj4string = sp::CRS("+proj=longlat")
)
# Transform to UTM by setting the EPSG to local UTM zone for the data.
cord.UTM <- as.data.frame(sp::spTransform(cord.dec, rasterCRS))
colnames(cord.UTM) <- c("NewEastingUTM", "NewNorthingUTM")
data <- cbind(data, cord.UTM) # 1308 x 7
# Construct movement models per individual.
# Some notes regarding the 'move package and construction of movement models: Several arguments
# need to be used to run the model. 1. Window size: corresponds with number of locations and moves
# along a given trajectory to estimate the MA parameter within defined subsections of the path.
# This increases the ability to detect breakpoints where changes in behaviour occur. The window
# size should relate to what kind of behaviours the model is desired to identify e.g., a window
# size of 23 means the sliding window is moved every 23 locations or every 23 hours (has to do
# with sampling interval) 2. Margin: motion variance based on only the middle section of the
# trajectory; the ends of the movement trajectory where no changes are allowed because at some
# stage you want to have a few locations to base your estimation of the variance on and how many
# locations in either side of the window we use for this, is called the margin. Smaller values for
# window size and margin is expected to give a higher frequency of behavioural changes; make these
# large for looking at migrations. 3. The raster dictates the grid cell size for the UD to be
# calculated per grid cell per individual. Create the raster of certain size that matches with
# coordinates used to make the move object 4. Extent: is incorporated if there are animal
# locations that border the edges of the raster.
# A dBBMM is not run if total detections of individual < window size (default value, 31).
# Below code checks and filters individuals with insufficient data.
# Below we set window size arbitrarily to 23
check1 <- data |>
dplyr::group_by(.data$ID) |>
dplyr::summarise(relocations = length(.data$Datetime))
check2 <- dplyr::filter(check1, .data$relocations >= dbbwindowsize) # filter: removed 2 rows (14%), 12 rows remaining
if (length(check1$ID) != length(check2$ID)) {
data <- dplyr::semi_join(data, check2) # Joining, by = "ID". semi_join: added no columns
check1 <- data |>
dplyr::group_by(.data$ID) |>
dplyr::summarise(relocations = length(.data$Datetime))
check2 <- dplyr::filter(check1, .data$relocations >= dbbwindowsize) # filter: no rows removed
length(check1$ID) == length(check2$ID)
}
# Create per-individual move object, project it, construct a dBBMM, calculate the volume area
# within the 50% and 95% contours, save as ASCII.
bb <- list()
bb.list <- list()
data <- tidyr::drop_na(data = data, ID) |>
dplyr::group_by(ID) |>
dplyr::distinct(Datetime, .keep_all = TRUE) |> # distinct (grouped): removed one row (<1%), 1,286 rows remaining
# prevents duplicate Datetime crash in move() later
dplyr::ungroup() # 1253 x 7 after removing as.numeric above
# Make a raster for the UD to plot into. Start with UTM.
# These coordinates need to be big enough to cover your data.
# May need to expand x and y ranges (i.e. the "e" variable below) if you encounter errors when constructing the DBBMM in the next code chunk.
# NOTE: "xUTM" grid should be larger than the "x" i.e. lower minima and higher maxima.
# The variable 'e' influences the extent [now done with buffpct]
# The resolution determines the grid size.
# Finer resolution/grid size (in m) i.e. lower value means longer computation time. May need/want to play with this value too.
# e <- 80 * 1000 # make e a function of range of extent as %
# Error in .local(object, raster, location.error = location.error, ext = ext, :
# Lower x grid not large enough, consider extending the raster in that direction or enlarging the ext argument
# make buffpct larger
if (!is.null(rasterExtent)) { # if xUTM object exists
xUTM <- sp::SpatialPoints(cbind(c(rasterExtent[1], rasterExtent[2]), c(rasterExtent[3], rasterExtent[4])), proj4string = sp::CRS("+proj=longlat"))
xUTM <- as.data.frame(sp::spTransform(xUTM, rasterCRS)) # Transform to UTM
xUTM <- raster::raster( # create a raster
xmn = xUTM[1,1] - ((xUTM[2,1] - xUTM[1,1]) * buffpct), # with xUTM's values as extents
xmx = xUTM[2,1] + ((xUTM[2,1] - xUTM[1,1]) * buffpct),
ymn = xUTM[1,2] - ((xUTM[2,2] - xUTM[1,2]) * buffpct),
ymx = xUTM[2,2] + ((xUTM[2,2] - xUTM[1,2]) * buffpct),
crs = rasterCRS, # +proj=utm +zone=27 +datum=WGS84 +units=m +no_defs
resolution = rasterResolution # 50
) # close raster
} else { # exists(xUTM), else create from data
xUTM <- raster::raster(
xmn = min(data$NewEastingUTM, na.rm = TRUE) - ((max(data$NewEastingUTM, na.rm = TRUE) - min(data$NewEastingUTM, na.rm = TRUE)) * buffpct),
xmx = max(data$NewEastingUTM, na.rm = TRUE) + ((max(data$NewEastingUTM, na.rm = TRUE) - min(data$NewEastingUTM, na.rm = TRUE)) * buffpct),
ymn = min(data$NewNorthingUTM, na.rm = TRUE) - ((max(data$NewNorthingUTM, na.rm = TRUE) - min(data$NewNorthingUTM, na.rm = TRUE)) * buffpct),
ymx = max(data$NewNorthingUTM, na.rm = TRUE) + ((max(data$NewNorthingUTM, na.rm = TRUE) - min(data$NewNorthingUTM, na.rm = TRUE)) * buffpct),
crs = rasterCRS, # +proj=utm +zone=27 +datum=WGS84 +units=m +no_defs
resolution = rasterResolution # 50
) # close raster
} # close else
# run move on all data together to generate global CRS for raster
alldata <- data |>
dplyr::arrange(Datetime) |>
dplyr::group_by(Datetime) |> # remove duplicates
dplyr::summarise(dplyr::across(where(~ is.numeric(.)), mean, na.rm = TRUE),
dplyr::across(where(~ is.character(.) | lubridate::is.POSIXt(.)), dplyr::first)) |> # library(lubridate)
dplyr::mutate(dplyr::across(where(~ is.numeric(.)), ~ ifelse(is.nan(.), NA, .)), #convert NaN to NA. POSIX needs lubridate
dplyr::across(where(~ is.character(.)), ~ ifelse(is.nan(.), NA, .))) |> # https://community.rstudio.com/t/why-does-tidyrs-fill-work-with-nas-but-not-nans/25506/5
dplyr::ungroup()
alldata <- as.data.frame(alldata)
moveall <- move::move(
x = alldata$Lon, # numeric vector with x coordinates if non-Movebank data are provided (e.g. data$x).
y = alldata$Lat, # numeric vector with y coordinates if non-Movebank data are provided (e.g. data$y).
time = alldata$Datetime, # vector of time stamps with POSIXct conversion if non-Movebank data are provided,
# i.e. as.POSIXct(data$timestamp, format="%Y-%m-%d %H:%M:%S", tz="UTC")
data = alldata, # extra data associated with the relocations
proj = proj, # projection method for non-Movebank data; requires a valid CRS (see CRS-class) object, e.g. CRS("+proj=longlat +ellps=WGS84")
# animal = data$ID,
sensor = sensor # Sensor name(s), either single character or a vector with length of the number of coordinates.
)
# Convert projection to Azimuthal Equi-Distance projection (aeqd)
rall <- sp::spTransform(moveall, center = center)
sp::proj4string(rall) # "+proj=aeqd +lat_0=44.99489997 +lon_0=-17.48575004 +x_0=0 +y_0=0 +datum=WGS84 +units=m +no_defs"
rallCRS <- sp::CRS(sp::proj4string(rall))
class(rallCRS) # CRS
write.csv(sp::proj4string(rall), file.path(savedir, "CRS.csv"), row.names = FALSE)
saveRDS(rallCRS, file = file.path(savedir, "CRS.Rds"))
rm(moveall)
rm(alldata)
# Reproject to AEQD. Make a dummy object to get the correct projection.
x.i <- raster::raster(
xmn = min(data$NewEastingUTM),
xmx = max(data$NewEastingUTM),
ymn = min(data$NewNorthingUTM),
ymx = max(data$NewNorthingUTM),
crs = sp::proj4string(rall) # if rasterCRS then same as above! trying proj4string(r.i) which is AEQD from spTransform
# could add here: resolution = rasterResolution # 50 ####
)
# Reproject UTM to AEQD
# ?projectExtent # project the values of a Raster object to a new Raster object with another projection (CRS), i.e. projectExtent(from, to, ...))
newTemplate <- raster::projectExtent(xUTM, sp::proj4string(x.i))
# change newTemplate to xAEQD. Change xAEQD res so x & y match. Kills values.
raster::res(newTemplate) <- rep(rasterResolution, 2)
# Give newTemplate some values. Make Rep equal to the ncell dimension
ones <- rep(1, raster::ncell(newTemplate))
xAEQD <- raster::setValues(newTemplate, ones)
rm(newTemplate)
rm(ones)
# get resolution from raster in rasterlist, assign it object, squared
rasterres <- (raster::res(xAEQD)[1]) ^ 2
write.csv(x = data.frame(rasterres = rasterres,
rasterResolution = rasterResolution),
file = file.path(savedir, "Resolutions.csv"))
# Loop through all unique tags
counter <- 0
for (i in unique(data$ID)) { # i <- unique(data$ID)[1]
# Print individual
counter <- counter + 1
message(
"processing ", which(unique(data$ID) %in% i),
" of ", length(unique(data$ID))
)
# Filter individual data
data.i <- as.data.frame(data[data$ID == i, ])
# create move object
move.i <- move::move(
x = data.i$Lon,
y = data.i$Lat,
time = data.i$Datetime,
proj = proj,
data = data.i,
animal = data.i$ID,
sensor = sensor
)
# Check the current projection
sp::proj4string(move.i) # "+proj=longlat +datum=WGS84 +no_defs"
# Incorporate uncertainty in the model by including a location error.
if ("moveLocError" %in% colnames(data.i)) {
move.i$LocationError <- data.i$moveLocError
} else {
if (exists("moveLocError")) {
if (length(moveLocError) == 1) {
move.i$LocationError <- moveLocError # single value, replicated down for each relocation
} else { # else if multiple values
if (length(moveLocError) == nrow(data.i)) { # should be same length as full dataset
move.i$LocationError <- data.i |> # take the full dataset,
dplyr::bind_cols(moveLocError = moveLocError) |> # cbind the full movelocerror
dplyr::filter(ID == i) |> # filter for just this ID
dplyr::pull(moveLocError) # and pull just the movelocerror for this ID
} else { # if not length 1 and not length of nrow(data)
stop(message("moveLocError must be either length 1 or length(nrow(data))")) # if not stop and tell user
} # close not length1 not same length as full dataset else
} # close length1 or else
} # close if exists movelocerror
}
# Convert projection to Azimuthal Equi-Distance projection (aeqd)
r.i <- sp::spTransform(move.i, CRSobj = rallCRS, center = center)
# Calculate time lag between consecutive detections.
# Obviously there is no time lag for the first detection, so we need to tell R this.
# Knowing time lags in acoustic telemetry is important, as the motion variance is based on the
# time difference b/w detections at consecutive locations i.e. larger time lag creates a wider
# bridge where the animal could've been and therefore potentially inflates the motion variance.
# Below code deals with this issue by identifying the location and number of detections with
# large time gaps. Below we will use an arbitrary value of 2 h.
TimeDiff <- move::timeLag(r.i, units = timeDiffUnits)
r.i$TimeDiff <- append(0, TimeDiff)
long <- which(r.i$TimeDiff > timeDiffLong)
# if no timediffs are longer than timedifflong, long is integer(0), a potentially dangerous object.
# Below we exclude the data points that are 'long' i.e. create a large time gap. move::burst ?
bursted <- move::burst(
r.i,
c("normal", "long")[1 + (move::timeLag(r.i, units = timeDiffUnits) > timeDiffLong)]
)
rm(r.i)
# There are 2 types of burst: "normal" and "long". You can select for these in the dbbmm by selecting the factor level in the burstType command.
if ("moveLocError" %in% colnames(data.i)) {
moveLocError.i <- data.i$moveLocError
} else {
if (exists("moveLocError")) {
if (length(moveLocError) == 1) {
moveLocError.i <- moveLocError # single value, replicated down for each relocation
# Robs: 1 m: gps loc of shark inferred from boat coord + bearing + distance estimate
} else { # else if multiple values
if (length(moveLocError) == nrow(data)) { # should be same length as full dataset
moveLocError.i <- data |> # take the full dataset,
dplyr::bind_cols(moveLocError = moveLocError) |> # cbind the full movelocerror
dplyr::filter(ID == i) |> # filter for just this ID
dplyr::pull(moveLocError) # and pull just the movelocerror for this ID
} else { # if not length 1 and not length of nrow(data)
stop(message("moveLocError must be either length 1 or length(nrow(data))")) # if not stop and tell user
} # close not length1 not same length as full dataset else
} # close length1 or else
} # close if exists moveLocError
} # close ifelse
# location error needs to be a positive number. Replace zeroes with 0.00001
moveLocError.i[which(moveLocError.i == 0)] <- 0.00001
# Construct the model. The time.step should reflect the ping frequency of the tag (in minutes)
bursted_dbbmm <- move::brownian.bridge.dyn(bursted,
burstType = "normal",
raster = xAEQD, # has to be AEQD for metre-based calculations, presuambly?
# Error:The projection of the raster and the Move object are not equal.
# Need bursted, r.i, to be the same projection as xAEQD
location.error = moveLocError.i,
ext = dbbext, # dbbext
margin = movemargin,
window.size = dbbwindowsize # must be >=2*margin which is 11 so >=22, but odd so >=23
)
# Build dBMvarianceBurst object with brownian.motion.variance.dyn
bursted_motionvar <- move::brownian.motion.variance.dyn(bursted,
location.error = moveLocError.i,
window.size = dbbwindowsize, # must be >=2*margin which is 11 so >=22, but odd so >=23
margin = movemargin
)
mtnvar <- as.data.frame(bursted_motionvar) # extract the data
mtnvar$motionVariance <- move::getMotionVariance(bursted_motionvar) # add motionvariance
write.csv(mtnvar,
file = file.path(savedir, paste0(i, "_MotionVariance.csv")),
row.names = FALSE)
rm(bursted)
rm(mtnvar)
# Re-standardize (Dr. Kranstauber's trouble shooting solution).
# Occurring errors are "due to limits of accuracy during calculations.
# Given the error is really small (<.000001 %) I would not worry to much and re-standardize".
# Then aggregate UD segments.
tmp <- raster::calc(bursted_dbbmm, sum) # raster::calc :Calculate values for a new Raster* object from another Raster* object
# using a formula. returns a RasterLayer if fun returns a single value (e.g. sum)
rm(bursted_dbbmm)
bb <- new(".UD", tmp / sum(raster::values(tmp))) # new() creates object from Class ("UD.")
rm(tmp)
# Calculate volume area (m^2) within 50% (core) and 95% (general use) contours. Note: absolute scale
# Note: Based on 'An introduction to the 'move' package', it's the opposite of what you might expect:
# "A cell with a very high value in the UD raster will have a very low value in the contour raster"
# area.50 <- round(sum(raster::values(move::getVolumeUD(bb) <= .50)), 4)
# area.95 <- round(sum(raster::values(move::getVolumeUD(bb) <= .95)), 4)
# area.50 <- sum(raster::values(move::getVolumeUD(bb) <= .50)) # 2024-01-08 remove rounding here. Seemingly had no effect
# area.95 <- sum(raster::values(move::getVolumeUD(bb) <= .95)) # 2024-01-08 remove rounding here. Seemingly had no effect
# Below calc introduced in 2022-10-08 commit, message =
#"changed code volume area: instead of using getVolumeUD() from the move package,
# we just did the calculations on the UD raster and not using the function.
# these seem to produce very plausible estimates!"
# This change creates unnaturally small areas.
# Changed back to previous calcs 2023-02-24. Added rounding
# area.50.new <- round(sum(raster::values(bb) >= (max(bb@data@values) * 0.5)), 4)
# area.95.new <- round(sum(raster::values(bb) >= (max(bb@data@values) * 0.05)), 4)
area.50.new <- sum(raster::values(bb) >= (max(bb@data@values) * 0.5)) # 2024-01-08 remove rounding here. Seemingly had no effect
area.95.new <- sum(raster::values(bb) >= (max(bb@data@values) * 0.05))# 2024-01-08 remove rounding here. Seemingly had no effect
# Combine in single df
area.ct <- data.frame(
core.use.new = area.50.new,
general.use.new = area.95.new
) # 2024-01-08 add rounding here if necessary. Maybe not.
# Add ID id
area.ct$ID <- i
# Put in list
bb.list[[counter]] <- area.ct
# need to reproject bb back to UTM else the resolution is unequal in x & y
# only an issue for ascii raster format
names(bb) <- i # puts name in slot, means scaleraster can name its output properly with lapply
# Export aggregated individual UDs to as ascii files for further exploration/spatial analysis in GIS software.
# Needed for when the aim is to plot population-level and normalized UDs per species.
raster::writeRaster(bb, # x has unequal horizontal and vertical resolutions. Such data cannot be stored in arc-ascii format
file.path(savedir, paste0(i, writeRasterExtension)),
format = writeRasterFormat, # "ascii"
datatype = writeRasterDatatype, # "FLT4S"
if (writeRasterFormat != "CDF") bylayer = TRUE, # bylayer kills ncdf4
overwrite = TRUE)
gc() # cleanup
} # close for i in unique data$ID
# Put everything in a data.frame
md <- dplyr::bind_rows(bb.list,
.id = "column_label"
) |>
dplyr::select(!"column_label") # remove column_label column
# 2023-08-30 quoted column_label to hopefully address gbm.factorplot: no visible binding for global variable ‘column_label’
md$core.use <- (rasterres * md$core.use.new) / 1000000 # convert from cells/pixels to metres squared area based on cell size, then to kilometres squared area
md$general.use <- (rasterres * md$general.use.new) / 1000000
write.csv(md,
file = file.path(savedir, absVolumeAreaSaveName),
row.names = FALSE
)
# 2023-10-04 Vital memory bug warning
if (all(md$core.use == md$core.use[1])) message("All core UDs identical. Possibly due to insufficient memory for raster calculations. Check rasterResolution")
if (all(md$general.use == md$general.use[1])) message("All general UDs identical. Possibly due to insufficient memory for raster calculations. Check rasterResolution")
if (length(which(md$core.use == max(md$core.use, na.rm = TRUE))) > 1) message("More than 1 individual share exactly the same max value for core use, possibly due to insufficient memory for raster calculations. Check rasterResolution")
if (length(which(md$general.use == max(md$general.use, na.rm = TRUE))) > 1) message("More than 1 individual share exactly the same max value for general use, possibly due to insufficient memory for raster calculations. Check rasterResolution")
if (length(which((md$core.use - md$general.use) == 0)) > 0) message("1 or more individuals have exactly the same value for core and general use, possibly due to insufficient memory for raster calculations. Check rasterResolution")
} # close function
Try the movegroup package in your browser
Any scripts or data that you put into this service are public.
R Package Documentation
Browse R Packages
We want your feedback!
Note that we can't provide technical support on individual packages. You should contact the package authors for that.
Embedding an R snippet on your website
Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.