detectionHistory: Species detection histories for occupancy analyses

View source: R/detectionHistory.R

detectionHistoryR Documentation

Species detection histories for occupancy analyses

Description

This function generates species detection histories that can be used in occupancy analyses, e.g. with package unmarked. It generates detection histories in different formats, with adjustable occasion length and occasion start time.

Usage

detectionHistory(
  recordTable,
  species,
  camOp,
  output = c("binary", "count"),
  stationCol = "Station",
  speciesCol = "Species",
  recordDateTimeCol = "DateTimeOriginal",
  recordDateTimeFormat = "ymd HMS",
  occasionLength,
  minActiveDaysPerOccasion,
  maxNumberDays,
  day1,
  buffer,
  includeEffort = TRUE,
  scaleEffort = FALSE,
  occasionStartTime = "deprecated",
  datesAsOccasionNames = FALSE,
  timeZone,
  writecsv = FALSE,
  outDir,
  unmarkedMultFrameInput
)

Arguments

recordTable

data.frame. the record table created by recordTable

species

character. the species for which to compute the detection history

camOp

The camera operability matrix as created by cameraOperation

output

character. Return binary detections ("binary") or counts of detections ("count")

stationCol

character. name of the column specifying Station ID in recordTable

speciesCol

character. name of the column specifying species in recordTable

recordDateTimeCol

character. name of the column specifying date and time in recordTable

recordDateTimeFormat

character. Format of column recordDateTimeCol in recordTable

occasionLength

integer. occasion length in days

minActiveDaysPerOccasion

integer. minimum number of active trap days for occasions to be included (optional)

maxNumberDays

integer. maximum number of trap days per station (optional)

day1

character. When should occasions begin: station setup date ("station"), first day of survey ("survey"), a specific date (e.g. "2015-12-31")?

buffer

integer. Makes the first occasion begin a number of days after station setup. (optional)

includeEffort

logical. Compute trapping effort (number of active camera trap days per station and occasion)?

scaleEffort

logical. scale and center effort matrix to mean = 0 and sd = 1?

occasionStartTime

(DEPRECATED) integer. time of day (the full hour) at which to begin occasions. Please use argument occasionStartTime in cameraOperation instead.

datesAsOccasionNames

If day1 = "survey", occasion names in the detection history will be composed of first and last day of that occasion.

timeZone

character. Must be a value returned by OlsonNames

writecsv

logical. Should the detection history be saved as a .csv?

outDir

character. Directory into which detection history .csv file is saved

unmarkedMultFrameInput

logical. Return input for multi-season occupancy models in unmarked (argument "y" in unmarkedMultFrame?

Details

The function computes a species detection matrix, either as a detection-by-date or a detection-by-occasion matrix. day1 defines if each stations detection history will begin on that station's setup day (day1 = "station") or if all station's detection histories have a common origin (the day the first station was set up if day1 = "survey" or a fixed date if, e.g. day1 = "2015-12-31"). If day1 is a date, as.Date must be able to understand it. The most suitable format is "YYYY-MM-DD", e.g. "2015-12-31".

output is analogous to spatialDetectionHistory. It makes the function return either counts of detections during occasions, or a binary indicator for whether the species was detected.

includeEffort controls whether an additional effort matrix is computed or not. This also affects the detection matrices. If includeEffort = FALSE, all occasions in which a station was not set up or malfunctioning (NA or 0 in camOp) will result in NAs in the detection history. If includeEffort = TRUE, the record history will only contain 0 and 1, and no NAs. The effort matrix can then be included in occupancy models as a (continuous) observation covariate to estimate the effect of effort on detection probability.

The number of days that are aggregated is controlled by occasionLength. occasionStartTime will be removed from the function. It has moved to cameraOperation, to ensure daily effort is computed correctly and takes the occasion start time into account. default). This may be relevant for nocturnal animals, in which 1 whole night would be considered an occasion.

The values of stationCol in recordTable must be matched by the row names of camOp (case-insensitive), otherwise an error is raised.

recordDateTimeFormat defaults to the "YYYY-MM-DD HH:MM:SS" convention, e.g. "2014-09-30 22:59:59". recordDateTimeFormat can be interpreted either by base-R via strptime or in lubridate via parse_date_time (argument "orders"). lubridate will be used if there are no "%" characters in recordDateTimeFormat.

For "YYYY-MM-DD HH:MM:SS", recordDateTimeFormat would be either "%Y-%m-%d %H:%M:%S" or "ymd HMS". For details on how to specify date and time formats in R see strptime or parse_date_time.

If the camera operation matrix (camOp) was created for a multi-season study (argument sesssionCol in cameraOperation was set, it will be detected automatically. Output can be for unmarkedMultFrame by setting unmarkedMultFrameInput = TRUE. Each row corresponds to a site, and the columns are in season-major, occasion-minor order, e.g. season1-occasion1, season1-occasion2, etc.).

Value

Depending on the value of includeEffort and scaleEffort, a list with either 1, 2 or 3 elements. The first element is the species detection history. The second is the optional effort matrix and the third contains the effort scaling parameters.

detection_history

A species detection matrix

effort

A matrix giving the number of active camera trap days per station and occasion (= camera trapping effort). It is only returned if includeEffort = TRUE

effort_scaling_parameters

Scaling parameters of the effort matrix. It is only returned if includeEffort and scaleEffort are TRUE

Warning

Setting output = "count" returns a count of detections, not individuals. We strongly advise against using it as input for models of animal abundance (such as N-Mixture models) models which use counts as input.

Please note the section about defining argument timeZone in the vignette on data extraction (accessible via vignette("DataExtraction") or online (https://cran.r-project.org/package=camtrapR/vignettes/camtrapr3.pdf)).

Author(s)

Juergen Niedballa

Examples



# define image directory
wd_images_ID <- system.file("pictures/sample_images_species_dir", package = "camtrapR")

# load station information
data(camtraps)

# create camera operation matrix
camop_no_problem <- cameraOperation(CTtable      = camtraps,
                                    stationCol   = "Station",
                                    setupCol     = "Setup_date",
                                    retrievalCol = "Retrieval_date",
                                    hasProblems  = FALSE,
                                    dateFormat   = "dmy"
)

## Not run: 
if (Sys.which("exiftool") != ""){        # only run this function if ExifTool is available
recordTableSample <- recordTable(inDir               = wd_images_ID,
                                 IDfrom              = "directory",
                                 minDeltaTime        = 60,
                                 deltaTimeComparedTo = "lastRecord",
                                 exclude             = "UNID",
                                 timeZone            = "Asia/Kuala_Lumpur"
)
}

## End(Not run)
data(recordTableSample)    # load the record history, as created above


# compute detection history for a species

# without trapping effort
DetHist1 <- detectionHistory(recordTable         = recordTableSample,
                            camOp                = camop_no_problem,
                            stationCol           = "Station",
                            speciesCol           = "Species",
                            recordDateTimeCol    = "DateTimeOriginal",
                            species              = "VTA",
                            occasionLength       = 7,
                            day1                 = "station",
                            datesAsOccasionNames = FALSE,
                            includeEffort        = FALSE,
                            timeZone             = "Asia/Kuala_Lumpur"
)

DetHist1                     # this is a list with 1 element
DetHist1$detection_history   # this is the contained detection/non-detection matrix


# with effort / using base R to define recordDateTimeFormat
DetHist2 <- detectionHistory(recordTable          = recordTableSample,
                             camOp                = camop_no_problem,
                             stationCol           = "Station",
                             speciesCol           = "Species",
                             recordDateTimeCol    = "DateTimeOriginal",
                             species              = "VTA",
                             occasionLength       = 7,
                             day1                 = "station",
                             datesAsOccasionNames = FALSE,
                             includeEffort        = TRUE,
                             scaleEffort          = FALSE,
                             timeZone             = "Asia/Kuala_Lumpur"
)

DetHist2$detection_history  # detection history  (alternatively, use: DetHist2[[1]])
DetHist2$effort             # effort (alternatively, use: DetHist2[[2]])

# with effort / using lubridate package to define recordDateTimeFormat
DetHist2_lub <- detectionHistory(recordTable          = recordTableSample,
                                 camOp                = camop_no_problem,
                                 stationCol           = "Station",
                                 speciesCol           = "Species",
                                 recordDateTimeCol    = "DateTimeOriginal",
                                 recordDateTimeFormat = "ymd HMS",
                                 species              = "VTA",
                                 occasionLength       = 7,
                                 day1                 = "station",
                                 datesAsOccasionNames = FALSE,
                                 includeEffort        = TRUE,
                                 scaleEffort          = FALSE,
                                 timeZone             = "Asia/Kuala_Lumpur"
)    

DetHist2_lub$detection_history  # detection history  (alternatively, use: DetHist2_lub[[1]])
DetHist2_lub$effort             # effort (alternatively, use: DetHist2_lub[[2]])


# multi-season detection history

# load multi-season data
data(camtrapsMultiSeason)
data(recordTableSampleMultiSeason)

# multi-season camera operation matrix
camop_season <- cameraOperation(CTtable          = camtrapsMultiSeason,
                                    stationCol   = "Station",
                                    setupCol     = "Setup_date",
                                    sessionCol   = "session",
                                    retrievalCol = "Retrieval_date",
                                    hasProblems  = TRUE,
                                    dateFormat   = "dmy"
)

# multi-season detection history
DetHist_multi <- detectionHistory(recordTable      = recordTableSampleMultiSeason,
                            camOp                  = camop_season,
                            stationCol             = "Station",
                            speciesCol             = "Species",
                            species                = "VTA",
                            occasionLength         = 10,
                            day1                   = "station",
                            recordDateTimeCol      = "DateTimeOriginal",
                            includeEffort          = TRUE,
                            scaleEffort            = FALSE,
                            timeZone               = "UTC",
                            unmarkedMultFrameInput = TRUE
)

DetHist_multi



camtrapR documentation built on May 29, 2024, 6:45 a.m.