geoEnvAccuracy: Coordinate accuracy
In occTest: Characterizing and Filtering Species Occurrence Data

geoEnvAccuracy

R Documentation

Coordinate accuracy

Description

Detect records with low accuracy in space and time

Usage

geoEnvAccuracy(
  df,
  xf,
  yf,
  af,
  dsf,
  ef,
  tf,
  method = "all",
  r.env,
  accept.threshold.cell = 0.5,
  accept.threshold.env = 0.5,
  bearing.classes = 10,
  distance.classes = 5,
  env.quantiles = c(0.3, 0.7),
  elev.threshold = 100,
  raster.elevation = NULL,
  verbose = FALSE,
  do = TRUE,
  doParallel = FALSE,
  mc.cores = 2
)

Arguments

`df`	data.frame of species occurrences
`xf`	character. column name in df containing the x coordinates
`yf`	character. column name in df containing the y coordinates
`af`	character. column name in df containing the coordinate uncertainty value (in the same)
`dsf`	character. column name in df containing the dataset to which the record belongs to (e.g. Forest Inventory of Spain)
`ef`	character. column name in df containing the registered elevation for the record.
`tf`	character. column name in df containing the dataset with the date/time where the species is recorded
`method`	character. Vector of methods to be used. See details. Default 'all'
`r.env`	raster. Raster with environmental data
`accept.threshold.cell`	numeric. Acceptance threshold for how much percentage of the Area of uncertainty in the cell we want to accept. Default to 0.5
`accept.threshold.env`	numeric. Default 0.5
`bearing.classes`	numeric. Default to 10.
`distance.classes`	integer. Default to 5.
`env.quantiles`	numeric. Default to c(0.3,0.7)
`elev.threshold`	numeric. Default to 100
`raster.elevation`	numeric. Default to 100
`verbose`	logical. Print messages? Default FALSE
`do`	logical. Should tests be performed? Default TRUE
`doParallel`	logical. Should computation use parallel functions? Default FALSE
`mc.cores`	numeric. How many cores to use? (used when doParallel = TRUE). Default 2

Details

Geoenvironmental accuracy function will implement differnt methods to assess occurrence accuracy in environmnental and geographic space.
Current implmented methods are: 'lattice' : tests for lattice arrangement in occurrence datasets. Borrowed from cd_round .
'elevDiff' : assess the elevation difference between a given raster (or automatically downloaded fro SRTM), and the elevation recorded. If differences >elev.threshold then the record is considered as a low accuracy threshold
'noDate' : assess whether there is a date or timestamp information in the record.
'noDateFormatKnown' : assess whether the information in the timestamp agrees with different formatting of Dates.
'outDateRange' : (not implemented) assess whether the record is within a user specified time frame.
'percDiffCell' : assess whether the record may be falling in a different raster cell given an information of coordinate accuracy.
'envDeviation' : assess whether the climate in a given record may be outside of the interval 30th-70th (default values) for a given variable due to coordinate uncertainty.