Skip to contents

Call density estimate from a parameter file (deprecated)

Source: R/callDensityEstimate.R
cdeFromParamFile.Rd

Deprecated. Use cde directly. This was a near-complete duplicate of cde's body, reading its inputs from a bundled parameter object instead of explicit arguments. Being a separate copy, it had drifted from cde: it called pa_CV() without transect area weights (cde always weights by truncationDistance^2), and it had no equivalent of cde's NcIsTruncated guard. This wrapper derives cde's arguments from p and calls cde directly, so it now gets exactly the same area-weighted, truncation-aware behaviour as cde itself, rather than a second, unmaintained implementation of the same calculation.

Usage

cdeFromParamFile(
  p,
  season,
  snrDetFun = NULL,
  truncationDistance = Inf,
  snrTruncationThreshold = -Inf,
  NL = NULL
)

Arguments

p

A data.frame containing parameters for call density estimation. Usually created by calling function defaultOutputFileNames. Expected to contain: detectorParams$fullYearDetectionCsv, detectorParams$fullYearEffortFile, capHistFile, slParams$slMean/slStd/slSampleSize, tlParams$tlFile, w, k, modelType, numKnots, output.resolution.m, outerloop, transectFile, simResultsFile, paFile, siteCode, densityResultsFile.

season

TimeCode specifying month, season, or year for outputs

snrDetFun

OPTIONAL linear-model like structure (GLM,GAM,SCAM,etc) specifying the SNR-detection function to use. If this is not included, then the SNR-detection function will be derived from the capture history table.

truncationDistance

scalar or matrix of truncation distances. If a matrix is provided, then the dimensions should be 1xN with N being the same as the number of transects

snrTruncationThreshold

scalar SNR in dB below which the probability of detection will be forcibly set to zero.

NL

data.frame containing distribution of noise level parameters. This data.frame must contain the rows mean, sd, and sampleSize (similar to SL).

Value

data.frame containing call density inputs, results, and CVs

Details

$D_c = \fracN_c*(1-c)kTP_a\piw^2$

where: $D_c$ is call density $N_c$ is number of calls $c$ is false discovery rate $k$ is number of sensors (here always 1) $T$ is the duration of data analysed $P_a$ is the probability of detection in the study area $\piw^2$ is the study area (in km\^2)