ctmm (version 0.5.3)

uere: Estimate UERE from calibration data

Description

Functions for estimating and assigning the User Equivalent Range Error (UERE) of a GPS device from calibration data.

Usage

uere(data)
uere(data) <- value
uere.fit(data,precision=1/2)
# S3 method for UERE
summary(object,level=0.95,...)

Arguments

data

telemetry object or list of telemetry objects, preferably with DOP columns.

value

UERE value(s) to assign to telemetry data (see details).

precision

Fraction of maximum possible digits of precision to target in categorical error fitting. precision=1/2 results in about 7 decimal digits of precision.

object

UERE object to summarize.

level

Confidence level for UERE estimate confidence intervals.

...

Further arguments are ignored.

Value

The UERE estimate.

Details

Often times GPS animal tracking devices return HDOP values but do not specifiy the device's UERE necessary to transform the HDOP values into absolute errors. uere.fit() allows users to estimate the UERE from calibration data, where the device was left fixed over a period of time. The calibration UERE can then be applied to tracking data with the uere()<- assignment method. Otherwise, when error=TRUE in ctmm, ctmm.fit will estimate the UERE simultaneously with the movement model, which is less reliable than using calibration data.

summary() applied to single UERE object will return UERE parameter estimates and confidence intervals, while summary() applied to a list of UERE objects will return a model-selection table, with AICc and reduced Z squared (goodness of fit) values.

See Also

as.telemetry, residuals.telemetry.

Examples

# NOT RUN {
# Load package and data
library(ctmm)
data(turtle)

# the first two datasets are calibration data
names(turtle)

# estimate UERE from calibration data
UERE <- uere.fit(turtle[1:2])
# inspect UERE estimate
summary(UERE)

# assign UERE to entire dataset
uere(turtle) <- UERE

# calculate residuals of calibration data
RES <- lapply(turtle[1:2],residuals)

# scatter plot of residuals with 50% and 95% quantiles
plot(RES,col.DF=NA,level.UD=c(0.50,0.95))

# check calibration data for autocorrelation using fast=FALSE because samples are small
ACFS <- lapply(RES,function(R){correlogram(R,fast=FALSE,dt=10 %#% 'min')})

# pooling ACFs
ACF <- mean(ACFS)

plot(ACF)
# }