prefilter (1) determines Argos data type (LS or KF);
(2) converts dates to POSIXt & identifies observations with duplicate dates;
(3) orders observations in time; (4) removes duplicate observations;
(5) removes observations occurring within 60 s of one another (keeps first);
(6) shifts longitudes that straddle -180,180 to 0,360 and vice-versa;
(7) projects lonlat coords to mercator x,y coords (in km);
(8) adds location error multiplication factors based on Argos location
class (for type LS);
(9) uses a trip::sda to identify potential outlier locations.
trip::sda is a fast, vectorized version of argosfilter::sdafilter
see ?argosfilter::sdafilter for details on implementation
prefilter(
data,
vmax = 5,
ang = c(15, 25),
distlim = c(2500, 5000),
spdf = TRUE,
min.dt = 60,
emf = NULL
)input data, must have 5 (LS), or 8 (KF) columns (see details)
max travel rate (m/s)
angles of outlier location "spikes" (default is c(15,25) deg); ang = NA turns off trip::sda filter in favour of trip::speedfilter
lengths of outlier location "spikes" (default is c(2500, 5000) m); distlim = NA turns off trip::sda filter in favour of trip::speedfilter. Either ang = NA or distlim = NA are sufficient.
turn speed filter on/off (logical; default is TRUE)
minimum allowable time difference in s between observations; dt < min.dt will be ignored by the SSM
optionally supplied data.frame of error multiplication factors for Argos location quality classes. see Details
an sf object with all observations passed from data and the following appended columns
keeplogical indicating whether observation should be ignored by sfilter (FALSE)
obs.typeflag indicating whether KF or LS measurement model applies
emf_xerror multiplication factors for x direction
emf_yerror multiplication factors for y direction
geometrysf POINT object giving x,y coordinates in km
called by fit_ssm.