prefilter: Prepare Argos data for fitting state-space model

In ianjonsen/foieGras: Fit Continuous-Time State-Space and Latent Variable Models for Quality Control of Argos Satellite (and Other) Telemetry Data and for Estimating Changes in Animal Movement

Prepare Argos data for fitting state-space model

Description

prefilter does the following: (1) flags (keep = FALSE) observations with duplicate dates & removes subsequent observations occurring within min.dt s of one another (2) determines Argos data type (LS, KF, G, or GL); (3) uses sda filter to identify extreme locations. sda is a fast, vectorized version of the now CRAN-archived argosfilter::sdafilter. sda is a native implementation of that found in the currently archived trip package (by MD Sumner: https://github.com/Trackage/trip). (4) projects lonlat coords to mercator x,y coords (in km) & shifts longitudes that straddle -180,180 to 0,360 and vice-versa; (5) adds location error multiplication factors based on Argos location class (for type LS);

Usage

prefilter(
  x,
  vmax = 5,
  ang = c(15, 25),
  distlim = c(2500, 5000),
  spdf = TRUE,
  min.dt = 0,
  emf = NULL
)

Arguments

x	input data, must have 5 (LS), or 8 (KF) columns (see details)
vmax	max travel rate (m/s)
ang	angles of outlier location "spikes" (default is c(15,25) deg); ang = NA turns off sda filter in favour of speedfilter
distlim	lengths of outlier location "spikes" in km (default is c(2.5, 5) m); distlim = NA turns off sda filter in favour of speedfilter. Either ang = NA or distlim = NA are sufficient.
spdf	turn speed filter on/off (logical; default is TRUE)
min.dt	minimum allowable time difference in s between observations; dt < min.dt will be ignored by the SSM. Default is NA: all time differences > 0 are allowed
emf	optionally supplied data.frame of error multiplication factors for Argos location quality classes. see Details

Details

called by fit_ssm.

Value

an sf object with all observations passed from data and the following appended columns

keep	logical indicating whether observation should be ignored by sfilter (FALSE)
obs.type	flag indicating whether KF or LS measurement model applies
emf_x	error multiplication factors for x direction
emf_y	error multiplication factors for y direction
geometry	sf POINT object giving x,y coordinates in km

ianjonsen/foieGras documentation built on Jan. 17, 2025, 11:15 p.m.