Computes the window for availability() for observers on a ship who
search a sector ahead of the bow. The window runs from the moment the animal
crosses the search radius to the moment its bearing passes the edge of the
search angle (abeam, for a full forward search).
Usage
view_window_vessel(radius, speed, angle = 90, distance = 0)
view_window_ship(radius, speed, angle = 90, distance = 0)Arguments
- radius
Numeric. Effective search radius ahead of the vessel, in metres. Must be positive.
- speed
Numeric. Vessel speed over ground, in metres per second. Must be positive.
- angle
Numeric. Search half-angle from the bow, in degrees. Default
90, a full beam-to-beam forward search. Must be above 0 and at most 90.- distance
Numeric. Perpendicular distance, in metres. Default
0, the trackline, which gives the widest window. Must not be negative.
Value
A numeric vector of windows in seconds, 0 where the animal is never
inside the searched sector. view_window_ship() is an alias.
Which platform this fits
A ship's observers search forward only. An animal behind the beam has
already been passed, and counting it would count it twice. This differs
from view_window(), which watches a full circle, and from
view_window_aerial(), whose window grows with perpendicular distance.
How it works
For a search radius \(r\), a search half-angle \(\theta\) measured from the bow, speed \(v\) and perpendicular distance \(x\):
$$w(x) = \frac{\sqrt{r^2 - x^2} - x/\tan\theta}{v}$$
The animal crosses the search radius when it is \(\sqrt{r^2 - x^2}\) ahead of the vessel (zero if \(x > r\)).
Its bearing passes the edge of the search angle when it is \(x/\tan\theta\) ahead. At
angle = 90this is set to exactly zero (abeam).The window is the track between the two, divided by speed, and
0when the second point is further ahead than the first.
At angle = 90 the window is exactly half of view_window()'s: the same
chord, with the aft half never searched.
Calibrating the inputs
radius is the effective search radius, the distance at which observers
actually detected animals, not the reach of the binoculars. A truncation
distance from the fitted detection function is a better starting point than
a nominal one. This window does not correct for animals moving toward or
away from the ship before detection; that is not an availability problem.
See g0() for how components combine.
References
Buckland, S.T., Anderson, D.R., Burnham, K.P., Laake, J.L., Borchers, D.L. and Thomas, L. (2001) Introduction to Distance Sampling. Oxford University Press. Shipboard search geometry, forward of the beam.
Laake, J.L., Calambokidis, J., Osmek, S.D. and Rugh, D.J. (1997) Probability of detecting harbor porpoise from aerial surveys: estimating g(0). The Journal of Wildlife Management 61:63-75. doi:10.2307/3802415 The availability formula this window feeds.
See also
availability() for what the window feeds, view_window_aerial()
for an aircraft, and view_window() for a circular area.
Examples
# A 6 km effective search radius at 10 knots (5.1 m/s), on the trackline
view_window_vessel(radius = 6000, speed = 5.1)
#> [1] 1176.471
# The window shrinks with perpendicular distance and is 0 beyond the radius
view_window_vessel(radius = 6000, speed = 5.1,
distance = c(0, 2000, 5000, 7000))
#> [1] 1176.4706 1109.1871 650.3186 0.0000
# A narrower search arc shortens the window more at larger distances
view_window_vessel(radius = 6000, speed = 5.1, angle = 45,
distance = c(0, 2000, 5000))
#> [1] 1176.4706 717.0302 0.0000
# A full forward search gives half the circular window
view_window_vessel(radius = 300, speed = 50) == view_window(300, 50) / 2
#> [1] TRUE