Assign equinox periods
assign_equinox_periods.RdAssign which days of the year estimated latitudes are unreliable due to equinox. Limits are based on
Which approximate latitude did the geolocator record light-data (objective; towards the poles changes in daylength is more dramatic around equinox than at equator = shorter equinox period).
how much uncertainty allowed (subjective) : uncertainty gets extreme right up to the equinox and less further away in time. ut off point of the equinox period aim to keep an average day-to-day variance in latitude below 1.5 degrees (165 km).
Results from sun angle calibrations (objective) which is a result of a) threshold (light-level) used to differentiate between night and daytime and b) the sensor's light sensitivity.
Arguments
- lats
Numeric value with latitudes from the tracked individual.
- dates
Dates (tfirst) when latitudes occur
- breedingloc_lat
Approximate latitude for the breeding locality.
- sun
Numeric value showing the sun's angle to the horizon at each twilight. Is a result based on the chosen threshold of light used in twilight estimations and the light sensor sensitivity
Value
A list saying if the equinox effect is present or not (1 = no equinox effect, 0 = equinox effect).
Details
Point 1-3 are compared to a reference table:
Apparent equinox (period midpoint) is defined from how sun angles compare to calculations in Hill & Braun (2001). Period start and end is based on estimations of variance in Hill & Braun (2001) and the mean day-to-day variance in latitudes observed in 16 000 tracked non-breeding seasons in SEATRACK.
The function performs the following steps:
Calculate median sun angle for a 5-day period at yday 41:45 (>30 days before spring EQ), yday 113:117 (>30 days after spring EQ), yday 227:231 (>30 days before autumn EQ), yday 300:304 (>30 days after autumn EQ).
If no sun angles available - use first sun angle
Calculate median latitudes for a 5-day period at yday 41:45 (>30 days before spring EQ), yday 113:117 (>30 days after spring EQ), yday 227:231 (>30 days before autumn EQ), yday 300:304 (>30 days after autumn EQ).
If no latitudes because of midnight sun - use breeding location
Matches e.g yday 41:45 median sun angle and median latitude to the reference table within the R package to identify the start of spring equinox period an so on.