CorrectDOY: Correct day-of-year time series
In greenbrown: Land Surface Phenology and Trend Analysis

Description Usage Arguments Value Author(s) See Also Examples

This function corrects a time series with days-of-years (e.g. start of growing season). For example, if the start of season occurs in one year at the end of the calendar year (doy > 305) and in another year at the beginning (doy < 60), the DOYs are corrected so that all values occur at the beginning of the year (e.g. negative DOYs will be produced) or at the end of the year (e.g. DOY > 365 will be produced). This function is applied in Phenology after phenology detection on sos, eos, pop and pot time series (see examples).

1	CorrectDOY(doy, check.outliers = TRUE)

`doy`	a vector or time series representing DOYs
`check.outliers`	Set outliers to NA after correction? Outliers are defined here as: `doy` < (median - IQR2) \| `doy` > (median + IQR2))

a vector or time series

Matthias Forkel <matthias.forkel@tu-dresden.de> [aut, cre]

Phenology

# imagine the following start of season DOYs in 10 years
sos <- ts(c(15, 10, 12, 8, 10, 3, 362, 2, 1, 365), start=1982) 
plot(sos)
# Visually, there seems to be big differences in the start of season. However, 
# there is actually only one day between the last two values (DOY 1 = 1st January, 
# DOY 365 = 31st December). Trend calculation fails on this time series:
plot(Trend(sos), ylab="SOS") 

# The DOY time series needs to be corrected to analyze 
# the true differences between days.
sos2 <- CorrectDOY(sos)
plot(Trend(sos2), ylab="SOS") 
# The correction now allows trend analysis.
# Negative DOYs indicate days at the end of the previous year!

# other example
sos <- ts(c(5, 12, 15, 120, 363, 3, 362, 365, 360, 358), start=1982) 
plot(sos) # one value seems like an outlier
sos2 <- CorrectDOY(sos)
plot(Trend(sos2), ylab="SOS") 
# The outlier is removed.
# DOYs > 365 indicate days in the next year!