Smooth a PhenoCam time series

Smooths time series iteratively using a Akaike information criterion (AIC) to find an optimal smoothing parameter and curve.

smooth_ts(
  data,
  metrics = c("gcc_mean", "gcc_50", "gcc_75", "gcc_90", "rcc_mean", "rcc_50", "rcc_75",
    "rcc_90"),
  force = TRUE,
  internal = TRUE,
  out_dir = tempdir()
)

Arguments

data

a PhenoCam data file or data structure

metrics

which metrics to process, normally all default ones

force

TRUE / FALSE, force reprocessing?

internal

return a data structure if given a file on disk (TRUE / FALSE = default)

out_dir

output directory where to store data

Value

An PhenoCam data structure or file with optimally smoothed time series objects added to the original file. Smoothing is required for `phenophase()` and `transition_dates()` functions.

Examples

if (FALSE) {
# with defaults, outputting a data frame
# with smoothed values, overwriting the original

# download demo data (do not smooth)
download_phenocam(site = "harvard$",
                  veg_type = "DB",
                  roi_id = "1000",
                  frequency = "3",
                  smooth = FALSE)

# smooth the downloaded file (and overwrite the original)
smooth_ts(file.path(tempdir(),"harvard_DB_1000_3day.csv"))

# the function also works on a PhenoCam data frame
df <- read_phenocam(file.path(tempdir(),"harvard_DB_1000_3day.csv"))
df <- smooth_ts(df)
}