Identify start and end of growing seasons for each year

Usage

identify_growing_seasons(
  dat,
  ...,
  full_season = TRUE,
  trend_threshold = 4,
  superchill_threshold = -0.7,
  max_season = 540
)

Arguments

dat: Data frame with at least three columns: TIMESTAMP (must be possible to convert to POSIXct), DEPTH, and VALUE. If column VARIABLE is included, it must have one unique entry. May also include columns with grouping variables passed to .... Other columns will be ignored.
...: Additional columns in dat to use as grouping variables. Results are automatically grouped by DEPTH.
full_season: Logical argument. The default, full_season = TRUE will return NA for START_SEASON and END_SEASON that occur outside of the date range of the data series. For example, if the data series starts when temperature is already above trend_threshold, START_SEASON = NA; if the season length is determined by max_season, which results in a date after the last observation, END_SEASON = NA. If full_season = FALSE, the START_SEASON and END_SEASON dates default to the first and last timestamp in ... and DEPTH.
trend_threshold: The threshold for "trending up". Default is trend_threshold = 4. The last observation above trend_threshold that does not return below trend_threshold triggers the beginning of the growing season for each DEPTH and group in ....
superchill_threshold: The threshold for "superchill". Default is superchill_threshold = -0.7. The first observation below superchill_threshold triggers the end of the growing season for each group in DEPTH and ....
max_season: The length of the growing season in months for groups that have no temperature observations below superchill_threshold. Default is max_season = 540 days (~18 months). Note: units are days because adding 18 months to August 30 or August 31 results in NA (because February 30 and February 31 are not real dates).

Value

Returns a tibble with the START_SEASON, END_SEASON, SEASON_DAYS (length of season in days), and SEASON_MONTHS

(season length in months; assumes 1 month = 30 days) for each group in DEPTH and group in ....

Details

The growing season starts when the temperature crosses trend_threshold and does not return below trend_threshold (e.g., 4-degrees trending up) and ends 1 minute before the first observation of superchill_threshold.

The function looks for START_SEASON in January to August to avoid temperature drops caused by hurricanes (typically in September and October) and Winter cooling (November, December).

The function looks for END_SEASON in January to May of the next year (spring superchill).

If the time series begins when the temperature is above trend_threshold, START_SEASON is assigned the minimum TIMESTAMP for the group in filter_growing_season().

If the time series does not go below superchill_threshold, the growing season is assumed to be max_season months long. END_SEASON is assigned the timestamp START_SEASON + days(max_season).

Growing seasons may overlap if max_season is used to determine END_SEASON.