Apply Threshold to Identify Events

Identifies and extracts events in a time series based on threshold criteria, with options for event selection and characterization.

Usage

apply_threshold(
  X,
  lim,
  where = "<=",
  what = "X",
  select = "all",
  Date = NULL,
  period = NULL
)

Arguments

X

Numeric vector of values to analyze (typically discharge or other hydro-meteorological variable)

lim

Numeric threshold value or vector of same length as X

where

String specifying threshold condition:

• "<", "<=", "==", ">=", ">" for standard comparisons

what

String specifying return type:

• "X": Values meeting condition

• "length": Duration of events (time steps)

• "first"/"last": First/last index of events

• Any function name (e.g., "which.max"): Applied to event values

select

String or numeric specifying event selection:

• "all": All events (default)

• "longest"/"shortest": Longest/shortest event

• Numeric value: Event containing this value

Date

Optional date vector for period filtering

period

Optional vector of 2 dates (start/end) for period filtering

Value

Depends on what parameter:

• Numeric vector of values ("X")

• Integer duration ("length")

• Single index ("first"/"last")

• Function application result

• NA if no events found

Details

Handles multiple threshold scenarios:

• For vector lim, uses most frequent value via rle()

• For period filtering, subsets data before analysis

• For event selection, identifies contiguous periods

Examples

set.seed(1)
Q <- rnorm(100, 50, 10) # Simulated discharge

# Get longest drought period (Q <= 45)
drought_days <- apply_threshold(Q, 45, where="<=", what="length", select="longest")

# Get first index of all flood events (Q >= 60)
flood_starts <- apply_threshold(Q, 60, where=">=", what="first")