ck_tx10p: Percentage of Cool Days (TX10p)

In climatekit: Unified Climate Indices for Temperature, Precipitation, and Drought

Percentage of Cool Days (TX10p)

Description

ETCCDI canonical index TX10p. Percentage of days where daily Tmax falls below the 10th percentile of the calendar-day distribution from a reference period (default 1961 to 1990). The threshold is computed using a 5-day window centred on each calendar day, pooled across the reference period.

Usage

ck_tx10p(
  tmax,
  dates,
  ref_start = 1961L,
  ref_end = 1990L,
  period = "annual",
  bootstrap = FALSE
)

Arguments

tmax	Numeric vector of daily maximum temperatures (degrees C).
dates	Date vector of the same length as tmax. Must contain data covering the reference period.
ref_start, ref_end	Integer. Reference period boundary years (inclusive). Defaults to 1961 and 1990.
period	Character. Aggregation period: "annual" (default) or "monthly".
bootstrap	Logical. If TRUE, apply the Zhang (2005) in-base bootstrap correction. Default FALSE for backward compatibility and speed.

Details

Set bootstrap = TRUE to apply the Zhang et al. (2005) in-base leave-one-out bootstrap, which removes the self-inclusion bias for years inside the reference period. The bootstrap is computationally expensive (roughly N^2 percentile fits for an N-year reference) but is the canonical climdex.pcic / climpact behaviour and is required for climate-change attribution work that spans the base.

Value

A data frame with columns period, value, index, and unit.

References

Zhang, X., Hegerl, G. C., Zwiers, F. W., & Kenyon, J. (2005). Avoiding inhomogeneity in percentile-based indices of temperature extremes. Journal of Climate, 18(11), 1641-1651. \Sexpr[results=rd]{tools:::Rd_expr_doi("10.1175/JCLI3366.1")}.

Examples

set.seed(1)
dates <- seq(as.Date("1961-01-01"), as.Date("1991-12-31"), by = "day")
tmax <- 15 + 10 * sin(2 * pi * as.integer(format(dates, "%j")) / 365) +
        rnorm(length(dates))
tail(ck_tx10p(tmax, dates))

climatekit documentation built on May 9, 2026, 5:08 p.m.