Skip to contents

Collapse a sub-daily meteo data frame to a daily time step

Source: R/calc_water_balance.R
collapse_met_daily.Rd

Aggregates a POSIXct-stamped (or otherwise sub-daily) meteorological data frame to one row per calendar day: every numeric column is averaged over the day, except MET_pprain / MET_ppsnow, whose aggregation is controlled by precip. Non-numeric columns other than Date are dropped, and Date is returned as a Date.

Usage

collapse_met_daily(obs_met, precip = c("mean", "sum"))

Arguments

obs_met

data frame; meteo forcing with a Date column (Date or POSIXct) and numeric MET_* columns.

precip

character; how to aggregate MET_pprain / MET_ppsnow over the day. "mean" (default) when they are already a mm/day rate; "sum" when they are per-time-step accumulations.

Value

A data frame with one row per day: Date (as Date) and the day-aggregated numeric columns.

Details

AEME defines MET_pprain / MET_ppsnow as a rate in mm/day (see standardise_met()). If the frame is already on that convention – e.g. the output of standardise_met(), which is what build_aeme() feeds the water balance – a daily mean of the rate is the daily rate, so precip = "mean" (the default) is correct. Raw sub-daily reanalysis / AWS products instead report precipitation as an accumulation per time step (mm that fell during that step); for those, use precip = "sum" to get the daily total.

See also

standardise_met()

Examples

hourly <- data.frame(
  Date = seq(as.POSIXct("2020-01-01", tz = "UTC"), by = "hour",
             length.out = 48),
  MET_tmpair = rnorm(48, 15, 3),
  MET_pprain = c(rep(0, 20), rep(0.5, 4), rep(0, 24))  # mm per hour
)
collapse_met_daily(hourly, precip = "sum")
#> # A tibble: 2 × 3
#>   Date       MET_tmpair MET_pprain
#>   <date>          <dbl>      <dbl>
#> 1 2020-01-01       15.7          2
#> 2 2020-01-02       15.2          0