est_ustar_scens: =============================================================================...

In grahamstewart12/tidyflux:

Usage

est_ustar_scens(temp, ctrl_est = control_ustar(),
  ctrl_sub = subset_ustar(), ustar = "ustar", NEE = "NEE",
  Tair = "Tair", Rg = "Rg", ..., sf = create_sf(data$timestamp, type
  = "month"), reps = 200L, probs = c(0.05, 0.5, 0.95),
  verbose = TRUE)

Arguments

ctrl_est	Control parameters for estimating ustar on a single binned series, see control_ustar.
ctrl_sub	Control parameters for subsetting time series (number of temperature and ustar classes), see subset_ustar.
ustar	Column name for ustar.
NEE	Column name for NEE.
Tair	Column name for air temperature.
Rg	Column name for solar radiation.
...	Further arguments to est_ustar_thr.
sf	Factor of seasons to split (data is resampled only within the seasons).
reps	Number of repetitions in the bootstrap.
probs	Quantiles of the bootstrap sample to return. Default is the 5 median and 95 \item verboseSet to FALSE to omit printing progress.

A data frame with columns agg_mod, year, and ustar estimate based on the non-resampled data. The other columns correspond to the quantiles of ustar estimate for given probabilities (argument probs) based on the distribution of estimates using resampled the data. Original name: sEddyProc_sEstimateUstarScenarios The choice of the criterion for sufficiently turbulent conditions (ustar > chosen threshold) introduces large uncertainties in calculations based on gap-filled Eddy data. Hence, it is good practice to compare derived quantities based on gap-filled data using a range of ustar threshold estimates.

This method explores the probability density of the threshold by repeating its estimation on a bootstrapped sample. By default it returns the 90 confidence interval (argument probs). For larger intervals the sample number must be increased (argument probs).

Quality Assurance

If more than ctrl_est$min_boot (default 40 report a threshold, no quantiles (i.e. NA) are reported.

TW

grahamstewart12/tidyflux documentation built on June 4, 2019, 7:44 a.m.