Skip to contents

Calculate grain-size percentiles

Source: R/percentiles.R
gs_d_values.Rd

gs_d_values() estimates D_p, the grain size at which p percent of a sample is finer. Interpolation is based on finite class boundaries from gs_cumulative(), not class midpoints.

Usage

gs_d_values(
  x,
  probs = c(5, 10, 16, 25, 30, 50, 60, 75, 84, 90, 95),
  interpolation_scale = c("phi", "log_um", "linear_um"),
  output_unit = c("um", "mm", "phi"),
  extrapolate = c("error", "warn_linear"),
  scale = NULL
)

Arguments

x

A valid gsd_tbl object.

probs

Numeric vector of percentiles on the 0-100 scale.

interpolation_scale

Interpolation scale. "phi" interpolates in phi units, "log_um" interpolates in log10 micrometers, and "linear_um" interpolates directly in micrometers.

output_unit

Preferred reporting unit. The returned table always includes micrometer, millimeter, and phi columns.

extrapolate

Behavior when a requested percentile falls outside the observed finite boundary curve. "error" throws an error, and "warn_linear" warns and linearly extrapolates on the selected scale.

scale

Compatibility alias for interpolation_scale.

Value

A tibble with one row per sample and requested percentile.

Details

Some samples contain a run of consecutive classes with zero retained mass (e.g. several sieve apertures with nothing caught between them), which produces an exact tie in cumulative percent finer across those boundaries. When a requested percentile falls between such a tied plateau and an adjacent distinct value, gs_d_values() resolves the tie deterministically: it brackets against the member of the tied plateau nearest the real transition (the finest boundary of a plateau being approached from below, or the coarsest boundary of a plateau being approached from above), rather than depending on incidental input row order. This is a fixed, documented rule, not an implementation detail that may vary between calls or package versions.