Skip to contents

Flag overly-gapfilled data

Source: R/flux_qc.R
flux_qc.Rd

Flags rows of data based on variables with associated _QC columns.

Usage

flux_qc(
  data,
  qc_vars,
  threshold = 0.5,
  operator = c("any", "all"),
  max_gapfilled = deprecated()
)

Arguments

data

A data frame created by flux_read().

qc_vars

A character vector of column names with associated *_QC columns to use for flagging.

threshold

For aggregated data (weekly, daily, monthly, annual), a number between 0 and 1. For hourly and half-hourly data, either 0, 1, 2, or 3 corresponding to 0 = measured, 1 = good quality gap-fill (MDS), 2 = medium quality gap-fill, 3 = poor quality gap-fill. Rows with _QC values greater than the number provided will be be flagged with qc_flagged = TRUE. Input can be length 1 or the same length as qc_vars to supply a different threshold for each variable.

operator

How to flag data when multiple qc_vars are supplied? If "any", the row will be marked as bad if any of the QC vars indicate gap-filling above their max_gapfill threshold. If "all" then the row will be flagged only if all of the QC vars are above their max_gapfill.

max_gapfilled

[Deprecated] Deprecated; use threshold instead.

Value

A tibble with the added columns p_gapfilled and qc_flagged. If operator = "any", qc_flagged = TRUE indicates that at least one of the supplied QC variables was more gapfilled than threshold and p_gapfilled will be the maximum proportion gapfilled across the QC vars for each row. If operator = "all", then qc_flagged = TRUE indicates that all of the supplied QC variables were more gapfilled than the thresholds supplies and p_gapfilled will be the minimum proportion gapfilled across all QC variables for each row.

Note

Most stations use Variable Ustar Threshold (VUT) method for NEE, but some use Constant Ustar Threshold (CUT), so when using NEE variables for QC flagging, it is a good idea to include both the VUT and CUT variant. E.g. qc_vars = c("NEE_VUT_REF", "NEE_CUT_REF").

Examples

if (FALSE) { # \dontrun{

# Flag rows where NEE_VUT_REF is more than 50% gapfilled
manifest <- flux_discover_files()
annual <- flux_read(manifest, resolution = "y")
annual_flagged <- flux_qc(
  annual,
  qc_vars = "NEE_VUT_REF",
  threshold = 0.5
)

# Use multiple variables each with a different threshold for QC
annual_flagged2 <- flux_qc(
  annual,
  qc_vars = c("NEE_VUT_REF", "TA_F"),
  threshold = c(0.4, 0.6)
)

# Same as above, but require *both* variables to be above their thresholds
# to consider that row a problem
annual_flagged2 <- flux_qc(
  annual,
  qc_vars = c("NEE_VUT_REF", "TA_F"),
  threshold = c(0.4, 0.6),
  operator = "all"
)

} # }