Read GLM netCDF output

Usage

read_glm_output(
  nc = NULL,
  vars_sim = NULL,
  depths = NULL,
  dates = NULL,
  date_index = NULL,
  incl_fluxes = TRUE,
  output_hour = 0,
  file,
  phyto_pars = NULL,
  load_all = TRUE,
  raw_output = FALSE,
  extra_vars = NULL,
  sed_layers = NULL,
  zone_summary = FALSE
)

Arguments

nc: An object of class ncdf4 (as returned by either function nc_open or function nc_create), indicating what file to read from.
vars_sim: Variables to extract in the AEME format e.g. "HYD_temp"
depths: Depths to extract. If NULL, extract all model layer depths. Defaults to NULL.
dates: Dates to extract. If NULL, extract all dates. Defaults to NULL.
date_index: Date index to extract. If NULL, extract all dates. Defaults to NULL.
incl_fluxes: Logical indicating whether to include flux variables. Defaults to TRUE.
output_hour: Hour of the day to extract (0-23). Defaults to 0.
file: File path to netCDF file. Only used if nc is NULL.
phyto_pars: Data frame with phytoplankton parameters from AED.
load_all: logical; also load every other variable present in the netCDF file, beyond the declared vars_sim set. Each such variable is keyed by its AEME var_aeme name if key_naming has a translation for it, or by its raw GLM/netCDF name otherwise. Variables shaped like GLM's usual (time) or (z, time) output are loaded the same way as any declared variable; variables with other dimensions (e.g. nzones, particle, sed_layers, lon, lat) are loaded as a new_grouped_var() object instead of being forced into the depth x time convention. Default TRUE.
raw_output: logical; if TRUE, return output as close to the raw netCDF file as possible instead of AEME's standardised format: (z, time) variables are left on GLM's own, time-varying model layer midpoints (no interpolation onto a common depth grid), variables requested via vars_sim are keyed by their raw GLM/netCDF name (e.g. "temp") rather than the translated AEME var_aeme name (e.g. "HYD_temp"), and no AED unit-conversion factors are applied. depths must not be supplied when raw_output = TRUE, since raw output has no common depth grid to interpolate onto. Default FALSE.
extra_vars: character; which of the variables swept up by load_all (those beyond the declared vars_sim set – sediment, diagnostic and flux output) to read. NULL (default) reads all of them. "none" reads none. Otherwise a vector of names – AEME names from glm_extra_vars (e.g. "SED_temp"), raw netCDF names (e.g. "sed_temp"), and/or the group keywords "sediment" and "diag" (every variable that would be stored in that sub-list). Variables not requested are never read from the file, which keeps ensemble output small. Names that are not in the file give a warning. Ignored if load_all = FALSE.
sed_layers: integer; for variables with a sed_layers dimension (sed_temp), keep only these sediment layers (indices, 1 = shallowest). NULL (default) keeps all.
zone_summary: logical; also add a lake-wide series for each requested sediment-zone variable (name ending _Z), the mean over zones weighted by the time-varying zone area (zarea). Stored in $sediment as <name>_lake (e.g. SDF_Fsed_oxy_Z_lake, or SED_Fsed_oxy_lake by AEME name), in the variable's own per-m2 units. Zones with no area (NA in zarea) are ignored. Naming a _lake series in extra_vars builds just that series, without zone_summary = TRUE (the zone array behind it is then not kept unless also requested). Default FALSE.

Value

List with AEME output variables. Also includes z, GLM's own raw layer-height matrix (height of each layer's top boundary above the lake bottom, per timestep, before conversion to LKE_depths) – unlike GOTM-WET/Simstrat-AED2, GLM's layer structure genuinely changes over time, so this is kept alongside the derived depth grid rather than discarded after use.