Skip to contents

Restrict a model's written output to the variables of interest

Source: R/set_output_vars.R
set_output_vars.Rd

By default every AEME model writes its full state at every output step. When only a few variables are needed - as in calibration or sensitivity analysis, where the objective is computed from one or two variables - the rest is wasted disk I/O (and, for Simstrat and GOTM, wasted files). set_output_vars() rewrites the output section of a model's configuration so that only vars, plus the handful of internals AEME always needs to read a result back (water level, the depth grid, temperature), are written.

Usage

set_output_vars(aeme, model, vars, mass_balance = TRUE, ens_n = 1)

Arguments

aeme

An Aeme object carrying a configuration for model.

model

Character. A single model, one of "glm_aed", "gotm_wet", "simstrat_aed", "simstrat_aed2", "dy_cd".

vars

Character. AEME variable names to keep, e.g. c("HYD_temp", "CHM_oxy"). Mapped to each model's own output names via key_naming; names with no mapping for model are dropped with a warning.

mass_balance

Logical. For "glm_aed" only: keep the GLMv4 &mass_balance diagnostic CSV? Default TRUE. Ignored for other models.

ens_n

Integer. Ensemble member whose configuration slot is updated. Default 1.

Value

aeme, with its configuration() updated.

Details

The change is made on the in-memory configuration; call write_configuration (or build_aeme) to write it to disk. build_aeme's output_vars argument applies this automatically at build time.

  • GLM always writes the full netCDF - its gridded variables cannot be sub-selected - so this only drops the &output csv_point_* keys (disabling the fixed-depth WQ_*.csv files) and, when mass_balance = FALSE, the whole &mass_balance block (disabling mass_balance.csv). The whole-lake lake.csv (csv_lake_fname) is left on: GLM 4.x only writes the netCDF diagnostic scalars (lake_level, ...) while that CSV is open, and AEME needs lake_level to read a GLM result back.

  • Simstrat writes one *_out.dat per variable; this sets Output$All = FALSE and lists only the needed variables.

  • GOTM replaces the /* (all-variables) output source with an explicit list.

  • DYRESM has a fixed output form and is returned unchanged.

See also

set_vars_sim, write_configuration, get_output_vars

Examples

aeme <- readRDS(system.file("extdata/aeme.rds", package = "AEME"))
path <- tempdir()
aeme <- build_aeme(path = path, aeme = aeme, model = "glm_aed",
                   model_controls = get_model_controls(), ext_elev = 5)
#> Warning: ! `SIL_rsi`: SIL_rsi is constant across all rows -- this may be a placeholder
#>   value.
#> ℹ Check raw data or unit conversion for this variable.
#> 
#> ── Calculating water balance ──
#> 
#> Resolving water level
#>   ℹ Using observed water level
#> ! Missing values in observed water level
#> ℹ Estimating surface water temperature
#> ✔ Estimating surface water temperature [26ms]
#> 
#> Estimating lake water levels for glm_aed
#>   ℹ Optimizing parameters for water balance
#>   ✔ Optimization Complete: C = 0.3343, h_inv = 23.4915, Final RMSE = 0.1431
#> ℹ Correcting water balance using estimated outflows (method = 2).
#> 
#> ── Building GLM-AED for lake wainamu ──
#> 
#> ✔ GLM nml validation completed - no issues detected.
aeme <- set_output_vars(aeme, "glm_aed", "HYD_temp", mass_balance = FALSE)
write_configuration(aeme, model = "glm_aed", path = path)