calibrate_sim: Calibrates GLM-AED2 variables to improve fit between observed...

In USGS-R/glmtools: Tools for working with the General Lake Model (GLM)

Calibrates GLM-AED2 variables to improve fit between observed and simulated data

Description

Starts a calibration run with multiple iterations using a GLM-AED2 setup. At the end a output csv.-file of all iterations and a diagnostic plot are created.

Usage

calibrate_sim(
  var = "temp",
  path,
  field_file,
  nml_file = "glm3.nml",
  glm_file = "glm3.nml",
  calib_setup = NULL,
  glmcmd = NULL,
  first.attempt = TRUE,
  period = NULL,
  scaling = TRUE,
  verbose = TRUE,
  method = "CMA-ES",
  metric = "RMSE",
  target.fit = 1.5,
  target.iter = 100,
  plotting = TRUE,
  output,
  conversion.factor = 1
)

Arguments

var	Character vector of valid variable names (see sim_vars)
path	String with the path to the GLM setup
field_file	CSV or TSV field data (see resample_to_fieldfor format)
nml_file	String of the namelist file that will be calibrated, default is 'glm3.nml'
glm_file	String of the glm-namelist file, default is 'glm3.nml'
calib_setup	Data frame containing information regarding the calibration (see get_calib_setup)
glmcmd	String containing the desired glm run command, default is GLM3r
first.attempt	Boolean, if TRUE a nml-template will be created, set it to FALSE after your first calibration run; default is TRUE
period	List that provides a start and a stop date for the simulation
scaling	Boolean, if TRUE variable values will be scaled on the space (0,10), recommended for CMA-ES, default is TRUE
verbose	should operations and output of GLM be shown. Default is TRUE.
method	String of the optimization method, default is 'CMA-ES' (Hansen 2009), alternatively you can also use 'Nelder-Mead'
metric	String of the calibration fit metric, default is RMSE
target.fit	Double of your preferred fit, calibration will stop after reaching that; default is 1.5
target.iter	Double of maximum amount of iterations, default is 150
plotting	Boolean, if TRUE plots all results as heat maps
output	Character array of the output folder path
conversion.factor	Double of the conversion factor between simulated and observed data, default is 1 for temp.

Author(s)

Robert Ladwig, Tadhg Moore

See Also

get_calib_setup, get_calib_periods, get_calib_init_validation

Examples

calib_setup <- get_calib_setup()
print(calib_setup)

#Example calibration
#Copy files into temporary directory
sim_folder <- tempdir() #simulation path
glmtools_folder = system.file('extdata', package = 'glmtools')

file.copy(list.files(glmtools_folder,full.names = TRUE), sim_folder, overwrite = TRUE)

field_file <- file.path(sim_folder, 'LakeMendota_field_data_hours.csv')
nml_file <- file.path(sim_folder, 'glm3.nml')
driver_file <- file.path(sim_folder, 'LakeMendota_NLDAS.csv')
period = get_calib_periods(nml_file = nml_file, ratio = 1)
output = file.path(sim_folder, 'output/output.nc')

var = 'temp' # variable to apply the calibration procedure
## Not run: 
calibrate_sim(var = var, path = sim_folder, field_file = field_file,
             nml_file = nml_file, calib_setup = calib_setup,
             glmcmd = NULL,
             first.attempt = TRUE, period = period, method = 'CMA-ES',
             scaling = TRUE, #scaling should be TRUE for CMA-ES
             verbose = FALSE,
             metric = 'RMSE',plotting = FALSE,
             target.fit = 1.5,
             target.iter = 50, output = output)

## End(Not run)

USGS-R/glmtools documentation built on March 26, 2024, 5:43 p.m.