Calibration of the parameters of one catchment or a network of sub-catchments

Calibration algorithm that optimizes the error criterion selected as objective function using the provided functions.

Usage

# S3 method for class 'GRiwrmInputsModel'
Calibration(
  InputsModel,
  RunOptions,
  InputsCrit,
  CalibOptions,
  useUpstreamQsim = TRUE,
  forceReservoirObs = TRUE,
  ...
)

# S3 method for class 'InputsModel'
Calibration(InputsModel, RunOptions, CalibOptions, ...)

Calibration(InputsModel, ...)

Arguments

InputsModel

[object of class InputsModel or GRiwrmInputsModel] see CreateInputsModel

RunOptions

[object of class RunOptions or GRiwrmRunOptions] see CreateRunOptions

InputsCrit

[object of class InputsCrit or GRiwrmInputsCrit] see CreateInputsCrit

CalibOptions

[object of class CalibOptions or GRiwrmCalibOptions] see CreateCalibOptions for details

useUpstreamQsim

boolean describing if simulated (TRUE) or observed (FALSE) flows are used for calibration. Default is TRUE

forceReservoirObs

boolean indicating if reservoir observed flows must be forced during the model run. Default is TRUE (See details of RunModel_Reservoir)

...

further arguments passed to airGR::Calibration, see details

Value

Depending on the class of InputsModel argument (respectively InputsModel or GRiwrmInputsModel object), the returned value is respectively:

• an OutputsCalib object (See airGR::Calibration for more details on this object)

• a GRiwrmOutputsCalib object which is a list of OutputsCalib objects with one item per modeled sub-catchment

Details

This function can be used either for a catchment (with an InputsModel object), for a network (with a GRiwrmInputsModel object), or for an ungauged node cluster (with a Ungauged object).

Argument classes should be consistent to the usage:

• a InputsModel argument of class InputsModel must be followed by a RunOptions argument of class RunOptions, an InputsCrit argument of class InputsCrit and a CalibOptions of class CalibOptions

• an InputsModel argument of class GRiwrmInputsModel must be followed by a RunOptions argument of class GRiwrmRunOptions, an InputsCrit argument of class GRiwrmInputsCrit and a CalibOptions of class GRiwrmCalibOptions

See also

CreateGRiwrm(), CreateInputsModel.GRiwrm(), CreateInputsCrit(), CreateCalibOptions()

Examples

# Loading catchment data
data(Severn)

#############################################################################
# EXAMPLE 1 - Basic calibration workflow                                   #
#############################################################################

# Creating catchment network
nodes <- Severn$BasinsInfo[, c("gauge_id", "downstream_id",
                               "distance_downstream", "area")]
nodes$model <- "RunModel_GR4J"
rename_columns <- list(id = "gauge_id",
                       down = "downstream_id",
                       length = "distance_downstream")
griwrm <- CreateGRiwrm(nodes, rename_columns)
griwrm
#>      id  down length    area         model donor
#> 4 54095 54001     42 3722.68 RunModel_GR4J 54095
#> 5 54002 54057     43 2207.95 RunModel_GR4J 54002
#> 6 54029 54032     32 1483.65 RunModel_GR4J 54029
#> 3 54001 54032     45 4329.90 RunModel_GR4J 54001
#> 2 54032 54057     15 6864.88 RunModel_GR4J 54032
#> 1 54057  <NA>     NA 9885.46 RunModel_GR4J 54057

# Preparation of InputsModel object
BasinsObs <- Severn$BasinsObs
DatesR <- BasinsObs[[1]]$DatesR
PrecipTot <- cbind(sapply(BasinsObs, function(x) {x$precipitation}))
PotEvapTot <- cbind(sapply(BasinsObs, function(x) {x$peti}))
Qobs <- cbind(sapply(BasinsObs, function(x) {x$discharge_spec}))
Precip <- ConvertMeteoSD(griwrm, PrecipTot)
PotEvap <- ConvertMeteoSD(griwrm, PotEvapTot)
InputsModel <- CreateInputsModel(griwrm, DatesR, Precip, PotEvap)
#> CreateInputsModel.GRiwrm: Processing sub-basin 54095...
#> CreateInputsModel.GRiwrm: Processing sub-basin 54002...
#> CreateInputsModel.GRiwrm: Processing sub-basin 54029...
#> CreateInputsModel.GRiwrm: Processing sub-basin 54001...
#> CreateInputsModel.GRiwrm: Processing sub-basin 54032...
#> CreateInputsModel.GRiwrm: Processing sub-basin 54057...

# Calibration period selection
# Set aside warm-up period and use the rest for calibration
IndPeriod_Run <- seq(
  which(InputsModel[[1]]$DatesR ==
        (InputsModel[[1]]$DatesR[1] + 365 * 24 * 60 * 60)),
  length(InputsModel[[1]]$DatesR)
)
IndPeriod_WarmUp <- seq(1, IndPeriod_Run[1] - 1)

# Preparation of RunOptions object
RunOptions <- CreateRunOptions(
  InputsModel,
  IndPeriod_WarmUp = IndPeriod_WarmUp,
  IndPeriod_Run = IndPeriod_Run
)

# Calibration criterion: preparation of the InputsCrit object
InputsCrit <- CreateInputsCrit(
  InputsModel = InputsModel,
  FUN_CRIT = ErrorCrit_KGE2,
  RunOptions = RunOptions,
  Obs = Qobs[IndPeriod_Run, ],
  transfo = "sqrt"
)

# Preparation of CalibOptions object
CalibOptions <- CreateCalibOptions(InputsModel)

if (interactive()) { # Too long for CRAN check...
  # Calibration
  OutputsCalib <- suppressWarnings(
    Calibration(InputsModel, RunOptions, InputsCrit, CalibOptions)
  )

  # Simulation
  OutputsModels <- RunModel(
    InputsModel,
    RunOptions = RunOptions,
    Param = extractParam(OutputsCalib)
  )
}