Skip to contents

Evaluate the quality of the simulation result

Source: R/core.R, R/hydrobudget-quality.R
evaluate_simulation_quality.Rd

From a simulation result, evaluate the quality by comparing with observations. The quality measurement can be used for model calibration (e.g. caRamel package) or sensitivity evaluation (e.g. sensitivity package).

Evaluates the simulated water budget with the average KGE.

Usage

evaluate_simulation_quality(obj, water_budget, ...)

# S3 method for default
evaluate_simulation_quality(obj, water_budget, ...)

# S3 method for hydrobudget
evaluate_simulation_quality(
  obj,
  water_budget,
  rcn_gauging,
  observed_flow,
  alpha_lyne_hollick,
  period = NULL,
  ...
)

Arguments

obj

The HydroBudget object with calibration parameters and column names mappings.

water_budget

The computed water budget. Input can be a data.frame/data.table or a path to a data file.

...

Other arguments passed to methods

rcn_gauging

The table with the list of RCN cells located in each gauging station watershed. Input can be a data.frame/data.table or a path to a data file.

observed_flow

The flow rates in mm/day. Input can be a data.frame/data.table or a path to a data file.

alpha_lyne_hollick

The Lyne and Hollick filter. Input can be a data.frame/data.table or a path to a data file.

period

The start and end years. If not provided, the start/end years will be extracted from the water budget data.

Value

The model-specific quality assessment.

The HydroBudget quality assessment.

Details

The columns of the water budget data set input are:

  • year

  • month

  • vi

  • t_mean

  • runoff

  • pet

  • aet

  • gwr

  • runoff_2

  • delta_reservoir

  • rcn_id

The columns of the observed flow data set input are:

  • year

  • month

  • day

  • one column per station (named by the station ID), the flow rates in mm/day

The columns of the RCN gauging stations data set input are:

  • rcn_id, the cell ID

  • station_id, the station ID

The columns of the Lyne and Hollick filter data set input are:

  • station_id, the station ID

  • alpha

Examples

if (FALSE) {
# Use input example files provided by the package
base_url <- "https://github.com/gwrecharge/rechaRge-book/raw/main/examples/input/"
input_rcn_gauging <- paste0(base_url, "rcn_gauging.csv.gz")
input_observed_flow <- paste0(base_url, "observed_flow.csv.gz")
input_alpha_lyne_hollick <- paste0(base_url, "alpha_lyne_hollick.csv.gz")

# Calibration parameters
HB <- rechaRge::new_hydrobudget(
  T_m = 2.1, # melting temperature (°C)
  C_m = 6.2, # melting coefficient (mm/°C/d)
  TT_F = -17.6, # Threshold temperature for soil frost (°C)
  F_T = 16.4, # Freezing time (d)
  t_API = 3.9, # Antecedent precipitation index time (d)
  f_runoff = 0.63, # Runoff factor (-)
  sw_m = 431, # Maximum soil water content (mm)
  f_inf = 0.07 # infiltration factor (-)
)

# ... compute the water budget ...

result <- evaluate_simulation_quality(
  HB,
  water_budget = water_budget,
  rcn_gauging = input_rcn_gauging,
  observed_flow = input_observed_flow,
  alpha_lyne_hollick = input_alpha_lyne_hollick,
  period = simul_period
)
}