RaDisMod

Description

RaDisMod is an R package developed by Guillaume Cinkus and Vincent Marc at Avignon Université. It contains one function that launches an R shiny application for modelling single flood events using production and transfer functions.

Installation

RaDisMod requires an installation of R 4.0.0. The instruction for the installation and the download of R can be found on the CRAN website.

Once R is installed, RaDisMod can be installed from GitHub.

if (!require("remotes")) install.packages("remotes") # install remotes package if needed
remotes::install_github("busemorose/RaDisMod") # install RaDisMod package

Launch

Once the package is installed, the application can be loaded with the RaDisMod() function.

library(RaDisMod)
RaDisMod()

RaDisMod::RaDisMod()

Data import

The data import field allows to load a file with the following conditions:

The data must be a plain text file
The file must have two columns representing precipitation and observed discharge, respectively named P and obs
The columns must be semicolon-separated and of the same length
The unit of observed discharge (obs) should be in [L T^-1]
Missing precipitation (P) values should be filled with 0
Decimal separator is a dot .

Type of unit hydrogram

Four unit hydrogram are available by default:

Socose
Gamma
Log-normal
Triangle

It is possible to import a custom unit hydrogram using Custom in the drop-down menu. The file for a custom unit hydrogram must be a unique column without header.

The length of the unit hydrogram to consider can be specified with Length of HU.

Production function

SCS, with $P_{cumul}$ the cumulated precipitation and $J$ a parameter to calibrate.

if $P_{cumul} < 0.2 * J$ then $P_{n-cumul} = 0$

else $P_{n-cumul} = \frac{(P_{cumul} - 0.2 * J)^2}{P_{cumul} + 0.8 * J}$

Then, to obtain the net precipitation $P_n$:

PN <- diff(c(0, PN_cum))

Model calibration

Two options:

Manual calibration using the three first sliders (J and two other parameters depending on the unit hydrogram)
Automatic calibration in the "Parameter optimisation" section using the Run button at the bottom. It will randomly sample parameters within the specified ranges a number of times equal to the Iterations number. The individual simulations are then evaluated with the mentioned performance criteria, either NSE, KGE, KGENP and RMSE.

The evaluation range can be specified using the appropriate slider.