This repository contains the data and scripts associated with "Ten strategies towards successful calibration of environmental models" by Mai (2023).

Figures 1, 5, and 11 are flowcharts and schematic overviews. They are derived using LaTeX. It requires LaTeX to be installed including several packages such as TikZ.

Figures 2-4, and 6-10 are produced through calibration experiments. The setups and results used for the publication are provided as scripts/example_XX.zip. Please unzip them and run the experiments again in case you want to test. Be aware that some examples require an executable of the hydrologic modeling framework Raven to be available. Please download a Raven executable from here or compile the code yourself. Then replace all Raven executables in the example folders with your executable.

However, all outputs of the examples 2-4 and 6-10 have been saved and stored in a way that the figures can be created without running the experiments, i.e. no need to run Raven. These pre-derived outputs were used in the publication and are called scripts/figure_XX.json where XX denotes the figure that will be produced using these data.

The codes to produce these figures are in Python. The Python environment can be build using requirements.txt or setup a Python environment manually using:

pyenv virtualenv 3.8.5 env-3.8.5-calib-pub
pyenv activate env-3.8.5-calib-pub

pip install argparse
pip install numpy
pip install scipy
pip install matplotlib
pip install lhsmdu

The script to produce the figures is plot.sh. Please select the figure you want to produce in the first lines of that scripts by setting dofigXX=1 to 1. If dofigXX=0 the figure will not be produced. The figues will be placed in figures/ after creation.

Mai, J (2023)
Ten strategies towards successful calibration of environmental models.
Journal of Hydrology, 620(A), 129414.
https://doi.org/10.1016/j.jhydrol.2023.129414

Mai, J (2023)
Ten strategies towards successful calibration of environmental models.
Zenodo