Discrete wavelet transform coupled with the active subspace method

The provided Python code represents the coupled framework between the discrete wavelet transform and the active subspace method. It has the goal to perform temporal scale dependent model parameter sensitivity analysis. In the provided case, the methodology is coupled to an R code containing the LuKARS model.

The folder named 'as_dwt' contains the entire source code of the methodology as well as the required data
of the Kerschbaum spring case study.

The subfolder uq_tools contains supplementary python scripts that can be used for analyses that go beyond
the methodology proposed in the WRR article.

The subfolder examples contains a folder called 'as_wavelets', in which the relevant python scripts are stored.

The data and the LuKARS model (R. file) can be found from this directory in 'scens/scen_main'.

The LuKARS model is given by the file 'main_exe.R.'

The precipitation and discharge data is stored in 'kerschbaum.txt'.

The monthly mean temperatures (needed for Thornthwaite's ET method) are stored in 'monthly_mean_temp.csv'.

The daily temperature values and snow depths are stored in 'snow_waidhofen.csv'.