Supplementary Material for "Equifinality, sloppiness, and emergent structures of soil biogeochemical models" by Marschmann et al. (2018)
-
to reproduce Fig.3 run PerfectDataSpec.py in Subfolders of Universally_Sloppy. Then call plot_spectra.py
-
for Fig.2 run the following scripts in Monod_Model: monod_model.py, generate_ensemble_predictions.py, generate_geodesics.py. Then plot with plot_timeseries.py, plot_landscape.py, plot_manifold_PCA.py
-
for Fig.4 run PlotSpectrum.py in the respective subfolders Functional_Gene_Data, Bulk_Data and Input_Output_Data. In order to check every iteration in the model reduction, run modelN_fit.py and modelN_reduce.py in the corresponding folder ModelN
-
for Fig. 5 run Performance/plot_calibrations.py. For model predictions, message me for the data files or uncomment lines in Functional_Gene_Data(MCPA_Data)/Model0/model0_fit.py and Functional_Gene_Data(MCPA_Data)/Model32/model32_fit.py to run the MCMC optimization (approx. 2 days)
-
for Fig. 6, the MATLAB R2017a Toolbox SAFE R1.1 is needed for Global Sensitivity Analysis
-
the R code for identifiability analysis of the minimal soil C model and corresponding MBAM reduction is in folder Universally_Sloppy/German