This repository accompanies the paper "Simulation-based comparison of multivariate ensemble post-processing methods" (Lerch et al., 2020, https://doi.org/10.5194/npg-2019-62) and provides R code with implementations of all simulation settings as well as code to reproduce the results presented in the paper and the supplemental material.

To run simulations and reproduce results from the paper and supplemental material, proceed as follows:

• .Rdata files for individual settings are generated by run_settingX.R in the folder /simulation code/, with X replaced by the setting of interest
• source code for the generation of ensemble forecasts and observations, as well as univariate and multivariate post-processing and evaluation can be found in the folder /simulation code/source/.
• The output files can be processed with code in the folder /processing code for simulation output/. The R scripts corresponding to the setting of interest will generate data frames with Diebold-Mariano test results.
• Code to reproduce all figures in the paper and the supplemental material is provided in the folder /reproducing results and figures/.