sensitivity.analysis.RdPerforms univariate sensitivity analysis and variance decomposition
sensitivity.analysis(trait.samples, sa.samples, sa.output, outdir)
| trait.samples | list of vectors, one per trait, representing samples of the trait value, with length equal to the mcmc chain length. Samples are taken from either the prior distribution or meta-analysis results |
|---|---|
| sa.samples | data.frame with one column per trait and one row for the set of quantiles used in sensitivity analysis. Each cell contains the value of the trait at the given quantile. |
| sa.output | list of data.frames, similar to sa.samples, except cells contain the results of a model run with that trait x quantile combination and all other traits held at their median value |
| outdir | directory to which plots are written |
results of sensitivity analysis
This function estimates the univariate responses of a model to a parameter for a set of traits, calculates the model sensitivity at the median, and performs a variance decomposition. This function results in a set of sensitivity plots (one per variable) and plot_variance_decomposition.
# NOT RUN { sensitivity.analysis(trait.samples[[pft$name]], sa.samples[[pft$name]], sa.agb[[pft$name]], pft$outdir) # }