Exploring posterior probability

Here’s a quick animation exploring the posterior probability density function of my mantle velocity model from Burdick & Lekic (2017).  The transdimensional Bayesian method gives us a giant ensemble of models that explain the data, so at any given location, we can plot up all of the potential values we find there.  A couple of fun things to note:

1. In complex regions where we find are sharp gradients in velocity, the distribution will often have multiple peaks, since the the boundaries of the model are allowed to move around.
2. In places without data (like beneath the ocean) the algorithm traces out the uniform prior that we assume at the beginning.
3. This should make it clear that the mean and standard deviation don’t always tell the whole story!