Thanks, Chris. I appreciate it. The sampler speed greatly improved when I narrowed the prior distribution for the sigma parameter.
I realized that my parametric function does not fit some of the data though. So, I am thinking of going to a more general probabilistic model, such as the GaussianRandomWalk, or the EulerMurayama SDE method. I am reading up on both methods to get a better sense of the underlying mechanics. Any advice on which approach might be a better fit here?