Modify .stan generation to include two station option

> CDI Proposal Tier 1: Modify the R function that creates Stan Bayesian model files, so that two-station models are written out alongside one-station models, with and without variable flow. Because the script loops over many options that are compatible with, yet orthogonal to, these decisions, this change will roughly double the available model configurations.

Bishop et al. (in review) describes a .stan file that has many similarities to our current model structure, but also includes some differences - we may want to implement the Bishop et al. method in a way that adds-to-bishop/retains-in-sM the priors on GPP, ER, and K600, optionally pools K across days, stays consistent with current sM methods for handling light (or adds another approach), potentially represents observation error on the upstream observations.

Start by locating the R function that generates the .stan files and inspect how it currently loops over one-station and variable-flow options. Compare the requested two-station combinations with the existing model outputs; done should mean two-station files are generated alongside the one-station variants, with and without variable flow. The Bishop et al. changes are described as possible follow-on design considerations rather than a defined implementation scope.

Written by the indexing model from the issue text.