Bayesian inference with multiple forward model with different numbers of provided measurements but with one common discrepancy model and parameter?

Thank you for pointing that out! Regarding your question on mean and MAP prediction - I think the way you introduce both terms in your post is very clear.