Your prior on these parameters is independent, but that doesn’t mean your posterior will be. You can compute the posterior covariances on these parameters and see.
If you’re happy with your model, I don’t see any reason to go about putting covariance in your priors.