[UWBCFD]

My system involves two partially closed cell zones of air with adiabatic boundary conditions. The cell zones are in contact and are allowed to interact at that contact region. Each cell zone starts out at different initial temperatures and is undergoing a transient, density based study with full buoyancy effects and viscid flow.

Any thoughts on these residuals?

[LuckyTran]

That is more or less how residuals behave in a transient simulation. What do you think is odd?

Residuals tend to decrease (hopefully) with each iteration. At the next time-step the most recent solution is used as the initial guess (which is not the solution to the new time-step) so residuals jump up on the first iteration of subsequent time-steps.