weird results for small layer height

Hello, I'm simulating a lake that's 9m deep and intermittently stratified. I decided to explore the effect of refining the layer height (DZ). For DZ=0.3m, I get reasonable results: internal waves along the thermocline cause temperatures to fluctuate throughout the water column for the entire 3-month duration of the simulation. However, if I reduce to DZ=0.1m, temperatures flatten out, i.e., stop varying in time, about half-way through the simulation in the bulk of the water column -- only the surface temperature fluctuates at all after this point. Is there some way a small layer height could result in numerical errors that cause effects like I'm seeing? Thanks, Allie