Anyone else dialing back 2D mesh density to speed up HEC-RAS 6.5 without wrecking results? I’ve had good luck moving from 2 m to 5 m cells in overbanks, keeping 1 m around structures with breaklines, and calibrating to June 2023 high-water marks — stage RMSE stayed under 0.08 m and runtime dropped from 48 to 11 minutes on a 6‑core laptop. What theta/Courant targets or roughness tweaks are you using to keep stability and avoid dry-cell oscillations when you coarsen?