-
Notifications
You must be signed in to change notification settings - Fork 29
Thetis meeting agenda 2021 11 26
Tuomas Kärnä edited this page Nov 26, 2021
·
3 revisions
Date/time of meeting 2021-11-26 11:00 GMT
Present: Stephan, Anastasia, Joe, Matt
Tuomas gave some updates on 2D Baltic Sea modeling. Can take long (~1 h) time steps on a coarse (>2 km) mesh. The Crank-Nicolson solver tends to get unstable as it is not L-stable. DIRK22 works much better. Also BDM2-DG1 discretization results in smoother elevation field compared to DG1-DG1, although BDM2-DG1 is ~2x slower. Simulating on the sphere vs. flat UTM coordinates does not make a big difference in terms of computational cost.
Currently the TPXOTidalBoundaryForcing method in Thetis uses uptide's OTPSncTidalComponentInterpolator to interpolate tidal velocities but this method is not in the uptide master version. Tuomas and Stephan will update uptide.