Greg, I re-ran the simulation saving restart files, then picked a time (2004112012) and then ran it for 30
minutes with model output every 2 minutes (instead of 6 hours as usual). This was done because in using 6 hour output, we
were assuming a constant diabatic pv change for 12 hours essentially. This reduces that to a 4 minute assumption. Also the radiation timestep was set to the model
timestep (120 seconds). Below are those same plots from the group meeting except with this new simulation. It looks much
better when I calculate advection this time...but the Lagrangian calculation is still pretty far off.
WRF PV profiles: Model output every 2 minutes
Ertel PV vertical profiles. The left plot shows the forward
differencing in time, and the right plot shows centered differencing in time. In both plots, the blue line is the pv and the
red line is the change in pv. Everything is in PV units.
Plots without the advection term calculated (Lagrangian
calculations). The plot on the left is from following the vortex at a single gridpoint, while the plot on the right is by
taking an average within the 280K closed tropopause theta contour.
Profiles with the advection term calculated. Uses centered
differencing in time and space to estimate. The plot on the left is following a single gridpoint at the core while the plot
on the right is by
taking an average within the 280K closed tropopause theta contour. The goal here is for the green and red lines to match.