Welcome to the new DiscussCESM forum!
We are still working on the website migration, so you may experience downtime during this process.

Existing users, please reset your password before logging in here: https://xenforo.cgd.ucar.edu/cesm/index.php?lost-password/

Lagrangian DSE budget closure

khartig

Kara
New Member
I'm running air parcel trajectories (with HYSPLIT) based on CAM5 output (CESM 1.2.2, FV dynamical core, F_1850_CAM5 compset, f19_f19 resolution) and would ultimately like to generate a closed temperature/dry static energy/potential temperature budget along those trajectories to study the contribution of various processes to temperature change as they pass from the Arctic over continental North America. I had hoped that PTTEND or some subset of its terms (QRL+QRS or QRL+QRS+DTCOND, for example) would close a Lagrangian DSE budget, but so far still have pretty large residuals, on the order of 0.5 K/hr when comparing change in instantaneous DSE (calculated from 'T' and 'Z3', avgflag_pertape='I') to averaged PTTEND (avgflag_pertape='A'). What variables would I need to close the DSE budget in a Lagrangian frame, and are those available with the finite volume dynamical core? It seems like there may still be diffusion contributions in DTCORE that I can't account for independently of advective and adiabatic contributions.
 

brianpm

Member
Hi Kara, I don't know the actual answer to your question. I can confirm that the temperature budget closes. I think that the total temperature tendency should equal DTCORE + PTTEND + TFIX, and can also be constructed by (TAP(n)-TAP(n-1))/(time(n)-time(n-1)). And on the physics side:
PTTEND = DTCOND + QRS + QRL + DTV + TTGWORO

Going to DSE in a Lagrangian frame, I guess there's a geopotential term that you'd need to deal with. Probably the harder part is to account for the mixing along the trajectory. I don't know whether HYSPLIT is able to capture the advection well enough to close the budget.

Maybe you can first assess the DSE budget in the Eulerian frame, which should be tractable, and then deal with transforming to the Lagrangian frame. Maybe the vertically integrated DSE would be easier to deal with, but that's just speculation.
 
Top