Page 1 of 1

Choosing variables to include - are some "better"?

Posted: June 24th, 2021, 6:52 pm
by khartig
When you're working with data from a simulation and have many options for variables to include, are some better than others? I am working with CESM/CAM output data on hybrid levels, converted to ARL format with MeteoInfo (and hopefully soon cesm2arl) and run through the trajectory model. The bare minimum for the trajectory model seems to be (upper-level: ) temperature, geopotential height, u-wind, v-wind, w- or omega-wind, specific or relative humidity, and (surface level: ) 2m temperature, surface pressure. I have the capacity to include many more, but am not sure which ones will "improve" the trajectory calculation; for example, is there any reason to prefer specific over relative humidity? is providing more variables always better? what will adding precipitation do?

After a glance through some of the available HYSPLIT variables at https://www.ready.noaa.gov/hysplitusersguide/S141.htm and mentioned elsewhere in the forum, I've identified quite a few that are available from CESM/CAM (below, with both the CAM variable identifier and the HYSPLIT counterpart) and was hoping someone could tell me if any of them are preferable (or redundant):

Surface Level:
surface geopotential: PHIS (m2/s2) -> SHGT (m2/s2 divided by g)
surface friction velocity: USTAR (m/s) -> USTR (m/s)
planetary boundary layer height: PBLH (m) -> PBLH (m)
sea level pressure: PSL (Pa) -> MSLP (hPa)
surface latent heat flux: LHFLX (W/m2) -> LHTF (W/m2)
surface sensible heat flux: SHFLX (W/m2) -> SHTF (W/m2)
2m relative humidity: RHREFHT (%) -> RH2M (%)
2m specific humidity: QREFHT (kg/kg) -> SPH2 (kg/kg)
vertically-integrated total cloud: CLDTOT (fraction) -> TCLD (%)
total (convective + large scale) precipitation rate, liquid + ice: PRECT (m/s) -> TPP? (depends on model output rate)

Upper-Level:
turbulent kinetic energy: TKE (m2/s2) -> TKEN (m2/s2)