Adds ERA5 reanalysis driven forcing at cfsites #3519
Conversation
|
|
||
| zc_gcm = Fields.coordinate_field(Y.c).z[colidx] | ||
|
|
||
| setvar!(ᶜdTdt_hadv, "tntha", colidx, zc_gcm, zc_forcing) |
There was a problem hiding this comment.
| setvar!(ᶜdTdt_hadv, "tntha", colidx, zc_gcm, zc_forcing) | |
| evaluate!(ᶜdTdt_hadv, dTdt_timevaryinginput, t) |
Julians42
force-pushed
the
j/ERA5Driven2
branch
2 times, most recently
from
1e32660 to
cd74d00
Compare
| moist: "equil" | ||
| config: "column" | ||
| z_max: 70e3 | ||
| truncation: 40000 |
There was a problem hiding this comment.
Can you document the "truncation" option or remove it from the config if it isn't used.
| cfsite_number::String | ||
| end | ||
|
|
||
| function (initial_condition::ERA5Driven)(params) |
There was a problem hiding this comment.
These seem similar to the GCMDriven initial condition functions - you can dispatch over functions that perform the same operations in both setups vs. dispatching just on ERA5.
| set_insolation!(insolation) | ||
| set_cos_zenith!(cos_zenith) | ||
|
|
||
| @. ᶜinv_τ_wind[colidx] = 1 / (6 * 3600) |
There was a problem hiding this comment.
Note it's not essential for the relaxation timescales to be exactly the same as the LES setup. I would add a note about that somewhere (i.e. "these are taken from Shen et al., 2022, but may vary")
| @@ -297,3 +297,200 @@ function external_forcing_tendency!(Yₜ, Y, p, t, ::ISDACForcing) | |||
| # total specific humidity | |||
| @. Yₜ.c.ρq_tot += Y.c.ρ * ᶜdqtdt_nudging | |||
| end | |||
|
|
|||
| # For ERA5 reanalysis forcing data | |||
| function external_forcing_cache(Y, external_forcing::ERA5Forcing, params) | |||
There was a problem hiding this comment.
I would try to dispatch at a lower level if possible, specifically for external_forcing_cache and external_forcing_tendency on the ERA5Forcing and GCMForcing types. For instance, instead of adding a external_forcing_cache dispatched on the ERA5Forcing, you could have one functions that's internally dispatch on setvar! and called for both ERA5Forcing and GCMForcing.
| parent(cc_field[colidx]) .= interp_vertical_prof( | ||
| zc_gcm, | ||
| zc_forcing, | ||
| era5_driven_profile(ds.group[cfsite_number], varname), |
There was a problem hiding this comment.
I would be more straightforwards to have new function names match the existing naming pattern. i.e. the setvar! for the GCMDriven cases uses the time-mean gcm_driven_profile_tmean profile, whereas the setvar! ERA5 calls era5_driven_profile, and it's not clear if it's time-mean or not. This will also be less confusing when there's time-varying forcings
Julians42
force-pushed
the
j/ERA5Driven2
branch
from
cd74d00 to
784da18
Compare
Julians42
force-pushed
the
j/ERA5Driven2
branch
from
291f4e7 to
4f2daca
Compare
Purpose
Adds ERA5 Driven reanalysis forcing. Forcing files are currently stored on HPC under
/central/groups/esm/jschmitt/era5/forcing. Eventually these will be made into an artifact but leaving as files for now.To-do
Content