New Processing Tasks: Collection of Possible Diagnostics

[valentinaschueller]

See discussion in valentinaschueller#90.

Summary:

Ocean Diagnostics

• a global map of the Barotropic stream function (depth-integrated volume transport). If XIOS can write out this variable directly, this can be treated directly. If not, we need a processing task á la nemo_depth_integral_all_mean_map or simlar. Regional maps of this variable will require (a processing task? and) a map type handling function that can deal with regional plots.
• AMOC: We are interested in the time series of the maximum of a variable in a lat-depth region, to create an AMOC index time series. A map of this lat-depth region can also be interesting. Comment from David: time series for maximum AMOC at 26N (using the Atlantic meridional stream function diagnostic, called zomsfatl in the outputs)
• mixed layer depth: an interesting diagnostic would be a Labrador sea mean of the mixed layer depth.
• ENSO diagnostics are less relevant for monitoring and more relevant for analysis. The ENSO is typically visualized using a frequency plot which does not fit into our current diagnostic types. Pasha had an idea for a better monitoring diagnostic that would also give an idea about ENSO

Atmosphere Diagnostics

From David: "It is probably useful to have a map of sea-level pressure here (variable 151 in IFS). Other useful diagnostics include the 10m wind speed (variables 165 and 166) and all the radiation fluxes (surface net solar [176] and thermal [177], top net solar [178] and thermal [179], surface sensible [146] and latent [147] heat flux)."

From the EC-Earth 3 Post-Processing Tool I found these relevant time series diagnostics which we can not treat yet:

• precipitation minus evaporation: tp+e, mm/day
• top of atmosphere net heat flux: tsr+ttr, Wm-2
• surface net heat flux with snowfall: sshf+slhf+ssr+str+sfhf, Wm-2
• surface net heat flux: sshf+slhf+ssr+str, Wm-2
  Maybe all of these can be treated by an OpenIFS proc task that can sum up different variables.

We should also try to figure out how we can deal with data that accumulates over time (e.g. tp, e)

From Tim:

• precipitation: maybe a 2D map, maybe a zonal mean. In both cases: climatology over the last x years
• global mean, annual (or maybe monthly) mean time series net radiation flux at top of atmosphere. Either the variable itself or the sum of rlut, rsdt, and rsut.
• analogous time series for net radiation flux at the surface. --> another formula

[valentinaschueller]

Something to keep in mind when this discussion starts up again: The tasks nemo_global_mean_year_mean_timeseries and oifs_global_mean_year_mean_timeseries should be able to deal both with 2D and 3D variables. (Otherwise, 2D and 3D task would have the same name according to the scheme)