Disappearing UM stash?


A bit of a strange one here. I’ve been running an ensemble nesting suite (u-cv811 - committed up to date) for some 3 day forecasts. Everything seems fine with the output, except for the fact I can’t find the total precipitation rate (stash 05 216) anywhere. This is the only stash that seems to be missing from the output.

I’ve double-checked the suite and it’s definitely included in my stash requests. I’m a bit lost about where to go with this. I have multiple runs of this suite and none of them have this stash included in the output.

According to the suite, the stash should be included in the same usage, domain and time profiles as the rest of my single-level variables (the “pa” stream). When I do a iris.load_cube("umnsaa_pa000.pp","m01s05i216"), I’m met with nothing. When I load the whole file (iris.load("umnsaa_pa000.pp"), the stash is nowhere to be found in the list of output diagnostics.

I went out on a limb and tried iris.load_cube("umnsaa_p*000.pp","m01s05i216") just in case it was erroneously packaged into another output stream, but it’s still nowhere to be found.

Do you have any suggestions of how to diagnose this problem?



/home/d02/ajohnson/cylc-run/u-cv811/log/job/20190323T0000Z/VERONICA_km2p2_ra2t_um_fcst_em11_cr3/01/job.out says
??????????????????????????????      WARNING       ??????????????????????????????
?  Warning code: -30
?  Warning from routine: PRELIM_MOD:PRELIM
?  Warning message:
?          Field - Section:5, Item:216 request denied.
?          Unavailable to this model version.
?  Warning from processor: 0
?  Warning number: 43

This rings a bell - it might be because you configuration does not have pc2 switched on (I can’t find where I’ve seen that before.) Perhaps your supervisor might know?


Ah thanks, Grenville. I was hoping the output was just hiding somewhere.

PC2 is actually turned on in the suite. It must be something to do with the convective precip. I’ve got the stratiform rainfall rate output but I didn’t request the convective (and I’m pretty sure the total is just stratiform + convective).

My suite is convection-permitting.

The only other option I can find is the l_conv_prog_precip (“include 3D prognostic field of recent convective precipitation”) which is currently turned to false. But this is described as an “experimental” variable and I’m not convinced it’s relevant to the single level rainfall.

I would’ve expected total precip to be a standard diagnostic!


Any luck with this - we can look more in depth if not?


Hi Grenville,

After conversations with people at the Met Office I’ve realised that the convective rainfall stash is only used for the convection scheme (parametrisation), which is why I’m getting this error.

Despite stash 4203 being labelled as the “large scale rainfall rate” or the “stratiform rainfall rate” (depending on where you look), it is presumably all of the rainfall that is derived from the model when the convection scheme is turned off. I’m under the impression that this stash captures the rainfall from the explicit convection, even if not stratiform(?)

Total rainfall with explicit convection = 4203
Total rainfall with parametrised convection = 5216 = 4203 + 5205

Thanks for getting back to me!

This topic was automatically closed 2 days after the last reply. New replies are no longer allowed.