topo_mask for Ensemble-Stat

[JiliDong-NOAA]

While working on verification for ensemble forecasts, I found Ensemble-Stat didn't take topo_mask dictionary, which works well for Point-Stat. EnsembleStat didn't crash but showed no impact of topo_mask. As we need to limit the ensemble verification to obs-model pairs with small terrain difference, I wonder if there is any way to use topo_mask for ensemble-stat or if my set up was incorrect? Here is the topo_mask section I added for MET config of EnsembleStat:

topo_mask = {
   flag               = TRUE;
   file_name          = [];
   field              = { name = "HGT"; level = "Z0"; }
   regrid             = { method = BILIN; width = 2; }
   use_obs_thresh     = ge-100&&le100;
   interp_fcst_thresh = ge-50&&le50;
}

Thanks,

Jili

[j-opatz]

Hello Jili,

And thank you for your question. Upon review of the documentation, I tend to agree with your assessment: topo_mask should work with ensemble-stat, as long as your observation dataset is point observations.

I'm going to tag @JohnHalleyGotway so he can provide a more succinct answer when he returns to the office, but in the meantime I do have two recommendations to provide.

The first is to double-check your message_type_group_map dictionary and ensure that the SURFACE key has the correct values attached to it. If your observation variable HGT does not come from one of the groups listed in the val entries (or if you forgot to include the message_type_group_map dictionary), then it will not be used for topography information.
Secondly, you may want to change out the level you're checking to L0. This is only to double check that Z0 will not cause unforseen errors and to better match the documentation examples; I am skeptical that this will suddenly allow the logic to work, but it's worth a check.

[j-opatz]

Hi @JiliDong-NOAA, I wanted to update you on the situation:

John HG provided the correct information that ensemble-stat does not support the use of topo_mask.

However, he agrees that this is something that can be added to ensemble-stat and has created an issue for the creation of this ability.

We are also looking to update the documentation so that details like these (i.e. tool-specific dictionaries and options) are easier to find and understand.

[JiliDong-NOAA]

Thanks for the info and help! I am looking forward to future updates on this.

Jili