nancy collins

hi kevin, you did post that already. i agree with points 2-4. i'm going to try one last time with point 1. if you still don't agree, please schedule a...

i think the intent was this would skip computing inflation on an item-by-item basis if any ensemble members had missing values for that item. if this is *not* skipping the...

i looked at this code again (and i updated my last comment). i'd forgotten how groups work. this code *is* trying to compute inflation on a per-item basis, but if...

in looking at the inflation code i believe for adaptive inflation kinds 2 and 5 that the ```update_inflation()``` routine is called for each group, so the adaptive inflation values are...

eons ago i think we decided that instead of filter setting the missing status, model_mod should have the control. that accomplishes two good things. 1) any executable that uses the...

does PMO use the same init code as filter? also dart_to_model and model_to_dart - all of those main modules will have to have the same flag or they won't treat...

this is an attempt at a brain dump on "invalid data in the state" issues. i'm not sure where to put this comment but i'll add it here and you...

great point about differentiating between all ensemble members having invalid data at a location vs some members having valid data and others not. here are two other thoughts about missing...

the alternative to removing this option is to fix the code in the assimilation loop so it calls convert_vertical_obs() explicitly before doing the broadcast of values and setting QC=8 if...

hi kevin, to answer your question about assim vs output files -- the difference is when filter cycles multiple times without exiting, like for the lower order models. the assim...