kdraeder

@jedwards4b reviewed the change and said to open a PR.

This was discussed extensively in #296 for cam-fv. One (partial?) fix was to move the discarding_high_obs test from model_interpolate to get_close_{obs,state}. But that apparently was not merged into main. That...

I had to remind myself that trusted obs are defined for obs_diag, not for the assimilation. So it's possible for the number used for the metric calculation to be _larger_...

The "I'm not a fan..." comment was somewhat leftover from when I hadn't remembered that "trusted" is a post-assimilation tool. It's good to read that designating them as "trusted" often...

I've tidied my local branch, in which I set up the previous BNRH tests. Then I set it to track the qceff_inflation_fix branch by fetching the _fix branch and rebasing...

My notes say that it failed during the first assimilation for ~26,000 variables while processing almost a million obs x number_close_variables. That was started from a Reanalysis restart set, so...

Using the old filter with the namelist negative fixer turned off led to a failure somewhere in the first 10,000 obs: Before observation assimilation TIME: 2024/12/31 10:23:22 ERROR FROM: source...

I set up a new experiment to start from the same date as the assimilation which successfully used fix_bounds_violations (2019-12-30-00000), but turned off fix_b... and used the qceff_inflation_fix code. I...

@hkershaw-brown wrote in email: So I'm on the same page here, this test is comparing 2 versions of a given code (Kevin's ~DART/Manhattan_git) KDM = Kevin's ~DART/Manhattan_git code pre Jeff's...

Here's question that might influence our choices: is it reasonably easy to store some kinds of data distributed across a single node, which is essentially the tasks we request from...