# Obs filtering

**URL:** <https://forums.jcsda.org/t/obs-filtering/112>\
**Category:** Uncategorized\
**Created:** [July 19, 2021, 8:43pm UTC](https://forums.jcsda.org/t/obs-filtering/112 "2021-07-19T20:43:14Z")\
**Posts on this page:** 2\
**Page:** 1

<div class="post-metadata">

**Author:** ![andytangborn](https://yyz1.discourse-cdn.com/flex031/user_avatar/forums.jcsda.org/andytangborn/32/48_2.png) [@andytangborn](https://forums.jcsda.org/u/andytangborn)\
**Post date:** [July 19, 2021, 8:43pm UTC](https://forums.jcsda.org/t/obs-filtering/112/1 "2021-07-19T20:43:14Z")

</div>

I’ve just run into a problem when I apply obs filtering in the hyb-3dvar\_gfs\_aero.yaml file that didn’t occur before. And I wonder if recent changes to the .nc4 file structure are the cause. When I try to filter obs using “Domain Check”, I get undefined (meaning e+38) values for oman and hofx1. Have there been changes that would cause the following yaml section to fail?

```
obs operator:
  name: AodCRTM
  Absorbers: [H2O,O3]
  obs options:
    Sensor_ID: v.viirs-m_npp
    EndianType: little_endian
    CoefficientPath: Data/crtm/
    AerosolOption: aerosols_gocart_default
obs error:
  covariance model: diagonal
obs filters:
- filter: Domain Check
  filter variables:
  - name: aerosol_optical_depth
    channels: 4
  where:
  - variable:
      name: latitude@MetaData
    minvalue: 0
```

---

<div class="post-metadata">

**Author:** ![stephenh](https://yyz1.discourse-cdn.com/flex031/user_avatar/forums.jcsda.org/stephenh/32/28_2.png) [@stephenh](https://forums.jcsda.org/u/stephenh)\
**Post date:** [July 20, 2021, 8:31pm UTC](https://forums.jcsda.org/t/obs-filtering/112/2 "2021-07-20T20:31:17Z")

</div>

@andytangborn there have been many recent changes to UFO and IODA which potentially have impacted your hyb-3dvar\_gfs\_aero testing. Since you are suspecting the .nc4 file structure, could you submit a ZenHub issue in ioda with more details on what the failure is and how to reproduce it. That would greatly help us debug this.

Thanks!