[BUG/ISSUE] Malformed OFFLINE_BIOVOC data file for 2017-12-28

[kilicomu]

What institution are you from?

Wolfson Atmospheric Chemistry Laboratories

Description of the problem

The HEMCO data file OFFLINE_BIOVOC/v2019-10/0.5x0.625/2017/12/biovoc_05.20171228.nc appears to malformed, and is causing my simulations to crash. Any chance it can be reprocessed?

Description of troubleshooting performed

Looked at the GCHP log files to point me at the cause of the crash, narrowed it down to an OFFLINE_BIOVOC file by looking at dates and file sizes - you can see that the file for 2017-12-28 is much smaller than all others for the month. I dumped the data from that file using ncdump and there looks to be an awful lot more zeroes in that file than in the others!

GEOS-Chem version

N/A

Description of modifications

N/A

Log files

N/A

Software versions

N/A

[kilicomu]

Thinking about it...I'd be surprised if nobody else had hit this problem before with GC Classic...

The specific error message that I got came through GCHP's allPEs.log:

0000: CAP.EXTDATA: ERROR: ExtData could not find bracketing data from file template ./HcoDir/OFFLINE_BIOVOC/v2019-10/0.5x0.625/%y4/%m2/biovoc_05.%y4%m2%d2.nc for side R

Would this be handled differently in GC Classic?

[lizziel]

We've had issues with the offline biovoc files in the past due to the timestamps. In your ncdump what does it say the time for the file is?

[lizziel]

See also https://github.com/geoschem/GCHP/issues/84.

[lizziel]

One other thing. If your simulation is at 12/28 then the right bracket is likely 12/29, so that would be the problematic file.

[kilicomu]

Yep, you got it. The timestamp on the file for the 29th is not correct. I'll change that and give it another go.

However, I am still concerned about the file for the 28th - it weighs about 1MB vs the ~35MB of all other files in the month and is largely 0 where other files around it are not!

[lizziel]

Have you tried re-downloading the data? @YanshunLi-washu and @LiamBindle, this was fixed for the data available to download, correct?

[kilicomu]

Yeah I did - the OFFLINE_BIOVOC index at http://geoschemdata.wustl.edu/ExtData/HEMCO/OFFLINE_BIOVOC/v2019-10/0.5x0.625/2017/12/ reports a file size of 932k for that file, compared with 35-39mb for the others.

[kilicomu]

I should say that touching up the timestamps in the files for the last few days of my simulation has got it working now, so, thanks for the pointer towards that!

[YanshunLi-washu]

Hi @kilicomu @lizziel , the PKU group recently produced a new version http://geoschemdata.wustl.edu/ExtData/HEMCO/OFFLINE_BIOVOC/v2021-12/0.5x0.625/2017/12/. Let me know whether this could help. Thanks

[kilicomu]

@YanshunLi-washu Thanks for the heads up - I actually grabbed this dataset already, so I'll see how it goes with my next simulation.

[lizziel]

@kilicomu, are you all set with this issue?

[kilicomu]

Sure, I've got my simulations running.