# UM 13.8 odd behaviour

**URL:** <https://cms-helpdesk.ncas.ac.uk/t/um-13-8-odd-behaviour/1787>\
**Category:** Unified Model\
**Tags:** ARCHER2\
**Created:** [24 August 2025 17:07 UTC](https://cms-helpdesk.ncas.ac.uk/t/um-13-8-odd-behaviour/1787 "2025-08-24T17:07:12Z")\
**Posts on this page:** 4\
**Page:** 1

<div class="post-metadata">

**Author:** ![SimonTett](https://avatars.discourse-cdn.com/v4/letter/s/3da27b/32.png) [@SimonTett](https://cms-helpdesk.ncas.ac.uk/u/SimonTett)\
**Post date:** [24 August 2025 17:07 UTC](https://cms-helpdesk.ncas.ac.uk/t/um-13-8-odd-behaviour/1787/1 "2025-08-24T17:07:12Z")

</div>

Dear CMS,

I have been trying to run u-dr496 with 1 month jobs starting 1/9/2010 and using a 360-day calendar. Model is vn13.8 using cylc 8. It was taken from a 13.8 UKESM1.1 AMIP workflow and has a much reduced diagnostic set – only producing one output stream of monthly mean data and running (in theory) for 17 months. 17 months gives me a 4 month spin up, 12 months for the data I am interested in and a final month to make the December 2011 pp data.

However, if I run in 1 month segement the model runs the September 2010 cycle. But when it runs the October 2010 cycle it starts again with 1/9/2010 and runs 2 months. Then for November cycle it starts again with 1/9/2010 and runs 3 months… For the December cycle, the model starts 1/12/2010 and runs a month. Next cycle it runs 2 months etc. Eventually the model fails in post-process with a complaint about 0 length header when it is expected 256 words… I am guessing this is because the post-process can’t find something because the model has started again and post-process runs fast enough to not find an expected file.

If I run in three month periods then everything (so far – april 2011) is OK. I suspect I am pushing the UM system too far but is there anything I can to make it run in 1 month segments? My goal is to get the Dec 2011 output converted to pp and archived without having to do a manual intervention and use the minimum computer time!

Simon

---

<div class="post-metadata">

**Author:** ![RosalynHatcher](https://dub1.discourse-cdn.com/flex013/user_avatar/cms-helpdesk.ncas.ac.uk/rosalynhatcher/32/11_2.png) [@RosalynHatcher](https://cms-helpdesk.ncas.ac.uk/u/RosalynHatcher)\
**Post date:** [24 August 2025 19:31 UTC](https://cms-helpdesk.ncas.ac.uk/t/um-13-8-odd-behaviour/1787/2 "2025-08-24T19:31:43Z")

</div>

Hi Simon,

It’s because you are only producing a restart dump every 90 days so cycle 2 has nothing to start from so has to start again from the beginning. In panel _um → namelist → Model input and output → Dumping and meaning_ change the dump period to 30.

Cheers,  
Ros.

---

<div class="post-metadata">

**Author:** ![SimonTett](https://avatars.discourse-cdn.com/v4/letter/s/3da27b/32.png) [@SimonTett](https://cms-helpdesk.ncas.ac.uk/u/SimonTett)\
**Post date:** [26 August 2025 11:43 UTC](https://cms-helpdesk.ncas.ac.uk/t/um-13-8-odd-behaviour/1787/3 "2025-08-26T11:43:01Z")

</div>

Thanks a lot Ros.

Simon

---

<div class="post-metadata">

**Author:** ![system](https://europe1.discourse-cdn.com/flex013/uploads/cms_support/original/1X/1fd2411499ffcbc299fe756cd5cdf26e44956558.png) [@system](https://cms-helpdesk.ncas.ac.uk/u/system)\
**Post date:** [25 September 2025 11:43 UTC](https://cms-helpdesk.ncas.ac.uk/t/um-13-8-odd-behaviour/1787/4 "2025-09-25T11:43:47Z")

</div>

This topic was automatically closed 30 days after the last reply. New replies are no longer allowed.