Page 1 of 2

HYSPLIT error

Posted: May 4th, 2018, 8:23 am
by Smaug
I am getting the following error:

forrtl:error(69): process interrupted (SIGINIT)

after the simulation starts. I have attached the error, SETUP, CONTROL and EMITIMES file.

Thanks

Re: HYSPLIT error

Posted: May 4th, 2018, 11:01 am
by ariel.stein
I do not see the SETUP.CFG in the attachments. Can you please send it?

You could start by reducing the number of sources in the EMITIMES file to only 2 records and see if your run proceeds.

Re: HYSPLIT error

Posted: May 4th, 2018, 2:20 pm
by Smaug
SETUP attached.

Re: HYSPLIT error

Posted: May 8th, 2018, 10:56 am
by Smaug
I was able to reduce the number of entries in the EMITIMES file and the simulation was able to complete. The simulation will complete if I stop releasing on the 21st (so a total of 22 entries) . Is there a variable that I need to change so I can use more sources in the EMITIMES?

Re: HYSPLIT error

Posted: May 8th, 2018, 1:18 pm
by ariel.stein
Try replacing all the 0024 entries by 0001 and the 2400 by 0100. I assume you want to change your emissions hourly.

Re: HYSPLIT error

Posted: May 9th, 2018, 3:35 am
by Smaug
The emission rates are averaged over 24 hours so I think I have to keep emission duration as 24 hours.

Re: HYSPLIT error

Posted: May 9th, 2018, 10:32 am
by Smaug
Is there an alternative solution for this problem?

Re: HYSPLIT error

Posted: May 9th, 2018, 12:54 pm
by alicec
Can you please attach txt files of your CONTROL, EMITTIMES and SETUP.CFG files.
The most efficient way for me to diagnose your problem is to use your files to create a run
and it is very time consuming to do that with your pdf's.
Also, please attach MESSAGE file you are getting from your run.
Thanks,
Alice

Re: HYSPLIT error

Posted: May 10th, 2018, 7:40 am
by Smaug
I've attached the input files. I might have to attach the pdf of the MESSAGE file as I am running this on a separate system.

Re: HYSPLIT error

Posted: May 10th, 2018, 10:44 am
by alicec
I was able to run with your files.
I did change the numpar in SETUP.CFG to 100 to make it run faster.

I set it back to 100,000 but it looks like that is going to take awhile. Also you have
another issue. You are releasing 100,000 particles every 24 hour emission cycle
but you have maxpar set at 300,000 and you have a lot more than 3 emission cycles.
So after, 3 cycles you are maxing out the number of particles your simulation will carry.
If you look at your message file you probably have Warning: emstmp -exceeding puff limit which
means that you have reached the maxpar number of particles and HYSPLIT won't release any more.

you need to either

1. increase maxpar
2. kill your older computational particles by setting KHMAX in the SETUP.CFG (to 72 for instance, so you only keep them around for 3 emission cycles).
3. decrease the size of your meteorological grid so that particles that go outside of your area of interest are removed from the simulation.
Remember that gdas1 is a global met grid, so as is, HYSPLIT will keep following those particles all around the globe. There is a utility that will extract a portion of gdas1 global grid for you.

I was not able to reproduce the error:
forrtl:error(69): process interrupted (SIGINIT)

Please let me know your compiler, HYSPLIT version and system you are running on (linux flavor? windows? Max?)

Thanks,
Alice