Why do web-based and installed HYSPLIT create different results?

General questions and postings pertaining to the use of HYSPLIT regardless of the platform. For platform specific questions, use the HYSPLIT Platform forums.
Post Reply
szabob
Posts: 9
Joined: February 20th, 2019, 12:58 pm
Registered HYSPLIT User: Yes

Why do web-based and installed HYSPLIT create different results?

Post by szabob » March 1st, 2019, 3:30 am

Dear All,
I registered and intalled HYSPLIT-4 and run a -197 hours back trajectory from two sources. I used GDAS0p5 meteorology from 17.10.2016 to 25.10.2016. The initial time of back trajectory was 2016.05.25 17:00. I also run this back trejectory with same parameters (initial time, meteorology, end points etc) in the web-based HYSPLIT and it created a different result. I checked the control files were prduced by both the installed and web-based model. They are also attached. I found that the web-base HYSPLIT uses one extra meteorology file (20161026_gdas0p5). I do not understand why? Can someone explain me and inform me which version is reliable?
Thank you for your response in advance!
Attachments
Control_File_WEB_HYSPLIT.txt
(1010 Bytes) Downloaded 49 times
Control_File_INSTALLED_HYSPLIT.txt
(448 Bytes) Downloaded 52 times

glenn.rolph
Posts: 361
Joined: November 7th, 2012, 1:39 pm
Registered HYSPLIT User: Yes
Contact:

Re: Why do web-based and installed HYSPLIT create different results?

Post by glenn.rolph » March 1st, 2019, 10:14 am

The extra meteorological file for the 26th is added on the web version so that users can run back trajectories starting late on the 25th. HYSPLIT needs the extra file so that it can interpolate meteorological data for the last few hours of the day. You ran 4 trajectories from the same location and at two different heights on the web versus two trajectories at two different heights on the PC.

szabob
Posts: 9
Joined: February 20th, 2019, 12:58 pm
Registered HYSPLIT User: Yes

Re: Why do web-based and installed HYSPLIT create different results?

Post by szabob » March 4th, 2019, 3:15 am

Dear glenn.rolph,
Thank you for your response! I run again the back trajectory with two trajectory from the same location with two different heights in web and I also run the back trajectory with same set up (using meteorology in 26th as well) on my PC but the results are still different. I do not know which can be the right? I attached the results.
Attachments
trajplot_HYSPLIT_web.pdf
(73.91 KiB) Downloaded 51 times

glenn.rolph
Posts: 361
Joined: November 7th, 2012, 1:39 pm
Registered HYSPLIT User: Yes
Contact:

Re: Why do web-based and installed HYSPLIT create different results?

Post by glenn.rolph » March 4th, 2019, 7:38 am

Please provide the 2 outputs to show how they are different as well as the CONTROL and SETUP.CFG files.

szabob
Posts: 9
Joined: February 20th, 2019, 12:58 pm
Registered HYSPLIT User: Yes

Re: Why do web-based and installed HYSPLIT create different results?

Post by szabob » March 4th, 2019, 8:01 am

The SETUP.CFG and the CONTROL files are the same (except the top of the model). I attached the MESSAGE files. The Message files are different. I do not know which one is reliable.
Attachments
MESSAGE_WEB.txt
(84.71 KiB) Downloaded 49 times
MESSAGE_PC.txt
(80.05 KiB) Downloaded 49 times

glenn.rolph
Posts: 361
Joined: November 7th, 2012, 1:39 pm
Registered HYSPLIT User: Yes
Contact:

Re: Why do web-based and installed HYSPLIT create different results?

Post by glenn.rolph » March 4th, 2019, 8:38 am

The versions of the model are different, but should not create that much difference. Please send us the exact CONTROL and SETUP.CFG files you are using as well as the graphics from both, as we need those to find why the runs are different. You might also want to do a reset on the main page of PC HYSPLIT and retry your simulation.

szabob
Posts: 9
Joined: February 20th, 2019, 12:58 pm
Registered HYSPLIT User: Yes

Re: Why do web-based and installed HYSPLIT create different results?

Post by szabob » March 4th, 2019, 8:45 am

I attached the Setup and Control files.
Attachments
Setup.zip
(375 Bytes) Downloaded 51 times
Control.zip
(875 Bytes) Downloaded 53 times

glenn.rolph
Posts: 361
Joined: November 7th, 2012, 1:39 pm
Registered HYSPLIT User: Yes
Contact:

Re: Why do web-based and installed HYSPLIT create different results?

Post by glenn.rolph » March 4th, 2019, 9:18 am

If you set the maximum height of the model on the PC to 15240.0 you will get the same results as the web version. HYSPLIT uses this number to scale the internal levels of the model and because your trajectories crossed some weather discontinuity in northern Italy, the small difference in the levels of the model made a large difference in the calculation of the backward trajectories.

szabob
Posts: 9
Joined: February 20th, 2019, 12:58 pm
Registered HYSPLIT User: Yes

Re: Why do web-based and installed HYSPLIT create different results?

Post by szabob » March 4th, 2019, 9:31 am

Thank you! You are right! It caused the differences.

Post Reply