One record vs. two record EMITIMES results

Topics about the HYSPLIT dispersion model.
Post Reply
sampls
Posts: 10
Joined: March 1st, 2022, 1:59 am
Registered HYSPLIT User: No

One record vs. two record EMITIMES results

Post by sampls »

Hello! I am working on simulating fire smoke transport, and I noticed that when I included more than one 24-hour emissions record, the result for the first 24 hours changed. It doesn't change significantly, but it does change a bit. I have included the two EMITIMES files I am using, as well as the concentration results from the same time period on the first day of the simulation. I ran both of these using the same control file. The only difference between them was the emissions file specified. If anyone could provide insight into why this happens, I would be very grateful! Let me know if I can provide any additional information.
Attachments
combo_EMITIMES_onerec.txt
(217 Bytes) Downloaded 1213 times
aug 22 one rec 2000.png
aug 22 one rec 1700.png
alicec
Posts: 418
Joined: February 8th, 2016, 12:56 pm
Registered HYSPLIT User: Yes

Re: One record vs. two record EMITIMES results

Post by alicec »

Hello,
Please provide more information.

The graphics you have included are for different time periods.
The first one says it is for 20:00 to 22:00 and the second one from 17:00 to 18:00.

The 'combo_EMITIMES_onerec.txt' file looks like this

YYYY MM DD HH DURATION(hhhh) RECORDS
YYYY MM DD HH MM DURATION(hhmm) LAT LON HGT(m) RATE(/h) AREA(m2) HEAT(w)
2013 08 22 00 0048 0001
2013 08 22 00 00 2400 37.90894 -120.0058 0. 1.13E+16 1.67E+08 0

and I don't see the other EMITIMES file.
sampls
Posts: 10
Joined: March 1st, 2022, 1:59 am
Registered HYSPLIT User: No

Re: One record vs. two record EMITIMES results

Post by sampls »

Thanks for the response. Apologies for the mixed-up files! I have attached the correct two images and the second EMITIMES file. The first EMITIMES you noted is correct.

The second EMITIMES file looks like this:

YYYY MM DD HH DURATION(hhhh) RECORDS
YYYY MM DD HH MM DURATION(hhmm) LAT LON HGT(m) RATE(/h) AREA(m2) HEAT(w)
2013 08 22 00 0048 0002
2013 08 22 00 00 2400 37.90894 -120.0058 0. 1.13E+16 1.67E+08 0
2013 08 23 00 00 2400 37.92899 -119.9570 0. 1.12E+16 1.22E+08 0
Attachments
combo_EMITIMES_tworec.txt
(292 Bytes) Downloaded 1313 times
aug 22 two rec 1700.png
aug 22 one rec 1700.png
alicec
Posts: 418
Joined: February 8th, 2016, 12:56 pm
Registered HYSPLIT User: Yes

Re: One record vs. two record EMITIMES results

Post by alicec »

I'm pretty sure this is just because of a difference due to particle number.

Look at your MESSAGE file (or attach here) and see how many computational particles were allocated all together.

The numpar setting in SETUP.CFG tells HYSPLIT how many computational particles to user per emission cycle.
In your first setup, all those computational particles will go to simulating the first 24 hour release.
In your second setup, half the computational particles will be allocated in the first 24 h and half in the second 24 h.
As you say, the differences are small. These look like differences that are expected if you run the same simulation with
different number of particles or different random seed.

You can try the following if you would like
1. increase numpar to twice what you had before
OR
2. create two emission cycles in your EMITIMES file as shown below.

YYY MM DD HH DURATION(hhhh) RECORDS
YYYY MM DD HH MM DURATION(hhmm) LAT LON HGT(m) RATE(/h) AREA(m2) HEAT(w)
2013 08 22 00 0024 0001
2013 08 22 00 00 2400 37.90894 -120.0058 0. 1.13E+16 1.67E+08 0
2013 08 23 00 0024 0001
2013 08 23 00 00 2400 37.92899 -119.9570 0. 1.12E+16 1.22E+08 0
Post Reply

Return to “Dispersion Model”