Particles not wrapping around with global met fields

[mneish]

Running HYSPLIT 5.3.0 back trajectories, with ERA5 global met data encoded with hysplit_metdata and era52arl tools and the default area of 90/-180/-90/180 and grid resolution 0.3. In HYSPLIT, when the particles hit the meteorology longitude bounds (-180), they disappear from the grid. One example of endpoints with this problem:

This seems to be an issue with version 5.3.0. Testing the exact same configuration (but with HYSPLIT 5.2.3), the particles wrap around without any issue:

[mneish]

Same issue happens with GDAS1 met data, so this seems to be a general problem with global met fields and HYSPLIT 5.3? Particles hit the "boundary" of the grid, and die.

[christopher.loughner]

Hi, I was able to successfully use HYSPLIT Version 5.3.0 to run a backward trajectory simulation and a backward dispersion simulation that passed through the dateline with a 1 degree GDAS file that is archived at NOAA ARL (https://www.ready.noaa.gov/data/archives/gdas1/). For dispersion simulations, make sure your concentration grid defined in your CONTROL file is configured to go past the dateline. Also, make sure you are using a global dataset that is not a subset of the global domain. Feel free to send me your CONTROL file if you need help. I recommend first trying to see if you can run a backward trajectory simulation to see if that works before moving on to running a backward dispersion simulation through the dateline.

[mneish]

Apologies for any confusion. For GDAS, I'm getting a different particle cut-off at the prime meridian instead of the dateline. Seems to be specific to the dataset being used. Could you confirm if your particles are getting across the 0 degree longitude in your GDAS test? Thanks!
We're using the full GDAS files, so they should be detected as global fields?
Attached is a plot of an annual mean footprint using GDAS, showing the particle cutoff issue for that dataset.

[christopher.loughner]

Thank you! I was able to reproduce your bug, found when and where this entered into the code, and created a bug fix for this issue. This bug occurs when running a dispersion simulation with IDSP=2 for HYSPLIT versions 5.3.0, 5.3.1, 5.3.2, and 5.3.3. We are currently putting together new HYSPLIT distributions and testing them with this new bug fix. We will reply back when the new distributions are ready.

[christopher.loughner]

Are you running HYSPLIT with Linux or Windows executables? Updated source code and executables for Linux are now available on Ready. This update is Version 5.3.4.

[mneish]

We're running with the Linux executables.
I re-ran our tests against the latest version 5.3.4, and can confirm that this version does fix the global boundary issue for our experiments.
Many thanks for the quick solution to this problem!