Describe the bug
When running GRAL (run 1), if the gff file doesn't exist, GRAL creates it, then creates the grz file. The concentrations are as expected.
However, when running GRAL a second time (run 2), if the same gff file from the exact above simulation is read from for the next GRAL simulation, the resulting concentrations are slightly different. However, if GRAL is run again for the same exact simulation (run 3), but this time I delete the gff file from run 1, the concentration results are equal to run 1.
To Reproduce
Steps to reproduce the behavior:
- GRAL ran with reproducible results option to verify the slight differences in results (depending on presence of existing gff).
- GRAL run with steady state mode and buildings and point sources.
- GRAL was ran outside the GUI to isolate the problem.
- Resulting concentrations were compared using the GUI. The concentrations files were slightly different between these 2 cases.
Expected behavior
It is expected that if the gff file already exists, GRAL would use it. I see this happening as I see the advection scheme is skipped. I would expect identical results between GRAL runs if the gff file is unavailable (run 1 - file is created during run), compared to if the gff file is available (from run 1 - file already has been created).
I have found that the results are consistently different. Meaning multiple GRAL runs from scratch result in max concentration of 931, whereas multiple GRAL runs with the gff already present from the first run yield max concentration of 933.
As you can see. This is a very small difference.
Desktop (please complete the following information):
Describe the bug
When running GRAL (run 1), if the gff file doesn't exist, GRAL creates it, then creates the grz file. The concentrations are as expected.
However, when running GRAL a second time (run 2), if the same gff file from the exact above simulation is read from for the next GRAL simulation, the resulting concentrations are slightly different. However, if GRAL is run again for the same exact simulation (run 3), but this time I delete the gff file from run 1, the concentration results are equal to run 1.
To Reproduce
Steps to reproduce the behavior:
Expected behavior
It is expected that if the gff file already exists, GRAL would use it. I see this happening as I see the advection scheme is skipped. I would expect identical results between GRAL runs if the gff file is unavailable (run 1 - file is created during run), compared to if the gff file is available (from run 1 - file already has been created).
I have found that the results are consistently different. Meaning multiple GRAL runs from scratch result in max concentration of 931, whereas multiple GRAL runs with the gff already present from the first run yield max concentration of 933.
As you can see. This is a very small difference.
Desktop (please complete the following information):