Emulsification and ADIOS problems

[amercatini]

Hi everyone,
First of all, thank you for the amazing work on OpenDrift and the OpenOil module. It is an incredibly useful tool for the community.
I am reaching out because I noticed that for several specific oils in the ADIOS database, the emulsification process is effectively disabled because the emulsion_water_fraction_max (total water) data is missing or set to zero.
To make the model more robust in these cases, would it be possible to implement a fallback mechanism?
For example, when the database returns zero or a null value, the model could:

• Allow the user to provide a manual default value (e.g., 0.7 or 0.8) via the configuration settings.

• Or use a simple empirical formula based on oil density to estimate a reasonable maximum water fraction.

This would prevent the simulation from skipping the emulsification phase entirely when a specific database entry is incomplete.
I am specifically seeing this issue with oils like IFO 180, where the lack of 'total water' data stops the emulsification process.
I would be happy to help test this feature or provide further details if needed.
Best regards,
Alessandro

[knutfrode]

Hi, we are happy if you have found OpenDrift useful for your work.

Indeed, I also find that there is no emulsion with IFO 180 oil.
Emulsion seems not to start at all since bullwinkle_time is not to be set for this oil, and also the max water fraction seems to be defaulting to 0. It seems that this comes not from the JSON files, but rather from the adios_db toolbox, namely this module:
https://github.com/NOAA-ORR-ERD/adios_oil_database/blob/production/adios_db/adios_db/computation/gnome_oil.py#L175
As this specific oil seems to be of type "Residual Fuel Oil", apparently ADIOS code will always give a default water fraction of 0. I am not sure if this is a bug or a feature.
https://github.com/NOAA-ORR-ERD/noaa-oil-data/blob/production/data/oil/AD/AD01676.json#L11

In the below sample code, I manage to activate emulsion by setting the properties bullwinkle_time (onset of emulsion) and emulsion_water_fraction_max explicitly after seeding.
You can use this trick as a temporary workaround, but I am not sure how "safe" it is.
It might be better to ask NOOA about these oil specific tings, on https://github.com/NOAA-ORR-ERD/adios_oil_database/issues

from datetime import datetime, timedelta
from opendrift.models.openoil import OpenOil

o = OpenOil()
o.set_config('environment:constant', {'x_wind': 10, 'y_wind': 0, 'x_sea_water_velocity': 0, 'y_sea_water_velocity': 0})
o.seed_elements(lon=0, lat=0, time=datetime.now(), number=1000, oil_type='IFO 180')

o.oiltype.gnome_oil['bullwinkle_time'] = 0  # So that emulsion shall start immediately; default here was -999, meaning no emulsion
o.oiltype.gnome_oil['emulsion_water_fraction_max'] = 0.8  # default max water content was also 0 for IFO 180

o.run(duration=timedelta(hours=24))
o.plot_oil_budget('oilbudget.png')

[amercatini]

Thank you for the advice.
I'm going to use the workaround you suggested, but I will also contact NOAA to clarify if this behavior can be addressed within ADIOS.