About wind_drift_factor

[Erispoe29]

Hi all,

I am trying to set up an oil spill scenario and understand how OpenOil takes wind and currents into account when calculating drift.
My attention was especially caught by the wind_drift_factor.

I understood from documentation that : u_drift = u_currents + wind_drift_factorAu10_atm
where A is a matrix taking into account deflection angle (Reed et al. 1994)

So I placed myself in two ideal cases :

• Case 1
  o = OpenOil(loglevel=20)
  o.set_config('environment:fallback:x_sea_water_velocity', 0)
  o.set_config('environment:fallback:y_sea_water_velocity', 0.1)
  o.set_config('environment:fallback:x_wind', 0)
  o.set_config('environment:fallback:y_wind', -1)

• Case 2
  o = OpenOil(loglevel=20)
  o.set_config('environment:fallback:x_sea_water_velocity', 0)
  o.set_config('environment:fallback:y_sea_water_velocity', 0.1)
  o.set_config('environment:fallback:x_wind', 0)
  o.set_config('environment:fallback:y_wind', -10)

To make sure I didn’t add any unwanted side effects, I added :

• o.set_config("drift:current_uncertainty", 0)
• o.set_config("drift:wind_uncertainty", 0)

I release 1000 oil particles at a single point and time and observe how they behave in open waters.

First picture is the result of the case 1, the second picture is the result of the case 2.
Why do all particles are not moving the same distance in the second case ?
Is there a critical ratio between wind and current that causes particles to drift in different directions ?

Thank you very much for your help.

[knutfrode]

Hi,

In the second case the wind speed is 10 m/s, and hence there will be wave breaking and some entrainment of oil.
So the elements that get entrained (parameterisation with random numbers) will move northwards with current as they are sheletered from windage, whereas surface elements will move southwards, since wdf*10 > 0.1 m/s.

You can see this by either plotting the depth vs time:

o.result.z.plot.line(x='time')
plt.show()

or make an animation with elements colored by depth:
o.animation(color='z')

or by plotting the oil budget
o.plot_oil_budget()

Btw, I recommend using constant instead of fallbackfor the config. Constant always take predominance, whereas fallback will be used if all other readers fail (though in your case there are no readers)

And also: OpenOil does not use a deflection angle as in Reed94.

[Erispoe29]

Thank you for your answer, I better understand the oil behavior now !

In my realistic scenario, I can observe same behavior as in idealised case when using Croco surface outputs.
After a fast check, the particles are going to different depths (so they have different paths).
Which currents data are then used to move oil particles ?

[knutfrode]

The Croco currents are interpolated to the element positions, and are used for advection (if I understand your question correctly).

[Erispoe29]

Sorry, I wasn't precise enough.
The Croco file used gives only surface currents. For particles that are entrained under the surface, how the currents are calculated at their positions ? If it is simply by interpolating, which kind of interpolation is it ?

[knutfrode]

Ok, if only surface currents are available, these will be used at all depths (i.e. barotropic).

If 3D currents are available, they will be interpolated bilinearly horizontally and linearly vertically and in time.

[Erispoe29]

Thank you very much for explanations !