Nudging/clamping/radiation/gradient conditions for tracers

[artem_vb]

Hi, can somebody explain different open boundary conditions for tracers or point to corresponding documentation/forum thread?

1) What is the difference between "nudging" and "clamping"?

2) What is the difference between "gradient" and "radiation"?

3) Do tracer flags like EAST_TNUDGING cover sediments as well?

4) What are the nudging time scales?

Artem

P.S. Actually, after looking at https://www.myroms.org/wiki/index.php/B ... Conditions I found the answer for 2). Still, other 3 items remain mystery. Also, setting Tnudg in ocean.in apparently affects nudging time scale (whatever it is) for passive tracer only. For sediments it remains zero. How to set it to non-zero value for sediments? In ana_nudgcoef.h ?

[kate]

1. Clamping is nudging with an infinitesimally small timescale or nudging so hard it can't deviate from the suggested value.

The sediment nudging timescales are set in sediment.in.

[artem_vb]

Clamping is nudging with an infinitesimally small timescale or nudging so hard it can't deviate from the suggested value

Thank you, just to clarify this -- are following statements correct?

1) Nudging means that from time to time (within the time scale set by Tnudg parameter) boundary values are set to the specified values.

2) If nudging time scale is less than size of the 2D time step in ROMS nudging and clamping are equivalent?

Best wishes
Artem

[kate]

What is it that you want to do? If you want it clamped, then clamp it - there are fewer computations.

[artem_vb]

> What is it that you want to do?

First, I would like to understand how the 'nudge' behaves, so I am able to make a knowledgeable choice -- why one would use it instead of clamping?

As to the problem itself -- I am doing the simulation of the point source of contaminant (inside the domain) in the basin with wind and tidal forces. It looks, like radiation boundary condition does not give reasonable results, when 'cloud' reaches the boundary and then flow of water reverses due to tide or wind change. I have got an impression, that in this case boundary becomes the source of the matter (passive scalar or sediments) which was on the way out of the domain, when velocity had changed the sign. This is understandable, as simulator has no way to know what happened to the part of the 'cloud' outside the domain. I am not quite sure, why it apparently happens along only _part_ of the boundary -- my guess is, that this effect shows itself if there is a local maximum of the concentration on the boundary i.e. gradient of concentration at the boundary is directed inside the domain.

Anyway, setting boundary values to background concentration seems logical for the point source problem. Nudging and clamping apparently behave quite similarly -- hence the question.

[kate]

This is the boundary condition we always use: https://www.myroms.org/wiki/index.php/B ... _condition

The two timescales are a few days for incoming fluid and a year for outgoing fluid.