Wind speeds very close to the surface

[Claus]

Let me set the caveat first off that I'm an acoustician, not a meteorologist, so it's possible that some of my questions may be naive.

My research group (and the outdoor acoustics community in general) has been using WRF with great success for several years now to generate high-res (by our standards anyway) atmospheric models for low-frequency sound propagation studies at short ranges (<200 km). We work at wavelengths on the order of hundreds to thousands of meters, so fine structure tends to have little effect and we're generally happy with ~5km horizontal resolution and vertical resolution on the order of hundreds of meters. We recently have begun looking at sound propagation over the ocean and have found that wind shear can be very strong in the first 10 meters, with noticeable changes in both speed and direction over that distance. This implies that sound ducting over water will vary in ways that a coarsely-gridded atmosphere will not capture; ideally we would be able to capture multiple windspeed and temperature results in the lowest 10 meters.

1. Is WRF an appropriate tool to obtain multiple vertical points at less than 10 meters height? We're only looking at propagation over water so topography isn't a concern. I can obviously set the eta output levels however I want, but if the physics doesn't support it then that doesn't matter.
2. If so, which physics modules would be appropriate, and what other namelist settings do I need to be careful of?
3. If not, is there a publicly-available model that would be better for this use case?

Thank you in advance for your consideration.

-Claus Hetzer
National Center for Physical Acoustics
The University of Mississippi

 

[kwerner]

Hi Claus,

Apologies for the delay. I needed to refer to a colleague for input on this one. They didn't specifically answer each of your questions, but they sent this, which I believe is a response to your first question:

I don't think the direction should change much below 10 m, but a good speed approximation below that is a log profile. u(z) goes in proportion to log(z/znt) where znt is the roughness length that should be in the output. Given u10 is proportional to log(10/znt), other heights below that can be estimated.
u(z)/u10 = log(z/znt) / log(10/znt).
Similarly for v(z).

2. I'm not aware of specific physics modules that would be better or worse for this, and if you've been running WRF with reasonable results for a while, it's likely the physics options you're using are good. In case you haven't already seen it, you can look at specifics (with links to references) on the physics options.

3. The MPAS model may be an option if WRF isn't resolving what you need.

 

[Claus]

Thank you for the answer, and for the link to the physics descriptions. I will try out the log scaling and see how that matches up with our acoustic observations.

-Claus Hetzer