Is there an advantage to not using anti-aliasing?

[Harrison Cater]

In GR products(or any radar program that allows it for that matter) is there any practical reason to disable velocity anti-aliasing? I see some people prefer their data that way but I can't figure out why or particular advantages to doing it that way.

 

[rdale]

Only if there's some massive error in the dealising. Otherwise you want it on all the time.

 

[Mark Blue]

It is a software technique used to smooth out jaggity or boxy appearing lines. In a radar program I have it enabled since I prefer the smooth look over the boxy look of the radar returns.

 

[rdale]

Mark - I think you are referring to the "smoothing" button in reflectivity? He's talking about velocity dealiasing which "fixes" Level II data in high wind situations. That was a big issue with RadarScope during Florence because it couldn't recognize the strong winds. There's no reason I can think of to turn that off.

 

[Harrison Cater]

Yea that's what I was thinking... I only asked because I occasionally see some credible individuals running with anti-aliasing off. Just wondered if they gained anything from doing that.

 

[rdale]

Their credibility should drop a notch or two

 

[Mark Blue]

My bad on this one with confusing it with smoothing. I don’t use GRAE as much as I used to because of mobile trends and I don’t use my laptop to chase anymore, so only when I armchair chase. There is scant information to be found on this option. Most of what I did find was on the owners forum and it pertains to wind profiles during hurricanes. I suppose there’s more somewhere, perhaps in the WSR-88D manual, but it would be interesting to find some basic written documentation to refer to.

 

[rdale]

No problem. Here you go in this note from Haby:

Suppose the maximum unambiguous velocity is 30 m/s. If the wind is blowing at 35 m/s, the radar will show the 35 m/s wind as a 5 m/s wind instead. All wind speeds above 30 m/s will be aliased in this case. Computer algorithms can correct this aliasing or increasing the PRF can correct this aliasing.

More detail at https://training.weather.gov/wdtd/courses/rac/principles/dealiasing/presentation_html5.html

 

[Jeff Duda]

THe only times I ever turn off dealiasing is, like what Rob said, when the dealiasing algorithm is clearly failing in an area of interest. Such situations include in winter storms with high wind speeds aloft and the local NWS radar meteorologist has set the radar to VCP 31 or 32, because the maximum ambiguous velocity is so low that huge chunks of the velocity field are improperly dealiased. This usually appears as sudden shifts in velocity sign between adjacent radials for a large along-radial distance, so that there are large circular-sector areas with obviously different velocity.

Another situation is in tropical cyclones. The larger-scale flow is quite predictable, so when you see a velocity signature that doesn't resemble a large rankine vortex, it's pretty obvious the dealiasing algorithm is failing.

The last situation is in/near the mesocyclone on intense or tornadic supercells. If the storm is close enough to the radar that it is able to start sampling the low-level meso or the tornado itself, the dealiasing algorithm can fail because it might legitimately not know if velocities are folded or not (or folded twice or just once). In these situations, however, it may not be obvious how to correct the aliased velocities, so even turning off the dealiasing algorithm may not help. I analyzed a really prominent example of this - from the 31 May 2013 El Reno beast - in a radar document I made a few years ago: http://www.caps.ou.edu/~jdduda/how_to_read_and_interpret_weather_radar_advanced_guide.pdf (start with page 21).