"Noisy" or erroneous VV fields produced by models

[Bill Schintler]

I’m looking at the NAM and GFS forecasts initializing at 00Z, April 20, 2005; and I’ve noticed the excessively “noisy†vertical velocity fields (700 and 500mb) produced by both the 24 and 30hr FCST (valid 00Z and 06Z, 04/21/05) in the western NE, and southwestern SD areas… This is apparently in association with a series of vorticity maxima approaching the area ahead of the trough in the west. I’m using this date and time as a specific example, although I’ve observed this phenomena in the past. I do not have a background in NWP, but I can assume these noisy fields are a result of weird data being ingested into the models that is causing the solution to “blow up†mathematically…? I can appreciate the fact that the VV fields are probably very sensitive to small perturbations in the data.

With this post, I’m hoping to entertain some discussion about two things:

1) Usefulness of “reading into†model vertical velocity fields when making a convective forecast. From personal forecasting experience (I’m an amateur – I have no background in meteorology), there have been many times when convection is accurately predicted by VV fields, while other times it makes no sense to take more then a casual glance at the forecast VV fields. I've noticed this mostly with short term RUC forecasts, but as this example points out, other models are prone to this.

2) Generally, what conditions tend to cause VV fields to behave strangely. I can assume that poorly sampled (temporally or spatially) vorticity advection would cause large errors in the vertical velocity fields.

Thanks in advance,
- bill

 

[Blake Michaleski]

IMO it's not really noise but a true bullseye of peak VV. one thing to remember, non-convective vertical velocity is measured in microbars per second. and an increase in one (1) yields an other color on the scale. One microbal is not much at all and an area of significant lifting can have a huge gradient in the VV field, there by appearing noisy.

To back this up as a legitimate peak we can look at our 24hr 250mb height and isotach analysis and by applying the ageostrophic wind equation, it is obvious that western NE will be in a very favorable environment for upward vertical motion.

Here is a screen capture of the 24hr forecast of the 500mb VV

[Broken External Image]:http://www.lockitupoffroad.com/gumbo/vv500mb.jpg

And of the 250mb Heights/Isotachs

[Broken External Image]:http://www.lockitupoffroad.com/gumbo/vag250mb.jpg

Besides the obvious diffluence thanks to the severely negatively tilted trough, there is a very significant exit region to the jet and thereby applying the speed divergence term of the Vag eqn, you will get an ageostphic Vector (Vag1 above) to the right of the mean flow removing mass in teh western Nebraska area. By applying the curvature term you end up with Vector Vag2 above in the negative direction of the flow again removing mass over W NE. This is the same term that often times gives sub-geostrophic flow in a trough axis. In this case the height gradient is rather tight and the subgeosptophic flow is hard to pick out but be assured the curvature term is at work in the region of cyclonic curvature. The same term is responsible for the Vector Vag3 above in the region of anticyclonic curvature. This yields an Ageostrophic vector in the positive direction of the flow.

So as you can see there is enough evidence to support what the 500 mb vv chat is showing. In fact, the equation I mentioned, along with the Q-G omega equation, are what goes into that product.


edit: excuse the typos/poor grammar/spelling

it;s 6 am and i havent been to bed yet