GOOD RIDDANCE NAM/ETA!!!!

After much gnashing of teeth and pulling hair this chase season, the horrible and pitiful ETA "engine" driving the NAM output is retired. Good riddance. When you can't even halfway rely on a model even 12 hours out, it's a serious problem. I can't even begin to count all of the setups that it totally missed this year. It wasn't just moisture problems but upper air and surface features as well. I quickly got to the point that by late April and early May, the NAM model was last on my forecasting list with the RUC (for Day 1 setups) and GFS ahead of it. Yep....I'm glad to see it go away.

In it's place, and as I noticed today on the famous UCAR models, the NAM is running with the WRF (Warf...as in the Klingon) "engine". In a nutshell, it is sure to be a major improvement. Alot of the test phase results I've read about are certainly promising. For a nice technical multimedia presentation targeted for operational forecasters, click here and click on "Begin" on the right. Some of it might be a tad too technical for some, but there is alot of "meat" to it. I strongly recommend it. In the future, I'll post other sources as I can find them. I consider this to be an extremely important subject for stormchasing since model performance is directly related to the costs incurred...not to mention the time involved along with commitment schedules. The old NAM/ETA output has cost chasers ALOT of money and sanity this season. And, of course, it certainly adversely affects those more concerned with the competitive/commercial side of stormchasing. :-)

Here is my plain english interpretation/highlights of the important changes and differences (I appreciate any corrections and additions in the blog "comments" section).

1) MUCH better analysis/initialization...similar to GFS. Alot more use of satellite derived data will be ingested. This allows a smaller grid initialization as well as covering gaping holes. I'm not sure about the all of the types of satellite data being used, but what I'm aware of is that temperatures and moisture will be part of that. I'm not clear if that includes all levels of the atmosphere or just the surface. I have to think that it is the entire atmosphere such as GOES soundings. I'm also not clear if satellite derived winds will be used or not. NEXRAD L2 winds (I assume the VAD profiler data) will be used as well. Speaking of moisture, soil moisture will now be used and the gross evapotranspiration bias of the ETA engine (which caused most of the problems for us this year) is now corrected. In general, the WRF engine will entail better moisture analysis and forecast of GOM return flow (YEAH!!!!). The other thing I like is that the WRF will use 3 hour analysis corrections of the "first guess" forecast output. This should result in much better resolution and handling of surface features closer to the mesoscale level. The mesoscale parameters should be more sharp and accurate. (YEAH!!!!). There will certainly be alot of performing the "Forbidden Sacred Dance of Chaser Merriment" by yours truly.

2) For the "upslope" and mountain terrain fans, the WRF is much bettter in this area. Better handling of moisture and mountain wave/flow dynamics will surely be noticed. Higher terrain forecast should be improved including windflow. They refer to this as "improved terrain flowing coordinates". I would think this would directly impact leeside troughing and cyclogenesis. I'll be keeping a close eye on this aspect.

3) Since the WRF will be implementing the NMM (Non-hydrostatic Mesoscale Model), certain algorithms come into play that handle vertical velocities and forcing MUCH better and in much finer detail/resolution. This will result in better forecasts of UVV parameters. This should directly correlate to improved convective initiation and helping predict if the cap can be busted....and the amount of linear forcing. Something else I'll be paying attention to as well.

4) Better handling of sea surface temperatures. This would seem to have a slight impact on the GOM return flow and moisture quality, but mostly for tropical systems I would think. Speaking of which, the WRF has demonstrated a signifcant improvment in handling tropical systems.

5) For model output sources that implement it, simulated NEXRAD reflectivities will be available.

Some biasis are noted though. One of which is that the WRF might show a slight dry bias with time as the forecast progresses. This is in part to the revised evapotranspiration algorithm. Also of note is that the WRF deepens troughs too much. I want to find out more information about this myself as this is a critical parameter. How much is "too much" for instance. When does it occur? Lost of questions there. Some of the convective triggering mechanisms are the same as the ETA. This might not be a bad thing if the model now has better moisture and UVV forecasts to work with. The WRF will still likely struggle with the shallow arctic airmasses in winter as the ETA did.

So, in summary, my take is that it will be a vast improvement over the ETA driven NAM. It will resemble the RUC to a degree. Of course, a Ouiji board, magic 8-ball and a crystal ball would be an improvement over the ETA engine. :-) We should have alot of time over the next 9 months to get familiar with it. Hopefully it will reduce ALOT of frustration next year for everybody.