NAM/GFS vs. UKMET/ECWMF and Low-level CAPE

[Glen Romine]

It's amazing how the NAM continues to trend toward the consistent UKMET/ECMWF. This goes to show how poor the NAM is beyond 36 hours. I rarely the NAM for anything beyond 36 hours.... or just use it's thermodynamic forecasts and mentally shift it towards the UKMET surface features..

Mike U

In fairness - with systems moving as quickly as they are this time of year combined with a zonal flow pattern - you have to think of long-range ETA as nothing more than a high resolution GFS - because all that you are seeing after 36 hours is what came in on the western boundary condition, which comes from the GFS forecast. NAM is only adding some better physics and terrain, not a whole lot more. Once the system is within the observing network - the NAM can finally start to get a better solution, and with this system this only became true this morning with just the nose of the upper jet sampled at VBG.

[mets only note]Also, in fairness the ideas suggested about the importance of low-level buoyancy were actually brought up by several I believe long before Jon Davies. I think of the work by McCaul and Weisman 1996 in particular. I also love to look at the 0-3 km CAPE for signs of potential tornadic activity - but I don't know how much I would trust a model forecast of it ~33 hours out.[/mets only note]

Glen

 

[Mike Umscheid]

In fairness - with systems moving as quickly as they are this time of year combined with a zonal flow pattern - you have to think of long-range ETA as nothing more than a high resolution GFS - because all that you are seeing after 36 hours is what came in on the western boundary condition, which comes from the GFS forecast. NAM is only adding some better physics and terrain, not a whole lot more. Once the system is within the observing network - the NAM can finally start to get a better solution, and with this system this only became true this morning with just the nose of the upper jet sampled at VBG.

You make very good points, Glen. I would respond by saying, then, that in situations involving robust mid-latitude baroclinic development, the European models UKMET and ECMWF have an edge over the NCEP GFS. I only say this based on my every day experience using these global spectral NWP. So, then, if you initialize the NAM boundary with GFS, then you are simply getting, what you said, a GFS-induced NAM forecast. The NAM is clearly different than the GFS, and it's boundary layer forecasts are generally good (aside from shallow cold airmasses). The NAM upper tropospheric predictions, however, are left to be desired for. Again, experience. All I can say is this.. if there is a dichotomy between the NAM and UKMET at 36 hrs and beyond, then I need some serious convincing that the NAM is the better prediction based solely on performance of the model alone. Within 36 hours, however, the NAM forecasts are generally "good enough" most of the time in my forecast process at least.

[mets only note]Also, in fairness the ideas suggested about the importance of low-level buoyancy were actually brought up by several I believe long before Jon Davies. I think of the work by McCaul and Weisman 1996 in particular. I also love to look at the 0-3 km CAPE for signs of potential tornadic activity - but I don't know how much I would trust a model forecast of it ~33 hours out.[/mets only note]

Jon Davies work certainly builds upon former work done on the topic, however, Jon has been the most aggressive in recent years research regarding the importance of background low level thermodynamics and tornadogenesis. It is not just 0-3km CAPE, and I absolutely *don't* believe that this is the only parameter that is important in low level stretching. This is why model sounding interrogation/modifications are extremely important. However, when models suggest an area of potential for such low level thermodynamics, combined with pattern recognition and everything else, then it is a strong signal in my book. Good discussion..

Mike U