RUC vs. GFS: CAPE

[John Hudson]

The air here in southern Manitoba's forecast to get real juicy here tomorrow, and I was entertaining the possibility of a local chase.

Looked at the Sat. 6Z RUC just for the fun of it, as I know it only goes out 12 hours, and it places an area of extreme CAPE (4353 j/kg) over western Manitoba. Meanwhile, the GFS for the same time frame forecasts only 1500.

Why the huge difference?


John
VE4 JTH

 

[Mikey Gribble]

I didn't look at the forecast, but I would bet the reason for the vast difference in CAPE between the two models has to do with differences in one or more of the following - moisture, temperature, and lapse rates. Those are what typically influence the amount of instability so I would look for differences in those three things between the two models if you want to pinpoint the reason for the difference in forecasted CAPE.

 

[afischer]

It looks like the RUC model was forecasting sfc dewpoints in the low 70s, while the GFS was forecasting sfc dewpoints in the mid 60s. Also, the GFS had a shallow surface-based inversion which reduced the CAPE even further. When CAPE is calculated using strictly a surface parcel (which I assume was being done on the website you were using), this can account for large differences between models since the computed value is so strongly dependent upon the sfc T & Td. The RUC has been known to forecast enormous values of surface CAPE at times, and this can actually happen under a few differing meteorological set-ups... so I tend to use it with caution.

100 mb mixed-layer (ML) CAPE accounts for vertical mixing effects in the boundary layer and has generally been shown to be more "realistic." Today, the 18Z RUC had MLCAPE forecast around 2000 j/kg by 03-06Z in the area with the 4000-5000 surface-based CAPE max... along with stronger CINH. This is because the forecast rich boundary layer moisture from the RUC model was very shallow.

Earl Barker's site has the 100 mb MLCAPE and MLCINH forecasts from the NAM and the RUC, which is what I like to use. Then there's the question of whether to apply the virtual temperature correction... that's a whole other issue by itself!