Forums (Forecast Builder)

Back

ProbRefreezeSleet AGL

RM
Ronald Miller, modified 7 Years ago.

ProbRefreezeSleet AGL

Youngling Posts: 6 Join Date: 9/24/12 Recent Posts

I'm looking at the NIC smartInits for the ProbRefreezeSleet grid.  I can see where it requires a depth of 2500' (762 m).  But if I'm reading the code correctly, the HeightFreezingLevel at that point is in MSL, not AGL.  I would think that this check should be for the depth of cold air in terms of AGL.  But I'm at the end of a midshift, so I could be completely missing something too.

That said, I find it a little strange that the 2500' criteria is a strict yes/no.  Since this is a probabilistic grid that only uses the Min Temp in the cold air, I would have thought that there would be some graduated scale on the cold air depth.  So if the depth was 1500' the ProbRefreezeSleet would be a lower (but non-zero) value than if the depth were 3000', if that makes sense.

Ron

AJ
Andy Just, modified 7 Years ago.

RE: ProbRefreezeSleet AGL

Youngling Posts: 89 Join Date: 6/2/15 Recent Posts

Ron,

  When developing the ProbRefreezeSleet grid about 4 years ago or so, when Dan and I looked at specifically refreeze cases from Roebber's 2001 data set, they all basically had a minimum depth threshold of 2500 ft.  However, as you approached -6 to -8C, the # of cases increased dramatically for refreezing.  Thus this is how the grid was developed, and why the probabilities are based solely on the MinT in the cold dome.  

  However, we have looked at another data set from Heather Reeves at NSSL and there are more cases with a minimum depth threshold going to as low as 1500-2000 ft.  So some adjustments to ProbRefreezeSleet are likely in the future. We just haven't had a chance to really adjust the grid. Additionally we want to be really cautious as the approach in place has worked pretty well over the past several years.