NSSL WRF

[Robert Dewey]

NSSL runs a WRF run at 4KM resolution for most of the US. I've been watching it over the past few storm systems, and I must say that it's accuracy is beyond impressive, especially the simulated radar in strong synoptic lift regions, as well as identifying areas of potentially severe weather (obviously it doesn't nail the storms with great accuracy, but it lets you know the general area and strength of convection). It's only run at 00Z and takes a long time to complete (6-10 hours).

Here it is: http://www.nssl.noaa.gov/wrf/

 

[Michael P. Morris]

I also like this one, run by NCEP every 12 hours. Very similar to the NSSL WRF in that it is good at predicting storm mode and timing of initiation, though it did not predict the development of the southern Panhandle storms last weekend.

http://www.emc.ncep.noaa.gov/mmb/mmbpll/cent4km/v2/

 

[beaudodson]

NSSL runs a WRF run at 4KM resolution for most of the US. I've been watching it over the past few storm systems, and I must say that it's accuracy is beyond impressive, especially the simulated radar in strong synoptic lift regions, as well as identifying areas of potentially severe weather (obviously it doesn't nail the storms with great accuracy, but it lets you know the general area and strength of convection). It's only run at 00Z and takes a long time to complete (6-10 hours).

Here it is: http://www.nssl.noaa.gov/wrf/

Is it updated four times a day on their site or just once per day?

 

[Jeff Snyder]

I also like this one, run by NCEP every 12 hours. Very similar to the NSSL WRF in that it is good at predicting storm mode and timing of initiation, though it did not predict the development of the southern Panhandle storms last weekend.

http://www.emc.ncep.noaa.gov/mmb/mmbpll/cent4km/v2/

I have found the 4.5km explicit convection WRF (the guidance from which is available in the link Michael gave, despite the fact that it says it's the 4.0km run) to be quite good in forecasting storm mode. I don't expect individual storms to be forecast well ~20-24hrs in advance, but I have been pleased with that model's ability to forecast storm mode.

But, I think that model is only run every 24 hours (starting at 00z), with output available around midnight CSD. I don't think I've ever seen 12z output from that model, but I could be wrong.

 

[rdale]

Is it updated four times a day on their site or just once per day?

Last line in his post:

"It's only run at 00Z and takes a long time to complete (6-10 hours)."

 

[afischer]

I have found the 4.5km explicit convection WRF (the guidance from which is available in the link Michael gave, despite the fact that it says it's the 4.0km run) to be quite good in forecasting storm mode. I don't expect individual storms to be forecast well ~20-24hrs in advance, but I have been pleased with that model's ability to forecast storm mode.

I've seen it do well in some cases and not so well in others w.r.t. convective mode. I recall that 24 hours out, it forecast linear convection for the Aug 24 SD event and the Sep 16 SD event. But you've probably looked at it more than I have. I still see it still a valuable "tool" alongside everything else we look at it.