Short-Range Prediction with the HRRR

The newest model in modeling suite at the National Centers for Environmental Prediction is the High Resolution Rapid Refresh or HRRR (pronounced Hur, although you get bonus points if you can let the r drag on for a while).

With 3-km grid spacing and radar data assimilation, the HRRR is the first operational US modeling system capable of near-storm-scale prediction.

Here's an example from today.  At 1445 UTC (0845 MDT) an area of precipitation was located over northwest Utah and was moving eastward.

Here's the HRRR forecast from 1300 UTC (0700 MDT) through 1800 UTC (1200 MDT).  For the first part of the loop, I've overlaid the HRRR reflectivity on top of the observed.  After that, you see the HRRR forecast reflectivity, which calls for showers to be located over the Salt Lake Valley at 1700 UTC (1100 MDT).  That's about the time I walk across campus for a swim.  I'll keep an eye on the radar and adjust my plans if necessary.

Radar data assimilation is challenging over the mountain west due to poor radar coverage and constraints posed by topography.  Nevertheless, I have been using the HRRR the past few months in my forecasting class and for planning personal activities and have found it to be quite helpful.  It is perhaps best at the timing of arrival and departure of well-resolved and strongly forced fronts and precipitation systems.  It does less well with precipitation intensity.  For instance, I'm not sure one can count on the precipitation system to decay and for the northern Wasatch Front to be rain free as advertised above by the HRRR.  I'd probably go with showers moving into that area later this morning based on the recent radar loop.

I was hoping to use the HRRR more extensively this past winter for short-range snowfall forecasting in the mountains, but Mother Nature wasn't very cooperative.  We'll have to check it out next winter when hopefully we see more snow.