Friday, March 2, 2018

Indeed, the Devil Is in the Details

Long-range model forecasts are physically plausible and easy to buy into, but one needs to be cautious about buying into a very narrow range of possible forecast outcomes.

On Wednesday, we discussed the warm-frontal feature passage that occurred yesterday, which most models generally agreed upon, but also some of the differences in the timing and placement of the frontal passage for today and tomorrow.

The warm-frontal feature came through as expected yesterday.  It wasn't a huge snow producer, but automated snow-depth sensors at Snowbasin-Boardwalk, Alta-Collins, and Sundance-Mid Mountain recorded 2, 2, and 6 inches respectively.  It's good to see our friends to the south get a little something.

This morning, also as anticipated, we are in the strong southerly flow ahead of a front that extends downstream from the Sierra Nevada into southwest Idaho.  This front boundary separates relatively dry, warm air that has circumscribed the southern high Sierra and traversed the southern Great Basin from cooler air that has moved across the northern Sierra and southern Cascades.

On Wednesday, the models diverged as to what would happen from here.  The 6Z GFS favored a solution in which the front pushed into northern Utah and gave us heavy precipitation late Friday and Friday night.  This was a beautiful and alluring solution if ever there was one but, while plausible, it does not appear that this will happen in the Salt Lake Valley and the central Wasatch.

In contrast, the 12Z NAM favored a solution in which the front stalled to our northwest, keeping us in the prefrontal southerly flow Friday and Friday evening, with the front retreating back to the northwest.  This is now looking closer to the likely outcome.

For example, here's the 13Z HRRR showing the surface wind shift reaching the Salt Lake City airport later this afternoon, but the precipitation remains to the north, even through 0700 UTC (Midnight MST).

Not all models bring the surface front all the way to Salt Lake City today.  The NAM, for example, keeps it to our north.

In addition, the frontal precipitation isn't in a hurry to get here and is likely to pivot northwestward tonight.  The 6Z NAM gives some precipitation to Alta tonight in the southerly to southwesterly flow, but the front doesn't come through until 11 PM Saturday night, as evinced very well by the wind direction forecast (green plusses) in the middle-right-column graph.

It's worth a look at the SREF plumes for Alta which show incredible spread in the timing and intensity of precipitation.  Look for periods where the lines slope steeply toward the top of the graph for periods of heavier precipitation.  Some members call for the heaviest precipitation tonight, others hold it off until, gulp, Sunday.  This largely reflects the sensitivity of the precipitation to the timing and position of the front.
Why all this uncertainty?  It has a lot to do with the characteristics of the large-scale flow, which has caused the upper-level trough to dig southeastward to the northwest coast, stall for a bit, and then finally progress eastward across the western U.S.  The stalling can be seen quite well in the loop below, which is a blend of analysis and forecast times (note the transition from radar imagery to model forecast precipitation as one transition from analysis to forecast times).

This has allowed the cold front to push to our doorstep where it sits this morning, but then wiggle around as the trough stalls along the coast and amplifies.  This amplification causes the flow to become more southerly and the front to retreat northward for a bit.  Finally, the trough and front progress eastward across our area.

These nuances are easy to diagnose in hindsight, but difficult to anticipate at long lead time.  The subtle variations make a huge difference in local weather.  For example, if the upper-level trough was just a bit further east, the front might have made it to us and pivoted over us, giving us the big dump later today and tonight.

Finally, a quick note that due to the frontal delay, the NWS has issued a high wind warning for Salt Lake and Tooele Counties. 


  1. Public television Nova broadcast has a new program called "prediction by the numbers" about frequentist and Bayesian probability. One program segment includes video of weather forecasting and the probabilistic nature of multiple weather models. Kurt

    1. That looks like a good show.

      One of the challenges of my profession (and it is not the only one) is that we need to communicate these uncertainties. Much to be learned in this area.