Back to winter...
Posted January 3, 2011
After the third coldest December on record at RDU (and a tie for coldest ever at Wilmington), we had quite the turn to milder weather to close out December and kick off the new year. At RDU, temperatures maxed out at 61, 63 and 62 degrees on Friday, Saturday and Sunday, all of which were more than ten degrees above normal, though well below records for those days. A cold front that helped bring some rain over the weekend, however, passed across the area during the early to mid afternoon hours yesterday, with our temperature falling from 62 to 52 at the airport between 1 and 3 pm. Even more dramatically, in the wake of the front we found ourselves at 7 am this morning with clear skies and a temperature 31 degrees colder than the same time Sunday morning!
You might recall that back in the fall, the Climate Prediction Center forecast a winter that would have a slightly better chance of being warmer and drier than normal for our region than being near normal or being colder and wetter than normal. I pointed out in an earlier blog that La Nina winters can be kind of volatile, so that confidence in such a forecast had to be tempered by the small differences in the probabilities of each category. Well, given the cold December, and the frequent negative phase of another climate pattern called the North Atlantic Oscillation, which tends to produce below normal temperatures in the eastern U.S., the CPC has revised their outlooks a bit, to favor cooler temperatures and an equal chance of above, below, or near normal precipitation for our area, as you will see in the first two images included here.
For the rest of this week, it appears we'll stay on the chilly side, with some potential for very cold temperatures (highs in the upper 30s to low 40s) toward the end of the week and beginning of the weekend. So far, a good chance of precipitation is hard to discern, and we have a pretty good chance of being mainly dry as that invasion of cold air occurs. There are some hints from the GFS (American) and ECMWF (European) forecast models that an upper level trough and surface low pushing a front through the area on Friday could trigger a few flurries or sprinkles, but when we look at groups of multiple model runs based on these same systems, the likelihood of measurable precipitation with that system so far looks quite small. The models further bring a surface low across and out of the Gulf of Mexico early next week that might have some potential to produce precipitation, but it is very hard to have any confidence about that system yet.
This low ability to have confidence in what otherwise looks like an active low is based both on how far out in time it is (8-10 days away) and also how ensemble forecasts area handling the system. As a reminder, an ensemble approach means running a given computer model many times with slight "perturbations" of the initial conditions and/or of some of the equations used in the models, to simulate the effects of imperfect observations of the atmosphere and of less than perfect assumptions and simplifications that have to be made in order to model something as complex as worldwide weather. To give you a sense of how this approach can be helpful in assigning confidence to a forecast, I included a couple of examples using the ECMWF ensemble prediction system, initialized last evening.
First, compare the left and right panels in the 3rd image above. They are sea level pressure forecasts showing the ensemble mean and deterministic patterns, respectively, 48 hours from the start time, in this case valid Tuesday evening Jan 4th at 7 pm. What is noticeable here is that the pattern, including the weak ridge of high pressure across the central and southeastern U.S., and the low pressure areas near the upper left and upper right corners of the image, looks very much the same on both panels. This indicates that the average of the 50 ensemble model runs is very close to that of the single deterministic run, giving some confidence that it is probably on target.
Contrast this with the panels in the 4th and final image. These are the same as before, except valid at 7 pm on Tue Jan 11th. This time, notice the strong low pressure center located just off the Virginia coast on the deterministic run. This low has moved into that position from the south/southwest, in a way that at least raises the possibility of some wintry precipitation for parts of our region. However, it's also worth checking the ensemble mean on the left panel. Unlike the one at 48 hours in the future, this one, at 216 hours out, shows significant differences from the deterministic run. There is some hint of low pressure northeast of us, but the very weak nature of that system is reflective of the variety of intensities and locations that resulted from those 50 model runs, leaving us with little confidence as to whether the low will form at all, whether it will pass too far southeast to produce precipitation around here, or will pass far enough west to produce just rain, or will end up in fact giving us a round of winter weather. It'll be interesting to watch for this feature, and how it evolves, as we head through the next week...