I like our chances of picking up several inches of mountain snow in the Cottonwoods later today, tonight, and early tomorrow as a cold, "juicy" northwesterly flow develops over northern Utah. Perhaps we'll do better if the lake effect can kick in with the right orientation. If we can get high precipitation rates, you might even see some of the white stuff on the benches.
Here are some thoughts on lake effect that are relevant for tonight's forecast. First, lake effect is relatively rare in September. During the latter half of the month, you can expect an event on average every 4-5 years. Events are more common later in the fall when there is a higher frequency of trough passages and cold air intrusions.
Although events are scarce this time of year they do happen. In addition, if we look at the environmental conditions in which lake effect occurs throughout the cool season, we find that during the fall (and spring), one generally needs a much larger temperature difference between the lake and the overlying airmass [we usually use the 700-mb (about 10,000 ft) temperature as indicative of the overlying airmass] to generate lake effect. The graph below is fairly busy, but the red box-and-whiskers show the range of lake-to-700-mb temperature differences during lake-effect periods and illustrate fall and spring events generally require larger differences than events in the
spring and fall winter.
The dashed red line reflects our best guess of the minimum lake-to-700-mb temperature difference needed to generate lake-effect in any given month. In September, it's about 21ºC. The good news is that we'll be beyond that tonight. Lake temperatures as of yesterday were around 20ºC, whereas the 700-mb temperature is forecast to be -5 to -7ºC (depending on what model you look at) late tonight and early tomorrow morning. Even though the lake will cool some over the next several hours, we should be above the necessary (but not sufficient) threshold.
The problem is that that threshold has been met before in September without producing lake-effect (the black box and whiskers show differentials without lake-effect, confirming this). This is because there are a number of additional factors at play, one of which is the relative humidity. It's simply harder (or impossible) to get lake effect from a dry low-level environment. If we look at all lake-effect periods, we find that the likelihood of lake-effect increases with the size of the lake-to-700-mb temperature difference (relative to the dashed red line above, something we call ΔT excess, or how far you are above the magic threshold) and the low-level (850-700 mb) relative humidity.
Of course, there are still other variables at play, such the time of day. There is a strong diurnal modulation of lake-effect with events more common at night and especially just before and around sunrise and less common in the afternoon. This effect is more pronounced in the Fall and Spring when solar heating is stronger than it is in the winter.
So if you put all this scentific mumble jumble together, what do you have. Well, we have several key ingredients coming together at that optimal time (late night and early morning): Big lake-to-700-mb temperature difference, high relative humidities, and flow from the west-northwest to northwest. We can extract information about the environment forecast by the computer models and estimate the likelihood of lake-effect and you get the numbers below from the GFS and NAM.
So, the NAM is positively giddy, the GFS a bit more subdued. In the case of the GFS, it's simply a little warmer and drier, so the odds are reduced a bit. Note that those are instantaneous probabilities. If you were to consider the odds of lake effect at any time during the late night or early morning, it would be higher than the maximum values provided in the tables above.
Put all this together, and you have a good chance of lake effect. Whether or not it goes and goes big is something we really can't forecast reliably. In addition, even if it goes, there's the issue of where and how much. This is why I don't bet the house on the Dreaded Lake Effect.
Dr. Rumsfeld, er I mean Jim, that's one hell of a great post on the DLE. So seeing that these are all things that we know that we know (we think), what are the things that we don't know that we don't know that could be the fly in the ointment? I'm guessing you don't know...lol
ReplyDeletePem
With such a relatively warm trough, the dendritic growth zone will be between 550 and 650 hPa. Looking at the GFS sounding for 12Z on Friday, your saturation water vapor mixing ratio at the surface is ~5 g/kg and falls to ~3 g/kg at 650 hPa in a convective updraft, so you could produce over 1 g/m^3 of cloud water before hitting the dendritic growth zone. It seems like this is a better setup for a lot of graupel production than more typical colder troughs later in the season. If that is true, then that would effect fallout, i.e. graupel falls faster than snow and would not travel as far downstream, and thus effect the distribution of precipitation. Do you have any thoughts on the dependence of hydrometeor type and precipitation distribution on let's say 700 hPa temperature?
ReplyDeleteAdam:
DeleteThis is an interesting aspect of the forecast problem that might be important, especially for snowfall amount forecasting. Let's see what happens - maybe the MASC is in operation at Alta, but I doubt it. You'll need to go up and take some obs yourself....
Jim
Definitely a good amount of graupel in the band this morning with some 40-50 dBZ echoes and lightning strikes over the lake and in northern Tooele County, but there is still a decent amount of snow production as well judging from fallout farther downstream over Utah County.
DeleteWe looked at this with WRF simulations of the 27 Oct 2010 event. It was a colder event, much less likely to be producing a lot of graupel. In that case, most of the snow growth was happening over the lake or near the shore. Due to slow fall speeds for dendrites and moderate NW winds, when the snow finally reached the ground it was over the Wasatch crest. In today's case, I suspect that the highest snowfall totals will be much closer to the shore (where we have seen 50+ dBZ echoes).
ReplyDeleteDry Fork SNOTEL (7160 feet) has been under the gun most of the morning. It's showing 7" since 5 AM, with 0.9" of SWE.
Echo tops in this storm are over 20,000 ft, so the storm extends through the dendritic growth zone. The storm depth, decent NW flow, and modest fall dendritic fall speeds (~1 m/s) are all consistent with the downstream band extension.
DeleteYeah, I think there was a lot of graupel near shore because snow is not going to give you 45-50 dBZ. This is also probably why "snow" was making it to the ground not completely melted near that location, but the updrafts probably aren't strong enough or large enough to hold it up very long, so with it falling out, dendrites are allowed to grow and fall downstream. The amount of snow falling downstream is probably negatively affected by the consumption and removal of water by graupel though.
Delete