Wednesday, November 26th 2014
I am always digging into the past to make a better forecast for the future. Simply put, that is what using analogs is all about: finding past years that have many similarities to the current year, and using the weather that occurred then, to make a projection into the future. In a previous article I listed the following years as likely analogs: 1957-58, 1976-77, 1986-87, 2002-03, and 2009-10. Recently, I have been looking for additional years and came across 1941-42. Lets look at the what the oceans were doing at that time:
Several things about it stand out about the image above.
PDO Phase: Strongly positive as it is now
AMO Phase: Strongly positive as it is now
ENSO Phase: El Nino in progress as it is now
While the El Nino that was occurring in the fall of 1941 was better organized than the weak El Nino that is present now, we will likely continue to see the current El Nino strengthen a bit. Here is a look at the current sea surface temperature structure:
What did the average sea surface temperature anomaly look like from November 1941 through May 1942? See below:
Here is a fresh look at what the JAMSTEC model has for a sea surface temperature anomaly forecast through May:
To its credit, the JAMSTEC has not changed the sea surface temperature anomaly outlook. The sea surface temperature anomaly forecast from the JAMSTEC looks very similar to that of what occurred from November 1941 through May of 1942. The most interesting part is that the warmest water in the fall of 1941 was centered from the west coast of South America westward through ENSO regions. However, that warm pool eventually reorganized into more of a hybrid or Central Pacific based El Nino. Likely not quite meeting Modoki criteria, because of the lack of cooling immediately off the west coast of South America. See the differences between traditional and Modoki El Nino below:
So what was the winter/spring of 1941-42 like for Denver? Lets take a look…
Looks pretty normal in terms of snowfall for Denver, but it was certainly distributed in a different way. Very little snowfall through November, then things started to take off with the most snow in February. Nearly 20” of snow in February is pretty unusual as the monthly normal is only 7.5” Bottomline, December through February saw 37.8” of snow, which is 15.5” above normal for that 90 day period. Here’s what occurred nationally during that time in terms of precipitation and temperature:
Overall, Colorado was cooler and wetter than normal from November 1941 through May of 1942. This doesn’t mean that this type of snowfall/temperature or monthly distribution will occur, I am just showing you what happened during that particular time.
New JAMSTEC Model info…
Temperature Outlook ( Blue = Cooler Than Normal Red = Warmer Than Normal White = Normal )
December - Feburary
March - May
Precipitation Outlook ( Green = Wetter Than Normal Brown = Drier Than Normal White = Normal )
December - February
March - May
Based on the new JAMSTEC model, one could infer that Denver would have normal snowfall with above normal temperatures for December through February. March through May would then have normal to slightly above normal moisture, with cooler than normal temperatures. This still fits with our thinking that the best snowfall and coldest temperatures for Denver will likely favor the latter part of winter into early spring. Overall, the forecast doesn’t look to change much, except maybe warming things up a bit during the first half of winter.
So, I am sure you are asking, “which analog fits us the best?” That is a tough question, as there are so many different variables. However, I like the analogs that best fit what the oceans are doing right now. 1) Positive/warm PDO 2) Positive/warm AMO 3) Weak, but likely to intensify a bit El Nino. And another oscillation that you have heard me talk about before is the MJO, which is fixing to get increasingly active. All of that adds up to my liking the analogs from best to least in this order:
Please reference my previous article to see how those years stacked up. Remember this isn’t an exact science, but is used to provide a roadmap to where we think we are going. I do think the first four on that list are all very close, with the latter two distancing themselves somewhat but not a great deal. I still think locations east of the mountains in southern Colorado will see more snow than locations east of the mountains in northern Colorado this year. This is due to southern Colorado likely being influenced by the southern storm track more than the northern part of the state. To be honest, I hope that works out as drought recovery continues…
Tuesday, November 25th 2014
This past week, we have been hit another meteorology buzz term of “Lake-effect Snow” (known in the meteorological community as LES). Buffalo, NY was inundated by heavy snow amount with some places receiving > 80” in a 5-day period. It snowed so much that the NFL had to postpone and relocate the Buffalo Bills football game this past Sunday to Monday night in Detroit. Even the governor of New York, Andrew Cuomo, got into the LES talk by blaming the National Weather Service (NWS) about their terrible weather prediction, about which he was totally incorrect. The NWS did a fantastic job on letting people know of the severity of this LES event ahead of time. Since, here in Colorado, we do not get to experience LES events, let’s talk briefly about what they are and what atmospheric processes cause them.
A radar animation of the recent LES event over Buffalo (image courtesy of College of DuPage)
What is Lake-effect snow?
A snowstorm occurring near or downwind from the shore of a lake or large body of water resulting from the warming (destabilization) and moistening of cold air during passage over a relatively warm body of water. Lake-effect snow occurs when mean lake temperatures exceed mean land temperatures; therefore, the lake is not usually frozen for this phenomena to occur.
Average Annual Snowfall Around the Great Lakes From 1971 – 2000.
Overview of the Lake-Effect Process
Lake-effect snow occurs to the lee (east) of the Great Lakes during the beginning of the cool season. When the cold air mass travels across a relatively warmer lake, the air mass acquires heat and moisture, which destabilizes the air mass. Cloud formation is further enhanced by thermal and frictional convergence and upslope along the lee shore.
Conceptual Model of Lake-Effect Processes (Heat and moisture from lake + frictional convergence + upslope flow = clouds and lake-effect precipitation) (image courtesy of COMET).
Lake-Effect Snow Characteristics
Lake-effect snow is considered a mesoscale convective snow phenomena. Destabilization of the cold air mass is sometimes strong enough to produce thundersnow. Although convective in nature, the clouds in LES are very shallow compared to their Cumulonimbus counterparts (see Figure below), but can still pack quite a convective punch. These LES bands can come in a single- or multiband variety (See Figures below).
A vertical comparison between a Lake-Effect convective cloud and a summertime thunderstorm (image courtesy of COMET).
Satellite image of a Lake-Effect Single Snow Band event (image courtesy of COMET).
Satellite image of Lake-Effect Multiple Snow Bands event (image courtesy of COMET).
Lake-effect snow occurs from late fall through winter, though lake-effect rain can occur from late summer through mid-fall. Tremendous snowfall amounts and snowfall gradients are associated with LES events (see Figure below). Lake-effect snowstorms rarely produce multiple fatalities directly, but are very disruptive to commerce and transportation, with the exception of the recent Buffalo event (November 2014), where at least 12 confirmed fatalities have been reported.
Image courtesy of Brad Panovich of WCNC via his facebook page.
While Buffalo is infamous for LES events, other parts of the Great Lakes region and North America and the world do experience this type of weather phenomena. The image below shows where these events can and do occur across North America.
Lake-Effect-type Phenomena across North America (image courtesy of COMET).
Typically, the rule of thumb is that once the Lakes are frozen over, the majority of the LES events will diminish or end the lake-effect season. This is not always the case. A frozen lake does not necessarily preclude a lake-effect event if ice cover is not continuous. Sensible heat flux can occur through the ice but not as efficiently; therefore, if enough moisture is available, LES bands can still occur.
Key Processes from a Forecasting Point of View
Localized instability - lapse rate and boundary layer depth
Fetch – the distance in which the cold air moves across the warmer lake water (the longer the fetch, the higher the snowfall potential)
Wind direction and shear
Cloud microphysics – snow crystal habits
Synoptic (large) scale forcing
Orography/topography -- lake shape and orientation and lee-shore topography
Snow/ice cover on the lake
Lake-effect snowstorms are difficult to observe and forecast due to several factors. First, LES bands are shallow systems (depth often < 3 km), and the lowest elevation radar scans “overshoot” the tops. Secondly, the onset, intensity, orientation, and exact location are very sensitive to wind shear/direction and inversions in the lower troposphere. Third, LES bands are difficult to distinguish from orographic influences in some locations (e.g., Great Salt Lake). Lastly, some operational NWP models do not have sufficient resolution, microphysics, etc. to resolve the scales of lake-effect snow bands quite yet. The HRRRX, the ESRL version of the HRRR, performed well compared to its NCEP counterpart keeping the heaviest snow just of south of Buffalo.
11-h forecast of simulated composite reflectivity from HRRRX (ESRL version) initiated on 1900 UTC at 11/17/2014.
Lake-effect snow is weather phenomena that can be difficult to forecast due to the sensitivity of many atmospheric variables, especially upstream fetch and wind direction. These events can last as long as the cold air moves across the “relatively” warm lake without much deviation from the overall wind field. In most cases, LES episodes do not create major societal distribution and multiple fatalities. The November 17-21, 2014 event near Buffalo, NY was an historic event and somewhat atypical as far as LES events go. Snowfall rates for this storm reached 5” per hour over many hours revealing the highly convective nature of the snow band. In the end, the NWS did a fantastic job in forecasting and anticipating societal impacts, but mitigation for an outlier event is nearly impossible.
Total snowfall from Nov. 17-21, 2014 lake-effect snow event in western New York (Image courtesy of the Weather Channel).
Sunday, November 23rd 2014
We’re heading into the final full week of November and Thanksgiving, 2014. As we discussed in our previous post, the weather pattern for eastern Colorado has been overwhelmingly ‘normal’ since early in the week, with ongoing mountain snows.
Snowfall across central and western Colorado will continue through Tuesday, with winter weather advisories and winter storm warnings up across the board. Snowfall totals will range from 10 to 20”+ through Monday evening. There will be a chance for a few snow showers across lower elevations through Tuesday as well, but we’re not expecting much impact from these showers. The latest hi-res NAM has a dusting of snow for Denver over this period, with a spotty inch or two across the Palmer Divide. Mean SREF plumes produce a ½” of snow at DIA, but likely most folks will see less or none. Greatest snow shower activity will likely be south and southeast of Denver Sunday. Temperatures will also be cooler over the next several days, with daytime highs near 40.
Here’s a look at the WPC’s snowfall probability map for 8” or more of snow through Wednesday -- probabilities of 80% for the north central mountains.
WPC 3 Day Snowfall Probabilities Colorado
The forecast for Thanksgiving across eastern Colorado has gone from what looked “interesting” regarding snow, to likely nothing. Late last week the EURO was indeed showing a pretty hefty cold front and chance for snow by Thursday for Denver, even then its own ensembles were a bit more finicky. Also, the Canadian (which had a few snowy runs) backed off quickly, and well, the GFS never really showed anything. Chalk one up for the GFS in the win column? Maybe.
It seems to me that we were way too far out for anyone to be getting too excited about anything in the first place. And the ECM tends to have a western bias at times with its troughing -- while the GFS can have an eastern bias as we saw much of early fall. It was worth watching if only because our preferred medium range model was showing snow -- but why anyone feels the need to share 10 day snowfall maps in any serious setting, is beyond me. Fact is, this time of year some model at any given run almost always shows a snowstorm 10 days out. Always look to see what the ensembles are trying to do, any if two of three models have no snow, maybe consider there’s something there. Get ready for a long winter of long-range maybe snows…
Most models and their ensembles now generally agree on ridging taking hold by midweek, which will lead to a more mild, dry Thanksgiving day. What we’ll likely need to watch Thursday is for a backdoor cold front that could make its way far enough west to knock temperatures down even for the Denver area, and maybe quite a bit across the eastern Plains of Colorado. Interestingly, it seems like a lot of the same folks calling for a Turkey-snow are now calling for warm, when we really need to watch the potential for cooler temps closely. The chance for snow in Denver, at this time, looks pretty low.
The GEM’s 5 day temperature anomaly from Nov 25 through Nov 30 actually looks pretty good to me. The coldest anomalies will be east of Colorado, but eastern Colorado will range from near normal to several degrees below normal through the period.
Beyond Thanksgiving things become a bit more murky. Last week the ECM and Canadian models both showed a very active pattern through the first week of December, but latest runs are less convincing. Plenty of time to see how things will shake out here…
In short -- cooler temperatures Sunday and Monday with a few snow showers drifting off the mountains across the Denver area. Greatest chance for any accumulation east of the mountains will likely be along the Palmer Ridge, and even then any accumulation will be very light. Still a few questions especially with regard to temperatures for Thanksgiving day, but snow chances generally look low for the I-25 urban corridor on Turkey Day. We’ll keep a close eye on things and offer updates as needed.
All for now!
Wednesday, November 19th 2014
We’ve snapped back to a normal weather pattern this week in Denver after an extended period of bitter cold. Afternoon temperatures in the 40s and low 50s is right on par for the time of year, and this trend will continue right into the weekend.
The afternoon “warmth” is due in part to downsloping northwesterly winds that have also created these periods of standing waveform clouds. The mountain waves are pronounced during times of a strong northwesterly fetch and humidity at the mid-levels.
To the visual forecaster, these cloud formations indicate more than dry/warm-ish Denver days. These also indicate that enough moisture is within the flow to be aware that mountain snow is also possible. Indeed that is the case. Here is the GFS, as a singular example, of the snowfall expected through 5pm Sunday. This assumes a 10:1 ratio, which is underestimate in this type of flow regime. 20:1 would produce double these amounts and would seem more appropriate. I am working on a post discussing our major modeling shortfalls when it comes to snow prediction; you’ll see it in the coming weeks.
The northern and central mountains with a north to westerly orientation fare well in these weather patterns. Resorts like Beaver Creek and Vail send lots of press releases on these days, too. Further, Steamboat Springs is known for “champagne powder” during the same periods of time that Denver is warm and under mountain wave clouds.
I chose to stop at 5pm Sunday night for a reason. That reason is that that is when this pattern will take a brief break. A trough, a weak one, will scoot over the state later Sunday through Monday and will bring a slight cooldown to Denver and a chance for our own snow. This is still the GFS, but through 5pm Wednesday night. As of this posting, we aren’t expecting a significant snowfall anywhere outside of the mountains. It’ll be another minor affair for the metro area.
As far as the temperature change. Denver will likely go from near 50 Saturday to the 30s for Sunday through Tuesday...also a relatively minor affair.
Bottom line, Thanksgiving skiers and travelers will come away with a pretty good week across Colorado. HOWEVER, I’m guessing many of you will be headed elsewhere for Thanksgiving...east perhaps?
Currently, Buffalo NY is getting walloped with over 5 feet of snow. The Great Lakes region will certainly be a snow covered Thanksgiving from what’s fallen there the past few days.
Carolyn Thompson/AP Photo
Additional snowfall through the next 7 days doesn’t look nearly as major, which is better news for flights to the major hubs back east. Through Thanksgiving Day we will likely see snowfall in these areas:
That’s a lot of snow to fall in Northern Minnesota, but for the most part the upcoming snowfall is quite manageable. HOWEVER, if you are to fly to the northeast Sunday, Monday, or Tuesday there could be great delays as that’s when most of the snow will fall and the wind will be awful around the Great Lakes with that system.
As the snow intensity drops, temperatures will be remain quite cold. If you are setting out that wardrobe for Grandma’s house, pack extra layers for those areas to the east. The quick skiff of snow and “cold” that Denver will feel early in the week will become jarring cold for the eastern half of the country by Thanksgiving.
This is showing above or below average temperatures for Thanksgiving morning. Yes this is showing nearly 20-degrees colder than average for the southeast near Atlanta, and 10-degrees colder than average for New York City. Ugh, we will see Al Roker bundled up again at the Macy’s Thanksgiving Day Parade. Meanwhile, the southwestern quarter of the country will have mild weather, including Denver. For those needing to warm-up, Phoenix, LA, or San Diego will do just that for you.
Denver’s Thanksgiving Day weather is still a bit more in question. Several models have been trying to develop our next system by next Thursday and Friday. The EURO/Canadian have been most consistent with this, but the the GFS is seeing it too (at times) to an extent. A long ways out, but something we’re watching, and if it turns out to be true -- those anomalous cold temperatures pictured above will likely be a bit further west as well.
I hope this gives you an early heads up on the coming week. We, of course, will get far more in depth the closer we get to the holiday week. I, for one, am tuning the skis right now and will be hitting Colorado’s fresh powder. I like what I see coming to our mountains this week.