As I was composing a blog on thunder-freezing rain in southwest Missouri on Thursday morning, February 21, 2013, I received a call from my fearless leader, Jeff Masters, who asked me to talk to National Public Radio about reports of thunder-snow coming out of the Middle West in concert. The thunder-snow and thunder-freezing rain were produced by the developing storm system over the southern Plains (check out the 12Z surface analysis below). At the time, I had already collected the METARS at Springfield, Missouri, from 13Z to 15Z on February 21 (check them out; I annotated important data for your convenience).
On the list of METARS, note the present weather (in red), which was freezing rain during this period. But there were also reports of lightning all around the Springfield airport (in blue). So thunder-freezing rain was ongoing at surface air temperatures of minus 1.7 degrees Celsius = 29 degrees Fahrenheit.
The 12Z surface analysis on February 21, 2013. At the time, a developing surface low-pressure system over the Texas Panhandle drew relatively warm, moist air from the Gulf of Mexico over a layer of cold air north of the low's warm front. Courtesy of HPC.
Many people naturally associate thunder and lightning with the warm season, so freezing rain, sleet, or snow reaching the ground during a thunderstorm in winter seems like an enigma, leading, not surprisingly, to lots of inquiries.
Let's face it, thunder-snow draws the most attention. Check out (below; larger image) the METARS from 12Z to 14Z at Topeka, Kansas, on February 21. I annotated the observation with heavy thunder-snow.
The METARs at Topeka, KS, from 12Z to 14Z on February 21, 2013. Larger image. Courtesy of UCAR.
I'm not shy, so I jumped at the opportunity to discuss weather on National Public Radio (here it is). When I was teaching at Penn State, I conducted many similar interviews with the media, so I knew going into the interview that I had to be succinct and not too "professorial" in order to convey something meaningful and understandable to the listening public. My goals in this blog will be a loftier and more comprehensive explanation as I cover the topic of elevated thunderstorms that produce wintry precipitation (sleet, freezing rain, or snow).
While talking to NPR, I didn't get too deep into the science, but I also didn't take any dramatic shortcuts in science like the national news did Thursday evening. One anchor succinctly pontificated that thunder-snow is caused by "the clash of warm and cold air" and that thunder-snow is "rare." In my opinion, the "clash" between air masses is a catch-all, generic argument frequently used by the media as a watered-down explanation for all kinds of weather, but I'll bet it really leaves some viewers wanting more.
The anchor followed up his "clash metaphor" by describing thunder-snow as "rare." Using "rare" in the same context as the "clash of warm and cold air" seems contradictory to me because the juxtaposition of warm and cold air masses is the hallmark of winter. I mean, it happens all the time during the cold season. So, using the news anchor's "logic," if cold and warm air clash frequently, shouldn't thunder-snow be just as frequent?
Again, I didn't jump right into an in-depth scientific discussion during my NPR interview. I began by stating that, in my view, thunder-snow is not really a rare event. I admit that I was being a bit coy by taking such an opposing view, but there's always a method to my madness.
The 14Z mosaic of composite reflectivity on February 21, 2013. Larger image. Courtesy of Penn State.
Here's the scoop. Almost all of the precipitation that falls over the middle latitudes (roughly 30 degrees to 60 degrees latitude) begins as snow at cold, high altitudes. Thus, on a hot, humid July day after surface temperatures surge into the 90s (Fahrenheit), for example, and thunderstorms subsequently erupt, there's thunder-snow four miles up in the atmosphere. From this standpoint, thunder-snow is not rare at all. Of course, I realize that the context of thunder-snow as a "rare" event was meant to apply to winter. Even so, strong low-pressure systems that emerge from the southern Rockies and establish a surface low over the southern Plains (like the one over Texas early Thursday), draw relatively warm, moist air from the Gulf of Mexico over a layer of cold air north of the low's warm front (revisit the 12Z surface analysis on February 21). This lifting of air can result in wintry precipitation from elevated thunderstorms, assuming the middle part of the troposphere is sufficiently unstable. For confirmation, check out the lightning display at 14Z on February 21 and the 14Z mosaic of composite reflectivity (see above; larger image).
On NPR, I stated, as succinctly as I could, that the thunder-snow four miles overhead during summer is often observed at the ground as rain with lightning and thunder (snowflakes melt into raindrops at, say, 10,000 feet above the ground). During winter, snow sometimes reaches down to the earth's surface where observers also see flashes of lightning and then hear thunder. Of course, it's not quite as simple as snow not melting into raindrops during winter, so I saved the real science for this blog.
The low I pointed out on the 12Z surface analysis over Texas produced warm advection north of the low's warm front. The 14Z Rapid-Refresh model analysis of 850-mb heights (dark contours), 850-mb isotherms (dashed red contours), and 850-mb temperature advection (color filled in degrees Celsius per hour) shows warm advection (red and pink color filled areas) surging northward into western Missouri. You can see the impact of the warm advection over the layer of cold air next to the ground on the 14Z Rapid-Refresh model skew-T at Springfield (below; larger image).
The 14Z Rapid-Refresh skew-T at Springfield, Missouri, on February 21, 2013. The line designating 0 degrees Celsius is highlighted in yellow. Courtesy of Penn State.
The lower troposphere was very stable...note the strong temperature inversion from roughly 925 mb to 800 mb on the Rapid-Refresh skew-T above. In effect, the relatively warm, moist air feeding the updrafts of thunderstorms began at 800 mb, not the ground, qualifying any developing thunderstorms as elevated convection (as opposed to surface-based convection).
The convective available potential energy (CAPE), which is a measure of the potential strengths of updrafts, is surface-based, and really should not be applied to elevated thunderstorms (in most cases of elevated convection during winter, surface-based CAPE is zero; it might fool you into believing that winter thunderstorms aren't possible). But, of course, they are. Instead of CAPE, weather forecasters use MUCAPE in situations where elevated thunderstorms are slated to occur. MUCAPE stands for Most Unstable CAPE, which essentially identifies the most unstable parcel of air near the top of the low-level stable layer. In other words, MUCAPE refers to the air parcel aloft that feeds into the updraft of an elevated thunderstorms and whose upward acceleration in the updraft is greatest. Keep in mind that the upward acceleration of an air parcel is governed by the difference between the parcel's temperature and the temperature of its environment. In a nutshell, the warmer the air parcel is compared to its environment, the greater its upward acceleration (the good old hot-air balloon analogy).
Here's the 14Z Rapid-Refresh model analysis of MUCAPE (the thick and thin red contours, expressed in Joules per kilogram) and the altitude of the most unstable air parcel (color-filled, dashed black contours, in meters). In southwest Missouri (near Springfield, for example), the altitude of the most unstable air parcel was a bit lower than 3000 meters (a bit greater than 700 mb). I realize this is a very rough estimate, but, for ease of presentation, let's assume the most unstable parcel lay near 740 mb. With this very rough estimate in mind, I went back to the 14Z Rapid-Refresh skew-T and I marked the path of the most unstable air parcel rising along the local moist adiabat from roughly 740 mb (see annotated skew-T below). The saturated parcel, whose path is indicated by the blue curve, lies to the right of the environmental temperature sounding (in red). In other words, the parcel was warmer than its environment and, thus, positively buoyant, rising to relatively high altitudes and setting the stage for deep convective clouds.
The MUCAPE (most unstable convective available potential energy) associated with this parcel was simply the area bounded by the blue curve and the red curve (according to the 14Z Rapid-refresh analysis, about 250 Joules per kilogram). That's relatively low by most CAPE standards, but sufficiently strong for updrafts to create pockets of positive and negative electrical charges in the convective cloud and to pave the way for lightning. In order to achieve an adequate separation of electrical charge for lightning to occur, the convective cloud must have an updraft that extends to altitudes where temperatures are well below minus 20 degrees Celsius. Moreover, the convective cloud must have a base warm enough for liquid drops (supercooled or otherwise) to exist.
The 14Z Rapid Refresh skew-T at Springfield, Missouri, on February 21, 2013. Courtesy of Penn State.
Farther to the northwest, Topeka, Kansas, was deeper in cold air, and, even though there was weak warm advection at 850 mb, temperatures in the lower troposphere stayed below 0 degrees Celsius, the melting point of ice, so, unlike Springfield, Missouri, there wasn't any melting layer and snowflakes made it to the ground while lightning flashed and thunder rumbled across the landscape. In case you're interested, here's the 13Z Rapid-Refresh skew-T at Topeka, KS on February 21 (around the time when heavy thunder-snow was reported).
I'm sure you're getting the picture for why I didn't go into as much scientific detail during my NPR interview. Hopefully, you've learned the meteorology of thunder-snow during winter. Thunderstorms are elevated, and temperatures throughout the troposphere are lower than the melting point of ice.
P.S. I will continue to call 0 degrees Celsius = 32 degrees Fahrenheit the melting point of ice and not "freezing" or the "freezing mark." I know I'm fighting a losing battle, but at least I'll go down fighting. :-)
The Cranky Old Man :-)
Updated: 1:06 PM GMT on February 27, 2013
A A A
Unprecedented Radar Reflectivities and Snowfall Rates
I have one last observation to make about NEMO before I move on, and it's a lulu. There were some 50+ dBZ reflectivities in Connecticut the evening of February 8, and, in my 40 years as a forecaster, I've never observed higher reflectivities associated with snow. For confirmation, check out the 0218Z base reflectivity from the radar at New York City / Upton, NY (KOKX), on February 8 below (larger image). Courtesy of NOAA.
The 0218Z image of base reflectivity from the radar at New York City / Upton, NY (KOKX) on February 9, 2013 (the evening of February 8). Larger image. The 50+ dBZ reflectivity of snow is the highest I've observed in my 40 years of forecasting.
Of course, I've seen high reflectivities associated with bright bands (wet snow near the top of a melting layer; stay tuned), so I'm referring exclusively to high reflectivities associated with snow that reaches the ground. Before I discuss Nemo and the storm features that led to such high reflectivities over Connecticut on the evening of February 8, 2013, I think a slight digression on bright banding is in order (I've heard too many erroneous explanations for relatively high reflectivity in wintry situations that never even consider bright banding as a possible cause). For the record, bright banding was discovered during military operations in World War II, and there were papers written in the 1940s that explained the underpinning science (example 1946 paper).
Here's how I explain bright banding. As falling snowflakes reach the top of the melting layer, where the temperature is 0 degrees Celsius, they began to melt (again, 0 degrees Celsius = 32 degrees Fahrenheit is NOT the "freezing mark"). Melting snowflakes soon become covered with a film of meltwater, and they look like large raindrops to the radar. Thus, the base reflectivity abruptly (and dramatically) increases. In rather quick fashion, however, snowflakes melt completely, shriveling in size as raindrops take shape. Since radar is very sensitive to particle size, reflectivity decreases rapidly once the melting process is complete. Moreover, raindrops now quickly accelerate earthward (raindrops fall faster than snowflakes). This rather abrupt acceleration after the water-covered snowflakes melt completely into raindrops decreases the number of radar targets just below the melting level (like the flow of traffic quickly opening up after cars accelerate away from a crowded toll booth on a super-highway). The combined effect of decreasing both the size and number of drops causes radar reflectivity to decrease just below the melting layer, creating a band of higher reflectivity (a bright band) above the ground.
To summarize my discussion on bright banding, check out this nifty flash animation of an idealized bright band, copyrighted by Penn State's online Certificate of Achievement in Weather Forecasting.
During the evening of February 8, 2013, the tops of the melting layer lay south of New England (check out the 02Z Rapid-Refresh model analysis of the height of the melting level, in meters. Note the bulge in the contours toward eastern Massachusetts, a sign of warmer air pressing northward on the eastern flank of the low-pressure system (02Z Rapid-Refresh model analysis of MSL isobars). Just to be sure, I grabbed the 02Z Rapid-Refresh model skew-T at New Haven, Connecticut, which shows the entire temperature sounding (in red) to the left of (lower than) 0 degrees Celsius, which I highlighted in yellow. So snowflakes were the dominant hydrometeor indicated by the 50+ dBZ shown on the 0218Z images of base reflectivity (near the top of this blog) and composite reflectivity.
The 0218Z image of Hydrometeor Classification from the radar at New York City / Upton, NY, on February 9, 2013. Note the red and dark pink, which indicate hail and graupel, respectively. To be fair, the Hydrometeor Classification algorithm identified some unknown precipitation in southern Connecticut (purple). Larger image. Courtesy of NOAA.
Reflectivity is much more sensitive to particle size than number, so I can only deduce that wet snow, which is more likely to form large aggregates (agglomerations) than dry snow, generated such large reflectivities. Like I said before, I don't think I've ever observed 50+ dBZ associated with snow. On the other hand, I've never heard of snowfall rates as high as 8 inches per hour that were sustained for two hours. The rarity of the reported snowfall rates seemed to fit the rarity of the radar data for this event. I think the radar reflectivities for snow were so large that the hydrometeor classification algorithm, a utility of dual-polarization radars, simply failed in this case. Indeed, the hydrometeor classification (HC) algorithm "concluded" that the heaviest snow had to be hail or graupel (ice). For confirmation, check out the 0218Z HC product (above; larger image) from the dual-polarization radar at KOKX). Presumably, the red (hail) and dark pink (graupel) shown on the 0218Z image of hydrometeor classification were erroneously selected because there must be a red flag in the HC algorithm that automatically eliminates snowflakes as the possible dominant hydrometeor when radar reflectivities are 40 to 50 dBZ or higher. In effect, these very high radar reflectivities likely "overruled" other dual-pol metrics (such as differential radar reflectivity) that probably indicated otherwise.
A quick look at the synoptic set-up during the evening of February 8, 2013, starts with the 02Z Rapid-Refresh model analysis of 850-mb heights, which showed a large gradient in 850-mb heights off the New England Coast. In response, a robust low-level jet stream, with wind speeds exceeding 80 knots (see the 02Z Rapid-Refresh model analysis of 850-mb isotachs and streamlines below; larger image), rapidly transported Atlantic moisture inland, setting the stage for unprecedented radar reflectivities and snowfall rates over Connecticut and other parts of New England.
The 02Z Rapid-Refresh model analysis of 850-mb isotachs (color-filled in knots) and 850-mb streamlines. Larger image. Courtesy of Penn State.
For me, this snowstorm will make my lifetime top-ten list. And I'm pretty old.
Updated: 6:06 PM GMT on February 19, 2013
A A A
Would the Groundhog Day Gale of 1976 Been a Named Winter Storm?
In my last blog, I presented research showing that the long-term average of strong East-Coast winter storms (wind speeds of at least 45 knots) was highest during February. I also noted that February had the greatest number of "full coast" storms, which provided more opportunities for the southern and northern branches of the jet stream to constructively interact and produce intense nor'easters. By "interact," I meant that two 500-mb short-wave troughs (one in each branch) phase or deliver a one-two punch. I had indicated that today's follow-up blog would elaborate on the one-two punch, but the meteorology is pretty straightforward (here's my definition).
The surface analysis at 12Z on February 2, 1976. Note the 964-mb low centered over Maine. Courtesy of the National Meteorological Center (NMC).
So I decided to switch gears, although I'm still sticking to the theme of strong East-Coast winter storms during February. Indeed, the topic of today's blog is the infamous Groundhog Day Gale of February 2, 1976 (see the analysis of mean sea-level isobars above at 12Z on this day; details will be forthcoming). I chose this storm for three reasons. First, the storm certainly fits the motif of strong February nor'easters that I wrote about in my last blog (and an interesting topic Stu Ostro broached earlier in the week). Second, two 500-mb short-wave troughs, one in the northern branch and one in the southern branch of the jet stream, merged (phased) to set the stage for explosive cyclogenesis. Last, but not least, I'm wondering whether the Groundhog Day Gale of 1976 would have been named by The Weather Channel.
The surface analysis above indicates an intense low-pressure system centered over southwestern Maine at 12Z on February 2, 1976 (the minimum pressure at the time was 964 mb). During the next 18 hours, the bottom fell out of local barometers as the minimum pressure plummeted to 944 mb by 06Z on February 3 as the storm headed northeastward into the Canadian Maritimes. To this day, all-time records for low barometric pressure set in Maine (957 mb at Caribou, for example) during the Groundhog Day still stand. At Boston, Massachusetts, the barometric pressure of 965 mb ranked as the second lowest on the city's all-time list.
It doesn't take a rocket scientist to infer that the rapidly deepening low produced very strong winds, ranking the Groundhog Day Gale of 1976 as one of the fiercest windstorms on record in the Canadian Maritimes (and rivaling the Saxby's Gale of 1869). According to Environment Canada, winds gusted to 188 kilometers an hour (117 miles an hour) at Saint John, New Brunswick. Waves as high as 12 meters (almost 40 feet) and swells to 33 feet were observed in the Bay of Fundy. There was also surge flooding in Maine (I recommend reading this compelling NOAA account). Freezing spray from crashing waves coated everything with ice as far as four miles inland from the bay, and eroded chunks of coastline disappeared into surging seas (YouTube documentary of damage).
Coarse analyses of 500-mb heights (solid contours) and 500-mb absolute vorticity (dashed contours) at 00Z on February 1, 1976 (upper left), 12Z on February 1 (upper right), 00Z on February 2 (lower left), and 12Z on February 2 (lower right). Short-wave troughs (circled in red) were traveling in the northern and southern branches of the 500-mb westerlies. The two troughs phased by 12Z on February 2, 1976. Larger image. Courtesy of Robert S. Gaza and Lance F. Bosart.
As I suggested in my last blog, strong, "full coast" storms typically involve the phasing of short-wave troughs traveling in the northern and southern branches of the 500-mb westerlies. The evolution of the surface pattern from 00Z on February 1 to 12Z on February 2. 1976 (larger image), confirms a low-pressure system developing along the Gulf Coast and eventually moving up the East Coast (so it was a "full coast" storm). The pattern of 500-mb heights and vorticity (above; larger image) shows distinct short-wave troughs and their associated vorticity maxima (circled in red) embedded in the northern and southern branches of the 500-mb westerlies. Note that the short-wave troughs moved closer to each other from 00Z on February 1 to 00Z on February 2, eventually phasing by 12Z on February 2 (lower-right panel) and setting the stage for explosive cyclogenesis.
On The Weather Channel's Web site, there are three guidelines for naming winter storms:
1) Significant impact due to snow or ice within three days
2) Significant disruption to road and air travel
3) Life-threatening conditions from wind, snow, ice and cold
Given that the Groundhog Day Gale of 1976 didn't make the NESIS Top-46 list and that the Groundhog Day Gale's primary impacts were strong winds and surge flooding, I'm wondering if the Groundhog Day Gale of 1976 would have been named (to be fair, there were some pretty heavy lake-effect snows in the wake of the storm). Maybe I'm just plain wrong, but, taking the three criteria used to name winter storms at face value, I don't see how this storm could have been named (I recognize that "Groundhog Day Gale of 1976" is a name; I'm obviously referring to a name on a TWC list). What do you folks think?
I'm not trying to stir the pot here (well, maybe a little). I'm asking this question respectfully. Indeed, naming Nemo was very helpful, in my view. But the Groundhog Day Gale of 1976 was also one heck of a storm. It gained its infamy, however, from wind and surge flooding, not from snowfall. So I think a discussion about whether the Groundhog Day Gale of 1976 would have been named is a compelling topic. I'm looking forward to reading your opinions.
Updated: 12:35 PM GMT on February 18, 2013
A A A
About February Nor'easters
In response to Stu Ostro's interesting blog on February 8, I point out that there have already been several climatological studies of East-Coast winter storms (ECWS). Whenever I taught East-Coast cyclogenesis in my synoptics course at Penn State, I frequently cited this paper published in the Journal of Climate in 2000 by Hirsch, et al.
For the record, the authors partitioned East-Coast winter storms from 1951-1997 and 1948 into three categories...those affecting the northern Atlantic Coast (poleward from Cape Hatteras, North Carolina), the southern Atlantic Coast (equatorward from Cape Hatteras, North Carolina), and the entire Atlantic Coast (the "full" coast).
The monthly distribution of East-Coast winter storms that affected the northern Atlantic Coast (black), the southern Atlantic Coast (white), and the entire Atlantic Coast (gray). The numbers within each column indicate the percentage of each type of ECWS during the specified month. Larger image. Courtesy of Hirsch et al.
In order to qualify as an East-Coast winter storm for this study, a low-pressure system had to satisfy the following four criteria. Each of the qualifying 930 lows from 1951-1997 and 1948...
1) had a closed circulation
2) was located along the East Coast of the United States within the quadrilateral bounded at 45 degrees North latitude by 65 degrees and 70 degrees West longitude, and at 30 degrees North latitude by 75 degrees and 85 degrees West (the solid line segments shown on this map mark the boundaries of the quadrilateral used to count East-Coast winter storms)
3) showed general movement from the south-southwest to the north-northeast, and
4) produced winds with speeds greater than 20 knots during at least one six-hour period
The bar graph shown above (larger version), which I adapted from the paper, shows the monthly distribution of qualifying storms affecting the northern Atlantic Coast (black), the southern Atlantic Coast (white), and the entire Atlantic Coast (gray). The numbers within each column indicates the relative percentage of storms affecting each of the three coastal strips during the specified month. During the month of February, for example, 50% of the qualifying East-Coast winter storms affected the entire Atlantic Coast ("full coast"), while 39% and 11% were confined to the northern and southern coasts, respectively.
Not surprisingly, the greatest number of East-Coast winter storms occurred in January (see graph above), but, as Stu Ostro suggested, more strong East-Coast winter storms occurred during February than any other month (graph not shown here; refer to the paper). In case you're wondering, the authors defined a strong ECWS as a deep low that produced maximum winds of at least 45 knots.
Perhaps the key to February's propensity for strong East-Coast winter storms lies in the observation that the second month of the year has the most "full coast" storms than any other month. I say this because, in my view, a "full coast" storm is more likely to involve both the southern and northern branches of the jet stream. More specifically, a disturbance in the subtropical jet stream (the southern branch) has the opportunity to phase (interact) with a disturbance in the polar jet stream (the northern branch). That's exactly what happened with Nemo (oh my, I swore I'd never buy into the practice of naming winter storms; I admit it's a name I'm not likely to forget).
For other "full coast" storms, disturbances in the two branches don't always phase, but sometimes deliver a one-two punch, a 500-mb pattern conducive to explosive cyclogenesis. It's characterized by a short-wave trough in the southern branch of the jet stream lowering surface pressure along the East Coast just before the arrival of a short-wave trough in the northern branch.
So what I'm suggesting here is that February has the most strong East-Coast winter storms because it has the most "full coast" storms. In such patterns, there are more opportunities for both branches of the jet stream to interact constructively. That's assuming, of course that there's an active southern branch (an active subtropical jet stream).
The long-term February average 200-mb wind vectors. Arrows indicate direction, and speeds are color-coded in meters per second. Larger image. Courtesy of ESRL.
Mind you, I'm thinking out loud here, but there really is a method to my madness. Check out (above; larger image) the long-term February average of 200-mb wind vectors (arrows indicate direction, and speed is color-coded in meters per second). I chose 200 mb because that's the typical pressure altitude of the subtropical jet stream (STJ). You can see the dramatic footprint of the STJ off Southeast Asia and southern Japan, where wind speeds exceed, on average, 65 meters per second (145 knots).
There's a break in the STJ over the eastern Pacific, on average, but mean speeds start to ramp up over Baja California and peak at roughly 45 meters per second (100 miles an hour) over the southeastern United States.
Now compare these average 200-mb winds during February to January (below; larger image). Focus your attention on the eastern Pacific and Baja California and note that 200-mb wind speeds are weaker, on average, during January (compared to February). I'm thinking that a more active southern branch (subtropical jet stream) lends itself to more disturbances that might pave the way for cyclogenesis along the southern Atlantic Coast (the low then moves northward along the Atlantic Seaboard, where it has the opportunity to interact with the polar jet stream).
The differences in 200-mb wind speeds over the eastern Pacific are subtle, I agree, but remember that we're dealing with monthly averages here. So what appears to be small differences can translate to a very active southern branch on any given day. Too see my point, check out the analysis of 200-mb vector winds on February 8, 2013 (a daily average), when the southern branch was doing its part to produce the historic storm called Nemo. Yes, the southern branch was pretty active as Nemo started to take shape, wouldn't you agree?
The long-term January average 200-mb wind vectors. Arrows indicate direction, and speeds are color-coded in meters per second. Larger image. Courtesy of ESRL.
The subtropical jet is driven primarily by the conservation of angular momentum as air parcels in the upper branch of the Hadley Cell move northward as they circle the earth (idealized, spiraling path of an air parcel looking down on the North Pole; courtesy of A World of Weather: Fundamentals of Meteorology). But north-south temperature contrasts aloft can enhance or weaken the subtropical jet stream. I say "aloft" here because the subtropical jet stream, unlike the polar jet stream, is not rooted in temperature contrasts at the earth's surface. Rather, the STJ has its roots around 500 mb / 400 mb, give or take. The takeaway from this discussion is that relatively large (small) temperature contrasts in the middle troposphere can enhance (weaken) the STJ.
To see what I mean, compare the long-term average 500-mb temperatures for February (below; larger image) to January's average 500-mb temperatures. Over the eastern Pacific and Baja California, the north-south temperature gradient is slightly greater in February (compared to January), probably accounting for the faster 200-mb winds over that neck of the woods. Note that a slight 500-mb thermal ridge over the eastern Pacific Ocean in January collapses in February, leading to a slightly enhanced temperature gradient (and a faster STJ). I'm not sure about the reason why this occurs, so I'll give it more thought.
The long-term average 500-mb temperatures during February. Larger image. Courtesy of ESRL.
The climatological study (Hirsch et al.) seems to support my theory that a robust southern branch plays an important role in the development of "full coast" storms. I quote: "Strong El Nino events are often characterized by a very active southern branch of the jet stream. This pattern may be conducive to the development of southern and especially full coast storms at the expense of the frequency in northern ECWS."
So I wanted to follow up Stu Ostro's interesting blog to let you know that there are indeed climatological studies that show February holding the distinction of producing the strongest East-Coast winter storms, on average, than any other month (including January, which produces more ECWS than any other month).
Updated: 4:52 PM GMT on February 14, 2013
A A A
Looming Snowstorm and the Fujiwhara
With an historic snowstorm looming over New England later today and tomorrow, I wanted to set the record straight with regard to the evolving interaction between the two 500-mb short-wave troughs in the southern and northern branches (see the 12Z NAM model analysis below; larger image). The explanations I heard on television have run the gamut from "merging surface lows" to a "phasing" of the 500-mb short-wave troughs.
The 12Z NAM model analysis of 500-mb heights, 500-mb absolute vorticity, and 500-mb wind barbs on February 8, 2013. Larger image. Courtesy of NCEP.
The merging of the two surface lows (12Z surface analysis) is not really feasible because the low associated with the 500-mb short-wave trough in the northern branch can't scale the Appalachians and remain intact. That's because an eastward-moving low attempting to "climb" a mountain will gain anticyclonic relative vorticity (vorticity is a measure of the rotation of an air parcel around the local vertical). For a low-pressure system whose longevity depends on maintaining or enhancing its cyclonic vorticity, gaining anticyclonic rotation while climbing a mountain means that the low will weaken dramatically. By the way, the notion of an air parcel gaining anticyclonic vorticity as it ascends a mountain is a consequence of the conservation of Ertel's Potential Vorticity, a topic which I intend to write about in a future blog.
The explanation that purports a phasing of the two short-wave 500-mb troughs is correct, although it's been my experience that most folks treat phasing as a simple merging of the two troughs. Phasing is not that simple. Indeed, two short-wave 500-mb troughs destined to phase typically perform a Fujiwhara, a cyclonic "dance" that occurs when the two short-wave troughs get sufficiently close together. I point out here that the context of a Fujiwhara typically involves tropical cyclones (when tropical cyclones get sufficiently close together, they start to cyclonically rotate about the same intermediate point). Obviously, the Fujiwhara also occurs over the middle latitudes during the phasing of two 500-mb short-wave troughs (details will be forthcoming). For the record, I have also observed the Fujiwhara in action when a 500-mb short-wave trough and a tropical cyclone get close enough to interact.
Let's assume that two short-wave 500-mb troughs get sufficiently close together to interact, and, for ease of explanation, are aligned from north to south. Check out the idealized juxtaposition of the two short-wave troughs on slide #1 from the synoptic course I used to teach at Penn State (the short-wave troughs are each designated by an "X" inside a red circle to mark the position of its vorticity maXimum). Okay, the northern short-wave trough starts to slow its eastward progression as it encounters the cyclonic circulation associated with the southern short-wave trough (see slide #2). Meanwhile, the southern-branch short-wave trough swings northeastward as it "senses" the cyclonic circulation of the northern short-wave (see slide #3). Note that the two vorticity maxima are even closer (see slide #4 below; larger image). At this point, the Fujiwhara continues, with one trough eventually emerging after this cyclonic dance comes to an end.
When two short-wave 500-mb troughs (marked by their vorticity maXimum in each case) get sufficiently close together to interact, the northern trough slows and the southern trough swings northeastward, setting the stage for the Fujiwhara to continue. Eventually, the two troughs phase. Larger image.
Look again at the 12Z NAM model analysis of 500-mb heights and absolute vorticity and note the initial positions of the two 500-mb short-wave troughs in the northern and southern branches. Now follow the progression of 500-mb NAM forecasts from 3 hours (valid at 15Z this morning) to 21 hours (valid at 09Z tomorrow morning). I'm listing them individually so you can focus more closely on the evolving interaction of the two short-waves. Can you see the Fujiwhara?
NAM Forecasts of 500-mb heights and absolute vorticity from the 12Z run on February 8, 2013:
Updated: 8:46 PM GMT on February 08, 2013
A A A
Narrow Cold-Frontal Rainbands
The 12Z surface analysis on January 30, 2013 (close-up analysis), indicated a "squall line" just ahead of a cold front slicing through the Ohio and Tennessee Valleys. Sure enough, the corresponding 12Z mosaic of composite reflectivity (below; larger image) displayed a very narrow line of relatively high reflectivity (formally, a narrow cold-frontal rainband...more in a few moments).
So this was not a squall line in any traditional sense, assuming that you accept the AMS definition hook, line and sinker (the AMS glossary defines a squall line as a continuous or broken line of active thunderstorms). Indeed, the lightning data from 11Z to 12Z on January 30 (courtesy of WSI) indicated that there weren't any cloud-to-ground strikes in the vicinity of this so-called squall line.
The 12Z mosaic of composite reflectivity on January 30, 2013, shows a classic narrow cold-frontal rainband (NCFRB) stretching from south-southwest to north-northeast across Ohio and Kentucky (focus your attention on the thin, red ribbon of roughly 50 dBZ reflectivity). Larger image. Courtesy of WSI and Penn State.
I'm wondering how literally the AMS requirement of thunder should be interpreted here. For the record, there are tornadic supercells that are bereft of CG (cloud-to-ground) and IC (in-cloud) lightning. So, if we follow the letter of the law, these tornadic supercells weren't even "thunderstorms" because they didn't produce lightning and thunder. Yet, dynamically speaking, they are, without reservation, supercells (deep, moist convection driven by rotating updrafts). Confused? I'm not surprised. Such aberrant supercellular storms develop in atypical environments (read more), where the microphysics in these convective clouds simply does not lend itself to a separation of charge necessary to spark lightning. The inner cores of hurricanes also have little to no lightning, yet they are typically described as large conglomerations of tropical thunderstorms.
The bottom line here is that the line of convection in Ohio and Kentucky on January 30, 2013, was not a squall line according to a strict interpretation of the AMS definition. In my view, however, this definition is ostensibly too strict. Why not substitute "deep, moist convection" for "thunderstorms?" I realize that "deep, moist convection" is a bit unwieldy to continually write or say, but, if the line of convection on January 30 doesn't qualify as a squall line, then I think there's a problem. That's because the term, "squall," generically refers to blustery, showery weather. So a "squall line" could be a line of blustery, showery weather. And that's exactly the kind of weather the line of convection produced on January 30, 2013 (note the reports of damaging winds from the Storm Prediction Center in southern Ohio, Kentucky, western Pennsylvania, etc.).
The 12Z NAM model analysis of 1000-mb streamlines on January 30, 2013. Note the confluence of streamlines (shift in wind direction) over Ohio and Kentucky. Larger image. Courtesy of Penn State.
The debate about the AMS definition of squall line aside, the line of convection on January 30 was, as I hinted earlier, a narrow cold-frontal rainband (NCFRB). As their name suggests, NCFRBs are noticeably narrow, with the distance measured perpendicular to the line as small as one to two kilometers. Narrow cold-frontal rainbands form along a surface cold front where there's a noticeable shift in wind direction. For confirmation, check out the NAM model analysis of 1000-mb streamlines at 12Z on January 30, 2013 (above; larger image). Updrafts in narrow cold-frontal rainbands can be intense, with speeds reaching 10 meters per second (about 22 mph) at relatively low altitudes (one kilometer or so). Despite these speedy updrafts in the lower troposphere, the depth of the convection is, on average, three kilometers.
To get a sense for the vertical structure of temperature in an environment conducive for a narrow cold-frontal rainband, check out (below) the 12Z skew-T at Wilmington, Ohio (KILN) on January 30, 2013, where the NCFRB passed at roughly this time (see meteogram). Note the "EL" on the left of the skew-T (circled in red)...it stands for the equilibrium level, which is the altitude at which the temperature of a rising air parcel (lifted from the surface) equals the temperature of its environment. In other words, the equilibrium level is the altitude where a rising air parcel becomes neutrally buoyant and stops accelerating upward. Put another way, the equilibrium level is a proxy for the tops of convective clouds. So, yes, the depth of convection associated with narrow cold-frontal rainbands is relatively shallow.
The 12Z skew-T at Wilmington, Ohio (KILN), on January 30, 2013. The temperature sounding is red, and the dew-point sounding is green. Note the altitude of the equilibrium level (EL), which is a proxy for the tops of the convective clouds associated with the narrow cold-frontal rainband over southwest Ohio at this time. Courtesy of the University Center of Atmospheric Research.
If you look closely at KILN's skew-T, you'll see an "M" at 500 mb. The "M" marks the local moist adiabat that an air parcel "follows" as it rises above the lifting condensation level (LCL). Obviously, a little background for my readers is in order here. For the record, the lifting condensation level is the altitude at which net condensation is set to begin inside a rising parcel of air. In other words, a rising air parcel becomes saturated at the LCL. Above the LCL, where net condensation is ongoing and clouds develop (see photograph), the air parcel still expands and cools as it rises, but the release of latent heat of condensation now reduces the parcel's cooling rate. The moist adiabats on a skew-T correspond to these reduced cooling rates (dashed green curves; revisit the primer flash animation that I showed you in an earlier blog). Not surprisingly, the rate of cooling associated with moist adiabats is formally called the moist adiabatic rate. As you can tell from the various slopes of the moist adiabats, the moist adiabatic lapse rate is not a constant.
At any rate, I drew, in black on this zoomed-in version, the local moist adiabat that rising parcels followed between the LCL and EL over Wilmington, Ohio, at 12Z on January 30, 2013. In this case, rising air parcels were slightly negatively buoyant (parcels were slightly colder than their environment). As a general rule, the typical environment conducive to a cold frontal rainband is nearly neutrally stratified. The bottom line here is that CAPE, the convective available potential energy, is zero (or nearly so) for environments where narrow cold-frontal rainbands form. Note that CAPE was zero over KILN at 12Z on January 30, 2013 (along the top of the skew-T).
With little or no CAPE, how do updrafts in narrow cold-frontal rainbands get so strong? Recall that low-level updrafts can reach 10 meters per second = 22 mph. As it turns out, the updrafts of NCFRBs are primarily driven by vertical gradients of pressure perturbations, a topic that's highly mathematical and beyond the scope of this blog.
The 1430Z base reflectivity from Pittsburgh on January 30, 2013, shows the core-and-gap structure of the narrow cold-frontal rainband over eastern Ohio and western Pennsylvania. Larger image. Courtesy of NCDC.
One of the telling characteristics of narrow cold-frontal rainbands is their tendency to develop kinks and breaks that are sometimes described as a core-and-gap structure. To see what I mean, check out the 1430Z base reflectivity from the radar at Pittsburgh (above; larger image) on January 30, 2012 (2.5 hours after the radar image at the top of this page). These gaps result from meso-gamma vortices (horizontal swirls of air whose spatial scale is on the order of several kilometers) that fracture the updrafts along the surface cold front. These fractures create the alternating precipitation cores and gaps in the narrow cold-frontal rainband shown on the Pittsburgh radar above.
Where do these meso-gamma vortices come from? Good question. As it turns out, a narrow cold-frontal rainband lies in a zone of strong horizontal wind shear (a change in wind speed over a specified horizontal distance) associated with a low-level jet stream that lies ahead, and nearly parallel to, the surface cold front. Check out, below, the NAM model analysis of 850-mb streamlines and isotachs (in knots) at 12Z on January 30, 2013 (larger image). Note that the low-level jet stream paralleled the surface cold front and that 850-mb wind speeds exceeded 60 knots over Ohio and Kentucky. The dramatic change in wind speed over the horizontal distance between the western edge and the core of the low-level jet stream created horizontal wind shear and meso-gamma vortices (see my idealized schematic). You can see the footprint of the horizontal wind shear associated with the low-level jet stream in the 12Z NAM model analysis of 850-mb vorticity (a measure of the rate of spin) on January 30, 2013.
The 12Z NAM model analsis of 850-mb streamlines and 850-mb isotachs (in knots) on January 30, 2013, shows a low-level jet stream just ahead of, and parallel to, the surface cold front. Larger image. Courtesy of Penn State.
These meso-gamma vortices are likely the ones being amplified to tornado strength whenever narrow cold-frontal rainbands spawn a twister. Tornadoes in NCFRBs are relatively rare. Indeed, the most serious threat posed by narrow cold-frontal rainbands is damaging straight-line winds.
I hope you learned something new. Narrow cold-frontal rainbands are not discussed very often on television because the term, squall line, is routinely used instead. That's a shame because "squall line" just doesn't do a narrow cold-frontal rainband any justice, in my humble opinion.
Updated: 4:55 PM GMT on February 08, 2013
A A A
Questionable Observations at Scranton
One of the graduates from the Penn State online certificate program posted the METARS from Scranton, PA, at The Milli Bar (for PSU certificate students) on January 28, kicking off an interesting discussion about the sporadic jumps in temperature at the Avoca Airport (KAVP).
In a nutshell, the airport's temperature at 15Z (10 A.M. local time) was 44 degrees (see the image of 15Z station models below; larger image). Meanwhile, temperatures at nearby airports were in the 20s or low 30s. The catalyst for the graduate's question was a television meteorologist who attributed the anomaly to a temperature inversion and the fact that the Scranton airport was one of the highest in the region (highest elevation). A bit of "sleuthing" quickly revealed that the explanation was faulty, proving that you can't always believe what you hear on television or what you read on the Internet.
The 15Z station models on January 28, 2013. The temperature observation at the Scranton Airport (KAVP), 44 degrees Fahrenheit, was an outlier compared to nearby temperatures. Larger image. Courtesy of Penn State.
To refresh your memory about the prevailing weather pattern, a warm front extending eastward from a low-pressure system centered over Lake Michigan was approaching Pennsylvania (15Z surface analysis). Ahead of the warm front, warm advection above the earth's surface produced stratiform precipitation over the region (check out the 15Z mosaic of composite reflectivity). By warm advection I mean that winds were blowing from areas of warmer air toward colder regions. You can get a better sense for the warm advection by studying the 12Z NAM model analysis of 850-mb isotherms (in degrees Celsius) and 850-mb streamlines below (larger image). I circled the region where relatively large warm advection was occurring at 850 mb (roughly 1500 meters above the earth's surface).
The 12Z NAM model analysis of 850-mb isotherms, in degrees Celsius, on January 28, 2013. Dashed contours indicate isotherms whose values are less than 0 degrees Celsius. Larger image. Courtesy of Penn State.
So, with warm air arriving aloft and cold air hanging tough near the ground, a strong temperature inversion formed in the lower troposphere, as the 15Z Rapid Refresh model analysis of the temperature and dew-point soundings at Scranton showed. So the television meteorologist was correct in asserting that there was a strong, very stable temperature inversion over Scranton (and much of the rest of the region ahead of the warm front). But the truth of what he was selling on television ended there.
First, Scranton's airport is not one of the highest valleys as he suggested. The elevation of KAVP is 962 feet. In contrast, the elevation at the airport at nearby Mount Pocono (KMPO), is 1915 feet, almost 1000 feet higher than KAVP. Moreover, temperatures at KMPO on the morning of January 28 did not get out of the 20's (for proof, check out KMPO's meteogram). So Mount Pocono, which is nearly 1000 feet higher than Avoca, was below the inversion. So it didn't make any sense to me that morning temperatures in the 40s at Avoca meant that KAVP was in the inversion (as advertised on TV).
Clearly, the television meteorologist's claim that KAVP was so warm because it's high elevation put it squarely in the temperature inversion was, for lack of a better word, hogwash. So he made his assertion without looking at a skew-T and without looking closely at surrounding observations.
I suspected that this was a case of instrument error. One of the certificate graduates sent an e-mail to the person who maintains the ASOS station at Avoca, who then sent a notice to the technicians at WFO BGM (the ticket number was 130130-137). It was promptly fixed.
Site Trouble Ticket Update
Site Identifier:aKAVP Office:aWFO BGM (Binghamton, New York) << Back
Location: a Wilkes-Barre/Scranton, PA AOMC/RMM ID: a AVP
Class: C2T Towered SvcLvl: C
Agency: NWS Region: Eastern
AFOS PIL / WMO Header: PHLMTRAVP LongLineComms: ADAS
ALDARS Site: No
Comments: ZNY (New York) ADAS/External
Number: 130130-137 Opened: 01/30/2013 18:42 GMT
Updated Trouble Ticket Information
Updated Problem Category(required when closing): Temp
Updated Problem Description(required): Closed ticket. Tech (DE) WFO BGM (Binghamton, New York) replaced temp sensor aspirator housing. sensor ops checked good.
Actual Return to Service: 01/30/2013 18:10 GMT
When I was teaching, I always tried to instill in students a scientific curiosity that motivated them to question things that they heard on television or read on the Internet. Being skeptical, within reason, is an attribute of any good scientist. And looking at real data to confirm or disprove what they hear or read is another quality I always tried to nurture in students. So I was impressed by how certificate graduates questioned and then discovered the truth about Scranton's temperature observation, even though a professional meteorologist had already offered an explanation.
And, yes, I also make mistakes sometimes, and my students are usually happy to point them out to me. :-)
Updated: 9:48 PM GMT on February 02, 2013
A A A