About the Coastal Occlusion on December 27
The secondary low-pressure system that developed east of the Appalachians and dumped heavy snow on northern New England on December 27 (snowfall map) occluded and deepened early on Thursday. For confirmation, compare the central pressure of the occluding low at 06Z surface analysis on December 27 with the barometric reading at 09Z. Note that the pressure fell from 989 mb to 987 mb in three hours.
The 09Z surface analysis on December 27, 2012. Note the central pressure of the low along the New Jersey Coast was 987 mb, 2 mb lower than it was 06Z (despite the fact that the low had entered the early stages of occlusion). Courtesy of HPC.
So the low-pressure system intensified during the early stages of occlusion (not unusual). Aloft, the closed 500-mb low associated with the negatively tilted short-wave trough was moving toward a vertical alignment with the surface low (compare the 12Z NAM model analysis at 500 mb with the 12Z surface analysis).
The movement toward a vertical "stacking" is characteristic of the occlusion stage (note the vertical alignment of the surface low and the 500-mb low at 12Z). Even after the vertical stacking takes place, the central pressure of the occluded low is still pretty close to its minimum central pressure. Indeed, check out the 18Z surface analysis and note that the minimum barometric reading was 988 mb.
When the 500-mb low and surface low are vertically stacked, the upper-level divergence that had supported the surface low during its mature state and the stages of early occlusion has all but vanished. Keep in mind here that the upper-level divergence that sustains the surface low typically occurs to the east of the 500-mb short-wave trough. So, when the system is vertically stacked, the surface low is flat out of luck when it comes to upper-level divergence.
How does the occluding surface low respond as it starts to run out of upper-level divergence? Simple answer: It starts to move back into the cold air. At first glance, such a move seems self-defeating. Why's that? Well, above the center of a low-pressure system, the air column that extends from the earth's surface to the top of the atmosphere weighs the least compared to all the neighboring air columns. Keep in mind that, when the atmosphere is hydrostatic, surface pressure corresponds to the weight of the overlying air column (for convenience, we assume the column has unit cross-sectional area).
At any rate, the column of air with the lowest weight (above the center of a low) has the highest average air temperature (more precisely, the highest virtual temperature) compared to neighboring air columns. In turn, the column with the highest average temperature also has the lowest average air density (which is consistent with the lowest weight). The bottom line here is that an occluding low moving back into the cold air, which has a higher density than warm air (at the same pressure), appears, at first glance, to be counterproductive for a low's well being.
An idealized cross section showing the movement of an occluding low back into the cold air and the tropopause undulation (TUL) associated with a lower tropopause in the cold air mass behind the low and a higher tropopause in the warm air mass ahead of the low. Larger image. Courtesy of the American Meteorological Society.
So how can an occluding low move back into the cold air and still have its central pressure decrease (like the low along the Northeast Coast on December 27)? To answer this question, check out the idealized cross section of an occluding low above (larger image) (from Tropopause Undulations and the Development of Extratropical Cyclones by Paul A. Hirschberg and J. Michael Fritsch). The rightmost "L" along the bottom represents a surface low about to enter the occlusion stage at some time, t sub 0. The two "L"s to the left indicate the position of the occluding low at two later times. Back deep in the cold air mass, the tropopause, indicated by the thin dark curves (solid and dashed) roughly between 200 mb and 500 mb, lies at a relatively low pressure altitude. I say "relatively" because the tropopause is higher to the east of the low (in the warm air mass). Meteorologists call these variations in the heights of the tropopause a tropopause undulation.
To understand why the tropopause lies at a lower altitude in the cold air mass, recall that pressure decreases relatively rapidly with height.
In the cold air mass, sinking motion associated with collapsing heights prompts air to sink from the lower stratosphere, dramatically warming on descent. This dramatic warming of sinking stratospheric air creates a warm pocket at roughly 200 mb (and higher up). In turn, winds at 200 mb blow warm air at 200 mb eastward. This warm advection at 200 mb transports relatively warm air into the column of air over the center of the occluding low. The arrival of warm air at 200 mb over the center of the low increases the average mean temperature in the air column over the low. Thus, the average air density in the column decreases. In other words, the air column over the center of the low loses some weight, and, as a result, the surface pressures decreases (even as upper-level divergence vanishes).
So the low cuts back into the cold air, making it seem like an ill-advised move. But the low is actually moving back toward the warm pocket of air at 200 mb, banking on warm advection at this altitude to increase the mean column temperature and, thus, lower the mean column air density and weight. Eventually, the low moves far enough back into the cold air where warm advection at 200 mb ceases. Now the stage is set for the low to rapidly weaken and eventually dissipate.
The 12Z GFS model analysis of 200-mb isotherms (dashed contours in degrees Celsius) and 200-mb streamlines (thin blue contours with arrows). Larger image. Courtesy of Penn State.
Let's get back to the coastal low on December 27. Revisit the 12Z surface analysis. Okay, let's look (above) at the 12Z GFS analysis of 200-mb isotherms (dashed, green contours in degrees Celsius) and 200-mb streamlines (thin, blue contours with arrows). Larger image. Note the pocket of relatively warm air off the DelMarva Peninsula (inside the closed contour of minus 48 degrees Celsius). Note how 200-mb winds blew warm air toward the surface low (warm advection), helping the surface low to deepen (or to maintain its strength) as it moved back into the cold air during the early stages of occlusion.
Lee Grenci
|
Updated: 4:20 PM GMT on December 31, 2012
|
Permalink |
A A A
|
Christmas Day Severe Weather
When teaching students how to forecast severe thunderstorms (like the outbreak on Christmas Day...see storm reports below), I've always stressed that they try their best to get a comprehensive handle on the large-scale weather pattern before they start to weigh the mesoscale details. Such a "big-picture" approach means looking at the surface for boundaries, areas of low-level confluence, etc.), 850 mb (low-level jet streams, temperature advection, etc.), 500 mb (short-wave troughs, wind maxima, etc.), and 300 mb (jet streaks, diffluence, etc.).
With regard to 300-mb jet streaks (or 250-mb in the warm season), I pretty much throw out the idealized (and sometimes unrealistic) four-quadrant conceptual model for a straight jet streak. Yes, I realize that this conceptual model is sometimes touted as gospel. Toward the end of my teaching career, however, I presented the four-quadrant conceptual model, but then I told my students to use it at their own risk whenever severe weather was lurking, especially in the case of curved jet streaks.
The storm reports for Christmas Day, 2012. Courtesy of the Storm Prediction Center.
Why would I take such an approach? A seasoned forecaster, Jack Hales, stated (probably more than once) that "People have died in the wrong jet quadrant." By this he meant that low-level uplift and any subsequent initiation of severe thunderstorms are not always constrained to occur in the two most statistically favored quadrants (right-entrance and left-exit regions).
Let's assume that there's an upper-level jet streak slated to pass over a region where ingredients in the lower troposphere appear to be coming together for an outbreak of severe thunderstorms. Focusing my attention on the corresponding quadrant of the 300-mb jet streak (the quadrant above where the "low-level ingredients" are coming together), I look for reasons why this specific quadrant might not be favorable for the development of storms. In other words, I automatically assume from the get-go that this quadrant will support deep, moist convection. Then I look for reasons why it might not be favorable. If I can't find any good reasons why it can't, red flags immediately go up in my mind.
Such was the case for the outbreak of severe weather over the Deep South on Christmas Day. A cyclonically curved jet streak (00Z NAM analysis of 300-mb isotachs, valid at 7 P.M. CST on December 25) was rounding the base of a vigorous short-wave trough. To see this relationship, check out the 00Z NAM model analysis of 300-mb isotachs (color filled this time) and 300-mb heights (lime green contours). Note that the streak's maximum wind speeds were greater than 120 knots at this time.
The 00Z NAM model analysis of 300-mb isotachs (green contours) on December 26, 2012 (7 P.M. CST December 25) and the corresponding 00Z mosaic of composite reflectivity. Larger image. Courtesy of Penn State.
The short wave supported a surface low-pressure system and its associated cold front (see 00Z surface analysis). To convince yourself that you can throw out any preconceived notion about the infallibility of the four-quadrant model of a straight jet streak, check out (above) the 00Z NAM model analysis of 300-mb isotachs and the corresponding mosaic of composite reflectivity (larger image). Yes, the severe thunderstorms occurred in the right-exit region, where the fallible four-quadrant model "maintains" that upper-level convergence occurs (and, concomitantly, downward motion beneath the right-exit region).
When I added the 300-mb streamlines to the isotachs (00Z NAM analysis), a diffluent pattern emerges (note how the streamlines in the vicinity of the jet streak spread out). Above the earth's surface, diffluent streamlines don't automatically translate to divergence. That's because the wind vectors can produce speed convergence (see idealized schematic), so we have to be careful here. Looking at the 00Z NAM model analysis of the 300-mb vertical motion field (blue contours) in concert with the 300-mb isotachs (green contours), we can see that there was upward motion in the right-exit region (so the diffluent pattern was divergent). To get your bearings, I point out that the dashed, blue contours represent negative values (in microbars per second) of upward motion, which is consistent with pressure decreasing with time. The elongated area of upward motion at 300 mb corresponds to the area where severe thunderstorms were occurring.
Here's a summary based on the NAM model analysis of the weather pattern at 00Z and the corresponding radar data. Note the thunderstorms occurred in the diffluent right-exit region of the 300-mb jet streak.
To be fair, note the bull's eye of strong upward motion in the left-exit region of the jet streak over Arkansas. Still, the lesson learned here is that the four-quadrant model is highly idealized and sometimes fails in outbreaks of severe thunderstorms like the one on Christmas day.
Here endeth the lesson.
Lee Grenci
|
Updated: 2:30 PM GMT on December 27, 2012
|
Permalink |
A A A
|
In the Chill of the Night
The winter storm that ended the record string of days without measurable snow in Chicago and several other Midwestern cities (see Jeff Masters' blog on December 19) affords me the opportunity to present a lesson on predicting nighttime temperatures using Model Output Statistics (MOS). For starters, the image below displays the snowfall that fell across the Upper Middle West on December 19-20. Note the large gradient in snowfall across northern Illinois and southern Wisconsin.
Snowfall for December 19-20, 2012. Courtesy of the National Weather Service at La Crosse, WI.
For this exercise, let's focus our attention on Joliet, Illinois (where a measly few tenths fell during the storm), and Madison, Wisconsin, where a 16.6 inches were measured during the period, December 18-20.
The meteograms for Joliet and Madison leading up to the night of December 21-22 show the clearing at both cities in the wake of the storm. I circled the 21Z (3 P.M. CST) temperatures on each meteogram to indicate the daytime differences, which were due, in part, to higher ground temperatures (cold advection was also a player; see the 3-hour GFS forecast of MSL isobars and 1000-mb isotherms, valid at 21Z).
Not surprisingly, the ground at Joliet was likely bare (or nearly so), allowing for more absorption of solar radiation. The bottom line is that, going into the nighttime hours, the daytime surface air temperature was higher at Joliet than Madison, which is always an important forecast consideration when predicting nighttime minima at places where there's snow cover versus bare ground. There's another factor forecasters always weigh in situations such as these. But first, let's take a look at Model Output Statistics at Joliet and Madison. Then I'll present some background on MOS before we get to the second issue I just mentioned.
Model Output Statistics for Joliet, Illinois, based on the 12Z run of the NAM on December 21, 2012. The predicted minimum temperature for the night of December 21-22 was 15 degrees Fahrenheit. Courtesy of NOAA's Meteorological Development Lab.
Check out the Model Output Statistics (MOS) for Joliet (above) and Madison (below) from the 12Z run of the NAM on December 21. I circled the MOS predicted minimum temperatures at both cities during the night of December 21-22 (15 degrees and 8 degrees Fahrenheit, respectively).
By way of introduction (or review if you're familiar with MOS), Model Output Statistics is a set of forecasts based on statistical regression equations derived for each airport (or observing station). These regression equations are created by relating weather observations to model data (I note that the precipitation regression equation is regional, but all other equations apply strictly to each airport).
Model Output Statistics for Madison, Wisconsin, based on the 12Z run of the NAM on December 21, 2012. The predicted minimum temperature for the night of December 21-22 was 8 degrees Fahrenheit. Courtesy of NOAA's Meteorological Development Lab.
To get these statistical forecasts, predictors are plugged into the regression equations (these predictors are raw output from the dynamic models). The most frequently selected predictors of temperatures, for example, are typically temperatures, dew points, and relative humidity at two meters and 1000 mb (again, raw data from the dynamic models). The list of possible temperature predictors include boundary-layer temperatures, low-level thickness, etc.
The bottom line is that a group of temperature predictors are plugged into regression equations to get MOS (statistical) forecasts for maximum temperature, minimum temperature, and temperatures at the synoptic times (00Z, 03Z. 06Z, etc.).
The advantages of using MOS instead of the raw output from the dynamic models is that there are often imperfect parameterizations that attempt to model 2-meter temperatures, for example. So the statistical predictions of MOS are, over the long haul, superior to output from the dynamic models. Looking to the future, I think MOS's days might be numbered, however, as ensemble forecasts that remove model biases improve (more on this topic in a later blog).
For sake of comparison, the GFS MOS predictions for the minimum temperatures at Joliet (GFS MOS for KJOT) and Madison (GFS MOS for KMSN) on the night of December 21-22 were 9 degrees and 10 degrees Fahrenheit, respectively.
As it turns out, MOS tends to have a high bias for nighttime minima on clear nights with light winds and snow cover (typically the bias becomes a forecasting issue when there's at least a couple of inches of snow on the ground, especially at airports where snow cover is not commonplace during winter).
So, during the daytime hours of December 21, I looked at NAM and GFS MOS for Madison, Wisconsin, and, with the bias in mind, immediately decided to lower the forecast for the nighttime low at Madison because both NAM MOS and GFS MOS at KMSN predicted a clear sky with light winds from 00Z to 12Z. To see that conditions were ideal for net radiational cooling, check out the MOS data to the right of "CLD" (Cloud Cover) and "WSP" (Wind Speed, in knots). Light winds and a clear sky were consistent with a sprawling ridge of high pressure slated to drift eastward during the night...check out the 00Z surface analysis (7 P.M. CST).
Taking snow cover into account was a good move because the nighttime minimum at Madison was 5 degrees Fahrenheit. For confirmation, check out the METARS at Madison below. The 2-group at 12Z indicates the observed minimum temperature between 06Z and 12Z (-15.0 degrees Celsius = 5 degrees Fahrenheit). I point out that most lows occur between 06Z and 12Z east of the Rockies (the low temperature sometimes occurs after 12Z in the Central Time Zone, which is why I included the 13Z and 14Z METARS).
The 2-group on the 12Z METAR at Madison, WI, on December 22 indicates that the low temperature at KMSN between 06Z and 12Z was -15.0 degrees Celsius = 5 degrees Fahrenheit. The temperatures after 12Z and before 06Z were higher. Courtesy of NOAA.
Meanwhile, the low at Joliet was 15 degrees (see the relevant METARS at KJOT; NAM MOS had a pretty darn good forecast). As I stated earlier, part of the difference between the nighttime lows was related to snow cover over Madison and bare ground (or nearly so) over Joliet (temperatures were higher over Joliet as sunset approached).
The second reason focuses squarely on properties of snow. The thermal conductivity of snow is appreciably lower than that of the ground. With this observation in mind, here's what happened. After sunset, the surface of the snow cooled by net radiation (net radiational cooling). Given the low conductivity of the snow, its surface temperature decreased without much of a compensating upward energy flux from below. The soil, however, with its higher conductivity, cools in such a way that its temperature is more uniform. Indeed, it's higher conductivity allows more of an upward flux of energy from soil below the surface, partially offsetting the net radiational cooling at the surface.
So we had two factors at work. The temperature of the bare ground at Joliet during the late afternoon was appreciably higher than the snow temperature at Madison. This gap in surface temperatures increased after dark because the surface temperature of the snow dropped more rapidly than the soil temperature.
On clear nights with light winds, the air temperature at two meters is largely determined by the temperature of the underlying surface. Thus, the stage was set for the minimum temperature at Madison to be dramatically lower than the minimum at Joliet. And, of course, knowing about the high bias of MOS on clear nights with light winds and snow cover alerts forecasters to adjust MOS predictions downward.
Here endeth the lesson.
Lee Grenci
|
Updated: 4:47 PM GMT on December 24, 2012
|
Permalink |
A A A
|
The Flying Circus of Physics
Beell asked me to follow up on my tuna-can experiment using RO water (reverse osmosis). As I mentioned, my wife put a hold on Lee's experiments until after Christmas.
If the truth be told, she probably remembers my last freezer experiment when, after a hailstorm after dark, I took a flashlight and crawled around on my hands and knees in the front yard looking for hailstones with an insect entombed at its core (updrafts in strong thunderstorms can sweep insects to high altitudes, where water deposits onto bugs and freezes). Needless to say, she was not very enthusiastic about frozen insects being stored in her freezer (she said it "bugged" her). :-) The bottom line is that my freezer experiments are not very popular in the Grenci household.
As you might have read in the comments section below my tuna-can blog, I told beell that I'm expecting more drops to remain unfrozen after several tens of minutes in my freezer. I'll let you know what happens.
Beell's interest in my experiments reminded me of one of my all-time favorite books, Jearl Walker's The Flying Circus of Physics. I just love this book!
One of Walker's experiments has always intrigued me. I bought his book at least 20 years ago, but, alas, I confess that I've never had the time or drive to try the experiment. To my credit, I've always had it on my list of fun things to do (at 65, I better get my rear in gear). My procrastination aside, Walker's experiment goes something like this...
Grab two identical containers (cups, etc.). By identical, I mean they both must have exactly the same thermal characteristics. Fill one with cool water (I think the temperature matters here, but I'll have to look this up) and the other with boiling water. Great care must be taken to insure that both containers have exactly the same amount of water, so measure, measure, measure! Now place both containers outside on your porch on a cold winter night. You must pay close attention because you want to determine the time it takes for the water in one of the containers to freeze. So you just can't walk away and watch a movie or order a pizza. In other words, you must check your experiment frequently.
Which container freezes first? You might be surprised to learn that the boiling water should freeze first. Can anybody explain why?
Lee Grenci
|
Updated: 8:20 PM GMT on December 22, 2012
|
Permalink |
A A A
|
Snowstorm and Severe Weather
I woke up this morning to the sound of my neighbor's trash can clanking around...the black bear that frequented our neighborhood last winter is back. The bear ripped down some bird feeders, and, fortunately, left our fence alone. We're expecting more bear close encounters because the mild weather so far in the East likely means that the black bear won't hibernate (it didn't last winter). We look at the bear visits as a natural consequence of winter life here in central Pennsylvania.
Although I could do without the bear this winter, I'm missing the snow and cold of central Pennsylvania. So I'm quite envious of the folks in the Upper Middle West, where a white Christmas now looks likely. At 12Z this morning (surface analysis below), a mature 988-mb low was centered to the north of St. Louis.
The 12Z surface analysis on December 20, 2012. Courtesy of HPC.
Snow was falling to its north and severe thunderstorms to the south (12Z mosaic of composite reflectivity; cropped version to show severe storms). Some of the storms over the Deep South were discrete and semi-discrete supercells, prompting the Storm Prediction Center to issue a tornado watch earlier in the morning.
The bulk vertical wind shear between the ground and six kilometers (roughly 500 mb) was greater than 40 knots (12Z Rapid Refresh model analysis of bulk shear), more than sufficient to support rotating updrafts in thunderstorms (supercells). A robust short-wave trough at 500 mb...here's the 12Z NAM model analysis of 500-mb heights (solid, green contours) and 500-mb isovorts (dashed, blue contours indicate isopleths of absolute vorticity) provided upper-level divergence for the surface low-pressure system. In the context of vertical wind shear, there was a speed maximum associated with the 500-mb trough (12Z NAM model analysis of 500-mb isotachs, in knots), which helped to produce strong shear in the lowest six kilometers.
The risk of tornadoes was heightened by a strong low-level jet stream (check out the 12Z NAM model analysis of 850-mb heights and 850-mb isotachs, color-filled in knots. Indeed, 850-mb wind speeds topped 70 knots over the Tennessee Valley, with an elongated maximum stretching southward to the Gulf of Mexico. Note also that the Gulf of Mexico was wide open for business, with a tongue of relatively high values of precipitable water extending northward ahead of the low-pressure system (12Z NAM model analysis of PWAT, in inches). For the record, Gulf moisture was likely getting involved in the moderate to heavy snow falling over parts of the northern Great Lakes and Upper Mississippi Valley...note the westward turn of the 850-mb streamlines over Wisconsin and surrounding areas.
Although the height gradient at 850 mb was fairly strong (12Z NAM model analysis at 850 mb), the low-level jet stream was also driven by a jet streak at 300 mb (see 12Z NAM model analysis below).
The 12Z NAM model analysis of 300-mb isotachs on December 20, 2012. Courtesy of Penn State.
Indeed, 850-mb heights fell under the left-exit region of a cyclonically curved, 110-knot+ jet streak at 300 mb (where upper-level divergence was favored). For convenience, I circled the region (approximate) where 850-mb height falls occurred...check out this annotated 12Z 300-mb chart of isotachs). So there was an ageostrophic contribution to 850-mb wind speeds (850-mb winds tended to accelerate toward the region of 850-mb height falls under the left-exit region of the 300-mb jet). In this way, the two jet streams (polar jet at 300 mb and low-level jet at 850 mb) were coupled.
At any rate, the strong low-level wind shear associated with the low-level jet stream heightened the risk of tornadoes spawned by discrete to semi-discrete supercells over the Deep South.
I should point out that thunderstorms tend to be discrete (or semi-discrete) when low-level convergence is relatively weak. When convergence is strong, however, storms tend to organize into lines (squall lines) or mesoscale convective systems. Here's the 12Z Rapid Refresh model analysis of convergence in the lowest few kilometers of the troposphere (red contours indicate convergence). Note the line of strong convergence associated with the squall line, while, farther east, convergence was weaker and less organized, creating an environment suitable for discrete or semi-discrete thunderstorms (some tornadic).
Lee Grenci
|
Updated: 11:51 AM GMT on December 21, 2012
|
Permalink |
A A A
|
Experiments on a Tuna Can
Whoa Nellie! I guess I caused a stir by announcing that 32 degrees Fahrenheit is not "freezing." To repeat, only pure water freezes at 32 degrees (at a pressure of one atmosphere). Moreover, finding pure water is rather difficult. Yes, even distilled water contains some impurities (contamination by particles in the air, etc.). Even the water from the reverse-osmosis system that my wife and I use regularly for cooking and drinking isn't pure.
Craig Bohren, now a retired professor from Penn State, performed an experiment which I duplicated for my textbook. I took an empty tuna can, greased it with oil, and then placed drops of tap water on the can (see photograph on the left below). Please note that I first boiled the water in an attempt to remove any dissolved air. At any rate, I placed the can in my freezer (about -11 degrees Celsius, which roughly equals 12 degrees Fahrenheit) and waited ten minutes...tick, tock, tick, tock...
(Left) Water drops on the bottom of a tuna can. (Right)After ten minutes in my freezer (-11 degrees Celsius), four drops resisted freezing. Courtesy of A World of Weather: Fundamentals of Meteorology
I removed the can, and, lo and behold, there were several drops that did not freeze (see photograph on the right above). Back into the freezer. A couple more drops froze. Even after another ten minutes, two drops refused to freeze. I can only deduce that the drops lacked freezing nuclei. I plan to repeat the experiment using reverse-osmosis water. I'm expecting more drops to resist freezing.
And so it is with tiny water drops that exist in high, cold clouds. Despite the temperature being well below 0 degrees Celsius, they resist freezing because they are bereft of freezing nuclei. At temperatures below -40 degrees Celsius, however, all bets are off, and the tiny drops spontaneously freeze.
A short blog, I admit, but I hope I presented sufficient empirical evidence that 32 degrees Fahrenheit is not the freezing point of water.
Here endeth the lesson.
Lee Grenci
|
Updated: 7:02 PM GMT on December 19, 2012
|
Permalink |
A A A
|
Frost and Freezing
georgevandenberghe broached the issue of frost in his reply to my blog yesterday (many thanks, George). Any mention of frost, as it relates to gardening, is almost always perceived in a pejorative context. Indeed, there are routinely a spate of frost advisories in spring and autumn around the beginning and end of the growing season (respectively). Portraying frost as a killer of tender plants always bothers me because frost is actually Nature's way of trying to protect them. Yes, frost gets an undeserved rap, in my opinion.
By definition, frost typically forms on clear nights with light winds and relatively low dew points as ice crystals deposit on the ground and other objects (tender plants, car wind shields, etc.). Frost can form when two-meter temperatures are as high as the mid 30's. That's because temperatures at ground level on such nights are sufficiently far below the melting point of ice (in other words, there's a nocturnal inversion extending from the ground upward).
In case my reference to the "melting point of ice" (instead of "freezing") caught you off guard, I confess that I stubbornly refuse to call 32 degrees Fahrenheit (0 degrees Celsius) "freezing." For the record, water almost always freezes at temperatures lower than 32 degrees Fahrenheit because impurities in the water introduce a "disorder" that interferes with the formation of the highly ordered lattice of ice. Only pure water freezes at 32 degrees Fahrenheit, and distilled water doesn't freely exist in Nature. Forgive me, but I regularly shake my head when I hear or read 32 degrees Fahrenheit being described as "freezing." I'm afraid it's hopelessly ingrained in our culture (more on freezing and my disdain in just a moment).
Let's get back to the issue of frost. When water vapor deposits as frost on a tender plant, latent heat of deposition acts to keep the plant warmer (a release of 680 calories per gram; see the energy staircase below). As long as deposition occurs, the temperature of the plant will not precipitously fall into the 20's. At such readings, ice crystals can form inside plants...if plant cells freeze, they can rupture and then there's real damage.
The energy staircase for water. When water vapor deposits onto objects as ice, 680 calories per gram are released (latent heat of deposition). Larger image Courtesy of A World of Weather: Fundamentals of Meteorology.
Are the owners of orchards crazy to spray their tender plants in order to avoid such damage? Facing the loss of their crop, would they willingly exacerbate matters? Orchard owners don't have to know squat about science. All they need to know is that spraying water on plants offers them some protection against dangerously low temperatures. And so it is with frost, which is Nature's way of trying to protect plants (via the release of latent heat of deposition). Killing plants is just not even in frost's purview.
Granted, frost advisories can alert gardeners that temperatures might become threateningly low (I have no problem with this kind of interpretation). Unfortunately, I suspect that most people interpret frost advisories literally...it's the frost that will kill. If gardeners perceive frost as a killer of plants (the appearance of ice crystals on the exterior of plants), please rid your brain of this notion.
Which leads me back to the issue of 32 degrees Fahrenheit being called "freezing." Honestly, meteorologists should know better. Almost all precipitation that falls over the middle latitudes begins as snow at cold, high altitudes. Here, ice crystals grow at the expense of surrounding tiny water drops. Yes, tiny water drops can resist freezing down to minus 40 degrees Fahrenheit (which equals minus 40 degrees Celsius). So I take issue with meteorologists who, on one hand, accept the Bergeron-Findeisen process (and its underlying assumption that tiny water drops resist freezing down to extraordinarily low temperatures), but who, on the other hand, describe 32 degrees Fahrenheit as "freezing." Like it or not, 32 degrees Fahrenheit is the melting point of ice. Indeed, all ice begins to melt at 32 degrees (actually, ever so slightly above 32 degrees), but water does not freeze at 32 degrees Fahrenheit unless it's pure (not present in Nature).
There, I feel a lot better.
Lee Grenci
|
Updated: 11:16 PM GMT on December 18, 2012
|
Permalink |
A A A
|
Elevated Thunderstorms
Early this morning, there was a mesoscale convective system (MCS) over southern Alabama (check out the the 12Z mosaic of composite reflectivity below; larger image). For the record, the thunderstorms associated with this MCS were elevated (as opposed to surface-based). Before we look more closely at this morning's elevated MCS over the Deep South, let's tackle some basic meteorology.
The 12Z mosaic of composite reflectivity at 12Z on December 17, 2012. Larger image. Courtesy of Penn State University.
For the record, an elevated thunderstorm (elevated convection) is a type of deep, moist convection whose updraft originates at some altitude above the ground (usually in the lowest three kilometers of the troposphere). In contrast, updrafts associated with surface-based convection originate at the ground.
To gain a better sense for elevated thunderstorms, check out this flash animation (courtesy of Penn State's certificate program) that shows the temperature sounding favorable for elevated thunderstorms (you should probably keep the flash open while you read the rest of this paragraph). As a general rule, elevated thunderstorms develop above a stable layer of air next to the ground (above a nocturnal inversion or above a stable layer on the cool side of anafronts). For this example, there's an inversion between roughly 870 mb and 740 mb (pressure located along the vertical axis). Okay, choose any air parcel (the blue, square-like object on the skew-T) anywhere between 870 mb and 740 mb. So that we're all on the same page, let's choose 825 mb. So click and drag the red arrow on the right of the image and lift it to 825 mb (there's a little tic mark midway between 800 mb and 850 mb). The blue line through the air parcel corresponds to the air parcel's temperature at 825 mb. Note that the parcel's temperature is lower than it's environment (the blue line lies to the left of the red environmental temperature sounding). At most, we would get a stratiform cloud near this altitude, not a cumuliform cloud like a thunderstorm. Indeed, any air parcel lifted from a pressure altitude between 870 mb and 740 mb would be negatively buoyant. In other words, it would resist rising through a deep layer of air.
But what if the air parcel at 750 mb was lifted from this level? Well, let's find out...click and drag the red pointer up to 750 mb. Now the temperature of the air parcel would be higher than the temperature of its environment (the orange line marking the parcel's temperature lies to the right of the red environmental temperature sounding at this pressure altitude). In other words, the air parcel at 750 mb would be positively buoyant and the stage would be set for the parcel to rise to great altitudes, thus contributing to the formation of a cumuliform cloud and, perhaps, a thunderstorm. Stated another way, the updraft of this thunderstorm would not begin at the ground but at 750 mb. Such a storm would indeed be elevated rather than surface-based.
Okay, now we're ready to tackle the issue of the elevated MCS over southern Alabama this morning. Remember that one of the places where elevated thunderstorms form is on the cool side of fronts. Well, there was a stationary front undergoing frontogenesis (increasing horizontal temperature gradient) at 12Z this morning (check out the 12Z surface analysis). Moreover, there was a low-level jet stream near 850 mb with speeds as high as 35 knots... check out the 12Z NAM model analysis of 850-mb streamlines and 850-mb isotachs below (color-filled in knots; larger image).
The 12Z NAM analysis of 850-mb streamlines and 850-mb isotachs (color-filled in knots). Larger image. Courtesy of Penn State.
In turn, this relatively fast-moving current of air produced a bull's-eye of moisture convergence over southern Alabama (the pocket of closed, red contours over Alabama on this 12Z Rapid Refresh model analysis indicate moisture convergence). Given that moist air was converging near 850 mb, the atmosphere, which works to mitigate any congestion of air, responded by producing upward motion, paving the way for elevated thunderstorms.
To seal the deal on the elevated nature of these storms, check out the 12Z NAM model temperature (red) and dew-point (green) soundings at a point in Alabama close to the center of the bull's-eye in moisture convergence and high composite reflectivity. Note the inversion near the ground. Yes, the inversion was rather shallow, so the thunderstorms here were only slightly elevated (but elevated nonetheless). Tornadoes are not usually associated with elevated thunderstorms that have rotating updrafts (too stable near the ground), but, when supercells are only slightly elevated, tornadoes can indeed occur.
Lee Grenci
|
Updated: 12:17 PM GMT on December 18, 2012
|
Permalink |
A A A
|
Adventures in Flying
The occluded low-pressure system over the North Atlantic that I talked about yesterday (in the context of open-cell convection over the ocean) also had a dramatic impact on aviation yesterday afternoon. Indeed, aircraft landing and taking off from the airport at Bilbao, Spain, had to cope with fierce wind gusts greater than 40 knots (I saw 43 knots on the Bilbao METARS; here's the relevant meteogram at Bilbao).
Check out this disturbing video (limited shelf life). Here's another video.
If you look closely at the two-minute sustained winds on the meteogram over Bilbao, speeds were relatively tame (roughly 20 knots) compared to the wind gusts (circled in black). These sustained winds were consistent with the relatively strong surface pressure gradient indicated on the 06Z surface analysis. The strong wind gusts, however, were consistent with the downward mixing of momentum from faster winds aloft by convective eddies (idealized schematic of mechanical eddies, courtesy of A World of Weather: Fundamentals of Meteorology).
Indeed, check out the GFS model analysis of the temperature (red) and dew-point (green) soundings at the Bilbao Airport (shown below).
The 06Z GFS model analysis of the temperature (red) and dew-point (green) soundings at Bilbao, Spain (LEBB) on December 14, 2012. Courtesy of Penn State.
Notice that the atmosphere was saturated (air temperature equals the dew point) between roughly 750 mb and 200 mb (the green dew-point sounding lies on top of the red temperature sounding so you only see green). Below 750 mb, there was a layer of unsaturated air (dew point less than the air temperature). As precipitation fell into the unsaturated air, precipitation evaporated, cooling the air in the layer. In turn, this cooling aloft enhanced negative buoyancy (more unstable), paving the way for the downward mixing of momentum of fast winds aloft to the earth's surface.
If you focus your attention on the vertical profile of winds on the right of the image (within the yellow ellipse), you see wind speeds as high as 55 knots. Of course, the downward transfer of momentum by convective eddies is not 100% efficient (energy is lost in the transfer), but momentum reaching the ground was sufficient to produce gusts of 43 knots or higher.
And making take-off and landing a real adventure.
Lee Grenci
|
Updated: 2:23 PM GMT on December 17, 2012
|
Permalink |
A A A
|
Open-Cell Convection
The surface analysis over the North Atlantic Ocean at 12Z today (courtesy of the Ocean Prediction Center) shows an occluded low-pressure system just south of Greenland. Extending eastward from its center, an occluded front stretched toward a triple-point low centered near Great Britain.
The corresponding 12Z visible satellite image from Meteosat shows the cloud structures associated with the occluded and triple-point lows (full disk). Note the enduring darkness of winter at high latitudes.
The most interesting features on the visible satellite image are the hexagonal, ring-like patches called open-cell cumulus clouds (or open-cell convection). To get a better sense for this open-cell convection, check out the cropped image below (here's a slightly pixelated, zoomed-in version).
A portion of the 12Z visible satellite image from Meteosat on December 14, 2012, shows an area of open-cell convection over the North Atlantic Ocean. Courtesy of NERC Satellite Receiving Station, Dundee University, Scotland.
The corresponding 12Z infrared satellite image from Meteosat (full disk) indicates that the convection associated with these open-celled cumulus clouds was relatively shallow (warmer cloud tops). As their shape and name suggests, open-celled convection appears as a ring of cumuliform clouds. Note that the centers of individual patches are indeed "open," indicating that air was sinking in the core of each cell and rising around the edges. Such a mesoscale-alpha pattern of vertical motion creates a lacy, hollow-looking array.
For the record, an individual ring in the array of open-cell convection has a diameter that typically ranges from 20 to 200 kilometers (so each cell is a meso-beta feature). Such intricate structures in cumulus clouds were unknown before the advent of weather satellites because the cell-like patterns were simply too small to be resolved by the fragmented observational network. And the closely packed mesh of clouds was just too large to be recognized from aircraft.
Open-cell convection forms over the oceans when the lower troposphere is unstable and local isobars (or height contours on a constant pressure surface) are cyclonically curved (courtesy of the Penn State Certificate of Achievement in Weather Forecasting). By way of clarification, cyclonically curved contours are a proxy for low-level convergence (which is consistent with synoptic-scale upward motion). The instability that supports this rather low-topped convection typically goes hand-in-hand with moderately strong cold advection over relatively warm water. Weather forecasters also look for winds with speeds greater than 25 knots in the marine boundary layer.
To seal the deal on the environment favorable for open-cell convection, check out the 12Z GFS model analysis (below; larger, annotated image) of 850-mb heights (black, thick contours), 850-mb wind barbs, and 850-mb isotherms (thin, colored contours). The circle I drew roughly indicates the area where open-cell convection occurred. The local 850-mb height lines are indeed cyclonically curved, 850-mb wind speeds are as high as 55 knots, and 850-mb wind barbs cross 850-mb isotherms from lower to higher values (cold-air advection). The bottom line is that all the criteria on the checklist for open-cell convection are met.
The 12Z GFS model analysis of 850-mb heights (black contours), 850-mb wind barbs, and 850-mb isotherms (thin, colored contours) on December 14, 2012. Larger image. Courtesy of the National Centers for Environmental Prediction
Just to convince myself, I retrieved the 12Z GFS model temperature and dew-point soundings (red and green, respectively) at 45 degrees North Latitude and 40 degrees West Longitude (where there was open-cell convection). Perhaps I'll teach beginners how to interpret skew-T's on a later blog. For now, it suffices to say that the temperature at 850 mb was -6.6 degrees Celsius and the temperature at 1000 mb was 6.1 degrees Celsius (essentially the model's sea surface temperature). The 850-mb height was 1349 meters. Thus, the lapse rate between 850 mb and the sea surface was 9.4 degrees per kilometer, which is strongly unstable.
Here endeth the lesson.
Lee Grenci
|
Updated: 1:14 PM GMT on December 15, 2012
|
Permalink |
A A A
|
Out of the Blue
Yesterday I gave you the sense that people often take shortcuts when they try to explain why nighttime temperatures can plummet on clear, calm nights with relatively low dew points. By way of review, any explanation that incorporates "a lot of radiational cooling" sends the wrong message because radiational cooling is greatest around the time of the daytime maximum temperature (radiational cooling is proportional to the fourth power of absolute temperature). That's why I prefer "nocturnal cooling," or, more precisely, "net radiational cooling," which indicates that the ground runs an energy deficit at night (radiational cooling exceeds the amount of energy absorbed by the ground).
Another "shortcut" that bothers me is the popular explanation for a blue sky. I think most of us were taught that the sky is blue because air molecules preferentially scatter blue light (in all directions). Hogwash! Before I reveal a more scientific explanation that doesn't take any shortcuts, let me first set the stage by examining the spectrum of energy emitted by the sun. This image displays the spectrum of electromagnetic energy according to wavelength, which is the distance from crest to crest (or trough to trough). Focus your attention on the rather limited range of visible light, whose wavelengths vary between 0.4 and 0.7 micrometers. For the record, a micrometer (or micron, if you prefer), is one-millionth of a meter.
Unlike similar figures in most textbooks, there isn't any specific color assigned to a specific wavelength...for good reason!
The spectrum of skylight measured by pointing a spectrophotometer at a cloudless sky away from the direct rays of the sun. Courtesy of A World of Weather: Fundamentals of Meteorology
To see what I mean, check out the spectrum of skylight (light from the sky) as a function of wavelength (above). I measured this spectrum by pointing a spectrophotometer at the blue sky (away from the direct rays of the sun). The screaming message from this spectrum of skylight is that all the wavelengths of visible light comprise what we see as a blue sky. In short, there really isn't any pure blue light in nature. Granted, this visible spectrum dramatically peaks in the shorter wavelengths, but, as you can see, there is no single wavelength associated with light from a blue sky.
Alas, "in nature" is an important qualifier here because we can rightfully consider a blue laser, for example, as a source of pure blue light. Don't let that dilute my message, folks. The spectrum of light from a blue sky contains all the wavelengths of visible light. Do a Google search on blue light and you'll find that there are a lot of references to 0.475 micrometers (or some other equivalent units). Just looking at my measurements from the spectrophotometer should convince you that assigning a specific wavelength to a primary color is not cool. Heck,the spectrum of skylight doesn't even peak at 0.475 micrometers.
My spectrophotometer measurements of skylight pave the way for disputing statements such as "Oranges are orange because they only reflect orange light." I'm sorry to say that such statements you probably learned in school are blatantly false. Indeed, check out my measurements of the spectrum of light from an orange (my wife thinks I'm nuts). Yes, the spectrum of light from an orange includes all the wavelengths of visible light. To be fair, your teachers never had a trusty spectrophotometer to make the correct scientific point.
To me, the most compelling aspect of skylight is that, despite its spectrum not peaking in the "blue" (0.475 micrometers according to some Web sources), we perceive skylight as blue. What's up with that?
Here's the REAL story. Air molecules (mostly oxygen and nitrogen) scatter all the wavelengths of visible light emitted by the sun (revisit the spectrum of skylight above). This "assortment" of wavelengths of visible light from a cloudless sky enter our eyes and get processed (integrated) by our brains. As children, we learned to call this brain-integrated spectrum of visible light, "sky blue" (or simply "blue").
Make sure you understand what I'm saying here. Technically speaking, our eyes don't see the sky as "blue." Rather, they merely act as collectors of light (skylight, in this case). Only our brains truly "see" the blue sky. If you're queasy about this explanation, I point out that we sometimes dream in color despite the absence of light striking our retinas.
The bottom line here is that the popular explanation of air molecules preferentially scattering blue light is not correct. Air molecules scatter all the wavelengths of visible light (albeit unevenly, as you observed above). Moreover, such an explanation removes our brains from the entire process. As I stated earlier, we perceive skylight as blue. Not to tell students (or readers) this irrefutable fact is to deprive them of knowledge about how their brain processes visual stimuli.
Here endeth the lesson.
Lee Grenci
|
Updated: 12:28 PM GMT on March 06, 2013
|
Permalink |
A A A
|
Arctic Air Masses
At 65, I'm one of those old timers who still enjoys cold and snowy weather. In the early 1980s while I was in graduate school at McGill University (Montreal, Quebec), I would chase lake-effect snow squalls downwind of Lakes Ontario and Erie, inserting myself into convective bands with snowfall rates of six inches per hour (or higher) with thunder and lightning. Those were the days, my friends.
Here in central Pennsylvania, cold weather has been hard to come by this autumn and this winter, so I looked enviously at the Arctic air mass that has dipped into the Upper Middle West this morning (09Z surface analysis).
Take a closer look at the 09Z temperatures and dew points on the station models over western and central Canada (09Z is 4 A.M. EST). With temperatures and dew points as low as minus 44 degrees and minus 47 degrees Fahrenheit respectively, there's no question that this is a genuine Arctic air mass. What is an air mass? Are polar air masses colder than Arctic air masses?
The source regions for air masses that affect North America. Credit: Courtesy of A World of Weather: Fundamentals of Meteorology.
For the record, an air mass is a large chunk of air with horizontal dimensions on the order of several hundred to a couple of thousand miles. Within any air mass, temperatures and dew points near the earth's surface (or at any other arbitrary altitude) vary only gradually with increasing distance away from the center of the air mass (a center of high pressure).
Winter's most frigid air masses, like the air mass shown on the 09Z surface analysis, are tagged continental-Arctic air masses (cA) to herald their extreme cold and very low dew points. There are a couple of tests for Arctic air masses. For starters, I avoid looking at nighttime temperatures because sometimes, nocturnal cooling can mislead you into thinking the air mass is colder than it really is (this happens sometimes in Nevada and other parts of the Intermountain West on clear, calm nights during the cold season).
No, I like to look at daytime temperatures in order to determine the presence of Arctic air. When Arctic air masses invade the United States, daytime high temperatures are typically in the single digits or they're below zero degrees Fahrenheit. Yesterday's high temperatures over central and western Canada (right panel) seal the deal on the Arctic nature of this air mass.
I also look at 850-mb temperatures (standard altitude is 1500 meters) to verify the presence of Arctic air. As a general rule, 850-mb temperatures equal to or lower than minus 20 degrees Celsius are consistent with Arctic air. Check out the 06Z GFS 3-hour forecast for 850-mb temperatures (valid at 09Z this morning) and note the large area over Canada where 850-mb temperatures were predicted to be minus 20 degrees Celsius or lower.
For your benefit, here's a close-up of the Arctic front on the 09Z surface analysis this morning, which marks the leading edge of Arctic air. In case you're wondering, the cold front shows where Arctic air was advancing, and the stationary front indicates where Arctic air stalled.
Despite its name, continental-Polar air masses (designated cP) are not as cold as continental-Arctic air masses during winter. Yes, I agree, the convention of using "polar" to describe these air masses is somewhat confusing, but revisit the source-region map above to drive home my point that continental-Polar air masses form farther equatorward than continental-Arctic air masses.
A Pet Peeve
Earlier, I made a reference to "nocturnal cooling." For the record, I always try to avoid "radiational cooling" whenever I'm discussing low nighttime temperatures. That's because I often hear people describe the reason for very low nighttime temperatures as "great radiational cooling." Aaaarrrgggghhhh...
If the truth be told, the greatest radiational cooling occurs around the time of the daytime maximum temperature. Indeed, radiational cooling is proportional to the fourth power of absolute temperature (in Kelvins), according to Stefan-Boltzmann's Law. So that's why I prefer "nocturnal cooling," a term that gets to the heart of the real reason for very low temperatures on a clear, calm night...yes, there is radiational cooling, but the amount of radiation from the atmosphere is relatively small. So the ground runs an energy deficit, and temperatures continue to fall until solar energy arrests the decline after sunrise. I'll have more to say about nocturnal cooling in future blogs.
So, yes, I'm sort of a stickler for the accuracy of science. There will be an underlying theme of science in all of my blogs, and I hope readers learn more about meteorology and atmospheric science as I delve into deeper topics.
I am very grateful for this opportunity.
Lee Grenci
|
Updated: 11:46 PM GMT on December 12, 2012
|
Permalink |
A A A
|
