The Weather Blog

Relationship between Solar Cycle Length and Global Temperature Anomalies

The most recent solar cycle, currently estimated to have reached a minimum in December 2008 based on a 12 month centered moving average of monthly sunspot numbers, was the longest solar cycle since 1798-1810, measured trough to trough.  The cycle lasted approximately 150 months or 12.5 years, two full years longer than the 20^th century average of 10.5 years.     The impact of a long solar cycle is two-fold.   First of all, there is a definite correlation between the length of one solar cycle and the peak sunspot number during the following cycle as the graph below shows.

Only two solar cycles since 1750 have lasted as long as or longer than the most recent cycle, and both were followed by the weakest solar maximums of the past 250 years, a period during the early 1800s known as the Dalton Minimum.   Given the length of the previous cycle, the upcoming solar cycle could be among the weakest on record since 1750 if the relationship holds.   Additionally, of the last five solar cycles that persisted twelve or more months longer than average, four were followed by another cycle lasting longer than average.  In every instance, however, the next solar cycle was somewhat shorter than the previous one in length.  

Since net solar radiation is slightly higher during periods of heightened sunspot activity (and lower during periods of little sunspot activity), the combination of long solar cycles and low sunspot numbers results in cumulatively more months on a decadal time scale with below average net solar radiation.   The result is that weak and long solar cycles lead to cooling on a global basis while short and intense solar cycles tend to result in warming on a global basis.  

The following plot illustrates the relationship between the length of a given solar cycle and the average temperatures during the following solar cycle.   Average global temperatures were calculated for each solar cycle (measured trough to trough) based on the GHCN-ERSST dataset (though the results are similar for the Hadley Center global temperature dataset) and compared to the length of the previous solar cycle.   As can be seen, there is a solid relationship between solar cycle length and average global temperatures during the following solar cycle.  This in turn leads to the possibility of using the solar cycle length as a means to at least partially predict temperature trends during the upcoming decade.

Besides the notable trend of cooler global temperatures following longer solar cycles and warmer global temperatures following shorter solar cycles, the other notable feature of the data is the distinctly higher temperature anomalies observed during the most recent three solar cycles.  A trend line fitted through the most recent three solar cycles is nearly parallel to a trend line through the previous nine solar cycles, but is displaced 0.3-0.4°C higher.   If the difference was purely the result of CO₂ increases, a steadily increasing departure from the previous trend would be expected, and not a distinct jump and then a continuation of the old trend at a warmer temperature level.  The timing of the data jump happens to be the mid-to-late 1970s, at a time of a noted global climate shift.  The Pacific Decadal Oscillation changed from a negative to a positive phase at that time, and could explain part of the discrepancy.  Improved ocean temperature estimates began during that time period as a result of satellite measurements, while at the same time a reduction in the number of land-based observation stations has occurred.   Both of those factors, along with issues such as urban heat island growth could also play a role in the observed differences.  There is also some chance that the apparent shift is simply an artifact of the limited number of data points.

Finally, note that the warm 1996-2008 period followed the shortest solar cycle during the last 150 years.  With the most recent solar cycle being the longest cycle of the past 200 or so years, the potential exists for significant global cooling during the upcoming decade.  Even if the warmer trends of the past 30 years are utilized, below average global temperatures would still be possible during the next decade if the past trends hold. 

While not quite as distinct as on a global scale, temperature trends for North America also have closely followed trends in solar cycle length as the plot below demonstrates.    While temperature anomalies will likely not decline as much as might be suggested by the solar cycle length curve, the long solar cycle could at least help to drive North America temperature anomalies back down towards the long-term average. 

In conclusion, the upcoming decade will provide key data for research into both solar impacts on global climate and human-induced global warming.   Since CO₂ levels have been rising during a time of increasing solar activity, untangling the impacts of each on the global climate system is difficult.  But, with rising CO₂ levels and decreasing solar activity during the upcoming decade, the impacts of both should become more evident. 

Data Sources

Global and North America Temperature Anomaly Data: http://www.ncdc.noaa.gov/gcag/index.jsp

 

Monthly Sunspot Number data: http://www.sidc.be/sunspot-data/

The Return of El Nino

It's official: El Nino is back. The warming taking place in the equatorial Pacific Ocean, which brought an end to La Nina earlier this year, has reached the degree that it now qualifies as an El Nino. The Climate Prediction Center issued their first advisory on the new El Nino on July 9.

The image at left shows the sea surface temperature (SST) anomalies across the Pacific basin, as of July 12 (click to open a full-size version in a new window). Most of the equatorial Pacific between Indonesia and the west coast of South America is painted in the reds indicating SSTs that are warmer than normal by 0.5-1.5 C. The SST anomalies near the South American coast are quite strong; the yellow shades indicate waters that are on the order of 2-3 C warmer than normal.

The SST anomalies, when averaged over a set of geographic regions within the equatorial Pacific, yield the standard temperature anomaly indices that are used to gauge the oceanic component of ENSO (El Nino-Southern Oscillation). Typically, the rule of thumb is that when these indices reach +0.5 C (or more), we are in El Nino. If the anomalies are -0.5 C (or less), we are in La Nina. In between is ENSO-neutral. Most of the SST indices are now in the range of +0.6 C to +0.9 C, so the situation now officially qualifies as El Nino.

This nascent El Nino may have some legs to it. Subsurface waters (down to about 1000 feet deep) across the equatorial Pacific are also warmer than normal, which tells us that there actually is a mass of anomalously warm water extending to some depth, and not just a thin skin of unusually warm waters at the surface. In addition, the forecast plumes from IRI also suggest this El Nino could be more than a passing fad. Most of the forecasts agree that the El Nino will continue into the fall season, and it could grow stronger. It isn't out of the question that the El Nino could grow to moderate strength.

The ramifications of El Nino on the weather in the U.S. are many. However, El Nino's impacts during summer are rather muted and it won't be until the fall and winter that the weather patterns over the U.S. begin to show a more definitive El Nino influence. What can be seen during the summer is the impact that El Nino has on the degree of tropical activity in the Atlantic basin. Typically, El Nino generates stronger winds than usual at high levels in the atmosphere. These stronger winds in turn lead to greater wind shear, and wind shear is the enemy of many a tropical cyclone. The development of El Nino was anticipated to a certain degree in many of the seasonal outlooks for hurricanes and tropical storms in the Atlantic, and is the primary reason why most of the outlooks call for near or below normal activity.

In our next entry, we'll take a look at some recently published research documenting a variant of El Nino, known colloquially as "modoki El Nino." In this form, the warm anomalies in the equatorial Pacific are displaced to the central Pacific as opposed to the eastern Pacific, which could affect the degree of suppression of tropical activity in the Atlantic.

Active Weather Pattern in Europe

A weather pattern more reminiscent of January than June is affecting our neighbors across the pond. Unseasonably cold air has spilled down into western Europe from Scandinavia and the Arctic, leading to below and much below normal temperatures and a plethora of unusual weather.

Snow was reported in northern England on Friday, June 5. A light dusting fell in the relatively higher elevations (approximately 1500-2000 feet) near the Scotland border, which was the first significant snow that had fallen in England during the month of June since 1975. The snow came on the heels of a very hot spell for the UK at the end of May.

Severe thunderstorms also broke out over the UK and other parts of western and central Europe during the past several days. Following the snow, a band of thunderstorms brought heavy rain to southern England on June 6 and 7, including funnel clouds and tornadoes in the southwestern parts of the country. Tornadoes were also reported in Italy and Russia, and severe thunderstorms also affected France, the Benelux, and Germany with severe wind and hail.

From the heat at the end of May to snow and severe thunderstorms at the start of June? What a radical pattern shift! However, the pattern shift over Europe is a somewhat common phenomenon, falling within the purview of the North Atlantic Oscillation, or NAO. The NAO is an approximate measure of the pressure differences between Iceland and the Azores. Typically the semipermanent Icelandic Low occupies the North Atlantic, with the semipermanent Bermuda-Azores High covering the Atlantic subtropics and mid-latitudes. In this situation, the NAO index has positive values. On occasion though, ridges of high pressure can build north into the higher latitudes, often associated with blocking patterns. Low pressure is shunted south, and the NAO index reverses and becomes negative.

Negative NAOs have been associated with cold outbreaks and stormy conditions during winters in Europe. With the blocking ridge over the North Atlantic, the flow of milder air off the ocean is displaced away. A further complication can ensue when the blocking ridge retrogrades, or moves west, contrary to the typical west-to-east flow. When the blocking ridge over the North Atlantic retrogrades away from the continent, cold upper lows are free to slide down from Scandinavia and northern Russia, bringing outbreaks of cold air and unsettled weather.

The change from unseasonable warmth at the end of May to the cold and unsettled pattern currently over Europe was precisely due to the retrogression of the blocking ridge over the North Atlantic. A cold upper level low dropped down the North Sea from Scandinavia. The cold air aloft with the upper low was enough to produce instability and snow showers in the north of England, and also contributed to the production of the severe thunderstorms.

The satellite image shown above shows the current weather situation over Europe on the morning of June 9. The cold upper low we mentioned is centered over the North Sea, marked by a well-formed spiral of clouds north of The Netherlands. The cyclonic (counterclockwise) flow around this low extends all the way into the south of France. Over the British Isles, the flow is generally out of the north, bringing relatively cold polar air over the UK.

Cold upper lows like this over the North Sea often exhibit some interesting "substructure." In this case, using the visible satellite image at left, we can pick out a few "swirls within the swirl". The most evident is the bright one over the North Sea that we mentioned earlier. A second counterclockwise swirl can be seen over the English Channel, with some of the cloud bands over northern France spiraling into it. Finally, a third swirl can be seen to the north of Scotland and west of Norway. The three smaller circulations are pinwheeling about a common center, the center of circulation of the larger parent upper low.

Incidentally, these small features can often be polar lows, small and often subtle circulations embedded within the cold polar air streams, not associated with any fronts. Considered by some meteorologists to be distant cousins of tropical storms and hurricanes, polar lows are difficult to anticipate, and their behavior is not easy to predict. They can produce squalls of rain or snow, sometimes accompanied by thunder with gale or storm force winds over a small area. Sometimes they can spin up rapidly, or decay unexpectedly. It is very unusual, though not unheard of, to see polar lows in June. They are much more frequent during the winter months when the overlying polar air masses are strongest.

Summary of Atlantic Tropical Seasonal Outlooks

The first week of June marks not only the start of the hurricane season in the Atlantic basin, but also the release of updated seasonal forecasts from NOAA, Colorado State, and other agencies. The pre-season outlooks issued in the previous December, and updated in April, often show little actual forecast skill. However, the outlooks released closer to the actual start of the season begin to do a little better, and are somewhat more useful and reliable.

One of the best known tropical seasonal outlooks comes from Colorado State University and the forecast team led by Dr. Phil Klotzbach and Dr. Bill Gray. This year's CSU forecast - updated June 2 - calls for 11 named storms, 5 hurricanes, and 2 major hurricanes (an average season has 10 named storms, 6 hurricanes, and 2 major hurricanes). In terms of Accumulated Cyclone Energy (ACE), which takes into account the strength and longevity of the tropical cyclones that do occur in addition to just the numbers of storms, the CSU forecast also predicts a near to below average season, calling for 2009's total ACE to be approximately 72% of average.

The seasonal outlook from NOAA is worded a bit more vaguely, but is fairly close to the CSU forecast. NOAA calls for a 70% chance of 9-14 named storms, with 4-7 hurricanes and 1-3 major hurricanes. NOAA also produced a range in their prediction for ACE, predicting the ACE for 2009 to be anywhere from 65-130% of normal.

Other forecasts from firms such as Tropical Storm Risk and AccuWeather - not to mention Frontier Weather's own outlook - are in line with the general expectation for a season that comes in near average for the numbers of named storms and hurricanes in the Atlantic basin.

Most of the current seasonal forecasts have trended downward with their predicted levels of activity. The earlier outlooks from CSU and Tropical Storm Risk forecast 2009 to be similar to other recent years with numbers similar to the 1995-present average of 15 named storms. Through the spring months, revisions to the outlooks have led the prognosticators to lower their expectations based on two primary factors: El Nino, and Atlantic sea surface temperatures.

The recent La Nina episode in the equatorial Pacific Ocean has recently come to an end, but a quick transition to El Nino is possible. Conditions are currently ENSO-neutral, lying in between the extremes of La Nina and El Nino. However, sea surface temperatures (SSTs) across the equatorial Pacific have been warming over the past few months. Although not warm enough to qualify as El Nino at this time, additional warming over the next couple of months would bring about the onset of an official El Nino. This scenario would be somewhat similar to the 2006 hurricane season in the Atlantic, which began with a bang under ENSO-neutral conditions, but finished with a whimper as a developing El Nino suppressed activity and effectively ended the season in October. The Climate Prediction Center, which monitors ENSO, issued the final La Nina advisory on May 24 and then has recently (June 4) initiated an El Nino watch.

The possibility of an El Nino developing has ramifications for later in the season, however the current state of Atlantic SSTs has implications beginning right now. We've already seen how cooler than normal SSTs have affected possible developments with Tropical Depression One failing to strengthen into a named storm owing to unsuitably cool waters. A large part of the tropical Atlantic has cooler than normal temperatures. This includes the areas immediately to the north of the Caribbean islands, as well as a large part of the Main Development Region, or MDR, which occupies the area between 10 and 20 degrees north latitude, and between the African coast and the Lesser Antilles. Cooler SSTs means less "fuel" is available, and the shortage of fuel translates into fewer storms.

Even though most of the outlooks for the 2009 Atlantic hurricane season have been revised downward, it still only takes one major storm hitting the U.S. to make a season. The classic example, which has been beaten to death but still maintains its relevance, is the 1992 season. That year only had 7 named storms with 4 hurricanes. Only one hurricane made landfall in the U.S. - but that one was Category 5 Hurricane Andrew.

Storm Names for the 2009 Atlantic Hurricane Season

Hurricane season in the Atlantic officially begins on June 1, so it's a good time to become familiar with the names that will be used for this year's tropical systems. As we like to do, we'll also take a look back at the history of this year's names.

Without further ado, here's the list:
---
Ana
Bill
Claudette
Danny
Erika
Fred
Grace
Henri
Ida
Joaquin
Kate
Larry
Mindy
Nicholas
Odette
Peter
Rose
Sam
Teresa
Victor
Wanda
---

And now, some noteworthy items:

• In 2003, Tropical Storm Ana was an out-of-season storm. And by quite a lot. 2003's Ana had subtropical origins and formed in the central Atlantic in April. We would have been overjoyed had the recently-demised Tropical Depression One been able to earn a name, as it would have been the second time in a row that the name Ana was assigned to an out-of-season storm.


• The name "Bill" has been used twice previously, joining the list after the retirement of Hurricane Bob from 1991. Interestingly, all the Bob's were hurricanes (1979, 1985, 1991), and all made landfall in the U.S. (1979 in Louisiana, 1985 in the Carolinas, and 1991 in New England).


• Claudette, like Ana, is an original member of this list of names, having been used five times previously.


• Danny was the first replacement name to be added to this list, since Hurricane David was retired after 1979. "David" was also the first male name to be retired, since men's names were not used for tropical storms and hurricanes in the Atlantic prior to 1979.


• In the El Nino-plagued year of 1997, Hurricane Erika was the only major hurricane of the season. In that year, there were only 8 named storms and 3 hurricanes. Erika joined the list following the retirement of Hurricane Elena after 1985.


• Fred is a new name to this list, since Fabian was retired after 2003. In 2003, Hurricane Fabian made a devastating direct hit on Bermuda as a Category 3 storm.


• The name "Grace" was first used in 1991 following the retirement of Hurricane Gloria in 1985. That year, Hurricane Grace played a seminal role in the development of the fabled "Perfect Storm".


• In 1979, Hurricane Henri formed in the Gulf of Mexico but did not make landfall as a tropical storm or hurricane. Henri weakened in the Bay of Campeche and arrived at the coast as a mere tropical depression.


• Ida is another new name this year, taking the place of Isabel. 2003's Hurricane Isabel was a picturesque Category 5 hurricane for a time in the Atlantic which eventually struck North Carolina at just under Category 3 intensity. Hurricane Isabel of 2003 was also the one and only time that the name Isabel had been used on a hurricane (the only other time the name had even been used at all was in 1985 for a tropical storm).


• The final new addition to this year's list of names is Joaquin, which replaces Juan. Hurricane Juan struck Nova Scotia as a Category 2 hurricane in 2003, one of the strongest on record for Atlantic Canada. The only other occasion where the name Juan was used was in the busy 1985 season, when Hurricane Juan affected the Louisiana and Mississippi coasts.


• Kate is an original member of this list but has only been used twice before: in both cases, 1985 and 2003, as Hurricane Kate. Hurricane Kate in 1985 became one of the strongest November hurricanes on record and went on to strike the Florida Panhandle as a Category 2 storm.


• Larry, Mindy, Nicholas, Odette, and Peter were all used for the first time in 2003. That year, all were tropical storms. Odette and Peter formed in December, after the official end of the season on November 30.


• The remaining names on the list (Rose, Sam, Teresa, Victor, and Wanda) have not been used at all since this list first went into service in 1979.

TD 1 Exits Quietly

Tropical Depression One formed this week in the Atlantic just off the Mid-Atlantic coast. It's becoming almost commonplace for out-of-season systems to form in the Atlantic these days, as this is the third consecutive year with a May storm. However, TD 1, though it had potential, was not able to earn a name.

An area of low pressure was noted on Wednesday off the Southeast coast between Cape Hatteras and the Bahamas. There was only sporadic convection with the system, which is understandable considering water temperatures were only in the 70-75 F (25 C) range. Still, the circulation of the disturbance was obvious from the start. During the overnight hours on the morning of May 28, a burst of thunderstorms erupted near the center and by morning it had become the season's first tropical cyclone as Tropical Depression One, with the possibility to become yet another out-of-season named storm.

TD 1 was not able to maintain its convection for long. Despite the well-organized circulation, the water temperatures were just not warm enough to create enough atmospheric moisture and instability to sustain a tropical cyclone. As the depression moved northeast away from the coast, it began to encounter cooler waters - although it appeared to hug the north wall of the Gulf Stream for a time. In addition, wind shear increased as the system advanced further north, which stripped away the meager convection on the afternoon of May 29. The depression quickly degenerated to a naked, skeletal swirl of low clouds which was declared extratropical by the National Hurricane Center. The satellite image above left shows the remnants of TD 1 late in the afternoon on May 29. Although it is tempting to see a large region of clouds spiraling into the center of the remnant circulation at the center of the image, the depression was actually never that big and consists only of the small swirl of low clouds. Note also the region of clouds and showers to the southeast of the center - this is the remnant thunderstorm activity that was stripped from the depression by strong wind shear. The region of cloudiness to the north of the former depression is actually a bank of low clouds and fog which has formed over the cool Atlantic waters to the north of the Gulf Stream.

With TD 1 having lost its tropical characteristics, the remnant circulation will spin down this weekend before getting absorbed into a frontal zone, and activity in the tropical Atlantic will return to the usual quiet conditions. Hurricane season in the Atlantic officially begins on Monday, June 1.

Cyclone Aila in the Bay of Bengal

Over the weekend, the a tropical cyclone developed in the Bay of Bengal - the second of the season so far. Cyclone Aila made landfall near the the border of India and Bangladesh, bringing torrential rains, hurricane force winds, and storm surge. Presently, there are over 200 fatalities from Aila, with over 1000 additional people unaccounted for.

Cyclone Aila formed in a manner not often seen in the Atlantic or the eastern Pacific. Most tropical storms and hurricanes in these regions form from tropical waves, subtle disturbances that occasionally spark convection and then grow and organize. In the northern Indian Ocean, often the progenitor of tropical cyclones is the monsoon trough, a large zone of converging air that migrates from the equator to the Indian subcontinent as part of the seasonal monsoons. Perturbations along the monsoon trough can acquire rotation, becoming large monsoon depressions, and occasionally these monsoon depressions can organize into tropical cyclones. It is during the months when the monsoon trough is migrating (May and November, give or take) that tropical cyclones are most often seen in the northern Indian Ocean.

The image at the start of this entry shows Cyclone Aila at its peak strength, with winds estimated at 75 mph. Note Also that Aila is a large system - not as big as last year's Hurricane Ike, but still expansive nonetheless. Compare that to the image at left, which is the monsoon depression that evolved into Aila, but two days earlier (click to open a full-size version in a new window, which will be helpful to get an idea of the scale of the system). The monsoon depression covers almost the entire Bay of Bengal (a body of water comparable in size to the Gulf of Mexico) with bands and clusters of strong but disorganized convection. The overall system has a hint of a broad cyclonic turning, but not much in the way of a clearly defined circulation.

There is some speculation that tropical cyclones in the northern Indian Ocean, which are closely tied to the monsoon trough and the formation of monsoon depressions, are somewhat more predictable than their cousins in other basins. This would make sense in that the monsoon depression, being a large feature, could be better resolved by most operational computer forecast models. Still, the transition from the monsoon depression to the tropical cyclone would be difficult to anticipate. A similar situation has been seen anecdotally in the Atlantic for systems that have a subtropical origin. Often times the forecast model can pick up on the larger scale upper low or occluded synoptic scale low days in advance, although it is difficult to pin down the subtropical and tropical transitions with significant lead time. Such was the case with 2007's Tropical Storm Barry. The GFS forecast model suggested the possibility of subtropical development 6-8 days in advance, but the model was not much help in determining when the transition to a tropical system would occur. Unfortunately the situation is not as cut-and-dried as that single example would indicate, and although forecast skill in anticipating the formation of tropical cyclones has improved, there is still much more work to be done.

La Nina Update

La Nina conditions have officially ended, as the Climate Prediction Center recently released the final advisory on our now-departed La Nina. Still, the atmosphere - especially over North America - remains in a very lethargic, La Nina-like configuration.

The image above left is a global plot of sea surface temperature (SST) anomalies. The area of interest is the equatorial Pacific, where there is now virtually no blue painted on our map from Indonesia to the South American coast. SSTs in the tropical Pacific are no longer cooler than normal, and are actually a little warmer than normal. Most of the equatorial Pacific surface waters are averaging about 0.3 degrees Celsius warmer than normal of late.

The warm anomalies also extend to some depth below the surface. The plot at left shows the depth of the temperature anomalies in the equatorial Pacific as of May 13. Longitude is on the horizontal axis, with the left edge of the plot representing Indonesia, and the right edge South America. Depth is on the vertical axis, with the surface at the top and greater depths toward the bottom. While the anomalies at the surface of the Pacific may not be terribly impressive, there is a substantial reservoir of warmer than normal water just below the surface, which extends across almost the entire basin. Much of this water is a solid 2 degrees Celsius warmer than normal. Some of this warm subsurface water is making its way to the surface in the eastern Pacific.

Despite the presence of weak warm anomalies at the surface, and the large pool of warmer water lurking below the surface, conditions are currently ENSO-neutral. However, given the warm subsurface water (which can often be a predictor of surface temperatures in a few months), it is likely that additional warming will continue heading into the summer months. At left is the latest ENSO forecast plume from the International Research Institute for Climate and Society, which displays a collection of ocean temperature forecasts from various dynamical and statistical models. Most of the forecasts show increasing SST anomalies, though most keep the values of the anomalies within the ENSO-neutral category (from -0.5 Celsius to +0.5 Celsius). Some of the models do predict the development of El Nino conditions later this year.

In the near term, the consensus is for the ENSO-neutral conditions to continue into the summer months. By the Fall, there is a chance that continued warming could bring about El Nino conditions. IRI pegs this chance at a little less than 50/50. This could have implications for the upcoming Atlantic hurricane season, which begins officially on June 1. Should El Nino conditions manifest earlier, or be stronger than the current thinking suggests, it could quash tropical activity in the Atlantic in the latter part of the season, similar to what was seen in 2006 when no storms were observed after October 2 due to the onset of El Nino. With this in mind, many of the recent updated tropical season outlooks have been revising their forecast numbers downward, to account for the possible appearance of El Nino later in the season. The consensus is still for a season with activity levels above the long-term average, but not as great as the recent (1995-present) average.

The May 8, 2009 Derecho - Part Three

After blasting across southeastern Kansas with widespread damaging straight line winds of 80 to 100 mph, the derecho roared into southwestern Missouri early in the morning of May 8, 2009. The dangerous system continued to produce severe wind gusts, but as it pressed east towards Springfield, the character of the system began to change. As the broad scale rotation in its comma head strengthened, low level wind shear was enhanced, making all modes of severe weather - hail as well as tornadoes to go along with the ongoing straight line wind - a very serious threat.

---

A large scale radar look at the derecho as it approached Springfield at approximately 7:30 AM local time is shown at left (click to open a full-size version in a new window). The bowing line of thunderstorms producing the widespread wind damage takes up just a small part of the total area encompassed by the system. The shield of precipitation extended north all the way to Kansas City, and then as far northeast as St. Louis. The line of strong to severe thunderstorms trailed back into northeastern Oklahoma and northern Arkansas.

From a quick glance at the radar reflectivity image, the overall appearance of the system was becoming more reminiscent of a tropical cyclone. A number of spiral banding features can be seen wrapping into and around the center of circulation, which was located northwest of Springfield. The most obvious of the bands was the bowing line of intense thunderstorms producing the wind damage of the derecho, while another secondary band (highlighted by yellows) can be seen to the north of the center, wrapping into it from the west. Radar animations clearly showed the rotation about the comma head, unfortunately we cannot host animations in our blog.

With the rotation in the system strengthening, the derecho became more than just a straight line wind threat. At about this time, reports of large hail began to become interspersed with the wind damage reports. Joplin and Carthage, which had received winds of up to 93 mph, also reported golfball sized hail. Ahead of the main line and circulation center, thunderstorms developed across central and eastern Missouri that exhibited supercell characteristics. Baseball sized hail fell from one such storm in Washington County, situated to the southwest of St. Louis and well northeast of the center. It was also at about this time, 8:00 AM, that the first reports of tornadoes began to be received.

The radar image at left is a high-resolution plot of Doppler velocities close in to the Springfield site. The city of Springfield is located near the far left edge of the image. Reds indicate outbounds, and greens are inbound winds. Note the two areas, at center and bottom center, where the greens penetrate slightly into the solid wall of red returns. In both cases, enhanced outbound velocities (pinks and oranges) were also measured nearby. These signatures of small scale rotations were tornadic circulations forming within the line of thunderstorms. Numerous such features dotted the main line as it passed through and east of Springfield. Tornadoes from a line (or quasi-line) of thunderstorms are not unheard of, but it is not at all a textbook scenario. Tornadoes spawned by lines of thunderstorms are very difficult to anticipate with very much lead time - they often spin up quickly, have short paths, and then dissipate not long after.

In all, the derecho produced 27 tornadoes in southwestern and south-central Missouri. The strongest was rated an EF3 with winds estimated at 165 mph. A long-tracked EF2 tornado also occurred, its 21-mile long path very unusual for a tornado spawned by a line. The National Weather Service office in Springfield performed admirably during the event. In a situation that was not a classic tornado environment (a line of storms, in the early morning hours no less), adequate warnings were provided for 18 of the tornadoes, giving residents an average of 20 minutes of lead time. Nevertheless, 1 death and 3 injuries were attributed to the tornadoes in the Springfield area.

The radar grab at above left shows a wider look at the system as it pushed east of Springfield at about 8:45 AM. North of Springfield (which is marked by the dot labeled "KSGF"), there is a large area of blues and greens in a roughly spiral shape; this is a signature of the strengthening and enlarging circulation of the complex. The increased size of the comma head circulation can at this point be more accurately called a mesoscale convective vortex, or MCV. The active line of thunderstorms is marked by the boundary between reds and greens that is situated east of Springfield. Note the little kinks and indentations within the line indicative of potential tornadic circulations. It's also interesting to point out the existence of a few velocity couplets (bright reds or pinks immediately next to greens) near the eastern edge of the radar's range. These correspond to supercell thunderstorms to the east of the main line of the derecho which were producing severe hail. By this point, it's probably apt to drop the term "derecho" for the complex, as it had added the impressive MCV as well as the threats for tornadoes and large hail to its arsenal.

Leaving the Springfield area behind, the complex continued eastward across southern Missouri as the morning went on. By 10:45 AM, the derecho appeared as shown in the image at left, which is a radar reflectivity image from St. Louis. St. Louis is located near and just above the center of the image, marked by "KLSX" and "TSTL". At this point, the spiral character of the convection was readily apparent. Two or perhaps three bands of strong thunderstorms were located in the complex's eastern quadrant, spiraling in toward the center. The region of steady, stratiform rain remained very large in areal coverage to the north and northeast of the center. Note also that an area of weak reflectivity, almost like the eye of a tropical cyclone, had formed in the vicinity of the center of circulation itself.

The remarkable part of the evolution of the MCV is shown at left. In the first image, taken from St. Louis at 12:08 PM local time, the system had taken on the structure of a comma head, but at a much larger scale. A large region devoid of echoes had punched in behind the leading edge of thunderstorms and partially wrapped around the MCV. In the second image, taken from St. Louis at 12:47 PM, note the MCV itself. An eye-like feature had developed in the middle of the ball of convection, giving the system a stunning appearance overall and an uncanny resemblance to a tropical cyclone.

Another interesting evolution within the MCV took place a short while later, and is captured in the two images at left taken by the Doppler radar in Paducah, Kentucky. The image at left shows the MCV with its eye-like feature at 1:05 PM. Note also the area of weak reflectivity (depicted by shades of blue) trailing to the west of the eye-like feature. In the next image, taken at 1:15 PM, the trailing area of weak reflectivity had become the new eye-like feature, with the old "eye" moving eastward and filling in with light echoes (greens). It is difficult to say what is going on during the two images as a new eye-like feature forms and becomes dominant, or what the physical processes are that drive such changes in structure, but it is remarkable nonetheless. There is some temptation to draw parallels to Tropical Storm Erin of 2007, which re-intensified over Oklahoma well after landfall in Texas a few days earlier, and exhibited a similar structure on radar complete with an eye-like feature.

The complex with its compelling MCV continued east, working gradually up the Ohio Valley. The radar grabs at left, taken from the Doppler radar in Evansville, Indiana at 1:59 PM reveal some further structural changes in the complex. First, the main line of thunderstorms, pushing east across Kentucky, had become well displaced away from the MCV, trailing behind over southern Illinois. Within the MCV itself, the eye-like feature had begun to expand, and was quite a bit larger than about an hour earlier when it was nearly a pinpoint. The severe threat associated with the complex was decreasing at this time. In the overnight hours, a separate complex of thunderstorms had affected Kentucky and Tennessee, and had since advanced east into the Carolinas. Although the storms themselves were gone, they had used up most of the atmospheric instability in the area, leaving behind a region of rain-cooled and relatively stable air.

Robbed of its fuel source, the complex weakened as it continued east across Kentucky. At left is a radar image from Evansville at just before 3:00 PM. Note how the convection associated with the MCV had shrunk dramatically. The circulation of the MCV was still apparent, but the structure had decayed, looking much more ragged and far less organized - almost skeletal. Meanwhile, the active thunderstorm line was also weakening as it encountered the more stable air. The line had become more broken, more ragged, and less intense. Although some reports of severe weather were received across western and central Kentucky and Tennessee, they were neither as significant nor as widespread as what occurred hours earlier over Kansas and Missouri.

The decreasing trend over western Kentucky brought on by the stable air left behind by previous convection was only temporary. The air mass over eastern Kentucky was less "worked-over" by earlier thunderstorms, and provided a modest amount of instability to help the thunderstorms in the complex reintensify in the afternoon hours. This was to be the "last gasp" of the long-lived thunderstorm complex, and it is characterized well by the radar images above from Louisville. In the reflectivity image at above left, we see the decaying MCV to the west of Louisville (which is marked by the "KLVX" and "TSDF" identifiers). The line of thunderstorms along the complex's leading edge remained diffuse and more broken in character. As in earlier phases of the complex's lifetime, the storms out ahead of the line were taking on supercell characteristics. This can be seen in the velocity image (the right image from the two above; click to open a full-size version in a new window). In the velocity image, the MCV circulation can still be readily seen to the west of the radar site, but notice to the south and southeast of the radar, near the Kentucky-Tennessee border, the velocity couplets typical of rotating supercell thunderstorms. There are five (perhaps six?) such couplets where reds and greens are in close proximity. These supercells produced 8 tornadoes in eastern Kentucky and Tennessee, along with a swath of both wind and hail damage reports.

---

All told, the derecho of May 8, 2009 was a remarkable weather event. It produced a swath of damage from Kansas to Kentucky - in the form of widespread damaging winds primarily, but also in tornadoes and large hail as well. The complex spawned 44 tornadoes, a large number for a quasi-linear system, including an EF3 and several with long tracks. In addition, there were over 200 reports of damaging winds (including 19 instances of winds of hurricane force), and approximately 50 reports of severe hail. As of this writing, there are 6 fatalities and 13 injuries from the storm. The large amount of real estate affected by the storm, as well as its rapid progression (it went from genesis in northwestern Kansas to its eventual decline in eastern Kentucky in approximately 18 hours) and the numerous unusual and exemplary phenomena observed in association with it, will make this derecho one that both the residents it affected and the meteorologists who observed it will remember for a long time to come.

The May 8, 2009 Derecho - Part Two

This is the second in a series of posts chronicling the derecho that affected a swath of the U.S. from Kansas to Kentucky on May 8, 2009. In the first series, we followed the derecho from its genesis in northwestern Kansas through its "adolescence" across the state as it produced severe hail and damaging straight line winds of up to 80 mph across a large part of western and central Kansas, including Dodge City and Great Bend, and was about to do the same in Wichita. In this post, we follow the mature derecho in its most destructive phase across southeastern Kansas and into southern Missouri.

---

At about 4:30 AM local time, the mature bow echo, having left a swath of hail and wind damage across Kansas, approached the Wichita metro area. The images at left are radar captures from the Doppler radar at Wichita (marked by the blue dot labeled "KICT" on the map). The left is reflectivity, and the right panel is velocity (click on either image to open a full-size version in a new window). Note the bright greens and blues to the north of the radar site in the velocity image. These denote strong inbound velocities and are the radar signature of the strong straight line winds. Also worth noting are two subtle features to the west of the radar site, located along the leading edge of the high reflectivity associated with the thunderstorms. On the reflectivity map, these features appear to look like little "stair-steps," where the yellows and reds jut southward from the main line. On velocity, the features appear as small regions of gray and red entangled with the predominant greens of the line. These features are small-scale circulations, possibly indicative of supercell-type thunderstorms, embedded within the larger line. Radar signatures such as this can be associated with small or brief tornadoes occurring at the leading edge of the line, and they only add to the concern operational meteorologists experienced as this event unfolded. No tornadoes were reported in the area around Wichita, however.

On the northeast side of Wichita, the derecho took the next step in its development, with a new downburst forming and pushing out to the east at just before 5:00 AM. The radar grabs at left (reflectivity, left; higher resolution velocity (thanks to the storm's close proximity to the radar site) at right) show this process and its immediate results. On reflectivity, the feature of interest is the crescent-shaped arc of reds extending approximately southeast of Newton, then curving back southwest toward Wichita. The apex of this short bowing structure, where straight line winds are often at their most fierce, is pointed directly at the town of El Dorado, located to the northeast of Wichita along the Kansas Turnpike. On the velocity image, it's easy to see that El Dorado was in trouble. The intense reds, pinks, and oranges signify outbound velocities of at least 75 mph, and not long after this image was taken, wind gusts of up to 81 mph were measured in the El Dorado area. One motorist was injured when the strong winds flipped his vehicle on the Kansas Turnpike.

The bow continued toward the east, with winds gusting to at least 80 mph, picking up speed along the way. One fatality occurred in Wilson County when a woman was killed in her mobile home, which was no match for the powerful straight line winds. The radar grabs at left (reflectivity on the left, velocity on the right) show the progress the storm had made by approximately 5:45 AM. Reaching a forward speed of almost 70 mph, the bow was racing east across southeastern Kansas. Note also the storm cluster that had formed out ahead of the main bow echo. If you look closely at the velocity image, you'll notice within this cluster a couplet where inbounds (green) are located very close to outbounds (bright reds). This thunderstorm out ahead of the main line was exhibiting some of the characteristics of a supercell thunderstorm, and was a threat for large hail and possibly a tornado for a brief time before it was caught and absorbed by the main line.

Note also in the reflectivity image how the portion of the storm complex to the south of Wichita was not faring well as it pressed further south toward Oklahoma. Compare with the first image of reflectivity shown above (or one of the reflectivity images from Part 1), and it's apparent that the weak yellows and oranges of the storms as they moved toward the border are not a match for the strong reds seen earlier in their lifetime. As the storms attempted to press farther to the south, they were fighting an increasingly uphill battle. The most conducive thermodynamic conditions for the derecho existed over Kansas. Over Oklahoma, the atmosphere was much less favorable. In fact, a strong capping inversion present over Oklahoma had prevented thunderstorms from developing there during the previous afternoon despite the presence of strong instability. As the southern extent of the line of storms neared Oklahoma, it began moving into this less favorable thermodynamic environment, and it had to "work uphill" against the inhibition presented by the cap. Eventually the cap would win out, and the storms were never able to make much progress to the south. This would become a theme to the mature stage of the derecho. Where earlier it surged south from near Goodland to Dodge City and Wichita in its adolescent phase, the increasing inhibition to the south over Oklahoma blocked its continued southward progress, and instead it propagated quickly eastward along the periphery of the stronger inhibition.

Further evidence of the derecho's struggles in southward propagation comes from the reflectivity image at left. Taken from the Doppler radar in Tulsa, Oklahoma, at approximately 6:50 AM, the image shows that the derecho has raced east; the apex of the bow echo was approaching the Kansas-Missouri border. The thunderstorms had been able to make some progress into northeastern Oklahoma, but notice the thin blue line running parallel to but out ahead of the storms near the center of the image. This line marks the gust front, the leading edge of the strong outflow winds. In mature derechos, the inflow and the outflow are balanced, which means that the leading edge of the outflow winds is collocated with the leading edge of the strongest reflectivity. When the outflow becomes dominant, the gust front will push out ahead of the main line, which is usually a sign that the storms are weakening slightly. The greater convective inhibition over Oklahoma continued to thwart the derecho's attempts to propagate south.

However, in Kansas the derecho raged eastward unhindered. Widespread wind gusts of 80 mph were reported all over southeastern Kansas, including near Buxton in Elk County, around the town of Altoona, and also the area around Parsons. The most severe winds in the damage swath were reported near the town of Cherryvale where considerable structural damage took place under winds gusting as high as 100 mph.

After Kansas, the next target of the derecho was southwestern Missouri. The radar grabs at left show the derecho as it approached the Kansas-Missouri border, as seen from Springfield at approximately 7:00 AM. The familiar bowed-out appearance of the line remained evident on the reflectivity image, and on velocity the rear inflow jet (discussed in Part 1) was also visible as the region of strong inbound velocities (bright greens and blues) appearing over eastern Kansas. Notice the couplet of inbounds (greens) next to outbounds (reds) approaching the Kansas-Missouri border. This circulation is the signature of what is called a bookend vortex. The bookend vortices form at either end of a bow echo as the line bulges out to form the bow. On the northern flank of the bow, a counterclockwise circulation forms, and a clockwise circulation forms at the southern side of the bow. The Springfield radar was able to see the northern bookend vortex with its cyclonic circulation (colloquially known as the "comma head") as the system approached. This feature would take on an increasing importance later on in the derecho's life as it moved across southern Missouri.

Despite its evolving character with the development of the pronounced cyclonic bookend vortex, the main impact from the derecho on southwestern Missouri was the same as in southeastern Kansas: destructive straight line winds. Wind gusts of 60-70 mph were common across the Joplin area (and also in the far northeastern corner of Oklahoma where some wind damage occurred in Ottawa County). In addition, hail up to the size of golfballs pelted the area, including the city of Joplin and neighboring Carthage. Carthage also reported a 93 mph wind gust as the bow arrived. Farther to the east, an additional swath of 60-80 mph wind gusts occurred, including Springfield with 70 mph wind gusts, and Branson with 60 mph gusts.

About the time the line approached Springfield, an additional region of strong winds began to affect southwestern Missouri near Joplin, well behind the active thunderstorms. These winds, which have been seen often on the back edge of thunderstorm complexes in the Plains this year, are caused in part by a phenomenon called the wake low. Often times a small area of low pressure will form quickly behind a thunderstorm complex - which may be influenced in part by the rear inflow jet above. Winds associated with the wake low can sometimes reach severe criteria (58 mph) on their own. The image at above left is an analysis of the surface pressure tendency in the area around southwestern Missouri. Oranges and reds indicate areas where the surface pressure was rising, and the blues and purples show falling pressures. Joplin is marked by the black dot labeled "JLN", and Springfield by the black dot labeled "SGF". This analysis comes from a large-scale forecast model, so it does not completely capture what was going on in the vicinity of the derecho (a smaller scale phenomenon by comparison), but it can illustrate some salient details. First off, at the time of this analysis (approximately 8:00 AM), the leading edge of the bow echo was pushing through Springfield. Joplin, well behind the active thunderstorms, was clearing out but also experiencing an area of strong pressure rises. These rises, in association with the developing wake low (a feature too small to be picked up by the model), created a strong pressure gradient over a small region, and the resulting accelerations led to the severe - but non-thunderstorm - winds behind the derecho. Winds gusted as high as 85 mph in Joplin in association with the wake low. In the surrounding areas, the winds may have peaked at nearly 90 mph. More remarkably, the severe winds in excess of 60 mph persisted for as much as 20-30 minutes in any given location.

---

The derecho of May 8, 2009 was mature by the time it blasted through southeastern Kansas and southwestern Missouri. However, it was not done yet. As its structure began to change and the "comma head" bookend vortex strengthened, the character of the damage changed from primarily a straight line wind threat to a combined threat of severe hail, straight line winds, and tornadoes. In our next post, we'll examine the tornadoes spawned by this system as it continued east across southern Missouri and take a close look at the phenomenal structural changes of the derecho and its comma head as it continued east across the Mississippi River.