Thunderstorms Reach Lake Michigan

Introduction
By 11:00 pm (0300Z) Sunday night, the system was poised to race across northern Lake Michigan. The overall system was characterized by several multi-cellular convective clusters at the northeastern end and a Mesoscale Convective System with a squall line along its leading edge stretching to the southwest.

Although there were slight differences, the thunderstorms comprising the system (figure 1) were generally moving southeast at 25 to 30 knots. Figure 2, the echo top imagery valid at 11:01 pm (0301Z), indicates that the strongest storms, indicated by dark red, were approximately 60,000 feet tall.

Figure 1: Base reflectivity from KAMX at 11:01 pm (0301Z). (larger image)
Figure 2: Echo tops from KAMX at 11:01 pm (0301Z). (larger image)

At 11:09 pm (0309Z), the NWS office at Gaylord, MI (KAPX) issued an additional Marine Weather Statement (link) that continued to warn of winds to 33 knots. The Weather Statement also expanded the affected areas to include Grand Traverse Bay and specifically mentioned South Fox Island and Point Betsie. At this point in the Chicago-Mackinac Race, most participants were located in the northern portion of Lake Michigan and therefore in the path of the approaching storms. Unfortunately, a technical problem with the yacht tracking system left race organizers uncertain of the exact location of the competitors.

A Reputation For Damaging Winds
By the time the system reached Lake Michigan, it had a track record of producing damaging wind gusts across northern Wisconsin and Michigan's upper peninsula. The map below (figure 3) shows the locations of significant wind reports received by the Storm Prediction Center.

Figure 3: SPC Storm Reports preceding 11:00 pm (0300Z).

Typically, these reports are submitted in response to structural damage, downed trees, and power lines and not on the basis of an observation by an automated weather station. As such, the speed of the wind prompting the report is usually listed as unknown. In some instances, winds from a severe thunderstorm are observed by a nearby weather station, documenting the conditions that produced the damage. Such was the case when the system roared through Rhinelander, WI and produced a gust of 56 knots. Elsewhere, automated observations from sparsely positioned weather stations were much lower -- in the range of 20 to 30 knots.

The System Begins Its Crossing
Below are the base reflectivity (figure 4) and base velocity (figure 5) radar images produced by the Gaylord, MI (KAPX) radar site at 11:01 pm (0301Z), just as the system was beginning its transit of the lake. The dark green area on the base velocity radar image just to the west of the Fox Islands indicates inbound winds of approximately 25 knots at a height of 7,000 feet, while the bright blue region near KMQT indicated nearly 55 knots at approximately 18,500 feet. Since brisk winds aloft may contribute to the strength of a storm's downdraft, an overall assessment of the damaging wind potential of a system can be derived by sampling the velocity data associated with a storm system. Velocity data can be very helpful in identifying specific features known to produce strong winds at the surface.

Figure 4: Base reflectivity from KAMX at 11:01 pm (EST) (0301Z) (larger image)
Figure 5: Base velocity from KAMX at 11:01 pm (EST) (0301Z) (larger image)

A Possible Collapsing Thunderstorm
Beginning at approximately 11:01 pm (0301Z) and continuing for the next several minutes (figures 6a, 6b, 6c and 6d), a downdraft from a strong thunderstorm over the southeastern portion of Michigan's Upper Peninsula prompted the development of new storms downstream of the parent thunderstorm (annotated image). These new storms formed at the northeastern end of a convective line stretching southwest into central Wisconsin.

Figure 6a: KAMX base reflectivity 11:01 pm (0301Z) (larger image)
Figure 6b: KAMX base reflectivity 11:06 pm (0306Z) (larger image)
Figure 6c: KAMX base reflectivity 11:10 pm (0310Z) (larger image)
Figure 6d: KAMX base reflectivity 11:15 pm (0315Z) (larger image)

Although not apparent on the base velocity imagery, a review of the echo top imagery suggests that the downdraft may have formed during the collapse of the towering thunderstorm over the southeastern portion of Michigan's upper peninsula. At 10:52 pm (0252Z), the top of the storm reached approximately 60,000 feet. Over the next thirty minutes, the height of the storm dropped dramatically to nearly 30,000 feet (link to images ). The storms that formed from the downdraft of the collapsing storm raced ahead of the leading edge of the system.

A Bowing Segment
At 11:24 pm (0324Z) (link) and 11:29 pm (0329Z) (link), not long after the system moved over Lake Michigan, the base velocity imagery showed an area of 50 knot winds at a height of approximately 13,000 feet approaching Washington Island, WI. These strong winds were associated with the bowing segment ("bow echo") in the line (annotated image) west of the Fox Islands and just south of the area of new storm formation discussed above. A few minutes later, the base reflectivity image at 11:38 pm (0338Z) (figure 7) shows that the bowing segment contained thunderstorms characterized by high reflectivity values (in orange and dark red) along the leading edge, with an area of weaker returns (yellows) immediately behind and upwind. The base velocity image at this same time (figure 8) shows the pocket of strong winds aloft approaching from the west, behind the leading edge of the bowing segment.

Figure 7: Base reflectivity from KAMX at 11:38 pm (0338Z) (larger image)
Figure 8: Base velocity from KAMX at 11:38 pm (0338Z) (larger image)

The combination of strong winds aloft, the bowing segment and the variations in reflectivity values suggests the existence of a rear inflow jet. As the name suggests, a rear inflow jet is a fast-moving elevated stream of relatively cold and dry air entering the rear of the storm. Rear inflow jets form in response to a mid-level area of low pressure located behind the leading edge of the system and promote the formation of a bowing segment in the leading edge of the thunderstorms. In a process that is not completely understood, a rear inflow jet may descend to the surface and produce the violent winds associated with a downburst. The appearance of a bowing segment on reflectivity imagery signals the potential for damaging straight-line winds at the surface.

Supercell Downdraft
At approximately the same time (11:33 pm (0333Z)), the base velocity image showed a small area of inbound 50 knot winds (at a height of 7,500 feet) northwest of the Fox Islands and west of Beaver Island. Five minutes later, (11:38 pm (0338Z)), the base velocity data (figure 10) indicated that this pocket of strong inbound winds (50 knots) associated with a robust thunderstorm (figure 9) had descended to approximately 7,000 feet above the surface. Storm-relative velocity imagery subtracts overall storm motion from the velocity data and provides for an analysis of small-scale circulations, such as a rotating updraft within a storm. The storm-relative velocity imagery at 11:38 pm (0338Z) (link) shows the close proximity of inbound and outbound winds within the storm. This rotation was weak, but it suggests that the thunderstorm was probably a supercell. This weak circulation persisted and was more evident on the storm-relative imagery at 12:01 am (0401Z) (link).

Figure 9: Base reflectivity at 11:38 pm (0338Z). (larger image)
Figure 10: Base velocity at 11:38 pm (0338Z). (larger image)

The appendage extending from the southern edge of the large storm northwest of Beaver Island (annotated image), is the result of several storms merging together into a single complex. This group of storms are those that initially formed along the gust front of the collapsing thunderstorm in the southeastern portion of Michigan's upper peninsula and then raced of the system (radar loop). A region of strong, rapidly descending air on the leading edge of a supercell thunderstorm indicates the presence of a downdraft. Downdrafts associated with supercells are easily capable of producing surface winds of 50 to 70 knots or more.

The Storms Reach The Fleet
By 11:47 pm (0347Z), the competitors near the Fox Islands were already coping with the arrival of a large supercell thunderstorm from the northwest and a bowing segment approaching from the west. Both of these features have long been associated with damaging winds at the surface and were poised to deliver a possible one-two punch to the fleet. The base velocity imagery at 11:47 pm (0347Z) (figure 12) indicated there were three areas of strong winds aloft (annotated image). West of the bowing segment was a large area of 35 to 40 knot winds at approximately 8,000 feet while northwest of the Fox Islands were two pockets of 50 knot winds at approximately 6,700 and 6,200 feet.

Figure 11: Base reflectivity at 11:47 pm (0347Z) (larger image)
Figure 12: Base velocity at 11:47 pm (0347Z) (larger image)

While radar is an excellent tool for assessing the overall structure of a system, the increasing height of the radar beam as distance from the site increases dimishes its ability to sample surface conditions just a few miles from the radar site. The Fox Islands are approximately 55 nautical miles from KAPX and at that distance, the base scan samples the atmosphere at a height of nearly 5,000 feet. Strong downdrafts also typically hit the surface and then race ahead of the precipitation accompanying the storm. These facts make it difficult to use radar to determine precisely when strong surface winds may arrive. Experience may allow a meteorologist to infer surface conditions based upon the radar data, however offering a definitive assessment regarding surface wind speeds based upon data several thousand feet above the surface would be inappropriate. In this case, information from Chicago-Mac participants along with observations from a NDBC buoy and a surface weather station help to complete the analysis.

Buoy Observations
Before reaching the Fox Islands, the bowing segment passed over NDBC Buoy 45002 (figure 13), the only moored weather platform in northern Lake Michigan. Buoy 45002 is located approximately 25 miles west-southwest of the Fox Islands. The buoy is outfitted with an array of instruments to take observations of air and water temperature, barometric pressure, wind speed, and wave heights. It was perfectly placed to intercept and record the passage of the bowing segment (annotated base reflectivity and base velocity radar imagery at 11:51 pm (0351Z)). Prior to the arrival of the bowing segment, the wind was generally from the southwest between 15 and 20 knots. The chart of observed wind speeds (figure 14) show there's no mistaking the arrival time of the leading edge of the storm, as the wind abruptly veered to the northwest and spiked to 45 knots at 11:50 pm (0350Z).


Figure 13: The location of NCDC Buoy 45002. Image from Google Earth.
Figure 14: Observed wind gusts (in knots) from NCDC Buoy 45002. (Subtract 4 hours from GMT time for local Eastern Standard Time)

Typically, a relatively rapid and short-lived increase in wave heights accompanies the passage of a strong thunderstorm. However, the observations from Buoy 45002 indicate that wave heights held steady at approximately 3 feet before, during, and after passage of the bowing segment.

As the bowing segment was passing over Buoy 45002 both KGRB and KAPX issued additional warnings for upper Lake Michigan. The Marine Weather Statement issued by KGRB at 11:48 pm (0348Z) (link) covered the southern portion of the system and warned of "dangerous winds to around 50 knots". Just a few minutes later (11:55 pm (0355Z), KAPX issued a Special Marine Warning (link) for a line of thunderstorms "capable of producing strong winds of 34 knots or greater" from Grand Traverse Bay south to Point Betsie.

The System Completes Its Crossing
At 12:06 pm (0406Z), the SPC issued Mesoscale Discussion #1647 (link) the final Discussion related to the storm system. The discussion made note of the "extensive corridor of strong thunderstorms propagating eastward at greater than 40 knots" and that "gusty winds appear to be the primary concern ... along the southern half of the MCS from Lake Michigan into east central Wisconsin". The leading edge of the system reached the Michigan shore at approximately 12:23 am (0423Z) (figure 15). The base velocity image (figure 16) at this time shows an area of 50 to 60 knots at a height of approximately 2,900 feet approaching the Charlevoix area (annotated radar). This pocket of brisk winds was the downdraft associated with the supercell thunderstorm that was previously northwest of the Fox Islands. This downdraft was readily apparent on the base velocity imagery as it travelled across the race course towards the Michigan shore from 11:47 pm (0347Z) to 12:23 am (0423Z) (radar loop).

Figure 15: Base reflectivity imagery from KAPX at 12:23 am (0423Z) (larger image)
Figure 16: Base velocity radar imagery from KAPX at 12:23 am (0423Z) (larger image)

The automated weather station at Charlevoix (KCVX) was perfectly situated to observe the surface wind associated with this area of brisk winds aloft. At 12:35 am (0435Z), the wind abruptly veered from southeast at 8 knots to northwest at 40 knots. Although not contained in the database of periodic automated weather observations, the NWS office at Gaylord, Michigan reported a peak gust of 64 knot at Charlevoix, MI. Twenty minutes later the wind had backed slightly to the west, but was still gusting to 24 knots.

Mackinac Island
By 11:38 pm (0338Z), several minutes before the bowing segment reached Buoy 45002, a strong thunderstorm had reached Mackinac Island. As the name of the event suggests, the finish line for the Chicago-Mackinac Race is near Mackinac Island, Michigan. The island's harbor and shore-side facilities provide opportunities for the competitors to recuperate from the rigors of racing a distance of 300 miles and to celebrate their achievement. The island wasn't spared by the storms. In response to this line of storms, KAPX issued a Special Marine Warning at 11:41 pm (0341Z) (link) for the area near Mackinac Island and the northern portion of Lake Huron.

Figure 17: Base reflectivity (left) and storm-relative velocity (right) radar imagery from KAPX at 11:51 pm (0351Z) (larger image)

An area of rotation associated with this storm was located west of the island and made its first appearance on storm-relative radar imagery (link). Mackinac Island is approximately 50 nautical miles from the radar site. At that distance, the base storm-relative scan was sampling at an altitude of approximately 5,000 feet. This area of rotation persisted and was stronger on the imagery produced at 11:51 pm (0351Z) (figure 17). The persistence of the storms, and their ability to produce strong gusts, prompted the NWS to reinforce the previous Warning with a Marine Weather Statement at 12:06 am (0406Z) (link). The circulation continued to appear on radar until approximately 12:19 am (0419Z) (link). The broad circulation at 5,000 feet did not produce a tornado, but the thunderstorm was responsible for the strongest gust recorded on the island -- a mere 21 knots.