Origins And Synoptic Setting For The 30- 31 May 1998 Minnesota To New York Derecho

 William E. Togstad
NOAA/ National Weather Service, Chanhassen, Minnesota


The 30-31 May 1998 Minnesota to New York derecho event is examined with special emphasis on the development of an elevated rear-inflow jet as viewed with 0.5 and 1.5 degree WSR-88D velocity data. Peak rear inflow jet velocities are used to estimate relative flow to both 2.5 km above ground level (AGL) ambient wind and to forward motion of the system gust front for the purpose of comparison with Weisman's numerical modeling work on long-lived convective systems with rear-inflow jets.  The larger scale environment of this system is examined to assess synoptic-scale forcing for this extreme weather event and to place it within the shear-buoyancy domain for long-lived convective systems. A coupled jet structure together with a unique CAPE field provide most of the insight.

1. Introduction

     Derechoes (or long-lived bow echoes, LBEs) account for many of the casualties and much of the damage from convectively induced nontornadic winds according to Johns and Hirt (1987), hereafter referred to as JH87. However, operational expertise remains rather limited in anticipating LBE development, or in forecasting duration, intensity or movement, once a Mesoscale Convective System (MCS) has formed. In fact, the JH87 check sheet, for forecasting derecho development, remains an important operational tool for anticipating these systems; numerical approaches enjoy only limited success.

     The difficulty in operationally dealing with derechoes becomes apparent upon reviewing recent literature concerning both the structure and motion of these systems. For example, Weisman (1992) emphasizes the importance of lower tropospheric shear in balancing the thermal circulation induced on the leading edge of the system's cold pool. Meanwhile, Evans (1998) provides substantial observational evidence that derechos can function over a wide variety of vertical wind shear, including environments with weak lower tropospheric shear. However, system-relative wind (SRW) does appear to be a key factor explaining how some derechos are able to maintain their strength and structure over many hours. This is particularly true when there is strong lower tropospheric inflow on the downshear side of an MCS.

     Forecasting derecho movement provides additional operational frustrations as the methods of Merritt and Fritsch (1984) and Corfidi (1996) break down for this special class of MCS. However, recent evidence provided by Corfidi (1998) casts a great deal of light on the role convective-scale downdrafts play in accelerating these systems downshear where system-relative convergence is maximized. Unsaturated air, either in the mid- levels or subcloud layer is also cited as a factor that enhances downdraft strength; the presence of unsaturated, environmental dry air around and downstream of derechos has also been noted by JH87.

     One of the most intriguing aspects of derecho motion is the ability of some systems to propagate at a rate considerably faster than the mean wind speed . This is due to the fact that the advective and propagation components of these systems are additive. However, this aspect of the derecho makes it particularly dangerous to the public, especially to those engaged in outdoor activities.

     The purpose of this paper is to place the extraordinary 30- 31 May 1998 derecho into the context of its dynamic and thermodynamic origins. Special emphasis is placed on comparing this system to similar MCSs from both numerical modeling and synoptic climatology methodologies. In order to accomplish this, two separate studies of the system were completed. Study (1) examines the first 6 hours of the storm with reflectivity and velocity data from the Chanhassen, MN WSR-88D (KMPX) for the purpose of confirming both the development and intensification of an elevated rear inflow jet. See Weisman, (1992) for additional details.

     In study (1), radar velocity data were judiciously combined with 00 UTC 31 May 1998 ETA model forecast winds to develop vertical cross sections of the developing rear inflow jet for this MCS. Low elevation (0.5-degree) reflectivity data timed the eastward progress of the bow echo into the Twin Cities Metro. In addition, this LBE is compared to other numerically induced LBEs with a shear/buoyancy nomogram provided by Weisman (1992).

     Study (2) examines the synoptic environment in which this system formed and evolved over an 18-hour period. Special emphasis is placed on how well this storm conformed to the synoptic climatology of JH87. In addition, this particular storm is discussed as a possible hybrid derecho due to an unusual arrangement between its CAPE field and a coupled jet structure.

     Before examining the results of these two studies, it is first useful to review some of the severe weather details of this storm. It is felt that completion of a detailed storm review is needed to underscore how devastating this storm was for residents of the upper midwest and great lakes regions. It is hoped that this study will finally motivate an effort to develop new forecast methodologies for LBEs that yield an increase lead time of perhaps 6 to 12 hours for the public.

2. Storm Summary for the 30-31 May 1998 Minnesota to New York derecho

     A widespread, exceptionally destructive derecho originated around 2020 UTC on 30 May 1998 over Brown County in southwest Minnesota. MCS formation was preceded by an F4 tornado 125 nm to the southwest at 0142 UTC. The tornado destroyed most of Spencer, South Dakota killing 6 persons and injuring 150 more. It is not clear if the South Dakota storm was dynamically implicated in the MCS development farther east.  The storm track and some of the more unusual storm reports are mapped out in Figure 1 and Figure 2; a text summary for each state follows below. The information for these summaries was obtained in STORM DATA (NCDC, 1998).

MINNESOTA: Storm damage reports were concentrated in high population areas within and immediately south of the Twin Cities. Property damage (excluding agricultural figures) totaled $48.85 million in south central and east central part of the state. However, $35 million of these damages were reported in Dakota County alone. The Burnsville Indoor Mall reported $1 million in damages with a brick wall collapsing at the height of the storm.

Wind gusts first peaked above 80 mph at Green Isle in Sibley county, about 45 miles southwest of downtown Minneapolis at 0317 UTC. Wind gusts ranging between 80 and 100 mph were noted from Sibley county eastward through Washington county on the state line, between 0317 UTC and 0400 UTC. A total of 22 injuries were reported in Minnesota but there were no storm-related deaths. Derecho movement averaged between 50 and 60 mph between Brown County and Washington County with 500,000 homes losing power by the time the system entered Wisconsin. Some residents waited 5 to 6 days for power to be restored due to a shortage of replacement utility poles.

WISCONSIN: Utility companies and Emergency Managers stated that this storm was the most damaging, widespread, straight-line wind event to affect southern Wisconsin in 100 years. A 20 county area in south central and southeast Wisconsin sustained damages of $55.852 million; again these figures do not include agricultural damage totals. A woman was killed in the town of Erin in Washington County, when a tree fell through her home, crushing her as she slept. It should be noted that the town of Erin was identified as being within one of five micro/macrobursts corridors that were mapped by the NWS. Damage surveys provided evidence of 100 to 125 mph wind gusts over these downburst corridors while an actual wind gust measurement of 128 mph was reported 1.5 miles northwest of Lebanon in Dodge county.

Damages in other parts of Wisconsin totaled $1.81 million in the northeast, $1.592 million in the northwest and $463K in the southwest. Statewide, a total of 46 injuries were reported. The derecho progressed across Wisconsin at average speeds between 50 and 60 mph with a total of 230,000 customers losing power across south central and southeast Wisconsin. As was the case in Minnesota, numerous families were without power for up to 6 days due to a shortage of utility poles.

MICHIGAN: By the time the derecho slammed into the west shoreline of Michigan around 0830 UTC, it had picked up speed to about 70 mph. Consumers Energy reported that this storm was the most destructive weather event in its history; about 600,000 of its customers were left without power. A total of 250 homes and 34 business with totally destroyed while 12,250 homes and 829 businesses sustained damage. Damage estimated totaled $166 million across western Michigan and this did not include severe damages to fruit tree orchards. A total of 4 storm-related deaths were reported along with 146 injuries.

Some of the most severe damage occurred in Spring Lake in Ottawa County and Walker in Kent county where field surveys estimated wind gusts between 120 and 130 mph. The most heart wrenching stories came from the small town of Walker, just west of Grand Rapids where $45 million in damages were sustained. Damaged and destroyed buildings left 8403 persons unemployed in Walker, while Michigan National Guard troops had to be called in to maintain a dusk to dawn curfew and clear roads of debris.

In total, some 16 Michigan counties received Federal Disaster Declarations; however, most of Detroit was spared major damage as the bow echo punched across counties immediately north of the city including the Michigan Thumb. East Michigan reported $5.586 million in damages (agricultural totals excluded) by the time the storm passed over Lakes Huron and Erie around 1130 UTC.

NEW YORK: The derecho passed across western New York between 1527 UTC and Noon EST with damage reported across a 14 county area. Storm progression rates in excess of 60 mph were noted along with $300K in damages. Considerable damage to fruit tree orchards was not included in the damage total.

3. Rear Inflow Jet Development: The first 6 hours of the 30-31 May 1998 Minnesota to New York derecho

     Weisman (1992) conducted a number of numerical simulations for developing mesoscale convective systems that demonstrated the modulating affect of CAPE and vertical wind shear on system growth and intensity. In addition, these experiments describe the role elevated rear inflow jets play on prolonging MCS life, by forcing regular regeneration of strong cells through deep lifting along the gust front.

     Both the distance and direction between the initiation point of this derecho and the KMPX RDA allowed for a detailed examination of the developing rear inflow jet at 2 km to 4 km agl. Radar inbound velocities (0.5 and 1.5 degree elevation) were mapped according to height Above Ground Level (AGL), while wind vectors were estimated with the help of temporally interpolated wind directions from the 0000 UTC 31 May 1998 ETA. It can certainly be argued that forecast wind directions from a numerical model would have limited accuracy within a developing MCS. However, the purpose of this exercise was not to achieve absolute accuracy but rather to access relative differences in the horizontal wind field and describe the temporal evolution of the rear inflow jet for this storm.

     Wind vector plots, over an approximate 15 minute interval, from 0159 UTC to 0315 UTC are provided by Figures 3a, 3b. 3c, 3d, 3e and 3f, respectively. Reflectivity images (0.5-degree elevation) appear below the wind plots. Note, in particular, the steady strengthening in wind speeds through 0245 UTC and the approach of peak winds along the gust front as the system first bowed out between 0245 and 0300 UTC. It is noteworthy that the development of the rear inflow jet with this system coincided with the first severe wind damage reports in Sibley County.

     A time series of vertical cross sections oriented along the developing bow echo is provided in Figure  4 for a more graphic view of the developing elevated rear inflow jet. See Weisman (1992) for additional details on numerical simulations of long-lived convective systems.

     An assessment of the CAPE and vertical wind profiles, at the initiation point for this storm was also completed, in order to place this MCS within the shear/buoyancy domain described by Weisman (1992). Figures 5 and 6 display these plots for Weisman's 2.5 km and 5.0 km shear simulations, respectively. These results show that this particular derecho was initiated in an area with average values of CAPE for numerical simulations of systems that develop elevated rear inflow jets.

    In terms of vertical wind profiles, the magnitude of the surface to 2.5 km agl shear was ranked on the lower range of shears for an MCS that was able to develop an elevated, rear-inflow jet. However, measured wind shear through 5 km was significantly above average when compared to figures obtained from Weisman's numerical modeling experiments. These results suggest that important clues on the dynamic forcing for this unusual event may be located in upper levels of the troposphere.

4. The synoptic environment of the 30-31 May 1998 Minnesota to New York derecho

     In this section, the synoptic environment for this case is examined in terms of how well it conformed to the synoptic climatology for derechos as outlined in JH87. In addition, an argument is made for classifying this storm as a hybrid derecho due to a unique arrangement between its CAPE field and the upper-level jet stream structure. Below is a summary of major findings in derecho climatology from JH87:

Key points in the JH87 Derecho Climatology

1). A majority develop under west to northwest flow; in 86 percent of cases studied 500 mb wind direction were 240 degrees or more. Furthermore, 500 mb wind speeds averaged a seasonally strong 41 kts.

2). A quasi-stationary east-west boundary played a key role in most derecho cases; 86 percent of cases examined began on or just north of a clearly defined thermal boundary. Generally, the progressive derechos moved nearly parallel to the thermal boundary with an angle slightly towards the warm sector.

3). Lower tropospheric warm advection was clearly evident in a majority of cases. For example, in 86 percent of cases, 850 mb warm advection could be identified at the point of derecho initiation and was evident in all cases within 320 km of origin; at 700 mb, warm advection could be located 96 percent of the time within 320 km of the point of derecho origin.

4). Extreme instability is in place along or near the path of most derechos; in 86 percent of cases the old SELS lifted index measured -6 or less.

5). Pooling of moisture was evident along or near the thermal boundary in 74 percent of cases as viewed on both the surface and 850 mb charts. Generally, pooling at 850 mb occurs along and north of the thermal boundary while surface dewpoints pool along and south of the boundary.

6). The winds in the lower to middle troposphere (defined as the 700 mb to 500 mb layer) average 38 kts for derecho events with 700 mb winds averaging 34 kts.

     A sequence of figures is provided for this case which address all six of the synoptic pattern signatures listed above. In a number of instances, different parameters are substituted for comparisons; however, the physical meaning of the comparisons remains the same.

     An ETA model generated sounding for Marshall, MN at 0000 UTC 31 May 1998 is displayed in Figure 7, a location very close to the initiation point of the MCS. The wind observation at 500 mb indicates a speed of 50 kts from 260 degrees. This figure is very much in line with (1) above. The average wind speeds between 700 mb and 500 mb layer on this sounding are 40 kts and this agrees well with (6). It should be noted that the 700 mb wind speed on the Marshall, MN sounding is almost 10 kts weaker than the average wind for this level in JH87.

     An analysis of CAPE from the 0000 UTC 31 May 1998 ETA is shown in Figure 8. This illustrates the extreme instability that was in place over southwest Minnesota when this MCS formed. The CAPE field serves as a proxy for the old SELS lifted index in (4). In addition to the CAPE analysis, reverse plotted wind shear vectors (in knots over the 0-5 km layer) are included to illustrate the orientation and magnitude of environment shear along the path of this LBE. The strong eastward component of environmental shear suggests that this derecho would be subject to strong low-level inflow on its leading edge as it propagated towards New York state.

     Initial analyses of equivalent potential temperature and moisture divergence from the 0000 UTC 31 May 1998 ETA for the 1000 mb and 850 mb levels are displayed in Figure 9 and Figure 10, respectively. The 1000 mb theta-e analysis serves to identify a strong east-west thermal boundary in ( 2) of JH87. When Figs. 9 and 10 are viewed in combination, points (3) and (5) of the JH87 synoptic climatology are satisfied.

     When examined against the JH87 synoptic climatology for LBEs, nothing really stands out from the standpoint of the initial synoptic environment assessment for this storm. The only unusual features are the extremely high levels of CAPE just south of the storm initiation point as well as an extremely sharp northern CAPE gradient extending eastward into Ohio. Note in particular, the ridge of CAPE of 4500 J/kg extending from northwest Iowa into extreme southeast South Dakota; this feature is oriented towards the Spencer, South Dakota F4 tornado.

     An analysis of 250 mb isotachs and divergence is superimposed on a satellite moisture channel image at 0000 UTC 31 May 1998 in Figure 11. The orientation of developing convective clusters is identical to the elliptically shaped field of upper-level divergence. The 250 mb divergence field on this image clearly extends through a coupled jet streak, a structure discussed by Uccellini and Kocin (1987). Figure 12 and Figure 13 provide 00 hr and 12 hr forecasts for the 250 mb jet from the 0000 UTC 31 May 1998 ETA run. It is clear from these figures that the coupled jet structure was ideally positioned to help force further convective development as the derecho progressed to the east. Figures 14, 15,  and 16 display ageostrophic wind vectors and divergence at 250 mb for the first 12 hours of the 0000 UTC 31 May 1998 ETA run which illustrate these points further.

     It can be argued that some of the ageostrophic flow generated by the ETA may very well be the result of convective feedback (See, e.g., Maddox 1983). However, any convective feedback generated for this case is still in the same sense as ageostropic winds generated by both along stream and cross stream shears between these two jet streaks. Thus, this particular MCS may very well have been locked within a region of greatly enhanced upper level dynamics as it propagated eastward. If this is the case, this storm probably represents a derecho hybrid.

5. Summary

     A two part study of the 30-31 May 1998 Minnesota to New York derecho was completed. Part (1) established that an elevated rear inflow jet developed with this system during the first 6 hours of its life while part (2) illustrated that the synoptic environment for this MCS conformed well to synoptic climatology for derechos as provided by JH87. It is suggested that a unique distribution between the CAPE field with this storm and a coupled jet structure provided an unusual dynamic/thermodynamic synergism and a hybrid LBE was the result. Additional research efforts should attempt to discriminate between LBEs that form under coupled jet structures with overlapping CAPE with systems that do not in order to determine if there are any systematic differences in system intensity or duration.

Acknowledgements: The author is appreciative of the support and encouragement to pursue this study along with critical review from Rich Naistat, SOO, MPX. In addition, the author extends thanks to Ms. Liz Page and Ms. Dolores Kiessling of the COMET staff for providing data sets.

6. References

Corfidi, S. F.,J. H. Merritt and J. M. Fritsch, 1996: Predicting the movement of mesoscale convective complexes. Weather and Forecasting, 11, 41-46.

Corfidi, S.F., 1998: Forecasting MCS mode and motion. Preprints, 19th Conf. Severe Local Storms, American Meteorological  Society, Minneapolis, MN, Amer. Meteor. Soc., 626-629.

Evans, J.S., 1998: An examination of observed shear profiles associated with long-lived bow echoes. Preprints, 19th Conference on Severe Local Storms. Minneapolis, MN, Amer. Meteor. Soc., 30-33.

Johns, R.H., and W.D. Hirt, 1987: Derechos: Widespread convectively-induced windstorms. Weather and Forecasting, 2, 32-49.

Maddox, R. D., 1983: Large-scale meteorological conditions associated with midlatitude,  Mesoscale convective complexes. Monthly Weather Review, 111, 1475-93.

Merritt, J. T., and J. M. Fritsch, 1984: On the movement of heavy precipitation areas of mid-latitude mesoscale convective complexes. Preprints, 10th Conf. on Weather Analysis and Forecasting, Clearwater, FL, Amer. Meteor. Soc., 529-536. 

NCDC, 1998: Storm Data, Vol.40, May, 408 pp. [Available from the National Climatic Data Center, Federal Building, Asheville, NC 28801-2696.]

Uccellini L.W., and P. J. Kocin, 1987: The interaction of jet streak circulations during heavy snow events along the east coast of the United States. Weather and Forecasting, 2, 289-308.

Weisman, M. L., 1992: The role of convectively-generated rear-inflow jets in the evolution of long-lived mesoconvective systems. Journal of Atmospheric Sciences, 49, 1826-1847. is the U.S. government's official web portal to all federal, state and local government web resources and services.