The Convective Wind Event
of 4 November 2001

Ray Brady

1. Introduction

On 4 November 2001 a line of showers, with embedded isolated thunder, crossed the Binghamton forecast area during the late afternoon and early evening hours. Convective wind gusts of 35 to 50 mph accompanied this line (Table 1) causing isolated, very minor damage in the county warning area. The showers and thunderstorms were well forecast. However, the gusty winds that accompanied them were somewhat of a surprise. The widespread and unexpected nature of these winds made this an interesting case to look at.  The purpose of this note is to investigate whether there were any indications from model and observational data of the potential for strong gusty winds if convective showers were able to develop (and that, in itself, was a good question on this day), so that we may better recognize these events in the future.  WSR-88D reflectivity and velocity images of these storms will also be presented.

2. Synoptic Setting

The overall pattern featured west-northwesterly upper-level flow (never trust northwest flow in potential convective situations!) with embedded short wave troughs. The 1200 UTC Eta analyses (Fig. 1)  indicated that one of these troughs, stretching from New England into the mid-Atlantic states, had recently passed the region. The next short wave trough was located over the western Great Lakes, with the associated surface trough entering lower Ontario. The BGM forecast area was located under a short wave ridge situated between the 2 troughs. The 850 mb analysis at 1200 UTC  (Fig. 2) indicated fairly significant warm advection was occurring at low levels, while the 500 mb analysis suggested cold advection would be occurring at mid levels. This type of differential thermal advection would act to increase mid level lapse rates, thereby enhancing the chances for convective development. This area of steep mid-level (850-700 mb) lapse rates was forecast to be over the region late in the day (Fig. 3), and was noted in the afternoon forecast discussion (afdbgm). Another important feature to note from the 300 mb upper air analysis was the area of enhanced upper level divergence over Michigan in advance of the short wave trough.

The 6 hr Eta forecast valid at 1800 UTC showed the impressive, digging Great Lakes trough over Michigan (Fig. 4). The surface trough was forecast to enter western New York by this time with a band of weak to moderate lift and some enhanced (but fairly nominal) moisture. The upper air forecasts (Fig. 5)   indicated that differential thermal advection, implied by the 1200 UTC analyses, would be continuing at this time. The area of strong upper level (300 mb) divergence, located in the left front quadrant of a 120 kt jet, was poised to enter western New York. Note the lack of significant frontal (or prefrontal) moisture at 850 mb; a much better-defined frontal moisture axis (at least in the relative humidity fields) was situated at 700 mb (and even 500 mb) and was forecast to enter far western New York State around this time. Strong post-frontal cooling/drying was suggested by 1000-850 mb theta-E advection field.  However, the 1800 UTC Eta convective parameters were not impressive (Fig. 6), with positive LI's, very weak 1000-850 moisture convergence, and K-indices only in the teens. The 1800 UTC  Eta BGM profile certainly reinforced the notion that convection was not likely at this time over the southern tier of New York (Fig. 7)! However, note the veering winds from the surface to 800 mb representing the layer of warm advection, with a backing of the flow in the 800 to 500 mb layer associated with cold advection,  nicely illustrating the differential thermal advection that was acting to destabilize the synoptic environment.

Unfortunately, I do not have the synoptic 4-panel chart valid at 0000 UTC from the 1200 UTC Eta run (D'oh! I mis-named that gif image and overwrote it). However, the upper air forecasts valid at 0000 UTC (Fig. 8) showed an increase in the 850 mb relative humidity near the front, with the 700 mb moist axis remaining well defined. The greatest 500 mb height falls were forecast to be over New York, and the area of upper level divergence continued to slide southeast with the 300 mb jet streak. Convective parameters appeared to show some potential for development (Fig. 9), with an area of relatively lower LI's, associated with the 500 mb cold pool, forecast to be over western New York at this time. The axis of higher K-indices, with values in the teens at 1800 UTC, now displayed values of 20-25, indicating a more favorable environment for frontally-induced convection. However, 1000-850 mb moisture convergence associated with the front was forecast to remain rather weak.

The 1800 UTC mesoEta analysis of convective parameters (Fig. 10) showed an axis of slightly stronger moisture convergence; however this may have been due to the greater spatial resolution of the model. The 1800 UTC mesoEta profile (Fig. 11) for BGM was quite similar to the 6 hr fcst profile from the 1200 UTC Eta run (see Fig. 7), indicating little if any potential for prefrontal convection. However, by 2100 UTC  (Fig. 12) , the mesoEta showed a rather significant intensification of the 1000-850 moisture convergence associated with the advancing front. Similar to the 1200 UTC Eta, K indices were forecast in the 20-25 range ( which really isn't too bad considering the time of year) with strong negative theta-E advection behind the front. The 0000 UTC mesoEta convective parameters (Fig. 13) continued to show the eastward progression of the front with negative 1000-850 theta-E advection now occurring across most of upstate New York. Frontal moisture convergence remained well defined, although not as strong as 2100 UTC. An axis of K-indices in the 25-29 range extended from the western Mohawk Valley into the Poconos of northeast Pennsylvania. These K index values are impressive for a November convective event, and the trend for increasing K-index values appears to have been indication of increasing convective potential as the front crossed the BGM forecast area.

The mesoEta profiles for 2200 UTC, near the beginning/midpoint of the wind event, are rather interesting. The profiles indicate the front is still west of BGM (Fig. 14) as winds the near surface are still SW. The SYR profile (Fig. 15) appears to be more representative of a post frontal sounding, with NW flow at the surface and deeper moisture. Given the westerly flow at ELM (Fig. 16) , it appears that the front was very close to this location. All of the profiles showed very nominal capes, generally in the 30-60 j/kg range. These values are not normally supportive of thunderstorm activity, and indeed, most of the convection was in the form of showers with only isolated, embedded thunder in the BGM forecast area. This instability was not surface based, but elevated and generally confined in the 800 to 600 mb layer, where the more significant moisture was found. The ELM and BGM soundings also displayed a weak yet discernible inverted-V signature. This signature, most frequently seen in warm season profiles over the western plains and mountain regions of the western U.S.,  are associated with "dry" microbursts.  In our case, this type of profile would have supported the enhancement of convective downdrafts due to evaporational cooling. The ambient wind field on this day, although not "strong", certainly could have enhanced storm ouflow due to momentum transfer; winds below 850 mb were 25 kts or less, increasing to 30 to 35 knots at 700 mb. Steep lapse rates from the surface to 800 mb, as seen in the inverted-V profiles, would have facilitated this downward momentum transfer.  It cannot be determined in this subjective type of study which process contributed most to the gusty winds which developed on this day, but it seems likely that both evaporative cooling and momentum transfer played significant roles in their development.

The satellite imagery at 2000 UTC and 2032 UTC (Fig. 17)  showed the frontal cloud band expanding in areal coverage as it advanced into central New York.  Thunderstorms were associated with the front, evident through the positive cloud to groud strikes along a NE-SW line near Rochester.  Based on the surface pressure analysis and the winds at Dansville (DSV), it appears that this activity was developing just behind the front. Temperatures ahead of the front ranged from the mid 50s into the lower 60s with dewpoints generally in the lower to middle 30s. The 2200 UTC mesoEta profiles (Figs 15-17) showed this large T/Td spread at low levels. Near surface dewpoints were correctly depicted in the lower to mid 30s, and surface temperatures were also in the ballpark, but too cool by a few degrees, so the "inverted-V" signature may have been slightly underdone in the model profiles. It is interesting to note that the showers and thunderstorms were fairly high-based (lowest ceiling was 5000 ft), as evidenced by the BGM observations (Table 2), and which is fairly typical of these type of "inverted-V" events.

3. Radar data

The organized line of convection moved rapidly southeast through the entire BGM forecast area (Fig. 18). Isolated stronger cells displayed maximum returns of 50 to 55 dbz, with most cells showing returns of 40 to 50 dBZ. The line appeared to consolidate and strengthen as it headed into the southern tier. A closer view of the reflectivity data as the convective line passed through the local area is shown here (Fig. 19). The corresponding velocity data show the passage of the enhanced outflow/front with maximum velocities in the 36-50 mph range (Fig. 20).

4. Summary and Discussion

This was an interesting case where a line of convection, consisting of showers and embedded (but isolated) thunderstorms, produced widespread wind gusts of 35 to 50 mph. Strong dynamic forcing, in the form of a potent, digging short wave trough, along with lift due to frontal convergence, helped initiate the convection. The pre-frontal environment was characterized surface dewpoints in the lower to middle 30s with dewpoint depressions of 25-30F. Model profiles displayed an inverted-V profile, a signature which has been associated with high-based ("dry") micro/downburst producing storms. Increasing convective potential during the afternoon was indicated by the 1800 UTC mesoEta run, which showed stronger frontal convergence and increasing K-index values as the front crossed the BGM forecast area.

These cases seem to be most likely in the spring and fall,  when stronger dynamic forcing can be present in a convective environment characterized by lower surface dewpoints and potentially greater T/Td spreads. I recall a similar event a few years back which ocurred in April where a weakening and apparently non-severe convective line was entering our western zones with reflectivities only in the 30-35 dBZ range; yet it produced damaging wind gusts as it moved eastward into a warm environment with large surface dewpoint depressions. Similar type events can also occur during the summertime, but it will likely be a situation where it is very hot with temperatures in the upper 80s to mid 90s and surface Td's in the lower to mid 60s (or lower). The main point is that  the potential for strong convective wind gusts will exist when we see convection developing in an environment that displays an inverted-V signature, and where surface dewpoint depressions are significant, e.g., on the order of 25-30F. The threat will likely be greater if the development occurs with some dynamic forcing (i.e., a short wave) in an environment characterized by moderate ambient flow, especially if  the convection is able to organize into a line.

If wind damage is occuring with a convective line or velocities are greater than 50 knots,  a "Severe Thunderstorm Warning" should be issued, even if there is no thunder associated with the line. Alternatively, a short duration "High Wind Warning" could be issued, but the preference would be to issue the severe thunderstorm warning if there is an active convective line. This event was handled with strongly worded nowcasts, which highlighted the the 40-50 mph wind gusts. Per current policy, we would issue an SPS highlighting the showers and isolated thunderstorms with strong gusty winds, if they are expected to remain below severe limits.

5. References

Here are just a few microburst, downburst and inverted-V references/abstracts which have direct forecast applications.  I've highlighted the main points in the abstracts, and have included the AMS link when available. Otherwise you'll have to pick up a hard copy from the conference room!! Not included in this list is Fujita's "The Downburst" monograph, copies of which are also available in the office library.

Atkins, Nolan T., Roger M. Wakimoto, 1991: Wet Microburst Activity over the Southeastern United States: Implications for Forecasting. Weather and Forecasting: Vol. 6, No. 4, pp. 470–482.

The thermodynamic properties of wet-microburst-producing days, as observed during the 1986 MIST (MIcroburst and Severe Thunderstorm) field project, conducted in northern Alabama, have been examined and are shown to exhibit common characteristics. The parent storms and environment for this microburst type are substantially different than those documented over the High Plains in that the cloud bases are warmer, the subcloud layer is shallower, the radar reflectivities are greater, and the thermal environment is more moist and stable. Analyses of the rawinsonde data, launched in the morning and afternoon, show that low-level moisture is present and is capped by a midlevel dry layer. This midlevel dry air is generally advected from the northwest, where a large area of dry air exists over the central United States. In addition, it appears to be possible to differentiate between microburst days and thunderstorm days producing no wet microbursts by plotting the vertical profile of the equivalent potential temperature (oe). The strong wind-shear days are potentially more unstable. The difference between the surface value of oe and the minimum value aloft (in the afternoon) is greater than 20 K for the microburst days, whereas it is less than 13 K for the thunderstorm days with no microbursts. Consequently, these results suggest that they may be used by the forecaster to issue, in a timely manner (2–12 hours), a “wind-shear alert” to the general population and, more importantly, to the aviation community. Analyses of microburst storm structures indicate that they are vertically deeper than those storms developing during days with no microbursts, and that the precipitation core is largely composed of ice. Convergence, or inflow of environmental air into the microburst storms, was also commonly observed near the level of minimum oe. These wet microburst soundings and oe profiles were compared to other well-documented events. In each case, the soundings and oe profiles were similar to those derived here.

Ellrod, Gary, 1989: Environmental Conditions Associated with the Dallas Microburst Storm Determined from Satellite Soundings. Weather and Forecasting: Vol. 4, No. 4, pp. 469–484.

The thermodynamic structure of the troposphere in the vicinity of the microburst storm at Dallas-Ft. Worth Airport (DFW), Texas on 2 August 1985 is described. The analysis was based principally on a set of vertical soundings from the Visible and Infrared Spin Scan Radiometer (VISSR) Atmospheric Sounder (VAS) onboard the Geostationary Operational Environmental Satellite (GOES), valid about 1 h before the occurrence of peak surface winds. Convection in the DFW area developed in a gradient of stability on the west side of a tongue of low lifted index and high precipitable water. The lapse rates in the 850 mb-700 mb layer were large (8–9C km1). Vertical profiles of VAS data showed that DFW was in a transition zone in which conditions became drier at all levels and slightly warmer near 500 mb to the south and southwest. The midlevel warming reduced the buoyant energy available above cloud base, thus acting as a capping mechanism for the unstable, northward- moving low-level air. The potential instability was released in the vicinity of DFW by low-level convergence, caused in part by an outflow boundary from earlier convection. The storm had characteristics of both the wet and dry types of microbursts based on current models. There was a large decrease with height in total static energy (inferred from equivalent potential temperatures) from the surface to 700 mb, resulting in a source of potentially cool air fairly close to the surface.

Ellrod, Gary P., James P. Nelson III, Michael R. Witiw, Lynda Bottos, William P. Roeder, 2000: Experimental GOES Sounder Products for the Assessment of Downburst Potential. Weather and Forecasting: Vol. 15, No. 5, pp. 527–542.

Several experimental products derived from Geostationary Operational Environmental Satellite (GOES) Sounder retrievals (vertical profiles of temperature and moisture) have been developed to assist weather forecasters in assessing the potential for convective downbursts. The product suite currently includes the wind index (WINDEX), a dry microburst index, and the maximum difference in equivalent potential temperature  from the surface to 300 hPa. The products are displayed as color-coded boxes or numerical values, superimposed on GOES visible, infrared, or water vapor imagery, and are available hourly, day and night, via the Internet. After two full summers of evaluation, the products have been shown to be useful in the assessment of atmospheric conditions that may lead to strong, gusty surface winds from thunderstorms. Two case studies are presented: 1) a severe downburst storm in southern Arizona that produced historic surface wind speeds and damage, and 2) multiple dry and wet downbursts in western Kansas that resulted in minor damage. Verification involved comparing the parameters with radiosonde data, numerical model first guess data, or surface wind reports from airports, mesonetworks, or storm spotters. Mean absolute WINDEX from the GOES retrievals differed from the mean surface wind gust reports by <2 kt (1 m s1) for 82 events, but underestimated wind gusts for 7 nighttime events by 22 kt (11 m s1). GOES WINDEX was also slightly better than that derived from the concurrent National Centers for Environmental Prediction’s Eta Model first guess. There are plans to incorporate these downburst parameters into a future upgrade of the National Weather Service’s Advanced Weather Interactive Processing System, with the option to derive them from either GOES Sounder data, radiosondes, or numerical model forecast data.

Knupp, Kevin R., 1996: Structure and Evolution of a Long-Lived, Microburst-Producing Storm. Monthly Weather Review: Vol. 124, No. 12, pp. 2785–2806.

This paper describes an analysis of a long-lived, microburst-producing storm that evolved within a relatively dry environment having a relatively low CAPE value of 450 J kg1. The storm displayed a variety of kinematic and echo formations over its 2.5-h lifetime, including 1) a near equality in the strength (10 m s1) of updrafts and downdrafts, 2) strong downdrafts over an extended time period of greater than 60 min, 3) a prevalence of up-down-type downdraft trajectories associated with the strong downdrafts, 4) a prominent echo overhang during the early mature stage, 5) a spearhead-like echo protrusion during the mature storm phase that was indirectly associated with strong downdrafts, and 6) a narrow bow echo and associated weak inflow jet at midlevels during the latter storm stage. An elongated ascending branch of the up-down downdraft circulation was associated with the echo protrusion. The prominence of the up-down trajectory is corroborated by surface data and 3D numerical simulations, both of which reveal comparable values of equivalent potential temperature in the low-level inflow and downdraft outflow air. Time series plots of saturation point reveal an evaporation line structure typical of evaporation of precipitation into the subcloud boundary layer. Thus, in this case there is little evidence to indicate that significant amounts of downdraft air originated above the atmospheric boundary layer during the sustained mature to dissipating stages.

McCann, Donald W., 1994: WINDEX—A New Index for Forecasting Microburst Potential. Weather and Forecasting: Vol. 9, No. 4, pp. 532–541.

Microbursts are small-scale phenomena that have been viewed by many meteorologists as difficult to predict. However, there exists sufficient knowledge of microburst evolution by some in the research and operational communities that can be applied on the mesoscale to provide some warning to the public and aviation. This paper introduces a wind index or WINDEX that is based on this knowledge. It can be easily computed from soundings. The WINDEX is calculated from soundings known to have been taken in microburst environments and previously presented in the literature. The WINDEX can also be computed from surface observations using appropriate assumptions. This paper shows how to use the hourly surface-based WINDEX information (data) by showing its application to the infamous DFW microburst on 2 August 1985 and for three consecutive days in August 1993. The surface-based WINDEX analyses reveal a common pattern first noted by Ladd (1989); that is, microbursts primarily occur with new convection on old thunderstorm outflow boundaries. When an outflow boundary moves perpendicular to the WINDEX contours, into an area of high WINDEX values, conditions are favorable for microbursts. With this conceptual model it is possible for forecasters to give one to two hours warning that microbursts are probable for a small area.

McNulty, Richard P., 1991: Downbursts from Innocuous Clouds: An Example. Weather and Forecasting: Vol. 6, No. 1, pp. 148–154.

A microburst occurred at Goodland, Kansas, on 4 March 1990. This event was unique in the sense that it occurred at 0348 AM MST (1048 UTC) and produced 14F (7.8C) of warming. The environment along the High Plains exhibited characteristics similar to those associated with dry microbursts documented in the literature. Examination of an interpolated sounding for Goodland implied that evaporational cooling would need to occur throughout most of the dry sub-cloud layer to support the observed surface temperature.

Wakimoto, Roger M., Cathy J. Kessinger, David E. Kingsmill, 1994: Kinematic, Thermodynamic, and Visual Structure of Low-Reflectivity Microbursts. Monthly Weather Review: Vol. 122, No. 1, pp. 72–92.

On 9 July 1987, a series of low-reflectivity microbursts were studied over Colorado using dual-Doppler analyses, cloud photogrammetry, and in situ measurements collected by aircraft. These types of wind-shear events are particularly hazardous to the aviation community since the parent cloud and pendant virga shafts appear innocuous. The microburst downdrafts are shown to develop at the location where the virga shafts are, visually, the lowest and opaque. As the downdraft intensifies, sublimation and evaporation (to a smaller extent) rapidly deplete the hydrometeors and result in a shift of the axis of maximum negative vertical velocities into a relatively low reflectivity and transparent region of the virga shafts. Comparisons with weak downdrafts or null null cases reveal that the maximum radar reflectivities within the parent clouds for the two cases are comparable; however, the microburst storm consistently exhibits a larger horizontal area encompassed by the 10-dBZ contour at midlevels prior to downdraft formation.