7. Summary and Operational Considerations

Key points in this module are:

1. Microbursts are generated from storms with elevated cloud bases (>3 km AGL). Research has shown that convergence in the cloud and a descending reflectivity core frequently occur relative to the elevated cloud base. Understanding which radar elevation slice is actually observing the approximate altitude of the elevated cloud base is extremely important. The lowest radar tilt may be well below the cloud base and sampling fictitiously high reflectivity values caused by melting ice conglomerates (e.g., graupel or hail). The higher tilts are frequently the more appropriate elevation slices to better observe the structure of the cloud. The forecaster should use sounding data to develop an estimate of the anticipated base of the storms. In Fig. 1 and Fig. 2, the cloud base is inferred to be approximately 10,000 ft AGL by modifying and lifting a surface-based parcel on the 1200 UTC sounding using expected afternoon conditions. In this case, the cloud base would mostly be sampled by the 2.4° elevation slice. Of course, as the cell moves closer (farther) from the radar site, higher (lower) elevation slices will have to used to sample the cloud base.

If there is reasonable certainty that cloud bases are the result of surface-based lifting, then a first guess approximation of cloud base heights can be made by dividing the temperature-dewpoint difference or "spread") by a lapse rate of -4.5° F/1000 ft or -2.5° C/1000 ft). This simple conversion is derived from the rate at which the dry adiabatic temperature (-5.4° F/1000 ft or -3.0° C/1000 ft) and dewpoint (-0.9° F/1000 ft or -0.5° C/1000 ft) lapse rates approach each other for unsaturated adiabatic ascent in the lower troposphere (Example: a temperature/dewpoint of 95° F/50° F yields a “spread” of 45° F which equates to an estimated cloud base height of 10,000 ft AGL).

2. WSR-88D velocity signatures may provide 0 to 10 minutes microburst warning lead time (Roberts and Wilson, 1989). However, those signatures are generally weak (~12 kt convergence over 5 nm or less) and relying solely upon convergence above the cloud base may result in false alarms. Stronger convergence signatures (~22-25 kt) occasionally occur prior to peaks in severe surface winds. Reflectivity values near the freezing level are an important parameter to monitor in predicting severe and near-severe microburst winds. Although descending reflectivity cores are relatively small in size, they tend to be reliable microburst predictors and forecasters need to monitor cell tendencies very closely.

3. The dry microburst potential gust application is a linear function; for downdraft depths different from 3800 m, there will be a corresponding percentage difference that must be added to or subtracted from the predicted gust value for 3800 m (Example: a 50 dBZ core falling through a 3800 m thick dry adiabatic (-9.8° C km^-1) sub-cloud layer could produce a potential gust of 64 kt; whereas, for a fall depth of 3000 m (4500 m), the potential gust would be 50.5 kt (75.8 kt).

4. After the initial isolated microburst, small clusters tend to form as new cells develop along the edge of the outflow boundary. Subsequent cores may be weaker in terms of reflectivity, but stronger mid-level convergence can compensate for the lack of strong thermodynamics and generate equal surface winds. This appears to be typical of many western U.S. dry microburst storms which presents a challenge to forecasters attempting to observe individual storm characteristics.

5. Storm reflectivity values were too weak for detection and identification by the WSR-88D SCIT (Storm Cell Identification and Tracking) algorithm. Although lowering the SCIT minimum reflectivity threshold will allow for more cells to be identified and tracked by the algorithm, it may also produce unanticipated consequences such as too many weak cells or spurious features being identified.

6. The morning soundings typically show high relative humidity at mid-levels and a dry boundary layer producing the classic "inverted-V" temperature and moisture profile. However, storms occasionally occur when very little mid-level moisture is indicated. The, reason for this apparent disparity is there are significant mesoscale variations in instability (and vertical winds) not measured by the very coarse upper-air observing network. This means that forecasters need to pay close attention to nearby moisture advection on constant pressure and isentropic charts and in water vapor satellite imagery.

7. Melting below the 0° C level can result in exaggerated reflectivity values and false alarms (i.e., excessively high wind gust predictions). Therefore, forecasters should always focus on reflectivity cores at or just above cloud base.

8. To quickly assess appropriate dBZ values for several pulse storms at various ranges from a radar site, it is recommended that the Layer Composite Reflectivity Maximum (LRM) lower product (surface to 24,000 ft MSL) be used, and that the lower bound of the LRM-low product be set at the mean sea-level height of estimated cloud bases or the height of the freezing/melting level. Changes to the LRM-low product are made at the WSR-88D Unit Control Position (UCP) by entering SE,*****,L and pressing RETURN, where ***** is the first level password. This displays the Layer Product Parameters edit screen. Type M and press RETURN and make the desired change under column L0. The lowest layer height cannot exceed 18,000 ft MSL.

9. Peak wind gusts obtained from the Salt Lake City gust nomogram are based on only water evaporating into the the sub-cloud air. If a significant concentration of ice particles (snow/graupel/small hail) is suspected to be the main cause of the high reflectivity echo above the freezing level, then the predicted gust will be too low due to the additional latent heat of melting (80 cal g^-1) of ice which results in a total latent heat of sublimation (melting plus evaporation) of 680 cal g^-1.

10. Microbursts can be very small producing localized horizontal effects less than 2-3 nm in diameter.

11. The potential gust application (nomogram) is sensitive to both reflectivity values and sub-cloud lapse rates.

12. This gust prediction application is just that -- a prediction of peak wind gusts only -- not a predcition of wind damage; many unknown engineering factors are involved when trying to predict occurrences of wind damage. Also, users should not focus in on the specific gust values, but instead consider a range of near-severe and severe winds.

13. Owing to the relatively small data set used in this attempt to quantify and validate Srivastava's model results, these gust prediction considerations should be treated as being empirical at this stage.

Acknowledgments

One-minute average surface wind data were provided by the Utah State Department of Environmental Quality Air Monitoring Center.

Radar data analyses and product displays were obtained from the NEXRAD Operational Support Facility's WATADS (WSR-88D Algorithm Testing and Display System) software. For more information on this system, the reader is referred to the WATADS home page.

Upper-air sounding plots and analyses were obtained from the SHARP -- Skew-T/Hodograph Analysis and Research Program -- software (Hart and Korotky, 1991; Hart, 1995).

Mesonet data are courtesy of the University of Utah mesonet.