SR/SSD 97-6 2-1-97

Technical Attachment

FORECASTING DRY MICROBURST POTENTIAL USING THE WINDEX

Greg Murdoch
NWSO Midland, Texas

1. Introduction

Conditions conducive for dry microburst events occur in the High Plains, including the plains of West Texas and eastern New Mexico (Wakimoto 1985). The dry microburst is one of two types of microscale downbursts. Wakimoto (1985) defined a dry microburst as:

A microburst that is accompanied by little or no rain between the onset and the end of the high winds, including the intermediate calm periods, if any. The dry microburst is usually associated with virga from altocumulus or high based cumulonimbus type clouds.

The WINDEX or Wind Index is based on observations (Wakimoto and Bringi 1988) and numerical models (Srivastiva 1987, Knupp 1989, and Proctor 1989) of microbursts that suggest strong downdrafts are initiated as a result of frozen precipitation falling through the melting level (McCann 1994). Thereafter, evaporative cooling maintains the negative buoyancy in the subcloud dry adiabatic layer to the surface where the intense wind gust is sometimes realized.

Most of the research with regard to the WINDEX has been directed toward wet microbursts. The purpose of this paper is to investigate known dry microburst events over West Texas and eastern New Mexico with the WINDEX in the hope of providing forecasters with a quantitative means of identifying the potential for dry microbursts. A detailed explanation of the development of WINDEX can be found in McCann (1994).

2. Data and Methodology

Dry microbursts used in this study were determined through descriptions in Storm Data (1983-1993), other dry microburst studies, and in a few instances witnessed by the author. Once dry microbursts were identified, upper air data were imported into the Skew-T/Hodograph Analysis and Research Program (SHARP) workstation (Hart and Korotky 1991) from the Radiosonde Data of North America CD-ROM (available from the National Climatic Data Center) and from NWSO Midland's local upper air data archives. The database consisted of a total of 19 upper air soundings taken at 1200 UTC between the months of March and October, from Midland, Amarillo, El Paso, and Albuquerque.

Since dry microbursts rarely occur at an upper air site, the closest site was used to represent the environment in which each event occurred. The only time constraint placed on events used in this study was that they occur during the daylight hours to ensure that the convective temperature was attained. Soundings were modified using observed and estimated surface data, and then were used to calculate the WINDEX value. The unmodified and modified database compared well with the typical 1200 and 0000 UTC microburst soundings in Fig. 1 (Wakimoto 1985). Variations between the dataset and Wakimoto's prototype soundings generally pertained to the depth of the moisture layer and the height to which the dry adiabatic layer extended.

According to Wakimoto, a typical 1200 UTC dry microburst sounding exhibits a radiation inversion, a near dry adiabatic layer extending to approximately 500 mb, a dry subcloud layer with mixing ratios of 3 to 5 g kg-1, and the presence of mid-level moisture. In addition, the convective temperature must be reached to initiate events.

WINDEX values, which are unitless, are calculated from the following equation (McCann 1994):

WI = 5[HmRq(2 - 30 + Ql - 2Qm)]

where Hm is the height of the melting level in km above ground level; Rq = Q l /12, but not greater than 1; is the lapse rate in deg C km-1 from the surface to the melting level; Q l is the mixing ratio in g kg -1 in the lowest 1 km above the surface; and Qm is the mixing ratio in g kg-1 at the melting level. In general, high WINDEX values correlate well with the wind speed associated with microbursts.

3. Results

In this study, WINDEX values agreed well with the observed or estimated wind gusts (in kt), yielding a mean error of -5.8. The negative mean error implies that the WINDEX had a tendency to underestimate wind gusts. An explanation in underestimated cases, where the cloud bases were high compared to Fig. 1, may be due to a loss of downward momentum as the absorption of latent heat ends and the adiabatic heating continues. Another explanation may lie in the fact that the average 1 km mixing ratio for the database was just below the median mixing ratio put forth by Wakimoto for a typical dry microburst sounding. Table 1 shows a sample of cases studied.

Table 1. Sample of data used and results found in assessing the use of the WINDEX in forecasting the potential for dry microburst.

4. A Brief Case Study

McCann (1994) found that secondary convection associated with boundaries moving into areas of relatively higher WINDEX are most likely to produce the strongest microbursts. McCann also suggests that convection developing in or moving into areas of higher WINDEX, can produce microbursts. A case study of a severe dry microburst event at Lubbock, Texas (July 4, 1995), illustrates the use of the WINDEX in a dry microburst forecast. The following methodology applies to the case:

  1. Determine if the potential for dry microburst exists.

  2. Provide an estimate of wind gusts based on the WINDEX.

  3. Attempt to enhance the forecast using a WINDEX analysis to depict the area most likely to experience a dry microburst.

Surface features: A weak baroclinic zone and a prefrontal trough in the vicinity of Lubbock, a dryline extending south from east of Lubbock to west of Midland and Sanderson, and surface dewpoints mostly in the 30s (F) west of the dryline (Fig. 2).

Mid-level features: Moderate westerly cyclonic flow, an approaching shortwave, moderate mid-level lift, and increasing mid-level moisture (Fig. 3).

Stability parameters: CAPE value around 100 J kg-1, LI around -2C, and 700-500 mb temperature difference of 24C. The modified 1200 UTC sounding (Fig. 4) indicated a classic Wakimoto dry microburst sounding (Inverted-V).

This information suggests that the potential for dry microbursts existed, especially the modified 1200 UTC Midland sounding and the high temperature differences between 700 and 500 mb (Sohl 1986). It does not, however, provide the forecaster with a favored area of development nor an estimate of wind gusts.

WINDEX values were calculated, plotted, and analyzed (Fig. 5) for the observing sites around West Texas and eastern New Mexico by modifying the Midland and Amarillo 1200 UTC soundings with the afternoon conditions. WINDEX estimated a wind gust of 46 kt in Lubbock. Using this WINDEX analysis in conjunction with the surface and mid-level features, along with McCann's findings and available satellite data indicated that the area most conducive for the development of dry microbursts was in the South Plains.

This was confirmed at 2207 UTC with a 50 kt wind gust at the Lubbock International Airport. A second, less intense dry microburst produced a wind gust of 37 kt at 2224 UTC, and a third produced a wind gust of 48 kt at 0055 UTC. Satellite and WINDEX analysis suggests there was an interaction between outflow boundaries and secondary convection leading to the development of this multiple dry microburst event. This tends to support the findings of McCann (1994) that severe microburst events are a result of the interaction between boundaries and convection moving into increasing areas of WINDEX. This is yet to be confirmed by examination of the Lubbock WSR-88D data, however.

5. Conclusions

Other methods of estimating the downdraft potential are also available to the forecaster. For instance, the DAPE (Downdraft Available Potential Energy), which is a measure of the negatively buoyant area (in m2 s-2), can be calculated from a plotted sounding. The DAPE is analogous to the way in which the CAPE is used to estimate updraft speed and incorporates more of the sounding data than does the WINDEX. During this study, the DAPE was used in only a few cases to estimate wind gusts, and a conclusion to whether the DAPE or the WINDEX does a better job in forecasting wind gusts associated with microbursts has not been drawn. However, one could hypothesize that since the DAPE uses more of the sounding data than the WINDEX, that the DAPE may be more representative. The most obvious advantage to the operational forecaster for using the WINDEX over the DAPE is the relative ease of calculation.

The forecaster will still have to ascertain from inspection of soundings and other data that the potential for dry microbursts exists. In doing so, forecasters need to recognize that strong updrafts need not be present; only mid-level moisture and weak updrafts (Wakimoto 1985) are necessary. In cases where virga is being reported in observations or indicated by radar, the forecaster's confidence in the likelihood for a dry microburst will be increased.

Once the forecaster determines that the potential for dry microbursts exists, the WINDEX can be easily calculated with the aid of the SHARP program and either a hand calculator or computer. Applying the findings of McCann (1994) with a WINDEX analysis and the WSR-88D may enhance the shortfuse dry microburst forecast across West Texas and eastern New Mexico by providing a quantitative method that identifies the potential for dry microbursts.

6. Acknowledgments

The author would like to thank Brian Francis (SOO) at NWSO Midland and Loren Phillips (SOO) at NWSFO Lubbock for their helpful comments and reviews.

References

Hart, J.A., and W.D. Korotky, 1991: The SHARP Workstation - V1.50 - A skew T/hodograph analysis and research program for the IBM and compatible PC. User's manual. NOAA/NWS Forecast Office, Charleston, WV, 62 pp.

Knupp, K.R., 1989: Numerical simulations of low-level downdraft initiation within precipitating cumulonimbi: Some preliminary results. Mon. Wea. Rev., 117, 1517-1529.

McCann, Donald W., 1994: WINDEX - A New Index for Forecasting Microburst Potential. Wea. Forecasting, 9, 532-541.

Proctor, F.H., 1988: Numerical simulations of the 2 August 1985 DFW microburst with the three-dimensional terminal Area Simulation Systems. Proc. 15th Conference on Severe Local Storms, Baltimore, MD, Amer. Meteor. Soc., J99-J102.

Sohl, C.J., 1987: West Texas dry microburst of 21 May 1986. NOAA Tech Memo NWS SR-121, Scientific Services Division, NWS Southern Region Fort Worth, TX, 1-9.

Srivastiva, R.C., 1985: A simple model of evaporative driven downdraft: Application to microburst downdraft. J. Atoms. Sci., 42, 1004- 1023.

Wakimoto, R.M., 1985: Forecasting dry microburst activity over the High Plains. Mon. Wea. Rev., 113, 1131-1143.

___, and V.N. Bringi, 1988: Dual-Polarization observations of microbursts associated with intense convection: The 20 July storm during MIST project. Mon. Wea. Rev., 116, 1521-1539.