The largely ignored stepchild of Doppler weather radar data is the second moment, the base Velocity Spectrum Width (SW). Operational applications of the SW, if attempted at all, have typically been restricted to data quality evaluations. When the SW values are broad, velocity data is viewed as questionable, especially when base reflectivity data is very weak implying that the ratio of peak signal power to noise power, Signal to Noise Ratio (SNR), is low. Operationally, data quality is also questionable and reflected in broad widths when areas of range overlaid echo are not properly identified and removed. This is especially apparent when the SW values bordering range-overlaid echo are significantly broadened over the surrounding values.

However, there appear to be a much wider variety of operational applications for these SW estimates. It is suggested here that SW estimates can, within convective storms, aid in identification of particularly broad and intense updrafts, to an extent downdrafts, three-body scattering (indicative of very large hail), and low level gust fronts and related Deep Convergence Zones. Apart from convective storms, it is also suggested that SW may be used to provide clues to trends in tropical storm strength. Moreover, the depth of orographic induced wind shear and turbulence may be estimated.

Of course, as with any of the three Doppler moments (see Section 2), but particularly true of velocity spectrum width, it can not be interpreted and applied independently of those moments.

Before discussing SW applications, we will examine briefly what the estimate is and what contributes to it. Of course, the echo power or the zero moment of the pulsed-Doppler weather radar, reflectivity, is an indicator (in precipitating systems) of the liquid water content within the pulse volume. The first moment, mean radial Doppler velocity, is the mean radial motion of the power-weighted scatterers within the pulse volume. Finally, and the subject of this paper, the velocity spectrum width, is the square root of the second moment about the first of the normalized spectrum (Doviak and Zrnic, 1984). More simply, it is a measure of the velocity dispersion within the pulse volume. The velocity spectrum width within the pulse volume has a number of contributors including wind shear, turbulence, particle fallspeed dispersion, and antenna rotation. Moreover, clutter, clutter residue, system noise and artifacts can contribute but these are generally minimized or eliminated. Because antenna elevation is typically 20^o or less for surveillance weather radars, precipitation fallspeed dispersion is ignored. Further, even though measured values are obtained as the antenna sweeps horizontally, which contributes to measurement decoralation, this contribution is typically small and also ignored. Thus, assuming sufficiently high SNR, SW is reduced to wind shear and turbulence within the pulse volume. It should also be apparent that in most situations SW will increase with range from the radar. This is a natural consequence of the typical shear of the horizontal winds with height and the beam broadening with range.

It is suggested here that one of the applications of low (narrow) spectrum widths is the identification of particularly intense updrafts in certain very severe storms. Browning (1978) and the many studies he sites, indicate that, while at cloud base organized hailstorm updrafts are typically negatively buoyant, aircraft and other measurements find air ascending there uniformly and smoothly at 5 – 15ms^-1. Further, these updrafts were found to extend upward through the full depth of the Weak Echo Region (WER). Musil, et. al, (1986) reported that during the penetration of a supercellular storm at 6 km AGL, the T-28 aircraft encountered a very smooth, broad, virtually unmixed, 50 ms^-1 updraft. During a study of the Lahoma, OK extremely severe supercellular storm by Lemon and Parker (1996), updrafts were found to be marked in WSR-88D observations by SWs less than 4ms^-1 and often near 1ms^-1. In fact, Lemon (1998) using these data and data obtained from another very severe supercellular storm, concluded that one of the updraft markers within these severe storms are low velocity SWs that extend upward 2/3 of storm depth within and above the Bounded WER (BWER). In fact, he found that in the case of the Lahoma storm, these narrow widths extended upward within the updraft all the way to storm summit. It is speculated that these undiluted updraft cores are often very smooth due to vertical air-stream acceleration which damps out turbulence, because condensation processes are inherently smooth, and perhaps because helicity within supercellular storms inhibits (Drogemeier, 1993) the downward cascade of energy within updrafts. Thus, in contrast to less severe storms, low SW observations in mid-storm levels, when combined with the presence of a high reflectivity core, and conservation of sub-cloud angular momentum, can be used to locate the unmixed, nearly adiabatic core of broad updrafts and to characterize these updrafts as being particularly intense.

Related to these particularly intense updrafts and the storm-associated mesodepresion is the near surface, 5 – 15 ms^-1 horizontal wind acceleration into these updrafts. Like the storm updrafts themselves, this low-level (~ 1 to 2 km deep) updraft inflow is itself also marked by particularly low SW values. Again, it is suggested that this inflow is smooth because of the air-stream acceleration damping out the frictionally induced turbulence as it arises.

Particularly broad SW can be used to help discern Three-body scattering and the "Three-body Scatter Spike" (TBSS). Lemon (1998) found that the TBSS is a radar-artifact signature of the presence of particularly large hail within hailstorms and can be used as the basis of a radar-based severe thunderstorm warning. The TBSS is a generally 10-30-km long region of weak (5 to 15 dBZ) echo aligned radially downrange from a mid-level highly reflective echo core. One of the characteristics of these signatures is their particularly broad SWs. In many instances, the TBSS can be easily identified as a narrow low reflectivity "spike" extending outward from the edge of the storm echo. However, in some instances these TBSSs are confined within the parent storm echo itself or are obscured by other adjacent echoes. In these cases, at times the only evidence for the existence of the TBSS is its broad SW confined to these down range radials behind the reflectivity core. Further, as explained by Lemon (1998), the TBSS-associated weak negative velocities can tend to obscure velocity signatures of the mesocyclone or Tornadic Vortex Signatures internal to the parent echo. Again, it may be only the broad SW confined to these down range radials that provide clues to velocity signature obscuration.

Another well-known use of broad spectrum widths is as an aid to front, gust front, or wind-shift identification. However, perhaps a less well known but closely related use is as an aid to identification of the "Deep Convergence Zone" (Lemon and Burgess, 1993; Lemon and Parker, 1996) or as others term this structure, the "Mid-Altitude Radial Convergence" velocity signature (O’Sullivan, et al., 1998). This is a zone that has its surface roots in the gust front but that extends upward to extraordinary depths, sometimes averaging 10 km in vertical extent with extremes of 13 km. As the names imply, the vertical curtain of intense convergence is a prominent characteristic but because it is also an intense mixing zone of two very different air-streams, very high SW throughout its depth also marks it. The importance, among other things, of the DCZ or MARC is that it seems to play a very important role in downdraft augmentation and damaging surface wind production (Lemon and Parker, 1996).

Lemon and Parker (1996) also emphasized the turbulent nature of downdraft, which was therefore marked by broad SW. Evaporational processes are by nature turbulent processes and may help explain the broad SWs found in downdraft regions. Further, over much of its lower portions the downdraft is decelerating into the earth’s surface. In contrast to accelerating fluid streams, decelerating streams are characterized by turbulent motions.

A final consideration for convective storms is the use of SW for recognition of all sizes of vortices. In fact, this was at one time believed to be the principal way to detect tornadoes. Further, this same thing has again been suggested recently. However, it must be emphasized that, as we have already seen, there are a variety of causes for broad SWs. In itself, it is a non-unique signature. Moreover, when examining only the SW field, these very small vortex-related regions of broad spectrum widths will rarely stand out in contrast to those associated with the DCZ or with the TBSS. For this reason, considerable caution should be exercised when attempting to apply the spectrum width radar products for vortex detection. The velocity signature associated with these vortices are far more useful (Burgess and Lemon, 1990).

Whereas confidence in the proceeding discussions concerning SW applications to convective storms was high, here, due to the limited data sets, confidence in tropical storm applications is lower. However, some observations noted here should be further investigated. Most inferences drawn here are from observations of super Typhoon Herb that struck the Island nation of Taiwan on July 31, 1996. The newly installed WSR-88D scanned the Typhoon from its mountain top (~ 800 meters MSL) location for many hours prior to radar failure when the radome gave way after experiencing sustained winds over 70 ms^-1 for a prolonged period.

As mentioned previously, fluid flow undergoing acceleration typically tends to be less turbulent than non-accelerating or decelerating flows. During a period of apparent typhoon intensification and deepening prior to landfall, and despite strong sustained and increasing winds over 60 ms^-1, WSR-88D observations suggested remarkably smooth (low SW) flow surrounding and including much of the eye wall itself.

4.2 Broad SW Applications

After super Typhoon Herb began to significantly interact with the mountainous Taiwan island, and began to weaken and fill somewhat, WSR-88D observations continued. These observations were typified by larger SW and much larger areas of apparent turbulent flow, even in areas where the flow had not yet intercepted the island.

Flow over and around the mountainous areas were also very apparent. In fact, the orographic-induced turbulence could easily be seen in the SW data. More over, as the antenna was elevated it was obvious that it was possible – within the height resolution of the radar – to estimate the turbulence depth and effective upper limit of the turbulence using SW data. Thus, in order for aircraft to avoid this high wind induced turbulence, ground control could recommend the proper flight altitude. Previous studies suggested SW values of 4 ms^-1 (~ 8 kts) or greater could be used to suggest the presence of severe turbulence.

Although at this time Doppler radar velocity spectrum width data are largely ignored, there are a variety of potential operational applications. Even now these SW data are used to a limited extent for data quality evaluations and even for detection of windshift lines of all types, e.g., fronts, gust fronts, etc. However, there are many other potential operational applications. For example, in meteorological situations where wind fields are undergoing significant accelerations, turbulence and shears within these fields are naturally damped. Examples of these smooth flows include horizontal inflow toward the updrafts and the essentially undiluted updraft cores themselves. Flow fields undergoing decelerations are characterized instead by broad widths. Downdrafts are characterized by broad widths and significantly more turbulence, especially during deceleration into the earth’s surface. But, the interfaces themselves between these up and down drafts are especially turbulent as indicated by the very broad SW values found there. Three-body scattering associated with very large hail is also easily identified owing to the very broad associated SW values. Extent and height of oragraphic turbulence and shear can also be identified by the associated broad SWs.

Browning, K. A., 1978: The structure and mechanisms of hailstorms. Amer. Meteor. Soc., Monograph, #38, 43 pp.

Burgess, D. W., and L. R. Lemon, 1990: Severe thunderstorm detection by radar. Chapt. 30a, Radar in Meteorology. Editor, D> Atlas, Amer. Meteor. Soc., Boston, 619-648.

Doviak, R., Zrnic, 1984: Doppler Radar and Weather Observations. Academic Press, Inc.,458 pp.

Droegemeier, K. K., and S. M. Lazarus, 1993: The influence of helicity on numerically simulated convective storms. Mon. Weath. Rev., 121, 2005-2029.

Lemon, L. R., 1998: The radar "three-body scatter spike": an operational large-hail signature. Weather and Forecasting, 13, 327-340.

---------, D. W. Burgess, 1993: Supercell deep convergence zone revealed by a WSR-88D. Preprints, 26^th International Conf. on Radar Meteor., Paris, France, Amer. Meteor. Soc., 206-298.

---------, and S. Parker, 1996: The Lahoma deep convergence zone: its characteristics, and role in storm dynamics and severity. Preprints, 18^th Conf. on Severe Local Storms, Boston, Amer. Meteor. Soc., 70-75.

Musil, D.J., A. J. Heymsfield , P. L. Smith, 1986:Microphysical characteristics of a well-developed weak echo region in a high planes Supercell thunderstorm. J. of Climate and Applied Meteor., 25, 1037-1051.

O’Sullivan, J. M., R. W. Przybylinski, and Y. Lin, 1998: The 27 may 1996 severe storm event over central and eastern Missouri, a challenge in warning for the initial onset of damaging winds and non-supercell tornadoes. Preprints, 19t^h Conf. on Severe Local Storms, Minneapolis, 518-521.