1.INTRODUCTION

In this paper we present analyses of airflow evolution in the tornadic Newcastle, Texas, storm of 29 May 1994 (Wakimoto and Atkins 1996, referred to as paper WA), one focus of intensive mobile mesoscale field observations during the Verification of the Origins of Rotation in Tornadoes EXperiment or VORTEX (Rasmussen et al. 1994). Our goal is to document the evolution of low level rotation of the parent mesoscale circulation (i.e. mesocyclone) of the Newcastle tornado and gain insights into the origins of mesoscale rotation preceding tornadogenesis.

2.OBSERVATIONS AND ANALYSIS

The Newcastle storm evolved from a line of deep convection that developed along a dryline in north-central Texas during mid-afternoon. In expectation of storm development, the NOAA P-3 was flying over north Texas as early as 1900 and began flying legs adjacent to developing cells around 2100 (all times are Universal Time). The P-3 flew a series of seven consecutive legs along the southwest flank of the Newcastle storm from 2240 until 2318, obtaining pseudo-dual Doppler tail radar data in the fore-aft scan or FAST mode (Jorgensen et al. 1996). Operated in 180° sector mode, the tail radar completed each fore or aft sweep in 3 seconds, corresponding to a nominal along-track data spacing of about 700 m.

In post-analysis, radar data was extensively edited to remove ground targets and regions of atmospheric reflectivity less than about 10 dBZ, and velocities were dealiased using the NCAR Solo software. Since subsequent analysis would focus on the lower levels of the storm, editing was concentrated on the lowest 5 km AGL.

In the next analysis step, radar data were spatially interpolated via the Barnes technique using a slightly modified form of the NCAR Reorder software. Radial velocities and reflectivities were interpolated to a 51 x 51 x 21 regular cartesian grid with dimension of 25 x 25 x 5 km (i.e. 0.5 km horizontal and 0.25 km vertical grid spacing). The Barnes weighting function takes the form w = exp(-r²/k ), where r is the distance from a gridpoint to a datum and k is a smoothing parameter. The frequency response of the Barnes weighting function may be expressed as D₀ = exp[-k * (p /l * )²], where k = k * L², L = 2D , and l * = l /L. Given a nominal horizontal (vertical) data spacing D = 0.7 km (0.4 km), and picking a constant k * = 0.799, the theoretical horizontal (vertical) response of the Barnes filter is ~ 50% at wavelength l = 4.8 km (2.7 km) and near zero at the Nyquist wavelength L.

In the final analysis step using the NCAR Cedric software, an iterative scheme was employed in combination with the two linear equations for the fore and aft scan radial velocities to vertically integrate the mass continuity equation from w = 0 at ground level. The following steps were taken to modify the interpolated radial velocities prior to integration: 1) hole filling; 2) two-step, 2-D Leise filtering of radial velocity fields.

3.RESULTS

The Newcastle storm began as an updraft on the west end of a bow-shaped echo which was driven by outflow convergence from the neighboring Graham storm (Fig. 1a). At this early stage the bow echo was anchored by two counter-rotating mesoscale vorticies (i.e. "bookend vorticies"), the Graham mesocyclone to the east with (x,y) coordinates (km) of (23,12), and an anticyclonic circulation at (16,13). The parent mesoscale circulation of the Newcastle tornado formed northwest of the inflow notch into the anticyclonic circulation at (14,14). Near ground and at all analysis times, the southwestern edge of the Newcastle-Graham complex marked the location of a westward-moving cold outflow boundary.

With subsequent development, (Figs. 1b-f), the Graham mesocyclone moved southward while the Newcastle cell moved slowly westward, increased in reflectivity, and developed an intense mesocyclonic circulation. The Newcastle mesocyclone continued to intensify through 2302, the approximate onset time of damage to vegetation from strong surface winds (WA), through the time of the first tornado photo from a ground team around 2309. During the development of the mesovortex, especially after 2253, strong convective updrafts were noted within and on the inflow side of the circulation center.

Vertical cross-sections through the location of maximum vertical vorticity in the developing Newcastle cell reveal the close correspondence of updraft and vertical mesocyclonic rotation (Fig. 2). As early as 2242 (Fig. 2a), a peak vertical vorticity of mesocyclone strength (i.e. 10 x 10^-3 s^-1) is located in the updraft around 5 km, with weaker rotation at lower levels. By 2250 (Fig. 2b), midlevel rotation in the updraft had more than doubled in strength while the first signs of a developing rear flank downdraft (RFD) were seen in the lowest three kilometers rough 1 km to the west of the mesocyclone. Vertical vorticity continued to increase through 2253 (Fig. 2c) and 2301 (Fig. 2d), the most notable features being the development of a secondary maximum of vertical vorticity near ground, the persistent colocation of the mesocyclone and updraft, and the vigorous RFD to the west. The mesocyclone continued to intensify at all levels through 2305 (Fig. 2e) to 2312 (Fig. 2f), when a maximum value of 34 x 10^-3 s^-1 was analyzed around the location of the tornado. The analysis at 2315 (not shown) reveals the same overall morphology and a slow westward movement of the mesocyclone.

4.DISCUSSION AND CONCLUSIONS

Motivated by WA who suggested that stretching of preexisting vertical vorticity along a mesoscale boundary intensified the Newcastle circulation and eventually initiated the tornado, we performed an analysis of vertical vorticity forcing by tilting and stretching. Departing somewhat from previous studies, we evaluated these forcing terms and the accumulated vorticity along short (~4 min) lagrangians approximating the motion of air entering the mesocyclone. The results suggest that air entered the updraft and intensifying mesocyclone at 2253 from the east (e.g. trajectory in Figs. 1c and 2c) and was subsequently stretched to the observed mesocyclone intensity (Fig. 3).

Fig. 3Time histories of vertical velocity, analyzed vertical vorticity (s^-1 x 10^-3), and predicted vertical vorticity along a short (4 min) streamline entering the Newcastle mesocyclone during its developing stage. Vertical vorticity is set to the observed value at the beginning of the trajectory, and both individual (TLT, STR) and total (TOT) tendencies are accumulated following the motion.

It is difficult to explain the observed mesocyclogenesis leading to the Newcastle tornado as stretching of vertical vorticity along a preexisting boundary, in this case the westward-moving cold outflow. Our trajectory analysis suggests that air was entering the developing mesocyclone from the east (i.e. behind the outflow boundary. A portion of these E-W trajectories (not shown) entered the edge of the mesocyclone, became ingested into the wrapping rear flank downdraft (RFD), and were ultimately lifted and stretched as they passed from the RFD into the updraft.

There are additional difficulties with the hypothesis that stretching of preexisting outflow boundary vorticity formed the Newcastle tornado. The Newcastle mesocyclone first appeared aloft, eventually building down to low levels where a secondary vorticity maximum developed prior to tornadogenesis. Moreover, the analysis revealed the updraft meso-anticyclone farther southeast along the outflow boundary that preceded the Newcastle mesocyclone. Obviously, any hypothesis of mesocyclogenesis would need to explain the observed sequence of development of the circulations of both signs.

We are presently investigating the origins of the rather weak vertical vorticity to the east the outflow boundary and Newcastle mesocyclone. Our analyses reveal that some of the inflow air moved from a roughly annular region of negative vertical vorticity into the region of weak positive vertical vorticity surrounding the Newcastle mesocyclone (e.g. Fig. 2f). We are studying the possible role of tilting to move inflow air across the interface, and plan to report additional findings at the conference.

5.ACKNOWLEDGEMENTS

We thank the NOAA Aircraft Operations Center (AOC) staff, who capably operated the P-3 aircraft and its data systems during VORTEX. We also thank NCAR staff, especially Dick Oye, Michelle Case, and Jay Miller, for assistance with the NCAR analysis software. NCAR is sponsored by the National Science Foundation.

6.REFERENCES

Jorgensen, D. P., T. Matejka, and J. D. DuGranrut, 1996: Multi-beam techniques for deriving wind fields from airborne Doppler radars. J. Met. and Atmos. Phys., 59, 83-104.

Rasmussen, E. N., J. M. Straka, R. Davies-Jones, C. A. Doswell III, F. H. Carr, M. D. Eilts, and D. R. MacGorman, 1994: Verification of the origins of rotation in tornadoes experiment: VORTEX. Bull. Amer. Meteor. Soc., 75, 995-1006.

Wakimoto, R.M., and N. T. Atkins, 1996: Observations of the origins of rotation: The Newcastle tornado during VORTEX 94. Mon. Wea. Rev., 124, 384-407.