Doug Green
Science Operations Officer
WFO Phoenix Arizona
Introduction
During the evening of 22 August 2003, powerful microbursts struck north-central Pinal County and extreme southeast Maricopa County, Arizona. Greatest damage occurred in Queen Creek, a small community in the far southeast portion of the Greater Phoenix metropolitan area, as well as the new Johnson Ranch subdivision, located several miles southeast of Queen Creek. Winds estimated at 80-100 mph blew down more than 200 power poles, destroyed many outbuildings (a horse was killed when one outbuilding collapsed), smashed windshields and windows on numerous vehicles, damaged or destroyed many roofs of homes, uprooted hundreds of mature trees and destroyed farm irrigation equipment. Fortunately, no serious human injuries were reported. This TA-Lite will provide a brief synoptic overview, describe storm development and evolution, and end with a discussion/conclusions section.
Synoptic Situation
The large-scale weather pattern over much of the United States was dominated by a ridge of high pressure in the mid-troposphere, with its western extremity over east-central Arizona (Fig. 1). Examination of 1200 UTC soundings over southern Arizona indicated that, in the 700-400 hPa layer, 20-35 kt southerly winds existed over Yuma, 10-15 kt southerly wind existed over Phoenix, and 15-20 kt southeasterly winds existed over Tucson. Unstable, moist conditions existed over south central Arizona; the 1200 UTC Phoenix sounding indicated that mixed-layer (sfc-700 hPa) CAPE was ~1150 j/kg, while surface-700 hPa mean mixing ratio was 11.3 g/kg (Fig. 2). Surface dew points over south central Arizona remained in the 60-65 F range during the day, 4-6 degrees above the climatological average. GOES precipitable water images, soundings, and GPS-IPW sensors indicated that a moisture ridge, with precipitable water between 1.5 and 1.75", was oriented north-south across western and central Arizona. Noticeably drier air existed over southern California and over far eastern Arizona (Fig. 3).
A short wave trough moved northeast across southern California and southern Nevada on 22 August, while an easterly wave progressed slowly west across northern and central Mexico. At the southern end of the easterly wave, a tropical disturbance (soon to develop into Hurricane Ignacio) was taking shape just south of the entrance to the Gulf of California. The easterly wave was expected to enhance thunderstorm development over far southern Arizona, with more significant storms expected over and south of the state of Sonora, Mexico, during the evening. No well-defined diffluent or divergent flow existed over south central Arizona at and above 300 hPa during the day on 22 August 2003; however, a 300-200 hPa anticyclone center, which existed near the Arizona-New Mexico border at 1200 UTC, was forecast by the eta and GFS to drift slowly southwest and be situated over southern Arizona during the night of 22-23 August.
Given the presence
of relatively large instability, southerly winds in the 700-400 hPa layer,
and no apparent upper level forcing mechanism, thunderstorms were a possibility
over much of the Phoenix CWA during the upcoming evening. The afternoon zone
forecast prepared on 22
August for south-central Arizona, including northern Pinal County and the
Greater Phoenix area, called for a 30% chance of evening showers and thunderstorms,
with the possibility of blowing dust, for the evening/night of 22-23 August
2003 (the entire Phoenix CWA was forecast to have a 20-30 percent chance for
storms overnight, with highest POP over south central Arizona). Outflow from
storms over southern Arizona was expected to move north and trigger isolated
to scattered storms over southwest and south-central Arizona. The morning
HWO had highlighted the possibility of locally heavy rainfall with any storms,
but made no mention of the possibility of damaging wind.
The large-scale weather pattern had changed very little by 0000 UTC 23 Aug 2003 (Fig. 4). The 0000 UTC Phoenix sounding (Fig. 5) revealed drying above the 850 hPa level, with precipitable water lowering by nearly two tenths of an inch in 12 hours, which led to a sharp reduction in surface-700 hPa based CAPE (from > 1000 j/kg to < 200 j/kg), even though surface-based CAPE was quite large (> 1700 j/kg). A special sounding taken at Phoenix at 0300 UTC (Fig. 6) indicated moistening had occurred, with precipitable water rising to 1.67". Surface-700 hPa based CAPE rose to > 900 j/kg, while the lifted index lowered to - 4 C. Rapid changes in moisture are typical over the lower desert of south central Arizona, even in environments characterized by weak flow aloft.
Storm Evolution During the Late Afternoon and Evening of 22 August 2003
Thunderstorm development occurred over the higher terrain of southern Gila County east of Phoenix during mid-afternoon of 22 August. Locally heavy rain (1-2 inches per hour) occurred over an unpopulated region of southern Gila County between 2300 UTC and 0100 UTC, and a small stream flood advisory was issued for that area at 0013 UTC 23 August (513 pm MST 22 August). Although these storms appeared to generate enough rain to produce a modest cold pool, no discernable outflow boundary (OB) was detected via the KIWA WSR-88D 0.5 degree base velocity products (however, the moistening observed on the 0300 UTC Phoenix sounding appears to have resulted as a consequence of atmospheric moistening just upstream from Phoenix due to the upstream convection).
A larger area of strong to severe thunderstorms developed near Tucson by late afternoon, evolved into a MCS, and propagated slowly toward the west and north between 2300Z and 0300Z (Figs. 7 8 9 10). The MCS generated a decent cold pool, with several outflow surges/segments noted on radar. The first OB pushed north and west from storms that affected the Tucson area, while a second OB moved north from storms over north-central Pima and far southwest Pinal Counties. The leading edges of the thunderstorm outflow boundaries were noted via the KIWA WSR-88D as they reached Pinal County (Figs.11-12).
Noting the degree of instability and low-level moisture on the 0300 UTC Phoenix sounding, as well as GOES-derived stability and precipitable water fields (Fig. 13), warning forecasters at WFO Phoenix expected the OB to trigger new storms over central and northern Pinal County, as well as western Maricopa County. New storm development occurred just prior to 0400Z over north-central Pinal County at what appeared to be intersecting OB segments, precisely where a pre-existing west-east oriented band of low-mid level cumuliform clouds existed. The interaction between the pre-existing cloud band and the intersecting outflows resulted in the rapid development of several severe multicell storms in a narrow corridor of north-central Pinal County and far southeast Maricopa County between 0350Z and 0500Z, primarily along the eastern periphery of the primary OB (Figs.14 15 16 17 18 19). At least three individual microbursts were observed on radar; the two strongest low-level divergent signatures occurred over north central Pinal County and extreme southeast Maricopa County, coincident with the most significant wind damage (Figs. 20-21). Since these storms developed very close to the KIWA WSR-88D, the KFSX and KEMX radars provided very useful information regarding storm development and intensity (primarily via VIL, LRM-middle and LRM-upper products).
Discussion and Conclusions
At least three intense microbursts struck a small portion of north-central Pinal and extreme southeast Maricopa Counties between 9 pm and 10 pm MST on 22 August 2003. The initial microburst-producing storm developed at the intersection of pre-existing low-middle level cumuliform clouds and intersecting thunderstorm OBs. Outflow associated with the first microburst interacted with the pre-existing OB to spawn a second storm, and the microburst produced by the second storm interacted with the OB to spawn a third storm, which quickly produced the most intense of the three microbursts, which affected the town of Queen Creek. Microburst "families" such as this have been observed prior to 22 August 2003, but the proximity of this event to the KIWA WSR-88D and the availability of WES provided a great opportunity for forecasters to examine the evolution of this event as it related to synoptic and mesoscale conditions.
Although severe thunderstorm
development was triggered over a small portion of south-central Arizona, areal
coverage of storms over the Phoenix CWA the night of 22-23 August 2003 was
less than expected, and no thunderstorms occurred west of Phoenix. It appeared
that meager upper level support, coupled with dry advection between 850 and
700 hPa, were key contributors to the paucity of storms over much of the CWA.
Severe storm development was restricted to the area where a localized area
of low level moistening (and resultant increase in CAPE and lowering of the
LFC) was affected by colliding outflows from distant storms.
Figure Captions
Figure 1. 500 hPa heights overlain on water vapor image at 1200 UTC 22 August 2003.
Figure 2. Phoenix sounding taken at 1200 UTC 22 August 2003.
Figure 3. Four panel display of GOES-based fields at 2000 UTC 22 August 2003: clockwise from lower left, CAPE, precipitable water, lifted index, and sounder lifted index.
Figure 4. 500 hPa heights overlain on water vapor image at 0000 UTC 23 August 2003.
Figure 5. Phoenix sounding taken at 0000 UTC 23 August 2003.
Figure 6. Phoenix sounding taken at 0300 UTC 23 August 2003.
Figure 7. Visible satellite image, overlain by metars and 15-minute cloud-to-ground lightning, valid 2200 UTC 22 August 2003.
Figure 8. Visible satellite image, overlain by metars and 15-minute cloud-to-ground lightning, valid 0000 UTC 23 August 2003.
Figure 9. Infrared satellite image, overlain by 1 hour cloud-to-ground lightning, valid 0000 UTC 23 August 2003.
Figure 10. Infrared satellite image, overlain by 1 hour cloud-to-ground lightning, valid at 0300 UTC 23 August 2003.
Figure 11. KIWA WSR-88D 0.5 degree 8-bit base velocity image, overlain by 5-minute cloud-to-ground lightning, valid at 0315 UTC 23 August 2003.
Figure 12. KIWA WSR-88D 0.5 degree 8-bit base velocity image, overlain by 5-minute cloud-to-ground lightning, valid at 0345 UTC 23 August 2003.
Figure 13. Four panel display of GOES-based fields at 0400 UTC 23 August 2003: clockwise from lower left, CAPE, precipitable water, lifted index, and sounder lifted index.
Figure 14. Four panel radar displays, valid at 0357 UTC 23 August 2003: clockwise from lower left, KFSX VIL, KIWA 0.5 degree 8-bit base velocity, KIWA 0.5 degree 8-bit base reflectivity, and KIWA 19.5 degree 8-bit base reflectivity.
Figure 15. Same as Figure 14, except valid at 0407 UTC 23 August 2003.
Figure 16. Same as Figure 14, except valid at 0417 UTC 23 August 2003.
Figure 17. Same as Figure 14, except valid at 0428 UTC 23 August 2003.
Figure 18. Same as Figure 14, except valid at 0438 UTC 23 August 2003.
Figure 19. Same as Figure 14, except valid at 0448 UTC 23 August 2003.
Figure 20. Microburst "footprint" near Magma, valid at 0422 UTC 23 August 2003. 85 knot radial velocity difference between peak inbound (57 knots) and peak outbound (28 knots) values.
Figure 21. Microburst
"footprint" in Queen Creek, valid at 0443 UTC 23 August 2003. 103
knot radial velocity difference between peak inbound velocity (69 knots) and
peak outbound velocity (34 knots).
End
