|
<br />.
<br />
<br />(20 kIn x 20 kIn) groups of northward-moving convec-
<br />tive cells, which were separated in time by approximately
<br />1-1.5 h, The final and heaviest period of rainfall oc-
<br />curred over FCL between 2000 and 2215 MDT in as-
<br />sociation with convection originally located along the
<br />foothills southwest ofFCL that moved north-northeast
<br />and eventually became quasi stationary over the city,
<br />Constant Altitude Plan Position Indicators
<br />(CAPPls), constructed from KCYS reflectivity data
<br />and covering approximately the last two hours of the
<br />flood event, are presented in Figs, 9a-i. These figures
<br />emphasize heavily raining convection located directly
<br />over FCL, with the origin of the coordinate system
<br />located at the intersection of Taft Hill and Drake
<br />Roads, Note that the area of moderate to heavy rain
<br />(reflectivities ~ 40 dBZ) in each of the panels is rela-
<br />tively small, and that the maximum reflectivities (55-
<br />58 dBZ) are not out of the ordinary for a typical
<br />summertime thunderstorm over the northeastern plains
<br />of Colorado,
<br />Several spatial and temporal features of the precipi-
<br />tation that were highly relevant to the resultant flOO<ling
<br />can be discerned in Fig, 9, Near 2000 MDT (Fig, 9a),
<br />the final round of heavy rain developed over the al-
<br />ready saturated ground of southwestern and central
<br />FCL. As time progressed, convective elements in the
<br />northern part of the storm continued to move slowly
<br />toward the north-northeast, dissipating as they moved
<br />north of town (e,g" Figs, 9c-t), However, the south-
<br />ern extension of the convection continually regener-
<br />ated west of FCL over both the foothills and a weak
<br />outflow boundary (Fig, 10), The outflow boundary (x
<br />= -7-1; y = -3-0; Fig, 10) was generally confined to
<br />levels ~ I kIn AGL, but created a zone of convergence
<br />and new cell growth on the south-southeastern flank
<br />of the storm where the ambient east-southeasterly flow
<br />(V, > 0) encountered the westerly component (V, < 0)
<br />of the outflow (e.g" Fig, 13; x = 0, y = -1.5), Once
<br />initiated, the new cells moved northeastward (-2200 at
<br />6-8 m S-I) and down the Spring Creek basin, producing
<br />heavy rainfall as they merged with the larger echo mass,
<br />Between 2030 and 2100 MDT, and coincident with
<br />new cell development occurring on the southern and
<br />southeastern flanks of the storm, the echo area con-
<br />tracted in size (heavy rainfall became concentrated in an
<br />area ~ lOx 10 kIn'), increased in intensity, and became
<br />quasi stationary over southwestern FCL (Figs, ge-i;
<br />Fig, 10), The area contraction of the storm, most evi-
<br />dent in the radar imagery after 2100 MDT (e,g"
<br />Figs, 9e-i), is consistent with visual observations of
<br />the convection taken east of FCL around 2030 MDT,
<br />
<br />,
<br />
<br />BuJ/effn of the American Meteorological Society
<br />
<br />These observations describe the storm clouds as
<br />"originally spreading northeastward, but then con-
<br />densing into a small but ominous cloud mass over and
<br />west of the city" (Doesken and McKee 1998), After
<br />2100 MDT (Fig, ge), heavy rainfall possessing maxi-
<br />mum reflectivities of 58 dBZ (indicated by the CSU-
<br />CHILL) and core diameters on the order of 1-2 kIn
<br />continued to form on the southern and southeastern
<br />flanks of the storm and move toward the northeast,
<br />directly down the Spring Creek drainage (e,g"
<br />Figs, 9e-i), This pattern of cell regeneration, move-
<br />ment, and merger over a given location has been ob-
<br />served in many previous flash flood events (Caracena
<br />et al, 1979; Miller 1978, Chappell 1986; Smith et al,
<br />1996; Doswell et al, 1996). The production of new
<br />convective cells ceased over FCL at approximately
<br />2215 MDT (not shown), and the rainfall stopped
<br />shortly thereafter (around 2230 MDT).
<br />There was no hail or severe weather (strong winds,
<br />tornado, or funnel clouds, etc,) reported or detected in
<br />association with the storm, Further, in contrast to sev-
<br />eral previous studies of flash flood events that have
<br />noted copious amounts of cloud-to-ground (CG) light-
<br />ning (e,g" Holle and Bennet 1997; Bauer-Messmer
<br />et al, 1997; Soula et al, 1998), only 20 CG lightuing
<br />flashes were detected by the NLDN in the vicinity of
<br />FCL over the 5-h duration of the event (Fig, II; cf,
<br />section 7c), During the final two hours of the storm, a
<br />period associated with 4-6 in, of rainfall (~ 50% of the
<br />6-h total), only 7 CGs were detected by the NLDN.
<br />Lightning (Fig, II), radar (e,g" Figs, 8a-d), and sat-
<br />ellite imagery (e,g., Fig, 8b) all indicated the presence
<br />of larger, more intense areas of convection to the
<br />southeast of FCL. Indeed, when viewed instanta-
<br />neously and on a regional scale (e,g" Fig, 7), the con-
<br />vection located over FCL on the night of 28 July 1997
<br />was rather innocuous in appearance,
<br />
<br />b, Dual-Doppler observations of the horizontal
<br />wind field
<br />Surface mesoscale analyses presented in section 5
<br />provided a regional view of the low-level horizontal
<br />flow field evolution, Interestingly, the mesoscale
<br />streamline analyses (Figs, 7c-d) suggest that surface
<br />easterly winds increased in magnitude between 2000
<br />and 2100 MDT near and to the southeast of FCL, co-
<br />incident with the northeastward movement of a band
<br />of convection originally situated to the south of FCL.
<br />The relatively coarse resolution of the surface wind
<br />data in the mesoanalysis makes it difficult to isolate
<br />the role that the increase in easterly wind had on the
<br />
<br />201
<br />
|