|
<br />and an isohyetal analysis was performed (e,g" Figs,
<br />3a,b), To extend the rain gauge analysis on the extreme
<br />western edge of the city where rain gauge data were
<br />sparse, radar rainfall estimates (CSU-CHILL and
<br />NEXRAD) were used to quantitatively adjust the over-
<br />all pattern of the precipitation isohyets, This was ac-
<br />complished in a two-step process, First, the ratio of rain
<br />gauge to CSU-CHILL radar -estimated rainfall (cf,
<br />section 7b) over the gauge sites was determined (and
<br />found to be nearly constant), Next, radar-estimated
<br />rainfall totals for multiple grid points located on the
<br />western side of FCL were multiplied by the ratio and
<br />the isohyetal analysis was completed.
<br />A key component of the investigation was the col-
<br />lection of recording rain gauge data to document the
<br />timing and intensity of rainfall, While no systematic
<br />network of recording rain gauges existed in the FCL
<br />area, a total of 14 recording gauges were located in
<br />the immediate vicinity, a few with l-min time reso-
<br />lution and O,OI-in, resolution, The data from these
<br />gauges provided the necessary information to deduce
<br />the timing of rainfall events, thus greatly facilitating
<br />the interpretation of the manual measurements, some
<br />of which were single storm accumulations and others
<br />that were measurements for specific time periods dur-
<br />ing the storm,
<br />Observed rainfall patterns for two periods are
<br />shown in Figs, 3a,b, The total rainfall from the late
<br />afternoon of 27 July through midday 28 July is shown
<br />in Fig, 3a, Most of the city of FCL recorded between
<br />0,6 and 2 in, during this period, However, as much as
<br />4 in, of rain fell over the far western part of FCL, and
<br />more than 9 in, ofrain fell over a small area 12,8 km
<br />northwest of the center ofFCL and immediately north-
<br />west of the town of Laporte. Rainfall accumulations
<br />from the evening of the flash flood (28 July) are shown
<br />in Fig. 3b. The observed rainfall accumulations for this
<br />5,S-h period exceeded 10 in, over extreme southwest-
<br />ern FCL (and over Spring Creek in particular; Fig, 3b)
<br />and were in excess of 6 in, over most portions of west-
<br />ern FCL. A striking rainfall gradient was present to
<br />the southeast of the point of maximum precipitation,
<br />where rainfall totals decreased from lOin, to less than
<br />2 in, in a distance of ouly 4 km,
<br />
<br />4. The synoptic environment
<br />
<br />a. Midtroposphere
<br />Expansive high pressure layover the west-central
<br />United States during the final days of July 1997 as
<br />
<br />Bulletin of the Amer;can Meteorolog;col Society
<br />
<br />evinced by the presence of a ridge with an axis that
<br />extended from the Mississippi Delta to the Pacific
<br />Northwest (Fig, 4), Within the western periphery of
<br />the ridge was a shortwave trough extending from Ne-
<br />vada into southern Wyoming, At the western side of
<br />the trough was a maximum in vorticity over the border
<br />of Nevada and Utah; the maximum had moved north-
<br />westward from its original location in south-central
<br />Utah 12 h earlier. The eastern part of the trough had a
<br />weak and narrow ann of vorticity extending into south-
<br />ern Wyoming, according to the Nested Grid Model
<br />initializations at 0600 and 1800 MDT (not shown), The
<br />easterly wind at Lander, Wyoming, is consistent with
<br />positive vorticity north of the shortwave trough in
<br />Colorado, away from the site of the flood,
<br />Light to moderate' southern and southwesterly
<br />winds advected a deep layer of moisture into Colorado
<br />(Fig, 4) along the spine of the southern Rocky Moun-
<br />tains, Dewpoint depressions at 500 hPa were IOC at
<br />all sounding sites in New Mexico, Colorado, and
<br />Wyoming (Fig, 4), Overall, the main synoptic features
<br />aloft resembled those that accompanied the 1972
<br />Rapid City and 1976 Big Thompson (hereafter referred
<br />to as RC and BT) flash floods (Maddox et al, 1978;
<br />Caracena et al, 1979),
<br />On the night of the flood, storms from Denver to
<br />FCL developed at about the same time that a cluster
<br />of relatively cold cloud tops crossed the northern Front
<br />Range after having traveled north-northeastward from
<br />central New Mexico within the monsoonal flow, A
<br />12-h series of the cluster's footprints, as determined
<br />from Geostationary Operational Environmental Sat.
<br />ellite (GOES-B) infrared imagery, is shown in Fig, 5,
<br />The northernmost footprint in Fig, 5 depicts the pe-
<br />rimeter of these cloud tops at 1800 MDT, just prior to
<br />the onset of flooding rains in FCL and coincident with
<br />strengthening convection located in the Denver area
<br />that shortly thereafter evolved into a bow echo (cf.
<br />sections 5, 6), Whether or not the cold cloud tops were
<br />dynamically linked with the shortwave trough men-
<br />tioned earlier is unclear.
<br />
<br />b, Suiface
<br />At the surface, a cool and exceptionally moist Ca-
<br />nadian air mass was lodged against the eastern face
<br />
<br />2Although the 500-hPa analysis from the National Centers for
<br />Environmental Prediction depicts winds of 2.5-10 m S-l over the
<br />southern Rockies, data from the ACARS indicate winds in the up-
<br />per troposphere approached 23 m S-l just south of Fort Collins at
<br />about 1800 MDT 28 July (Fig, 6),
<br />
<br />195
<br />
|