|
<br />"
<br />
<br />Those signatures included 1) the presence of a nega-
<br />tively tilted 500-hPa ridge over the area and some de-
<br />gree of forcing associated with a weak shortwave
<br />trough that moved northward in the western side of the
<br />ridge; 2) postfrontal moist easterly upslope flow at low
<br />levels; 3) a veering, but weak to moderate, moist,
<br />south-southwesterly flow aloft; 4) slow system move-
<br />ment and training convection; and 5) a deep, moist,
<br />warm layer in the sounding conducive to precipitation
<br />production via warm-rainlcollision--{;oalescence pro-
<br />cesses, However, unlike the 1976 Big Thompson flood
<br />(Caracena et al, 1979), the environment of the Fort
<br />Collins flood possessed only modest thermodynamic
<br />instability and a lower LFC, This situation led to storm
<br />development where the easterly flow encountered its
<br />fIrst abrupt lift on the western side of Fort Collins, as
<br />opposed to the Big Thompson flood, where parcels
<br />required lifting to higher elevations in order to reach
<br />the LFC.
<br />Intensive radar sampling coupled with gauge and
<br />lightning observations provided a detailed view of the
<br />cloud and precipitation morphology, For example, the
<br />heaviest convective rainfall occurred in an area on the
<br />order of 10 x 10 kIn'; the heaviest precipitation cores
<br />were only 1-3 kIn wide. Relatively little lightning and
<br />no hail occurred in association with the storm over Fort
<br />Collins, consistent with the presence of only modest
<br />thermodynamic instability and the tropical nature of the
<br />sounding, Polarimetric radar observations suggest that
<br />a coupling between warm-rain/collision-coalescence
<br />processes and precipitation ice processes played an
<br />important role in the rainfall production associated
<br />with the flood convection, Dual-Doppler observations
<br />and mesoscale wind analyses revealed that the low-
<br />level mesoscale flow fIeld associated with a bow echo
<br />may have caused a brief acceleration in the easterly
<br />wind component at low levels during the last 1,5 h of
<br />the storm, The enhanced easterly flow apparently in-
<br />teracted with convection over Fort Collins, resulting
<br />in quasi-stationary convection and heavy rainfall.
<br />Radar estimates of storm total precipitation were
<br />computed using the NEXRAD Z-R relationship and
<br />both CSU-CHILL and KCYS NEXRAD reflectivity
<br />data, The resultant NEXRAD Z-R estimates of the
<br />maximum rainfall accumulation were approximately
<br />one-half of that measured by the rain gauge network,
<br />Two tropical Z-R relationships were also utilized to
<br />estimate the storm's accumulated rainfall, Application
<br />of the tropical Z-R' s yielded a mixed result (e,g., Table
<br />2), as either lower, equal, or higher rainfall totals
<br />were produced relative to the rain gauge network de-
<br />
<br />Bulletin of the Americon Meteorological Society
<br />
<br />pending on both the radar and the relationship
<br />used, Alternatively, multiparameter radar rainfall es-
<br />timation techniques such as the R(Kop' t;,R) method,
<br />which incorporates more information on the tempo-
<br />ral and spatial variation of the drop size distribution,
<br />provided a reasonable estimate (20% low) of the storm
<br />total precipitation with a minimal amount of tuning,
<br />
<br />9. Topics for future research
<br />
<br />Although sampling of the FCL storms was quite
<br />extensive, several questions and topics for further re-
<br />search have arisen out of the analysis conducted thus
<br />far and include the following,
<br />
<br />1) Radar estimation of rainfall: (a) The NEXRAD
<br />Z-R relationship provided a significant underesti-
<br />mate of the STP maximum regardless of the radar
<br />used, Application of two tropical Z-R relationships
<br />provided either accurate, excessive, or slight un-
<br />derestimates of the STP maximum depending on
<br />the radar and relationship used, Can ancillary me-
<br />teorological data be used in real time to provide
<br />gnidance for the selection of an appropriate Z-R
<br />relationship? If so, over what temporal and spatial
<br />scales might this work? To what extent could real-
<br />time lightning information be used in this process?
<br />(b) While polarimetric techniques such as the
<br />R(Kop' ZOR) and the blended product yielded rea-
<br />sonable estimates of STP for the FCL flood case
<br />with minimal processing, multiparameter tech-
<br />niques such as R(Kop) did not appear to work as
<br />well, What are the sensitivities of each technique
<br />to changes in the assumed drop size distribution,
<br />drop shape, and the spatial variability of rainfall?
<br />How did the method of calculating Kop (e,g"
<br />Hubbert et al. 1993) for this case affect the R(Kop)
<br />and R(Kop, ZOR) precipitation estimates?
<br />2) Physical processes: The coexistence of both mov-
<br />ing (storms to the south of Fort Collins) and quasi-
<br />stationary (the Fort Collins storm) convective sys-
<br />tems in the same synoptic environment presents a
<br />formidable challenge for forecasting, To what ex-
<br />tent did mesoscale processes, topographic effects,
<br />interactions between mesoscale convective systems,
<br />or other factors, influence the quasi-stationary char-
<br />acter of this storm? What factors control whether
<br />flash floods along the east slopes of the Rocky
<br />Mountains occur at the base of the foothills or far-
<br />ther up in the mountains? The strength of the
<br />
<br />213
<br />
|