<br />Normalized dBZ profiles: 7128-Bold; COARE-light; GATE-dash
<br />14
<br />13
<br />12
<br />11
<br />10
<br />9
<br />~ 8
<br />7
<br />i 6
<br />:I: 5
<br />4
<br />3
<br />2
<br />1
<br />o
<br />
<br />
<br />,
<br />":~....,
<br />,
<br />,
<br />,
<br />,
<br />,
<br />,
<br />
<br />o
<br />
<br />0,2
<br />
<br />0,4 0,6 0,8
<br />Norm.aRenectlvity
<br />
<br />1
<br />
<br />FIG, 16, Height profiles of nonnalized radar reflectivity (nor,
<br />malized to profIle maxima) for the FCL case (bold, solid line), and
<br />the mean convective profiles from the TOGA eOARE (light,
<br />solid; DeMott and Rutledge 1998) and GATE (dashed; Szoke and
<br />Zipser 1986) tropical oceanic regimes. Short solid horizontal
<br />(dash) lines indicate the heights of the 00 and -200e temperature
<br />levels for the (tropical oceanic) FeL environment.
<br />
<br />siluated near and jusI above the height of the OOC level.
<br />Reflectivities of 35 dBZ extended to temperatures
<br />colder than -lOoC al this time, with the first detected
<br />CG lightning occurring 15 min later, Cross sections
<br />of 2;,R and Kop in Fig, 17a combined with the cross
<br />section of radial velocity in Fig, 17b suggest that 2-
<br />3-mm diameter raindrops (Zo.'s approaching 2 dB,
<br />with Kop's of 2.5" km-I; e,g" Bringi et ai, 1996) were
<br />being formed between the 2- and 3-km levels (x
<br />= I km, Fig, 17b) in the updraft located on the eastern
<br />edge of the convective core (x = 0.5-1.5), The verti-
<br />cal extent of ZOR values> I dB and Kop values
<br />> 10 km-I also suggests that millimeter-sized raindrops
<br />were being lofted by the updraft to temperatures colder
<br />than OOC, followed by freezing (Bringi et ai, 1996), as
<br />indicated by reflectivities near 50 dBZ in the updraft
<br />and a rapid falloff in the ZOR and Kop above the 4-km
<br />level. To investigate this drop freezing process further,
<br />we calculated the reflectivity-weighted ice fraction (f)
<br />using ZH and ZOR as outlined in Carey and Rutledge
<br />(1996), As shown in Fig, 17c,J;increased from 0,1 to
<br />0.5 just above the 4-km level in the updraft However,
<br />mixed-phase precipitation existed up to at least 5 km
<br />as indicated by the 0,9 ice fraction line (and hence 0,1
<br />rain fraction), This observation is further supported by
<br />the presence of an "LDR cap" of -23 to -24 dB over-
<br />lapping the top of the ZOR core, consistent with drop
<br />
<br />208
<br />
<br />1.2
<br />
<br />freezing in a mixed-phase environment (e,g" Bringi
<br />et al, 1996; Bringi et al, 1997; Jameson et ai, 1996),
<br />The raindrops were lofted through the freezing
<br />level along a northwestward trajectory and as they
<br />froze likely underwent substantial accretional growth
<br />prior to descending on the northwest side of the convec-
<br />tive core, The location of the 57 dBZ reflectivity core
<br />near and just above the 3,8-km level, situated in horizon-
<br />tal gradients of radial velocity, ZOR and Kop' suggests
<br />that the descending ice particles were likely growing
<br />by accretion before they melted, Microphysically, the
<br />observations for the FCL case seem conceptually simi-
<br />lar to that of an "accumulation wne" model of precipi-
<br />tation production, This concept, which involves
<br />efficient precipitation production through a coupling
<br />of wann-rain and ice-particle accretion processes, has
<br />been previously invoked to explain hail growth (e,g"
<br />Sulakvelidze et ai, 1967) and microphysical observa-
<br />tions of rainfall production in tropical cumulonimbi
<br />(e,g" Takahashi 1990; Takahashi and Kuhara 1993),
<br />A similar microphysical process was proposed by
<br />Bringi et al, (1996) in their multiparameter radar study
<br />of a storm that produced heavy rain, hail, and minor
<br />flooding in Fort Collins in June 1992,
<br />A horizontal cross section of Zw ZOR' and Kop at
<br />1.2 km AGL (Fig, 18a) during the time correspond-
<br />ing to Figs, 17a-<: reveals interesting microphysical
<br />structure that is relevant to the radar estimation of rain-
<br />fall (Fig, 18b), Cells of high reflectivity exceeding
<br />50 dBZ were approximately collocated with cores of
<br />elevated 2;,R > 1.5 dB, suggesting the presence of drops
<br />in excess of 3 mm (Herzegh and Jameson 1992),
<br />Comparison of Figs, 17b and 18a reveals that these
<br />cells containing large raindrops, likely resulting from
<br />a collision-<:oalescence process, were located in the
<br />updraft along the leading edge of the convective com-
<br />plex, In contrast, the largest values of Kop (<': 1.50 km-I)
<br />were centered primarily in the downdraft and further-
<br />more were displaced to the northwest of the peak val-
<br />ues of ZH and ZOR in the updraft by 1-4 km, This
<br />juxtaposition between maxima in Kop and Z.lZOR has
<br />important implications for storm microphysics and the
<br />radar estimation of rainfall, as further discussed below,
<br />The corresponding horiwntal cross section of radar-
<br />derived rainfall rate is presented in Fig, 18b, We uti-
<br />lize the NEXRAD Z-R and a blended R(f4,p, 2;,R)/Z-R
<br />algorithm utilizing CSU-CHILL polannetric data (cf,
<br />section 7d), The blended R(Kop, 2;,R) estimate pro-
<br />duced peak rain rates of 110 mm h-I, while the peak
<br />rain rate from the Z-R alone is only 75 mm h-t, In ad-
<br />dition, the region of heavy rain (R> 50 mmh-I) is sig-
<br />
<br />"
<br />
<br />Vol, 80, No, 2, Februory 1999
<br />
|