1. Introduction
The National Aeronautics and Space Administration’s (NASA’s) Tropical Rainfall Measuring Mission (TRMM) was designed to measure and monitor precipitation throughout the tropics using a combination of a Precipitation Radar (PR) and a TRMM Microwave Imager (TMI; Kummerow et al. 1998). Extensive research has been carried out on understanding the distribution and variability of tropical precipitation from climate scales to storm scales based on TRMM observations (e.g., Adler et al. 2000; Cecil et al. 2005). Although the TRMM satellite samples precipitation systems at lower midlatitudes (the TMI coverage extends from ∼38°S to ∼38°N), it is primarily an observing system for the tropics. Consequently, few studies have focused on analyzing the precipitation structure of midlatitude winter storms using TRMM observations. This paper investigates the distribution of precipitation within a midlatitude frontal system that was observed by three consecutive overpasses of the TRMM satellite on 19 February 2001, during the Pacific Coastal Jets Experiment (PACJET).
The organization of precipitation within extratropical cyclones has been extensively studied using in situ and remotely sensed observations and numerical simulations (e.g., Browning and Harrold 1969; Houze et al. 1976; Matejka et al. 1980; Hobbs and Persson 1982; Braun et al. 1997; Jorgensen et al. 2003). In their studies of 11 occluded cyclones in the Pacific Northwest, Houze et al. (1976) classified six categories of mesoscale rainbands according to their location and morphology. The narrow cold-frontal rainband (NCFR; ∼5 km in width) is typically observed along the sharp gradient associated with the surface (or low level) cold front. It behaves much like a density current with the vertical circulation characterized by an approximate balance between the baroclinically generated vorticity of the cold front and the ambient vertical wind shear (Rotunno et al. 1988; Parsons 1992; Jorgensen et al. 2003). It usually produces intense convective rainfall (∼10–50 mm h−1) and occasional severe weather (Carbone 1982). Updrafts associated with NCFRs can be intense (up to 18 m s−1 as reported by Carbone 1982) and are often nearly vertical. Despite the vigorous strength of the rising motion in NCFRs, the rainbands usually only reach an altitude of approximately 3.5–5 km (e.g., Houze et al. 1976; Hobbs and Persson 1982; Carbone 1982; Jorgensen et al. 2003). Many studies (e.g., Hobbs and Persson 1982; Jorgensen et al. 2003) have focused on variations of the substructure shown in radar reflectivity fields of NCFRs (i.e., “precipitation cores” and “gap regions”) and have related these features to horizontal shear instability. Jorgensen et al. (2003) investigated the variation of the vertical circulation along a precipitation core in the 19 February storm. They found that variations in the strength of the prefrontal vertical wind shear led to either upshear-, erect-, or downshear-oriented updrafts, which resulted in alongfront variations of the precipitation structure along the precipitation core.
The wide cold-frontal rainband (WCFR; ∼50 km in width) is a region of enhanced stratiform precipitation. The upward motion of the WCFR is aloft, characterized by enhanced mean ascent (Houze 1993, 481–484), but also including shallow convective generating cells that contribute to precipitation growth via a feeder–seeder process. WCFRs are also often observed without NCFRs in the coastal region of the Pacific Northwest (Evans et al. 2005; Bond et al. 2005). Because the ascent occurs at upper levels, the WCFR can move at speeds that differ from the NCFR. The WCFR typically forms behind the low-level NCFR but frequently moves faster, so that it often overtakes the NCFR, temporarily straddling it for a period of time, before moving ahead of the low-level front. Locatelli et al. (1994) found that the precipitation distribution, intensity, and movement of a WCFR were closely related to the topography of the frontal surface.
This paper investigates the distribution of precipitation associated with an intense midlatitude cold front observed by the TRMM satellite over the eastern Pacific Ocean. The TMI brightness temperatures suggest significant alongfront mesoscale to synoptic-scale variability of the precipitation structure. Compared to the northern and southern ends of the frontal rainband,1 a greater concentration of precipitation ice was present along the central portion of the front where the front bowed out. This structure suggests that the forcing mechanisms for ascent along the frontal rainband varied along the band. The main objective of this study is to document and explain such variability, which occurs at a scale much larger than the focus of many previous studies [e.g., the small-scale cores and gaps within the NCFRs studied by Hobbs and Persson (1982) and by Jorgensen et al. (2003)].
A numerical simulation is conducted to examine the microphysical, thermodynamic, and dynamic variations related to the precipitation distribution in the frontal rainband. The synoptic setting and the structure of the rainband observed by TRMM are described in section 2. Section 3 summarizes the numerical model configuration, the evolution of the simulated synoptic environment and rainbands, and comparisons to observations. Using the simulation results, section 4 provides an analysis of the mesoscale to synoptic-scale alongfront variability of the precipitation including the microphysical and kinematic structures of the rainband. This section also examines the factors responsible for this variability, including the upper-level and surface fronts, a stratospheric potential vorticity (PV) intrusion, and related quasigeostrophic forcing of vertical motion.
2. Observations
An intense frontal precipitation system approached the coast of California on 19 February 2001, during the National Oceanic and Atmospheric Administration’s (NOAA’s) PACJET experiment in January–February 2001 (background information on the PACJET experiment can be found online at http://www.esrl.noaa.gov/psd/programs/2001/pacjet/pacjet.html). The synoptic environment (Fig. 1a) is characterized by a cutoff low at 500 hPa over the eastern Pacific Ocean, with high absolute vorticity within the trough. A strong upper-tropospheric jet streak was progressing around the base of the 300-hPa trough (Fig. 1b), while at the surface (Fig. 1c), an occluding cyclone with a strong surface cold front was present. A Geostationary Operational Environmental Satellite (GOES) satellite visible image (Fig. 1d) shows a tightly wrapped cyclone moving eastward over the eastern Pacific Ocean near 40°N. The bow-shaped cloud band near 130°W and between 30° and 40°N was associated with the intense surface cold front. The TRMM satellite coincidentally flew over the same precipitation system at a time nearly concurrent with the GOES image (see Fig. 1d), providing detailed observations of the precipitation structure of the rainband.
The TRMM PR observed the cold-frontal rainband during three consecutive overpasses (Figs. 2a–c) at approximately 0023, 0200, and 0337 UTC 19 February 2001. Because the swath width of the PR is only 247 km, the PR captured only small segments of the rainband. At 0023 UTC, the PR showed that the central to southern portion of the rainband (just southward of the apex of the front) consisted of a leading intense (up to 50 dBZ) line associated with a NCFR and a trailing region of weaker stratiform precipitation (∼30 dBZ); namely, a WCFR, that generally paralleled the leading NCFR. South of 32.7°N, the intense precipitation in the NCFR weakened. At 0200 UTC, the PR observed the central, or apex, portion of the frontal rainband. Precipitation cores and gaps were readily apparent within the NCFR. Unlike the segment of the front farther south, there was no obvious precipitation minimum between the intense line and the stratiform precipitation. At 0337 UTC, the PR viewed a portion of the NCFR well south of the apex, which was characterized by a very thin and much weaker line of reflectivity, followed by some scattered light precipitation. When the three PR plots are considered together, they suggest that precipitation was stronger in the central segment than in the more southern segment. The precipitation minimum that separated the NCFR and the stratiform precipitation was most obvious as the precipitation intensity decreased toward the south.
The TMI polarization-corrected brightness temperatures at 19, 37, and 85 GHz are used to show the structure of the frontal rainband and its evolution (Figs. 2d–l) over a broader region. The 19-GHz brightness temperature measures emission of microwave energy due to rain; the 85-GHz channel measures the scattering of energy by large ice particles (i.e., the brightness temperature depression associated with precipitation ice such as graupel and snow); while the 37-GHz channel brightness temperature detects both precipitation ice and liquid (Spencer et al. 1983, 1989; Cecil and Zipser 2002). Since the background water body also contributes to the brightness temperature depression because of its low emissivity, the polarization-corrected temperature (PCT) is computed to discriminate between the precipitation and the background ocean surface. The PCTs at channels 37 and 85 GHz are calculated based on the formulas of Spencer et al. (1989) and Cecil et al. (2002). The 19-GHz PCT is estimated using a similar approach, but with an appropriate coefficient2 to differentiate the precipitation from the background ocean surface.
With a much wider swath (878 km) than the PR, the TMI was able to observe much of the entire rainband (Figs. 2d–l). Because the horizontal resolution of TMI’s lower-frequency channels is ∼20 km, the TMI observations were generally not able to clearly differentiate the NCFR and WCFR. However, the elongated signature of the NCFR was still apparent in the southern portion of most of the TMI plots. At 0023 UTC (Fig. 2d), the structure of the 19-GHz PCT within the PR swath region shows good agreement with the PR observations; for example, the area of stratiform precipitation and the NCFR observed by PR near 33.5°N, −128.5°W (Fig. 2a) appear as a “hatchet”-shaped PCT depression outlined by the 285-K contour in Fig. 2d. Extending the view farther northward and southward from the PR swath, one can infer the broader structure and intensity of the frontal rainband. The frontal rainband formed an arc-shaped precipitation feature that bowed out to the east near 34.5°N. The PCTs were lowest north of the apex and quickly diminished as the front trailed to the southwest. This implies that the heaviest rainfall was near and northward of the apex with much less precipitation to the southwest, consistent with the observed reflectivity structure. With stratiform precipitation diminishing southward, the less intense NCFR became a very distinct feature. At 0200 UTC, TRMM obtained its best view of the major precipitation area along the front with the PR sampling the most intense portion along the bowed out section of the front (Figs. 2b,e). The precipitation in the northern section decreased gradually away from the apex, while it decreased more rapidly to the south. In the final overpass at 0337 UTC (Fig. 2f), the TMI scanned through the southern portion of the frontal rainband and captured the weakening NCFR and the diminishing stratiform precipitation.
The 37-GHz PCT depressions shown in Figs. 2g–i illustrate a structure and evolution of the rainband similar to that seen at 19 GHz. However, in addition to detecting precipitation liquid, the 37-GHz channel adds information about precipitation ice. It suggests that the total precipitation ice and liquid was strongest in the bowed section of the front, somewhat weaker to the north, and much weaker to the south.
The 85-GHz channel is predominantly affected by the distribution of precipitation ice associated with the frontal rainband. The 85-GHz PCTs in Figs. 2j–l clearly show that the largest ice concentrations were in the bowed out section of the front (e.g., approximately 33°–37°N at 0200 UTC), with the lowest PCT values within or slightly westward of the precipitation cores in the NCFR. Very little ice was present along the southern portion of the front.
The distribution of the precipitation ice and liquid along the frontal rainband observed by TRMM indicates that the kinematics and dynamics of the cold front vary in the alongfront direction. As described in the following section, a numerical simulation of this event has been carried out in order to examine the precipitation structure and microphysical variability along the cold-frontal rainband and to study the kinematic and dynamic features that contribute to this variability.
3. Numerical model description and simulated synoptic environment and rainband
a. Numerical model configuration
The fifth-generation Pennsylvania State University–National Center for Atmospheric Research (PSU–NCAR) Mesoscale Model (MM5), version 3.6, is used to conduct a 48-h simulation of the cold-frontal rainband from 0000 UTC 18 February to 0000 UTC 20 February 2001. The model grids consist of four two-way interactive nested domains (Fig. 3). The outermost domain contains 100 × 85 grid points in the x and y directions with a grid spacing of 45 km. The nested domains contain 148 × 121, 199 × 199, and 448 × 283 grid points with grid spacings of 15, 5, and 1.7 km, respectively. Fifty-one vertical sigma levels are used in each domain with 22 levels located below 850 hPa. The National Centers for Environmental Prediction Aviation (AVN) model analyses with 1° horizontal and 6-h temporal resolution are used to provide the initial and boundary conditions. The outer three domains are initialized at 0000 UTC 18 February and run for 48 h, whereas the innermost domain is initialized at 1700 UTC 18 February and run for 22 h.
Cloud processes in the model include the Grell cumulus parameterization (Grell et al. 1995) on the 45- and 15-km domains to account for unresolved convective processes and the Goddard three-ice phase (i.e., cloud ice, snow, and graupel) explicit microphysics parameterization on all grids (Tao and Simpson 1993; Braun and Tao 2000). Boundary layer processes are represented by the Eta parameterization, a Mellor–Yamada scheme developed by Janjic (1990, 1994). Longwave and shortwave radiative processes are based on the cloud–radiation scheme of Dudhia (1989).
Since this paper focuses on the mesoscale to synoptic-scale features of the precipitation distribution along the cold front, not the smaller-scale precipitation cores and gaps (e.g., Jorgensen et al. 2003), only data from the outer three domains are used to examine the structure and evolution of the cold-frontal rainband. A detailed comparison between the simulation and aircraft observations using the fourth domain will be presented in a future paper (O. Persson 2008, unpublished manuscript).
b. Simulated synoptic environment
The synoptic evolution over the 12-h period centered on the times of the TRMM observations, from 2000 UTC 18 February to 0800 UTC 19 February, shows that the cold front was associated with an occluding cyclone. The simulated sea level pressure, equivalent potential temperature (θe) and wind barbs at 975 hPa are shown in Fig. 4. At 2000 UTC 18 February (Fig. 4a), the cyclone central pressure was 985 hPa. Cold-air advection occurring southeast of the cyclone center undercut a tongue of relatively warm and moist air (>306 K) to the east of the low, indicating the onset of the occlusion process. An intense surface cold front can be identified from the tight gradient of θe and the corresponding sharp changes in the pressure and wind fields. The resolvable θe gradient on domain 1 (45-km horizontal resolution) associated with this front was ∼12°C (100 km)−1, whereas the gradient was much tighter, ∼10°C (10 km)−1, in domain 3 (5 km; figure not shown). At 0200 UTC 19 February (Fig. 4b), the central pressure remained constant while a pocket of warm air wrapped around to the south of the low. The tongue of warm and moist air (>306 K) narrowed as the occlusion process continued. At 0800 UTC 19 February (Fig. 4c), the cyclone central pressure rose by 3 hPa. By this time, the warm moist tongue (>306 K) had been pinched off and lifted above the 975-hPa level.
The evolution of the upper-tropospheric flow is displayed in Fig. 5, which shows a developing cutoff low associated with a negatively tilted trough. The low centers from the surface (see Fig. 4) to 300 hPa were nearly vertically aligned from 2000 UTC 18 February to 0800 UTC 19 February. Confluent flow occurred upstream of, or near, the trough axis with a diffluent ridge located downstream. During the 12-h period, a strong jet maximum, initially located upstream of the trough with maximum wind speeds above 60 m s−1, progressed downstream through the trough axis, with wind speeds decreasing to between 50 and 60 m s−1 by the end of the 12-h period. The jet maximum was associated with an upper-level baroclinic zone (discussed in the next section) and is thus referred to as an “upper-level jet-front system” (e.g., Keyser and Shapiro 1986; Schultz and Doswell 1999).
The upper-level synoptic environment is also depicted through the distribution of potential vorticity (PV) on the 320-K potential temperature (θ) surface (Fig. 6). A pool of high PV (>4 PVU) air was located behind the surface cold front (indicated by the 306-K θe contour). During the 12-h period, it rotated cyclonically and progressed toward the east-northeast. This positive PV anomaly was connected with a region of high PV at higher latitudes. The connecting strip of high PV gradually narrowed and formed a “treble clef” shape of high PV (Figs. 6b,c), an indicator of the occlusion process (Martin 1998). The PV anomaly also suggests a possible tropopause folding event in that region, with an upper-level front capable of providing enhanced ascent on the eastern side of the PV anomaly. The potential role of such ascent in determining the alongfront variability of precipitation is examined in section 4b.
Comparison of Figs. 4b and 5b to the NCEP analyses in Figs. 1c and 1b, respectively, shows that the synoptic-scale evolution in the simulation is in good general agreement with the analyses. The simulated sea level pressure is somewhat lower than analyzed, but the position of the low is quite good. Although the simulated θe values are slightly larger than analyzed, particularly south of 30°N, the overall θe pattern is in good agreement with the NCEP analyses. The simulated 300-hPa height and winds are also in very good agreement with the analyses.
c. Simulated rainband structure and model validation
The simulated radar reflectivity is calculated based on the hydrometeor mixing ratios (Fovell and Ogura 1988; Braun 2006).3 Figure 7 shows the structure and evolution of the cold-frontal rainband over a 10-h period from domain 3 (5 km). Comparison of the simulated reflectivity (Figs. 7a–c) with the TRMM-observed radar reflectivity and the PCT fields (Fig. 2) shows that the simulation reasonably captures the general structure and evolution of the rainband. The simulation replicates the bow shape of the whole frontal rainband and successfully differentiates the NCFR and the trailing stratiform precipitation. The stratiform precipitation in the central to northern portion is stronger than along the southern tail of the band. At later times (Figs. 7d–f), the simulation shows that the rainband weakens as it moves northeastward toward the coast of central California.
Comparison between the observed and simulated radar reflectivity at 0200 UTC (hour 26 of the simulation) shows that the simulated rainband (Fig. 7b) lags the observed band (Fig. 2b) by ∼100 km. The simulated transition in structure from a more continuous line to a series of precipitation cores and gaps in the NCFR near 34°N is very similar to the observations (Figs. 2a, b). The simulation also shows a fairly persistent region of stratiform precipitation between 32.5° and 38°N trailing the NCFR, similar to that observed by the PR.
The simulated synoptic evolution and mesoscale precipitation structure have been shown to agree well with NCEP analyses and TRMM observations. A detailed verification of the thermodynamic structure will be given in a future paper by O. Persson. Therefore, the consistency between the model simulation and the observations suggests that the MM5 simulation has sufficiently replicated the structure and evolution of the rainband, as well as its synoptic and thermodynamic environment. In the next section, the simulation is used to examine the dynamics and thermodynamics of this cold-frontal rainband and the mechanisms for the alongfront mesoscale variability of precipitation.
4. Alongfront precipitation variability and the processes responsible for it
In this section, two times, 0200 and 0800 UTC, are chosen to further illustrate the detailed microphysical and kinematic structures of the rainband and to determine the underlying dynamical processes impacting the alongfront variability.
a. Microphysical and kinematic structures of the rainband
There are three categories of precipitating hydrometeors (rain, graupel, and snow) in the Goddard microphysics scheme used in this simulation; of which graupel and snow are highly effective scatterers at higher frequencies (e.g., Adler et al. 1991; McFarquhar et al. 2006; Biggerstaff et al. 2006). In this simulation, the total precipitation ice is dominated by graupel, and thus the distribution of the simulated column-integrated graupel (Fig. 8) is chosen for comparison with the TRMM 85-GHz PCT fields. At 0200 UTC (Fig. 8a), the simulation shows peak concentrations (approximately 3.5–4.5 mm) of graupel within portions of the NCFR, qualitatively similar to the distribution of maximum PCT depression seen in the TRMM data (Fig. 2k). Between 33° and 37°N, two regions of enhanced graupel concentrations associated with the WCFR are seen in the simulation, one between 33° and 35°N and another between 35.5° and 37°N. In both regions, smaller concentrations of graupel are found in the WCFR to the northwest of the peak graupel amounts within the NCFR. In general, the observed PCT structure and the simulated graupel distribution both show that precipitating ice is most prevalent along the central part of the front. Taken with the results shown in Fig. 7, the results suggest that while the NCFR extends along much of the length of the front, the stratiform WCFR is concentrated more along the central part of the front as a result of the generation of significant precipitation ice there. By 0800 UTC, the graupel amount associated with the rainband decreases significantly as the band moves northeastward and approaches the coast of central California.
The vertical distributions of the three precipitation hydrometeors, as well as the thermodynamic characteristics of the front, are examined in cross sections AA′ and BB′ for 0200 UTC and CC′ and DD′ for 0800 UTC (see Figs. 7 and 8 for their locations). Sections AA′ and CC′ pass through the portion of the rainband where the precipitation is generally most intense at 0200 and 0800 UTC, respectively. Sections BB′ and DD′ cut through the southern portion of the rainband where the precipitation is generally weaker. Data in the cross sections depict section-normal average quantities derived by averaging over 100 km (50 km to each side of the cross section) in order to show representative features of the mesoscale structure.
At 0200 UTC, two regions of intense potential temperature, θ, gradient are seen in cross section AA′ (Fig. 9a). The first is near x = 370 km, extending vertically from the surface to ∼800 hPa, and is the low-level cold front associated with the NCFR. The second intense θ-gradient region, located near x = 200–275 km from approximately 700 to 450 hPa, is an upper-level baroclinic zone. The upper-level front is separated from the low-level cold front by a layer of more moderate θ gradient. The observed thermodynamic field also shows an enhanced θ gradient at upper and lower levels, separated by a layer of weaker gradient (O. Persson 2008, unpublished manuscript). The maximum value of rainwater mixing ratio is above 0.5 g kg−1 and is collocated with the surface cold front, thus identifying it as the NCFR. Just above the rain maximum, at midlevels (from approximately 800 to 650 hPa), there is a narrow zone of high concentrations of graupel and snow (greater than 0.5 and 0.1 g kg−1, respectively), indicating that the NCFR only reaches ∼650 hPa at this time. A secondary maximum of rain near x = 320 km is associated with even higher graupel and snow concentrations, 0.7 and 0.3 g kg−1, respectively, and corresponds to the WCFR. The graupel and snow maxima are located just ahead of the upper-level front, with smaller amounts within or underneath the upper-level front, suggesting that the upper-level front plays a role in the formation of the WCFR. In contrast to the NCFR, the precipitation hydrometeors associated with the WCFR show a much deeper vertical extent, reaching 400 hPa at this time.
In the southern portion of the rainband (cross section BB′, Fig. 9b), the low-level front, although slightly weaker than its northern counterpart, is still prominent with a strong θ gradient extending up to ∼800 hPa. The upper-level front is evident in the relatively flatter θ contours, with smaller values of the horizontal θ gradient (not shown) but increased stability between approximately 700 and 500 hPa. However, the upper-level front in this portion of the rainband is not dynamically active (thus, its leading edge is not indicated in the figure), as will be discussed later. The vertical distribution of hydrometeors varies dramatically from cross section AA′ to BB′, with graupel and snow hydrometeors nearly absent in BB′. The trailing stratiform precipitation is also nearly absent, with only a small postfrontal secondary rainwater maximum occurring. However, the rainwater concentration in the NCFR in BB′ is as large as its northern counterpart.
By 0800 UTC (Figs. 9c,d), the width of the rainband as a whole decreases as it approaches the California coast. The concentrations of the three hydrometeors in the northern portion of the band (CC′) are still large with peak concentrations of rainwater and graupel greater than 0.5 and 0.7 g kg−1, respectively. The NCFR no longer leads the stratiform precipitation as lighter precipitation is now found ahead of the NCFR. The variation of the distribution of the precipitation hydrometeors from the northern cross section (CC′) to the southern cross section (DD′) is similar to that seen at 0200 UTC. Comparison of the θ field between the two cross sections shows that the intensity of the surface cold front is comparable between the northern and southern portions at this time. In the northern cross sections, the distance between the upper- and lower-level fronts has decreased during the time interval between 0200 and 0800 UTC, indicating a faster movement of the upper-level front compared to the lower-level front. The upper-level front to the south (DD′) remains dynamically inactive, consistent with results seen at 0200 UTC.
An alternative view of the structure of the frontal rainband is provided through the equivalent potential temperature (θe) and the vertical motion (ω) fields (Fig. 10). The θe contours associated with the low-level narrow cold front (at x = ∼370 km) are oriented nearly vertically at 0200 UTC in cross section AA′. The associated θe gradient is extremely sharp (i.e., ∼10 K) over a horizontal distance of 10 km at 1000 hPa from domain 3 of the simulation (5-km horizontal resolution). The magnitude of this gradient is particularly notable given that the fields have been averaged over 100 km (50 km to each side of the cross section). The local θe gradient without averaging is even sharper, ∼16 K (10 km)−1. Although the θe gradient associated with the upper-level front is not as sharp as that of the low-level cold front, it is still quite intense in the layer from approximately 700 to 450 hPa. The two fronts are separated from each other by a layer of weaker θe gradient, suggesting that the upper- and the lower-level cold fronts are distinct entities.
The rising motion associated with the NCFR is intense, with maximum rising motion stronger than −120 dPa s−1 at ∼850 hPa in the averaged fields (Fig. 10a). The horizontal distribution of ω at 850 hPa (not shown) reveals that, without averaging, the maximum upward motion associated with the NCFR is as large as −200 to −300 dPa s−1 from domain 3 of the simulation. The finest domain (1.7-km horizontal resolution) simulation resolves even more intense upward motion, −600 to −800 dPa s−1 (not shown). The magnitude of the simulated rising motion is consistent with the pseudo-dual-Doppler analyses (presented by Jorgensen et al. 2003, approximately 5–10 m s−1) for the same case. A broad region of weaker elevated ascent (x = 250–350 km, 750–400 hPa) occurs behind the NCFR and is collocated with the significant snow and graupel concentrations of the WCFR.
In the southern portion of the rainband (BB′), the θe gradient associated with the low-level cold front is as intense as its northern counterpart. The average vertical motion associated with the low-level cold front is greater than −140 dPa s−1 in BB′ (Fig. 10b), while only very weak ascent occurs above 700 hPa. An interesting feature is the existence of a localized region of high θe gradient near x = 240 km that is accompanied by moderate ascent and light precipitation, and is apparently associated with a postfrontal shower (Fig. 7b).
At 0800 UTC, in the northern portion of the rainband (CC′, Fig. 10c), the low-level cold front is oriented vertically up to ∼700 hPa, with a somewhat weaker θe gradient compared to that at 0200 UTC. The high θe-gradient layer between ∼650 and 450 hPa near x = 200–250 km corresponds to the upper-level front, as seen in the θ field (see Fig. 9c). Although weaker, the updrafts associated with the lower- and upper-level fronts appear to be more vertically aligned than that seen earlier, as a result of the faster movement of the upper-level front relative to the low-level front. Consequently, the ascent responsible for the NCFR extends to a much higher altitude. In addition, a region of elevated ascent (near x = 330 km) possibly associated with the upper-level front moves ahead of the low-level front and is responsible for the light stratiform precipitation ahead of the NCFR (see Fig. 9c). In the southern portion of the rainband (DD′; Fig. 10d), the frontal structure is similar to that at the earlier time. The intensity of the θe gradient and the associated updraft are slightly decreased compared to 0200 UTC; however, they are both somewhat stronger than their northern counterparts at this time.
The distribution and evolution of the microphysical and kinematic properties of the rainband demonstrate that 1) the intense and shallow low-level cold front extends along much of the length of the arc-shaped cloud shield (see Fig. 1d) and 2) the upper-level cold front is a separate feature from the low-level front, with the former mainly active along the central to northern portion of the cloud system. The intense updraft associated with the low-level front is primarily responsible for the NCFR. The updraft associated with the upper-level front appears to be responsible for the stratiform precipitation, or WCFR, following or straddling the NCFR. In the areas where the lower- and upper-level cold fronts coexist, the ascent is deeper and stronger. This enhanced ascent supports greater condensate production, generally leading to the formation of more snow and at least small amounts of supercooled liquid water (not shown), which eventually results in greater production of graupel through accretion processes. In the next section, the dynamical mechanism responsible for the ascent ahead of the upper cold front is examined more closely.
b. Potential vorticity structure and quasigeostrophic forcing
In this subsection, the dynamical role of the upper-level front is investigated using model output from domain 2 (15 km). The upper-level baroclinic zone marks the leading edge of an intrusion of stratospheric high-PV air. Figure 11 illustrates the potential vorticity and wind barbs at 300 hPa in a zoomed-in view of the structure of the treble clef–shaped PV anomaly shown in domain 1 (Fig. 6). The leading edge of the high-PV air is located behind the central to northern portion of the low-level cold front at 0200 UTC when the warm, moist surface air is confined to a narrow zone along a rapidly occluding front. Near the central portion of the low-level front (the bowed out region), strong southwesterly winds produce positive PV advection at 300 hPa (Fig. 11a). By 0800 UTC, this area of positive PV advection moves eastward over the triple point of the surface fronts.
Figure 12 shows the distribution of Q vectors, ωQG, and geopotential height at 500 hPa. The geostrophic flow is generally confluent upstream and diffluent downstream of the trough axis. At 0200 UTC, Q-vector convergence occurs near or immediately downstream of the trough axis where a broad area with strong QG rising motion (<−6 dPa s−1) is induced. The location along the frontal rainband of the enhanced QG upward motion (<−12 dPa s−1) agrees reasonably well with the distribution of the column-integrated graupel (Fig. 8). Significant graupel concentrations are generally confined to the region between approximately 33° and 38°N, where enhanced QG upward motion is found. In contrast, QG upward motion is much weaker (or nonexistent) along the southern part of the front, in agreement with the lack of graupel and snow to the south (see Figs. 8 and 9b,d). This consistency between QG rising motion and precipitation ice concentration continues as the trough progresses northeastward by 0800 UTC (Fig. 12b).
While the locations of the maximum 500-hPa ωQG and column-integrated graupel concentrations along the frontal rainband are in good agreement, their locations in the east–west direction are not exactly collocated. The maximum QG updraft is about 1° longitude west of the graupel maximum, although a weaker region of QG ascent is collocated with the graupel maximum. Such a difference in the correspondence of the locations of the two fields is not surprising given that the major production of graupel occurs below 500 hPa in the layer between approximately 850 and 650 hPa (Fig. 9a). Results for 700 hPa (figure not shown, but refer to Fig. 14) indicate a better correspondence in the locations of the maximum QG ascent and vertically integrated graupel concentration. The enhanced QG ascent at 500 hPa likely does not directly drive the production of graupel, but instead contributes to more graupel indirectly by enhancing snow production aloft, with this snow being accreted by graupel at lower levels. This conclusion is consistent with the analysis in section 4a, where vertical cross sections of the hydrometeor and updraft distributions were compared. Overall, the good correspondence between the region of large precipitation ice concentration and the enhanced QG ascent suggests that the QG forcing associated with the upper-level front played a significant role in determining the mesoscale to synoptic-scale variation of the precipitation ice in the simulation and as observed by TRMM.
The vertical distribution of PV, θ, and vertical motion from domain 2 are investigated in four cross sections: LA–LA′, LB–LB′, LC–LC′, and LD–LD′ to further illustrate the dynamical role of the fronts on the distribution of precipitation ice indicated by the TRMM observations. The cross sections (see Fig. 11 or 12 for their locations) contain, but are longer than, the cross sections (AA′, BB′, CC′, and DD′) from domain 3. Arrows in Figs. 13 and 14 indicate the locations of the corresponding shorter cross sections (Figs. 9 and 10). Similar to the earlier cross sections, the magnitudes of all the fields in the cross sections have been averaged in the direction normal to the cross section, in this case over 150 km (75 km to each side of the cross section).
Figure 13 shows the vertical distribution of PV and rising motion (ω) while Fig. 14 gives the potential temperature and quasigeostrophically forced rising motion (ωQG). The lower- and upper-level fronts are readily seen in the θ field, at the leading edges of the high θ gradient regions in Fig. 14 (see also Fig. 9). At 0200 UTC, the northern part of the rainband (LA–LA′, Fig. 13a) is characterized by a dramatic vertical displacement of the stratospheric PV anomaly, with the high-PV (>2 PVU) air extending into the troposphere to below 450 hPa. At its leading edge (x = 675 km), a particularly strong intrusion of high-PV air, as evidenced by the 4-PVU contour, locally reaches below 500 hPa. A strong gradient of θ associated with the upper-level front begins in this filament of high-PV air and extends downward to ∼700 hPa (Fig. 14a). Strong rising motion (peak of −42.7 dPa s−1, Fig. 13a) at the upper levels is located immediately ahead of the PV anomaly.
The QG lifting (Fig. 14a) is weaker and occurs over a wider region than the simulated upward motion (Fig. 13a), largely as a result of the smoothing needed for calculating ωQG, the neglect of diabatic effects, and limitations of QG theory (e.g., Stoelinga et al. 2000; Locatelli et al. 2002; Han et al. 2007). Two regions of stronger ωQG are seen within the broader region of ascent. The first is located within the upper frontal zone near 400 hPa and corresponds to an area of weak ascent behind the PV filament (x = ∼600 km; Fig. 13a). The second occurs ahead of the upper-level front, above and slightly rearward of the low-level font, and approximately coincides with the region of simulated mid- to upper-level ascent in Fig. 13a. It also coincides with the region of enhanced snow and graupel production associated with the WCFR in Fig. 9a. Below 700 hPa, the updraft associated with the NCFR is more intense (peak of −53.6 dPa s−1; Fig. 13a) and is narrower than the updraft at upper levels. The absence of this strong low-level updraft in the QG-derived vertical motion (Fig. 14a) may be because the NCFR is behaving like a density current (Carbone 1982; Parsons et al. 1987; Parsons 1992; Houze 1993), which occurs at a scale too small for QG dynamics to be valid.
To the south, cross section LB–LB′ cuts through the PV anomaly from near its center to its southern flank (see Fig. 11a). Figure 13b shows a narrow tail of high PV stretching down to ∼600 hPa. Although the upper PV intrusion is prominent, the advection of PV (Fig. 11a) and the convergence of the Q vectors (Fig. 12a) are weak, leading to little QG forcing of ascent (Fig. 14b). As a result, the vertical motions at upper levels are much weaker along this part of the front (Fig. 13b) compared to its northern counterpart. In contrast, the low-level upward motion remains intense in LB–LB′, comparable to its counterpart in LA–LA′. It appears that the lack of mid- to upper-level QG ascent prevents the formation of appreciable precipitation ice so that precipitation is limited to the low-level NCFR (Fig. 9b).
Following the movement of the rainband, the northern cross section (LC–LC′) at 0800 UTC cuts through the maximum QG forcing. The PV, the upper- and lower-level fronts, and the upward motion (Figs. 13c and 14c) are similar in structure to that seen at the earlier time. Ascent in the NCFR consists of two centers of rising motion at low and midlevels that are somewhat weaker than at 0200 UTC, but lead to a deeper layer of ascent there (Fig. 13c), consistent with the structure seen in Fig. 10c. This deeper feature is attributed to the fact that the upper PV anomaly is moving faster than the low-level front (the distance separating the upper PV intrusion and the NCFR ascent decreases from ∼200 to ∼130 km), so that the upper-level QG forcing of ascent is more aligned with the lift associated with the low-level frontal density current. The QG ascent associated with the upper-level front is maximized near 500 hPa (Fig. 14c). A considerable portion of the ascent extends well ahead of the low-level front, which provides a possible explanation for the elevated ascent seen ahead of the NCFR in Fig. 10c. Similar to that seen in LB–LB′, the upper-level front in LD–LD′ (Fig. 14d) is dynamically inactive, while the updraft associated with the low-level front remains prominent (Fig. 13d).
The diagnosed QG ascent is too weak to support the growth of graupel due to the limitations described above, and this weaker ascent might remain even without these limitations. Even if the QG ascent is too weak to produce graupel, it may assist in the release of instabilities, such as conditional symmetric instability (CSI), which might enhance ascent on smaller scales to magnitudes capable of producing graupel. In a future paper (O. Persson 2008, unpublished manuscript) evidence will be shown from pseudo-dual-Doppler analyses and in situ measurements of more concentrated ascent on smaller scales within several updraft cores, with mean (instantaneous) ascent up to 1.2 m s−1 (3 m s−1) and a spacing of about 17–20 km, occurring within the broad region of QG ascent identified in this study. These updraft cores appear to be associated with generating cells near the top of the WCFR. This region is characterized by negative moist potential vorticity and is therefore susceptible to CSI or frontal circulations associated with small viscous moist symmetric stability (Bennetts and Hoskins 1979; Xu 1989a,b).
The analysis of the PV and QG forcing suggests that the upper-level front is generally responsible for the upward motions at higher altitudes, while the low-level front forces intense ascent below ∼700 hPa. In the central to northern portion of the front, where the two forcings combine, the enhanced upward motion intensifies the snow and graupel growth and contributes to high concentrations of precipitation ice. The upper-level forcing makes a large contribution to the stratiform precipitation (WCFR) that initially follows and later precedes the NCFR. Lacking significant upper-level forcing along the southern portion of the NCFR, the narrow precipitation band is generally shallow and consists of rain only.
5. Conclusions
Three consecutive overpasses of the TRMM satellite observed an intense midlatitude frontal rainband over the eastern Pacific Ocean on 19 February 2001. The PR captured the detailed structure of the rainband, including an intense leading line (i.e., NCFR) and a trailing region of enhanced stratiform precipitation (i.e., WCFR). The TMI PCT distribution at three different frequencies suggested that the precipitation was generally most intense along the central to northern portion of the rainband. The PCT fields at 85 GHz specifically indicated that the central portion of the rainband had enhanced precipitation ice concentrations compared to areas farther north and south. These mesoscale to synoptic-scale variations in the distribution of precipitation along the front were investigated using a finescale MM5 simulation. The structure of the simulated rainband agreed well with that observed by TRMM. In particular, the simulated distribution of precipitation ice was highly consistent with that inferred from the TMI 85-GHz data.
The dynamic mechanisms responsible for the alongfront precipitation variability were examined using the multiscale information from the different model grids. At fine scales, the simulation showed that ascent along the low-level NCFR was vigorous, but shallow, generally extending no higher than ∼700 hPa, but occurring along much of the length of the front. Along the central portion of the front, ascent occurred ahead of an upper-level cold front that marked the leading edge of a stratospheric PV intrusion. This ascent was generally weaker, but more widespread than that associated with the NCFR and contributed to the growth of the region of stratiform precipitation described herein as a WCFR. This area of ascent moved eastward somewhat faster than the NCFR so that by the end of the simulation period, the stratiform precipitation in the WCFR began to advance ahead of the NCFR. In the central portion of the front, the synergy of the lifting along the NCFR with the more limited (in the north and south directions) lifting along the upper cold front produced a region of enhanced and deep upward motion that favored the production of snow and graupel, while regions farther northward and southward, lacking the upper-level ascent, generally produced little snow and graupel.
Synoptic-scale information from the coarser grids indicated that the rainbands were part of an occluding cyclone associated with a negatively tilted midlevel trough. An upper-tropospheric jet maximum progressed through the trough from a position upstream of the trough axis on 18 February to slightly downstream of the axis by midday on 19 February, with the most intense part of the rainband located in the left jet exit region. An analysis of the Q vectors and quasigeostrophically forced vertical motion showed strong midlevel convergence of the Q vectors and mid-to-upper-tropospheric QG rising motion in the region where both the TRMM observations and simulation indicated substantial amounts of precipitation ice. Thus, the mesoscale to synoptic-scale variability of the precipitation along the front can generally be explained by the combination of the low-level rising motion along the NCFR and the quasigeostrophically forced ascent associated with the upper-level front.
It is important to note that, not too surprisingly, the quasigeostrophically forced lifting only accounted for part of the total ascent simulated by the model, particularly because of the neglect of moist processes, the necessity to significantly smooth meteorological fields used as input to the ωQG calculations, and the limitation of QG theory. We anticipate that the effect of moisture would narrow, strengthen, and perhaps change the vertical location of the maximum QG updrafts, so that they agree better with the simulated updrafts, as shown by some previous studies (e.g., Locatelli et al. 2002). Despite its limitations, the dry QG theory reasonably accounts for the general mesoscale to synoptic-scale variability of precipitation along the front and elucidates the role of the upper-level front.
Even if the quasigeostrophically forced ascent was determined with the effects of moisture included and with information on smaller scales retained, it might still be too weak to produce graupel. However, it could still play a major role in determining the mesoscale to synoptic-scale precipitation variability by favoring the release of instabilities such as CSI. The large amount of precipitation ice in the WCFR may therefore have resulted from a combination of mechanisms at different scales.
Although not designed with observing midlatitude winter precipitation systems in mind, the TRMM satellite is capable of providing very unique measurements of precipitation in winter storms at lower midlatitudes. While significant investment has occurred under TRMM to relate microwave brightness temperatures in tropical convection to surface rainfall, much less effort has been made for midlatitude winter storms. Unlike tropical convection, which tends to form in relatively uniform environments with little variation of the melting level, winter frontal systems involve complex interactions between air masses of very different origins, interactions between frontal and jet stream circulations, and large variations in the height of the melting level. The result is that a wide variety of types of rainbands, formation mechanisms, vertical structures, and latent heating profiles are possible. In addition to studies of individual storm systems, TRMM offers a great opportunity to study the climatological characteristics of winter precipitation (e.g., see Kodama et al. 2007) in a way that was previously difficult or impossible prior to its launch in 1997. More detailed studies of winter storms will be possible in the future with the expected launch of NASA’s Global Precipitation Measurement mission in 2013 that will provide radar measurements up to higher latitudes (approximately 65°S–65°N) and, with radar measurements at two frequencies, will provide improved information on the microphysical properties (e.g., mean particle diameter) of winter precipitation systems.
Acknowledgments
The authors gratefully thank two anonymous reviewers and the editor, Dr. James D. Doyle, for their valuable comments, which led to significant improvements of the paper. This work was supported by Dr. Ramesh Kakar at NASA Headquarters with funds from the NASA Precipitation Measurement Mission science program.
REFERENCES
Adler, R. F., H-Y. M. Yeh, N. Prasad, W-K. Tao, and J. Simpson, 1991: Microwave simulations of a tropical rainfall system with a three-dimensional cloud model. J. Appl. Meteor., 30 , 924–953.
Adler, R. F., G. J. Huffman, D. T. Bolvin, S. Curtis, and E. J. Nelkin, 2000: Tropical rainfall distributions determined using TRMM combined with other satellite and rain gauge information. J. Appl. Meteor., 39 , 2007–2023.
Bennetts, D. A., and B. J. Hoskins, 1979: Conditional symmetric instability—A possible explanation for frontal rainbands. Quart. J. Roy. Meteor. Soc., 105 , 945–962.
Biggerstaff, M. I., E-K. Seo, S. M. Hristova-Veleva, and K-Y. Kim, 2006: Impact of cloud model microphysics on passive microwave retrievals of cloud properties. Part I: Model comparison using EOF analyses. J. Appl. Meteor. Climatol., 45 , 930–954.
Bluestein, H. B., 1992: Synoptic-Dynamics Meteorology in Midlatitudes. Vol I, Principles of Kinematics and Dynamics, Oxford University Press, 431 pp.
Bond, N. A., B. F. Smull, M. T. Stoelinga, C. P. Woods, and A. Haase, 2005: Evolution of a cold front encountering steep quasi-2D terrain: Coordinated aircraft observations on 8–9 December 2001 during IMPROVE-2. J. Atmos. Sci., 62 , 3559–3579.
Braun, S. A., 2006: High-resolution simulation of Hurricane Bonnie (1998). Part II: Water budget. J. Atmos. Sci., 63 , 43–64.
Braun, S. A., and W-K. Tao, 2000: Sensitivity of high-resolution simulations of Hurricane Bob (1991) to planetary boundary layer parameterizations. Mon. Wea. Rev., 128 , 3941–3961.
Braun, S. A., R. A. Houze Jr., and B. F. Smull, 1997: Airborne dual-Doppler observations of an intense frontal system approaching the Pacific Northwest coast. Mon. Wea. Rev., 125 , 3131–3156.
Browning, K. A., and T. W. Harrold, 1969: Air motion and precipitation growth in a wave depression. Quart. J. Roy. Meteor. Soc., 95 , 288–309.
Carbone, R. E., 1982: A severe frontal rainband. Part I: Stormwide hydrodynamic structure. J. Atmos. Sci., 39 , 258–279.
Cecil, D. J., and E. J. Zipser, 2002: Reflectivity, ice scattering, and lightning characteristics of hurricane eyewalls and rainbands. Part II: Intercomparison of observations. Mon. Wea. Rev., 130 , 785–801.
Cecil, D. J., E. J. Zipser, and S. W. Nesbitt, 2002: Reflectivity, ice scattering, and lightning characteristics of hurricane eyewalls and rainbands. Part I: Quantitative description. Mon. Wea. Rev., 130 , 769–784.
Cecil, D. J., S. J. Goodman, D. J. Boccippio, E. J. Zipser, and S. W. Nesbitt, 2005: Three years of TRMM precipitation features. Part I: Radar, radiometric, and lightning characteristics. Mon. Wea. Rev., 133 , 543–566.
Dudhia, J., 1989: Numerical study of convection observed during the winter monsoon experiment using a mesoscale two-dimensional model. J. Atmos. Sci., 46 , 3077–3107.
Eliassen, A., 1962: On the vertical circulation in frontal zones. Geofys. Publ., 24 , 147–160.
Evans, A. G., J. D. Locatelli, M. T. Stoelinga, and P. V. Hobbs, 2005: The IMPROVE-1 storm of 1–2 February 2001. Part II: Cloud structures and the growth of precipitation. J. Atmos. Sci., 62 , 3456–3473.
Fovell, R. G., and Y. Ogura, 1988: Numerical simulation of a midlatitude squall line in two dimensions. J. Atmos. Sci., 45 , 3846–3879.
Grell, G. A., J. Dudhia, and D. R. Stauffer, 1995: A description of the fifth-generation Penn State/NCAR Mesoscale Model (MM5). NCAR Tech. Note NCAR/TN-398+STR, 122 pp.
Han, M., R. M. Rauber, M. K. Ramamurthy, B. F. Jewett, and J. A. Grim, 2007: Mesoscale dynamics of the trowal and warm-frontal regions of two continental winter cyclones. Mon. Wea. Rev., 135 , 1647–1670.
Hobbs, P. V., and P. O. G. Persson, 1982: The mesoscale and microscale structure and organization of clouds and precipitation in mid-latitude cyclones. Part V: The substructure of narrow cold-frontal rainbands. J. Atmos. Sci., 39 , 280–295.
Hoskins, B. J., I. Draghici, and H. C. Davies, 1978: A new look at the ω equation. Quart. J. Roy. Meteor. Soc., 104 , 31–38.
Houze Jr., R. A., 1993: Cloud Dynamics. Academic Press, 573 pp.
Houze Jr., R. A., P. V. Hobbs, K. R. Biswas, and W. M. Davis, 1976: Mesoscale rainbands in extratropical cyclones. Mon. Wea. Rev., 104 , 868–878.
Janjic, Z. I., 1990: The step-mountain coordinate: Physical package. Mon. Wea. Rev., 118 , 1429–1443.
Janjic, Z. I., 1994: The step-mountain eta coordinate model: Further development of the convection, viscous sublayer, and turbulent closure schemes. Mon. Wea. Rev., 122 , 927–945.
Jorgensen, D. P., Z. Pu, P. O. G. Persson, and W-K. Tao, 2003: Variations associated with cores and gaps of a Pacific narrow cold frontal rainband. Mon. Wea. Rev., 131 , 2705–2729.
Keyser, D., and M. A. Shapiro, 1986: A review of the structure and dynamics of upper-level frontal zone. Mon. Wea. Rev., 114 , 452–499.
Kodama, Y-M., H. Okabe, Y. Tomisaka, K. Kotono, Y. Kondo, and H. Kasuya, 2007: Lightning frequency and microphysical properties of precipitating clouds over the western North Pacific during winter as derived from TRMM multisensor observations. Mon. Wea. Rev., 135 , 2226–2241.
Kummerow, C., W. Barnes, T. Kozu, J. Shiue, and J. Simpson, 1998: The Tropical Rainfall Measuring Mission (TRMM) sensor package. J. Atmos. Oceanic Technol., 15 , 809–817.
Locatelli, J. D., J. E. Martin, and P. V. Hobbs, 1994: A wide cold-front rainband and its relationship to frontal topography. Quart. J. Roy. Meteor. Soc., 120 , 259–275.
Locatelli, J. D., M. T. Stoelinga, and P. V. Hobbs, 2002: Organization and structure of clouds and precipitation on the Mid-Atlantic coast of the United States. Part II: Diagnosis of a nonconvective rainband associated with a cold front aloft. Mon. Wea. Rev., 130 , 278–297.
Martin, J. E., 1998: The structure and evolution of a continental winter cyclone. Part I: Frontal structure and occlusion process. Mon. Wea. Rev., 126 , 303–328.
Matejka, T. J., R. A. Houze Jr., and P. V. Hobbs, 1980: Microphysics and dynamics of clouds associated with mesoscale rainbands in extratropical cyclones. Quart. J. Roy. Meteor. Soc., 106 , 29–56.
McFarquhar, G. M., H. Zhang, G. Heymsfield, R. Hood, J. Dudhia, J. B. Halverson, and F. Marks Jr., 2006: Factors affecting the evolution of Hurricane Erin (2001) and the distribution of hydrometeors: Role of microphysical processes. J. Atmos. Sci., 63 , 127–150.
Parsons, D. B., 1992: An explanation for intense frontal updrafts and narrow cold-frontal rainbands. J. Atmos. Sci., 49 , 1810–1825.
Parsons, D. B., C. G. Mohr, and T. Gal-Chen, 1987: A severe frontal rainband. Part III: Derived thermodynamic structure. J. Atmos. Sci., 44 , 1615–1631.
Rotunno, R., J. B. Klemp, and M. L. Weisman, 1988: A theory for strong, long-lived squall lines. J. Atmos. Sci., 45 , 463–485.
Schultz, D. M., and C. A. Doswell, 1999: Conceptual models of upper-level frontogenesis in south-westerly and north-westerly flow. Quart. J. Roy. Meteor. Soc., 125 , 2535–2562.
Smith, P. L., 1984: Equivalent radar reflectivity factor for snow and ice particles. J. Climate Appl. Meteor., 23 , 1258–1260.
Spencer, R. W., W. S. Olson, W. Rongzhang, D. W. Martin, J. A. Weinman, and D. A. Santeck, 1983: Heavy thunderstorms observed over land by the Nimbus 7 Scanning Multichannel Microwave Radiometer. J. Climate Appl. Meteor., 22 , 1041–1046.
Spencer, R. W., H. M. Goodman, and R. E. Hood, 1989: Precipitation retrieval over land and ocean with the SSM/I: Identification and characteristics of the scattering signal. J. Atmos. Oceanic Technol., 6 , 254–273.
Stoelinga, M. T., J. D. Locatelli, and P. V. Hobbs, 2000: Structure and evolution of winter cyclones in the central United States and their effects on the distribution of precipitation. Part VI: A mesoscale modeling study of the initiation of convective rainbands. Mon. Wea. Rev., 128 , 3481–3500.
Tao, W-K., and J. Simpson, 1993: The Goddard Cumulus Ensemble Model. Part I: Model description. Terr. Atmos. Oceanic Sci., 4 , 35–72.
Xu, Q., 1989a: Extended Sawyer–Eliassen equation for frontal circulations in the presence of small viscous moist symmetric stability. J. Atmos. Sci., 46 , 2671–2683.
Xu, Q., 1989b: Frontal circulations in the presence of small viscous moist symmetric stability and weak forcing. Quart. J. Roy. Meteor. Soc., 115 , 1325–1353.
(a) Geopotential height (contours, 60-m intervals), wind (full barb = 5 m s−1, short barb = 2.5 m s−1), and absolute vorticity (shading) at 500 hPa from NCEP global analyses at 0000 UTC 19 Feb 2001. (b) Geopotential height (contours, 60-m intervals) and wind speed (shading) at 300 hPa. The thick dashed line outlines the 306 K equivalent potential temperature at 975 hPa as shown in (c) Sea level pressure (contours, 4-hPa intervals), winds, and equivalent potential temperature (shading) at 975 hPa. (d) Visible satellite image showing the frontal cloud system of the cyclone at 2352 UTC 18 Feb 2001. The inset box indicates the approximate location of the TRMM view at 0023 UTC 19 Feb as shown in Fig. 2. Solid and dashed lines outline the PR and TMI swath edges, respectively.
Citation: Monthly Weather Review 137, 3; 10.1175/2008MWR2465.1
TRMM PR reflectivity (dBZ) at 2-km altitude MSL and TMI PCT (K) for three channels (19, 37, and 85 GHz) at (left) 0023, (middle) 0200, and (right) 0337 UTC 19 Feb 2001. Lines in the PCT panels outline the swath of the PR.
Citation: Monthly Weather Review 137, 3; 10.1175/2008MWR2465.1
Geographic locations of the four domains used in the numerical simulation.
Citation: Monthly Weather Review 137, 3; 10.1175/2008MWR2465.1
Domain 1 sea level pressure (thick contours, every 4 hPa), equivalent potential temperature (thin contours and shading, every 6 K), and wind barbs (full barb is 5 m s−1) at 975 hPa at (a) 2000 UTC 18 Feb, (b) 0200 UTC 19 Feb, and (c) 0800 UTC 19 Feb 2001. The shading of the equivalent potential temperature is as follows: <300 K (white), 300–306 K (light shading), 306–312 K (medium shading), and >312 K (dark shading).
Citation: Monthly Weather Review 137, 3; 10.1175/2008MWR2465.1
Domain 1 geopotential height (contours, every 60 m) and wind speed (shading) at 300 hPa at (a) 2000 UTC 18 Feb, (b) 0200 UTC 19 Feb, and (c) 0800 UTC 19 Feb 2001. The thick dashed line is the 306-K θe contour at 975 hPa.
Citation: Monthly Weather Review 137, 3; 10.1175/2008MWR2465.1
Domain 1 potential vorticity (shading, also contoured at every 1 PVU, dashed contour is 0 PVU) on the 320-K θ surface at (a) 2000 UTC 18 Feb, (b) 0200 UTC 19 Feb, and (c) 0800 UTC 19 Feb 2001. The thick solid line is the 306-K θe contour at 975 hPa.
Citation: Monthly Weather Review 137, 3; 10.1175/2008MWR2465.1
Domain 3 simulated radar reflectivity at 2-km altitude at (a) 0000, (b) 0200, (c) 0400, (d) 0600, (e) 0800, and (f) 1000 UTC 19 Feb 2001. Straight lines indicate the locations of the cross sections AA′, BB′, CC′, and DD′ in Figs. 9 and 10.
Citation: Monthly Weather Review 137, 3; 10.1175/2008MWR2465.1
Domain 3 column-integrated graupel at (a) 0200 and (b) 0800 UTC 19 Feb 2001. The lines AA′, BB′, CC′, and DD′ represent the same cross sections as in Fig. 7. (Note: In this simulation the total precipitation ice is dominated by graupel. The horizontal distribution of graupel is consistent with the overall distribution of both snow and graupel)
Citation: Monthly Weather Review 137, 3; 10.1175/2008MWR2465.1
Domain 3 snow (dashed contours), graupel (shading), and rain (thick solid contours) mixing ratios and potential temperature (thin solid contours) at cross sections (a) AA′ and (b) BB′ at 0200 UTC 19 Feb and (c) CC′ and (d) DD′ at 0800 UTC 19 Feb 2001. The snow, graupel, and rain mixing ratios are contoured every 0.2 g kg−1 starting at 0.1 g kg−1. The bold lines in (a) and (c) mark the leading edge of the upper-level front. The arrow along the bottom in each panel marks the leading edge of the low-level front. Fields represent a 100-km average (50 km to each side) in the direction normal to the cross section.
Citation: Monthly Weather Review 137, 3; 10.1175/2008MWR2465.1
As in Fig. 9, but for equivalent potential temperature (contours, 2-K intervals) and vertical motion (only rising motions are shaded, dPa s−1).
Citation: Monthly Weather Review 137, 3; 10.1175/2008MWR2465.1
Domain 2 potential vorticity (shading) and wind barbs (full barb is 5 m s−1, flag is 25 m s−1) at 300 hPa at (a) 0200 and (b) 0800 UTC 19 Feb 2001. The bold contour is the 306-K θe contour at 975 hPa. The straight lines indicate the locations of cross sections LA–LA′, LB–LB′, LC–LC′, and LD–LD′ in Figs. 13 and 14. LA–LA′, LB–LB′ LC–LC′, and LD–LD′ (in domain 2) are through, but longer than AA′, BB′, CC′, and DD′ (in domain 3, heavy bold lines), respectively.
Citation: Monthly Weather Review 137, 3; 10.1175/2008MWR2465.1
Domain 2 quasigeostrophic rising motion (shading, 6 dPa s−1), geopotential height (contoured at 60-m intervals), and Q vectors (10−6 m s−3 hPa−1) at 500 hPa at (a) 0200 and (b) 0800 UTC 19 Feb 2001. The straight lines show the same cross sections as in Fig. 11.
Citation: Monthly Weather Review 137, 3; 10.1175/2008MWR2465.1
Domain 2 potential vorticity (shading) and rising motion (contoured, 6 dPa s−1) along cross sections (a) LA–LA′ and (b) LB–LB′ at 0200 UTC 19 Feb, and (c) LC–LC′ and (d) LD–LD′ at 0800 UTC 19 Feb 2001. The straight lines with arrows indicate the cross sections AA′, BB′, CC′, and DD′ in domain 3. See Figs. 11 or 12 for the locations of the cross sections.
Citation: Monthly Weather Review 137, 3; 10.1175/2008MWR2465.1
Domain 2 quasigeostrophic rising motion (shading, 6 dPa s−1) and equivalent potential temperature (contours, 2-K intervals) along cross sections (a) LA–LA′ and (b) LB–LB′ at 0200 UTC 19 Feb, and (c) LC–LC′ and (d) LD–LD′at 0800 UTC 19 Feb 2001. The bold lines mark the leading edges of the upper- and lower-level fronts. The arrows are the same as shown in Fig. 13.
Citation: Monthly Weather Review 137, 3; 10.1175/2008MWR2465.1
As described in section 2, the NCFR (with clear core and gap regions) and WCFR can be readily differentiated in the TRMM PR observations. In contrast, because of its coarser resolution, the TMI is generally unable to clearly separate the NCFR and WCFR. Thus, this paper will occasionally refer to the NCFR and WCFR simply as “the frontal rainband.”
Here β = 0.65 is applied to the equation PCT = (βTBh − TBv)/(β − 1) (Spencer et al. 1989) for 19 GHz, where TBh and TBv are the horizontally and vertically polarized brightness temperatures, respectively.
The constants used here are identical to Braun (2006) except that the value of α, the ratio of the backscattering coefficients for the reflecting ice particles and water, was changed from 0.213 to 0.224 following Smith (1984).
The Read–Interpolate–Plot software developed by M. T. Stoelinga at the University of Washington is used to calculate ωQG, using a simple relaxation method to invert the Q vector form of the ωQG equation [see Stoelinga et al. (2000) for the description of the method].