1. Introduction
Observations of cloud-top precipitation generating cells1 (GCs) in winter cyclones date back to the 1950s (Marshall 1953). These convective cells at cloud top are generally 0.75–1.5 km wide and 1–2 km deep (Langleben 1956; Rosenow et al. 2014; Kumjian et al. 2014) and form above frontal layers in the upper troposphere (Douglas et al. 1957) where potential instability is favored (Wexler and Atlas 1959). Within GCs, updrafts of 0.75–3.00 m s−1 are common (Wexler 1955; Douglas et al. 1957; Wexler and Atlas 1959; Carbone and Bohne 1975; Rosenow et al. 2014; Kumjian et al. 2014). Precipitation fallstreaks emanating from the base of GCs seed underlying clouds where ice crystals grow and aggregate (Plummer et al. 2014, 2015). In winter cyclones, these fallstreaks have been traced from GCs to either the surface as individual streaks or merging into heavy banded precipitation common in the comma head of midlatitude cyclones (Evans et al. 2005; Cunningham and Yuter 2014; Plummer et al. 2014; Rauber et al. 2014a,b, 2015; Rosenow et al. 2014). Recent observations from the 14 cyclones sampled with the Wyoming Cloud Radar (WCR; Wang et al. 2012) during the Profiling of Winter Storms (PLOWS) campaign (Rosenow et al. 2014; Plummer et al. 2014; Rauber et al. 2014a,b, 2015) indicate that GCs are likely ubiquitous at cloud top in the warm-frontal and comma-head regions of winter cyclones. It is clear from these observations that an understanding of GC dynamics is important to understanding winter cyclone precipitation processes.
In recent years, it has been suggested that the dynamics of GCs could be analogous to that of stratocumulus clouds (Syrett et al. 1995; Kumjian et al. 2014; Rauber et al. 2014a,b), where radiative forcing favors destabilization at cloud top and development of convection (Wood 2012). Keeler et al. (2016, hereafter Part I) directly addressed this hypothesis by performing idealized simulations with nighttime, daytime, and no radiative forcing under shear and stability conditions representative of the comma head of a cyclone observed during the PLOWS field campaign. Kinematic and bulk microphysical properties of simulated GCs in Part I compared favorably with those in Rosenow et al. (2014) and Plummer et al. (2014) only when nocturnal radiative forcing representative of conditions in the actual cyclone was included in the model. Updrafts in the nighttime simulation exceeding 2.00 m s−1 were associated with high ice supersaturation (RHice > 165%), which favored rapid depositional growth of ice, with ice precipitation mixing ratio maxima occasionally >0.27 g kg−1. For the shear and stability conditions present in that cyclone, GCs were only maintained in simulations where radiative forcing was present. The range of the vertical air velocity (w) spectrum (defined as the difference between the 99th and 1st percentiles of w) was largest for nighttime radiation, when shortwave warming did not offset longwave cooling and its associated cloud-top destabilization.
In this paper, the model initial conditions from Part I are modified to examine the evolution of GCs under a broad range of upper-tropospheric stability conditions with nighttime, daytime, and no radiative forcing. In particular, this paper will assess the role of radiative forcing in either maintaining convection in environments where instability is preexisting or destabilizing the cloud-top region to the point where GCs can develop in environments that are either neutral or stable.
2. Idealized simulation methodology
a. Initial conditions
In Part I, initial conditions for idealized simulations of GCs consisted of a profile of θ, qυ, u, and υ from a WRF, version 3.3.1, storm-scale simulation of a winter cyclone observed during PLOWS. As discussed in Part I, this profile was representative of the shear and instability conditions in which GCs were observed overnight in the 14–15 February 2010 cyclone (Part I, their Figs. 5–7). The idealized simulation in Part I with nighttime radiative forcing resulted in simulated GCs with w and qi characteristics that compared favorably with observations from PLOWS (Part I, their Tables 2 and 3). Two layers of potential instability were present in their initial conditions, with θei decreasing between 4.9–5.3 and 5.9–6.7 km at 0.98 and 0.40 K km−1, respectively, corresponding to CAPE with respect to ice of 20.2 and 9.5 J kg−1.
Simulations presented in this paper were run for eight simplified stability profiles (Fig. 1), where the lower level of instability was removed by modifying θ so that θei increased with height at a constant rate from its former maximum of 298.2 K at 4.9 km to 298.4 K at 5.9 km, the base of the upper layer of potential instability. The original θei profile from Part I is shown in Fig. 1b in black, with modifications shown in blue. Above 5.9 km, the stability profiles were in one of three categories: potentially unstable, potentially neutral, and stable. For the potentially unstable profiles, θ was modified so that θei decreased by 0.53, 0.40, 0.27, and 0.13 K km−1 from 5.9 to 6.7 km, above which θei increased so that it intersected the original profile at 7.1 km (Fig. 1, profiles a–d). In the potentially neutral profile (Fig. 1e), θ was modified so that θei was constant at 298.4 K between 5.9 and 7.1 km. In the stable profiles θ was modified so that θei increased with height at 0.15, 0.30, and 0.45 K km−1 between 5.9 km and the altitude at which the modified θei profile intersected with the original profile (Fig. 1, profiles f–h). Above 7.6 and below 4.9 km, θ was not modified from the profile used in Part I; qυ was not modified for any of the profiles used in these simulations. The vertical wind profile for all simulations is the same as in Part I (their Fig. 7). The wind profile between 5.0- and 9.0-km altitude is shown in Fig. 2.
(a) Equivalent potential temperature with respect to ice for each stability profile (labeled a–h) displayed between 5- and 7.75-km altitude. (b) Profile of θei for 0–8 km from Part I (black line), with modifications to profiles shown in (a) given in blue. The extent of (a) is indicated by the gray box.
Citation: Journal of the Atmospheric Sciences 73, 4; 10.1175/JAS-D-15-0127.1
Vertical profile of the u (black line) and υ (blue line) components of the horizontal wind in the model initial conditions. Altitude is displayed on the vertical axis.
Citation: Journal of the Atmospheric Sciences 73, 4; 10.1175/JAS-D-15-0127.1
b. Model settings
Aside from differences in the initial θ profile, all model settings for the simulations in this paper are the same as those for the idealized simulations discussed in section 4b of Part I. In summary, idealized WRF 3.3.1 simulations were integrated for 180 min on a 50.1 × 50.1 × 15 km3 grid with horizontal grid spacing of 100 m, vertical level spacing of ~50 m near the GC level, and a dynamical time step of 0.5 s. Thompson microphysics and RRTMG radiation parameterizations were used, with a radiation time step of 30 s. The deep layer of nimbostratus below GCs was maintained using the WRF-LES module (Yamaguchi and Feingold 2012) using the prescribed ascent profile shown in Part I (their Fig. 9).
3. Results
Idealized simulations are presented for all combinations of the 8 initial stability profiles (Fig. 1) and all three radiation settings (nighttime, daytime, and no radiation), yielding a total of 24 simulations. Analyses are organized into sections on (a) bulk statistical analysis and (b) cross-sectional analysis of representative GCs. Bulk analyses include contoured frequency by time diagram (CFTD) analyses of the vertical air velocity w and Hovmöller-like time–height cross sections of the domain-averaged longwave, shortwave, latent, and net diabatic heating rates. As in Part I, cross sections presented later in the section are along the shear vector from west-southwest to east-northeast, are 10 km long, and are centered on the model grid.
a. Bulk properties of simulations
1) Vertical velocity spectrum
CFTDs of w for all nighttime radiation simulations are shown in Fig. 3, with panel labels corresponding to the stability profiles in Fig. 1. Development of convection, indicated by divergence of the w spectrum from 0 m s−1, delayed from t = 15 to 80 min as instability was decreased in the model initial conditions from the most unstable to neutral profile (Figs. 3a–e). The time from the initial divergence of the w spectrum until when the w spectrum reached its maximum range also increased as instability was decreased, indicating a slower development of convection under weaker instability. For the most unstable profile (Fig. 3a), the w spectrum range increased rapidly from t = 15 to 60 min, when the 99th percentile of w reached a maximum of 2.20 m s−1. The w spectrum range (with stability profile a) decreased slightly to a near steady state following the initial development of convection, with a 1st–99th percentile range from approximately −1.60 to 1.70 m s−1 for the remainder of the simulation. Both the w spectrum maximum range and steady-state range decreased as stability was increased in the model initial conditions. For example, the maximum 99th percentile decreased to 2.10, 2.00, and 2.00 m s−1, and the steady-state 99th-percentile w decreased to 1.60, 1.60, and 1.40 m s−1 for the less unstable stability profiles (Figs. 3b–d). For the neutral stability profile (Fig. 3e), the maximum 99th-percentile w decreased to 1.90 m s−1 and the steady-state 99th-percentile w decreased to 1.30 m s−1. This trend continued for the stable profiles (Figs. 3f–h), with maximum 99th-percentile w of 1.70, 1.60, and 1.40 m s−1, respectively. The decrease in the w spectrum range following its initial maximum nearly leveled off near the end of the simulation at t = 175–180 min, with the 99th-percentile w at 1.20, 1.10, and 1.10 m s−1 for the stable profiles.
CFTD analysis of the vertical air velocity data at the GC level (6–8 km) every 5 min with 0.1 m s−1 bin width for the nighttime simulations. Panels correspond to the stability profile for the simulations (see Fig. 1). Contours in individual panels show the 99th, 95th, 90th, 75th, 50th, 25th, 10th, 5th, and 1st percentiles of w. (b),(e),(h) Vertical bars at t = 80 and 140 min indicate the time for which cross sections of model output are shown later in Figs. 14–17 and 20–21.
Citation: Journal of the Atmospheric Sciences 73, 4; 10.1175/JAS-D-15-0127.1
Evolution of the w spectrum in the daytime radiation simulations (Fig. 4) was similar to that of the nighttime simulations, with early peaks in w spectrum range followed by near steady-state spectra for the unstable and neutral initial stability profiles (Figs. 4a–e). While the peak 99th percentiles of w were only ~0.10 m s−1 lower than in the nighttime radiation simulations for profiles a–c (Figs. 4a–c), with values of 2.10, 2.00, and 1.90 m s−1, differences in the w spectrum were more substantial both later in the simulations and for the more stable profiles. For profiles d–h (Figs. 4d–h), the peak 99th percentiles of w were 1.60, 1.60, 1.40, 0.90, and 0.70 m s−1, a decrease of 0.40, 0.30, 0.40, 0.70, and 0.70 m s−1 relative to the nighttime simulations. The near steady-state 99th-percentile w was 1.20, 1.30, 1.30, 1.10, and 1.00 m s−1 for the unstable and neutral profiles (profiles a–e), a decrease of 0.50, 0.30, 0.30, 0.30, and 0.30 m s−1. In the weakly stable simulation (profile f), the 99th percentile of w decreased to 1.00 m s−1 by the end of the simulation (t = 180 min), but it is not clear whether it had reached a steady state. The w spectrum for simulations with the two most stable profiles (profiles g and h) reached its maximum range at t = 150 min and maintained that range through the end of the simulation. These results show for a wider range of conditions than in Part I that GCs with the same initial stability profile will have weaker updrafts and downdrafts when shortwave warming offsets longwave cooling, resulting in weaker destabilization near cloud top.
As in Fig. 3, but for the daytime simulations.
Citation: Journal of the Atmospheric Sciences 73, 4; 10.1175/JAS-D-15-0127.1
Evolution of the w spectrum for the no-radiation simulations (Fig. 5) was heavily dependent on the model initial stability profile. For the initially unstable simulations (Figs. 5a–d), release of instability resulted in peak 99th percentiles of w of 1.70, 1.40, 1.10, and 0.90 m s−1. Note that even at early times in these simulations, the w spectrum ranges were lower than for the nighttime and daytime simulations, where cloud-top longwave cooling enhanced the preexisting instability. While the w spectrum range did decrease following the initial release of instability, convection did not dissipate completely. In the initially neutral simulation (Fig. 5e), the 99th percentile of w increased from 0.20 to 1.00 m s−1 from t = 135 to 180 min. The reason for the broadening of the w spectrum late in the initially neutral no-radiation simulation will be discussed in section 3b(2). Under stable conditions with no radiation (Figs. 5f–h), the 99th percentile of w did not exceed 0.10 m s−1 for the duration of those simulations.
As in Fig. 3, but for the no-radiation simulations.
Citation: Journal of the Atmospheric Sciences 73, 4; 10.1175/JAS-D-15-0127.1
2) Longwave heating rate
Time–height plots of the domain-averaged longwave heating rate
Time evolution of the domain-averaged longwave heating rate for the nighttime simulations. Panels correspond to the stability profile for the simulations (see Fig. 1). As in Fig. 3, contours in individual panels show the 99th, 95th, 90th, 75th, 50th, 25th, 10th, 5th, and 1st percentiles of w.
Citation: Journal of the Atmospheric Sciences 73, 4; 10.1175/JAS-D-15-0127.1
As in Fig. 6, but for the daytime simulations. As in Fig. 4, contours in individual panels show the 99th, 95th, 90th, 75th, 50th, 25th, 10th, 5th, and 1st percentiles of w.
Citation: Journal of the Atmospheric Sciences 73, 4; 10.1175/JAS-D-15-0127.1
3) Shortwave heating rate
The domain-averaged shortwave heating rate
Time evolution of the domain-averaged shortwave heating rate for the daytime simulations. Panels correspond to the stability profile for the simulations (see Fig. 1). As in Fig. 4, contours in individual panels show the 99th, 95th, 90th, 75th, 50th, 25th, 10th, 5th, and 1st percentiles of w.
Citation: Journal of the Atmospheric Sciences 73, 4; 10.1175/JAS-D-15-0127.1
4) Latent heating rate
Time–height cross sections of the domain-averaged latent heating rate
Time evolution of the domain-averaged latent heating rate for the nighttime simulations. Panels correspond to the stability profile for the simulations (see Fig. 1). As in Fig. 3, contours in individual panels show the 99th, 95th, 90th, 75th, 50th, 25th, 10th, 5th, and 1st percentiles of w.
Citation: Journal of the Atmospheric Sciences 73, 4; 10.1175/JAS-D-15-0127.1
As in Fig. 9, but for the daytime simulations. As in Fig. 4, contours in individual panels show the 99th, 95th, 90th, 75th, 50th, 25th, 10th, 5th, and 1st percentiles of w.
Citation: Journal of the Atmospheric Sciences 73, 4; 10.1175/JAS-D-15-0127.1
As in Fig. 9, but for the no-radiation simulations. As in Fig. 5, contours in individual panels show the 99th, 95th, 90th, 75th, 50th, 25th, 10th, 5th, and 1st percentiles of w.
Citation: Journal of the Atmospheric Sciences 73, 4; 10.1175/JAS-D-15-0127.1
5) Net diabatic heating
The net diabatic heating rate for the nighttime simulations
Time evolution of the domain-averaged net diabatic heating rate for the nighttime simulations (longwave + latent heating). Panels correspond to the stability profile for the simulations (see Fig. 1). As in Fig. 3, contours in individual panels show the 99th, 95th, 90th, 75th, 50th, 25th, 10th, 5th, and 1st percentiles of w.
Citation: Journal of the Atmospheric Sciences 73, 4; 10.1175/JAS-D-15-0127.1
Time evolution of the domain-averaged net diabatic heating rate for the daytime simulations (longwave + shortwave + latent heating). Panels correspond to the stability profile for the simulations (see Fig. 1). As in Fig. 4, contours in individual panels show the 99th, 95th, 90th, 75th, 50th, 25th, 10th, 5th, and 1st percentiles of w.
Citation: Journal of the Atmospheric Sciences 73, 4; 10.1175/JAS-D-15-0127.1
In the daytime simulations, minima in
In this section, the bulk evolution of diabatic processes and their relation to evolution of the upper-tropospheric w spectrum were shown. The following section will discuss the evolution of GC structure and kinematics in the context of the bulk analyses presented in the current section.
b. Properties of individual generating cells
In this section evolution of GCs (or lack thereof) in individual simulations are discussed using cross sections of w, RHice, ice precipitation mixing ratio qi, and the net diabatic heating rate QN. In this paper, ice precipitation mixing ratio is defined as the sum of the cloud ice and snow mixing ratios. The net diabatic heating rate is defined as the sum of the longwave and shortwave radiative heating rates (when present) and the latent heating rate. The evolution of model fields is shown in Figs. 14–17 (see also Figs. 20 and 21) for daytime, nighttime, and no-radiation simulations for the following profiles: an initially unstable profile (Fig. 1, profile b), an initially neutral profile (Fig. 1, profile e), and an initially stable profile (Fig. 1, profile h). Fields are shown for 10-km-long cross sections in the plane of the shear vector, centered on the model grid at t = 80 and 140 min, times that are indicated by vertical bars in Figs. 3b,e,h, 4b,e,h, and 5b,e,h in the w CFTDs.
(a)–(c) Vertical air velocity, (d)–(f) relative humidity with respect to ice, (g)–(i) ice precipitation mixing ratio, and (j)–(l) net diabatic heating rate at t = 80 min for simulations with stability profile b (see Fig. 1) and (left)–(right) nighttime, daytime, and no radiation.
Citation: Journal of the Atmospheric Sciences 73, 4; 10.1175/JAS-D-15-0127.1
(a)–(c) Vertical air velocity, (d)–(f) relative humidity with respect to ice, (g)–(i) ice precipitation mixing ratio, and (j)–(l) net diabatic heating rate at t = 140 min for simulations with stability profile b (see Fig. 1) and (left)–(right) nighttime, daytime, and no radiation.
Citation: Journal of the Atmospheric Sciences 73, 4; 10.1175/JAS-D-15-0127.1
(a)–(c) Vertical air velocity, (d)–(f) relative humidity with respect to ice, (g)–(i) ice precipitation mixing ratio, and (j)–(l) net diabatic heating rate at t = 80 min for simulations with stability profile e (see Fig. 1) and (left)–(right) nighttime, daytime, and no radiation.
Citation: Journal of the Atmospheric Sciences 73, 4; 10.1175/JAS-D-15-0127.1
(a)–(c) Vertical air velocity, (d)–(f) relative humidity with respect to ice, (g)–(i) ice precipitation mixing ratio, and (j)–(l) net diabatic heating rate at t = 140 min for simulations with stability profile e (see Fig. 1) and (left)–(right) nighttime, daytime, and no radiation.
Citation: Journal of the Atmospheric Sciences 73, 4; 10.1175/JAS-D-15-0127.1
1) Potentially unstable simulations
Well-defined GCs were present at t = 80 min in the stability profile b, nighttime radiation simulation (Fig. 14a). The upward vertical velocity exceeded 2.00 m s−1 at this time in some GCs as shown in Fig. 3b, with extreme values of w > 3.00 m s−1. Downdraft magnitudes between GCs were generally from 1.00 to >1.50 m s−1 and were collocated with RHice minima as low as 50% (Fig. 14d). Strong supersaturation with respect to ice was present in the most vigorous updrafts, with RHice > 165%, favoring qi maxima > 0.24 g kg−1 (Fig. 14g). The most pronounced diabatic cooling maxima were >2.00 K h−1 and were associated with longwave cooling above 7.0 km (Fig. 14j). Latent heating maxima > 4.00 K h−1 were present within the strongest updrafts, while heating > 2.00 K h−1 was common in weaker updrafts.
As discussed earlier, cloud-top convection appeared to reach a steady state prior to t = 140 min in the nighttime stability profile b simulation, with |w| > 1.50 m s−1 (Fig. 15a). The RHice range was similar to earlier in the simulation, from as low as 55% in the stronger downdrafts while exceeding 170% in GCs with the strongest updrafts (Fig. 15d). Ice precipitation mixing ratios in GCs were also similar, with some maxima in excess of 0.24 g kg−1 (Fig. 15g). Fallstreaks can be seen emanating from GCs at both t = 80 and 140 min in this simulation (Figs. 14g, 15g). Latent heating contributed to QN of 1.00–1.50 K h−1 in the stronger updraft areas (Fig. 15j). While longwave cooling maxima were slightly lower (1.25–2.00 K h−1; Fig. 15j), cooling was more uniform at cloud top than earlier, which contributed to the strong domain-averaged net diabatic cooling shown earlier in Fig. 12b.
Individual updrafts were weaker at t = 80 min in the daytime unstable simulation cross section (Fig. 14b), with typical w maxima > 1.00 m s−1 and one updraft > 1.50 m s−1. RHice was slightly lower than in the nighttime unstable simulation, with peak values near 155%–165% (Fig. 14e). This contributed to slightly lower ice precipitation mixing ratios, with qi maxima > 0.21 g kg−1 in the GC with w > 1.50 m s−1 (Fig. 14h). The offset of longwave cooling by shortwave warming is evident in Fig. 14k, with maximum net cooling of ~1.00 K h−1 near the top of some GCs. Net heating within GCs is slightly lower than in the nighttime simulation with maxima near 3.00 K h−1 and typical values of 1.50–2.00 K h−1 in most GCs.
GCs persisted, but had weakened somewhat by t = 140 min, with updrafts of 0.50–1.00 m s−1 in the cross section shown in Fig. 15b and one downdraft with a magnitude > 1.00 m s−1. RHice maxima were near 145%–155%, which supported qi maxima in most GC > 0.12 g kg−1 and qi maxima > 0.18 g kg−1 in one GC (Fig. 15h). As in the unstable nighttime simulations, fallstreaks emanated from GCs at both t = 80 and 140 min in this simulation. Net diabatic cooling was less focused near cloud top late in this simulation, with peak cooling of 0.75–1.00 K h−1. Net diabatic heating had also decreased by t = 140 min, with peak values near 1.00 K h−1. Since the updrafts shown in Fig. 15b were only near the 75th percentile of w at that time (Fig. 4b), typical RHice, qi, and QN maxima for this time may be higher than shown in Figs. 15e, 15h, and 15k.
In the stability profile b and no-radiation simulation, updrafts and downdrafts at t = 80 min were weaker than in simulations with radiation parameterized, with |w| > 1.00 m s−1 in the strongest convection (Fig. 14c). These weaker updrafts were associated with RHice maxima of ~145%–160% and precipitation cores with maximum qi > 0.15 g kg−1 (Figs. 14f,i). In the no-radiation simulations net diabatic heating only consists of latent heating; thus, no longwave cooling was present at cloud top to favor maintenance of instability. Cooling present between GCs of up to 0.75 K h−1 was associated with downdrafts and minima in RHice, while heating within GCs of up to ~1.00 K h−1 was associated with RHice maxima.
Convection had weakened substantially by t = 140 min, with a few small areas of |w| > 0.50 m s−1 (Fig. 15c). While the updrafts are associated with RHice maxima of ~135%–140% (Fig. 15f), they are either on the edge of, or completely outside of, areas with qi > 0.03 g kg−1 (Fig. 15i). Residual fallstreaks are present at lower altitude (below ~6.2 km), associated with former GCs (Fig. 15i). Latent heating and cooling maxima were much weaker overall, with cooling up to 0.25 K h−1 between fallstreaks and heating up to 1.00 K h−1 in isolated ~100-m patches within fallstreaks.
2) Potentially neutral simulations
While the w spectrum had diverged from a range near 0 m s−1 prior to t = 80 min in the nighttime radiation initially neutral simulation (Fig. 3e), the peak magnitude of w had not yet exceeded 0.50 m s−1, the lowest w contour in Fig. 16a. The weak, developing convection is evident in the small RHice maxima of 120%–125% (Fig. 16d) and in the small perturbations in qi near cloud top (Fig. 16g). The convection was developing as a result of the destabilization present in the QN field, which was characterized by broad cooling of 0.00–0.25 K h−1 above 7.0 km and broad heating between 5.0 and 6.0 km of 0.25–0.50 K h−1 (Fig. 16j). Between 6.0 and 7.0 km, heating > 0.25 K h−1 was present in the highest RHice maxima, where updrafts were developing. Elsewhere between 6.0 and 7.0 km, cooling maxima locally > 0.50 K h−1 within the top few hundred meters of qi > 0.30 g kg−1 contributed to destabilization.
By t = 140 min, the effect of this destabilization led to development of well-defined GCs, with typical updraft maxima near 1.50 m s−1 and an isolated maxima > 2.50 m s−1 in the cross section (Fig. 17a). RHice maxima associated with the strongest of the updrafts were ~160%–165%, and RHice minima were the lowest of any of the cross sections presented in this paper at ~40% (Fig. 17d). Maxima in qi exceeded 0.21 g kg−1 near the strongest updrafts, and clearly defined fallstreaks emanated from the GCs (Fig. 17g). High-precipitation content favored strong longwave cooling, which is evident in the QN field with cooling > 2.50 K h−1 above isolated GCs; cooling maxima of 1.00–2.00 K h−1 were typical near cloud top elsewhere (Fig. 17j). Within GCs, heating of 1.50–2.00 K h−1 was common, with an isolated peak heating > 4.00 K h−1 in one GC with w > 2.00 m s−1 and RHice > 165%.
GCs were just beginning to develop at t = 80 min in the daytime radiation, initially neutral stability simulation, when the 99th percentile of w exceeded 0.50 m s−1 (Fig. 4e). The destabilization began slightly earlier than in the nighttime simulations because of the shortwave heating below cloud top (Fig. 16k). In the cross section, w exceeded 0.50 m s−1 for two isolated developing updrafts (Fig. 16b). These updrafts were associated with RHice maxima of 130%–140% (Fig. 16e), perturbations in qi near cloud top, and QN maxima of 0.50–0.75 K h−1 (Fig. 16k). Broad heating of 0.25–0.50 K h−1 below 6.0 km was the result of both shortwave warming and latent heating. Heating within developing updrafts was ~0.50–0.75 K h−1, and cooling above and between developing updrafts occasionally exceeded 0.25 K h−1 (Fig. 16k).
Convection strengthened by t = 140 min, with updrafts and downdraft magnitudes commonly > 1.00 m s−1 (Fig. 17b), weaker than for the corresponding nighttime radiation simulation. Accordingly, RHice and qi maxima were also lower with values of 145%–160% and 0.15–0.18 g kg−1 (Figs. 17e and 17h). Longwave cooling was offset by shortwave warming, thereby leading to weaker destabilization, based on the QN field, with cooling maxima from only 0.50 to 0.75 K h−1 common above GCs. Net heating of 1.00–1.50 K h−1 was present in GCs, with heating of 0.25–0.50 K h−1 common in fallstreaks where diffusional growth of ice was occurring.
In the initially neutral, no-radiation simulation, the w spectrum had very little range at t = 80 min, with a 99th percentile of 0.10 m s−1 (Fig. 5e), well below the 0.50 m s−1 contour interval in Fig. 16c. As a result, there was very little horizontal variability in the RHice field, with range of 105%–115% below 7.8 km and a layer of RHice ~ 125% near 7.9 km (Fig. 16f). There was also little horizontal variability in qi (Fig. 16i). Latent heating of ~0.25 K h−1 at 6.0–7.0 km increased to ~0.50 K h−1 at 5.0–6.0 km, deeper in the stratiform cloud layer (Fig. 16l).
The w spectrum range increased slightly by t = 140 min to 0.20 m s−1 (Fig. 5e), but this was still well below the 0.50 m s−1 contour interval in Fig. 17c. Variation in the RHice field had increased somewhat, with maxima up to 130%. Weak perturbations were present in the qi field, but no closed contours in qi were present, nor were any fallstreaks evident (Fig. 17i). Latent heating within the qi perturbations at cloud top was only ~0.25 K h−1 (Fig. 17l). By t = 180 min in the initially neutral, no-radiation simulation, the 1st–99th-percentile w range increased to ±1.00 m s−1 (Fig. 5e). While this range in w is slightly larger than when GCs were present late in the initially neutral daytime radiation simulation (Fig. 4e), the majority of updrafts and downdrafts were outside of precipitation.
The presence of these updrafts and downdrafts under conditions where destabilization is not favored is clarified by examining the 7.5-km plan view of qi at t = 180 min in the context of other simulations (Fig. 18f). Maxima in qi are generally less than 0.05 g kg−1 and are organized into bands oriented northwest–southeast. These bands are normal to the 6.3–7.8-km shear vector (black arrow in Fig. 18f), consistent with the presence of gravity waves and not buoyancy-driven convection. In Part I, upper-tropospheric waves formed in the unstable no-radiation simulation. This occurred as a result of prescribed ascent that was stronger in the upper troposphere than it was in the lower stratosphere, which compressed the tropopause region and increased upper-tropospheric shear. Gravity waves do not form in the simulations where instability and radiative forcing are present—for example, with stability profile b and nighttime radiation (Fig. 18a). Convection in that simulation mixed the upper troposphere and reduced the wind shear, whereas lack of convection through much of the neutral no-radiation simulation allowed shear to increase to the point where gravity waves were favored (Fig. 19). Mixing also decreased as convection weakened in the potentially unstable no-radiation simulations, resulting in vertical wind shear profiles only slightly weaker than the initially neutral no-radiation simulation, where gravity waves clearly developed. Given the favorable shear environment for waves, the spread in the w spectra late in the initially unstable no-radiation simulations was also consistent with the presence of gravity waves (see Figs. 5a–d).
Ice precipitation mixing ratio plan views at 7.5 km for t = 180 min. Panels are shown for simulations with all combinations of radiative forcing and stability profiles b (unstable), e (neutral), and h (stable). (f) The 6.3–7.8-km shear vector for t = 180 min in the neutral no-radiation simulation is shown. (i) Given the lack of GCs in the stable, no-radiation simulations, qi did not exceed 0.01 g kg−1 at 7.5 km.
Citation: Journal of the Atmospheric Sciences 73, 4; 10.1175/JAS-D-15-0127.1
The V wind component at t = 180 min for the nighttime radiation, stability profile b simulation (blue line), the no-radiation, stability profile e simulation (red line), and the no-radiation, stability profiles a–d simulations (black lines).
Citation: Journal of the Atmospheric Sciences 73, 4; 10.1175/JAS-D-15-0127.1
3) Stable simulations
In the nighttime radiation, stability profile h simulation, variation in the w field was less than 0.10 m s−1 at 6.0–8.0 km at t = 80 min (below the lowest nonzero w contour in Fig. 20a), when the w spectrum began to diverge from 0 m s−1 (Fig. 3h). Horizontal variability was nearly nonexistent in the qi field (Fig. 20g), and the RHice field had low variability near 6.0–6.5 km (~110–115%; Fig. 20d). Broad weak cooling was present above ~6.3 km, with the strongest cooling > 0.25 K h−1 from 6.8 to 7.2 km (Fig. 20j).
(a)–(c) Vertical air velocity, (d)–(f) relative humidity with respect to ice, (g)–(i) ice precipitation mixing ratio, and (j)–(l) net diabatic heating rate at t = 80 min for simulations with stability profile h (see Fig. 1) and (left)–(right) nighttime, daytime, and no radiation.
Citation: Journal of the Atmospheric Sciences 73, 4; 10.1175/JAS-D-15-0127.1
The effect of the net cooling at cloud top on destabilization is clearly evident later in the nighttime simulation, where the 99th percentile of w increased to 1.40 m s−1 at t = 140 min (Fig. 3h). Individual updrafts and downdrafts typically exceeded vertical velocity magnitudes of 1.00 m s−1, and one updraft in the cross section exceeded 1.50 m s−1 (Fig. 21a). RHice ranged from 150% to 160% in GCs and was as low as 50% in downdrafts between them (Fig. 21d). High RHice and strong w supported rapid growth of ice crystals, with qi maxima typically > 0.15 g kg−1 and occasionally > 0.21 g kg−1, from which fallstreaks are evident (Fig. 21g). The QN consisted of (longwave) cooling of 1.25–1.75 K h−1 above GCs and (latent) heating of 1.75–2.50 K h−1 within most GCs, with heating > 4.00 K h−1 in one extreme case (Fig. 21j).
(a)–(c) Vertical air velocity, (d)–(f) relative humidity with respect to ice, (g)–(i) ice precipitation mixing ratio, and (j)–(l) net diabatic heating rate at t = 140 min for simulations with stability profile h (see Fig. 1) and (left)–(right) nighttime, daytime, and no radiation.
Citation: Journal of the Atmospheric Sciences 73, 4; 10.1175/JAS-D-15-0127.1
Cross-sectional analysis indicated very similar structure in the w, RHice, and qi fields at t = 80 min in the daytime radiation initially stable simulation (Figs. 20b, 20e, and 20h), with no GCs present and w weaker than the lowest nonzero contour. Net cooling was weaker in the daytime simulation because of inclusion of shortwave heating; however, net cooling of 0.00–0.25 K h−1 was present near cloud top between 6.8 and 7.2 km. Shortwave heating penetrated deeper into the cloud on average (Fig. 8h), which enhanced net warming below the cooling and favored destabilization.
This destabilization led to development of GCs by t = 140 m, when the 99th percentile of 6.0–8.0 km w was 0.60 m s−1 (Fig. 4h). GCs typically had w maxima > 0.50 m s−1 and RHice maxima of 135%–140%, which supported development of qi maxima > 0.12 g kg−1 (Figs. 21b, 21e, and 21h). Even though maxima in these fields were less pronounced than in other simulations with GCs, fallstreaks were clearly evident emanating from GCs. Net diabatic heating within GCs of ~1.25 K h−1 was common, with weak cooling above GCs and cooling > 0.50 K h−1 in between and occasionally near the top of GCs (Fig. 21k).
As discussed earlier, the w spectrum had very little range for the entire stable, no-radiation simulation, with the 99th percentile of 6.0–8.0-km w remaining below 0.10 m s−1 (Fig. 5h). The lack of GCs is also clear in the cross sections of w, RHice, and qi at both t = 80 and 140 min (Figs. 20c,f,i and Figs. 21c,f,i), where horizontal variation of these fields is minimal. Latent heating within the stratiform cloud deck (Figs. 20l and 21l) was not sufficient to destabilize cloud top under initially stable conditions.
4. Discussion
Cloud-top precipitation GCs were first observed in the early 1950s (Marshall 1953). As noted in the introduction, in the more than six decades since their discovery, subsequent research has established their typical dimension, kinematics, and microphysical properties. While several studies have noted their presence in layers characterized by potential instability (Wexler and Atlas 1959), and some recent studies have suggested that radiative forcing may favor their maintenance (Syrett et al. 1995; Kumjian et al. 2014; Rauber et al. 2014a,b), these hypotheses remained untested. Part I assessed the influence of radiative forcing on GC maintenance in idealized WRF simulations with shear and potential instability conditions representative of those in the 14–15 February 2010 cyclone, which was observed during the PLOWS field campaign. Under those conditions, radiative forcing was required for maintenance of GCs. In this paper, the influence of radiative forcing on cloud-top destabilization and subsequent development, maintenance, or absence of GCs was assessed under a wide range of cloud-top stability profiles (Fig. 1) under the same shear conditions as in Part I (Fig. 2).
Under potentially unstable conditions, GCs developed early in simulations regardless of radiative forcing. Development of GCs was manifested as an increase in the range of the w spectrum (Figs. 3–5) and domain-averaged increases in the peak longwave cooling rate at cloud top (Figs. 6 and 7) and latent heating rate within GCs (Figs. 9–11). On the scale of individual GCs, high ice supersaturation (RHice maxima in GCs frequently > 150%) and latent heating > 2.00 K h−1 were consistent with rapid depositional growth of ice crystals, as noted in observations by Plummer et al. (2014). Ice precipitation mixing ratio maxima within GCs were commonly > 0.15 g kg−1, with fallstreaks emanating from their bases that seeded the underlying stratiform cloud deck as described in Plummer et al. (2015). The maintenance of GCs and spread of their associated w spectra through the remainder of the simulations depended on the radiative forcing parameterized by the model. Nighttime radiative forcing favored the strongest and most persistent GCs, since the continuous destabilization at cloud top due to longwave cooling was not offset by shortwave heating. The magnitude of cloud-top shortwave heating was not large enough to completely offset longwave cooling, as evidenced by the persistence of GCs, albeit weaker, in the daytime simulations. Greater instability favored more vigorous convection and higher qi, which led to more focused longwave cooling at cloud top, and thereby maintenance of greater radiatively forced instability later in the simulations. This positive feedback was particularly evident late in the evolution of the w spectra in nighttime radiation simulations, when the spread of w was greater for simulations with larger initial instability. In simulations with initial instability and no radiative forcing, GCs weakened substantially following the initial release of potential instability. Compression of the layer in the vicinity of the tropopause late in the simulations, an artifact of the prescribed-ascent method, paired with weaker mixing of momentum near the tropopause in the absence of GCs, favored an increase in both shear and stability. This resulted in development of gravity waves that maintained a spread in the w spectrum later in the simulation even in the absence of GCs.
The development or absence of GCs was highly dependent on radiative forcing in the initially neutral and stable simulations. When radiative forcing was present, longwave cooling at cloud top was sufficient to destabilize the atmosphere near cloud top even in the most stable simulations (Fig. 1, profile h) to the point where GCs developed. Convection developed later and the w spectrum ranges were narrower in the simulations with greater initial stability, since there was more convective inhibition to overcome (Figs. 3, 4, 17a,b, and 21a,b). As in the initially potentially unstable simulations, the w spectra had greater range with nighttime radiation, when shortwave heating did not offset cloud-top destabilization through longwave cooling. GCs did not develop under initially neutral or stable conditions in the absence of radiative forcing (Figs. 5, 17c, and 21c). The spread in the w spectrum late in the neutral no-radiation simulation was due to the development of gravity waves in the layer near the tropopause, as discussed earlier.
5. Summary
In this paper, the sensitivity of generating cells (GCs) to upper-tropospheric stability under conditions of different radiative forcing was assessed using idealized Weather Research and Forecasting (WRF) Model, version 3.3.1, simulations. Input model fields near cloud top were modified from a base sounding (see Part I) representative of the comma-head region of a midlatitude winter cyclone, so that the evolution or absence of GCs could be investigated under potentially unstable, neutral, and stable initial conditions with nighttime, daytime, and no radiative forcing. Analysis of these simulations has resulted in the following conclusions:
Generating cells develop in the presence of cloud-top potential instability and persist when radiative forcing is present.
Under neutral and even stable cloud-top conditions, radiative forcing will destabilize the cloud top and generating cells will develop. This result provides a physical explanation for the ubiquity of generating cells in field observations.
Generating cells consist of stronger updrafts and higher ice precipitation mixing ratios at night, when destabilization due to longwave cooling is not offset by shortwave warming.
Generating cells do not develop in the absence of radiative forcing unless cloud-top potential instability is present. In the case when potential instability is present, the generating cells will not persist after the potential instability is exhausted.
The above results demonstrate the crucial role of radiative forcing in development and maintenance of generating cells under the shear conditions present in the 14–15 February 2010 cyclone. The third paper in this series will explore the influence of wind shear on the dynamics of GCs.
Acknowledgments
Collaborations with National Center for Atmospheric Research coauthors were made possible through an NCAR Advanced Study Program Graduate Visitor Program fellowship received by the lead author. Code for the WRF-LES module was provided by Takanobu Yamaguchi of NOAA/ESRL. Funding for this research was provided by NSF Grants ATM-0833828 and AGS-1247404 to the University of Illinois. All simulations were run on the Stampede supercomputer with support from XSEDE Grant TG-ATM050014N.
REFERENCES
American Meteorological Society, 2015: Generating cell. Glossary of Meteorology. [Available online at http://glossary.ametsoc.org/wiki/Generating_cell.]
Carbone, R. E., and A. R. Bohne, 1975: Cellular snow generation—A Doppler radar study. J. Atmos. Sci., 32, 1384–1394, doi:10.1175/1520-0469(1975)032<1384:CSGDRS>2.0.CO;2.
Cunningham, J. G., and S. E. Yuter, 2014: Instability characteristics of radar-derived mesoscale organization modes within cool-season precipitation near Portland, Oregon. Mon. Wea. Rev., 142, 1738–1757, doi:10.1175/MWR-D-13-00133.1.
Douglas, R. H., K. L. S. Gunn, and J. S. Marshall, 1957: Pattern in the vertical of snow generation. J. Meteor., 14, 95–114, doi:10.1175/1520-0469(1957)014<0095:PITVOS>2.0.CO;2.
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, doi:10.1175/JAS3547.1.
Keeler, J. M., B. F. Jewett, R. M. Rauber, G. M. McFarquhar, R. M. Rasmussen, L. Xue, C. Liu, and G. Thompson, 2016: Dynamics of cloud-top generating cells in winter cyclones. Part I: Idealized simulations in the context of field observations. J. Atmos. Sci., 73, 1507–1527, doi:10.1175/JAS-D-15-0126.1.
Kumjian, M. R., S. A. Rutledge, R. M. Rasmussen, P. C. Kennedy, and M. Dixon, 2014: High-resolution polarimetric radar observations of snow-generating cells. J. Appl. Meteor. Climatol., 53, 1636–1658, doi:10.1175/JAMC-D-13-0312.1.
Langleben, M. P., 1956: The plan pattern of snow echoes at the generating level. J. Meteor., 13, 554–560, doi:10.1175/1520-0469(1956)013<0554:TPPOSE>2.0.CO;2.
Marshall, J. S., 1953: Precipitation trajectories and patterns. J. Atmos. Sci., 10, 25–29, doi:10.1175/1520-0469(1953)010<0025:PTAP>2.0.CO;2.
Plummer, D. M., G. M. McFarquhar, R. M. Rauber, B. F. Jewett, and D. C. Leon, 2014: Structure and statistical analysis of the microphysical properties of generating cells in the comma head region of continental winter cyclones. J. Atmos. Sci., 71, 4181–4203, doi:10.1175/JAS-D-14-0100.1.
Plummer, D. M., G. M. McFarquhar, R. M. Rauber, B. F. Jewett, and D. C. Leon, 2015: Microphysical properties of convectively generated fall streaks in the comma head region of continental winter cyclones. J. Atmos. Sci., 72, 2465–2483, doi:10.1175/JAS-D-14-0354.1.
Rauber, R. M., and Coauthors, 2014a: Stability and charging characteristics of the comma head region of continental winter cyclones. J. Atmos. Sci., 71, 1559–1582, doi:10.1175/JAS-D-13-0253.1.
Rauber, R. M., M. K. Macomber, D. M. Plummer, A. A. Rosenow, G. M. McFarquhar, B. F. Jewett, D. Leon, and J. M. Keeler, 2014b: Finescale radar and airmass structure of the comma head of a continental winter cyclone: The role of three airstreams. Mon. Wea. Rev., 142, 4207–4229, doi:10.1175/MWR-D-14-00057.1.
Rauber, R. M., and Coauthors, 2015: The role of cloud-top generating cells and boundary layer circulations in the finescale radar structure of a winter cyclone over the Great Lakes. Mon. Wea. Rev., 143, 2291–2318, doi:10.1175/MWR-D-14-00350.1.
Rosenow, A. A., D. M. Plummer, R. M. Rauber, G. M. McFarquhar, B. F. Jewett, and D. Leon, 2014: Vertical velocity and physical structure of generating cells and convection in the comma head region of continental winter cyclones. J. Atmos. Sci., 71, 1538–1558, doi:10.1175/JAS-D-13-0249.1.
Syrett, W. J., B. A. Albrecht, and E. E. Clothiaux, 1995: Vertical cloud structure in a midlatitude cyclone from a 94-GHz radar. Mon. Wea. Rev., 123, 3393–3407, doi:10.1175/1520-0493(1995)123<3393:VCSIAM>2.0.CO;2.
Wang, Z., and Coauthors, 2012: Single aircraft integration of remote sensing and in situ sampling for the study of cloud microphysics and dynamics. Bull. Amer. Meteor. Soc., 93, 653–668, doi:10.1175/BAMS-D-11-00044.1.
Wexler, R., 1955: Radar analysis of precipitation streamers observed 25 February 1954. J. Atmos. Sci., 12, 391–393, doi:10.1175/1520-0469(1955)012<0391:RAOPSO>2.0.CO;2.
Wexler, R., and D. Atlas, 1959: Precipitation generating cells. J. Atmos. Sci., 16, 327–332, doi:10.1175/1520-0469(1959)016<0327:PGC>2.0.CO;2.
Wood, R., 2012: Stratocumulus clouds. Mon. Wea. Rev., 140, 2373–2423, doi:10.1175/MWR-D-11-00121.1.
Yamaguchi, T., and G. Feingold, 2012: Technical note: Large-eddy simulation of cloudy boundary layer with the Advanced Research WRF model. J. Adv. Model. Earth Syst., 4, M09003, doi:10.1029/2012MS000164.
The term “generating cell” describes a small region of locally high radar reflectivity at cloud top from which an enhanced reflectivity trail characteristic of falling snow particles originates (American Meteorological Society 2015).