1. Introduction
The snowfall distribution within the comma-head region of cool season extratropical cyclones can vary widely depending on the presence, intensity, and duration of bands of heavy precipitation, or snowbands. Some of these intense bands can produce 75–100 mm h−1 snowfall rates (Novak et al. 2004, 2008), thus causing major societal impacts, especially in the urban areas of the mid-Atlantic and Northeast United States (Picca et al. 2014). Even small errors in the location and intensity of these bands can result in large errors in the location and intensity of the resulting snowfall. It is thus important to understand snowband physical processes to help improve numerical weather prediction models and improve quantitative precipitation forecast (QPF) skill for winter storms.
Precipitation bands are common features within extratropical cyclones (e.g., Houze et al. 1976; Sanders and Bosart 1985; Geerts and Hobbs 1991; Jurewicz and Evans 2004). There is a spectrum of snowband structures within the comma head of extratropical cyclones (Novak et al. 2004), ranging from the primary band (Novak et al. 2008), typically located a few hundred kilometers to the northwest of the surface cyclone, to multibands that often exist to the north and northeast of the surface cyclone (Ganetis et al. 2018). Ingredients for the primary band have been well documented (e.g., Novak et al. 2004), which is defined as a radar reflectivity feature > 250 km in length, 20–100 km in width, and with intensities of >30 dBZ maintained for at least 2 h. Multibands have been defined as two or more midsized bands (length < 200 km, width 10–50 km, and ratio of width to length < 0.5) existing parallel to each other (Ganetis et al. 2018). Climatological studies have shown that bands that do not meet these primary band criteria tend to have weaker frontogenesis or a shallower frontal slope with stronger stability (Novak et al. 2004, 2008), while multibands can develop with little or no frontogenesis (Ganetis et al. 2018). The primary band often develops as the cyclone matures and when there is a well-defined deformation zone and frontogenesis around 600–700 hPa (Novak et al. 2010), which leads to an ageostrophic circulation and a relatively narrow (<50 km) band of snow that is often a few hundred kilometers long. A variety of instabilities exist for these bands, from conditional instability (CI; Trapp et al. 2001), conditional symmetric instability (CSI; Schultz and Schumacher 1999), inertial instability (II; Schultz and Knox 2007), and weakly stable environments (Novak et al. 2010). Theoretical studies have shown that this band becomes narrower and more intense as the stability weakens, with multibands more favored as either vertical or slantwise instability occurs.
Convective-allowing models can struggle to predict snowbands given the large sensitivity of band quantity and intensity to model initial conditions and physics (Connelly and Colle 2019). One challenge is forecasting the genesis of these bands. Ganetis et al. (2018) used an object-oriented approach to identify snowbands, and the environments were then diagnosed around the bands from multiple events along the U.S. East Coast. They found that over half of the bands did not have any well-defined frontogenesis. Therefore, other mechanisms may be possible, such as elevated convection or generating cells (Kumjian et al. 2014; Plummer et al. 2014, 2015; Rauber et al. 2017; Rosenow et al. 2014; Keeler et al. 2017).
Microphysical processes are also important for these snowbands. For several coastal winter storms over the western Atlantic impacting Long Island, New York, Colle et al. (2014) showed large variations in the degree of riming across a primary band ranging from more rimed dendrites on the upwind (east) side, where there were large vertical motions aloft, to less rimed dendrites and plates on the downwind (west) side of the band. Lackmann and Thompson (2019) showed that elevated convection in highly sheared flow can concentrate the snow hydrometeors aloft before they fall out within the band. Other recent studies of winter storms, such as the Olympic Mountains Experiment (OLYMPEX) in Washington State (Naeger et al. 2020) and studies in the Northeast United States (Molthan et al. 2016), have shown that models have difficulties predicting the correct amount of riming, which impacts the snow fallout and precipitation rate.
More observations are needed to better understand snowbands in the cyclone comma head and improve their predictions. The Investigation of Microphysics and Precipitation for Atlantic Coast-Threatening Snowstorms (IMPACTS) is a NASA-sponsored field campaign to study East Coast snowstorms (McMurdie et al. 2022) with the expressed goals of understanding the multiple processes contributing to snowband structures and to apply this understanding to improve numerical modeling of snowfall. The IMPACTS observational strategy includes a satellite-simulating aircraft, the ER-2, that flies above storms with a suite of remote sensing instruments including cloud and precipitation radars, lidar, and passive microwave radiometers and an in situ aircraft, the P-3, outfitted with instrumentation measuring microphysical and environmental quantities within clouds.
On 7 February 2020 the IMPACTS team sampled a heavy snow event across west-central New York State associated with a well-defined cyclone that tracked just east of this region. Between 0000 UTC 7 February and 0000 UTC 8 February 2020, 20–40 cm of snow fell across west-central New York State, with much of this falling between 1200 and 1800 UTC 7 February (Fig. 1). Finlon et al. (2022) investigated the observed microphysics of this event using quadruple-frequency radar observations from aircraft, while Varcie et al. (2023) showed air mass origins in the comma head as well as the precipitation growth processes and particle types using aircraft in situ data. Both studies noted some key synoptic conditions, including an axis of sloping frontogenesis to the north of the surface cyclone that produced elongated areas of locally higher reflectivity as observed by the Next Generation Weather Radar (NEXRAD) Weather Surveillance Radar-1988 Doppler (WSR-88D) network. This swath of heavy snowfall occurred without the presence of any well-defined snowbands as defined by Novak et al. (2004) or Ganetis et al. (2018). The goals of this study are to explore whether the model can realistically simulate the precipitation evolution, and to compare model microphysics predictions against observations. We will explore the following questions:
What were the precipitation structures aloft and the microphysical variations within this heavy snow area?
How well did the Weather Research and Forecasting (WRF) Model simulate the large-scale environment, precipitation structures, and microphysics?
How did the WRF microphysical sensitivities compare to the initial condition uncertainty for the precipitation?
(top) Map of the study region showing the storm-total (24 h) snowfall analysis (color shaded; cm; from the NOAA National Snowfall Analysis) from 0000 UTC 7 Feb to 0000 UTC 8 Feb 2020, New York Mesonet site locations, rawinsonde locations, WSR-88D (NEXRAD) sites, and ER-2/P-3 flight track locations (see inset legend). (bottom) The times (UTC), altitudes (m MSL), and direction of the P-3 are shown.
Citation: Monthly Weather Review 151, 9; 10.1175/MWR-D-22-0296.1
2. Data and methods
a. Observational and reanalysis datasets
Several model analyses and reanalysis datasets were used for synoptic and regional context and as model initialization and boundary condition fields. The analyses used for the WRF Model initializations and boundary conditions include the European Center for Medium-Range Weather Forecasting (ECMWF) Reanalysis version 5 (ERA5; Hersbach et al. 2020), the National Centers for Environmental Prediction (NCEP) Global Forecasting System (GFS; National Centers for Environmental Prediction/National Weather Service/NOAA/U.S. Department of Commerce 2015a), and the Rapid Refresh analyses (RAP; Benjamin et al. 2016). The RAP, with 13-km horizontal grid spacing and 50 vertical levels, was used to compare with the WRF forecasts.
The New York State Mesonet (NYSM; Brotzge et al. 2020) data were used in this study to better identify the low-level frontal location and precipitation gradients. The NYSM is a network of 126 standard environmental monitoring stations across New York State with an average spacing of 27 km (Fig. 1). The NOAA National Snowfall Analysis (National Operational Hydrologic Remote Sensing Center 2004) was used to obtain the 24-h snowfall accumulation for this event (Fig. 1). Radar data from the WSR-88D radar network for the lowest elevation scan (0.5°) were composited and gridded across the northeastern United States (Fig. 1) (Tomkins et al. 2022) using the Python ARM Radar Toolkit (Py-ART; Helmus and Collis 2016) with a Cressman weighting function (Cressman 1959) to determine the reflectivity values on a 2-km grid. Enhanced reflectivity in the composited radar data can indicate both heavy precipitation of a single type (snow, rain) and mixed precipitation as particle sizes increase at the melting level (Rauber and Nesbitt 2018). To aid in identifying regions of heavy snowfall, we apply the image muting technique from Tomkins et al. (2022) to mask those reflectivity grid points where the radar correlation coefficient is less than 0.97. Lower correlation coefficient values suggest that the higher reflectivity values are due to melting and mixed precipitation (Straka et al. 2000), and thus muting them allows remaining regions of high reflectivity to be interpreted as heavier snowfall.
A variety of in situ and remote sensing data collected from aircraft during IMPACTS (McMurdie et al. 2022) were used for this study. The ER-2 flew above the storm at ∼20 km MSL and had three nadir-looking radars, including the ER-2 X-band (9.6 GHz) radar (EXRAD; Heymsfield et al. 2016). The EXRAD consists of a nadir-pointing beam and a conically scanning beam, with only the nadir beam used for this study to provide a cross section view of the cloud and precipitation echo. In situ microphysics data were obtained from the P-3 aircraft. The 2D Stereo (2D-S; Lawson et al. 2006) and High Volume Precipitation Spectrometer (HVPS; Lawson et al. 1998) optical array probes (OAPs) on the P-3 provided information on the sizes, shapes, and concentrations of particles ranging from 100 µm to 3 cm. OAP data were processed using the University of Illinois/Oklahoma OAP Processing Software (UIOOPS; McFarquhar et al. 2018) and particle size distributions (PSDs) and bulk microphysical properties obtained every 5 s of flight following the methodology outlined in Finlon et al. (2022). The two aircraft were coordinated (ER-2 above P-3) flying from west to east across the area of heavy precipitation (Fig. 1), with the P-3 flying at different altitudes (temperatures) to measure the microphysical properties. Meanwhile, in addition to routine radiosonde launches (0000 and 1200 UTC), there were supplemental radiosondes launched at 1500 UTC 7 February at Syracuse, New York, and Albany, New York (points SYR and ALY on Fig. 1).
b. Modeling approach
The Weather Research and Forecasting (WRF; Skamarock et al. 2019) Model version 4.0 was utilized to simulate this event, with 40-vertical sigma levels spaced from ∼100 m within the boundary layer to 500 m around echo top (7–8 km), and a quadruple one-way nesting configuration at 18-, 6-, 2-, and 0.67-km grid spacing centered over the eastern United States (Fig. 2). Some of the initial condition errors and model errors were explored by initializing the model with three different initial and boundary conditions derived from 6-hourly RAP, GFS, and ERA5 reanalyses. The goal was not to create a diverse or calibrated ensemble, but to include enough members to explore some of the initial condition (IC) uncertainty as compared to model microphysics uncertainty. The 0.67-km domain may not resolve some of the small-scale generating cells aloft, but there have been limited WRF runs at <200-m grid spacing for winter storms. Thus, this will be saved for future work. Thus, unless specified, most runs had an innermost domain with 2-km grid spacing.
GOES-16 infrared brightness temperature (shaded; K) and 500-hPa geopotential height analysis (black every 6 dam) from the Rapid Refresh (RAP) at 1500 UTC 7 Feb 2020. The locations of the 18- (d01), 6- (d02), 2- (d03), and 0.67-km (d04) WRF domains are shown by the colored boxes.
Citation: Monthly Weather Review 151, 9; 10.1175/MWR-D-22-0296.1
Since the RAP analysis does not contain surface variables, soil moisture and temperature were taken from the NCEP North American Mesoscale Forecast System (NAM; National Centers for Environmental Prediction/National Weather Service/NOAA/U.S. Department of Commerce 2015b) analysis, while the NCEP Real-Time Global (RTG) at 0.083° grid spacing was used for the sea surface temperatures (Thiébaux et al. 2003). The RAP analysis run was initialized at two different times (1800 UTC 6 February and 0000 UTC 7 February 2020) and run with two different microphysics schemes: 1) the Thompson scheme (Thompson et al. 2008), which includes separate categories for ice, snow, and graupel; and 2) the P3 microphysics scheme (Morrison and Milbrandt 2015), which uses one ice category to represent the three ice species. All other runs used the P3 microphysics scheme. More details of these two schemes can also be found in other recent papers that have also validated these two schemes [e.g., Table 2 in Naeger et al. (2020)]. One WRF run nested down to 0.67-km grid spacing and initialized with the RAP at 1800 UTC 6 February highlights the impact of additional resolution. The 18- and 6-km domains used the Grell–Dévényi convective scheme (Grell and Dévényi 2002). The MYNN 2.5-level TKE PBL scheme (Nakanishi and Niino 2009) with the Unified NOAH Land Surface model (Livneh et al. 2011) was used for all simulations. Table 1 summarizes the different configurations used for the six simulations of the event as well as the WRF run abbreviations used in the text below. The model evolution discussion focuses on the WRF-RAP18Z run based on its ability to simulate the forcing for ascent more realistically as will be highlighted below. Other runs are shown as needed to illustrate any important differences.
The various WRF runs showing the initial and boundary conditions and the microphysical parameterization used. The italicized run abbreviations refer to the nomenclature used in the text. Those with “18Z” and “00Z” in the run names were initialized on 6 and 7 Feb, respectively. The “control” run is bold for reference.
3. Storm structure and evolution
a. Pressure, thermodynamic, and precipitation evolution
At 1500 UTC 7 February 2020 (Fig. 2), there was a high-amplitude 500-hPa trough over the eastern United States with its axis centered over the Great Lakes. At the surface at 1200 UTC 7 February (Fig. 3a), the cyclone had deepened to 983 hPa and was located over southeast Pennsylvania. There was a warm frontal boundary to the east of the cyclone, while an inverted surface trough was to the north of the cyclone over eastern New York, associated with the 0°C freezing line and a relatively strong 2-m temperature gradient of 5°–7°C over 200 km to the west of this boundary. The 18-km WRF-RAP18Z realistically simulated these features, although the cyclone was 2–3 hPa too deep and centered ∼150 km north of the RAP analysis. The analyzed storm continued to deepen to 974 hPa and moved northeast into western Massachusetts by 1800 UTC 7 February (Fig. 3c). The pressure trough extended northward from the low into northern New York and Vermont, and surface temperatures remained below freezing to the west of the trough. The 18-km WRF cyclone was ∼4 hPa deeper than the analyzed low (Figs. 3c,d), with above-freezing temperatures extending 100–150 km further northward into New England than analyzed.
Surface maps for the (a) RAP analysis and (b) 18-km WRF-RAP18Z showing sea level pressure (solid every 2 hPa), 2-m temperature (shaded; °C), and 10-m wind barbs (full barb = 5 m s−1) at 1200 UTC 7 Feb 2020. (c),(d) As in (a) and (b), but for 1500 UTC 7 Feb 2020. (e),(f) As in (a) and (b), but for 1800 UTC 7 Feb 2020.
Citation: Monthly Weather Review 151, 9; 10.1175/MWR-D-22-0296.1
There were some variations and larger errors in the other WRF runs valid at 1500 UTC 7 February (not shown). The WRF-ERA18Z has a cyclone center of ∼970 hPa, which is ∼10 hPa deeper than observed. The WRF-GFS18Z has a minimum pressure of 976 hPa, but it is 3°–4°C warmer than observed around Albany, New York. The WRF-RAP00Z is also 2°–3°C warmer than observed over southeast New York. In contrast, the cyclone in the WRF-THOM18Z is within 1 hPa of the P3 scheme and with similar temperatures, so changing the microphysics has much less impact on cyclone location and surface temperatures than using different initial conditions.
A vertical profile of equivalent potential temperature at Syracuse and Albany at 1500 UTC 7 February shows a shallow potentially unstable layer just above the surface (Fig. 4) capped by a layer of high potential stability. At Syracuse (Fig. 4a), the stable layer between 950 and 800 hPa, near neutral layer from 800 to 700 hPa, and the weak stable layer from 700 to 400 hPa were all well simulated by the WRF-RAP18Z. Closer to the warm sector at Albany, WRF simulations agreed with observations of an unstable layer from 625 to 725 hPa, thus supporting more elevated convection. All WRF members were too warm by 2–3 K below 800 hPa at Albany (Fig. 4b).
Vertical profile of equivalent potential temperature (K) at (a) Syracuse, NY, and (b) Albany, NY, for the observed profile (black) and 2-km WRF members (see legend) at 1500 UTC 7 Feb 2020.
Citation: Monthly Weather Review 151, 9; 10.1175/MWR-D-22-0296.1
At 700 hPa at 1200 UTC 7 February, there was a trough from eastern Ontario to northern Ohio evident in the analysis and the WRF simulations (Figs. 5a,b). The 2D Petterssen frontogenesis (Petterssen 1956) was calculated, and then compared between the analysis and the WRF runs to assess any differences. Frontolysis existed along the trough axis, while there was frontogenesis to the east of the trough from central Pennsylvania northward toward western New York. By 1500 UTC 7 February (Figs. 5c,d), which is the time of the start of the coordinated P-3/ER-2 observations, the 700-hPa low was centered over northern Lake Ontario with a trough and associated frontogenesis oriented southwest–northeast across central New York. As this trough moved to eastern New York by 1800 UTC the frontogenesis had weakened as the flow increased on the east side of the midlevel cyclone (Figs. 5e,f).
700-hPa frontogenesis [color shaded; K (100 km)−1 (3 h)−1], winds (full barb = 10 kt; 1 kt ≈ 0.51 m s−1), geopotential heights (black lines every 3 dam), and temperatures (red lines every 5°C) for the (a) RAP analysis and (b) 18-km WRF-RAP18Z at 1200 UTC 7 Feb 2020. (c),(d) As in (a) and (b), but for 1500 UTC 7 Feb 2020. (e),(f) As in (a) and (b), but for 1800 UTC 7 Feb 2020.
Citation: Monthly Weather Review 151, 9; 10.1175/MWR-D-22-0296.1
The 700-hPa frontogenesis is shown for some other members at 1500 UTC 7 February (Fig. 6). Compared to the analysis (Fig. 5a), the WRF-GFS18Z (Fig. 6a), WRF-ERA18Z (Fig. 6b), WRF-RAP00Z (Fig. 6c), and WRF-THOM18Z (Fig. 6d) overpredict either the intensity or coverage of the frontogenesis at this time, and thus likely overpredict the forcing for ascent. The frontogenesis is too strong in these runs because they either over-deepen the cyclone or are too warm to the east of the surface cyclone (not shown). The WRF-THOM00Z (Fig. 6e) is similar to the WRF-RAP00Z run using the P3 scheme (Fig. 6c), thus the WRF-THOM18Z more realistically predicts the frontogenesis than the other runs. Therefore, for much of the remainder of this paper, the analysis will focus on the WRF initialized at 1800 UTC using the P3 and Thompson microphysics schemes, but other members are sometimes shown to illustrate how these ICs impact the cloud structure, microphysics, and surface precipitation.
As in Fig. 5c at 1500 UTC 7 Feb, but for the (a) WRF-GFS18Z, (b) WRF-ERA18Z, (c) WRF-RAP00Z, (d) WRF-THOM18Z, and (e) WRF-THOM00Z simulations.
Citation: Monthly Weather Review 151, 9; 10.1175/MWR-D-22-0296.1
Figure 7 shows the observed radar mosaic evolution and model-derived reflectivity at 1.0 km MSL for the WRF-RAP18Z from 1200 to 1800 UTC 7 February. The observed radar reflectivity field at 1200 UTC (Fig. 7a) has an enhanced area of 30–40 dBZ due to heavy snow from northern New York to south-central New York, while a broad area of lighter snow (<30 dBZ) is evident over western New York. The axis of heaviest precipitation is collocated with the region of 700-hPa frontogenesis (Fig. 5a), but there is no well-defined snowband at this time. Many enhanced precipitation features are amorphous and transient (not shown), and thus do not fit the 2-h band duration criteria by Novak et al. (2004). The area of enhanced reflectivity expanded by 1500 UTC across central New York State (Fig. 7c), but there was still no concentrated snowband structure, and then it shifted to eastern New York and began weakening by 1800 UTC (Fig. 7e). Although the WRF-RAP18Z run evolution is similar to the observations, the precipitation coverage is somewhat less than observed at 1200 UTC (Fig. 7b), is weaker than observed over western New York State at 1500 UTC (Fig. 7d), and is stronger than observed over southeast New York at 1800 UTC (Fig. 7f).
Reflectivity (shaded; dBZ) for the (a) 0.5° composite radar analysis and (b) 2-km WRF-RAP18Z at 1.0 km MSL for 1200 UTC 7 Feb 2020. Sea level pressure is also shown (black lines every 2 hPa). (c),(d) As in (a) and (b), but for 1500 UTC 7 Feb 2020. (e),(f) As in (a) and (b), but for 1800 UTC 7 Feb 2020. Observed reflectivity values are masked in gray where dual-polarization correlation coefficients < 0.97.
Citation: Monthly Weather Review 151, 9; 10.1175/MWR-D-22-0296.1
b. Cross sections and frontogenesis structure
The ER-2 aircraft flew from west to east at 20 km MSL across the heavy precipitation area over New York between 1500 and 1600 UTC (Fig. 1). Figure 8 shows the reflectivity from the nadir EXRAD (X-band) radar beam for two periods between 1508 and 1613 UTC 7 February. Between 1508 and 1524 UTC (Fig. 8a), there is a deep cloud with 6–8 km MSL echo top heights across the region. There are narrow echo top generating cells above 6 km, and the sharp increase in reflectivity indicates snow growth from 6 to 4 km MSL. There is a broad area of precipitation enhancement (30–40 dBZ) from 75.5° to 74.5°W (∼100 km), but no concentrated snowband. The fall streaks extend downward from echo top (∼8 km) down to 2–3 km MSL from 74° to 75°W, but below this level and sloping upward to the west these streaks are less apparent and appear to be deformed from vertical wind shear along and normal to the cross section. The melting layer is apparent around 2 km MSL over the eastern cross section, and then rapidly descends to the surface around 74.3°W. From 1550 to 1613 UTC (Fig. 8c), similar structures were apparent, but the echo top is about 5 km MSL over the eastern part as the dry intrusion starts to enter this region.
Reflectivity (shaded; dBZ) for the (a) ER-2 at 1508–1524 UTC 7 Feb and (b) 2-km WRF-RAP18Z at 1515 UTC 7 Feb 2020. (c),(d) As in (a) and (b), but for 1550–1613 UTC for the ER-2 and 1545 UTC for the 2-km WRF-RAP18Z. The locations for the early and later cross-section times are given by the green E1 and E2 lines, respectively, in Fig. 1. The thick bright red line between 0 and 0.5 km in (a) and (c) represents the ground.
Citation: Monthly Weather Review 151, 9; 10.1175/MWR-D-22-0296.1
The WRF-RAP18Z realistically simulated the upper extent of the precipitation-sized particles (as defined by the ER-2 X-band radar) over the western two-thirds of the domain at both of these times (Figs. 8b,d), but it was too dry over the eastern part of the cross section as a result of the midlevel dry air intrusion moving too far north in the model (not shown). The 2-km WRF also does not resolve the echo top generating cells, but does capture the sharp increase in reflectivity from 4 to 6 km MSL as well as the broad area of precipitation enhancement over the central two-thirds of the section.
Figure 9 compares model reflectivity from the higher resolution 0.67-km WRF-RAP18Z run and other ensemble members in Table 1 at 1545 UTC 7 February, which corresponds to the E1 (1508–1524 UTC) ER-2 leg shown in Fig. 8a. The 0.67-km WRF-RAP18Z better resolves the echo-top generating cells as compared to the 2-km WRF-RAP18Z run (Figs. 9a and 8b). However, to simulate the smaller cells would require grid spacings of ∼100 m as in idealized WRF runs by Keeler et al. (2016a,b, 2017). Some of the vertical motions in the convective cores are 15%–20% larger (1–2 m s−1) in the 0.67-km run than the 2-km run (not shown). The WRF-THOM18Z run has greater precipitation coverage over the eastern part of the leg (Fig. 9b), which is more similar to the observations (Fig. 8a), but the intense reflectivity values (>25 dBZ) are more shallow than the WRF-RAP18Z run and the observations (Figs. 8a,b). In contrast, the depth of the >25-dBZ echo is deeper than observed in the WRF-ERA18Z run over the western part of the domain. The echo top in the WRF-RAP00Z and WRF-ERA18Z runs is too deep to the west and more intense over the central part of the section than observed (Figs. 9c,d and 8a).
Model reflectivity at 1545 UTC (shaded; dBZ) corresponding to the 1508–1524 UTC 7 Feb ER-2 leg for the (a) 0.67-km WRF-RAP18Z, (b) WRF-THOM18Z, (c) WRF-ERA18Z, and (d) WRF-RAP00Z members.
Citation: Monthly Weather Review 151, 9; 10.1175/MWR-D-22-0296.1
At 1500 UTC 7 February, a cross section along the E1 flight leg of the ER-2 (cf. Fig. 1) shows two sloping frontal zones in both the RAP analysis and WRF-RAP18Z (Figs. 10a,b). In the analysis (Fig. 10a), there is a shallow stationary frontal zone from 900 to 800 hPa, and a frontal zone sloping upward from 800 to 650 hPa to the west, labeled as a warm frontal zone in Varcie et al. (2023). Backward trajectory analysis of Varcie et al. (2023) also noted that air below the stationary front largely remained at low levels and experienced very little vertical motion along its pathway toward the low. There are separate regions of frontogenesis and associated stable layers with both frontal zones. The WRF-RAP18Z has both frontal zones merged to the east and the zone is 50–100 hPa shallower than the analysis (Fig. 10b), but two frontal zones are distinct at the midpoint of the flight leg. There is a region of negative saturated moist potential vorticity (MPV*; where the trailing asterisk here and elsewhere denotes evaluation of a quantity under saturated conditions) from around 500 to 350 hPa over western parts of the section in both the analysis and WRF-RAP18Z (Figs. 10a,b), while negative MPV* also exists from 500 to 700 hPa in the eastern part of the section in the WRF-RAP18Z run only. The cross section along the E2 ER-2 flight leg shows that this instability extends down to 700 hPa to the east in both the analysis and the WRF-RAP18Z run (Figs. 10c,d). Much of these negative MPV* areas are associated with a decrease in saturated equivalent potential temperature (thetaE*) with height, indicating CI. There are a few regions in which this is not the case in the analysis and WRF run, such as in the analysis around 500–400 hPa over the eastern one-third of the leg. There is no potential instability (PI) or CI in this region, but there is negative MPV*, which implies conditional symmetric instability (CSI) in this region in the WRF and analysis. This is consistent with the convective turrets and plumes, which are lower and more intense over eastern portions of this leg. The CI or PI instability was present for potential banding, but unlike past case studies with a well-defined primary band, the frontogenesis for this event slopes upward from 900 to 600 hPa over a few hundred kilometers, rather than concentrated over a 50–100-km region [cf. Fig. 12 in Novak et al. (2008)]. We hypothesize that this lack of concentrated frontogenesis is the reason for the lack of primary band in this event, although this is an issue for future research by comparing with other events.
Cross section showing frontogenesis [warm color shaded; K (100 km)−1 (3 h)−1], saturated equivalent potential temperature (red lines every 2 K), and regions of negative saturated moist potential vorticity (MPV*; shaded in blue using legend) for the (a) RAP analysis and (b) 18-km WRF-RAP18Z at 1500 UTC 7 Feb 2020 for the E1 track shown in Fig. 1. (c),(d) As in (a) and (b), but for 1500 UTC 7 Feb 2020 for the E2 track shown in Fig. 1.
Citation: Monthly Weather Review 151, 9; 10.1175/MWR-D-22-0296.1
Further analysis of the structure of this frontal zone and WRF validation is shown by the P-3 flight-level temperature measurements versus the WRF-member temperatures for the legs above and within the frontal zone at 5.1 and 3.6 km MSL, respectively (Fig. 11). The temperature was relatively constant (−15.5°C) for the western two-thirds of the 5.1-km flight leg (Fig. 11a), and then it increased for the easternmost part of this leg as also evident in the cross sections (Figs. 10c,d). Many of the WRF members are too cold at this altitude, such as in the WRF-RAP18Z with a 2–2.5°C cold bias. This frontal feature is too fast in the WRF, as evinced by its lack of precipitating cloud at midlevels observed over eastern parts of this leg (cf. Fig. 8b). From west to east at 3.6 km (Fig. 11b), there is an increase in temperature within the frontal zone from −12° to −9°C, then a slight decrease to −10°C midway through the leg, and then another sharp increase from −10° to −6°C toward the east. Many of the WRF members realistically simulated the 3°–4°C increase over the western part of the leg; however, many members are 1.5°–2.5°C too warm in this western region, while none of the members simulated the rapid temperature increase with the front further to the east.
Time series showing temperature (°C) for the observed (black) and the various 2-km WRF members (see bottom legend) for the (a) E1 (at 5.1 km from 1508 to 1524 UTC) and (b) E2 (at 3.6 km from 1530 to 1551 UTC) flight legs for the P-3 aircraft.
Citation: Monthly Weather Review 151, 9; 10.1175/MWR-D-22-0296.1
c. Microphysical and precipitation validation
Figure 12 shows the dry snow, rimed snow or graupel, and liquid (rain and cloud) water mixing ratios for four of the WRF simulations along the same cross sections in Figs. 9 and 10 and illustrates the impact of model resolution (WRF-RAP18Z versus 0.67-km WRF-RAP18Z), microphysics parameterization (WRF-RAP18Z versus WRF-THOM18Z), and initial/boundary conditions (WRF-RAP18Z versus WRF-ERA18Z). The greatest rimed snow (>0.6 g kg−1) is located within the convective plumes, which are better resolved and more intense in the 0.67-km WRF-RAP18Z run (Fig. 12b) as compared to the 2-km WRF-RAP18Z run (Fig. 12a). Within some of the convective cells and precipitation to the east there is a mix of rimed snow and liquid as high as 5 km (∼−15°C) in all the runs. This additional water is likely from the lack of seeder ice aloft as the midlevel dry intrusion moved into this region. Over the western half of this section there was primarily snow aloft but little rimed snow or graupel. In general, there is more rimed snow or graupel in the WRF-RAP18Z than the WRF-THOM18Z, especially in the convective cell region. The snow in the WRF-ERA18Z cloud is also deeper and more intense than the WRF-RAP18Z, which is likely because of the stronger frontogenesis using the ERA5 IC/BCs (Fig. 6b).
WRF cross section along E1 valid at 1515 UTC 7 Feb showing mixing ratios of snow (yellow every 0.3 g kg−1), rimed snow or graupel (purple every 0.15 g kg−1), rain and cloud water combined (g kg−1; shaded green using color bar) for the (a) WRF-RAP18Z, (b) 0.67-km WRF-RAP18Z, (c) WRF-ERA18Z, and (d) WRF-THOM18Z. The dashed blue lines in (a) and (b) represent the P-3 aircraft flight tracks and times (UTC).
Citation: Monthly Weather Review 151, 9; 10.1175/MWR-D-22-0296.1
The amount of supercooled liquid water (SCW) estimated from the P-3 Cloud Droplet Probe (CDP) and the presence of riming, as determined by oscillating frequency changes of the P-3 Rosemount Icing Detector (RICE), was compared to the WRF runs. Figure 13a shows a time series of SCW for the 5.1 km MSL flight leg, where the temperatures were ∼−15°C for much of this leg (Fig. 11a). There is little or no cloud water observed over much of this leg, with the exception of a few spikes to 0.1–0.3 g m−3 to the east, likely associated with some of the convective plumes in this region. All the WRF runs predicted the lack of SCW over the western one-third of this 5.1-km leg; however, the WRF-ERA18Z has spikes to 0.9–1.1 g m−3 that are too large over the eastern half of this leg. The results are similar for the 3.6-km P-3 flight leg, where observed temperatures increased from −12° to −7°C from west to east and little SCW was observed except for the easternmost part of the leg which had spikes to 0.35 g m−3 (Fig. 13b). However, there is more overprediction of SCW by more WRF members near the middle of the leg, and larger overprediction by the WRF-ERA18Z and WRF-RAP00Z further to the east.
As in Fig. 11, but for supercooled water (SCW; g m−3).
Citation: Monthly Weather Review 151, 9; 10.1175/MWR-D-22-0296.1
The riming also varied from west to east across the 3.6-km leg, which was noted by high-resolution ice crystal images and the RICE frequency (not shown), and summarized in ice habit observations in Varcie et al. (2023). There is little or no riming observed to the west, and light riming to the east at this level. A qualitative comparison was done with WRF using the riming fraction output from the P3 scheme and a ratio of the graupel to the total ice, snow, and graupel mixing ratios in the Thompson scheme. The riming fraction is near zero for the WRF-THOM18Z and <0.1 in the WRF-RAP18Z over the western and central part of the leg (not shown), while it is 0.1–0.3 in the WRF-ERA18Z and WRF-RAP00Z runs. The fractions increase to 0.2–0.5 for the WRF-RAP18Z to the east and generally <0.1 in the WRF-THOMP18Z. The riming fractions are on average ∼0.1 larger for the 0.67-km P3 run than the 2-km P3 run. The greater riming in the P3 scheme compared to the Thompson scheme is consistent with Naeger et al. (2020), which noted that the single snow category in P3 allows for a longer period of rime growth than the two category (dry snow versus graupel) approach in the Thompson scheme. The WRF-ERA18Z riming fractions of 0.7–0.9 to the east are overpredicted since the modeled vertical motions exceed 1 m s−1 at this level as compared to <0.5 m s−1 in the P-3 observations and other WRF runs (not shown).
The ice and snow size distributions derived from the P-3 aircraft and averaged along each flight leg are also compared with the WRF runs using the approach outlined in Naeger et al. (2020). The shape of the modeled size distributions for the P3 scheme simulations correspond well to size distributions from the P-3 observations for diameters less than 10 mm (Fig. 14). The size distributions for the Thompson runs are too narrow for sizes > 1 mm at the 5.1-km level, with too few larger particles by several order magnitudes less than observed (Fig. 14a). The Thompson size distribution improves somewhat at lower altitudes and higher temperatures. The P3 scheme better predicts the size distribution at 5.1 km, albeit it is somewhat too narrow of a distribution for the WRF-RAP18Z run, while the WRF-ERA18Z and WRF-GFS18Z runs are slightly too broad for sizes > 2 mm by at least an order of magnitude than observed. All P3 runs underpredict the number of smaller (<0.5 mm) particles, while the Thompson runs overpredict the number concentration for particles < 0.3 mm by around an order of magnitude. The higher resolution 0.67-km WRF-RAP18Z also improves the size distribution slightly as compared to the 2-km run at 5.1 km. At the 3.3 km and 3.6 km MSL levels (Figs. 14b,c), there is an overprediction of larger (>2 mm) particles in the P3 runs by an order of magnitude, and an underprediction < 2 mm. Since the P3 scheme goes from underpredicting the number of largest ice particles at 5.1 km to slightly overpredicting the largest ice at 3.6 km, this suggests that the scheme is overpredicting the representation of aggregation in this layer.
Observed P-3 (black and shaded) and WRF (see bottom legend) snow size distribution for the (a) 1508–1524 UTC at 5.1 km MSL (1515 UTC WRF), (b) 1530–1551 UTC at 3.6 km MSL (1545 UTC WRF), and (c) 1556–1613 UTC at 3.1 km MSL (1600 UTC WRF) flight leg periods.
Citation: Monthly Weather Review 151, 9; 10.1175/MWR-D-22-0296.1
Fig. 15 shows box-and-whisker plots for the ice water content along each of the 5.1-, 3.6-, and 3.3-km legs for the various WRF runs and derived values from the P-3. At 5.1 km MSL (Fig. 15a), the observed median ice water content is ∼0.6 g m−3. There is general underprediction (median around 0.25 g m−3) in the P3 runs using the RAP initialization and the WRF-THOM18Z, which is consistent with the underprediction in the number concentrations in Fig. 14a. The best correspondence is for the WRF-GFS18Z, WRF-ERA18Z, and WRF-THOM00Z, albeit the range is much larger than observed in the latter two runs. However, this better agreement may be for the wrong reason, since there was greater frontogenesis than in the RAP analysis (cf. Figs. 6c,e) and thus increased precipitation production. All the runs are better at predicting the median observed ice mass of ∼0.70 g m−3 at the 3.6-km flight level compared to the 5.1-km level (Figs. 15a,b). However, the Thompson runs realistically predicted the ice mass yet underpredicted the number concentration for most sizes (cf. Fig. 14b), suggesting the snow densities may be too large. On the other hand, a number concentration overprediction in the P3 scheme and a realistic mass amount suggest snow densities that are too small. As compared to farther aloft, there is larger WRF variability and errors at 3.3 km given the observed median of ∼0.8 g m−3 (Fig. 15c).
(a)–(c) Box-and-whisker plots of ice water content (g m−3) during the three coordinated flight legs, showing the interquartile range (IQR; box edges), median (line), and mean (filled dot). The whisker edges mark values at 1.5 × IQR above or below the upper and lower quartile, respectively, and the individual open dots show values beyond that range. The altitudes and times segments for (a)–(c) correspond to the legs in Fig. 14. The order of boxplots: (i) WRF-GFS18Z, (ii) WRF-ERA18Z, (iii) WRF-THOM18Z, (iv) WRF-RAP18Z, (v) WRF-THOM00Z, (vi) WRF-RAP00Z, (vii) 0.67-km WRF-RAP18Z, and (viii) derived values from the P-3 probes.
Citation: Monthly Weather Review 151, 9; 10.1175/MWR-D-22-0296.1
There is an overprediction in WRF-GFS18Z and WRF-THOM00Z, and a larger range than observed in the WRF-ERA18Z. The overprediction and larger range in these runs is once again likely from the larger forcing (Figs. 6a,c) and greater instability than observed (Fig. 4a), and thus more convective plumes.
Figure 16 shows the accumulated liquid-equivalent precipitation (in mm) from 1200 to 1800 UTC 7 February for the New York Mesonet sites and four WRF simulations. There is a broad area of heavy precipitation (>20 mm) across the central part of New York State from south to north. The lack of a concentrated area of heavy precipitation is evidence of the lack of persistent snow banding in this case. The WRF simulations realistically simulated this broad region of precipitation, although the WRF Model places the axis of heaviest precipitation slightly to the west of the New York Mesonet maxima. Both the 1800 UTC P3 and Thompson runs at 2-km grid spacing underpredict the precipitation by 5–10 mm (Figs. 16a,b). The Thompson scheme produces 3–5 mm less precipitation than the P3 scheme in the areas of maximum precipitation, which is likely the result of the lesser snow production aloft as well as riming (cf. Figs. 12a,d) and thus less rapid precipitation fallout in the Thompson run. In contrast, the WRF-ERA18Z simulation with its stronger forcing for ascent, produced 10%–20% more precipitation over the region, resulting in overprediction.
WRF Model 6-h accumulated precipitation (shaded; mm) from 1200 to 1800 UTC 7 Feb for (a) WRF-RAP18Z, (b) WRF-THOM18Z, (c) WRF-ERA18Z, and (d) WRF-RAP18Z for the 0.67-km domain as shown with inset box in (a). The annotated circles show liquid accumulated precipitation (mm) from the NY Mesonet during the same 6-h period.
Citation: Monthly Weather Review 151, 9; 10.1175/MWR-D-22-0296.1
Binning the 12–18 UTC precipitation by 5-mm intervals for the New York Mesonet illustrates how the precipitation validation varies by threshold and WRF member (Fig. 17). The 6-h precipitation at each NYS Mesonet station was binned by every 5 mm, and the model precipitation error at each location was calculated, with the median error for each bin shown. Since the 0.67-km domain does not cover the entire NYS Mesonet region, the median error values were calculated for both the whole domain (right bars) and for the 0.67-km domain (left bars) for each model simulation. The most severe underprediction occurs for the highest precipitation thresholds, especially for the 1800 UTC initialization runs. For example, the WRF-RAP18Z has a median error of 4–6 mm (18%–23%) in the 20–25-mm range and a median error of almost 10 mm (37%) in the 25–30-mm range. The 0.67-km WRF-RAP18Z reduces some of the underprediction for the 10–20-mm thresholds by 5%–10%. The WRF-RAP18Z run also has 3%–5% more precipitation for the 10–20-mm thresholds than the WRF-THOM18Z. The WRF-ERA18Z has an overprediction at these moderate thresholds, as well as 0000 UTC WRF-RAP runs. This overprediction is likely because the frontogenesis is greater than in the RAP analysis for these other runs, which in turn favors more ascent and greater cloud water (and thus riming) than observed and thus more precipitation fallout.
Median WRF precipitation error (mm) for the 1200–1800 UTC period vs the NY Mesonet stations for various bin thresholds along x axis (every 5 mm) for the (a) WRF-GFS18Z, (b) WRF-THOM18Z, (c) WRF-RAP18Z, (d) WRF-RAP18Z at 0.67-km grid spacing, (e) WRF-ERA18Z, (f) WRF-THOM00Z, and (g) WRF-RAP00Z runs. The left bar for each threshold denotes stations within the 0.67-km domain (Fig. 16a) and the right bar is for all NYS Mesonet stations.
Citation: Monthly Weather Review 151, 9; 10.1175/MWR-D-22-0296.1
Overall, the WRF runs with the forcing for ascent closest to the RAP analysis and vertical cloud structure most similar to the ER-2 do not have the best precipitation forecast (e.g., 0.67-km WRF-RAP18Z). There is a general underprediction of ice aloft for these runs where the heaviest precipitation fell. In contrast, the WRF runs that have more accurate surface precipitation (WRF-GFS18Z or WRF-RAP00Z) or overpredict the precipitation (WRF-ERA18Z) have too deep of a cloud layer or too much frontogenesis, riming, or ice mass compared to the analysis and observations. Many of these differences can be attributed to errors in the initial and boundary conditions. There are some microphysical uncertainties, but most of this uncertainty is over the eastern part of the flight legs where the heaviest precipitation did not fall (cf. Fig. 1). Naeger et al. (2020) showed that the P3 scheme produces more realistic riming than the Thompson scheme for a heavy precipitation event over the Olympic Mountains, but the riming during that event was more associated with heaviest precipitation. The heaviest precipitation for this 7 February event is in an area of little or no riming, thus the underprediction in the P3 and Thompson schemes suggests deficiencies in the scheme limiting ice production and snow growth aloft, such as the lack of cloud top generating cells (Fig. 8) and thus too little ice production aloft (Fig. 15), or issues related to snow density and aggregation.
4. Summary and conclusions
This paper investigates a Northeast U.S. winter cyclone that occurred on 7 February 2020, in which despite midlevel frontogenesis, low static stability, and a relatively deep surface pressure center (∼980 hPa), failed to produce a coherent snowband within the swath of heaviest precipitation (20–40 mm of liquid equivalent) over western and central New York State. This event was sampled by the Investigation of Microphysics and Precipitation for Atlantic Coast-Threatening Snowstorms (IMPACTS) field campaign. Both the ER-2 and P-3 aircraft flew multiple west-east flight legs across New York State within the comma head of this cyclone, of which three legs were coordinated. Several WRF Model runs were used to explore the multiple aspects of the storm. All simulations were run on a nested grid down to 2-km and one run at 0.67-km grid spacing. Three different initial and boundary conditions (RAP analysis, GFS analysis, and ERA5 analysis) were explored as well as two different microphysics parameterizations (Thompson and P3). Although a control run (WRF-RAP18Z) was identified, results from the other runs were also highlighted to demonstrate the initial condition and model uncertainties.
As in the radar observations, all the WRF members did not produce a well-defined band despite significant differences in the intensity of the frontogenesis. This event featured a sloping front from the surface up to midlevels and thus relatively broad cross-front frontogenesis over 150 km, as opposed to a distinct and more concentrated region of frontogenesis along a midlevel trough as in past snowband studies (e.g., Novak et al. 2010) in this region. Additional work is needed to understand the lack of well-defined banding in this case. The frontogenesis peaked over a 6-h period (1200–1800 UTC) and decreased as the flow increased on the east side of the midlevel cyclone, thus resulting in less frontogenesis.
There is a large variation in the precipitation and microphysics from west to east across New York State as sampled by the aircraft. The precipitating cloud was deeper to the west where there was relatively deep frontogenesis, and it was somewhat more shallow and more convective to the east, where there was more potential instability aloft. The WRF had a midlevel dry intrusion advancing faster than observed, and thus many members are too dry over the eastern part of the precipitation area. The 2-km WRF members failed to produce the convective cells near the echo top as observed by aircraft radar. The 0.67-km WRF domain model run represented these narrow precipitation features aloft better than the other coarser 2-km WRF domain runs.
Aircraft data suggest rapid ice growth via aggregation aloft with this amorphous band. The convective cells aloft likely enhanced the ice production aloft; however, the impacts of using a different microphysics scheme on the surface precipitation is modest (5%–10% increase). For example, the 0.67-km domain with its greater cell production only produces small surface precipitation increases compared to the 2-km grid.
Overall, using different IC and BCs produced much larger precipitation sensitivity (10%–40%) than the two microphysical schemes with a 12–24-h lead time (5%–10%). Changes in the initial conditions resulted in larger variations in the frontogenesis, and this had a larger impact than the snow and riming differences in the microphysics schemes for this event. Thus, although microphysics are important, more observations in concert with data assimilation of the regional environment can be even more important for short-term heavy snow predictions. The greater IC/BC sensitivity compared to microphysical scheme sensitivity should be explored for additional cool season cyclones with field data to generalize the results in this study.
Acknowledgments.
We thank the support of the NASA Earth Science Division (ESD), Earth-Venture Suborbital Program under the NASA Airborne Science Program. We also thank the three reviewers for their comments and suggestions to help improve the manuscript. We acknowledge support for IMPACTS by NASA Grants 80NSSC19K0338 (UW) and 80NSSC19K0394 (SBU).
Data availability statement.
All IMPACTS quick-look images and mission scientist reports from the 2020 Deployment are highlighted in the field catalog at http://catalog.eol.ucar.edu/impacts_2020 and the data can be obtained from the Global Hydrology Resource Center Distributed Active Archive Center at https://ghrc.nsstc.nasa.gov/uso/ds_details/collections/impacts.html and McMurdie et al. (2019). Since each of the model runs are ∼30 GB in size, these data can be made available upon request.
REFERENCES
Benjamin, S. G., and Coauthors, 2016: A North American hourly assimilation and model forecast cycle: The Rapid Refresh. Mon. Wea. Rev., 144, 1669–1694, https://doi.org/10.1175/MWR-D-15-0242.1.
Brotzge, J. A., and Coauthors, 2020: A technical overview of the New York State mesonet standard network. J. Atmos. Oceanic Technol., 37, 1827–1845, https://doi.org/10.1175/JTECH-D-19-0220.1.
Colle, B. A., D. Stark, and S. E. Yuter, 2014: Surface microphysical observations within East Coast winter storms on Long Island, New York. Mon. Wea. Rev., 142, 3126–3146, https://doi.org/10.1175/MWR-D-14-00035.1.
Connelly, R., and B. A. Colle, 2019: Validation of snow multi-bands in the comma head of a cyclone using a 40-member ensemble. Wea. Forecasting, 34, 1343–1363, https://doi.org/10.1175/WAF-D-18-0182.1.
Cressman, G. P., 1959: An operational objective analysis system. Mon. Wea. Rev., 87, 367–374, https://doi.org/10.1175/1520-0493(1959)087<0367:AOOAS>2.0.CO;2.
Finlon, J. A., L. A. McMurdie, and R. J. Chase, 2022: Investigation of microphysical properties within regions of enhanced dual-frequency ratio during the IMPACTS field campaign. J. Atmos. Sci., 79, 2773–2795, https://doi.org/10.1175/JAS-D-21-0311.1.
Ganetis, S. A., B. A. Colle, S. E. Yuter, and N. P. Hoban, 2018: Environmental conditions associated with observed snowband structures within Northeast U.S. winter storms. Mon. Wea. Rev., 146, 3675–3690, https://doi.org/10.1175/MWR-D-18-0054.1.
Geerts, B., and P. V. Hobbs, 1991: Organization and structure of clouds and precipitation on the mid-Atlantic coast of the United States. Part IV: Retrieval of the thermodynamic and microphysical structures of a frontal rainband from Doppler radar data. J. Atmos. Sci., 48, 1287–1305, https://doi.org/10.1175/1520-0469(1991)048<1287:OASOCA>2.0.CO;2.
Grell, G. A., and D. Dévényi, 2002: A generalized approach to parameterizing convection combining ensemble and data assimilation techniques. Geophys. Res. Lett., 29, 1693, https://doi.org/10.1029/2002GL015311.
Helmus, J. J., and S. M. Collis, 2016: The Python ARM Radar Toolkit (Py-ART), a library for working with weather radar data in the Python programming language. J. Open Res. Software, 4, e25, https://doi.org/10.5334/jors.119.
Hersbach, H., and Coauthors, 2020: The ERA5 global reanalysis. Quart. J. Roy. Meteor. Soc., 146, 1999–2049, https://doi.org/10.1002/qj.3803.
Heymsfield, A. J., S. Y. Matrosov, and N. B. Wood, 2016: Toward improving ice water content and snow-rate retrievals from radars. Part I: X and W bands, emphasizing CloudSat. J. Appl. Meteor. Climatol., 55, 2063–2090, https://doi.org/10.1175/JAMC-D-15-0290.1.
Houze, R. A., Jr., P. V. Hobbs, K. R. Biswas, and W. M. Davis, 1976: Mesoscale rainbands in extratropical cyclones. Mon. Wea. Rev., 104, 868–878, https://doi.org/10.1175/1520-0493(1976)104<0868:MRIEC>2.0.CO;2.
Jurewicz, M. L., Sr., and M. S. Evans, 2004: A comparison of two banded, heavy snowstorms with very different synoptic settings. Wea. Forecasting, 19, 1011–1028, https://doi.org/10.1175/WAF-823.1.
Keeler, J. M., B. F. Jewett, R. M. Rauber, G. M. McFarquhar, R. M. Rasmussen, L. Xue, C. Liu, and G. Thompson, 2016a: 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, https://doi.org/10.1175/JAS-D-15-0126.1.
Keeler, J. M., B. F. Jewett, R. M. Rauber, G. M. McFarquhar, R. M. Rasmussen, L. Xue, C. Liu, and G. Thompson, 2016b: Dynamics of cloud-top generating cells in winter cyclones. Part II: Radiative and instability forcing. J. Atmos. Sci., 73, 1529–1553, https://doi.org/10.1175/JAS-D-15-0127.1.
Keeler, J. M., B. F. Jewett, R. M. Rauber, G. M. McFarquhar, R. M. Rasmussen, L. Xue, C. Liu, and G. Thompson, 2017: Dynamics of cloud-top generating cells in winter cyclones. Part III: Shear and convective organization. J. Atmos. Sci., 74, 2879–2897, https://doi.org/10.1175/JAS-D-16-0314.1.
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, https://doi.org/10.1175/JAMC-D-13-0312.1.
Lackmann, G. M., and G. Thompson, 2019: Hydrometeor lofting and mesoscale snowbands. Mon. Wea. Rev., 147, 3879–3899, https://doi.org/10.1175/MWR-D-19-0036.1.
Lawson, R. P., R. E. Stewart, and L. J. Angus, 1998: Observations and numerical simulations of the origin and development of very large snowflakes. J. Atmos. Sci., 55, 3209–3229, https://doi.org/10.1175/1520-0469(1998)055<3209:OANSOT>2.0.CO;2.
Lawson, R. P., D. O’Connor, P. Zmarzly, K. Weaver, B. Baker, Q. Mo, and H. Jonsson, 2006: The 2D-S (Stereo) probe: Design and preliminary tests of a new airborne, high- speed, high-resolution particle imaging probe. J. Atmos. Oceanic Technol., 23, 1462–1477, https://doi.org/10.1175/JTECH1927.1.
Livneh, B., P. J. Restrepo, and D. P. Lettenmaier, 2011: Development of a unified land model for prediction of surface hydrology and land–atmosphere interactions. J. Hydrometeor., 12, 1299–1320, https://doi.org/10.1175/2011JHM1361.1.
McFarquhar, G. M., J. A. Finlon, D. M. Stechman, W. Wu, R. C. Jackson, and M. Freer, 2018: University of Illinois/Oklahoma Optical Array Probe (OAP) Processing Software. Zenodo, accessed 1 December 2021, https://doi.org/10.5281/zenodo.1285969.
McMurdie, L. A., G. Heymsfield, J. E. Yorks, and S. A. Braun, 2019: Investigation of Microphysics and Precipitation for Atlantic Coast-Threatening Snowstorms (IMPACTS) collection. NASA EOSDIS Global Hydrology Resource Center Distributed Active Archive Center, accessed 1 November 2021, https://doi.org/10.5067/IMPACTS/DATA101.
McMurdie, L. A., and Coauthors, 2022: Chasing snowstorms: The Investigation of Micro-physics and Precipitation for Atlantic Coast-threatening Snowstorms (IMPACTS) campaign. Bull. Amer. Meteor. Soc., 103, E1243–E1269, https://doi.org/10.1175/BAMS-D-20-0246.1.
Molthan, A. L., B. A. Colle, S. E. Yuter, and D. Stark, 2016: Comparisons of modeled and observed reflectivity and fall speeds for snowfall of varied riming degree during winter storms on Long Island, NY. Mon. Wea. Rev., 144, 4327–4347, https://doi.org/10.1175/MWR-D-15-0397.1.
Morrison, H., and J. A. Milbrandt, 2015: Parameterization of cloud microphysics based on the prediction of bulk ice particle properties. Part I: Scheme description and idealized tests. J. Atmos. Sci., 72, 287–311, https://doi.org/10.1175/JAS-D-14-0065.1.
Naeger, A. R., B. A. Colle, N. Zhou, and A. Molthan, 2020: Evaluating warm and cold rain processes in cloud microphysical schemes using OLYMPEX field measurements. Mon. Wea. Rev., 148, 2163–2190, https://doi.org/10.1175/MWR-D-19-0092.1.
Nakanishi, M., and H. Niino, 2009: Development of an improved turbulence closure model for the atmospheric boundary layer. J. Meteor. Soc. Japan, 87, 895–912, https://doi.org/10.2151/jmsj.87.895.
National Centers for Environmental Prediction/National Weather Service/NOAA/U.S. Department of Commerce, 2015a: NCEP GFS 0.25 degree global forecast grids historical archive. Dept. of Commerce/NOAA/NWS/NCEP, National Center for Atmospheric Research Computational and Information Systems Laboratory, accessed 1 December 2021, https://doi.org/10.5065/D65D8PWK.
National Centers for Environmental Prediction/National Weather Service/NOAA/U.S. Department of Commerce, 2015b: NCEP North American Mesoscale (NAM) 12 km Analysis. Dept. of Commerce/NOAA/NWS/NCEP, National Center for Atmospheric Research, Computational and Information Systems Laboratory, accessed 1 December 2021, https://doi.org/10.5065/G4RC-1N91.
National Operational Hydrologic Remote Sensing Center, 2004: Snow Data Assimilation System (SNODAS) data products at NSIDC, version 1. National Snow and Ice Data Center, accessed 1 December 2021, https://doi.org/10.7265/N5TB14TC.
Novak, D. R., L. F. Bosart, D. Keyser, and J. S. Waldstreicher, 2004: An observational study of cold season-banded precipitation in Northeast U.S. cyclones. Wea. Forecasting, 19, 993–1010, https://doi.org/10.1175/815.1.
Novak, D. R., B. A. Colle, and S. E. Yuter, 2008: High-resolution observations and model simulations of the life cycle of an intense mesoscale snowband over the northeastern United States. Mon. Wea. Rev., 136, 1433–1456, https://doi.org/10.1175/2007MWR2233.1.
Novak, D. R., B. A. Colle, and A. R. Aiyyer, 2010: Evolution of mesoscale precipitation band environments within the comma head of northeast U.S. cyclones. Mon. Wea. Rev., 138, 2354–2374, https://doi.org/10.1175/2010MWR3219.1.
Petterssen, S., 1956: Motion and Motion Systems. Vol. I, Weather Analysis and Forecasting, McGraw-Hill, 428 pp.
Picca, J. C., D. M. Schultz, B. A. Colle, S. Ganetis, D. R. Novak, and M. J. Sienkiewicz, 2014: The value of dual-polarization radar in diagnosing the complex microphysical evolution of an intense snowband. Bull. Amer. Meteor. Soc., 95, 1825–1834, https://doi.org/10.1175/BAMS-D-13-00258.1.
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, https://doi.org/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 within the stratiform comma-head region of continental winter cyclones. J. Atmos. Sci., 72, 2465–2483, https://doi.org/10.1175/JAS-D-14-0354.1.
Rauber, R. M., and S. W. Nesbitt, 2018: Radar Meteorology: A First Course. John Wiley & Sons, 496 pp.
Rauber, R. M., S. M. Ellis, J. Vivekanandan, J. Stith, W.-C. Lee, G. M. McFarquhar, and B. F. Jewett, 2017: Finescale structure of a snowstorm over the northeastern United States: A first look at high-resolution HIAPER cloud radar observations. Bull. Amer. Meteor. Soc., 98, 253–269, https://doi.org/10.1175/BAMS-D-15-00180.1.
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, https://doi.org/10.1175/JAS-D-13-0249.1.
Sanders, F., and L. F. Bosart, 1985: Mesoscale structure in the megalopolitan snowstorm, 11–12 February 1983. Part II: Doppler radar study of the New England snowband. J. Atmos. Sci., 42, 1398–1407, https://doi.org/10.1175/1520-0469(1985)042<1398:MSITMS>2.0.CO;2.
Schultz, D. M., and P. N. Schumacher, 1999: The use and misuse of conditional symmetric instability. Mon. Wea. Rev., 127, 2709–2732, https://doi.org/10.1175/1520-0493(1999)127<2709:TUAMOC>2.0.CO;2.
Schultz, D. M., and J. A. Knox, 2007: Banded convection caused by frontogenesis in a conditionally, symmetrically, and inertially unstable environment. Mon. Wea. Rev., 135, 2095–2110, https://doi.org/10.1175/MWR3400.1.
Skamarock, W. C., and Coauthors, 2019: A description of the Advanced Research WRF Model version 4. NCAR Tech. Note NCAR/TN-556+STR, 145 pp., https://doi.org/10.5065/1dfh-6p97.
Straka, J. M., D. S. Zrnić, and A. V. Ryzhkov, 2000: Bulk hydrometeor classification and quantification using polarimetric radar data: Synthesis of relations. J. Appl. Meteor., 39, 1341–1372, https://doi.org/10.1175/1520-0450(2000)039<1341:BHCAQU>2.0.CO;2.
Thiébaux, J., E. Rogers, W. Wang, and B. Katz, 2003: A new high-resolution blended real-time global sea surface temperature analysis. Bull. Amer. Meteor. Soc., 84, 645–656, https://doi.org/10.1175/BAMS-84-5-645.
Thompson, G., P. R. Field, R. M. Rasmussen, and W. D. Hall, 2008: Explicit forecasts of winter precipitation using an improved bulk microphysics scheme. Part II: Implementation of a new snow parameterization. Mon. Wea. Rev., 136, 5095–5115, https://doi.org/10.1175/2008MWR2387.1.
Tomkins, L. M., S. Yuter, M. Miller, and L. Allen, 2022: Image muting of mixed precipitation to improve identification of regions of heavy snow in radar data. Atmos. Meas. Tech., 15, 5515–5525, https://doi.org/10.5194/amt-15-5515-2022.
Trapp, J. R., D. M. Schultz, A. V. Ryzhkov, and R. L. Holle, 2001: Multiscale structure and evolution of an Oklahoma winter precipitation event. Mon. Wea. Rev., 129, 486–501, https://doi.org/10.1175/1520-0493(2001)129<0486:MSAEOA>2.0.CO;2.
Varcie, M. M., and Coauthors, 2023: Precipitation growth processes in the comma-head region of the 7 February 2020 Northeast snowstorm: Results from IMPACTS. J. Atmos. Sci., 80, 3–29, https://doi.org/10.1175/JAS-D-22-0118.1.