Snow Band Research Articles

Dual adaptive differential threshold method for automated detection of faint and strong echo features in radar observations of winter storms

Abstract. Radar observations of winter storms often exhibit locally enhanced linear features in reflectivity, sometimes labeled as snow bands. We have developed a new, objective method for detecting locally enhanced echo features in radar data from winter storms. In comparison to convective cells in warm season precipitation, these features are usually less distinct from the background echo and often have more fuzzy or feathered edges. This technique identifies both prominent, strong features and more subtle, faint features. A key difference from previous radar reflectivity feature detection algorithms is the combined use of two adaptive differential thresholds, one that decreases with increasing background values and one that increases with increasing background values. The algorithm detects features within a snow rate field rather than reflectivity and incorporates an underestimate and overestimate of feature areas to account for uncertainties in the detection. We demonstrate the technique on several examples from the US National Weather Service operational radar network. The feature detection algorithm is highly customizable and can be tuned for a variety of data sets and applications.

Western Montana Snow Squall Climatology and Synoptic Patterns, 2000–2020

The rapid onset of whiteout conditions, from strong winds, blowing snow, and heavy snow, associated with snow squalls often results in hazardous travel conditions. Snow squalls bring pronounced impacts to the transportation sector, leading to accidents, fatalities, and significant travel delays. Given these impacts, historical frequencies, and the synoptic environments associated with snow squalls in western Montana were examined to develop forecast aides, primarily through synoptic pattern recognition, for meteorologists to identify and communicate snow squalls. A 20-yr climatology was constructed to identify events characterized by wind gusts ≥26 kt (30 mph) and visibility ≤0.8 km (0.5 mi). 74 snow squall events were identified from 2000 through 2020 across three ASOS sites: Missoula (KMSO), Kalispell (KGPI), and Butte (KBTM). The median duration of heavy (moderate) snow during snow squalls was 26 (34) min. A composite study utilizing these events was conducted using the NCEP North American Regional Reanalysis. The composite study along with surface observations, revealed three distinct snow squall patterns and types: orographic snow bands, Pacific cold fronts, and Canadian arctic fronts.

A Kinematic Modeling Study of the Reorganization of Snowfall between Cloud-Top Generating Cells and Low-Level Snowbands in Midlatitude Winter Storms

Abstract This paper explores whether particles within uniformly spaced generating cells falling at terminal velocity within observed 2D wind fields and idealized deformation flow beneath cloud top can be reorganized consistent with the presence of single and multibanded structures present on WSR-88D radars. In the first experiment, two-dimensional wind fields, calculated along cross sections normal to the long axis of snowbands observed during three northeast U.S. winter storms, were taken from the initialization of the High-Resolution Rapid Refresh model. This experiment demonstrated that the greater the residence time of the particles in each of the three storms, the greater particle reorganization occurred. For experiments with longer residence times, increases in particle concentrations were nearly or directly collocated with reflectivity bands. For experiments with shorter residence times, particle reorganization still conformed to the band features but with less concentration enhancement. This experiment demonstrates that the combination of long particle residence time and net convergent cross-sectional flow through the cloud depth is sufficient to reorganize particles into locations consistent with precipitation bands. Increased concentrations of ice particles can then contribute, along with any dynamic forcing, to the low-level reflectivity bands seen on WSR-88D radars. In a second experiment, the impact of flow deformation on the reorganization of falling ice particles was investigated using an idealized kinematic model with stretching deformation flow of different depths and magnitudes. These experiments showed that deformation flow provides for little particle reorganization given typical deformation layer depths and magnitudes within the comma head of such storms. Significance Statement Past research with vertically pointing and scanning radars presents two different perspectives regarding snowfall organization in winter storms. Vertically pointing radars often observe cloud-top generating cells with precipitation fallstreaks descending into a broad stratiform echo at lower altitudes. In contrast, scanning radars often observe snowfall organized in quasi-linear bands. This work attempts to provide a connection between these two perspectives by examining how two-dimensional convergent and deformation flow occurring in winter storms can contribute to the reorganization of snowfall between cloud top and the ground.

Snowpack Aging, Water Isotope Evolution, and Runoff Isotope Signals, Palouse Range, Idaho, USA

A snowpack’s δ2H and δ18O values evolve with snowfall, sublimation, evaporation, and melt, which produces temporally variable snowpack, snowmelt, and runoff isotope signals. As a snowpack ages, the relatively depleted δ2H and δ18O values of snow will become less depleted with sublimation and evaporation, and the internal distribution of isotope signals is altered with melt moving through and out of the snowpack. An examination of δ2H and δ18O values for snowpack, snowmelt, and ephemeral creek water in the Palouse Range of northern Idaho indicated an evolution from variably depleted snowpack to enriched snowmelt and relatively consistent isotope signals in springtime ephemeral creeks. Within the primary snow band of the mountain range and during the winter–spring period of 2019–2020, the snowpack had an isotope range of −130 to −75‰ for δ2H and −18 to −10.5‰ for δ18O with resulting snowmelt values of −120 to −90‰ for δ2H and −16.5 to −12.5‰ for δ18O. With runoff of snowmelt to ephemeral creeks, the isotope values compressed to −107 to −104‰ for δ2H and −15.5 to −14.5‰ for δ18O. Aging of the snowpack produced increasing densities in the base, middle, and upper layers along with a corresponding enrichment of isotope values. The highest elevation site indicated the least enrichment of δ2H and δ18O in the snowpack base layer, and the lowest elevation site indicated the strongest enrichment of δ2H and δ18O in the snowpack base layer. Deuterium excess decreased with snowpack aging processes of accumulation and melt release, along with the migration of water vapor and snowmelt within the snowpack. It is likely that winter melt (early depleted signal) is a primary contributor to creeks and groundwater along the Palouse Range, but the strong variability of snowpack isotope signals provides a wide range of possible isotope signals to surface-water and groundwater systems at the mountain front.

Climatology of Lake-Effect Snow Days Along the Southern Shore of Lake Michigan: What Is the Sensitivity to Environmental Factors and Snowband Morphology?

The Laurentian Great Lakes have substantial influences on regional climatology, particularly with impactful lake-effect snow events. This study examines the snowfall, cloud-inferred snow band morphology, and environment of lake-effect snow days along the southern shore of Lake Michigan for the 1997–2017 period. Suitable days for study were identified based on the presence of lake-effect clouds assessed in a previous study and extended through 2017, combined with an independent classification of likely lake-effect snow days based on independent snowfall data and weather map assessments. The primary goals are to identify lake-effect snow days and evaluate the snowfall distribution and modes of variability, the sensitivity to thermodynamic and flow characteristics within the upstream sounding at Green Bay, WI, and the influences of snowband morphology. Over 300 lake-effect days are identified during the study period, with peak mean snowfall within the lake belt extending from southwest Michigan to northern Indiana. Although multiple lake-effect morphological types are often observed on the same day, the most common snow band morphology is wind parallel bands. Relative to days with wind parallel bands, the shoreline band morphology is more common with a reduced lower-tropospheric zonal wind component within the upstream sounding at Green Bay, WI, as well as higher sea-level pressure and 500-hPa geopotential height anomalies to the north of the Great Lakes. Snowfall is sensitive to band morphology, with higher snowfall for shoreline band structures than for wind parallel bands, especially due south of Lake Michigan. Snowfall is also sensitive to thermodynamic and flow properties, with a greater sensitivity to temperature in southwest Michigan and to flow properties in northwest Indiana.

The Severe Multi-Day October 2019 Snow Storm Over Southern Manitoba, Canada

ABSTRACT A devastating storm struck southern Manitoba, Canada on 10–13 October 2019, producing a large region of mainly sticky and wet snow. Accumulations reached 75 cm, wind gusts exceeded 100 km h−1, and surface temperature (T) remained near 0°C (−1°C ≤ T ≤ 1°C) for up to 88 h. It produced the largest October snowfall and was the earliest to produce at least 20 cm since 1872 in Winnipeg. These factors led to unparalleled damage and power restoration challenges for Manitoba Hydro and, with leaves still largely on vegetation, the most damaging storm to Winnipeg’s trees ever recorded. The storm’s track was uncommon, and produced elevated convection related to buoyancy-driven instability and conditional symmetric instability (CSI), with a moist absolutely unstable layer (MAUL) near 500 hPa. Instabilities were released via lift through lower-tropospheric warm advection and frontogenesis, differential cyclonic vorticity advection, and jet streak dynamics. Precipitation bands, elevated convection, and lake effect snow bands enhanced local snowfall. Snow adhering to structures was not always wet but, when present, it sometimes occurred because of incomplete freezing of particles partially melted aloft in a near-surface (<100 m deep) inversion. Although other storms over the historical record have produced a similar combination of severe precipitation, temperature and wind conditions, none have done this for such a long period.

Heavy Snowfall Produced by Topographically-Induced Winds in the Snake River Plain of Eastern Idaho. Part I: Observational Analysis

This study is the first of two papers concerning the dynamics of a heavy snowfall event in the Snake River Plain of eastern Idaho on 26 November 2005. Heavy snowfall occurred along the southern perimeter of the wide (~100 km) valley and in three to four mesoscale bands aligned across the valley and parallel with the post cold-frontal northwesterly flow. This event was driven by two main topographic forcing mechanisms. First, widespread precipitation resulted from a low-level barrier jet (oriented down a tight horizontal pressure gradient) ascending the Snake River Valley and spilling over the southern edge of the valley. Second, the mesoscale snow bands in the valley resulted from boundary layer convergence forced by flow through the upstream barrier, the Idaho Central Mountains, in three narrow valleys aligned with the 700 hPa wind. This case study illustrates how the unique topography of eastern Idaho contributed to near-blizzard conditions in a post-frontal regime.

Analysis of a vortex causing sea-effect snowfall in the western part of the Black Sea: a case study of events that occurred on 30–31 January 2012

In this study, snow events in the Marmara region were identified for the 2009–2018 period, and then the times at which a vortex causing the formation of a sea-effect snow (SES) band was seen over the Black Sea were determined from Terra, MSG-2 and MSG-3 satellite imagery. During this period, the SES band was found only between 30 and 31 January 2012 over the western Black Sea. This band maintained its dominance over the region for approximately 24 h. The cause and effects of this band and the precipitation that resulted were investigated. Meteorological data published by 12 airports, 16 meteorological stations and 5 radiosonde stations belonging to five countries were analysed within the two-day period to describe the relevant meteorological conditions. In addition, sea surface temperature data and synoptic charts of the upper atmospheric levels were investigated. According to the results, the maximum amount of snow (24 cm) was observed at the Inebolu Station (in Turkey). As a result of the effect of the vortex-type SES band moving over the Marmara and western Black Sea regions, a total of 102 flights belonging to a private airline company were cancelled, sea transportation was disrupted, and vehicular accidents occurred in Istanbul. In all radiosonde measurements, inversion layers were observed between the surface and the 700 hPa level, and wind direction changes occurred mostly between 0° and 50° within the same level range. The temperature difference between the sea surface and the 850 hPa level reached 25.8 °C where the vortex-type SES band core was located. In the synoptic level charts, the dominance of a low-pressure centre during the two-day period was observed in the area where the vortex-type SES band was formed.

Classification and analysis of sea‐effect snowbands for Danube Sea area in Black Sea

AbstractIn this study, for the Danube Sea Area (Danube Region—western part of the Black Sea), the structure of the sea effect snow (SES) bands, meteorological conditions in which SES bands occurred, and SES band types are analysed for the period 2009–2018. In the 10‐year period, snowy times were determined using aviation reports published by four airports within the borders of the Marmara Region. SES events and SES bands were determined by examining the several satellite images (Channel 9‐IR, Natural Colour‐RGB, Airmass‐RGB, Channel 6‐VW) obtained from the Spinning‐Enhanced Visible and Infrared Imager (SEVIRI) device on MSG‐2 (Meteosat‐9) and MSG‐3 (Meteosat‐10) satellites. The Danube Region is divided into six sub‐regions. Five different types of SES bands were analysed separately for each sub‐region. Some meteorological conditions that are known to play an important role in the formation of SES bands (presence, thickness and level of the inversion layer; wind structure in the upper levels over the sea; sea surface temperature—SST) were analysed for the snowy times with sea‐effect. According to the preliminary results; a total of 186 days were found to be snowy, 75 days of these snowy days occurred with sea‐effect, and SES bands were seen in 8 days but did not cause snowfall in the Marmara Region. Type‐2 (T‐2) band was the most common band type (85%) over the Danube Region. The inversion layer(s) were observed between the level of 850–700 hPa in 52 days of 75 snowy days with sea effect. The majority of the direction of the upper‐level winds (at 850 and 700 hPa) was found to be northern (North‐N, Northwest‐NW, Northeast‐NE). The lowest average SST value (6.02°C) observed in the sub‐region ‘A’ and the highest average SST value (8.66°C) was observed in the sub‐region ‘F’.

Statistics of sea-effect snowfall along the Finnish coastline based on regional climate model data

Abstract. The formation of convective sea-effect snowfall (i.e., snow bands) is triggered by cold air outbreaks over a relatively warm and open sea. Snow bands can produce intense snowfall which can last for several days over the sea and potentially move towards the coast depending on wind direction. We defined the meteorological conditions which statistically favor the formation of snow bands over the north-eastern Baltic Sea of the Finnish coastline and investigated the spatio-temporal characteristics of these snow bands. A set of criteria, which have been previously shown to be able to detect the days favoring sea-effect snowfall for Swedish coastal area, were refined for Finland based on four case study simulations, utilizing a convection-permitting numerical weather prediction (NWP) model (HARMONIE-AROME). The main modification of the detection criteria concerned the threshold for 10 m wind speed: the generally assumed threshold value of 10 m s−1 was decreased to 7 m s−1. The refined criteria were then applied to regional climate model (RCA4) data, for an 11-year time period (2000–2010). When only considering cases in Finland with onshore wind direction, we found on average 3 d yr−1 with favorable conditions for coastal sea-effect snowfall. The heaviest convective snowfall events were detected most frequently over the southern coastline. Statistics of the favorable days indicated that the lower 10 m wind speed threshold improved the representation of the frequency of snow bands. For most of the favorable snow band days, the location and order of magnitude of precipitation were closely captured, when compared to gridded observational data for land areas and weather radar reflectivity images. Lightning were observed during one third of the favorable days over the Baltic Sea area.

Assessment of coupled CRCM5–FLake on the reproduction of wintertime lake-induced precipitation in the Great Lakes Basin

This study assesses the high-resolution, 0.11° (12 km), Canadian Regional Climate Model Version 5 (CRCM5), interactively coupled to the one-dimensional Freshwater Lake model (FLake), to predict wintertime precipitation along the Canadian snowbelts of Lake Superior and Lake Huron. CRCM5–FLake was compared against various datasets to evaluate the 20-year (1995–2014) SWE and wintertime precipitation, seven lake-induced precipitation events and lake effect snowfall (LES) predictor variables during the months of December and January. The findings of SWE along both snowbelts in December and January show MBD ≤− 10 mm and ≤− 30 mm, respectively. Similarly, precipitation results along both snowbelts in December and January show MBD ≤− 5 mm and ≤− 10 mm, respectively. The negative biases in simulated SWE and precipitation, predominantly along the snowbelts, suggest that the model may un-realistically represent lake effect processes. Comparison of lake-induced precipitation events also indicates that the model mostly under-predicts the daily accumulated precipitation associated with each event but tends to accurately capture the timing and the general location of the squalls along the snowbelts, though not for highly localised snow bands. Furthermore, lake-wide results of LES predictor variables indicate that the model over-estimates lake surface temperature (LST) for both lakes during December and January and under-estimates ice cover concentrations for both lakes in December. The resultant biases could be attributed to limitations within the coupled RCM because the quality of reproducing lake-induced precipitation in this region is highly dependent on the performance of FLake.

Comprehensive accuracy assessment of MODIS daily snow cover products and gap filling methods

The accuracy of the standard Moderate Resolution Imaging Spectroradiometer (MODIS) daily snow cover products (Collection 5) and several of the most common and computationally frugal gap-filling methods were validated using 12 years of daily observations from over 800 Snow Telemetry (SNOTEL) stations. While several factors affect snow cover accuracy to some extent, the largest controls are associated with land cover type; accuracy fell from the maximum in croplands (94.9%) to the minimum for observations with a land cover type of water (79.5%). The second largest impact on accuracy is attributed to the changes in solar zenith angle, where accuracy decreased from a maximum of ∼96% at 42° to a minimum of ∼84% at 58°. Based on the results of this work, the highest accuracy binary daily snow cover dataset can be achieved by reclassifying the fractional snow band using a no-snow/snow threshold of 10 from the MODIS Terra sensor. Gap-filling is best accomplished by a temporal filter on the same dataset, and the window length depends on the desired completeness of the final dataset and the amount of missing values in the area of interest. Decreases in accuracy caused by this gap-filling method taper off after 7 days. Therefore, if a larger temporal window is needed to adequately fill in a time series it is best to extend the window to the largest gap present.

Forecasting the Inland Extent of Lake Effect Snow Bands Downwind of Lake Ontario

Forecasting the Inland Extent of Lake Effect Snow Bands Downwind of Lake Ontario

Characteristics of convective snow bands along the Swedish east coast

Abstract. Convective snow bands develop in response to a cold air outbreak from the continent or the frozen sea over the open water surface of lakes or seas. The comparatively warm water body triggers shallow convection due to increased heat and moisture fluxes. Strong winds can align with this convection into wind-parallel cloud bands, which appear stationary as the wind direction remains consistent for the time period of the snow band event, delivering enduring snow precipitation at the approaching coast. The statistical analysis of a dataset from an 11-year high-resolution atmospheric regional climate model (RCA4) indicated 4 to 7 days a year of moderate to highly favourable conditions for the development of convective snow bands in the Baltic Sea region. The heaviest and most frequent lake effect snow was affecting the regions of Gävle and Västervik (along the Swedish east coast) as well as Gdansk (along the Polish coast). However, the hourly precipitation rate is often higher in Gävle than in the Västervik region. Two case studies comparing five different RCA4 model setups have shown that the Rossby Centre atmospheric regional climate model RCA4 provides a superior representation of the sea surface with more accurate sea surface temperature (SST) values when coupled to the ice–ocean model NEMO as opposed to the forcing by the ERA-40 reanalysis data. The refinement of the resolution of the atmospheric model component leads, especially in the horizontal direction, to significant improvement in the representation of the mesoscale circulation process as well as the local precipitation rate and area by the model.

Retrieving soil surface temperature under snowpack using special sensor microwave/imager brightness temperature in forested areas of Heilongjiang, China: an improved method

Soil surface temperature ( T s ) is an important indicator of global temperature change and a key input parameter for retrieving land surface variables using remote sensing techniques. Due to the masking in the thermal infrared band and the scattering in the microwave band of snow, the temperature of soil surfaces covered by snow is difficult to infer from remote sensing data. We attempted to estimate T s under snow cover using brightness temperature data from the special sensor microwave/imager. T s under snow cover was underestimated due to the strong scattering effect of snow on upward soil microwave emissions at 37 GHz. The underestimated portion of T s is related to snow properties, such as depth, grain size, and moisture. Based on the microwave emission model of layered snowpacks, the simulated results revealed a linear relationship between the underestimated T s and the brightness temperature difference (TBD) at 19 and 37 GHz. When TBDs at 19 and 37 GHz were introduced to the T s estimation method, accuracy improved, i.e., the root mean square error and bias of the estimated T s decreased greatly, especially for dry snow. This improvement allows T s estimation of snow-covered surfaces from 37 GHz microwave brightness temperature.

Snow bands over the Gulf of Finland in wintertime

Large shore-parallel, quasi-stationary snow bands are occasionally observed over the Gulf of Finland during wintertime when the sea is not frozen. On the basis of Weather Research and Forecasting mesoscale model experiments and radar observations of snow bands formed in January 2006 and February 2012, we show that their dynamics share common characteristics: (1) the sea gulf that produces the known lake effect, (2) cold easterly large-scale flow along the gulf and (3) a cold local flow from the two near and opposite coastlines of Estonia and Finland in the form of two land-breeze cells which collide offshore. The associated fronts, which have strong rising motions, are maintained by the convergence of the land-breeze cells. In addition to these factors, the concave shape of the coast in the eastern part of the Gulf of Finland promotes offshore convergence and the formation of several secondary bands of precipitation that are adjacent to the eastern part of the main band. When the easterlies turn to southerlies, horizontal convective rolls appear over the sea. The Estonian land breeze is enhanced while the cold air remains stagnant inland over the Finnish coast, acting as an orographic barrier lifting the marine air mass upwards. Consequently, a line of convective precipitation composed of several cells is formed along the Finnish coast. In both events, the simulations also show two low-level jets generated by the combined effects of the land-breeze cells and baroclinicity over the coast of Finland and Estonia.

Cold air outbreaks along a non-frozen sea channel: effects of wind on snow bands

Wintertime cold air outbreaks along a non-frozen sea channel or a long lake can become destructive if the related bands of heavy snowfall hit onto land. The forcing for such bands is studied with a 2D numerical model set across an east–west sea channel at 60oN (‘Gulf of Finland’), varying the basic geostrophic wind Vg. Without any Vg opposite coastal land breezes emerge with convergence. This results in a quasi-steady rising motion wmax ~ 7.5 cm/s at 600 m in the middle of the gulf, which can force a snow band. During weak Vg, the rising motion is reduced but least so for winds from 60o to 80o (~ENE), when modest alongshore bands could exist near the downstream (Estonian) coast. During Vg of 4–6 m/s from any direction, the land breezes and rising motions are reduced more effectively, so snow bands are not expected during moderate basic flow. In contrast, during a strong Vg of 20–25 m/s from 110o to 120o (~ESE) the land breeze perturbations are intense with wmax up to 15–18 cm/s. The induced alongshore bands of heavy snowfall are located in these cases at the sea but quite close to the downstream (Finnish) coast. They can suddenly make a landfall if the basic wind turns clockwise.

Synoptic scale outbreak of self‐initiated upward lightning (SIUL) from tall structures during the central U.S. blizzard of 1–2 February 2011

AbstractA major central U.S. winter cyclone on 1–2 February 2011 produced a band of high winds, up to 75 cm of snow, and numerous reports of thundersnow from Oklahoma into Ontario over a 26 h period. The National Lightning Detection Network (NLDN) recorded 282 flashes comprised of 1153 events which were &gt;96% negative polarity. Hopes of imaging winter sprites associated with energetic positive cloud‐to‐ground events that sometimes accompany such winter storms did not materialize. However, the lack of lightning over the Great Lakes waters, plus media reports of numerous thundersnow events in downtown Chicago, prompted a detailed analysis of the NLDN data. This revealed that &gt;93% of all lightning in the snow band was likely or possibly associated with self‐initiated upward lightning (SIUL) events from a variety of tall, and some not so tall, structures. In addition to 43 events from two Chicago skyscrapers, many shorter structures were involved, including wind turbines (13.1% of the total) and transmission line towers (6.7%). Wind speeds for all events exceeded the 8 m s−1 minimum threshold associated with SIULs in Japanese winter lightning storms. Radar reflectivities at the event locations had a mean of 28 dBZ and were almost always &lt;35 dBZ. While conventional radar displays suggested stratiform precipitation in the thundersnow region, detailed analysis of 3‐D‐gridded NMQ (National Mosaic and Multi‐Sensor Quantitative Precipitation Estimation) radar reflectivity data confirmed elevated embedded cellular convection spanning the −10°C region associated with isentropic lifting above a frontal surface, evidence of noninductive charge generation sufficient to allow upward leader initiation from tall objects.

Formation and maintenance mechanisms of a thick snow band along the Okhotsk Sea coast of Hokkaido Island, Japan

Formation and maintenance mechanisms of a thick snow band along the Okhotsk Sea coast of Hokkaido Island, Japan

Observed Microphysical Evolution for Two East Coast Winter Storms and the Associated Snow Bands

AbstractThis paper presents the observed microphysical evolution of two coastal extratropical cyclones (19–20 December 2009 and 12 January 2011) and the associated passage of heavy snowbands in the cyclone comma head. The observations were made approximately 93 km east of New York City at Stony Brook, New York. Surface microphysical measurements of snow habit and degree of riming were taken every 15–30 min using a stereo microscope and camera, and snow depth and snow density were also recorded. A vertically pointing Ku-band radar observed the vertical evolution of reflectivity and Doppler vertical velocities. There were rapid variations in the snow habits and densities related to the changes in vertical motion and depth of saturation. At any one time, a mixture of different ice habits was observed. Certain ice habits were dominant at the surface when the maximum vertical motion aloft occurred at their favored temperature for depositional growth. Convective seeder cells above 4 km MSL resulted in relatively cold (less than −15°C) ice crystal habits (side planes, bullets, and dendrites). Needlelike crystals were prevalent during the preband period when the maximum vertical motion was in the layer from −5° to −10°C. Moderately rimed dendritic crystals were observed at snowband maturity associated with the strongest frontogenetical ascent on the warm (east) side of the bands. Riming rapidly decreased and more platelike crystals became more numerous as the strongest ascent moved east of Stony Brook. Snow-to-liquid density ratios ranged from 8:1 to 13:1 in both events, except during the period of graupel, when the ratio was as low as 4:1.