1. Introduction
Occasionally, meteorological events occur that do not meet warning, watch, or advisory criteria, but have a significant impact on life, loss, and property. DeVoir (2004) pointed out that events “characterized by extremely heavy but short-lived snow bursts or squalls” fit into such a category. Snow bursts or squalls have the potential to create very low visibility and dangerous driving conditions, in addition to anxiety and confusion for the general public, owing to the rapid and often unexpectedly changing environment (DeVoir 2004). Additionally, aviation can often be affected, as winter precipitation and low visibility can greatly impact airport operations, albeit for relatively short periods of time during events such as the one discussed here (i.e., compared to a winter storm or blizzard). This paper focuses on one such event that impacted eastern Ontario and western Quebec in Canada, and northern New York State on 28 January 2010. For example, the Ottawa, Ontario (CYOW), region reported dozens of automobile accidents that resulted in at least one person being critically injured, which is further detailed in section 3. So as not to confuse the reader with “snow squalls” that would typically be associated with lake-effect precipitation regions, we will subsequently refer to these events as “snow bursts,” following Pettegrew et al. (2009).
To our knowledge, there exists little in the way of published literature on the topic of cold-season, high-impact, limited-moisture events (Pettegrew et al. 2009). Pettegrew et al. (2009) is a recent study that included a synoptic–dynamic analysis on an event similar to the one in this paper, which created near-whiteout conditions in Iowa and Illinois in 2003, while only depositing 4 cm of snow. DeVoir (2004) briefly described two such cases in the eastern United States, but did not discuss the details of the dynamics and thermodynamics associated with the events. Instead, the focus was more on the human impacts of the snow bursts and the need for the National Weather Service (NWS) to establish a subadvisory-scale warning criterion. In an NWS technical bulletin, Lundstedt (1993) analyzed snow squalls in northern New England not associated with lake-effect snow. Lundstedt (1993) computed a wintertime instability index (WINDEX) to establish the potential for snow squalls typically associated with the passage of an Arctic cold front. WINDEX was based on a combination of low-level lapse rates, boundary layer relative humidity, and the 12-h change in the lifted index. However, to our knowledge, there are no references to Lundstedt (1993) in the refereed literature and it is unclear whether WINDEX was ever implemented operationally.
Three studies have focused on thundersnow events. Market et al. (2002) established a synoptic-scale climatology of such events, Market et al. (2006) performed a composite analysis of proximity soundings during thundersnow occurrences, and Crowe et al. (2006) analyzed the relationship between thundersnow occurrences and deep 24-h snow accumulations. Market et al. (2002) showed that the majority of thundersnow events occur in association with a synoptic-scale cyclone. Market et al. (2002) also showed that thundersnow did occasionally occur in association with other types of events, such as orographically forced, Arctic front passages, and lake-effect snow, but at a much lower frequency than events associated with a synoptic-scale cyclone. Crowe et al. (2006) showed that thundersnow occurrences within an extratropical cyclone tended to indicate a cyclone capable of producing large snowfall accumulations somewhere in its path, although not necessarily collocated with the thunder observations themselves.
Other work has been mostly limited to examining the dynamic and thermodynamic properties of specific cold-season events involving lightning or severe weather (Holle and Watson 1996; Schultz 1999; Hunter et al. 2001; Trapp et al. 2001; van den Broeke et al. 2005; Corfidi et al. 2006). In particular, Schultz (1999) analyzed the difference between lake-effect events with and without lightning for both northern Utah and western New York State, two regions susceptible to frequent lake-effect events (Niziol et al. 1995; Steenburgh et al. 2000; Steenburgh and Onton 2001). Overall, the large majority of case studies regarding occurrences of snow bursts are directly related to either heavy lake-effect snow events (e.g., Niziol et al. 1995; Steenburgh and Onton 2001; Payer et al. 2007), or heavy mesoscale bands within synoptic-scale heavy snowfall events (e.g., Nicosia and Grumm 1999). In fact, events marked by small amounts of snow are rarely analyzed in the literature (Homan and Uccellini 1987); since, in general, they have a relatively minor impact on human life and property.
Many of the lessons learned from dynamic and thermodynamic analyses of lake-effect snow squalls and heavy mesoscale precipitation bands within larger synoptic-scale heavy snowfall events can be applied to the analysis in this paper. However, this paper differs from all previous work other than Pettegrew et al. (2009) and Lundstedt (1993) in that we are analyzing the dynamics and thermodynamics associated with a low–quantitative precipitation forecast (QPF), high-impact precipitation event in a region not prone to lake-effect snow.
In section 2, an overview of the data used in this study is provided. Section 3 details the surface observations and human impacts of the event on 28 January 2010, while section 4 contains the dynamic and thermodynamic analysis of the snow bursts. Finally, section 5 provides a summary, an outline of future work, and brief recommendations for the operational handling of future similar events.
2. Data
The radar imagery in this paper was obtained using the Environment Canada (EC) historical radar database (available online at http://www.climate.weatheroffice.gc.ca/radar/index_e.html). The radar imagery shown in this paper is entirely from the Franktown (CXFT) radar in eastern Ontario. EC radars are C-band radars with a wavelength of 5 cm and a beamwidth of 0.65°. These radars operate in a continuous scanning mode with the typical volume scan that lasts 5 min. Additionally, EC radars have a Doppler coverage area that is approximately 256 km in diameter. In the winter season, the radar generally operates in snow precipitation mode. (Meteorological Service of Canada 2010b).
The surface observations incorporated into the meteorograms were obtained from Iowa State University’s Iowa Environmental Mesonet archive (available online at http://mesonet.agron.iastate.edu/archive/). Precipitation data were acquired from the EC historical climate database (located online at http://www.climate.weatheroffice.gc.ca/climateData/canada_e.html). EC forecasts were obtained in real time from the EC Weatheroffice Web site (http://www.weatheroffice.gc.ca/canada_e.html). All dynamic and thermodynamic analyses performed for this manuscript were completed using the National Centers for Environmental Prediction (NCEP) North American Regional Reanalysis (NARR), with a horizontal resolution of 32 km, and a 3-hourly temporal resolution (Mesinger et al. 2006). NARR fields were compared with the ½° horizontal resolution NCEP Global Forecast System (GFS) analysis and the NCEP–NCAR Global Reanalysis (Kalnay et al. 1996) and were found to be qualitatively similar in terms of both quasigeostrophic (QG) dynamic and thermodynamic structures. The calculations and analyses in this study are displayed using the General Meteorological Package (GEMPAK) version 5.11.1 [updated from the original package devised by Koch et al. (1983)], a data manipulation and visualization software package.
3. Case overview
This paper focuses on the causes and impacts of the snow burst event of 28 January 2010, which occurred in association with the passage of an Arctic front (detailed in section 4) through parts of eastern Ontario, western Quebec, and northern New York State. For our surface station diagnostics we concentrate on two aviation routine weather report (METAR) stations: 1) CYOW and 2) Massena, New York (KMSS), the latter of which is located 70 km southeast of the former; these stations are marked with a black star and red oval, respectively, in Fig. 1a. On the afternoon of 28 January 2010 two separate snow bursts affected CYOW, KMSS, and their surrounding regions, with one moving through CYOW at approximately 1800 UTC and the other just after 1900 UTC. For the remainder of this manuscript we define t = 0 h as the approximate hour that the first snow burst moved through CYOW, at 1800 UTC.
The EC radar imagery from the Franktown (CXFT) for 28 Jan 2010 at (a) 1510, (b) 1610, (c) 1710, (d) 1810, (e) 1910, and (f) 2010 UTC. The approximate location of CYOW (KMSS) is marked in each panel with a black star (red oval). The first snow burst is marked SB1 with a blue arrow and the second snow burst is marked SB2.
Citation: Weather and Forecasting 26, 3; 10.1175/2010WAF2222432.1
Figure 1 shows a time series of radar imagery from the EC Franktown (Ottawa) radar. Images are shown every hour from 1510 to 2010 UTC as the snow bursts crossed the area. In Fig. 1a, a line of moderate to heavy snow is observed in Ontario along the St. Lawrence River (from east of Kingston northeastward to Cornwall) at t = −3 h. This band of snow persisted for approximately two more hours before moving into the Adirondack Mountains of New York State. Although this first snowband did reduce visibility to less than 1 statute mile (sm) at Kingston, Ontario (not shown), and KMSS (Fig. 2b) prior to t = 0 h, it is not the focus of the paper, as the two later snow bursts (which appear as convective squall lines in Figs. 1c–e) had a substantially larger impact on the region (Jackson 2010; Spears 2010).
Meteorograms for 28 Jan 2010 at (a) CYOW and (b) KMSS. In the top panels of (a) and (b), temperature (dewpoint) is plotted in red (green) in °C. In the bottom panels of (a) and (b), visibility (sm) is on the vertical axis (blue, not shown for values >3 sm), and wind [in knots (kt) where 1 kt = 0.514 m s −1] is represented by barbs. Time on the horizontal axis is relative to t = 0 h (1800 UTC).
Citation: Weather and Forecasting 26, 3; 10.1175/2010WAF2222432.1
The first convective snow burst (hereafter SB1, and marked in Fig. 1) is first observed on the radar imagery at t = −3 h, located approximately 100 km west of CYOW (Fig. 1a). By t = −2 h and t = −1 h it extends in a north–south-oriented line from just north of the Ottawa River (located on the Ontario–Quebec border) to the St. Lawrence River just west of Brockville, Ontario (Figs. 1b and 1c). Also at t = −1 h (Fig. 1c) the second convective snow burst (hereafter SB2 and marked in Fig. 1) is first evident just west of where SB1 originally formed an hour earlier (Fig. 1b). By t = 0 h (Fig. 1d) SB1 is moving through CYOW and SB2 is approaching from the west at approximately 50 km h−1. SB2 moves through CYOW at t = +1 h (Fig. 1e) at approximately the same time that SB1 passes through KMSS. Finally, as SB1 moves eastward into Quebec by t = +2 h (Fig. 1f), SB2 approaches KMSS and moves through the station around t = +3 h (not shown). The maximum observed reflectivity value for both of the convective lines was between 40 and 45 dBZ, observed in both SB1 and SB2 at t = 0 h (Fig. 1d).
Figure 2 presents meteorograms for both CYOW (Fig. 2a) and KMSS (Fig. 2b) for 28 January 2010. Prior to the passage of the snow bursts (and associated Arctic front), the 10-m winds at both stations were observed to be mainly southerly. At CYOW (Fig. 2a), a wind shift (to westerly) is observed at t = 0 h in association with the passage of SB1. The visibility also drops to 1.5 sm and snow showers are reported at t = 0 h (Fig. 2a). Radar imagery at t = 0 h (Fig. 1d) strongly suggests that SB1 moved through CYOW around t = 0 h. This is substantiated by a special METAR report (Table 1) issued at 1803 UTC that reported a visibility of ½ sm (with snow showers) and by a second report at 1812 UTC that reported a visibility of ⅜ sm and a wind that shifted strongly to the west (with reported snow showers). Consequently, the observed visibility during the passage of SB1 at CYOW was actually substantially lower than seen in Fig. 2a, since it occurred in between hourly surface observations (Table 1). Given that SB2 passes through CYOW just after t = +1 h (Fig. 1d, Table 1), we believe the report of ¼-sm visibility at t = +2 h (Fig. 2a) is mostly representative of blowing snow, in association with the 25-kt 10-m wind (Fig. 2a). This is evidence that visibility was hampered both during and after the passage of SB1 and SB2, which is further supported by the drifting snow observations at CYOW at t = +3 and +4 h (Fig. 2a). Finally, both the temperature and dewpoint show a marked decrease shortly after the passage of the convective snow bursts and associated Arctic front, with the temperature dropping more than 5°C between t = −1 h and t = +4 h (Fig. 2a).
Raw METAR reports from CYOW from 1700 (t = −1 h) to 2100 UTC 28 Jan 2010 (t = +3 h), during the passage of the snow bursts. Special reports issued between hourly observations are included.
KMSS experienced more continuous light precipitation than CYOW, primarily in association with the aforementioned moderate snowband observed along the axis of the St. Lawrence River well before t = 0 h (Fig. 2b, Table 2). This is reflected in the meteorograms, as relatively low visibilities were observed while the wind was still southerly, from t = −6 to 0 h (Fig. 2b). SB1 moved through KMSS between t = +1 and +2 h, during which time a wind shift from southerly to westerly occurs (Fig. 2b). METAR reports issued by the ASOS station between these observation times indicate visibilities of less than 1 sm, including ¼-sm visibility at 1911 UTC (Table 2). The lack of a low-visibility observation in the meteorogram (Fig. 2b) during this time is due to the fact that SB1 passed through KMSS between hourly surface observations. By t = +2 h, drifting snow is reported (Fig. 2b) in association with a 20-kt westerly wind. The passage of SB2 through KMSS is evident at t = +3 h in Fig. 2b and in Table 2, with observed ¼-sm visibility and moderate snow. The METAR reports between the hourly observations (Table 2) suggest that low visibility at KMSS was a more persistent concern than can be observed in the meteorograms, which only show the hourly observations. Finally, the temperature and dewpoint did exhibit a similar drop (Fig. 2b) to that observed at CYOW (Fig. 2a) following the shift to westerly winds.
Raw METAR reports from KMSS from 1700 (t = −1 h) to 2100 UTC 28 Jan 2010 (t = +3 h), during the passage of the snow bursts. Special reports issued between hourly observations are included.
It is important to emphasize that while this event had a large impact on human interest, it was a relatively low-precipitation event. The total measured snow accumulation for the day at CYOW was 3.6 cm, very similar to the precipitation accumulation in Illinois in the event studied by Pettegrew et al. (2009). At KMSS, only a trace of liquid equivalent precipitation was recorded by the Automated Surface Observing System (ASOS) on 28 January 2010, although it is likely that this is in part due to the high winds and blowing snow observed at KMSS during the time of heaviest snowfall. Although KMSS is not a manual observation station and snow accumulation reports are not available, it is very possible that this event was associated with high snow to liquid water ratios. To that end, the measured liquid equivalent precipitation at CYOW was 1.6 mm, suggesting a snow to liquid water ratio of approximately 22:1. Precipitation data from a cooperative observation station in Ogdensburg, New York (58 km southwest of KMSS and 80 km south of CYOW), was downloaded from the National Climatic Data Center (NCDC) and showed that 2 in. (5.1 cm) of snow was observed on 28 January 2010. Therefore, it is reasonable to assume that KMSS received a similar total amount of snow compared to CYOW. At the Ogdensburg cooperative observation station, 0.04 in. of liquid equivalent was recorded, suggesting an even higher snow to liquid ratio (approximately 50:1).
The impact of these snow bursts was large, particularly in the CYOW area, where several multicar pileups and dozens of automobile accidents were reported (Jackson 2010). Moreover, several minor injuries were reported and at least one person (a 13-year-old boy) was critically injured when he was struck by a car in Ottawa around 1500 local time, or just after the passage of SB2 (Jackson 2010). In all, 18 vehicle accidents were reported in Ottawa during SB2 alone (Spears 2010), prompting the Ottawa police to issue a press release for motorists to use extreme vigilance in the area (Spears 2010).
Environment Canada’s Ontario Storm Prediction Centre did not issue any warnings for eastern Ontario during this event, as snow bursts did not fit their warning criteria for any type of wintertime event in the province. However, a special weather statement was issued between t = −1 and 0 h that called for “brief, but intense bursts of snow over eastern Ontario.” Meanwhile, the Quebec Storm Prediction Centre in Montreal issued a “snow squall warning” for regions in western Quebec around t = 0 h, while SB1 was passing through CYOW. Warning criteria vary from one province/territory to another, which explains the different handling of the same event by these two storm prediction center; however, neither is ideal in effectively communicating the risks associated with snow bursts. The significant societal impacts that resulted from the 28 January 2010 event clearly highlight the need for the development of a broadly defined advisory or warning that can be applied to events such as this one, as first mentioned by DeVoir (2004). These points are discussed further in section 5, following the dynamic and thermodynamic analysis of the event.
4. Dynamic and thermodynamic analysis
Both Doswell et al. (1996) and Schultz and Schumacher (1999) argue for the use of an ingredients-based methodology in the diagnosis of particular varieties of moist convection (gravitational and slantwise). Moreover, Wetzel and Martin (2001) made a similar argument for the analysis and operational prediction of midlatitude winter precipitation events, highlighting five specific ingredients. Three ingredients are common to all three studies: lift, instability, and moisture. We apply this approach to the analysis in this section, because even though southern Canada in January is not a region in which one would typically consider occurrences of moist convection, this case has all the markings of such an event. Furthermore, while this event clearly did not produce a large amount of precipitation (section 3), it contained enough moisture and high snow to liquid water ratios to create havoc for the general public. Since the largest impacts to the public occurred in the CYOW region, the focus of the following analysis will be on this location.
Figure 3 displays 850–500-hPa Q-vector divergence, sea level pressure (SLP), and 925-hPa equivalent potential temperature (θe). Figures 3a and 3b show a broad area of 850–500-hPa layer-averaged Q-vector convergence, forcing QG ascent, located west of CYOW and just downstream of a midtropospheric trough (Fig. 4). By t = −6 h (Fig. 3c), CYOW is located in the center of the large area of Q-vector convergence, which has strengthened relative to t = −18 h (Fig. 3a). Simultaneously, a large low-level θe gradient is present over the region, supporting the earlier assertion of an Arctic cold front approaching and subsequently moving through the CYOW region. By t = −3 and 0 h (Figs. 3d and 3e), CYOW remains in the region of ascent forcing, as a weak sea level cyclone develops in this area. In fact, at t = 0 h, CYOW is clearly located in an area of cold-air advection, which, in considering Eq. (1), suggests that the main mechanism responsible for QG ascent is cyclonic vorticity advection. Regardless, it is clear that CYOW is located in a large and persistent region of Q-vector convergence (and thus QG ascent) from t = −12 to 0 h, when SB1 passed through the station. By t = +3 h (Fig. 3f), most of the Q-vector convergence moves into Quebec as the snow burst event in eastern Ontario comes to an end. Additionally, there is a large sea level pressure gradient between a strong high pressure system over the central United States and a weak low pressure system over eastern Ontario and western Quebec; this gradient certainly suggests relatively strong surface winds at the time of the snow bursts, which likely contributed to the low visibilities in the region by causing blowing and drifting snow. Finally, it is clear from Fig. 3 that CYOW and KMSS were located in a region of quasigeostrophic forcing for ascent. Thus, we now examine possible focus mechanisms for ascent below, to help explain why the forcing for ascent is so broad and the response (i.e., the snow bursts) is relatively narrow.
Time series of NARR SLP (hPa, solid contours), 925-hPa equivalent potential temperature (K, dashed), and 850–500-hPa layer-averaged Q-vector divergence × 10−18 K m−2 s−1 (shaded cool colors for convergence, warm colors for divergence) on 28 Jan at (a) 0000 UTC (t = −18 h), (b) 0600 UTC (t = −12 h), (c) 1200 UTC (t = −6 h), (d) 1500 UTC (t = −3 h), (e) 1800 UTC (t = 0 h), and (f) 2100 UTC (t = +3 h). The approximate location of CYOW is marked with a black star.
Citation: Weather and Forecasting 26, 3; 10.1175/2010WAF2222432.1
As in Fig. 3, but for potential temperature (K, shaded) on the dynamic tropopause (2-PVU surface), 10-m wind (barbs), and coupling index (solid, every 4 K from 0 to +16).
Citation: Weather and Forecasting 26, 3; 10.1175/2010WAF2222432.1
Pettegrew et al. (2009) found substantial low-level (925 hPa) frontogenesis associated with the snow burst in Illinois. Furthermore, frontogenesis has been suggested to play an important role in intense mesoscale bands of snow, as mentioned by Nicosia and Grumm (1999) and Hunter et al. (2001). Frontogenesis helps to serve as a focusing mechanism for ascent through a thermally direct circulation, supporting rising motion on the warm side of the cold front (Martin 2006, p. 201). Figure 5 displays NARR 925-hPa total wind frontogenesis, 10-m wind, 1000–500-hPa thickness, and steep values (≥8 K km−1) of the 925–700-hPa lapse rate. At t = −18 h (Fig. 5a), two bands of 925-hPa frontogenesis are present: one is located in a line stretching from northeastern Ontario southward through Lake Michigan and into the Mississippi River valley, while the second is located farther to the northwest in northwestern Ontario. Both are indications of a strengthening cold front (Fig. 3) ahead of an Arctic air mass. By t = −6 h (Fig. 5c), the first band of frontogenesis has intensified on the leading edge of the frontal zone and is located just west of CYOW. Simultaneously, steep (8 K km−1) low-level lapse rates are present in the area of largest frontogenesis (Fig. 5c). A small area of weak frontogenesis in the St. Lawrence River valley (south of CYOW) is also evident at t = −6 h (Fig. 5c). While a full diagnosis is beyond the scope of this paper, we believe this frontogenesis is at least partially related to the local topography, which typically channels surface northeasterly winds down the river valley toward lower pressure, often creating low-level frontogenesis when opposed by geostrophic southerlies. Such a pressure pattern is evident in Fig. 3c and this process might help to explain the pre-SB1 band of snow along the St. Lawrence River that is seen in Figs. 1a–c. By t = −3 and 0 h (Figs. 5d and 5e), the frontogenesis throughout Ontario has markedly increased and the 925–700-hPa lapse rates have also steepened (>8 K km−1). Both the strongest frontogenesis and steepest lapse rates are located in or near the CYOW area at these times (Figs. 5d and 5e). This suggests that at the time of the snow bursts there was substantial forcing for ascent through frontogenesis in the CYOW area, as well as (at least) a conditionally unstable lower troposphere. These conditions proceeded to move eastward and by t = +3 h (Fig. 5f) were firmly entrenched near KMSS and extreme western Quebec, coinciding with the occurrence of the snow bursts in the KMSS and western Quebec region around t = +3 h. Finally, a band of strong frontogenesis is located in a north–south line in eastern New York State and western New England at t = −3 and 0 h (Figs. 5e and 5f). This observation coincides with several lines of intense snow that moved through this area at these times, and affected parts of western and central Massachusetts during rush hour (not shown).
As in Fig. 3, but for 925-hPa frontogenesis × 10−2 K (100 km)−1 (3 h)−1 (shaded), 925–700-hPa lapse rate (K km−1 with solid contours starting at −8 with an interval of 0.5), 1000–500-hPa thickness (dam, dashed), and 10-m wind (kt, barbs). A horizontal red rectangle is placed in (a) to represent the cross-section analysis in Fig. 6.
Citation: Weather and Forecasting 26, 3; 10.1175/2010WAF2222432.1
In Figs. 3–5, we have established that SB1 and SB2 in the CYOW and KMSS regions were clearly associated with strong QG forcing for ascent and strong values of low-level frontogenesis. Furthermore, we suggested that these snow bursts occurred in an environment characterized by low tropospheric stability, which we quantify in Figs. 6–8. Emanuel (1983) pointed out that conditional symmetric (slantwise) instability (CSI) can be largely responsible for some mesoscale precipitation bands within larger-scale storms. However, in cases when convective (potential) instability [dθe/dz < 0; Bluestein (1992), p. 222] is collocated with regions of CSI, the convective instability mode tends to dominate over time (Emanuel 1983; Moore and Lambert 1993; Schultz and Schumacher 1999). To assess both convective instability and CSI, a west–east cross-section analysis is presented in Fig. 6, along 45.4°N, roughly the latitude of CYOW; a line representing the cross-section area is included in Fig. 5a. The cross section runs from Petawawa, Ontario (approximately 170 km west of CYOW), to Granby, Quebec (approximately 370 km east of CYOW), allowing for the stability to be assessed across the entire region impacted by the snow bursts.
As in Fig. 3, but west–east cross-section plots from 45.4°N, −77.2°W to 45.4°N, −72.4°W, or approximately 170 km west of CYOW to approximately 150 km east of Montreal’s Trudeau International Airport (CYUL). A horizontal red rectangle is in Fig. 5a to represent the cross-section area. Plotted are saturated equivalent geostrophic potential vorticity (×10−7 m2 s−1 K kg−1 shaded for negative values) and equivalent potential temperature (K, solid contours). The location of CYOW is marked on each panel.
Citation: Weather and Forecasting 26, 3; 10.1175/2010WAF2222432.1
NARR soundings for CYOW on 28 Jan at (a) 0000 UTC (t = −18 h), (b) 0600 UTC (t = −12 h), (c) 1200 UTC (t = −6 h), (d) 1500 UTC (t = −3 h), (e) 1800 UTC (t = 0 h), and (f) 2100 UTC (t = +3 h). Temperature (dewpoint) is plotted in red (green).
Citation: Weather and Forecasting 26, 3; 10.1175/2010WAF2222432.1
Figures 6a–d show a layer of
Areas of frontogenesis and CSI are often collocated in regions of intense mesoscale precipitation bands (Schultz and Schumacher 1999; Nicosia and Grumm 1999). Here, we follow the guidelines of Schultz and Schumacher (1999), who state that the separation of forcing (frontogenesis) and response (the release of CSI) is challenging. This point is mitigated slightly by the apparent dominance of the convective instability mode. However, generally speaking, we conclude that both the forcing for ascent (QG forcing and frontogenesis) and instability (convective and conditional symmetric) are present at CYOW and KMSS during the passage of SB1 and SB2.
Bryan and Fritsch (2000) state that, traditionally, five different states of static stability are considered (where γ is the observed environmental lapse rate, Γs is the moist-adiabatic lapse rate, and Γd is the dry-adiabatic lapse rate):
1) absolutely stable, γ = Γs;
2) saturated neutral, γ = Γd;
3) conditionally unstable, Γs < γ < Γd;
4) dry neutral, γ = Γd; and
5) dry absolutely unstable, γ = Γd.
Bryan and Fritsch (2000) subsequently argue that a sixth type of static stability exists in nature: a “saturated lapse rate (γs) that is steeper than the moist-adiabatic lapse rate,” defined as
6) moist absolutely unstable, γs > Γs.
Such a state is suggested to be shallow, relatively short lived, and rare, as Bryan and Fritsch (2000) have found that soundings with a moist absolutely unstable layer (MAUL) of 100 hPa or greater composed only 1.1% of over 100 000 examined soundings in their dataset. Additionally, Bryan and Fritsch (2000) found that nearly all deep MAULs occurred in close proximity to moist convection, and that they are likely created by (and are indications of) intense mesoscale vertical motions. Bryan and Fritsch (2000) defined a deep MAUL as a layer of 100 hPa or more with a lapse rate greater than the moist-adiabatic lapse rate and a dewpoint depression ≤1°C. For the purpose of this study, we have allowed for a higher dewpoint depression given that the temperatures are substantially below zero and the lower troposphere is almost certainly saturated with respect to ice.
In Figs. 7 and 8 we show a time series of NARR soundings for CYOW and KMSS, respectively, on the day of the snow bursts. From t = −18 h (Fig. 7a) to t = −6 h (Fig. 7c), it is evident that a deep MAUL [according to the definition of Bryan and Fritsch (2000)] exists at CYOW in the near-surface layer up to about 850 hPa. However, by t = −3 h (Fig. 7d) and t = 0 h (Fig. 7e), this MAUL has grown to encompass approximately 400 hPa, from the surface to around 600 hPa. Figures 7 and 8 also serve to confirm the steep low-level lapse rates seen in Fig. 5. Additionally, the environmental conditions were quite similar at CYOW and KMSS, with a MAUL of approximately the same depth (400 hPa) as analyzed by the NARR at both stations. Finally, while most thundersnow events (Market et al. 2002) have been shown to occur in the presence of elevated instability, the case documented by Pettegrew et al. (2009) showed evidence of being rooted in the boundary layer with the sort of structure one would expect from a warm-season squall line. Figures 7 and 8 show a similar result to Pettegrew et al. (2009) for the cases presented in this paper. It would be interesting to determine which one of these situations is climatologically more common during intense wintertime convection, in association with the passage of an Arctic front, but that is beyond the scope of this manuscript. Finally, we note that while rawinsonde sounding observations took place at times (0000 and 1200 UTC) before and after the snow bursts moved through the area, the 0000 UTC 29 January sounding (not shown) from Maniwaki, Quebec (CWMW), showed structures very similar to the aforementioned NARR soundings.
5. Concluding discussion and future work
On the afternoon of 28 January 2010, two intense snow bursts (SB1 and SB2) moved through eastern Ontario, western Quebec, and extreme northern New York State, creating low visibilities, falling temperatures, and strong winds, in association with the passage of an Arctic cold front. In following with the ingredients-based methodology for moist convection, proposed by Doswell et al. (1996) and Schultz and Schumacher (1999), we have established that there were large amounts of lift and instability in this case, with just enough moisture (3.6 cm of snow at CYOW) and high snow to liquid water ratios to create havoc on area roads and several injuries throughout eastern Ontario. Specifically, with regard to lift, Ottawa, Ontario (CYOW), and Massena, New York (KMSS), were located in an environment of strong Q-vector convergence (e.g., QG ascent, see Fig. 3), ahead of a trough with very low-θ air on the dynamic tropopause (DT, Fig. 4). Moreover, Fig. 5 shows a rapidly intensifying band of low-level frontogenesis in the region just before the two snow bursts passed through CYOW and KMSS.
In terms of instability, Fig. 4 shows relatively low values of the coupling index (CI) present in the region, suggesting very low tropospheric stability. Figure 6 displays an area of convective instability in the near-surface layer at least 18 h prior to SB1 moving through CYOW. As the time of snow burst passage (t = 0 h) approached, this layer of convective instability expanded and intensified, and was concurrent with a layer of conditional symmetric instability (CSI) above, throughout all of eastern Ontario, western Quebec, and northern New York State. While Fig. 6 suggests that the convective instability mode was dominant, it is possible for convective instability and CSI to both be factors over short periods of time (Schultz and Schumacher 1999). Finally, the soundings in Figs. 7 and 8 corroborate the steep low-level (925–700 hPa) lapse rates in Fig. 5 at the time of the snow bursts, by showing that a moist absolutely unstable layer (MAUL) of approximately 400 hPa was present from the surface to 600 hPa at both CYOW and KMSS. MAULs of such depth are rare, difficult to maintain, and almost always associated with moist convection and very large values of mesoscale ascent (Bryan and Fritsch 2000). Therefore, it is clear from the diagnostics in this paper that this event occurred in a markedly unstable environment, and one that differed from the thundersnow climatology of Market et al. (2002) in which elevated convection was the primary observation, but similar to the case of Pettegrew et al. (2009), in which convection was rooted in the near-surface boundary layer.
We suggest that the snow bursts occurred in the narrow region (including CYOW and KMSS) where the large area of Q-vector convergence (Figs. 3e and 3f) and low CI values (Figs. 4e and 4f) intersect. This may help to explain why the snow bursts were relatively narrow in scope, despite the large area of synoptic-scale forcing for ascent (Q-vector convergence). This corresponds to the area of convective instability observed in Fig. 6, particularly at t = 0 h and t = +3 h, when the snow bursts crossed the region (Figs. 6e and 6f).
As previously stated, little in the way of published literature exists on this topic and much work remains to be done. Specifically, many more events similar to this one should be examined, as not all Arctic front passages are associated with intense convective snow bursts. For example, a climatology of similar and null events would allow diagnostics to determine what tools an operational forecaster can use to better assess the occurrence potential of such an event, both on the short- and medium-range time scales.
While the forecasting of such an event remains a challenge given the unique conditions required and the infrequency of wintertime convection in Canada, it is not impossible to predict the occurrence of snow bursts. An examination of the NCEP Global Forecast System (GFS) forecasts (not shown) shows that it accurately predicted the dynamic and thermodynamic structures evident in the NARR up to 72 h prior to the occurrence of the snow bursts. Given that Environment Canada uses the Global Environmental Multiscale (GEM) model to produce their forecasts, we acknowledge that using the GFS to analyze operational model performance is this particular case may be a little misleading and that ideally model verification should be performed using the GEM model grids. Regardless, detailed operational model verification is beyond the scope of this paper and should be a topic of future investigation.
The issue, however, is less so that these types of events are predictable and more so that within the wintertime advisories and warnings available, forecast offices cannot adequately express the expected conditions and threats associated with snow bursts. For instance, the Ontario Storm Prediction Center (OSPC) accurately called for 2–4 cm of snow throughout much of eastern Ontario on the day of the snow bursts (3.6 cm was measured at CYOW), but the forecast implied gradual “flurries over the course of the day.” Apart from the high winds also forecast, there was no reason for the public to expect brief whiteout conditions. A special weather statement was issued for CYOW 20 min before SB1 hit the city stating that “narrow but intense bands of snow are moving through southeastern Ontario… Because of the short duration, snowfall accumulations are not expected to be significant; however visibilities may be reduced to a few hundred metres [meters] or less at times.” While the special weather statement accurately expressed the conditions observed in CYOW, these types of statements are not given the same priority as watches and warnings, and presently are not disseminated to the public through the media and Weatheradio, etc. It is unlikely, especially given the short 20-min window, that this special weather statement available on the Weatheroffice Web site reached a wide audience prior to SB1 moving through the area. For the same event, shortly after both squalls were evident on radar and moving through eastern Ontario, the Quebec Storm Prediction Centre (QSPC) issued a snow squall warning for western Quebec. A snow squall warning, as defined by EC (Meteorological Service of Canada 2010a), is “when, to the lee of the Great Lakes or other large lakes, snow squalls are expected and 15 cm or more of snow is likely to fall within 12 hours, OR the visibility is likely to be near 0 for 4 or more hours, even without warning level accumulations of snow.” This event was not forecast to meet the warning criteria yet the QSPC issued the warning anyway to ensure the public was informed of the hazardous weather conditions approaching the area. Neither approach adequately warns for, nor expresses, the threats associated with snow bursts.
Our goal is not to criticize the operational forecasters, or to try to differentiate between model forecast and human forecast errors. Instead, we would like to emphasize the fact that, as DeVoir (2004) pointed out for the United States, no warning or advisory exists that is specific to a high-impact, but short-duration and low-precipitation event such as this one. Much like advisories and warnings are issued and understood by the public regarding squalls (either winter or summer), similar messaging is needed for snow bursts. That being said, the issuance of an advisory or warning alone only goes part way to effectively communicating the threats and associated impacts of this type of event to the general public. However, combined with forecaster, public, and media education of these types of events and their impacts, for example through the engagement of the Warning Preparedness Meteorologists (WPM) of Environment Canada with Emergency Measures Organizations (EMOs) and media across Canada, a greater understanding of snow bursts would encourage appropriate actions to be taken by both the media and the public in their respective roles to reduce loss of life and property.
In conclusion, we strongly encourage the creation of a broadly applied warning or advisory (e.g., a “snow burst warning”) that is specific to this type of high-impact but short-duration and low-precipitation event, as proposed by DeVoir (2004), so that the public can be aware that a wintertime convective event is possible on a given day. If such warnings exist for warm-season convective events, we see little reason why they should not be available for cold-season convection. Alternatively, current warnings or advisories such as the snow squall warning could be amended so that a forecaster feels comfortable issuing such a warning or advisory for an event like the one studied here, without having to “stretch” the language of the written operational definition. Regardless of how such a change is achieved, it is crucial that the operational forecasters have all the necessary tools to alert the general public to forthcoming or imminent hazardous weather conditions.
Acknowledgments
This research has been supported by grants from the Natural Sciences and Engineering Research Council of Canada and the Canadian Foundation for Climate and Atmospheric Sciences. Thanks to the National Centers for Environmental Prediction (NCEP) and the National Climatic Data Center (NCDC) for providing access to the GFS analyses, and NARR, NCEP global reanalysis, and cooperative station data. The authors would also like to acknowledge the insightful and helpful comments of the three reviewers. Finally, thanks to Environment Canada for access to their online historical radar and climate information databases.
REFERENCES
Bluestein, H., 1992: Principles of Kinematics and Dynamics. Vol. 1, Synoptic–Dynamic Meteorology in Midlatitudes, Oxford University Press, 431 pp.
Bosart, L. F., and Lackmann G. M. , 1995: Postlandfall tropical cyclone intensification in a weakly baroclinic environment: A case study of Hurricane David (September 1979). Mon. Wea. Rev., 123, 3268–3291.
Bryan, G. H., and Fritsch J. M. , 2000: Moist absolute instability: The sixth static stability state. Bull. Amer. Meteor. Soc., 81, 1207–1230.
Corfidi, S. F., Corfidi S. J. , Imy D. A. , and Logan A. L. , 2006: A preliminary study of severe wind-producing MCSs in environments of limited moisture. Wea. Forecasting, 21, 715–734.
Crowe, C. C., Market P. S. , Pettegrew B. P. , Melick C. J. , and Podzimek J. , 2006: An investigation of thundersnow and deep snow accumulations. Geophys. Res. Lett., 33, L24812, doi:10.1029/2006GL028214.
DeVoir, G. A., 2004: High impact sub-advisory snow events: The need to effectively communicate the threat of short duration high intensity snowfall. Preprints, 20th Conf. on Weather Analysis and Forecasting/16th Conf. on Numerical Weather Prediction, Seattle, WA, Amer. Meteor. Soc., P10.2. [Available online at http://ams.confex.com/ams/pdfpapers/68261.pdf.]
Doswell, C. A., III, Brooks H. E. , and Maddox R. A. , 1996: Flash flood forecasting: An ingredients-based methodology. Wea. Forecasting, 11, 560–581.
Emanuel, K. A., 1983: On assessing local conditional symmetric instability from atmospheric soundings. Mon. Wea. Rev., 111, 2016–2033.
Galarneau, T. J., Jr., and Bosart L. F. , 2005: An examination of the long-lived MCV of 10–13 June 2003. Preprints, 11th Conf. on Mesoscale Processes/32nd Conf. on Radar Meteorology, Albuquerque, NM, Amer. Meteor. Soc., JP5J.7. [Available online at http://ams.confex.com/ams/pdfpapers/96598.pdf.]
Holle, R. L., and Watson A. I. , 1996: Lightning during two central U.S. winter precipitation events. Wea. Forecasting, 11, 599–614.
Homan, J., and Uccellini L. W. , 1987: Winter forecast problems associated with light to moderate snow events in the mid-Atlantic states on 14 and 22 February 1986. Wea. Forecasting, 2, 206–228.
Hoskins, B. J., Draghici I. , and Davies H. C. , 1978: A new look at the ω-equation. Quart. J. Roy. Meteor. Soc., 104, 31–38.
Hunter, S. M., Underwood S. J. , Holle R. L. , and Mote T. L. , 2001: Winter lightning and heavy frozen precipitation in the Southeast United States. Wea. Forecasting, 16, 478–490.
Jackson, K., 2010: Winter havoc blankets city, 13-year-old boy critically injured in one of many accidents. Ottawa Sun, 29 January, 4.
Kalnay, E., and Coauthors, 1996: The NCEP/NCAR 40-Year Reanalysis Project. Bull. Amer. Meteor. Soc., 77, 437–471.
Koch, S., DesJardins M. , and Kocin P. , 1983: An interactive Barnes objective map analysis scheme for use with satellite and conventional data. J. Climate Appl. Meteor., 22, 1487–1503.
Lundstedt, W., 1993: A method to forecast wintertime instability and non-lake effect snow squalls across northern New England. NWS Eastern Region Tech. Attach. 93-11A, 13 pp.
Market, P. S., Halcomb C. E. , and Ebert R. L. , 2002: A climatology of thundersnow events over the contiguous United States. Wea. Forecasting, 17, 1290–1294.
Market, P. S., and Coauthors, 2006: Proximity soundings of thundersnow in the central United States. J. Geophys. Res., 111, D19208, doi:10.1029/2006JD007061.
Martin, J. E., 2006: Mid-Latitude Atmospheric Dynamics. John Wiley and Sons, 324 pp.
Mesinger, F., and Coauthors, 2006: North American Regional Reanalysis. Bull. Amer. Meteor. Soc., 87, 343–360.
Meteorological Service of Canada, cited 2010a: PUBPRO: Manual of standard operational procedures for issuing public weather forecasts and warnings. [Available online at http://www.msc.ec.gc.ca/msb/manuals/pubpro/index_e.cfm.]
Meteorological Service of Canada, cited 2010b: The National Radar Program, frequently asked questions. [Available online at http://www.msc-smc.ec.gc.ca/projects/nrp/faq_e.cfm.]
Moore, J. T., and Lambert T. E. , 1993: The use of equivalent potential vorticity to diagnose regions of conditional symmetric instability. Wea. Forecasting, 8, 301–308.
Nicosia, D. J., and Grumm R. H. , 1999: Mesoscale band formation in three major northeastern United States snowstorms. Wea. Forecasting, 14, 346–368.
Niziol, T. A., Snyder W. R. , and Waldstreicher J. S. , 1995: Winter weather forecasting throughout the eastern United States. Part IV: Lake effect snow. Wea. Forecasting, 10, 61–77.
Payer, M., Desrochers J. , and Laird N. F. , 2007: A lake-effect snowband over Lake Champlain. Mon. Wea. Rev., 135, 3895–3900.
Pettegrew, B. P., Market P. S. , Wolf R. A. , Holle R. L. , and Demetriades N. W. S. , 2009: A case study of severe winter convection in the Midwest. Wea. Forecasting, 24, 121–139.
Roebber, P. J., and Gyakum J. R. , 2003: Orographic influences on the mesoscale structure of the 1998 ice storm. Mon. Wea. Rev., 131, 27–50.
Schultz, D. M., 1999: Lake-effect snowstorms in northern Utah and western New York with and without lightning. Wea. Forecasting, 14, 1023–1031.
Schultz, D. M., and Schumacher P. N. , 1999: The use and misuse of conditional symmetric instability. Mon. Wea. Rev., 127, 2709–2732.
Spears, T., 2010: Boy, 13, critical after struck by vehicle; young pedestrian one of many victims of snow squalls. Ottawa Citizen, 29 January, C1.
Steenburgh, W. J., and Onton D. J. , 2001: Multiscale analysis of the 7 December 1998 Great Salt Lake–effect snowstorm. Mon. Wea. Rev., 129, 1296–1317.
Steenburgh, W. J., Halvorson S. F. , and Onton D. J. , 2000: Climatology of lake-effect snowstorms of the Great Salt Lake. Mon. Wea. Rev., 128, 709–727.
Trapp, R. J., Schultz D. M. , Ryzhkov A. V. , and Holle R. L. , 2001: Multiscale structure and evolution of an Oklahoma winter precipitation event. Mon. Wea. Rev., 129, 486–501.
van den Broeke, M. S., Schultz D. M. , Johns R. H. , Evans J. S. , and Hales J. E. , 2005: Cloud-to-ground lightning production in strongly forced, low-instability convective lines associated with damaging wind. Wea. Forecasting, 20, 517–530.
Wetzel, S. W., and Martin J. E. , 2001: An operational ingredients-based methodology for forecasting midlatitude winter season precipitation. Wea. Forecasting, 16, 156–167.
