Abstract

Lake-effect snowstorms east of Lake Ontario are frequently intense and contribute to substantial seasonal accumulations, especially over the Tug Hill Plateau (hereafter Tug Hill), which rises at a gentle 1.25% slope to ~500 m above lake level. Using a variety of datasets including radar imagery from the KTYX (Fort Drum, New York) WSR-88D, this paper examines the characteristics of lake-effect precipitation east of Lake Ontario over 13 cool seasons (16 September 2001–15 May 2014). During this period, days with at least 2 h of lake effect account for 61%–76% of the mean cool-season snowfall and 24%–37% of the mean cool-season liquid precipitation. Mean monthly lake-effect frequency and snowfall peak in December and January. The highest lake-effect frequency and snowfall occur over the western and upper Tug Hill, with an arm of relatively high lake-effect frequency and snowfall extending to the southeast shore of Lake Ontario. To the east (lee), lake-effect frequency and snowfall decrease abruptly over the Black River valley, although relatively high frequency and snowfall extend downstream into the western Adirondack Mountains. Broad coverage and long-lake-axis-parallel (LLAP) bands dominate the lake-effect morphology throughout the region. There is no diurnal modulation of lake-effect frequency during winter, but weak modulation in fall and spring, especially of LLAP bands.

Collectively, these results quantify the role that lake effect plays in the cool-season hydroclimate east of Lake Ontario. The increase in lake-effect frequency and snowfall over Tug Hill suggest an inland/orographic intensification of many lake-effect systems, with evidence for shadowing in the lee.

1. Introduction

The Great Lakes of North America produce frequent, sometimes intense, lake-effect snowstorms during the cool season (e.g., Niziol et al. 1995). High snowfall rates, low visibility, and heavy accumulations impact commerce and transportation, but also contribute to a vibrant winter-sports economy (Tug Hill Commission 2014). The region east of Lake Ontario, in particular, observes some of the most intense snowstorms in the world, many enhanced over the Tug Hill Plateau (hereafter Tug Hill). World-record snowfall accumulations observed in this region include 30.5 cm (12 in.) in 1 h at Copenhagen, New York; 44.5 cm (17.5 in.) in 2 h at Oswego, New York; and 129.5 cm (51 in.) in 16 h at Bennetts Bridge, New York (Burt 2007). A 24-h snowfall of 195.6 cm (77 in.) occurred on Tug Hill from 11–12 January 1997, but was based on six measurements instead of four and is considered unofficial (Leffler et al. 1997). Much of Tug Hill averages over 500 cm (200 in.) of snow per year, with 1173 cm (462 in.) observed in Hooker, New York, during the 1976–77 cool season.

Lake effect occurs when cold air passing over a warm body of water leads to the initiation of moist convection (Markowski and Richardson 2010). Most wintertime (December–February) lake-effect precipitation in the Great Lakes region falls as snow, although fall (September–November) events can produce rain, especially prior to early November (Miner and Fritsch 1997). On average, lake-effect frequency and snowfall over the Great Lakes increase through the fall, peak in early winter, and then decrease as the lakes cool and, in some winters, become partially or fully ice covered (Niziol et al. 1995). Kristovich and Spinar (2005) suggest that the lake-effect precipitation frequency in the Great Lakes region is highest in the overnight/early morning hours and lowest in the afternoon. Steenburgh et al. (2000) and Alcott et al. (2012) describe a similar diurnal modulation of lake-effect frequency over the Great Salt Lake of Utah.

Synoptic, mesoscale, lake-surface, and land surface conditions influence the areal coverage, intensity, and organization of lake-effect precipitation systems, leading to a rich morphological spectrum that includes the following:

  1. Wind-parallel bands generated by land-breeze convergence when the prevailing flow is oriented along the long axis of an elongated body of water (e.g., Peace and Sykes 1966; Passarelli and Braham 1981; Braham 1983; Hjelmfelt 1990; Niziol et al. 1995; Steenburgh et al. 2000; Alcott et al. 2012). Sometimes called midlake or shoreline bands (Laird and Kristovich 2004), following Steiger et al. (2013), we refer to this morphological regime as long-lake-axis-parallel (LLAP) bands. LLAP bands can extend well inland and typically produce the largest snowfall rates and accumulations (Niziol et al. 1995; Steiger et al. 2013).

  2. Broad coverage events that feature open-cellular convection or multiple, quasi-periodic wind-parallel bands produced by horizontal roll convection (e.g., Kelly 1982, 1984, 1986; Kristovich 1993; Laird and Kristovich 2004).

  3. Hybrid events that have characteristics of LLAP bands and broad coverage events, typically with an apparent connection to a LLAP band from an upstream lake (e.g., Lake Huron or Georgian Bay). This morphological regime is synonymous with the hybrid classification of Niziol et al. (1995), who referred to LLAP band and broad-coverage events as type I and II, respectively. Although it is possible to have other regime hybrids, this is the most common.

  4. Shoreline bands that are generated by land-breeze convergence during cold, relatively calm conditions near the center of anticyclones. Although Laird and Kristovich (2004) merge shoreline and LLAP bands into a single category, we maintain a separate category since shoreline bands remain close to the shoreline and exhibit little movement due to weak flow (e.g., Kelly 1986; Hjelmfelt 1990; Niziol et al. 1995). We distinguish shoreline bands from LLAP bands by requiring the former to penetrate no more than 20 km inland and feature slow or nonexistent movement of cells within the band.

  5. Mesoscale vortices that form during weak flow, often where the lakeshore has a bowl-shaped configuration (Forbes and Merritt 1984; Niziol et al. 1995; Laird 1999).

Within many lake-effect regions, there exists the compounding influence of orography. For example, cold, persistent flow across the Sea of Japan during the Asian winter monsoon encounters numerous mountains over 1500 m MSL (some over 3000 m MSL) and produces heavy snow in western Japan (e.g., Magono et al. 1966; Hozumi and Magono 1984; Matsuura et al. 2005; Steenburgh 2014). In northern Utah, lake-effect systems interact with mountains that rise up to 2300 m above the Great Salt Lake (Steenburgh et al. 2000; Steenburgh and Onton 2001; Onton and Steenburgh 2001; Alcott et al. 2012; Yeager et al. 2013). Alcott and Steenburgh (2013) showed that the influence of mountains during one Great Salt Lake–effect event was multifaceted. Upstream terrain forced a drying, foehnlike flow that was detrimental to lake-effect formation, whereas the concave shape created by topography downstream of the lake reinforced the lake-breeze-induced convergence zone, greatly enhancing the storm intensity. Climatological analysis by Yeager et al. (2013) shows a factor of four increase in mean precipitation produced during lake-effect periods from the lowlands to the mountains southeast of the Great Salt Lake.

Although more modest, the orography of the Great Lakes region also has an important influence on lake-effect storms. As noted by Niziol et al. (1995), “the greatest snowfall occurs where the prevailing winds blow across the longest fetch of the lake, particularly where orographic features enhance precipitation processes.” Wilson (1977) utilized radar and precipitation gauge data to produce a year-long precipitation analysis for the Lake Ontario basin, and similar to Muller (1966) and Hill (1971), notes that annual snowfall increases considerably with elevation in the terrain downstream of Lake Ontario. Using numerical simulations, Hjelmfelt (1992) showed that even the minor terrain downstream of Lake Michigan enhances lake-effect precipitation.

Over a distance of ~40 km, Tug Hill rises at a gradual slope of ~1.25% to ~500 m above Lake Ontario (Fig. 1). Mean September–May liquid precipitation equivalent (LPE) based on National Weather Service Cooperative Observer (COOP) data (1994–2014) increases from 83.7 cm (33.0 in.) at Watertown [WTN (151 m MSL)] just north of Tug Hill to 109 cm (42.9 in.) at Hooker [HKR (448 m MSL)] on Tug Hill (see Fig. 1 for locations). Mean September–May snowfall increases from 288 cm (113 in.) to 571 cm (225 in.), although these sites may not be well situated to identify the orographic precipitation gradient or the area of heaviest snowfall on Tug Hill. In addition to weather impacts, precipitation and snowfall enhancement over Tug Hill affect water availability and hydroelectric power generation (Norton and Bolsenga 1993), as well as spatial patterns of forests and other ecological systems (e.g., Henne et al. 2007).

Fig. 1.

Topographic and geographic features of the study region. Abbreviations include Bennetts Bridge COOP site (BBR), Big Moose 3SE COOP site (BMO), location of NARR data used for conditional climatologies (GP1), Hooker 12NNW COOP site (HKR), Highmarket COOP site (HMK), WSR-88D (KTYX), Lowville COOP site (LWV), North Redfield NWS spotter (NRD), Oswego East COOP site (OSW), and Watertown COOP site (WTN). Elevation (m MSL) is based on the scale at the left. The yellow line circumscribes the Maple Ridge Wind Farm.

Fig. 1.

Topographic and geographic features of the study region. Abbreviations include Bennetts Bridge COOP site (BBR), Big Moose 3SE COOP site (BMO), location of NARR data used for conditional climatologies (GP1), Hooker 12NNW COOP site (HKR), Highmarket COOP site (HMK), WSR-88D (KTYX), Lowville COOP site (LWV), North Redfield NWS spotter (NRD), Oswego East COOP site (OSW), and Watertown COOP site (WTN). Elevation (m MSL) is based on the scale at the left. The yellow line circumscribes the Maple Ridge Wind Farm.

Nevertheless, despite dramatic influences on individual storms and climatological snowfall and precipitation, the influence of Tug Hill on lake-effect systems remains poorly documented and understood. In this research, we develop a radar-based climatology to provide new insights into lake-effect precipitation east of Lake Ontario and over Tug Hill, with relevance for operational forecasting, regional climate applications, and improved knowledge of lake-effect and orographic precipitation processes. The methods used for the radar-based climatology are described in section 2, with detailed analysis of the regional lake-effect characteristics and influence of Tug Hill presented in section 3. Conclusions and future work are summarized in section 4.

2. Data and methods

a. Event identification and morphological classification

Lake-effect periods (LEPs) were identified subjectively using all available lowest-elevation tilt (0.5°) base reflectivity imagery from the KTYX (Ft. Drum, New York) WSR-88D during the cool season (16 September–15 May) from 16 September 2001–15 May 2014. Radar data were obtained from the National Climatic Data Center Hierarchical Data Storage System in level III format (Crum et al. 1993). Data were missing for 4.3% of the study period. Data from LEPs that were partly affected by an outage were included in the analysis, but those with start or end times during an outage were omitted from the calculations of diurnal behavior. No effort was made to identify or include LEPs during long outages.

Following Laird et al. (2009) and Alcott et al. (2012), LEPs were identified as periods ≥2 h (the aforementioned studies used ≥1 h) during which precipitation features were 1) coherent and quasi stationary with a distinct connection to the lake, 2) shallow and distinguishable from large transitory synoptic features, and 3) exhibiting increasing depth and/or intensity in the downwind direction. This definition includes periods when lake-effect precipitation features occur concurrently with synoptically forced and other non-lake-effect precipitation. The ≥2-h threshold was adopted to simplify the identification process as there are a large number of short-lived (≤2 h) precipitation features that develop over Lake Ontario but produce little precipitation.

Although the manual identification process is subjective, the criteria listed above were developed to make the process as objective as possible. Lake-effect identification was frequently straightforward, but we consulted surface observations, upper-air soundings, and satellite images when event identification was ambiguous based on radar alone. For example, a few events appeared to marginally meet the radar characteristics of lake effect, but examination of the regional upper-air soundings indicated that lake-surface air parcels would likely have no positive buoyancy, and thus any precipitation features observed near this time would not be lake effect. These ambiguous cases were infrequent, short lived, and have little influence on the overall statistics.

During LEPs, the lake-effect morphology was identified based on the categories described in the introduction [i.e., LLAP (usually 1 but in some cases 2 bands), broad coverage, hybrid, shoreline, and mesoscale vortex], plus lake orographic and miscellaneous (no clear fit into any of the aforementioned categories). Examples of all but miscellaneous are shown in Fig. 2. The lake-orographic category featured a stationary precipitation area confined to Tug Hill with only weak or intermittent lake-effect features developing over the Lake Ontario. Start and end times for the existence of a morphological category were recorded, with evidence of the morphological structure for at least 1 h required for inclusion. Two or more morphological types were noted in instances where they occurred concurrently. For example, LLAP bands sometimes occurred concurrently with broad coverage.

Fig. 2.

Lowest-tilt (0.5°) base reflectivity [dBZ, based on inset scale in (f)] examples of lake-effect morphological types: (a) LLAP, (b) broad coverage, (c) hybrid, (d) mesoscale vortex, (e) lake orographic (concurrent with LLAP band to north), and (f) shoreline.

Fig. 2.

Lowest-tilt (0.5°) base reflectivity [dBZ, based on inset scale in (f)] examples of lake-effect morphological types: (a) LLAP, (b) broad coverage, (c) hybrid, (d) mesoscale vortex, (e) lake orographic (concurrent with LLAP band to north), and (f) shoreline.

b. Radar capabilities, limitations, and statistics

The KTYX WSR-88D is located at 562 m MSL on Tug Hill (Fig. 1). The fairly continuous data availability over the 13 cool-season study period, combined with the proximity to and view over Tug Hill and the eastern shore of Lake Ontario, allows for a unique perspective of lake-effect precipitation in the region. Nevertheless, some problems are caused by the Maple Ridge Wind Farm along the eastern side of Tug Hill (Fig. 1), which reduces the down-radial signal power and produces multipath/interturbine scattering, indirect echoes, and spurious high reflectivity values (Radar Operations Center 2014; National Weather Service Forecast Office Burlington, VT 2014a). This affects radar returns near the wind farm and in down-radial areas over the Black River valley and western Adirondack Mountains (hereafter Adirondacks, see Fig. 1 for locations), as discussed where necessary later in the analysis. Unrelated to the wind farm, there are occasionally blocked radials or clutter removal in the vicinity of the ~135° and ~340° azimuths. Bragg scattering was observed near the radar during a small number of LEPs. We elected not to remove these events since they were infrequent and thus have minimal impact on results. Finally, although beam height is low enough to detect precipitation falling over and near Tug Hill, lake-effect convection is sometimes confined to below 2 km MSL, and the centroid of the 0.5° scan reaches this altitude at about 100-km range (Fig. 3). Therefore, lake-effect features may go undetected or only partially fill the beam at longer ranges (Brown et al. 2007). In addition, low-level hydrometeor growth or sublimation may be undetected.

Fig. 3.

(a) Height of the KTYX beam centroids with horizontal range at various elevation angles. (b) Height of the KTYX 0.5° beam centroid above ground level. Adapted from Brown et al. (2007).

Fig. 3.

(a) Height of the KTYX beam centroids with horizontal range at various elevation angles. (b) Height of the KTYX 0.5° beam centroid above ground level. Adapted from Brown et al. (2007).

Spatial statistics were generated using the 0.5° base-reflectivity scans to better understand the distribution and intensity of precipitation during the LEPs. Although the time between volume scans can vary from 5–12 min due to changes in radar scan mode, no attempt was made to apply a time-weighting algorithm because the events with longer-interval scans typically feature weaker, less persistent echoes and have a correspondingly small influence on the results.

c. Additional datasets

Conditional radar statistics were generated using data from the North American Regional Reanalysis (NARR; Mesinger et al. 2006), which was obtained from the National Climatic Data Center (NCDC). Specifically, we used NARR wind data interpolated to 43.7°N, 76.2°W (GP1 in Fig. 1), ~2.5 km upstream of the eastern shore. Although the use of reanalysis data has its limitations, low-frequency, remoteness, and questionable representativeness precludes the use of upper-air observations from the nearest regular upper-air sounding sites in Buffalo and Albany, New York. The NARR has been used for similar climatological studies and provides a reasonable representation of the large-scale atmospheric flow conditions (e.g., Alcott and Steenburgh 2010; Kennedy et al. 2011; Walters et al. 2014).

Snow depth and snowpack SWE analyses were obtained from the National Operational Hydrologic Remote Sensing Center (NOHRSC) Snow Data Assimilation System (SNODAS; National Operational Hydrologic Remote Sensing Center 2004). SNODAS uses RUC/RAP snowfall analysis data, ground observations, data from airborne platforms, and spaceborne microwave remote sensing to estimate snow depth and SWE at 1-km spatial and 24-h temporal resolution with data coverage beginning in 2003. Station snowfall and LPE data for COOP stations were obtained from NCDC. In addition, we obtained and used a temporally complete record of daily snowfall observations during the study period from a National Weather Service spotter at North Redfield, New York (NRD), which is not included in the NCDC archive.

3. Results

a. LEP characteristics and intraseasonal variability

A total of 636 LEPs were identified during the 13 cool-season study period. The mean (median) LEP duration was 19.5 (13.2) h, with the longest event lasting 237 h (9.9 days; 3–12 February 2007) and producing 358 cm (141 in.) of snow at NRD on the western slope of Tug Hill (National Weather Service Forecast Office Buffalo, NY 2014b). Although short LEPs are the most common, with the frequency decreasing with increasing duration, there were 42 that lasted ≥48 h, yielding a mean of 3.2 per cool season (Fig. 4).

Fig. 4.

LEP durations.

Fig. 4.

LEP durations.

The most active month for lake effect was January, with 2862 h (29.5% of the hours during that month), followed closely by December (29.4% of the hours; Fig. 5). May was the least active month, with only 2 LEPs during the study period (0.6% of the hours). The decline in lake-effect frequency from January through the end of the cool season is generally attributed to decreasing lake-surface temperatures and increasing air temperatures (e.g., Niziol et al. 1995). Lake Ontario rarely develops significant ice cover, so freezing of the lake is not likely a major factor in the decrease, although the freezing of upstream lakes could have some influence. This seasonal cycle is consistent with other studies in the Great Lakes region (e.g., Ruhf and Cutrim 2003; Kristovich and Spinar 2005), but contrasts with that found over the Great Salt Lake of Utah where the lake-effect frequency features a bimodal distribution with peaks in the fall and spring (Alcott et al. 2012). This unique bimodal distribution reflects the shallow (3-m-average depth), hypersaline composition of the Great Salt Lake, which prevents freezing and leads to rapid warming in the spring, in contrast to the deeper, freshwater Lake Ontario.

Fig. 5.

Mean frequency (%) of lake effect by month. Asterisk indicates half month (i.e., 1–15 May or 16–30 Sep).

Fig. 5.

Mean frequency (%) of lake effect by month. Asterisk indicates half month (i.e., 1–15 May or 16–30 Sep).

b. Interannual variability and contributions to hydroclimate

Lake-effect hours varied from 577 during the 2012 cool season (defined based on the ending calendar year) to 1308 during the 2007 cool season (Fig. 6), with a mean of 956 and a standard deviation of 177. The 2007 cool season also featured the aforementioned 237-h LEP. The number of LEPs per cool season ranged from 41 to 66 (not shown). Alcott et al. (2012) documented a larger interannual variability in Great Salt Lake effect, with cool seasons experiencing anywhere from 3 to 20 LEPs.

Fig. 6.

Hours of lake effect by cool season.

Fig. 6.

Hours of lake effect by cool season.

The hours of lake effect, however, do not account for coverage and intensity, which ultimately play an important role in the snow climate of the region. Therefore, data from available COOP sites and the NRD NWS spotter site were analyzed for the 13 cool-season study period to estimate the lake-effect fraction [i.e., the fraction of snowfall or LPE generated on lake-effect days (LEDs), which are defined as any day with at least 2 h of lake effect]. Because only daily (i.e., 24 h) snowfall and LPE observations are available (0800–0800 EST for COOP stations and 0000–0000 EST for North Redfield), two types of LEDs are identified, the first with lake effect occurring for the entire 24-h observation day (LED24) and the second with lake effect occurring for ≥2 h but <24 h (LED2). Lake-effect fractions based solely on LED24 observation days are likely underestimates of the total cool-season or monthly lake-effect fractions since they exclude some LEPs, whereas lake-effect fractions based on the combined fraction produced on LED24 and LED2 observation days (i.e., LED24 + LED2) are likely overestimates since they may include some non-lake-effect precipitation.

The mean cool-season snowfall (i.e., that produced on LEDs and non-LEDs) is significantly greater over Tug Hill compared to lowland sites (Table 1). The two sites over 400 m MSL [HKR and Highmarket (HMK)] have a mean cool-season snowfall >500 cm and LPE > 110 cm, whereas WTN at 152 m MSL observes 304 and 88 cm, respectively (Table 1). Further, on the western slope of Tug Hill, NRD (399 m MSL) observes a mean cool-season snowfall of 718 cm, the highest of any site in the area (LPE data not available). As shown later, the greater snowfall at NRD agrees well with spatial radar statistics, which place the area of greatest snowfall on the western Tug Hill near NRD, south of HKR, and north of HMK.

Table 1.

Surface observing site information and climatology. Mean cool-season LPE and snowfall are for the study period (2001–14 cool seasons).

Surface observing site information and climatology. Mean cool-season LPE and snowfall are for the study period (2001–14 cool seasons).
Surface observing site information and climatology. Mean cool-season LPE and snowfall are for the study period (2001–14 cool seasons).

The cool-season lake-effect snowfall fraction based on LED24 + LED2 observation days is ≥47% for all sites and cool seasons and reaches as high as 90% during the 2002 cool season at HMK and 2013 cool season at NRD (Fig. 7). The study period (2001–14) cool-season mean at near-lake and Tug Hill sites exceeds 70%, with a lesser but still sizeable mean of 61% at Big Moose 3SE (BMO) in the western Adirondacks (Table 1). Although the lake-effect snowfall fractions based solely on LED24 observation days are lower, they still represent a large portion of the snowfall for most sites and cool seasons (Fig. 7). The study period cool-season mean LED24 lake-effect snowfall fraction is ≥49% at all sites except BMO in the western Adirondacks (31%), and is as high as 55% at HKR and HMK (Table 1).

Fig. 7.

Cool-season snowfall during the 13-yr study period at (a) WTN, (b) HKR, (c) LWV, (d) NRD, (e) BMO, (f) OSW, (g) BBR, and (h) HMK. Portion produced on LED24, LED2, and non-lake-effect observing days indicated in light blue, dark blue, and black, respectively. Combined fraction produced on LED24 and LED2 observing days (i.e., LED24 + LED2) annotated.

Fig. 7.

Cool-season snowfall during the 13-yr study period at (a) WTN, (b) HKR, (c) LWV, (d) NRD, (e) BMO, (f) OSW, (g) BBR, and (h) HMK. Portion produced on LED24, LED2, and non-lake-effect observing days indicated in light blue, dark blue, and black, respectively. Combined fraction produced on LED24 and LED2 observing days (i.e., LED24 + LED2) annotated.

In contrast, the cool-season lake-effect LPE fractions based on LED24 + LED2 observation days range from only 13% at BMO during the 2013 cool season to 63% at HKR during the 2009 cool season (Fig. 8). The study period cool-season LPE means are also much lower than found for snowfall with the highest values of 38%, 37%, and 42% at HMK, Bennetts Bridge (BBR), and HKR, respectively (Table 1).

Fig. 8.

As in Fig. 7, but for cool-season LPE. North Redfield not available.

Fig. 8.

As in Fig. 7, but for cool-season LPE. North Redfield not available.

The monthly lake-effect snowfall fractions based on either LED24 or LED24 + LED2 observation days are highest in the fall and lowest in the spring (Fig. 9). Nearly all of the snow in October and November falls on LED24 and LED2 observation days, which is plausible given that anomalously cold air is needed to generate snow instead of rain during that time of year and lake effect is frequently generated during such cold-air intrusions. In contrast, during the spring, the monthly lake-effect snowfall fractions based on either LED24 or LED24 + LED2 observation days are lower as lake effect becomes less frequent and synoptic systems dominate snowfall accumulations. In contrast, the monthly lake-effect LPE fractions peak in December and January instead of October and November, likely due to the large amounts of non-lake-effect rain observed in the fall (Fig. 10). One notable non-lake-effect contribution during the study period came from the remnants of Tropical Storm Nicole, which brought widespread rainfall amounts >7 cm to the region on 1 October 2010.

Fig. 9.

As in Fig. 7, but for mean monthly snowfall. The asterisk indicates that September and May include data only for the first and latter half of each month, respectively.

Fig. 9.

As in Fig. 7, but for mean monthly snowfall. The asterisk indicates that September and May include data only for the first and latter half of each month, respectively.

Fig. 10.

As in Fig. 9, but for mean monthly LPE.

Fig. 10.

As in Fig. 9, but for mean monthly LPE.

The larger cool-season (Fig. 7) and monthly (Fig. 9) lake-effect snowfall fractions compared to lake-effect LPE (Figs. 8 and 10) fractions reflects at least three factors. First, lake-effect snow is typically low density, with mean snow-to-liquid ratios of 14:1–16:1 (Baxter et al. 2005), so the snowfall contribution from these events is prodigious for a given LPE. Second, the study region experiences regular synoptic rain events throughout the cool season (including midwinter), which dilutes the lake-effect contribution to LPE. Third, many precipitation gauges are unshielded, and undercatch may artificially lower LPE during lake-effect snow events (Rasmussen et al. 2012; Yeager et al. 2013). Nonetheless, the fact that in some seasons, >50% of the cool-season LPE falls on LED24 + LED2 observation days at locations like HMK, BBR, and HKR, much in the form of snow, shows the importance of lake effect for the regional hydroclimate.

c. Diurnal variability

During spring and fall, there is a tendency for lake effect to initiate near/after sunset (Figs. 11a,c) that is significant at the 95% level, but this trend is absent during winter (Fig. 11b).1 For the dissipation time, there is little signal in either the spring or winter (Figs. 11e,f), with only fall showing a tendency for LEPs to end near or following sunrise (significant at the 95% level, Fig. 11d). Overall, the frequency of lake effect shows a weak, broad minimum in the afternoon [~1800–2300 UTC (1300–1800 EST)] and a weak, broad maximum overnight [~0400–0800 UTC (2300–0300 EST)] in the fall and spring, both of which are significant at the 95% level (Figs. 12a,b). There is no significant signal (at the 95% level) during the winter (Fig. 12c).

Fig. 11.

Initiation time of LEPs relative to sunset during (a) fall (SON), (b) winter (DJF), and (c) spring (MAM). (d)–(f) As in (a)–(c), but for dissipation time relative to sunrise.

Fig. 11.

Initiation time of LEPs relative to sunset during (a) fall (SON), (b) winter (DJF), and (c) spring (MAM). (d)–(f) As in (a)–(c), but for dissipation time relative to sunrise.

Fig. 12.

Number of lake-effect events observed by time of day during (a) fall (SON), (b) winter (DJF), and (c) spring (MAM). Range of sunrise (SR) and sunset (SS) times annotated.

Fig. 12.

Number of lake-effect events observed by time of day during (a) fall (SON), (b) winter (DJF), and (c) spring (MAM). Range of sunrise (SR) and sunset (SS) times annotated.

Previous work from Kristovich and Spinar (2005) describes a similar afternoon minimum and overnight maximum in precipitation on lake-effect days over Lakes Superior and Michigan. They define a lake-effect day (0000–0000 EST) based on the presence of lake-effect features in visible satellite imagery during the illuminated portion of the day, with hourly precipitation data used to examine the diurnal cycle. Their diurnal signal appears to be stronger than found in our radar-based analysis, which simply requires the presence of lake-effect precipitation structures with no requirement for accumulation, so our findings are not equivalent. Alcott et al. (2012) found a pronounced late night/early morning maximum and afternoon minimum over the Great Salt Lake, with the strongest diurnal variation in the spring and weakest in the winter. Collectively, these results suggest that the diurnal modulation of lake effect is weakest during the winter and strongest during the fall and spring, especially for smaller bodies of water like the Great Salt Lake. Interestingly, Miner and Fritsch (1997) found that fall lake-effect rainfall events downwind of Lake Erie are more intense in the afternoon and early evening. Their lake-effect definition, however, was very broad and included any persistent (≥6 h) echoes occurring with lower-tropospheric cold advection within a 400 × 400 km2 box centered on Buffalo.

As discussed by Kristovich and Spinar (2005), a variety of factors may be responsible for this diurnal modulation, but they point to the sensible heat flux maximum (caused by a maximum in lake–air temperature difference) in the early morning hours as likely being the dominant mechanism. Other studies have also pointed to this as a possible cause (e.g., Hjelmfelt 1990). On the other hand, the diurnal modulation of land-breeze circulations and the afternoon decline in relative humidity accompanying mixed-layer growth have been suggested as contributors over the Great Salt Lake (e.g., Steenburgh et al. 2000; Alcott et al. 2012). The prevailing atmospheric conditions over eastern Lake Ontario are often quite different, however, than those over the Great Salt Lake. The eastern United States and Canada observe higher relative humidity on an annual mean basis than Utah, many lake-effect events on Lake Ontario feature an airflow and boundary layer that have been modified by upstream lakes (e.g., Rodriguez et al. 2007), and Lake Ontario is much larger, enabling a greater modification of air masses compared to the Great Salt Lake. Further work is needed to fully understand the diurnal modulation of lake effect and its variability across differing climates and water bodies.

d. Spatial characteristics derived from radar

Radar echoes ≥10 dBZ during LEPs are located most frequently over the western and upper Tug Hill, with an arm of locally high frequencies extending to the southeast shore of Lake Ontario near Mexico Bay (Fig. 13a). The latter reflects the high frequency of LLAP events near Mexico Bay. Unfortunately, beam blockage and clutter removal produce unrealistically low frequencies along some radials near the 135° and 340° azimuths and the Maple Ridge Wind Farm produces nonmeteorologically high frequencies over the eastern Tug Hill, as well as high-frequency spikes over the western Adirondacks due to multipath scattering and indirect echoes. The wind farm artifacts are not present in a comparable analysis for the period prior to the Maple Ridge Wind Farm installation (i.e., 16 September 2001–15 May 2004), but the overall pattern remains intact (Fig. 14). The frequency of reflectivities ≥30 dBZ for 2001–14 reflects the most intense lake-effect precipitation and features a similar pattern (Fig. 13b), although the influence of blocking and clutter removal is even more apparent over much of the upper Tug Hill.

Fig. 13.

(a) Frequency of echoes ≥10 dBZ during LEPs with mean (2001–14) cool-season snowfall on LED24 + LED2 observation days annotated. (b) As in (a), but for echoes ≥30 dBZ.

Fig. 13.

(a) Frequency of echoes ≥10 dBZ during LEPs with mean (2001–14) cool-season snowfall on LED24 + LED2 observation days annotated. (b) As in (a), but for echoes ≥30 dBZ.

Fig. 14.

As in Fig. 13a, but for the cool seasons in the period before the installation of the Maple Ridge Wind Farm (2001–04).

Fig. 14.

As in Fig. 13a, but for the cool seasons in the period before the installation of the Maple Ridge Wind Farm (2001–04).

NRD, on the western slope and near the highest frequency of echoes ≥10 and ≥30 dBZ, has a mean cool-season snowfall on LED24 + LED2 observation days of 539 cm (Table 1 and Figs. 13a,b, see Fig. 1 for station locations). Although falling in the region affected by blocking and clutter removal, HKR and HMK appear to lie on the edge of the core of highest reflectivities ≥10 and ≥30 dBZ and have mean cool-season snowfalls on LED24 + LED2 observation days of 439 and 413 cm, respectively. Collectively, these results suggest that the maximum in lake-effect precipitation and snowfall likely lies on the western or upper Tug Hill near or east of NRD, south of HKR, and north of HMK.

The frequency of echoes ≥10 dBZ decreases sharply to the immediate lee of Tug Hill, reaching a minimum over the Black River valley, and then increasing slightly over the western Adirondacks. Frequencies are much lower over the Adirondacks than over Tug Hill, although multipath scattering and indirect echoes from the Maple Ridge Wind Farm, as well as ground clutter, produce some spurious high values (Fig. 13a). The frequency of radar echoes ≥30 dBZ is very low over most of the western Adirondacks, implying that intense lake-effect echoes are rare in this region, although this may reflect partial filling of the radar beam at longer ranges due to overshooting (Fig. 13b).

Given the significant fraction of regional snowfall produced on LEDs, the mean maximum snowpack SWE from SNODAS generally exhibits a spatial pattern similar to the above radar-based analysis (cf. Figs. 13 and 15; SNODAS analysis based on data since 2003). High SWE is found over Tug Hill, with an extension of locally high values to the southeast shore of Lake Ontario near Mexico Bay. Farther east, a minimum is found over the Black River valley. SWE values over the western Adirondacks, however, are comparable to those found on Tug Hill, in contrast to the frequencies of radar echoes ≥10 and ≥30 dBZ, which are lower. It is possible that overshooting causes lake effect to be poorly sampled over the western Adirondacks, or that the greater SWE reflects a larger contribution of precipitation from non-lake-effect events. The latter is likely important since the mean cool-season snowfall on LED24 + LED2 observation days is only 216 cm at BMO, the lowest of all sites in the region (cf. Figs. 13 and 15).

Fig. 15.

Mean (2003–14) maximum snowpack SWE from SNODAS with mean (2001–14) cool-season snowfall on LED24 + LED2 observation days annotated.

Fig. 15.

Mean (2003–14) maximum snowpack SWE from SNODAS with mean (2001–14) cool-season snowfall on LED24 + LED2 observation days annotated.

The frequency of echoes ≥10 dBZ was subdivided based on the mean 975–850-hPa wind direction in the NARR on the eastern shore of Lake Ontario (see Fig. 1 for location). Frequencies for each subdivision are relative only to that subdivision, not relative to LEPs as a whole. For flows from 240°–249° to 250°–259°, the northern and central Tug Hill observe the highest echo frequencies, but high values extend well to the lee of Tug Hill to the northwest Adirondacks (Figs. 16a,b). For flows from 260°–269° to 270°–279°, the central Tug Hill is favored, but the echo frequency decreases more abruptly over the lee slope, with lower values over the Black River valley and western Adirondacks (Figs. 16c,d), although this decline may be affected by beam blockage by the Maple Ridge Wind Farm and overshooting over the western Adirondacks.

Fig. 16.

Frequency of echoes ≥10 dBZ for volume scans during which the mean 975–850-hPa wind direction is (a) 240°–249°, (b) 250°–259°, (c) 260°–269°, (d) 270°–279°, (e) 280°–289°, and (f) 290°–299°. The middle value in each direction range is indicated by a black arrow.

Fig. 16.

Frequency of echoes ≥10 dBZ for volume scans during which the mean 975–850-hPa wind direction is (a) 240°–249°, (b) 250°–259°, (c) 260°–269°, (d) 270°–279°, (e) 280°–289°, and (f) 290°–299°. The middle value in each direction range is indicated by a black arrow.

As the flow veers from 240°–249° to 270°–279°, contributions from LLAP events become more substantial, leading to an increasingly prominent arm of high frequency over southeast Lake Ontario and the adjoining lowlands (Figs. 16a–d). As the flow veers farther to 280°–289° and 290°–299°, however, echo frequencies weaken, indicating less intense or spatially coherent LEPs, with the maximum shifted to lowland areas downstream of southeast Lake Ontario (Figs. 16e,f).

For most flow directions, there is a clear broadening of the echo frequency maximum over Tug Hill, illustrating that precipitation over Tug Hill tends to not only be more intense and persistent during LEPs, but also more extensive geographically. For example, a LLAP band may broaden or be accompanied by an area of orographic precipitation over Tug Hill to the north or south. The broadening of the echo frequency maximum over Tug Hill is especially evident with flow directions from 280°–289° and 290°–299° when the area of most frequent echoes is displaced well to the south of Tug Hill, but there remains an area of moderate echo frequency over the central and northern Tug Hill. There is very low echo frequency over Lake Ontario upstream of the central and northern Tug Hill with these flow directions, suggesting orographic generation/intensification of precipitation over Tug Hill (and outside of the main lake-effect precipitation band or bands) during some events.

Contoured frequency by distance diagrams (CFDD) of radar reflectivity along zonal transects crossing Tug Hill further highlight the lake and orographic influences during LEPs (Fig. 17). The southernmost transect, Ta, exhibits an increase in reflectivity frequency at all intensities >~5 dBZ from Lake Ontario to a broad maximum over the western Tug Hill, somewhat west of the crest (Fig. 17a). For reflectivities <~5 dBZ, there is a weak minimum over the western slope that could reflect the more persistent nature of stronger (i.e., >~5 dBZ) echoes in this region. An abrupt decline for all reflectivities occurs over the lee slopes with a minimum over the Black River valley. A weak secondary maximum lies over the Adirondacks. Although these results are generally consistent with the earlier analysis of observed snowfall and snowpack SWE observations, the low reflectivity frequencies at longer ranges likely partially reflect the partial or full overshooting of shallow lake-effect systems. In addition, echo frequencies are also affected by blockage and clutter removal over upper Tug Hill and by spurious returns produced by the Maple Ridge Wind Farm from upper Tug Hill to the Adirondacks. Two pronounced echo frequency spikes over the Adirondacks in Fig. 17 provide evidence of the latter.

Fig. 17.

Contoured frequency of reflectivity by distance (CFDD) diagrams along zonal transects at (a) 43.47°N, (b) 43.63°N, and (c) 43.86°N. Black dotted line indicates mean reflectivity value. The lower inset shows the surface elevation (m MSL, brown) and the height of the 0.5° beam centroid under standard atmospheric refraction conditions (black). The lower-right inset shows position of each transect and surface elevation (m MSL). The location of NRD is indicated in (b).

Fig. 17.

Contoured frequency of reflectivity by distance (CFDD) diagrams along zonal transects at (a) 43.47°N, (b) 43.63°N, and (c) 43.86°N. Black dotted line indicates mean reflectivity value. The lower inset shows the surface elevation (m MSL, brown) and the height of the 0.5° beam centroid under standard atmospheric refraction conditions (black). The lower-right inset shows position of each transect and surface elevation (m MSL). The location of NRD is indicated in (b).

The central transect, Tb, which runs through NRD, features an increase in reflectivity frequency at intensities >~5 dBZ from Lake Ontario to a pronounced peak near NRD (Fig. 17b). Recall that NRD is the snowiest observing site in the region with a mean total cool-season snowfall of 718 cm and a mean cool-season snowfall on LED24 + LED2 observation days of 539 cm. A local minimum in the frequency of reflectivities <5 dBZ is also found near NRD and, as suggested previously, could reflect the more persistent nature of stronger (i.e., >~5 dBZ) echoes on the western slope of Tug Hill during LEPs. The Maple Ridge Wind Farm has a stronger influence along this transect, causing spurious returns over the eastern edge of Tug Hill. In addition, the decline in echo frequency at all intensities just west of the upper Tug Hill could reflect beam blockage and clutter removal. Thus, it is unclear if lake-effect frequency and snowfall maximize over the western slope near NRD or farther east on the upper Tug Hill.

Structurally, the pattern is similar over the northern transect, although the echo frequency maximum appears to be centered nearer the crest (Fig. 17c). Unfortunately, artifacts related to the Maple Ridge Wind Farm strongly influence results over and east of Tug Hill.

e. Morphological characteristics

Using the morphological classification scheme described in section 2a, broad coverage was found to be the most common morphology, followed by LLAP, with 10 626 and 3018 h, respectively (Fig. 18). This includes 8900 h of broad coverage occurring in isolation, 1291 h of LLAP in isolation, and 1726 h with both occurring simultaneously. The other five morphology types, miscellaneous, shoreline, lake orographic, hybrid, and MCV, were less common, with a total of 259, 254, 226, 106, and 20 h, respectively.

Fig. 18.

Hours of lake effect for broad coverage (BC), broad coverage only (*BC), LLAP (LL), BC/LL (both occurring concurrently), LLAP only (*LL), miscellaneous (MSC), shoreline (SL), lake orographic (LO), hybrid (HYB), and mesoscale vortex (MCV).

Fig. 18.

Hours of lake effect for broad coverage (BC), broad coverage only (*BC), LLAP (LL), BC/LL (both occurring concurrently), LLAP only (*LL), miscellaneous (MSC), shoreline (SL), lake orographic (LO), hybrid (HYB), and mesoscale vortex (MCV).

Diurnal behavior was analyzed for periods when broad coverage or LLAP occurred in isolation. Although neither broad coverage nor LLAP exhibit a significant diurnal signal in the winter, LLAP exhibits a stronger diurnal signal in the fall and spring compared to broad coverage and LEPs in general (cf. Figs. 12 and 19). During the fall, when the signal is strongest, LLAP is 4 times more likely during the morning maximum (0800–0900 LST) than during the afternoon minimum [1500–1600 LST (Fig. 19d)]. We hypothesize that the dependence of LLAP on land-breeze circulations (e.g., Peace and Sykes 1966; Hjelmfelt 1990; Steiger et al. 2013), which typically develop or strengthen at night and weaken during the day, contributes to this diurnal modulation.

Fig. 19.

As in Fig. 12, but for (a)–(c) broad coverage periods and (d)–(f) LLAP periods.

Fig. 19.

As in Fig. 12, but for (a)–(c) broad coverage periods and (d)–(f) LLAP periods.

The frequency of echoes ≥10 dBZ for broad coverage periods closely resembles the frequency of echoes ≥10 dBZ for all events, likely due to the fact that broad coverage comprises ~70% of the total hours of lake effect (cf. Figs. 13a and 20a; frequencies calculated for broad coverage and other morphological types in Fig. 20 are relative only to each type, not relative to LEPs as a whole). The analysis for LLAP periods shows the favored location for these bands, with a strip of high frequency near Mexico Bay that broadens inland and over Tug Hill (Fig. 20b). Although this strip may seem a bit south of that expected based on conceptual models showing convergent land breezes meeting in the center of the lake (e.g., LaDue 1996), the southern shoreline near Mexico Bay actually lies closer to the central axis of Lake Ontario as a whole. Nevertheless, further work is needed to fully elucidate the causes of the high frequency of echoes during LLAP events near Mexico Bay.

Fig. 20.

Frequency of echoes ≥10 dBZ during (a) broad coverage only, (b) LLAP only, (c) shoreline only, and (d) lake-orographic-only periods.

Fig. 20.

Frequency of echoes ≥10 dBZ during (a) broad coverage only, (b) LLAP only, (c) shoreline only, and (d) lake-orographic-only periods.

A zonal CFDD for LLAP periods illustrates some of the orographic effects over Tug Hill during these intense events, with a very pronounced peak at all reflectivities >~10 dBZ over the western slope near NRD. For reflectivities <~10 dBZ, there is a local minimum over the western slope near NRD, suggesting precipitation is more intense and persistent over Tug Hill than over the lake and coastal lowlands. Compared with LEPs as a whole, LLAP events feature a higher frequency of the strongest reflectivities and a lower frequency of the weakest reflectivities, especially near NRD (cf. Figs. 17b and 21).

Fig. 21.

As in Fig. 17b, but for LLAP-only periods.

Fig. 21.

As in Fig. 17b, but for LLAP-only periods.

The echo frequency for shoreline periods shows a frequency maximum just offshore and parallel to the central portion of the southern shore of Lake Ontario (Fig. 20c). Precipitation features during these periods tend to form along the land-breeze front and, with weak flow, rarely penetrate more than ~10 km inland, although there is a localized coastal protrusion of frequent returns southeast of Oswego, New York (OSW, see Fig. 1 for location).

Echoes corresponding to the lake-orographic morphology exhibit a frequency maximum centered on Tug Hill (Fig. 20d). The signatures from LLAP bands that accompanied some of these periods are evident to the north and south of Tug Hill. The difference in echo frequency near the lake shore (<30%) and over the upper Tug Hill (> 80%) is the largest observed for any morphology type and illustrates the dramatic orographic enhancement process that can occur over Tug Hill.

4. Conclusions

This study has examined the characteristics of lake-effect precipitation east of Lake Ontario and over Tug Hill. During the 13 cool-season (16 September–15 May) study period (2001–14, with the year based on the ending cool-season month), we identified 636 lake-effect periods (LEPs). Lake-effect hours totaled 12 434 (a mean of 956 per cool season), increased through the fall to a broad maximum in December and January, and then decreased through the spring. The mean (median) LEP length was 19.5 (13.2) h, with a mean of 3.2 events per year lasting >48 h. By cool season, lake-effect hours ranged from 577 in 2012 to 1308 in 2007.

Although not apparent in winter, a diurnal cycle of decreased (increased) lake-effect frequency is evident in the afternoon (night/early morning) during the fall and spring. This is in broad agreement with previous studies such as Kristovich and Spinar (2005) over the Great Lakes and Alcott et al. (2012) over the Great Salt Lake, although the amplitude of the signal found here is weaker, especially compared to the Great Salt Lake. The diurnal variability for LLAP events is more amplified than that for LEPs in general, although even this morphological type shows little diurnal modulation during the winter. Overall, broad coverage is the most common morphological type (10 626 h), followed by LLAP (3018 h), with the other five types (miscellaneous, shoreline, lake orographic, hybrid, and MCV) considerably less common.

Our analysis confirms that lake effect is the dominant contributor to snowfall in the region, with a lesser, but still sizeable, contribution to liquid precipitation equivalent (LPE). Days with ≥2 h of lake effect account for 61%–76% of the mean cool-season snowfall and 24%–37% of the mean cool-season LPE, depending on location. The fraction of monthly snowfall produced on these lake-effect days peaks in the fall, whereas the fraction of monthly LPE peaks in December and January.

Spatial patterns of radar reflectivity frequencies during all LEPs exhibit an arm of high frequency along the southeast shore of Lake Ontario near Mexico Bay with a broadening and intensification to a maximum over Tug Hill. These radar reflectivity frequency analyses, combined with surface observations and snowpack SWE analyses, suggest that the maximum in lake-effect snowfall and snowpack SWE lies on the western or upper Tug Hill near or east of North Redfield (NRD). Although full and partial overshooting affect radar statistics at longer ranges, these datasets also show a clear increase in snowfall, LPE, and snowpack SWE from the coastal lowlands to the western and upper Tug Hill where lake-effect precipitation systems frequently become steadier, stronger, and more geographically extensive. To the lee, blocked radials and scattering/signal power extinction from the Maple Ridge Wind Farm complicate the analysis, but there is evidence for a sharp decrease in echo frequency and snowfall over the Black River valley, followed by a slight increase farther east over the western Adirondacks.

Windward and leeward influences of Tug Hill are apparent across a wide range of low-level 975–850-hPa flow directions (Figs. 16a–f), with echo frequency calculations for winds from 280° to 299° (Figs. 16e,f) suggesting orographic generation/intensification of precipitation over central/northern Tug Hill, even when more coherent lake-effect features are located well to the south. This is important, as the influence of Tug Hill is not limited to areas within the primary band(s) of lake-effect precipitation. Nevertheless, further study is needed to elucidate the role of orographic processes in storm intensification over Tug Hill. The recently completed Ontario Winter Lake-Effect Systems (OWLeS) project should provide valuable data for examining orographic processes and their contribution to individual storms east of Lake Ontario and over Tug Hill.

Acknowledgments

We thank Carol Yerdon for providing a comprehensive record of snowfall observations at North Redfield and gratefully acknowledge the provision of datasets, software, and/or computer time and services by NCDC, NCEP, NCAR, Unidata, and the University of Utah Center for High Performance Computing. Comments and suggestions by Leah Campbell, John Horel, Jeff Massey, Court Strong, and two anonymous reviewers aided this research and improved the manuscript. This material is based upon work supported by National Science Foundation Grants AGS-0938611 and AGS-1262090 and NOAA/National Weather Service C-STAR Grant NA13NWS5680003. Any opinions, findings, conclusions, or recommendations expressed are those of the authors and do not necessarily reflect those of the National Science Foundation or the NOAA/National Weather Service.

REFERENCES

REFERENCES
Alcott
,
T. I.
, and
W. J.
Steenburgh
,
2010
:
Snow-to-liquid ratio variability and prediction at a high-elevation site in Utah’s Wasatch Mountains
.
Wea. Forecasting
,
25
,
323
337
, doi:.
Alcott
,
T. I.
, and
W. J.
Steenburgh
,
2013
:
Orographic influences on a Great Salt Lake–effect snowstorm
.
Mon. Wea. Rev.
,
141
,
2432
2450
, doi:.
Alcott
,
T. I.
,
W. J.
Steenburgh
, and
N. F.
Laird
,
2012
:
Great Salt Lake–effect precipitation: Observed frequency, characteristics, and associated environmental factors
.
Wea. Forecasting
,
27
,
954
971
, doi:.
Baxter
,
M. A.
,
C. E.
Graves
, and
J. T.
Moore
,
2005
:
A climatology of snow-to-liquid ratio for the contiguous United States
.
Wea. Forecasting
,
20
,
729
744
, doi:.
Braham
,
R. R.
,
1983
:
The Midwest snow storm of 8–11 December 1977
.
Mon. Wea. Rev.
,
111
,
253
272
, doi:.
Brown
,
R. A.
,
T. A.
Niziol
,
N. R.
Donaldson
,
P. I.
Joe
, and
V. T.
Wood
,
2007
:
Improved detection using negative elevation angles for mountaintop WSR-88Ds. Part III: Simulations of shallow convective activity over and around Lake Ontario
.
Wea. Forecasting
,
22
,
839
852
, doi:.
Burt
,
C. A.
,
2007
: Extreme Weather. 2nd ed. W. W. Norton and Company, 320 pp.
Crum
,
T. D.
,
R. L.
Alberty
, and
D. W.
Burgess
,
1993
:
Recording, archiving, and using WSR-88D data
.
Bull. Amer. Meteor. Soc.
,
74
,
645
653
, doi:.
Forbes
,
G. S.
, and
J. H.
Merritt
,
1984
:
Mesoscale vortices over the Great Lakes in wintertime
.
Mon. Wea. Rev.
,
112
,
377
381
, doi:.
Henne
,
P. D.
,
F. S.
Hu
, and
D. T.
Cleland
,
2007
:
Lake-effect snow as the dominant control of mesic-forest distribution in Michigan, USA
.
J. Ecol.
,
95
,
517
529
, doi:.
Hill
,
J. D.
,
1971
: Snow squalls in the lee of Lakes Erie and Ontario. NOAA Tech. Memo. NWS ER-43, 20 pp.
Hjelmfelt
,
M. R.
,
1990
:
Numerical study of the influence of environmental conditions on lake-effect snowstorms over Lake Michigan
.
Mon. Wea. Rev.
,
118
,
138
150
, doi:.
Hjelmfelt
,
M. R.
,
1992
:
Orographic effects in simulated lake-effect snowstorms over Lake Michigan
.
Mon. Wea. Rev.
,
120
,
373
377
, doi:.
Hozumi
,
K.
, and
C.
Magono
,
1984
:
The cloud structure of convergent cloud bands over the Japan Sea in winter monsoon period
.
J. Meteor. Soc. Japan
,
62
,
522
533
.
Kelly
,
R. D.
,
1982
:
A single Doppler radar study of horizontal-roll convection in a lake-effect snow storm
.
J. Atmos. Sci.
,
39
,
1521
1531
, doi:.
Kelly
,
R. D.
,
1984
:
Horizontal roll and boundary-layer interrelationships observed over Lake Michigan
.
J. Atmos. Sci.
,
41
,
1816
1826
, doi:.
Kelly
,
R. D.
,
1986
:
Mesoscale frequencies and seasonal snowfalls for different types of Lake Michigan snow storms
.
J. Climate Appl. Meteor.
,
25
,
308
312
, doi:.
Kennedy
,
A. D.
,
X.
Dong
,
B.
Xi
,
S.
Xie
,
Y.
Zhang
, and
J.
Chen
,
2011
:
A comparison of MERRA and NARR reanalyses with the DOE ARM SGP data
.
J. Climate
,
24
,
4541
4557
, doi:.
Kristovich
,
D. A. R.
,
1993
:
Mean circulations of boundary-layer rolls in lake-effect snow storms
.
Bound.-Layer Meteor.
,
63
,
293
315
, doi:.
Kristovich
,
D. A. R.
, and
M. L.
Spinar
,
2005
:
Diurnal variations in lake-effect precipitation near the western Great Lakes
.
J. Hydrometeor.
,
6
,
210
218
, doi:.
LaDue
,
J.
,
1996
: COMET course notes and satellite meteorology. [Available online at http://www.comet.ucar.edu/class/smfaculty/byrd/sld027.htm].
Laird
,
N. F.
,
1999
:
Observation of coexisting mesoscale lake-effect vortices over the western Great Lakes
.
Mon. Wea. Rev.
,
127
,
1137
1141
, doi:.
Laird
,
N. F.
, and
D. A. R.
Kristovich
,
2004
:
Comparison of observations with idealized model results for a method to resolve winter lake-effect mesoscale morphology
.
Mon. Wea. Rev.
,
132
,
1093
1103
, doi:.
Laird
,
N. F.
,
J.
Desrochers
, and
M.
Payer
,
2009
:
Climatology of lake-effect precipitation events over Lake Champlain
.
J. Appl. Meteor. Climatol.
,
48
,
232
250
, doi:.
Leffler
,
R. J.
,
R. M.
Downs
,
G. W.
Goodge
,
N. J.
Doesken
,
K. L.
Eggleston
, and
D.
Robinson
,
1997
: Evaluation of the reported January 11–12, 1997, Montague, New York, 77-inch, 24-hour lake-effect snowfall. National Weather Service Special Rep., 60 pp. [Available online at http://www1.ncdc.noaa.gov/pub/data/cmb/extremes/ncec/mantague-ny-snowfall-24hour.pdf.]
Magono
,
C.
,
K.
Kikuchi
,
T.
Kimura
,
S.
Tazawa
, and
T.
Kasai
,
1966
:
A study on the snowfall in the winter monsoon season in Hokkaido with special reference to low land snowfall
. J. Fac. Sci. Hokkaido Univ. Ser. 7,
11
,
287
308
.
Markowski
,
P.
, and
Y.
Richardson
,
2010
: Mesoscale Meteorology in Midlatitudes. Wiley-Blackwell, 407 pp.
Matsuura
,
S.
,
K.
Matsuyama
,
S.
Asano
,
T.
Okamoto
, and
Y.
Takeuchi
,
2005
:
Fluctuation of the seasonal snowpack in a mountainous area of the heavy-snow district in the warm-temperate zone of Japan
.
J. Glaciol.
,
51
,
547
554
, doi:.
Mesinger
,
F.
, and Coauthors
,
2006
:
North American Regional Reanalysis
.
Bull. Amer. Meteor. Soc.
,
87
,
343
360
, doi:.
Miner
,
T. J.
, and
J. M.
Fritsch
,
1997
:
Lake-effect rain events
.
Mon. Wea. Rev.
,
125
,
3231
3248
, doi:.
Muller
,
R. A.
,
1966
:
Snowbelts of the Great Lakes
.
Weatherwise
,
19
,
248
255
, doi:.
National Operational Hydrologic Remote Sensing Center
,
2004
: Snow Data Assimilation System (SNODAS) data products at NSIDC. Snowpack snow-water equivalent, Boulder, CO, National Snow and Ice Data Center. [Available online at http://nsidc.org/data/g02158.]
National Weather Service Forecast Office Burlington, VT
,
2014a
: National Weather Service WSR-88D radar and wind farm impacts. [Available online at http://www.erh.noaa.gov/btv/research/Wind_Farm/.]
National Weather Service Forecast Office Buffalo, NY
,
2014b
: Summary of lake-effect snow event over the Tug Hill February 3–12, 2007. [Available online at http://www.erh.noaa.gov/buf/locust/.]
Niziol
,
T. A.
,
W. R.
Snyder
, and
J. S.
Waldstreicher
,
1995
:
Winter weather forecasting throughout the eastern United States. Part IV: Lake effect snow
.
Wea. Forecasting
,
10
,
61
77
, doi:.
Norton
,
D. C.
, and
S. J.
Bolsenga
,
1993
:
Spatiotemporal trends in lake-effect and continental snowfall in the Laurentian Great Lakes, 1951–1980
.
J. Climate
,
6
,
1943
1956
, doi:.
Onton
,
D. J.
, and
W. J.
Steenburgh
,
2001
:
Diagnostic and sensitivity studies of the 7 December 1998 Great Salt Lake–effect snowstorm
.
Mon. Wea. Rev.
,
129
,
1318
1338
, doi:.
Passarelli
,
R. E.
, and
R. R.
Braham
,
1981
:
The role of the winter land breeze in the formation of Great Lake snow storms
.
Bull. Amer. Meteor. Soc.
,
62
,
482
492
, doi:.
Peace
,
R. L.
, and
R. B.
Sykes
,
1966
:
Mesoscale study of a lake effect snow storm
.
Mon. Wea. Rev.
,
94
,
495
507
, doi:.
Radar Operations Center
,
2014
: Wind farms: Coming soon to a WSR-88D near you. [Available online at http://www.roc.noaa.gov/wsr88d/NNOW/WindFarmSoon.aspx.]
Rasmussen
,
R.
, and Coauthors
,
2012
:
How well are we measuring snow: The NOAA/FAA/NCAR winter precipitation test bed
.
Bull. Amer. Meteor. Soc.
,
93
,
811
829
, doi:.
Rodriguez
,
Y.
,
D. A. R.
Kristovich
, and
M. R.
Hjelmfelt
,
2007
:
Lake-to-lake cloud bands: Frequencies and locations
.
Mon. Wea. Rev.
,
135
,
4202
4213
, doi:.
Ruhf
,
R. J.
, and
E. M.
Cutrim
,
2003
:
Time series analysis of 20 years of hourly precipitation in southwest Michigan
.
J. Great Lakes Res.
,
29
,
256
267
, doi:.
Steenburgh
,
W. J.
,
2014
: Secrets of the Greatest Snow on Earth. Utah State University Press, 244 pp.
Steenburgh
,
W. J.
, and
D. J.
Onton
,
2001
:
Multiscale analysis of the 7 December 1998 Great Salt Lake–effect snowstorm
.
Mon. Wea. Rev.
,
129
,
1296
1317
, doi:.
Steenburgh
,
W. J.
,
S. F.
Halvorson
, and
D. J.
Onton
,
2000
:
Climatology of lake-effect snowstorms of the Great Salt Lake
.
Mon. Wea. Rev.
,
128
,
709
727
, doi:.
Steiger
,
S. M.
, and Coauthors
,
2013
:
Circulations, bounded weak echo regions, and horizontal vortices observed within long-lake-axis-parallel–lake-effect storms by the Doppler on Wheels
.
Mon. Wea. Rev.
,
141
,
2821
2840
, doi:.
Tug Hill Commission
,
2014
: Environment and economy. [Available online at http://www.tughill.org/environment-economy/].
Walters
,
C. K.
,
J. A.
Winkler
,
S.
Husseini
,
R.
Keeling
,
J.
Nikolic
, and
S.
Zhong
,
2014
:
Low-level jets in the North American Regional Reanalysis (NARR): A comparison with rawinsonde observations
.
J. Appl. Meteor. Climatol.
,
53
,
2093
2113
, doi:.
Wilks
,
D. S.
,
2006
: Statistical Methods in the Atmospheric Sciences. 2nd ed. Elsevier, 627 pp.
Wilson
,
J. W.
,
1977
:
Effect of Lake Ontario on precipitation
.
Mon. Wea. Rev.
,
105
,
207
214
, doi:.
Yeager
,
K. N.
,
W. J.
Steenburgh
, and
T. I.
Alcott
,
2013
:
Contributions of lake-effect periods to the cool-season hydroclimate of the Great Salt Lake basin
.
J. Appl. Meteor. Climatol.
,
52
,
341
362
, doi:.

Footnotes

This article is included in the Ontario Winter Lake-effect Systems (OWLeS) Special Collection.

1

Statistical significance based on permutation resampling (Wilks 2006).