Syracuse University SURFACE
Dissertations - ALL SURFACE
December 2019
Snowstorms in Upstate New York: synoptics, spatial modeling and temporal variability
Justin Joseph Hartnett Syracuse University
Follow this and additional works at: https://surface.syr.edu/etd
Part of the Social and Behavioral Sciences Commons
Recommended Citation Hartnett, Justin Joseph, "Snowstorms in Upstate New York: synoptics, spatial modeling and temporal variability" (2019). Dissertations - ALL. 1120. https://surface.syr.edu/etd/1120
This Dissertation is brought to you for free and open access by the SURFACE at SURFACE. It has been accepted for inclusion in Dissertations - ALL by an authorized administrator of SURFACE. For more information, please contact [email protected]. ABSTRACT
This dissertation examines the characteristics of snowstorms that affect Central New York, a subsection of the eastern Great Lakes region, in a series of chapters organized as journal articles. The first article develops a classification scheme to categorize snowstorms in
Central New York from the 1985/86 season to the 2014/15 season. Twelve different snowstorm types were classified by their connection to the Great Lakes, the presence or absence of a synoptic low, or their area of cyclogenesis.
The second article uses the 2055 classified snowstorms to examine their relative contribution to seasonal snowfall totals. Although lake-effect snowstorms are the dominant snowfall contributor for most of Central New York, their contributions vary considerably across the region. These storms contribute approximately 50% of the seasonal snowfall totals in the Tug Hill, and only about 25% in southeastern Central New
York. Instead, Nor’easters are the dominant snowfall contributor in southeastern Central
New York. Model results can accurately estimate seasonal snowfall contributions using a location’s latitude, longitude, elevation and distance from the lake, or its latitude, longitude, and 5 km elevation exposure.
The third article examines the typical snowfall distributions and synoptic conditions associated with the different snowstorms. Localized snowfall patterns are most common when there is a surface high pressure over the United States and a low over northeastern
Canada. This setup combined with cold air (< 20˚C), often initiates the formation of lake- effect or lake-enhanced snow, potentially leading to the localized snowfall patterns.
Regional-wide snowfall was most common with cyclonic snowstorms (Nor’easters and
Rocky lows). These storms are often associated with an omega-blocking pattern, and heavier snowfall is common when air trajectories pass directly over the long-axis of Lake
Ontario.
The fourth and fifth articles examine how snowfall totals for the different snowstorm types have changed over time, and potential causes for these changes. This is the first study to directly assess seasonal snowfall trends for individual snowstorm types. Lake-effect snowfall significantly ( ≤ 0.05) increased in areas furthest from the lake from 1985/86 –
2014/15, while snowfall from clippers decreased across the entire region. Snowfall from lake snowstorms also increased in Region 3, but trends were inconsistent. Snowfall significantly increased in the late-1980s and late 1990s, but significantly decreased in the mid-1990s. The variability in trends may be linked to environmental conditions, as air temperatures were incorporated in 21/35 of the significant models. The Great Lakes also influenced seasonal snowfall totals mostly in Regions 1 and 3, while precipitation and average temperatures are the most influential factors in Regions 4 and 5. Teleconnections affect seasonal snowfall the most for Nor’easters, as above normal snowfall often occurs with the positive phases of the AO and PDO and the negative phase of the NAO. Lake-effect snowfall is mostly influenced by the WP, while teleconnections in the Atlantic Ocean and
Arctic largely affect snowfall totals from storms originating in western Canada.
Together, these articles highlight the importance of examining individual snowstorm types in the Great Lakes region and showcase potential forcings behind seasonal variations. This
study also highlights the importance of understanding the seasonal snowfall contribution of different snowstorm types and how it is changing, so that accurate predictions can be made for future climate scenarios.
SNOWSTORMS IN UPSTATE NEW YORK: SYNOPTICS, SPATIAL MODELING AND TEMPORAL VARIABILITY
by
Justin Joseph Hartnett B.S., Coastal Carolina University, 2011 M.S., University of South Florida, 2013
DISSERTATION
Submitted in partial fulfillment of the requirements for the degree of Doctor of Philosophy in Geography
Syracuse University December 2019
Copyright © Justin J. Hartnett 2019 All Rights Reserved
ACKNOWLEDGEMENTS
This work has been a long, arduous process, one that could not have been completed without the help from a number of people. First, I would like to thank all the people that have helped mold this work and my education. I would especially like to thank Jennifer Collins for always believing in me and helping me become the successful academic I am today. Your guidance has been invaluable, and I cannot thank you enough for it. To my advisor, Susan Millar, thank you for all the guidance on this work and the countless hours spent reading it and providing feedback. I am proud of this work, and you helped shaped it tremendously.
To Wendy Lascell and those at SUNY Oneonta that gave me a chance to thrive as an educator and helped mold my dissertation, I thank you as well. Oneonta has taught me so much on how to be an academic and I cherish all the connections I have made so far. I would also like to thank Jake Bendix who has acted as my second advisor throughout this process. I appreciate all the work that we have done together, but more importantly the mentoring that you provided me. This dissertation has also been enriched immensely by my committee members Adam Burnett, Peng Gao, and Jane Read. I especially want to thank Adam and Art Samel for all of their insights into lake-effect and for taking me under their tutorage.
I would like to thank all the people that have supported me through this process. Without each and every one of you, I would not have accomplished this tremendous feat. I could not have asked for better friends to surround myself with, encouraging me and helping me through the process. I especially want to thank Emily Bukowski, Pat Oberle, and Maddy Hamlin for all the conversations, venting sessions, and fun times that we’ve had. I would also like to thank everyone that helped me enjoy my time at Syracuse including Forrest Young, Joe (Philly) Hunter, Alyssa Almeida, Erik and Tori, Maeve Lanning, the Santoferrara family, and many more. All of you helped provide some sanity throughout this process.
Lastly, I want to give a special thank you to my family. Thank you to everyone that has supported me through this process, while also helping me enjoy it. Thanks to all my cousins, aunts and uncles, and grandparents. I especially want to thank my parents, Ted and Chrissy, for always supporting me and giving me the tools necessary to succeed in life. Thank you to my brother, Derrick, for always being my inspiration. To Jerry and Sulley, I cannot thank you two enough for putting up with my stress, late-night sessions, and sitting in front of the computer for countless hours. I appreciate everything that you both did for me to keep me sane and I love you more than can be described. I can gladly say that it is now over, and it would not have been possible without any of you.
vi
TABLE OF CONTENTS
List of Tables ...... x
List of Figures ...... xii
Chapter 1: Introduction ...... 1 1.1 Purpose of the Research ...... 8 1.1.1 Resolving the Percent Contribution of Lake-Effect Snow to Seasonal Snowfall Totals ...... 8 1.1.2 The Influence of Atmospheric Variability on Snowfall Contributions ...... 11 1.1.3 Storm Trajectories ...... 14 1.1.4 Historical Snowfall Trends ...... 16 1.1.5 The Influence of External on Forcings on Seasonal Snowfall Contributions ...... 19 1.1.6 The Influence of Internal Forcings on Seasonal Snowfall Contributions ...... 22 1.2 Research Objectives ...... 27
Chapter 2: Data and Methods ...... 29 2.1 Study Area ...... 29 2.2 Snowstorm Identification and Magnitude ...... 31 2.3 Snowstorm Classification ...... 36 2.3.1 Data ...... 36 2.3.2 Identification of Non-Cyclonic Snowstorms ...... 37 2.3.3 Identification of Cyclonic Snowstorms ...... 40 2.3.3.1 Clippers ...... 44 2.3.3.2 Great Lakes Lows ...... 45 2.3.3.3 Hudson Lows ...... 46 2.3.3.4 Nor’easters ...... 46 2.3.3.5 Colorado Lows ...... 49 2.3.3.6 Oklahoma Hooks ...... 50 2.3.3.7 Texas Hooks ...... 50 2.3.3.8 Tropical Cyclones ...... 51 2.4 Snowfall Contributions from Different Snowstorm Types ...... 51 2.5 Synoptic Classification of Different Snowstorm Types and Magnitudes ...... 57 2.6 Trends in Snowfall Contributions from Different Snowstorm Types ...... 59 2.6.1 Environmental Variables...... 60 2.6.2 Teleconnection Patterns ...... 62 2.7 Conclusion ...... 65
vii
Chapter 3: The Influence of Snowstorm Type on the Spatial Distribution of Snowfall and its Relative Contribution to Seasonal Snowfall Totals – A Scale Issue in the Great Lakes Region ...... 66 3.1 Introduction ...... 66 3.2 Methods ...... 70 3.2.1 Data ...... 70 3.2.2 Analysis ...... 72 3.3 Results and Analyses ...... 76 3.3.1 Magnitude and Frequency of Storm Types at the Regional Scale ...... 76 3.3.2 Magnitude and Frequency of Storm Types at the Subregional Scale ...... 86 3.3.3 Magnitude and Frequency of Storm Types at the Local Scale ...... 94 3.3.4 Modeling the Effects of the Physical Characteristics of a Location on Snowfall Contributions ...... 99 3.4 Discussion ...... 102 3.5 Conclusion ...... 109
Chapter 4: The Synoptic Conditions Associated with Different Snowstorms within Central New York ...... 112 4.1 Introduction ...... 112 4.2 Methods ...... 117 4.3 Results and Analyses ...... 121 4.3.1 Canadian Lows ...... 124 4.3.2 Lake-Effect Snowstorms ...... 131 4.3.3 Nor’easters ...... 137 4.3.4 Rocky Lows ...... 141 4.3.5 Non-Cyclonic Snowstorms ...... 149 4.4 Discussion ...... 156 4.5 Conclusion ...... 160
Chapter 5: Seasonal Trends in Snowfall Contributions from Different Snowstorm Types and the Environmental Factors Influencing Those Trends in Central New York ...... 162 5.1 Introduction ...... 162 5.2 Methods ...... 165 5.3 Results and Analyses ...... 167 5.3.1 Modeling the Secular Trends in Seasonal Snowfall Contributions ...... 167 5.3.2 Trend Reversal ...... 169 5.3.3 Modeling the Environmental Effects on Snowfall Contributions ...... 186 5.4 Discussion ...... 194 5.5 Conclusion ...... 199
viii
Chapter 6: The Influence of Atmospheric Low-Frequency Variability on the Seasonal Snowfall Contributions from Different Snowstorm Types Affecting Central New York ...... 202 6.1 Introduction ...... 202 6.2 Methods ...... 205 6.3 Results and Analysis ...... 207 6.3.1 Linear Fixed-Effects Models ...... 207 6.3.2 Variable Correlations ...... 218 6.4 Discussion ...... 224 6.5 Conclusion ...... 226
Chapter 7: Conclusions ...... 228
Chapter 8: Appendices ...... 241 8.1 Appendix 8.1 ...... 241 8.2 Appendix 8.2 ...... 245 8.3 Appendix 8.3 ...... 246 8.4 Appendix 8.4 ...... 248 8.5 Appendix 8.5 ...... 250 8.6 Appendix 8.6 ...... 254
Chapter 9: References ...... 255
ix
LIST OF TABLES
Chapter 2: Table 2.1. The source, agency, and date for data used to identify and classify snowstorms in Central New York ...... 32
Chapter 3: Table 3.1. Storm type classification applied in this chapter and their average seasonal (July – June) frequency and snowfall ...... 71 Table 3.2. Central New York COOP stations by region ...... 75 Table 3.3. The average frequency and snowfall for different magnitude snowstorms to influence Central New York from 1985/86 – 2014/15 ...... 77 Table 3.4. The average frequency of occurrence, and snowfall produced by light, moderate, and heavy lake snowstorms and non-lake snowstorms from 1985/86 – 2014/15 ...... 79 Table 3.5. The average seasonal frequency and seasonal snowfall of different snowstorm types in Central New York ...... 80 Table 3.6. Average seasonal frequency of occurrence, and average seasonal snowfall produced by the five snowstorm types identified to influence Central New York from 1985/86 – 2014/15 ...... 85 Table 3.7. Average percent contributions and average seasonal snowfall (cm) from each storm type for each subregion ...... 89 Table 3.8. Significance of Bartlett Tests of Variances and Analysis of Variances (ANOVAs)/Kruskal-Wallis Tests between the five subregions for each snowstorm type ...... 90 Table 3.9. ρ-values of Student t-tests and Mann-Whitney tests comparing the average percent contribution or average snowfall contribution of different snowstorms across regions ...... 90 Table 3.10. The influence of a location’s environmental conditions on the snowfall contribution (cm) of different snowstorms in Central New York ...... 100
Chapter 4: Table 4.1. Zonal statistics for the average snowfall (cm) per heavy snowstorm by snowfall type ...... 121 Table 4.2. The percentage of Central New York covered by various snowfall thresholds ...... 122 Table 4.3. Average trajectory length of snowstorms 72 hours prior to maturation ...... 122
Chapter 5: Table 5.1. Linear regression results for seasonal snowfall totals (cm yr-1) for different snowstorm types within Central New York from 1985/86 – 2014/15 for the five snowfall subregions...... 168
x
Table 5.2. The modeled results for the influence of a single environmental variable on seasonal snowfall totals from different snowstorm types for the five subregions of Central New York ...... 188 Table 5.3. Correlations between the seasonal snowfall totals from different snowstorm types (Storm) and the environmental parameters (Env. Variable) for models significantly explained by a single variable ...... 189 Table 5.4. AIC table for the influence of multiple environmental variables on seasonal snowfall totals from different snowstorm types within the five subregions of Central New York ...... 191 Table 5.5. Correlations between the seasonal snowfall totals from different snowstorm types (Storm) and the environmental parameters (Env. Variable) for models significantly explained by at least two variables ...... 193
Chapter 6: Table 6.1. AIC table for the teleconnection predictors of seasonal snowfall totals for the different snowstorm types in Region 1 ...... 208 Table 6.2. AIC table for the teleconnection predictors of seasonal snowfall totals for the different snowstorm types in Region 2 ...... 210 Table 6.3. AIC table for the teleconnection predictors of seasonal snowfall totals for the different snowstorm types in Region 3 ...... 212 Table 6.4. AIC table for the teleconnection predictors of seasonal snowfall totals for the different snowstorm types in Region 4 ...... 215 Table 6.5. AIC table for the teleconnection predictors of seasonal snowfall totals for the different snowstorm types in Region 5 ...... 217 Table 6.6. The correlation of all teleconnection variable against seasonal snowfall totals for the top models identified in Section 6.3.1 ...... 219
xi
LIST OF FIGURES
Chapter 1: Figure 1.1. Map of the Great Lakes Basin ...... 5 Figure 1.2. The development of lake-effect and lake-enhanced snow ...... 8 Figure 1.3. Mean surface global temperature anomalies for 2001-2005 compared to 1951-1980 ...... 19
Chapter 2: Figure 2.1. Central New York Study Area ...... 30 Figure 2.2. Average seasonal snowfall totals in Central New York from 1931/32 – 2011/12 ...... 30 Figure 2.3. Location of the 60 COOP stations used for analysis in Central New York ...... 34 Figure 2.4. Location of stations reporting hourly precipitation in Central New York ...... 35 Figure 2.5. Snowstorm Classification Scheme Diagram ...... 38 Figure 2.6. Areas of cyclogenesis in North America from Whittaker and Horn (1981) ...... 41 Figure 2.7. Location of cyclogenesis ...... 42 Figure 2.8. Cyclogenesis counts per 5˚x5˚ grid cell ...... 43 Figure 2.9. Getis G Ord Hot Spot analysis of the locations of cyclogenesis ...... 43 Figure 2.10. Zones of cyclogenesis for cyclonic storms affecting Central New York ...... 43 Figure 2.11. Typical track of extratropical cyclones affecting Central New York ...... 44 Figure 2.12. Elevation exposure (m) for different grid cells within Central New York ...... 54 Figure 2.13. Correlation plots between teleconnection patterns used in the analysis for model development ...... 64
Chapter 3: Figure 3.1. Five Central New York snowfall subregions ...... 67 Figure 3.2. Seasonal snowstorm frequency and seasonal snowfall totals for light, moderate and heavy snowstorms affecting Central New York from 1985/86 – 2014/15 ...... 77 Figure 3.3. Seasonal frequency and seasonal snowfall totals (cm) for lake- snowstorms (LS) and non-lake snowstorms (NLS) in Central New York from 1985/86 – 2014/15 ...... 78 Figure 3.4. Seasonal frequency (storms season-1) of the five snowstorm types identified to influence Central New York from 1985/86 – 2014/15 ...... 81 Figure 3.5. Seasonal snowfall totals (cm) from the five snowstorm types identified to influence Central New York from 1985/86 – 2014/15 ...... 81 Figure 3.6. The average percent frequency for each snowstorm type based on the storm magnitude ...... 83
xii
Figure 3.7. The average percent snowfall contribution for each snowstorm type based on the storm magnitude ...... 84 Figure 3.8. Average seasonal snowfall totals (cm) per snowstorm type for the five snowfall regions of Central New York from 1985/86 – 2014/15 ...... 87 Figure 3.9. Average seasonal snowfall contributions (%) per snowstorm type for the five snowfall regions of Central New York from 1985/86 – 2014/15...... 88 Figure 3.10. The percent contribution of lake snowstorms and non-lake snowstorms to seasonal snowfall totals in Central New York from 1985/86 – 2014/15 ...... 95 Figure 3.11. The percent contribution of seasonal snowfall totals associated with different snowstorm types to affect Central New York from 1985/86 – 2014/15 ...... 96 Figure 3.12. Daily weather map from the NOAA/ESRL’s 20th Century Reanalysis V2 representing a Nor’easter on 14 March 1993 at 06 Z ...... 108
Chapter 4: Figure 4.1. Average snowfall (cm) and air trajectories per heavy, moderate, and light Hudson lows, clippers and Great Lakes low in Central New York ...... 125 Figure 4.2. Average atmospheric conditions during heavy-snowfall Canadian lows: Hudson lows, clippers, and Great Lakes lows ...... 128 Figure 4.3. Average snowfall (cm) and air trajectories per heavy, moderate, and light LES-H and LES-UL snowstorms in Central New York ...... 132 Figure 4.4. Average atmospheric conditions during heavy-snowfall lake-effect storms: high-induced LES (LES-H) and upper atmospheric disturbance induced LES (LES-UL) ...... 134 Figure 4.5. Average snowfall (cm) and air trajectories per heavy, moderate, and light east coast storms and Gulf Coast storms (bottom) in Central New York ...... 138 Figure 4.6. Average atmospheric conditions during heavy-snowfall Nor’easters: east coast storms and Gulf Coast storms ...... 139 Figure 4.7. Average snowfall (cm) and air trajectories per heavy, moderate, and light Colorado lows, Texas hooks and Oklahoma hooks in Central New York ...... 142 Figure 4.8. Average atmospheric conditions during heavy-snowfall Rocky lows: Colorado lows, Texas hooks, and Oklahoma hooks ...... 145 Figure 4.9. Average snowfall (cm) and air trajectories per heavy, moderate, and light upper atmospheric disturbances, cold fronts and stationary fronts in Central New York ...... 150 Figure 4.10. Average atmospheric conditions during heavy-snowfall non- cyclonic storms: upper atmospheric disturbances, cold fronts, and stationary fronts ...... 153
xiii
Chapter 5: Figure 5.1. Seven-year seasonal snowfall trends from 1985/86 – 2008/09 for lake snowstorms and non-lake snowstorms ...... 171 Figure 5.2. Region 1 seven-year seasonal snowfall trends from 1985/86 – 2008/09...... 173 Figure 5.3. Region 2 seven-year seasonal snowfall trends from 1985/86 – 2008/09...... 175 Figure 5.4. Region 3 seven-year seasonal snowfall trends from 1985/86 – 2008/09...... 177 Figure 5.5. Region 4 seven-year seasonal snowfall trends from 1985/86 – 2008/09 ...... 180 Figure 5.6. Region 5 seven-year seasonal snowfall trends from 1985/86 – 2008/09 ...... 182
xiv
1.0 INTRODUCTION
As winter approaches, high latitudes and altitudes begin bracing for the drastic change from lush, green vegetation and warm temperatures, to a barren landscape often covered with centimeters of snow. Snowfall during these months plays a critical role in the social, economic, ecological, hydrological, and climatological processes in cold-winter areas.
The economies of many high latitude and altitude locations rely on snowfall to generate revenue. For example, during the 2015/16 winter season, recreational winter sports in the
United States accounted for over $11 billion in direct and indirect revenue and over
191,000 jobs (Hagenstad et al. 2018). Most of this revenue was generated through alpine and cross-country skiing, snowmobiling, and snowshoeing (Falk 2010; Hopkins and
Maclean 2014; Scott et al. 2008; Lazo et al. 2011; Pütz et al. 2011; Burakowski and
Magnusson 2012; Steiger and Stötter 2013; Rutty et al. 2015; Wobus et al. 2017). Extreme snowfall events and seasons are also potentially hazardous to life and property. A single- day snowstorm in the United States can cost upward of a million dollars (Zhu and Wang
2016; Kocin and Uccellini 2004a). Snowstorms halt ground and air traffic, disrupt businesses and corporations, and require plow trucks for snow removal and brine dispersion (Rooney 1967). Snowstorms can also become life threatening as the risk of automobile accidents (Andreescu and Frost 1998) and heart attacks (Persinger et al. 1993) increases during heavy-snowfall events.
Snowfall can also influence the ecology of an area, as the abundance of parasites, such as ticks, in the spring and summer is highly correlated with winter air temperatures and
1 snowpack conditions in northern latitudes (Dobson and Carper 1992; Harvell et al. 2002;
Rohr et al. 2012; Bradley et al. 2010). In return, this influences the survival rates of native
Mammalia, including moose (Rempel 2011) and white-tailed deer (Cortinas and Kitron
2006). Snowfall patterns can influence flora, as the insulative properties of a substantial snowpack (> 10.2 cm) help protect vegetation from freezing and root damage (Sharratt et al. 1992; Grippa et al. 2005; Zhang 2005; Brown and DeGaetano 2011; Kreyling and Henry
2011; Campbell et al. 2014). Changes in snowfall can favor invasive species, which threatens the survival of native species (Ziska et al. 2011; Stachowicz et al. 2002; Walther et al. 2002). Freezing air temperatures and snowfall also dictate the growing season, as early and late-season snowfall events can reduce its length (Schmidlin and Dethier 1985;
Yu et al. 2013; Vitasse et al. 2009; Christiansen et al. 2011).
Snow also has a considerable influence on the hydrologic cycle, as its accumulation and subsequent melt replenishes the water supply for more than one-sixth of the Earth’s population (Barnett et al. 2005). Snowmelt also saturates soils throughout the spring and early summer (Mastin et al. 2011). This moisture fuels vegetation growth, which increases biomass productivity (Ekstrand and Wallenberg 2010), and in return reduces the likelihood of wildfires (Groisman et al. 2004; Westerling et al. 2006). Economically, snowmelt helps maintain shipping industries (Millerd 2011), sewage plants (Kaczor and
Bugajski 2012), hydroelectric power (Fortier et al. 2011), and agriculture (Andersen and
Shepherd 2013; Ekstrand and Wallenberg 2010). However, snowmelt is not always beneficial. Extreme flooding and erosion are common following snowfall events since
2 soils are often frozen (Changnon and Changnon 2006; Pelletier and Rasmussen 2009; Olson and Morton 2012).
Climatologically, the high albedo of snow helps regulate global air temperatures (Qu and
Hall 2007; Robock and Robock 1980; Warren and Wiscombe 1980; Mote 2008; Namias
1963; Ellis and Leathers 1998). Snow and ice cover also influence precipitation patterns as they chill and stabilize overlying air masses. This enhances the strength of anticyclones and weakens cyclones (Cayan 1996; Turner and Gyakum 2011).
Precipitation patterns are affected by snow due to its influence on the moisture content of soils and the water levels of streams, rivers, and lakes (Groisman et al. 2004; Barnett et al.
2005; Westerling et al. 2006). If snowfall decreases, those moisture sources are diminished, if not lost, decreasing the amount of moisture available to the atmosphere.
This can lead to severe droughts, crippling both the natural and human environment
(Pederson et al. 2006; Bumbaco and Mote 2010; Mishra et al. 2010). Lastly, snowfall patterns in high latitudes can affect the characteristics of the polar jet stream. For example, severely diminished snowfall can lessen the meridional pressure gradient across the jet stream (Francis and Vavrus 2015; Kretschmer et al. 2016), which can decrease the frequency and strength of midlatitude cyclones (van den Brink et al. 2004), and increase the frequency of cold air troughing in the middle latitudes (Francis and Vavrus 2015).
Previous research suggests that antecedent snow cover can influence the location and intensity of the jet stream (Ross and Walsh 1986; Serreze et al. 1998; Rydzik and Desai
3
2014; Walsh et al. 1982), as storm centers tend to track just south of the snow extent line in the United States (Rydzik and Desai 2014).
Since snowfall plays such an important role in the characteristics of an area, understanding its spatiotemporal trends is crucial. As the climate changes, improvements in our understanding of how snowfall totals will respond are necessary. Research is also needed to better understand the influence of different snowstorm types on an area. Finally, improvements to seasonal snowfall projections are necessary to better prepare societies for an upcoming winter season.
Winter storms in the northern United States are frequently accompanied by snowfall.
However, due to the extensive size of the United States, these storms can exhibit large variability in their moisture content and snowfall totals. In the mountainous west, snowfall totals are relatively high due to elevated terrain and the occurrence of frequent extratropical cyclones originating in the northern Pacific Ocean (Thomas and Martin 2007).
To the lee of the Rocky Mountains, in the Mississippi River Valley, snowfall totals are severely reduced. Snowfall here, typically occurs from extratropical cyclones forming to the lee of the Rocky Mountains over flat terrain and in air masses with a low moisture content. In the eastern United States and the Great Lakes Basin, snowfall totals are unusually high compared to similar latitudes (Kocin and Uccellini 2004a). The considerable snowfall totals in the east coast are largely due to its proximity to the Atlantic
Ocean and the presence of the Appalachian Mountains. Comparatively, the Great Lakes
4
Basin averages even more snowfall, with some of the highest totals east of the Rocky
Mountains (Peace and Sykes 1966a).
Hydrologically, the Laurentian Great Lakes Basin is defined by watersheds that drain into one of the five Great Lakes: Lake Erie, Lake Huron, Lake Michigan, Lake Ontario, or Lake
Superior (Figure 1.1). These lakes provide vital freshwater resources for the United States and Canada, accounting for approximately 95% of all surface freshwater in the United
States alone (Lofgren 2004; Wang et al. 2012). Due to differences in the lake dimensions, atmospheric conditions over/near the lakes, water temperatures, and ice cover extent vary from lake-to-lake. Lake Erie tends to develop the earliest ice cover and the largest maximum ice cover due to its shallow depth and small volume (Assel et al. 2003; Bai et al.
2012; Wang et al. 2012). Despite Lake Ontario’s small surface area, it averages the smallest annual ice cover extent at 24.7%. This is largely because Lake Ontario is on average the deepest of the Great Lakes (Bai et al. 2012).
Figure 1.1. Map of the Great Lakes Basin
5
The term Great Lakes Basin, also referred to as the Great Lakes region, is also used in the atmospheric sciences, but its boundary is not clearly defined (e.g. Norton and Bolsenga
1993; Notaro et al. 2015; Grover and Sousounis 2002; Gula and Richard Peltier 2012).
Generally, it is recognized as the area surrounding the lakes in which the atmosphere is directly influenced by at least one of the lakes. The influences of midlatitude lakes, such as the Laurentian Great Lakes, on the atmosphere are well documented (Eichenlaub 1970;
Lofgren 1997; Notaro et al. 2013b; Changnon and Jones 1972; Bates et al. 1993; Bonan
1995; Angel and Isard 1998; Small et al. 1999; Long et al. 2007). As air masses advect over these lakes, they are frequently altered due to differences in moisture, heat content, and friction between the lake surfaces and the upwind areas (Notaro et al. 2013b; Changnon and Jones 1972). Due to the large size of the Laurentian Great Lakes they are a constant source/sink of moisture and energy to the overlying atmosphere, as long as there is an absence of an ice cover (Notaro et al. 2013b; Bonan 1995; Scott and Huff 1996). These influences are most prominent from September – March, as lake surface temperatures are often warmer than the overlying air masses (Angel and Isard 1998; Eichenlaub and Hodler
1979). This can lead to the destabilization of the air column, which regularly results in the formation of lake-effect clouds and precipitation.
Lake-effect snow describes the formation of snowfall that occurs due to the advection of a polar or Arctic air mass over a relatively warm and moist surface, typically a lake (Figure
1.2; Peace and Sykes 1966; Kunkel et al. 2000). The advection of the cold air initiates a heat and moisture transfer from the lake to the air (Norton and Bolsenga 1993). This transfer destabilizes the air, causing it to rise and form convective cells above the lake, which
6 results in the formation of relatively low (3000 m cloud tops) stratocumulus clouds (Pease et al. 1988). Light to moderate surface winds (10-20 kts) are then necessary to push clouds toward the shoreline (Laird and Kristovich 2004), where frictional forces are increased causing convergence zones and dynamic uplift. This enhances the convection within the storm cells, increasing condensation and cloud production (Dewey 1979a; Niziol et al.
1995; Laird and Kristovich 2004); yet, often not enough for the formation of precipitation.
As the clouds advect inland, they are met by increased elevations, inducing orographic uplift (Dewey 1979a; Laird and Kristovich 2004). This initiates precipitation, resulting in some of the snowiest locations in the world (Peace and Sykes 1966). Hill (1971) illustrates the significance of topography in inducing lake-effect snow by documenting a 25-50 cm increase in seasonal snowfall for every 100-meter rise in elevation downwind of the Great
Lakes. Thus, due to the presence of lake-effect snow, seasonal snowfall totals are greater in the Great Lakes region than most of North America (Minder et al. 2015; Eichenlaub and
Hodler 1979; Reinking et al. 1993; Hartnett et al. 2014).
Compared to other snowstorm types, lake-effect snowstorms are localized, 5-20 km wide and 50-300 km long (Niziol 1987). Since these storms are usually generated in stable air masses, they are dependent on the complexities of the advection of air over a waterbody including factors such as the surface and upper-level wind direction and speed, the amount of vertical wind shear in the atmosphere, the fetch, the surface lake/air temperature difference, the shape of the shoreline, the elevation and exposure, and convergence zones
(Peace and Sykes 1966; Niziol 1987). With low wind shear and a substantial fetch, a single lake-effect snowstorm can produce meters of snow (Reinking et al. 1993; Niziol et al. 1995;
7
Lackmann 2001). However, due to the relatively small size of lake-effect snowbands, snowfall totals can vary tremendously. For example, a single storm can result in one location receiving over 100 cm of snow, while a second location only kilometers away may barely receive a trace (Niziol 1987; Ellis and Leathers 1996; Ballentine et al. 1998).
Figure 1.2. The development of lake-effect and lake-enhanced snow. Produced by Joe Stoll, Syracuse University.
1.1 Purpose of the Research
1.1.1 Resolving the Percent Contribution of Lake-Effect Snow to Seasonal Snowfall
Totals
Lake-effect snowstorms are believed to be the dominant snowstorm type in the Great
Lakes region, despite differences in the lakes’ surface area, depth, and average ice extent
(Norton and Bolsenga 1993; Wang et al. 2012). However, there is considerable disagreement as to exactly what the contribution of lake-effect snow is to seasonal snowfall totals. Eichenlaub (1970) estimates that lake-effect snow accounted for at least 30% of the seasonal snowfall in Michigan from 1957/58 – 1961/62. These results contrast with those
8 of Veals and Steenburgh (2015), in which lake-effect snow accounted for 61-76% of the mean cool-season snowfall in the Tug Hill from September 2001 – May 2014. Further estimates suggest that within the Great Lakes region, lake-effect snow accounts for approximately half of the seasonal snowfall (Miner and Fritsch 1997; Liu and Moore 2004).
Understanding the seasonal contribution of lake-effect snowstorms to seasonal snowfall totals is further complicated by the occurrence of lake-enhanced snowfall. Lake-enhanced events are those which exhibit the characteristics of a lake-effect snowstorm but are linked to convective activity upwind of the lake or a synoptic scale system (Tardy 2000).
Although these events would likely not produce snowfall on their own, they enhance snowfall totals during other storms.
Recent research (e.g. Norton and Bolsenga 1993; Burnett et al. 2003; Kunkel et al. 2009a;
Notaro et al. 2015; Ellis and Johnson 2004) shows that seasonal snowfall totals from lake- effect snowstorms and non-lake effect snowstorms are trending in opposite directions. As lake-effect snow has been increasing since the early 20th century, snowfall from non-lake effect snowstorms has slightly decreased (Burnett et al. 2003; Kunkel et al. 2009a; Norton and Bolsenga 1993; Leathers and Ellis 1996). A crucial need is then to resolve the percent contribution of various snowstorm types to seasonal snowfall totals. This will provide a better understanding of how seasonal snowfall totals may change for a given location. In
Chapter 3, Central New York, a subsection of the Great Lakes region that receives considerable lake-effect snow, is examined in detail. The purpose of examining snowfall in central New York State is to specifically address the controversy of how much snow can be
9 defined as originating from a lake-effect or lake-enhanced storm versus other types of snowstorms.
Environmental conditions play a significant role in snowfall variability; therefore, in
Chapter 3, their specific effects are examined. For example, greater snowfall totals typically occur in areas with higher elevations due to enhanced orographic uplift and cooler temperatures (Grünewald et al. 2014; Spreen 1947; Johnson and Hanson 1995; Liu et al.
2011). The effects of elevation are especially prominent downwind of the Great Lakes, as elevated areas near the lakes favor high seasonal snowfall totals (Muller 1966; Wilson
1977; Minder et al. 2015; Reinking et al. 1993a; Hartnett et al. 2014; Eichenlaub and Hodler
1979; Burt 2007). However, high elevations are not always linked to higher snowfall totals, as the exposure of a location strongly influences precipitation totals (Perry et al. 2007;
Brown and Peck 1962; Veals and Steenburgh 2015). For example, a highly elevated location surrounded by similarly elevated locations, tends toward lower annual precipitation totals compared to a moderately elevated area surrounded by lower elevations (Brown and Peck 1962). The location of a station relative to the Great Lakes also influences snowfall totals. Niziol et al. (1995) note that snowfall totals from lake-effect storms are the greatest when the fetch over the lake is the longest. Chapter 3 discusses the application of geostatistical modeling to estimate the seasonal snowfall contributions from different snowstorm types for a given location based on its elevation, latitude, longitude, exposure, and distance from Lake Ontario.
10
1.1.2 The Influence of Atmospheric Variability on Snowfall Contributions
The synoptic-scale atmospheric conditions play a crucial role in determining the type and intensity of snowstorms that influence the United States (Liu and Moore 2004; Mote et al.
1997; Lawrimore et al. 2014; Changnon et al. 2006; Hjelmfelt 1990; Changnon et al. 2008;
Jurewicz and Evans 2004; Mullens et al. 2016). Since midlatitude extratropical cyclones derive their energy from the polar jet stream, a storm’s trajectory and the mean flow variabilities in the jet stream are closely related (e.g. Chang 2006; Lau 1988; Cai and van den Dool 1992). Thus, the latitudinal position and strength of the Northern Hemisphere jet stream modulates the strength and frequency of different snowstorm types in the United
States (Belmecheri et al. 2017). However, warming global temperatures are expected to increase the frequency and magnitude of meridional shifts in the jet stream (Barnes and
Simpson 2017; Delcambre et al. 2013). Shifts in the jet stream alter the steering winds of a storm, and can potentially change the frequency of snowstorms in an area, and ultimately the seasonal snowfall contributions from those storms. Uccellini and Kocin (1987) found that larger meridional shifts in the jet stream increase the potential for higher snowfall producing storms in North America. Thus, ironically, warmer global temperatures may increase the strength of heavy-snowfall producing snowstorms, but at the same time, there may be a decrease in light snowfall producing storms due to the advection of warmer air further northward (Lawrimore et al. 2014; Changnon et al. 2006).
In addition to an amplified meridional shift in the jet stream, the frequency of quasi- stationary planetary waves in the jet stream, also referred to as blocking patterns, have also increased (Kretschmer et al. 2016; Coumou et al. 2015; Screen and Simmonds 2014;
11
Belmecheri et al. 2017). Blocking patterns favor extreme weather in the midlatitudes, and are considered responsible for recent extreme weather events such as the flooding event of
Hurricane Harvey, the 2015 South Carolina floods, the 2010 Russian heat wave, the 2010
Pakistan floods, and the extreme cold of the 2014/15 Northeast United States’ winter
(Carrera et al. 2004; Whan et al. 2016; Brunner et al. 2017; Sillmann et al. 2011). As the climate changes, there has been an increase in the frequency and magnitude of blocking patterns, resulting in more frequent extreme storms, including heavy snowstorms
(Barriopedro et al. 2006; Croci-Maspoli et al. 2007; Scaife et al. 2010; Sillmann et al. 2011).
The occurrence of lake-effect snowstorms is not linked as closely to the jet stream; therefore, a great deal of research has examined the synoptic conditions associated with these storms (e.g. Suriano and Leathers 2017a; Barthold and Kristovich 2011; Kristovich et al. 2018; Sousounis and James 2003; Liu and Moore 2004; Leathers and Ellis 1996; Ellis and Leathers 1996). Leathers and Ellis (1996) and Ellis and Leathers (1996) identified five unique synoptic conditions associated with lake-effect snowstorms in Syracuse, NY.
Suriano and Leathers (2017b) identified two additional synoptic conditions associated with lake-effect snowfall leeward of Lakes Erie and Ontario. The seven lake-effect synoptic types correspond to surface wind directions leeward of the lakes and are associated with a prominent high-pressure system, generally over the central or western United States and a low-pressure system over New England or southeastern Canada. The presence of a high- pressure system during winter for the majority of North America signifies cold, but relatively stable air with little precipitation. However, to the lee of the Great Lakes, these high-pressure systems are often associated with lake-effect snowstorms (Onton and
12
Steenburgh 2001; Notaro et al. 2013b; Strong 1972). Thus, the presence and persistence of high-pressure systems brought on by an omega blocking pattern, can potentially increase seasonal snowfall totals in the Great Lakes region.
Recent research also indicates the influence of multi-lake interactions on the formation of lake-effect snowstorms (e.g. Mann et al. 2002; Rodriguez et al. 2007; Laird et al. 2017; Lang et al. 2018). In general, the position of high and low-pressure systems can generate lake- to-lake (L2L) snowbands, which are snowbands that develop over an upstream lake, extend over an intermittent landmass, and connect to or form snowbands over a downstream lake (Laird et al. 2017; Lang et al. 2018). Lang et al. (2018) note that snowfall totals after lake-effect snowstorms were greater from L2L days than non-L2L days. They attributed the increased snowfall to a more favorable environment for the development of lake-effect snow, including greater instability over the upwind lake, more near-surface moisture availability, faster wind speeds, and larger surface heat fluxes over the upstream lake.
In Chapter 4 I focus on the average synoptic conditions associated with different snowstorm types, and of different snowfall magnitudes. This provides a clearer understanding of the relationship between the synoptic conditions and the spatial patterns of snow. In this chapter, I examine the influence of the jet stream on the distribution of snowfall following a storm. Therefore, this research improves snowfall forecasting by better understanding the atmospheric conditions which promote certain snowstorm types in Central New York. In addition, this research helps better understand how snowstorm
13 frequency may change in the future based on the expected response of synoptic conditions to climate change.
1.1.3 Storm Trajectories
In addition to influencing wind and temperature patterns, the synoptic atmospheric conditions influence the trajectory of storms, which in return influence snowfall totals
(Peace and Sykes 1966; Changnon et al. 2008; Perry et al. 2007). The most intense snowfall tends to occur downstream and to the left of the central low pressure (Goree and Younkln
1966; Changnon et al. 2008). Storm trajectories also influence the temperature, moisture content, and stability of the air due to conditions imparted by the region it passes over
(Katurji and Zhong 2012; Zhu et al. 2005; Perry et al. 2007). For example, air that passes over Lake Ontario prior to reaching New York State in winter, tends to have a higher moisture and heat content than air advecting from Ontario or Quebec (Fuhrmann and
Konrad 2013). Comparatively, air that passes over elevated terrain tends to have a lower moisture content due to orographic precipitation extracting moisture prior to reaching the area (O’Handley and Bosart 1996; Schumacher et al. 1996; Perry et al. 2007; Barros and
Kuligowski 1998). Thus, the upwind terrain of a storm considerably influences snowfall totals and the distribution of that snowfall.
Source regions of air are especially important in the Great Lakes region, as lake-effect clouds rarely form over a single lake. Instead, air regularly travels across multiple lakes, increasing the potential for L2L snowbands (Mann et al. 2002; Rodriguez et al. 2007; Laird et al. 2017). Due to the Westerlies, L2L snowbands develop most frequently between north
14
(e.g. Lake Superior and Lake Huron) and south lakes (i.e. Lake Michigan and Lake Ontario) from December – February (Kristovich and Steve III 1995; Rodriguez et al. 2007; Laird et al. 2017). The aggregate effects of heat and moisture transfer from the lakes to the overlying air change the large-scale winds, temperature, moisture, and stability characteristics over the individual lakes (Agee and Gilbert 1989; Hjelmfelt 1990; Sousounis and Fritsch 1994; Ballentine et al. 1998; Weiss and Sousounis 1999; Sousounis and Mann
2000; Laird et al. 2017). Although there is little consensus as to the effects of multi-lake interactions, research does suggest that these interactions alter downstream lake-effect snowbands. Some researchers have found a strengthening of lake-effect snowbands due to the influence of multiple lakes (Yuen and Young 1986; Agee and Gilbert 1989; Niziol et al.
1995; Ballentine et al. 1998; Sousounis and Mann 2000; Rodriguez et al. 2007). Others have observed smaller snowfall totals associated with L2L snowbands due to a reduction in convective instability (Sousounis and Fritsch 1994; Sousounis and Mann 2000; Mann et al.
2002). The influence of L2L snowbands makes the understanding of lake-effect snowstorms in the eastern Great Lakes basin considerably difficult (Niziol et al. 1995).
Therefore, in Chapter 4, I also examine the influence of storm trajectories on snowfall. The purpose of this section is to determine if certain trajectories favor larger magnitude storms and to determine how the distribution of snowfall differs for different trajectories.
Although previous studies (e.g. Perry et al. 2007; Changnon et al. 2008) have analyzed the influence of various storm tracks on snowfall totals, they have not examined the effects of snowstorm tracks on all snowstorm types within the Great Lakes region.
15
1.1.4 Historical Snowfall Trends
Temporal trends and interannual and interdecadal variability of snowfall totals have been well-documented (Kunkel et al. 2016; Walsh et al. 1982; Ropelewski and Halpert 1986;
Diaz et al. 1989; Groisman and Easterling 1994). Although precipitation totals determine the potential amount of snowfall in an area, air temperatures determine whether that precipitation falls as snow (Knowles et al. 2006; Dai et al. 2001). Therefore, annual snowfall totals are particularly sensitive to climate change. Climate proxies from the last
10,000 years suggest that global average temperatures have increased the most during the
20th and 21st centuries, with temperature anomalies exceeding +1⁰C during the 21st century relative to average air temperatures from 1880-1920 (Wang et al. 2007; CRU 2007; NRC
2012). Some of these changes are attributed to natural climate variability such as teleconnection patterns, volcanic eruptions and solar flares; however, the most significant external climate forcing in the 21st century is an increase in atmospheric greenhouse gases
(Overland et al. 2007). The IPCC's 5th Assessment Report (2013) suggests that there is a high confidence (95%) that the increase in temperatures during the 20th and 21st century is due to an increase in anthropogenically sourced greenhouse gases.
Changes in air temperatures and snowfall patterns also vary spatially across the earth
(Hansen et al. 2006). With the establishment of the National Weather Service’s
Cooperative Observer Program (COOP), more than 10,000 volunteers take daily weather observations across the United States. With observations dating back to the late-1800s, snowfall patterns across the United States have been widely studied (e.g. Groisman and
Easterling 1994; Ellis and Johnson 2004; Hartnett et al. 2014; Kunkel et al. 2009b; Knowles
16 et al. 2006; Serreze et al. 1998; Clark et al. 2016; Norton and Bolsenga 1993; Burnett et al.
2003; O’Hara et al. 2009; Karl et al. 1993; Mote 2006; Knowles 2015; Mote et al. 2005;
Durre et al. 2013). The results of these studies suggest that snowfall has mostly decreased across the continental United States since the early 20th century, but it largely depends on the area of study and the dominant type of snowstorm the region experiences.
The western United States has experienced greater warming than the central and eastern
United States from 1951 - 2005 (Figure 1.3; Hansen et al. 2006). This warming has resulted in a reduction in the precipitation that falls as snow, which has decreased snowfall and snowpack totals (Dyer and Mote 2006; Knowles et al. 2006; Robinson and Henderson-
Sellers 1999; Groisman and Easterling 1994; Scott and Kaiser 2004; Mote et al. 2005). In the eastern United States, winter precipitation totals have generally increased (Groisman and Easterling 1994), but have not necessarily coincided with an increase in snowfall (e.g.
Burnett et al. 2003; Kunkel et al. 2009a). Lawrimore et al. (2014) found that severe snowstorms in the eastern and central United States are trending toward an earlier date of occurrence due to warmer temperatures in the spring, which has resulted in decreased snowfall totals during spring months. Burnett et al. (2003) concluded that between
1931/32 – 2001/02, snowfall totals for locations outside of the Great Lakes region remained relatively constant, or slightly decreased. However, within the Great Lakes region, the consensus is that since the early 20th-century, there has been a significant increase in snowfall (Norton and Bolsenga 1993; Burnett et al. 2003; Braham and Dungey
1995; Ellis and Johnson 2004; Kunkel et al. 2009a; Hartnett et al. 2014). For example,
Norton and Bolsenga (1993) documented a significant increase in snow in the Great Lakes
17 basin from the 1950s to 1980s, compared to no significant change in areas outside of the basin. Likewise, Burnett et al. (2003) concluded that lake-effect stations experienced an average snowfall increase of 1.5 cm yr-1 between 1931 and 2001. Using quality assessed data, Kunkel et al. (2009) suggests that there was a significant increase in snowfall at stations downwind of Lakes Superior and Michigan. They further indicated that the shorter-term record showed a significant increase in snowfall downwind of Lake Erie, whereas the longer record had a significant increase downwind of Lake Ontario (Kunkel et al. 2009a).
The regional differences in snowfall trends likely reflect changes to the dominant snowstorm types affecting an area. It is generally assumed that an increase in lake-effect snowfall is responsible for the increased snowfall in the Great Lakes region (Burnett et al.
2003; Kunkel et al. 2009a; Hartnett et al. 2014). Snowfall from extratropical cyclones however, has mostly decreased since the early 20th century (Thomas and Martin 2007;
Jeglum et al. 2010), especially contributions from Alberta Clippers and Colorado lows. This has had the greatest effect on snowfall and snowpack in the western United States, where
Alberta Clippers and Colorado lows are the dominant snowfall producers. Although these storms also affect the eastern United States, snowfall totals have not decreased as noticeably and in some cases increased due to the influence of additional storms (Hirsch et al. 2001; Harrington et al. 1987). Multiple snowstorm types often contribute to the seasonal snowfall totals of a region. Since snowfall from one or more of these storms may be changing over time, it is necessary to understand the contribution of different snowstorm types to seasonal snowfall totals. Without such information, the accuracy of
18 future snowfall predictions is diminished. Thus, the purpose of Chapter 5 is to examine an area that is influenced by a variety of snowstorm types, and to determine how snowfall contributions have changed over time for the different storms.
Figure 1.3. Mean surface global temperature anomalies for 2001-2005 compared to 1951- 1980. Warming is generally greatest over land and high latitudes in the Northern Hemisphere. Figure from Hansen et al. (2006).
1.1.5 The Influence of External Forcings on Seasonal Snowfall Contributions
As the IPCC (2013) suggests, the most significant external forcing on ice and snow conditions is anthropogenic climate change, due to increased greenhouse gases. Since the
1980s, regional warming has impacted the physical characteristics of the Great Lakes, including their water levels, precipitation and evaporation patterns, water temperatures, and winter ice extent and thickness (Bolsenga and Norton 1993; Dietz and Bidwell 2011;
Vavrus et al. 2013). Since the lakes are warming faster than the air (Lofgren 2004;
Trumpickas et al. 2009; Dietz and Bidwell 2011), lake-effect snowfall patterns are changing due to alterations to the annual ice cover on the lakes and the average lake-air temperature
19 difference. Generally, there has been a reduction in the overall ice extent and a successively earlier date of ice departure on the Great Lakes during the 20th and 21st centuries (Hanson et al. 1992; Assel et al. 2003; Wang et al. 2012). The largest decrease was recorded on Lake Ontario, which experienced an 88% decrease in ice extent from
1973-2010 (Wang et al. 2012). Ice thickness has also decreased on the Great Lakes (Wang et al. 2012), and this combined with reduced ice cover, can affect moisture transfers, lake dynamics, and local wind patterns, which in turn can alter the frequency and severity of lake-effect snow (Tsuboki et al. 1989; Segal and Kubesh 1996). For example, an earlier ice departure leads to an earlier stratification of the lake, resulting in a warm cap that heats rapidly during the spring and summer. The warm cap results in magnified lake temperatures relative to the air, which enhances the sensible flux to the atmosphere in the fall, increasing the potential for more evaporation over the lakes and precipitation downwind of the lakes (Hanson et al. 1992; Wang et al. 2012). Enhanced evaporation rates throughout winter, without a corresponding increase in precipitation, has led to a decrease in the water volume of the lakes. This has the potential to change the thermodynamics of the lakes and alter snowfall patterns (Hanson et al. 1992; Sellinger et al. 2008; Trumpickas et al. 2009).
Associations between lake characteristics and snowfall have been documented by several researchers. Burnett et al. (2003) examined twentieth century calcite oxygen isotopes
[δ18OCaCO3] records from sediments in several Finger Lakes south of Lake Ontario. Based on a comparison with δ18OCaCO3 from snowfall samples and their association with storm tracks, the authors conclude that the decreasing δ18OCaCO3 in core samples is a strong indication of
20 increasing lake-effect snow attributed to a warming lake surface. Since the Great Lakes surface temperatures are increasing more rapidly than air temperatures, the temperature difference between the lake surface and the 850 hPa air level is greater therefore enhancing the likelihood and magnitude of lake-effect snowstorms. The increased temperature difference combined with a smaller ice extent and shorter ice season, has the potential for more seasonal lake-effect snow and a later shift in peak snowfall (Burnett et al. 2003; Vavrus et al. 2013).
As air temperatures increase, they have the potential to transition snow to rain (Groisman and Easterling 1994; Knowles et al. 2006; Schmidlin and Dethier 1985; Notaro et al. 2014;
Robinson and Henderson-Sellers 1999; Scott and Kaiser 2004; Mote et al. 2005; Mote
2006). Areas most vulnerable to this transition are those where winter air temperatures are already near freezing such as the Mid-Atlantic, northwestern, and southern United
States (Solomon et al. 2007; Pierce and Cayan 2013; Kluver and Leathers 2015). However, as temperatures warm, regionally-focused Global Climate Models (GCMs) predict that seasonal snowfall totals will drastically decrease within the Great Lakes region by the middle-end of the 21st century (Suriano and Leathers 2016; Notaro et al. 2013b; Kunkel et al. 2002). Thus, in Chapter 5 I use geostatistical models to determine how the environmental conditions of the Great Lakes (e.g. surface temperatures) and the overlying air masses (e.g. surface air temperatures) influence snowfall contributions from the different snowstorm types. By understanding the influence of air temperatures and lake surface conditions on seasonal snowfall contributions per storm type, seasonal snowfall predictions can be improved by incorporating any future changes to these conditions.
21
1.1.6 The Influence of Internal Forcings on Seasonal Snowfall Contributions
In addition to anthropogenic forcings, snowfall patterns have been linked to natural variations in atmospheric and oceanic circulation patterns, most notably teleconnection patterns (e.g. Serreze et al. 1998; Ge and Gong 2009). According to the CPC (2018), teleconnection patterns are preferred modes of low-frequency variability in pressure and circulation that extend over a large area. These large-scale changes in the atmospheric wave and jet stream patterns have been shown to influence temperature and precipitation patterns, storm tracks, and the jet stream location/intensity over North America (Barnston and Livezey 1987; CPC 2012). In return, these influences have been shown to alter snowfall patterns (e.g. Serreze et al. 1998; Ge and Gong 2009; Ghatak et al. 2010; Baxter et al. 2014; Gan and Wu 2015; Wise et al. 2015). The teleconnection patterns shown to have the most significant influence on seasonal snowfall in North America include the El Niño
Southern Oscillation (ENSO), the North Atlantic Oscillation (NAO), the Pacific North
American (PNA) pattern, the Pacific Decadal Oscillation (PDO), the Arctic Oscillation (AO), the East Atlantic (EA) pattern, and the West Pacific (WP) pattern.
ENSO is a large-scale ocean-atmosphere climate phenomenon linked to changes in sea surface temperatures in the central and eastern equatorial Pacific (Barnston 2015). The links between ENSO and snowfall patterns in North America are well documented (Kunkel et al. 2009b; Hirsch et al. 2001; Bai et al. 2012; Patten et al. 2003; Smith and O’Brien 2001;
Mason and Goddard 2001; Ropelewski and Halpert 1986; Sittel 1994; Yarnal and Diaz
1986; Rohli and Vega 2011; Kahya and Dracup 1993; Piechota and Dracup 1996; Kunkel and Angel 1999; Groisman and Easterling 1994; Eichler and Higgins 2006; Wise et al. 2015;
22
Seager et al. 2010b; Dai and Wigley 2000; Gutzler et al. 2002; McCabe and Dettinger 2002;
Hidalgo and Dracup 2003; Goodrich and Walker 2011). The consensus is that during the El
Niño (La Niña) phase, seasonal snowfall totals are anomalously low (high) over the
Northeast and Great Lakes region of the United States.
The NAO is another major source of interannual and decadal-scale variability in the winter atmospheric circulation over North America. It is defined by surface sea-level pressure differences between the Azores High and the Subpolar Low, which affect average temperature patterns around the North Atlantic (Ghatak et al. 2010; Hurrell 1995; Notaro et al. 2006; Bai et al. 2012; Wallace and Gutzler 1981; Barnston and Livezey 1987; Wise et al. 2015; Walker and Bliss 1932; Wettstein and Mearns 2002; Hartley and Keables 1998;
Athanasiadis et al. 2017; Sobolowski and Frei 2007; Roller et al. 2016; Archambault et al.
2008; Fereday et al. 2012; Kalra and Ahmad 2012; Seager et al. 2010b; Osborn 2011;
Coleman and Budikova 2013). Strong positive phases of the NAO typically correspond to above-normal temperatures in the eastern United States, decreasing the likelihood of snowfall.
The PNA is linked to low-frequency variability in the extratropical Northern Hemisphere and is defined by geopotential heights near Hawaii. A positive phase means above-average heights and a negative phase is below average heights. Generally, the positive phase of the
PNA shifts the exit region of the jet stream over the western United States (Ghatak et al.
2010; Ge and Gong 2009; Wallace and Gutzler 1981; Barnston and Livezey 1987; Leathers et al. 1991; Roller et al. 2016; Notaro et al. 2006; Wise et al. 2015; Henderson and Leathers
23
2010; Ewen et al. 2008; Coleman and Rogers 2003; Mock 1996). This leads to above- average temperatures over western Canada and the United States and below average temperatures over the southern and southeastern United States. The PNA is believed to be a dominant mode of winter atmospheric variability in North America (Ghatak et al. 2010) and is strongly tied to surface regional temperature and precipitation anomalies (Leathers et al. 1991).
The PDO describes temperature anomalies in the northeast and tropical Pacific Ocean. The positive phase is characterized by anomalously warm sea surface temperatures along the
Pacific Coast, with abnormally cold water in the interior North Pacific (Newman et al.
2016). The positive phase of the PDO tends to lead to anomalously warm water along the west coast of North America, resulting in above average snowfall in the Northeast United
States (Ge and Gong 2009; Kunkel et al. 2009b; Gutzler et al. 2002; McCabe and Dettinger
2002; Hidalgo and Dracup 2003; Goodrich and Walker 2011).
The AO is a large-scale mode of climate variability characterized by counterclockwise circulating winds around 55˚N in the Arctic. The positive phase of the AO has been linked to anomalously strong winds circulating the Arctic, which confines the cold air to polar regions (Bai et al. 2012; Rohli and Vega 2011; Zhu and Wang 2016). The negative phase is associated with weaker winds and so colder air can penetrate further south, increasing midlatitude storminess and snowfall.
24
The EA is another low-frequency pattern over the North Atlantic. It is characterized by a north-south dipole of anomaly centers spanning the North Atlantic from east to west. The positive phase often results in below-average surface temperatures over the southern
United States from January-May, and north-central United States from July-October (Wise et al. 2015; Davis and Benkovic 1994; Seierstad et al. 2007; Woollings and Blackburn 2012;
Moore et al. 2013; Strong and Davis 2008; Barnston and Livezey 1987).
The WP is the primary mode of low-frequency variability over the North Pacific. It is characterized by a north-south dipole over the Kamchatka Peninsula and southeastern Asia and the western subtropical North Pacific. Strong negative and positive variations in the
WP correspond to pronounced zonal and meridional variations in the location and intensity of the entrance region of the Pacific jet stream. The positive phase has been associated with above average temperatures over the lower latitudes of the western North
Pacific in both winter and spring (Barnston and Livezey 1987; Wise et al. 2015; Lau 1988;
Sui and Lau 1992; Linkin and Nigam 2008; Baxter and Nigam 2015; Tanaka et al. 2016).
This phase is also linked to above average precipitation during all seasons across the high latitudes of the North Pacific and below average precipitation across the central North
Pacific during the winter and spring.
Although these teleconnection patterns are hemispheric in scale, they can influence the seasonal frequency of midlatitude cyclones passing through the Great Lakes region and affect the orientation of the wind field relative to the axes of the lakes. In the Northeast and
Great Lakes regions, snowfall totals are closely linked to the phases of ENSO and the NAO
25
(Serreze et al. 1998; Ge and Gong 2009; Ghatak et al. 2010; Allen and Zender 2011; Grise et al. 2013; Baxter et al. 2014; Yu et al. 2014; Gan and Wu 2015). The strength of the relation is dependent on the spatial and temporal resolution evaluated. For example, Grimaldi
(2008) subdivided the winter into two segments to reveal that during the El Niño phase of
ENSO, early winter months in Syracuse, New York are generally warmer with anomalously low snowfall totals, compared to mid-winter which is colder and snowier than normal.
Grimaldi (2008) contends that the warm early winter during El Niño, pre-conditions the lake surfaces so that they remain ice free and warmer, therefore can enhance the occurrence of large magnitude lake-effect snowfall events associated with mid-winter storms during an El Niño. At a broader spatial scale, Kocin and Uccellini (2004b) noted that the NAO is negatively (p < 0.05) correlated with increased seasonal snowfall in the eastern
United States, including the Great Lakes.
Since researchers have shown that hemispheric teleconnection patterns can influence snowfall totals within the Northeast and Great Lakes regions of the United States, the purpose of Chapter 6 is to determine the relative influence of several teleconnection patterns on seasonal snowfall contributions from different snowstorm types. The influence of teleconnections on seasonal snowfall totals is also examined spatially, to determine if the effects are homogenous throughout the region. By determining the relative influence of different teleconnections on seasonal snowfall totals, the results from this chapter can then be used to enhance the accuracy of seasonal snowfall projections.
26
1.2 Research Objectives
The overarching objective of this research is to understand more fully the nuanced spatial and temporal patterns of snowfall within Central New York. By understanding the average contributions of different snowstorm types, the general synoptic conditions of those different storms, and the external and internal forcings influencing those storms, future snowfall projections for this area can be improved immensely. The improvement of future seasonal snowfall projections is especially important in this area, which relies heavily on stable seasonal snowfall totals for agriculture, business, water resources, and recreation
(Falk 2010; Hopkins and Maclean 2014; Brown and DeGaetano 2011; Mastin et al. 2011).
Thus, by unraveling the nuanced spatial and temporal patterns of snowfall within Central
New York, several uncertainties and controversies can be addressed.
In Chapter 2, I present a full description of the study area’s geography and its appropriateness for examining the topic of lake-effect snow. In addition, I present the general data-processing methodology by providing details of what observational, radar, and reanalysis data were used to classify snowstorms and snowstorm types, in addition to any other datasets collected, and how they were processed and classified for use in the remaining chapters. In this chapter I also develop a methodology for classifying snowstorms within the Great Lakes region, that has applicability in other areas.
Chapter 3 provides a complete analysis of the primary storms that affect the study area in order to address the key uncertainty of the actual contribution of lake-effect snow to seasonal snowfall totals in Central New York. This chapter highlights the large spatial and
27 temporal variations in snowfall contributions from different snowstorm types. Chapter 4 on the other hand, examines specifics of the storm paths and their interaction with larger scale atmospheric dynamics which ultimately affect snowfall patterns in the study area.
Chapter 5 observes long-term snowfall patterns from different snowstorm types and uses environmental variables to model those contributions to provide some understanding of the key regional variables that interact with different storm types. Chapter 6 provides a final substantive contribution by examining how teleconnection patterns impact the relative frequencies of different storm types and their ultimate contributions to snowfall in the study area. To conclude, Chapter 7 provides a summary, the larger implications of the findings, as well as important areas of future research.
28
2.0 DATA AND METHODS
In this dissertation I use many data sets and derive variables from observational data. A key variable is storm type based on a detailed classification scheme, which is used in various forms in later chapters. The specific form and application of the data are discussed in detail in relevant chapters, however the initial processing and the construction of the storm classification methodology and description that are used throughout, are discussed here. The storm classification is specific to snowstorms in the Great Lakes region, but with modifications and consideration of local conditions, it is transferrable to any location.
2.1 Study Area
In this research I examine snowstorms in twelve upstate Central New York counties
(Figure 2.1). The elevation in Central New York varies from 85 m in the Erie-Ontario and
Hudson-Mohawk Lowlands to 915 m in the Adirondack Mountains and Tug Hill Plateau, further referred to as the Tug Hill (Figure 2.1). Elevated terrain (365 – 915 m) is also present throughout southern Central New York in the Allegheny Plateau and Southern
Hills. Partially situated within the Great Lakes Basin, this part of New York receives greater seasonal snowfall totals than most areas at similar latitudes (Norton and Bolsenga 1993;
Hartnett et al. 2014). The greater snowfall totals are on account of the elevated terrain and nearby moisture sources, including the Great Lakes and the Atlantic Ocean.
Due to Central New York’s proximity to Lake Ontario, lake-effect snowstorms are often considered the most frequent snowstorm type. These storms occur most frequently to the lee of the lake over the Tug Hill, where seasonal snowfall totals commonly exceed 635 cm
29 and are some of the greatest east of the Rocky Mountains (Figure 2.2). Seasonal snowfall totals are also generally higher in the Tug Hill because of Lake Ontario’s east-west orientation which favors a longer fetch (Peace and Sykes 1966; Niziol 1987) and because of its smaller ice cover and ice extent compared to the other Great Lakes (Assel et al. 2003;
Bai et al. 2012; Wang et al. 2012). The eastern position of the Tug Hill on the Great Lakes also allows for the formation of multi-lake snowbands from Lake Huron, and to a lesser extent from Lakes Michigan and Superior (Mann et al. 2002; Rodriguez et al. 2007; Laird et al. 2017). Although seasonal snowfall totals are generally lower throughout the rest of
Central New York, lake-effect snowstorms are still considered significant contributors to
Figure 2.1. Central New York Study Area. The twelve counties of interest are shaded in beige. Included are the geographic features of Central New York and the hydrological Great Lakes Basin. Figure 2.2. Average seasonal snowfall totals in Central New York from 1931/32 – 2011/12.
30 seasonal snowfall totals. However, they are not the sole contributor as Central New York, unlike the western Great Lakes region, is highly influenced by both the Great Lakes and the
Atlantic Ocean. Thus, Central New York offers a unique opportunity to examine the complexity of seasonal snowfall totals for an area frequently influenced by both lake-effect and synoptic snowstorms.
2.2 Snowstorm Identification and Magnitude
Intense (≥ 2.5 cm hr-1) and heavy (≥ 25 cm) snowstorms receive the bulk of attention in the media, as they are the most disruptive to society (Call 2005). In snow-hardy regions, little attention is given to smaller snowstorms (≤ 15.2 cm) because there are fewer accidents, schools and businesses remain open, transportation is minimally disrupted, and businesses and structures are less susceptible to damage (Call 2005). However, smaller snowstorms significantly contribute to seasonal snowfall totals, especially in lake-effect dominated regions. Thus, in this study I examined all snowstorms to influence Central New York from
1 July 1985 through 30 June 2015. This period was chosen because a climatological period is defined as 30 years of data as defined by the World Meteorological Organization
(Glickman 2000) and because recent research suggests that snowfall within the Great
Lakes region underwent a trend reversal in the late 1970s – early 1980s (Bard and
Kristovich 2012; Hartnett et al. 2014; Suriano and Leathers 2016).
Snowstorms were defined using the guidance of Perry et al. (2007), whereby they were considered a snowfall event if at least 0.3 cm of daily snow was recorded for at least two
COOP stations within Central New York. Snowstorms were identified using daily snowfall
31 records for COOP stations located in Central New York from the National Centers for
Environmental Information’s (NCEI) online server at http://www.ncdc.noaa.gov/cdo- web/search (Table 2.1). Since 1 July 1985, 295 COOP stations recorded daily snowfall in
Central New York for a least one day.
Table 2.1. The source, agency, and date for data used to identify and classify snowstorms in Central New York. Data Source Agency Date Daily Snowfall Cooperative Observer Program NCEI 1985 – 2015 Hourly Precipitation Local Climatological Data NCEI 1985 - 2015 Daily Weather Maps NCEP 1871 - 2015 Surface Reanalysis NOAA NWS Reanalysis Data Display by NCEP WPC NCEP 1948 - 2011 Surface Analysis Archive WPC 2005 - 2015 GOES Infrared Imagery International Satellite Cloud Climatology Project NCEI 1983 - 2015 Radar NEXRAD Data Archive, Inventory and Access NCEI 1994 - 2015 G.Lakes SFC Temp. CoastWatch Great Lakes Node GLERL 1995 - 2015 G.Lakes Ice Cover CoastWatch Great Lakes Node GLERL 2008 - 2015 Syracuse Int’l Climate Data Online NCEI 1938 - 2015 Atmospheric Data AO, ENSO, NAO, PDO & Climate Monitoring Online – Teleconnections NCEI 1985 - 2015 PNA EA & WP Northern Hemispheric Teleconnection Patterns CPC 1985 - 2015 Composite Reanalysis Earth Systems Research Laboratory NOAA 1948 - 2015 Air Trajectories HYPSLIT Model NOAA 1948 - 2015
The quality of data varies by station, therefore each station was scrutinized for inconsistencies using the methods detailed in Kunkel et al. (2009c). The most common limitation is missing observations. This is especially problematic at volunteer-based stations because they typically have less complete records compared to first-order stations, such as those at an official government site, like an airport. Thus, stations were only analyzed if daily data were recorded for at least 90% of the snowfall season (1 October – 31
May) and for at least 25 of the 30 snowfall seasons (Kunkel et al. 2009c; Hartnett et al.
2014). In instances where a station was reporting observations, but failed to report a daily
32 snowfall total, the snowfall total was set to zero. This was done because COOP observers readily report days with snowfall, but commonly fail to report days of no snow (Rasmussen et al. 2012). Station data were also omitted if the data were flagged by the NCEI for failing at least one of the quality control practices outlined by the National Weather Service’s Snow
Measurement Guidelines (NWS 2012). Finally, inconsistencies in station records can emerge through station relocations. Identifying station relocations is complicated by the use of geolocating satellites, which often improved but modified the geographic coordinates of stations (Kunkel et al. 2009c). Therefore, any stations with an elevation change greater than 10 meters or a change in latitude or longitude greater than 0.15ᵒ were not used in climatological analyses. After filtering for inhomogeneities in daily snowfall records, 60 COOP stations were retained (Figure 2.3; Appendix 9.1).
In addition to dealing with inconsistencies in data, the use of daily snowfall observations can create major challenges when identifying individual snowstorms. First, if snowfall is only measured during a single 24-hour period, then there is a likelihood that the snowfall total may reflect more than one snowstorm. A second challenge is the ability of COOP station observers to choose the observation time (Appendix 9.1). For example, Station A may record observations every 24 hours at 0700 EST, while Station B may record observations every 24 hours at 2300 EST. Both stations record snowfall on the same day; however, snowfall totals for Station A better reflect the previous day’s snowfall, while totals for Station B reflect the current day’s snowfall. Lastly, although flexibility in the timing and frequency of snowfall observations increases observer participation, it is
33
Figure 2.3. Location of the 60 COOP stations used for analysis in Central New York.
potentially problematic when measuring snowfall from storms with a mix of precipitation types (Doesken and Judson 1996). For example, during rain on snow events, rainfall and above freezing temperatures have the potential to melt any accumulated snow. Therefore, depending on the timing and frequency of observations, the observer could sample prior to snowfall, yet after rainfall which may result in an absence of a snowfall measurement.
Thus, to improve the temporal resolution of snowstorms, hourly surface observation summaries were used from fifteen first-order COOP stations within Central New York, accessed at https://www.ncdc.noaa.gov/cdo-web/datatools/lcd (Table 2.1; Figure 2.4).
These data were used to determine the onset, maturation, and dissipation times of storms in Central New York (Perry et al. 2007). The onset of a storm is defined as the hour in
34 which any precipitation is first reported within the study area. Storm maturation corresponds to the hour with the most intense precipitation across the most first-order stations. Whereas the dissipation corresponds to the last hourly report of precipitation in which there was at least a six-hour gap in precipitation reports at any of the fifteen first- order stations (Perry et al. 2007). Thus, if there was an absence of precipitation between storms of less than six hours, then those two storms were considered a single storm. A total of 2055 snowstorms were identified that influenced Central New York between
1985/86 – 2014/15.
Figure 2.4. Location of stations reporting hourly precipitation in Central New York.
The hourly data provided the delineation of individual snowstorms that could then be applied to station data from the daily COOP. In some instances, single-storm snowfall totals for a station included multiple 24-hour totals. In those cases, the snowfall total for the 35 storm was the summation of multiple 24-hour snowfall measurements for that station. The station with the greatest snowfall total was identified for each storm, and this total was used to categorize the magnitude of the storm as either a light (< 10.2 cm), moderate (10.2 cm ≤ x < 25.4 cm), or heavy (≥ 25.4 cm) snowstorm (Kocin and Uccellini 2004b).
2.3 Snowstorm Classification
2.3.1 Data
The National Centers for Environmental Prediction’s (NCEP) “Daily Weather Maps” archive was used to classify snowstorms based on their storm type and zone of cyclogenesis (Table
2.1). Daily Weather Maps were accessed through NOAA’s Central Online Library at https://www.lib.noaa.gov/ collections/imgdocmaps/daily_weather_maps.html and are available from 1 January 1871 – present. Operational weather maps include daily surface weather, 500 hPa heights, maximum and minimum temperatures, and precipitation totals.
Daily Weather Maps were the primary source for classifying snowstorms; however, since observations only occur once every 24 hours, reanalysis maps with a higher temporal resolution were also necessary to identify the exact zone of cyclogenesis. NCEP’s reanalysis data were accessed at www.wpc.ncep.noaa.gov/ncepreanal/ and included archived reanalyses at 12-hr intervals from 00Z 1 January 1948 through 12Z 31 December 2011
(Table 2.1). Reanalysis charts included 200 hPa heights and isotachs; 500 hPa heights and standardized height anomalies; 850 hPa heights, temperatures, and standardized temperature anomalies; and 1000 hPa heights, precipitable water, and standardized precipitable water anomalies. A limitation of this data is that it concluded in December
2011, and thus reanalyses from the Weather Prediction Center (WPC) had to be used for
36 more recent storms (Table 2.1). Reanalysis images were found under the Surface Analysis
Archive at www.wpc.ncep.noaa.gov/archives/web_pages/sfc/sfc_archive.php. The “United
States (CONUS)” and the “U.S. Analysis/Radar Composite” charts were used to display 3-hr intervals from 00Z 1 May 2005 to present. Although this data provides the highest temporal resolution, they were not used to examine all snowstorms due to their limited timespan.
2.3.2 Identification of Non-Cyclonic Snowstorms
Non-cyclonic snowstorms were defined as snowstorms without the presence of a surface low pressure (< 1013 mb) within 150 km of the study area (Figure 2.5; Kelly et al. 2012).
The absence of a cyclone was determined using surface reanalysis charts (Table 2.1). If the storm was classified as non-cyclonic, NCEP/NCAR reanalysis data and data from Syracuse
Hancock International Airport were used to determine if there was a freezing surface air temperature; a temperature gradient of at least 13˚C between the lake surface and the 850 hPa layer; a wind direction with a favorable fetch (e.g. westerly wind) over Lake Ontario; directional shear less than 30˚ between the surface and 850 hPa winds; and 850 hPa winds greater than 5 m s-1, yet less than 20 m s-1 (Niziol et al. 1995; Suriano and Leathers
2017a,b).
If the snowstorm met the previous criteria, U.S. composite surface NEXRAD data from the
NCEI’s NEXRAD Data Archive were obtained at www.ncdc.noaa.gov/data-access/radar- data (Table 2.1). These data were used to examine whether non-cyclonic snowstorms had quasi-stationary, coherent precipitation with a notable connection to the lake; a distinct
37
If Yes:
From Syracuse Hancock International Airport and Surface and 850 hPa reanalysis chats, is there: If Yes: Was If Yes: 1. Freezing surface air? precipitation 2. Temp. gradient between lake surface and 850 hPa >13⁰C? Categorized as a separated by at Categorized a Lake- 3. Winds favorable for a fetch over Lake Ontario or Lake Erie? Lake Snowstorm least 6 hours? Effect Snowstorm 4. Directional shear < 30⁰ between the surface and 850 hPa? 5. 850 hPa winds between 5 and 20 m s-1?
From GOES infrared imagery for storms from 1985/86-1993/94, is there: If No: 6a. Partially visible upwind shore of the lake? Categorized as a Presence of an Categorized as an 7a. Cloud structure not noticeably linked with other cloud masses? Non-Cyclonic Storm upper level Upper Atmospheric and disturbance? Disturbance From NEXRAD data for storms from 1994/95 – 2014/15, is there: Categorized as a 6a. Presence of quasi-stationary precipitation with connection to the lake? Non-Lake Surface low Snowstorm Presence of a 7a. Precipitation distinct from other mesoscale precipitation with cloud Categorized as a (< 1013 mb) front through heights below 2 km Frontal Storm with a closed 8a. Mesoscale precipitation bands that increase in strength downwind of From surface and CNY? isobar within the lake? 500 mb charts: 150 km of If Yes: Categorized as study area? Zone D: If No: Was Did cyclogenesis a Miller Type B storm Categorized as an cyclogenesis occur in Gulf of If No: Categorized as Categorized as a East Coast Storm north of 35⁰N? Mexico? a Miller Type A storm Nor'easter If No: If Yes: Categorized as a Gulf Coast Storm
Categorized as a Non-Lake Zones B, C, or H: Zone B: Categorized as a Hudson Low Snowstorm Zone C: Categorized as a Clipper Categorized as a From surface and upper Canadian Low Zone H: Categorized as a Great Lakes Low atmospheric reanalysis Zone E: Categorized as a Colorado Lows charts, which zone did Zones E, F, or G:
cyclogenesis occur? Zone F: Categorized as a Texas Hook Categorized as a Rocky Low Zone G: Categorized as an Oklahoma Hook Atlantic Ocean as a Categorized as a Tropical Cyclone tropical storm? Figure 2.5. Snowstorm Classification Scheme Diagram
38 mesoscale structure of the precipitation identifiable from other areas of precipitation, with cloud heights often below 2 km; and mesoscale precipitation bands that originate (< 10 km) and increase in strength (i.e. increased reflectivity, depth, or spatial coverage) downwind of the lake with precipitation extended over the lake (Sobash et al. 2000; Laird et al. 2009, 2010). Due to limitations in the length of NEXRAD data, data were only used to classify storms from 1994/95 – 2014/15. Instead, GOES infrared images were obtained for snowstorms from 1985/86 – 1993/94 from the NCEI’s International Satellite Cloud
Climatology Project (ISCCP) at www.ncdc.noaa.gov/isccp/isccp-data-access. Images were used to determine if the upwind shore of the lake was partially visible and if the cloud structure was not noticeably linked with nearby cloud masses (Kelly 1986). Although this data spans the entire study period, the occurrence of clouds does not ensure the production of snowfall. Thus, the use of NEXRAD data to classify snowstorms were preferred over GOES satellite images.
Snowstorms that satisfied the above criteria were categorized as lake snowstorms and all others were classified as non-lake snowstorms (Figure 2.5). If the previous criteria were not satisfied, yet the storm was non-cyclonic, then the storm was named a non-cyclonic storm. For storm’s classified as lake snowstorms, if the precipitation was separated from all other snowstorm’s precipitation by at least six hours, then the storm was also classified as a lake-effect snowstorm. Non-cyclonic storms were further examined using surface and
500 hPa charts. If there was the presence of an upper level disturbance (e.g. low, trough, ridge, etc.) storms were categorized as upper atmospheric disturbances, or a frontal system, storms were categorized as frontal storms. Of the 2055 snowstorms, 814 were classified as lake snowstorms and 1241 were non-lake snowstorms. Of the 1133 39 snowstorms classified as non-cyclonic, 721 were lake-effect snowstorms, 233 upper atmospheric disturbances, and 179 frontal storms. These classifications will be used in the following chapters to analyze lake snowstorms versus non-lake snowstorms, and lake- effect snowstorms versus other types of non-cyclonic and cyclonic snowstorms. Although storms may be classified into multiple categories, the specific storm delineation applied will be driven by the specific research questions being addressed.
2.3.3 Identification of Cyclonic Snowstorms
Since extratropical cyclones are driven by the transient polar jet stream, their formation is highly variable in time and space (Klein 1958; Whittaker and Horn 1981; Jones and Davis
1995). The location of a cyclone’s initial formation will affect the trajectory of its path and therefore its internal characteristics. According to the Glossary of Meteorology, cyclogenesis is the development of cyclonic circulation, or its strengthening around an existing cyclone or depression (Glickman 2000). For this study, cyclogenesis is the moment a closed isobar surrounds a surface low pressure center less than 1013 hPa. Since cyclonic storms periodically strengthen and weaken, barring complete cyclolysis, the area of initial formation was used to define the storm type.
Areas regularly influenced by the jet stream (Robinson and Henderson-Sellers 1999), with strong baroclinicity (Changnon 1969; Whittaker and Horn 1981; Hoskins and Hodges
2002) are common zones of cyclogenesis. From fall to spring, cyclogenesis over North
America is most frequent between 35-40⁰N, with a secondary peak from 50-55⁰N (Klein
1958; Whittaker and Horn 1981). These latitudes correspond to the location of the polar
40 jet stream, which interacts with permanent and semi-permanent anticyclones centered near 30⁰N, creating shear (Klein 1958; Whittaker and Horn 1981). Several zones of cyclogenesis in North America have been identified that produce extratropical cyclones
(e.g. Klein 1957; Changnon 1969; Reitan 1974; Zishka and Smith 1980; Whittaker and Horn
1981; Beckman 1987; Jones and Davis 1995; Mote et al. 1997; Zielinski 2002; Changnon et al. 2008). Shown in Figure 2.6, these zones can be simplified into: the east coast (E), coastal Texas (G), lee of the mountainous western United States (B & C), and lee of the northern Rocky Mountains (A & N).
Figure 2.6. Areas of cyclogenesis in North America from Whittaker and Horn (1981). Cyclogenesis primarily occurs in six locations: (A) Alberta, (B) Great Basin, (C) Colorado, (E) East Coast, (G) Gulf of Mexico, and (N) Northwest Territories.
41
The latitude and longitude of cyclogenesis were determined for the 922 cyclonic storms using surface reanalysis maps (Table 2.1) and were plotted on a 5˚x5˚ grid (Figure 2.7).
Cyclogenesis counts per grid cell (Figure 2.8) and a Getis G Ord hotspot analysis implemented in ArcGIS were used to determine typical areas of cyclogenesis (Figure 2.9).
The analyses identified four primary zones of cyclogenesis that produce storms affecting
Central New York and were assigned a unique storm type: Canadian lows (Zones B, C, and
H), Nor’easters (Zone D), Rocky lows (Zones E, F, and G), and Tropical Cyclones (Zone A).
Due to variability within Canadian lows and Rocky lows, these storms could be further categorized into Hudson lows (Zone B), clippers (Zone C), Colorado lows (Zone E), Texas hooks (Zone F), Oklahoma hooks (Zone G), and Great Lakes lows (Zone H). These eight primary snowstorm types are discussed later and are shown in Figure 2.10.
Figure 2.7. Location of cyclogenesis. The points represent the latitude and longitude of cyclogenesis for the 921 identified cyclonic storms to affect Central New York.
42
Figure 2.8. Cyclogenesis counts per 5˚x5˚ grid cell. (Left) Figure 2.9. Getis G Ord Hot Spot analysis of the locations of cyclogenesis. (Right)
Figure 2.10. Zones of cyclogenesis for cyclonic storms affecting Central New York.
43
Figure 2.11. Typical track of extratropical cyclones affecting Central New York.
2.3.3.1 Clippers
Clippers are categorized as a type of Canadian low, and are defined by the Glossary of
Meteorology as lows that form to the lee of the Rocky Mountains, centered near the province of Alberta or the Northern Territories (Zone C, Figure 2.10; Reitan 1974; Chung et al. 1976; Zishka and Smith 1980; Whittaker and Horn 1981; Thomas and Martin 2007;
Glickman 2000; Petterssen 1956; Nielsen and Dole 1992). Clippers are low-moisture storms since the Rocky Mountains deplete the air of moisture and there are no major nearby moisture sources (Glickman 2000). Therefore, clippers typically fill as they move across the continent (Thomas and Martin 2007). Clippers have a relatively small areal extent, high (> 990 mb) central pressure, and move quickly (13 m s-1), especially compared to other extratropical storms (Thomas and Martin 2007). Even with frequent filling and a
44 high pressure, strong winds (18 m s-1) typically develop due to a strong pressure gradient between the clipper and anticyclones flanked to the south and northwest.
Winter cyclogenesis is more frequent in Alberta than the Northwest Territories (Whittaker and Horn 1981). Once formed, the cyclone tracks to the southeast toward the north-central
United States, and progresses eastward toward the Atlantic Ocean (Figure 2.11; Reitan
1974; Zishka and Smith 1980; Hoskins and Hodges 2002; Hutchinson 1995; Thomas and
Martin 2007). Their low moisture content and fast movement typically result in lower snowfall totals and snow water equivalencies compared to other cyclonic storms. Snowfall totals are often enhanced to the lee of the Great Lakes because of an influx of low-level moisture, initiating lake-effect snowfall after the clipper’s passage (Harms 1973; Vinzani and Changnon 1981; Silberberg 1990; Angel and Isard 1997; Thomas and Martin 2007).
2.3.3.2 Great Lakes Lows
Great Lakes lows are the second type of Canadian low, and form throughout the Great
Lakes and Upper Midwest of the United States (Zone H, Figure 2.10). Typically, they originate as clippers or Colorado lows that have undergone complete cyclolysis (Figure
2.11). As the storm approaches the Great Lakes, baroclinic conditions enhance the formation or reformation of the cyclone. Since these storms form relatively close to the study area, their central pressure is relatively high as they move into Central New York, and their areal extent is small compared to other cyclonic storms. In their classification, Hirsch et al. (2001) did not include these as individual storms most likely because they are often remnants of previous cyclones.
45
2.3.3.3 Hudson Lows
Hudson lows are the final Canadian low and form due to the baroclinicity caused by the
Hudson Bay (Zone B, Figure 2.10). Since they also tend to form from remnant clippers, they display many of the same characteristics, and so were not identified as separate storms by
Hirsch et al. (2001). Hudson lows are associated with polar or Arctic air masses; therefore, they are often accompanied by strong cold fronts, which can extend south from the low, well into the continental United States. The contrast in surface temperatures between air masses is often very strong, with differences exceeding 10˚C (Curry 1983).
2.3.3.4 Nor’easters
A Nor’easter, also referred to as an east coast storm, is any closed low pressure circulation with winds exceeding 10.3 m s-1, which generally tracks to the north or northeast for at least six hours within Zone D of Figure 2.10 (Hirsch et al. 2001; Glickman 2000). Due to the large size of Zone D, Nor’easters can exhibit strong variations in their storm characteristics, so are further categorized into three types: Gulf Lows, Miller Type A Nor’easters, and
Miller Type B Nor’easters (Miller 1946; Mather et al. 1964; Colucci 1976; Dickson and
Namias 1976; Zishka and Smith 1980; Whittaker and Horn 1981; Douglas et al. 1982; Davis et al. 1993; Jones and Davis 1995; Hirsch et al. 2001; Zielinski 2002; Changnon et al. 2008).
Gulf Lows are cyclones forming along or near the Gulf Coast of the United States (Figure
2.11). Cyclogenesis is common from coastal Texas to the Gulf Coast of Florida, due to humid air and a strong baroclinicity associated with the these areas (Jacobs et al. 2005).
Regardless of the longitude, all Gulf Lows form south of 35⁰N. Miller Type A and Miller
46
Type B Nor’easters form along or over the east coast of the United States, and typically track north/northeast parallel with the eastern shore (Figure 2.11). The strong baroclinic conditions created by the Appalachian Mountains and Gulf Stream often cause these storms to undergo rapid development and bombogenesis. Bombogenesis is the rapid deepening of a low pressure, dropping at least 24 millibars over 24 hours and is associated with heavy precipitation storms (Zishka and Smith 1980; Sanders and Gyakum 1980; Jacobs et al.
2005; Cione et al. 1993). The prominent difference between Miller Type A and Miller Type
B Nor’easters is the latitude of cyclogenesis.
Miller Type A Nor’easters form south of 35˚N, over or near the Carolinas, Georgia, or
Florida and just east of the Gulf Stream temperature axis (Figure 2.10; Miller 1946;
Changnon et al. 2008). Whittaker and Horn (1981), Kocin and Uccellini (1990), and
Petterssen (1941) have suggested that cyclogenesis is most frequent lee of the Appalachian
Mountains around 32˚N, because the mountains impede the progress of strong cold fronts to the south-southeast. The synoptic conditions associated with Miller Type A Nor’easters include a cold anticyclone over most of the eastern United States flowing off the continent; the advection of warm maritime air from the Gulf or western Atlantic; a distortion of the cold front into a wave-like pattern; and middle cloud and precipitation formation along the distorted portion of the cold front (Miller 1946). Though Miller Type A Nor’easters are less frequent than Miller Type B Nor’easters (Kocin and Uccellini 1990), research suggests that their southern formation results in higher precipitation totals and snow water equivalencies (Kocin and Uccellini 1990; Davis et al. 1993).
47
Miller Type B Nor’easters originate north of 35˚N between Norfolk, Virginia and Cape Cod,
Massachusetts (Figure 2.10; Miller 1946; Whittaker and Horn 1981). These storms are unique to the eastern United States, and form along the boundary of a remnant cyclone’s warm front (Miller 1946). Though less severe than their southern counterpart, their northern and western formation makes them more likely to affect the Northeast United
States (Miller 1946; Branick 1997; Kocin and Uccellini 2004a; Changnon et al. 2008). The synoptic conditions conducive to the formation of Miller Type B storms consists of an occluded frontal boundary in the Great Lakes; cold continental air trapped between the
Appalachian Mountains and the Gulf Stream; warm maritime air advecting northward into the trapped cold air; cloud and precipitation formation within the trapped cold air; and an area of decreasing pressure dissociated with falling pressure from the primary cyclone center (Miller 1946).
Forming along troughs in the jet stream, prolonged cold outbreaks in the eastern United
States are often associated with heavy snowfall producing Nor’easters (Jones and Davis
1995). Nor’easters are usually large storm systems, bringing heavy snowfall with a high moisture content, and cold and windy conditions throughout the Northeast (Kocin and
Uccellini 2004b). Hirsch et al. (2001) noted that Nor’easters affect the Northeast United
States approximately 11.8 times per winter season (October – April). Historically, the frequency of Nor’easters has fluctuated, with increased activity in the 1950s, fewer storms in the 1980s, and pronounced activity in the 1990s (Mather et al. 1967; Hayden 1981;
Davis et al. 1993; Hirsch et al. 2001; Kocin and Uccellini 2004b).
48
2.3.3.5 Colorado Lows
Colorado lows are a type of Rocky low, and the Glossary of Meteorology defines them as lows that form a definite center near Colorado, on the eastern slopes of the Rocky
Mountains (Zone E, Figure 2.10). Cyclogenesis of these storms is common over the Great
Basin and Pacific Northwest, as the storms typically intensify on the leeward side of the
Rocky Mountains within the Colorado Front Range (Whittaker and Horn 1981; Zielinski
2002; Changnon and Changnon 2006). Cyclogenesis involves the presence of continental polar air emanating from Canada, and maritime tropical air originating over the Gulf of
Mexico (Lahey et al. 1960; Whittaker and Horn 1981). The stronger the temperature and moisture gradients, the more energy for storm development, and the stronger the storm.
Thus, less (more) snow and ice cover in the Arctic during the fall results in a(n) reduced
(increased) pressure gradient force, and ultimately a slower (faster) jet stream and weaker
(stronger) Colorado lows (Whittaker and Horn 1981).
Once formed, Colorado lows track to the northeast, across the central United States and into the Great Lakes region (Figure 2.11; Changnon 1969; Zielinski 2002; Branick 1997;
Changnon et al. 2008). Their large areal extent and curved track results in the potential for severe weather over a large swath of the United States. Though snowfall is common in the northwest sector of the storm, other sectors typically produce sleet, freezing rain, rain, thunderstorms, high winds, and even tornadoes (Saylor and Fawcett 1965; Kreitzberg and
Brown 1970; Galway and Pearson 1981; Rydzik and Desai 2014; Changnon et al. 2008).
49
2.3.3.6 Oklahoma Hooks
Oklahoma hooks are the second type of Rocky low and are sometimes referred to as
Panhandle hooks. These storms form within Zone G (Figure 2.10) and are low-pressure systems originating in the panhandle regions of Texas or Oklahoma. These storms initially move east, then recurve northeast towards the upper Midwest or Great Lakes region
(Figure 2.11; NWS 2004). The more pronounced trough in the jet stream during these storms compared to Colorado lows, often results in a greater influence from the Gulf of
Mexico. This often leads to storms with lower surface pressures, stronger winds, and greater precipitation totals (Bentley and Horstmeyer 1998; Zielinski 2002; Changnon et al.
2008).
2.3.3.7 Texas Hooks
Cyclogenesis is also common to the lee of the southern Rocky Mountains and over coastal
Texas, identified as Zone F in Figure 2.10. The storms forming here are referred to as Texas hooks, and are also a type of Rocky low. Due to the southern formation of these storms, when they track east, they pass, albeit briefly, directly over the Gulf of Mexico (Figure 2.11).
This passage results in an influx of moisture and heat from the Gulf to the overlying air, triggering heavier precipitation and a higher potential for severe weather than Oklahoma hooks. Texas hooks are also distinct from Gulf Lows due to the pronounced curve in their track, directing them to the west of the Appalachian Mountains, similar to the tracks of
Oklahoma hooks and Colorado lows. This strong curvature causes more intense precipitation over the Great Lakes region than the coastal Atlantic.
50
2.3.3.8 Tropical Cyclones
Tropical cyclones are the final snowstorm type that has affected Central New York during the study period. According to the NWS (2004), a tropical cyclone is a warm-core, non- frontal synoptic-scale cyclone, originating over tropical or subtropical waters, with organized deep convection and a closed surface wind circulation about a well-defined center. If these storms do produce snowfall it is after they have transitioned into extratropical cyclones, in which case they are cold-core systems forming along an air mass boundary (Hart and Evans 2001; Evans et al. 2017; Ritchie and Elsberry 2007). However, since this classification is based on the area of cyclogenesis, storms originating as tropical cyclones were assigned to Zone A (Figure 2.10).
2.4 Snowfall Contributions from Different Snowstorm Types
A specific question that I address in this research is how the different source regions contribute to the seasonal snowfall in Central New York. To do this, I assessed the total seasonal snowfall associated with a specific storm class and compared it to the relative percent contribution of other snowstorms. For example, the question of which storms contribute lake snow versus non-lake snow can be directly addressed by examining the seasonal total and percent contribution of the storms classified as “lake snowstorms” with those classified as “non-lake snowstorms.” For this study, the seasonal snowfall for each class of storms was defined as the summation of the largest snowfall totals for each storm within the same class of storms from 1 July to 30 June.
51
The amount of snowfall produced by a storm depends on the observed scale. Therefore, I calculated snowfall contributions at a variety of spatial scales. In Chapter 3, I calculate snowfall contributions for Central New York, and then for the five snowfall subregions identified by Hartnett et al. (2014). Contributions at the local level were also found using interpolated surfaces derived from the cokriging interpolation method.
Several geographic and locational factors play an important role in determining the amount of seasonal snow a station receives. Due to the sensitivity of snowfall accumulations to environmental factors, linear mixed-effect models were developed capable of estimating seasonal snowfall contributions for any location within Central New York. Station data for the models were obtained from the NCEI’s Climate Data Online server and included the subregion, elevation, latitude, and longitude for each station (Table 2.1).
Subregion was used as a random effect in the model, and is incorporated to suggest that stations within the same region behave more similarly than stations within different subregions (Hartnett et al. 2014). Elevation was included because it is widely understood that higher elevations tend toward larger snowfall totals due to colder temperatures and orographic enhancement (Peace and Sykes 1966; Hill 1971; Dewey 1979a; Laird and
Kristovich 2004). Latitude and longitude were incorporated to represent the locational attributes of a location. Latitude has a greater bearing on the physical processes, as higher latitudes tend toward cooler average air temperatures. In addition, higher latitudes suggest that the location is east of Lake Ontario, which is prime territory for lake-effect snow (Peace and Sykes 1966). Longitude was also incorporated to represent location.
52
Since Central New York is situated between the Lake Ontario and the Atlantic Ocean, a location’s proximity to each can have a considerable effect on its snowfall. Therefore, a location’s east-west position can dictate whether snowfall is more influenced by the Great
Lakes or the Atlantic Ocean.
Distance from Lake Ontario and exposure were also derived for use in the model. The distance from Lake Ontario was calculated in ArcGIS by measuring the distance between a station and an arbitrary point over eastern Lake Ontario (43.6307˚N, 76.7962˚W). Since no single point on Lake Ontario is responsible for lake-effect snow, this point was chosen to represent the center point of eastern Lake Ontario, the half of the lake most responsible for lake-effect snow in Central New York. Although longitude and distance from the lake may measure similar attributes, they were both included because there are locations, particularly south of the lake, that may be a similar distance to the lake and near the same latitude, but their longitudes are quite different. In addition, longitude only incorporates the east-west position of a location, but ignores its proximity to Lake Ontario.
To calculate exposure, three fishnet grids (2.5 km, 5 km, and 10 km) were created and overlaid on an elevation raster of the study area. The average elevation for each grid cell was determined, along with that of nearby cells. The average elevation of nearby cells was calculated by averaging the elevation of all grid cells within a distance twice the resolution of the original cell. For example, the average adjacent elevation for a 2.5 km x 2.5 km grid cell was calculated by averaging the elevation of all cells within 5 km of the original cell.
53
Figure 2.12. Elevation exposure (m) for different grid cells within Central New York. Three different grid cells resolutions are shown: 2.5 km (top left), 5.0 km (top right), 10.0 km (bottom). 54
The exposure of an individual grid cell was calculated by subtracting the average elevation of the grid cell by the average elevation of adjacent grid cells (Figure 2.12). A positive
(negative) number indicates a relatively higher (lower) elevation than nearby areas. The exposure of each station was then extracted from the corresponding grid cell and used as a potential explanatory variable. The previous methods were used rather than limiting exposure from a particular direction because snowstorms can come from all different directions in Central New York. Although, lake-effect snowfall predominately comes from the west or northwest, snow can often come from the northeast or southeast during
Nor’easters, from the south during Rocky lows, and from the north during upper atmospheric disturbances. By limiting the exposure to one direction, the exposure to other wind directions is ignored. Thus, future improvements to the model may incorporate a more specific exposure variable that is adjusted to the snowstorm type.
After deriving the variables used in the model, the significance of the predictor variables in influencing seasonal snowfall contributions for a location was tested using a linear regression incorporating the suite of variables. If the -value of a predictor variable was less than or equal to 0.05, then the variable was assumed to significantly influence snowfall contributions. Using the predictor variables that significantly influence snowfall from each class of storms, linear mixed-effects models were constructed using a combination of the significant predictor variables with county and snowfall subregion as random effects
(Symonds and Moussalli 2011). For example, if three variables (A, B and C) were shown to significantly influence snowfall from a storm, then eight different models were created.
55
푌̂ = 푏0+ region Null Model 푌̂ = 푏0 + 푏1퐴 + region Model 1 푌̂ = 푏0 + 푏2퐵 + region Model 2 푌̂ = 푏0 + 푏3퐶 + region Model 3 푌̂ = 푏0 + 푏1퐴 + 푏2퐵 + region Model 4 푌̂ = 푏0 + 푏1퐴 + 푏3퐶 + region Model 5 푌̂ = 푏0 + 푏2퐵 + 푏3퐶 + region Model 6 푌̂ = 푏0 + 푏1퐴 + 푏2퐵 + 푏3퐶 + region Model 7
The Akaike information criterion (AIC) was then used to rank models from the best model to the worst model. Although used increasingly in ecological studies, AIC has been seldomly used in the atmospheric sciences (Chowdhury and Sharma 2009; Kharin and
Zwiers 2002; Woolhiser 2008; Wagenmakers and Farrell 2004; Nakagawa and Schielzeth
2013; Symonds and Moussalli 2011). The AIC is a novel model selection method that uses information theory to compare and rank multiple competing models. The top model estimates the parameters which best reflect the ‘true’ processes under examination
(Burnham and Anderson 2002; Symonds and Moussalli 2011). Model selection for the AIC incorporates the variance explained by the model, but penalizing models that explain only a slightly greater variance after adding an addition explanatory variable. AIC improves upon traditional hypothesis testing because it is multivariate and can test a suite of variables that may be associated with a particular process, rather than simply a single predictor variable. By concurrently comparing multiple models, model uncertainty can be quantified, and selection can include a set of models rather than a single, non-descriptive model (Symonds and Moussalli 2011). AIC is calculated as,
AIC = -2 ln(L) + 2k Equation 1 where the maximum likelihood estimate for the model (L) and the number of fitted parameters (k) are used (Akaike 1974). A drawback of this method is that models are only
56 an approximation of the process, as the user dictates the variables that are included. In addition, AIC is a comparison of the models created, thus, if model creation is poor, the ‘top’ model may be insignificant.
Once the top model was determined using the AIC, an ANCOVA was used to extract the R2 value of the top model. The R2 is used to represent the variance explained by the model
(Nakagawa and Schielzeth 2013), and the model output was used to assess the geographic factors which significantly influence seasonal snowfall contributions for a given location.
Therefore, even though this method has not been widely used in the atmospheric sciences, its use in ecological studies showcase its applicability.
2.5 Synoptic Classification of Different Snowstorm Types and Magnitudes
The synoptic classification of the atmospheric conditions during individual snowstorms allows for a better understanding of the conditions which promote the development and growth of snowstorms in Central New York (Suriano and Leathers 2017b). It also enables better predictions of snowfall totals throughout the region based on the synoptic conditions of the atmosphere.
Synoptic classification techniques have been applied to examine patterns in snowfall in
North America (Ellis and Leathers 1996; Leathers and Ellis 1996; Karmosky 2007; Suriano and Leathers 2017b; Leathers et al. 2002). Ellis and Leathers (1996) identified five synoptic types which generate lake-effect snow in Syracuse, New York, while Suriano and
Leathers (2017b) found two additional synoptic types associated with lake-effect snow in
57 the eastern Great Lakes region. Leathers and Ellis (1996) found that the synoptic conditions strongly influence the frequency and snowfall intensity of snowstorms leeward of the Great Lakes. However, no studies have examined the differences in synoptic conditions that promote certain types of snowstorms, nor the synoptic conditions that favor heavy snowstorms over moderate and light snowstorms. Thus, Chapter 4 examines the synoptic atmospheric conditions and the air trajectories of different snowstorm types with different magnitudes.
Synoptic conditions were examined using composite NCEP reanalysis data (Kalnay et al.
1996). Composite plots were created for North America using the Earth Systems Research
Laboratory - Physical Sciences Division’s server found at http://www.esrl.noaa.gov/psd/
(Table 2.1). Daily mean composite reanalysis images were constructed using twenty snowstorms of the same type and magnitude. Reanalysis plots for the upper atmosphere included the 850 hPa geopotential height, 850 hPa air temperatures, and the 850 hPa vector wind speeds. Surface analyses included sea level pressure, surface air temperatures, and surface vector wind speeds.
Storm trajectories across the study area were examined using ArcGIS. To maintain consistency, the storms used for the reanalysis composite plots were again used. The relatively large spatial extent of Central New York means that air patterns may vary across the region. Thus, upper air trajectories for snowstorms were examined for Syracuse
(43.02481˚N, 76.1474˚W), Watertown (43.9748˚N, 75.9108˚W), and Utica (43.1009˚N,
75.2327˚W). Air trajectories were obtained from NOAA’s web-based Air Resource
58
Laboratory Hybrid Single Particle Lagrangian Integrated Trajectory (HYSPLIT) model found at https://ready.arl.noaa.gov/HYSPLIT_traj.php (Table 2.1; Stein et al. 2015; Rolph et al. 2017). According to NOAA, the HYSPLIT model is “a complete system for computing simple air parcel trajectories, as well as complex transport, dispersion, chemical transformation, and deposition simulations.” The model uses a combination of Eulerian and Lagrangian methods to calculate air trajectories (Stein et al. 2015). Since this research examines airflow into Central New York, ‘normal’ 72-hour back trajectories were obtained using the global meteorological reanalysis. Model specificities included vertical velocity, a level 1 height of 500 hPa (meters AGL), and the output trajectory in a GIS shapefile. The trajectory start time was the hour of storm maturation. These data were then used to assess the differences in air trajectories between the different snowstorm types, and how differences influence storm magnitude.
2.6 Trends in Snowfall Contributions from Different Snowstorm Types
Assessments of seasonal changes in North American snowfall since the early 20th century show that snowfall within the Great Lakes region has increased, whereas snowfall in areas less prone to lake-effect snow have generally decreased (Kunkel et al. 2009a; Serreze et al.
1998; Berger et al. 2002; Kunkel et al. 2013b; Norton and Bolsenga 1993; Kunkel et al.
2009c; Burnett et al. 2003; Knowles et al. 2006; Hartnett et al. 2014; Leathers and Ellis
1996). The assumption is that the increased snowfall in the Great Lakes region is due to an increase in lake-effect snow, while the decrease is associated with decreased non-lake- effect snow. However, none of these studies directly assess trends for different storm types. To address this gap, the purpose of Chapter 5 is to examine seasonal snowfall totals
59 for individual storm types in Central New York for the period 1985/86 – 2014/15. This will provide a better understanding of how and why seasonal snowfall totals are changing in this region, and better predict how they may change in the future.
To do this, seasonal snowfall totals were examined for the different snowstorm types identified in Section 2.3 for the five snowfall subregions of Central New York. Since data did not violate normality (Kolmogorov-Smirnov test) nor homoscedasticity (Bartlett test), linear regressions were used to determine regional trends in seasonal snowfall contributions and seasonal percent snowfall contributions for each snowstorm type. Data were also examined for non-linearity in the seasonal snowfall trendlines (Bard and
Kristovich 2012; Hartnett et al. 2014), using seven-year snowfall trends with a one-year moving window. Although long-term trends highlight how snowfall contributions have changed over time, they provide little information as to why these changes occur. Previous researchers have noted that seasonal snowfall totals are strongly influenced by air temperatures, lake surface temperatures, and ice cover on the lake (Tsuboki et al. 1989;
Segal and Kubesh 1996; Hanson et al. 1992; Wang et al. 2012; Notaro et al. 2015). Thus, in
Chapter 5 and Chapter 6 I examine the environmental conditions associated with the air and lake and the teleconnections that have the greatest influence on seasonal snowfall contributions from different snowstorm types.
2.6.1 Environmental Variables
The Great Lakes and atmosphere have an important influence on the observed seasonal snowfall totals. Lake conditions are examined in Chapter 5 using data from the NOAA’s
60
CoastWatch: Great Lakes Node at http://coastwatch.glerl.noaa.gov/ (Table 2.1). The average surface water temperature (⁰C) from 1995-2015, and the Great Lakes’ ice concentration (%) from 2008-2015 were acquired. For seasonal atmospheric conditions, data were obtained from the NCEI’s Climate Data Online server at https://www.ncdc.noaa. gov/cdo-web/ for Syracuse Hancock International Airport from 1 July 1985 – 30 June 2015
(Table 2.1). Data include the number of days the minimum temperature reached at least
0˚C, number of days the minimum temperature reached at least -17.8˚C, number of days the maximum temperature was at most 0˚C, the average temperature, the average winter
(November – March) temperature, the average maximum temperature, the average maximum winter temperature, the average minimum temperature, the average minimum winter temperature, the number of precipitation days with at least 0.25 cm, the number of winter precipitation days with at least 0.25 cm, average wind speed, average winter wind speed, the number of precipitation days with at least 1.3 cm, the number of winter precipitation days with at least 1.3 cm, the number of precipitation days with at least 2.5 cm, and the average number of winter precipitation days with at least 2.5 cm.
The normality and homoscedasticity of the data were tested to examine the appropriateness of linear regression models. If violated, data were transformed to satisfy the assumptions needed. I then used these data to test which predictor variables most influence seasonal snowfall totals from different snowstorms in Central New York using similar procedures outlined in Section 2.4. I used three separate linear regressions to determine the significance of the potential predictor variables including a model with all the lake temperature variables (Appendix 8.3), a model with ice cover on the Great Lakes
61
(Appendix 8.4), and a model with air temperature and precipitation data (Appendix 8.5).
Using the predictor variables that significantly influence snowfall from each class of storms, mixed-effects linear models were constructed using a combination of the significant predictor variables with region as a random effect (Symonds and Moussalli 2011). The AIC was then used to rank models and an ANCOVA was used to extract the R2 value of the top model.
2.6.2 Teleconnection Patterns
Numerous teleconnections have been shown to influence North American snowfall (e.g.
Serreze et al. 1998; Wise et al. 2015; Ghatak et al. 2010; Ge and Gong 2009). The main modes of variability and the influence of teleconnections on North American snowfall have been mostly examined through principal component analyses (e.g. Ge and Gong 2009;
McCabe and Dettinger 2002; Suriano and Leathers 2017a; Ellis and Leathers 1996;
Kalkstein and Corrigan 1986; Gutzler and Rosen 2002; Siegert et al. 2016). Although previous studies have found linkages between teleconnections and North American snowfall, results are not always consistent nor substantial. Thus, the purpose of Chapter 6 is to examine the influence of different teleconnection patterns on seasonal snowfall totals from different snowstorm types using linear mixed-effects models and the AIC.
Although not widely used, the AIC has been used to examine the influence of teleconnection indices on the climate. Woolhiser (2008) used the AIC to determine the top models representing the combined effects of ENSO and the PDO on precipitation in the Southwest
United States. Villarini et al. (2010) also used the AIC to examine the influence of
62 teleconnections (ENSO and NAO), in addition to tropical Atlantic SSTs and tropical mean
SSTs, on the frequency of landfalling hurricanes in the United States. Beaulieu and Killick
(2018) suggest that an advantage of the AIC is that it can distinguish between changes due to climate change from those due to underlying processes, by recognizing abrupt changes from trends. Another advantage is that it can correct for bias in random small-scale variability due to spatial and temporal variations in precipitation (Wong et al. 2014). This highlights the applicability of this method for estimating seasonal snowfall within Central
New York. Therefore, the use of the AIC in this research is as much exploratory, as it is a proven method to observe snowfall.
Teleconnection indices for the models were obtained at monthly intervals for the AO, EA,
ENSO, NAO, PDO, PNA, and WP from July 1985 – June 2015. Data were obtained for the AO,
NAO, PDO, and PNA from NCEI’s Climate Monitoring server at https://www.ncdc.noaa.gov/ teleconnections/ (Table 2.1). Since ENSO data are represented by equatorial sea surface temperatures that vary throughout the equatorial Pacific, data for three Niño regions were acquired: Niño 3, Niño 4, and Niño 3.4 (Bjerknes 1969; Rasmusson and Carpenter 1982;
Wyrtki 1985). Data for the EA and WP were obtained from the Climate Prediction Center’s online server at http://www.cpc.ncep. noaa.gov/data/teledoc/teleintro.shtml (Table 2.1).
The data were processed by testing for normality, homoscedasticity and collinearity. Since data were normally distributed and homoscedastic, the assumptions for the use of linear regressions were satisfied. Minimal collinearity between two teleconnection patterns was assumed if there was a correlation between -0.6 and 0.6. If collinearity did exist, then only
63 one of the teleconnection patterns was used in the model development. Correlation plots examining potential collinearity between teleconnection patterns are shown in Figure 2.13.
Figure 2.13. Correlation plots between teleconnection patterns used in the analysis for model development.
To determine the teleconnection patterns to incorporate into model construction, I used a similar set of procedures outlined in Section 2.4. I first tested the significance of each teleconnection pattern on influencing seasonal snowfall totals for the different subregions.
Using the significant teleconnection patterns, I then created a combination of linear mixed- effects models using subregion as a random effect. The top model was then identified using
64 the AIC and an ANCOVA was used to extract the R2 value of that top model. The model results were then used to evaluate the influence of different teleconnection patterns on the seasonal snowfall totals from different snowstorms types affecting Central New York.
2.7 Conclusion
Overall, the purpose of this research is to examine in detail the influence of individual snowstorms on seasonal snowfall totals in Central New York. This research helps fill in a gap of our understanding of snowfall at the regional level within the Great Lakes region.
The methods used to classify snowstorms in this chapter can be used at a broader scale for storm classification in any region. Updates will have to be made, regarding the types of storms to influence an area, but the general procedure for classifying storms can be applied to other regions. The data described in this chapter are used throughout the dissertation to address the research objectives introduced in Chapter 1. A blend of standard techniques and new techniques adapted for this research are used throughout the dissertation.
Additional details will be provided in each chapter as necessary and this chapter will be referred back to where appropriate to identify the dataset(s) used for analysis.
65
3.0 THE INFLUENCE OF SNOWSTORM TYPE ON THE SPATIAL DISTRIBUTION OF SNOWFALL AND ITS RELATIVE CONTRIBUTION TO SEASONAL SNOWFALL TOTALS – A SCALE ISSUE IN THE GREAT LAKES REGION
3.1 Introduction
Located at the eastern extent of the Great Lakes snowbelt, Central New York has been the focus of a great deal of snowfall research (Veals and Steenburgh 2015; Suriano and
Leathers 2016; Kunkel et al. 2009a; Niziol 1987; Reinking et al. 1993; Suriano and Leathers
2017a,b; Hartnett et al. 2014). Seasonal snowfall totals in this region are frequently the highest recorded totals east of the Rocky Mountains. This is in large part due to the contribution from lake-effect snowstorms, the passage of midlatitude cyclones, and the proximity of the Atlantic Ocean as a further moisture source. This study provides a detailed picture of snowfall at multiple spatial scales in order to determine where and how localized the large snowfall totals are. To examine the relative importance of moisture sources and other environmental factors that influence the spatially varying snowfall totals, a predictive model is developed.
Previous research generally indicates that lake-effect snow and lake-enhanced snow account for approximately half of the seasonal snowfall in the Great Lakes region as a
(Eichenlaub 1970; Miner and Fritsch 1997; Liu and Moore 2004). Veals and Steenburgh
(2015) note a slightly higher contribution, 61-76%, to the lee of Lake Ontario over the Tug
Hill. However, the Tug Hill is unique compared to other areas in the Great Lakes region, because of its exceptionally high seasonal snowfall totals (Saslo and Greybush 2017;
Minder et al. 2015; Campbell et al. 2016). The high seasonal snowfall totals in the Tug Hill are partly responsible for Hartnett et al. (2014) classifying snowfall in Central New York
66 into five distinct subregions (Figure 3.1). The authors suggest that snowfall totals and trends within these regions behave similarly, thus two stations within the same region have more similar snowfall totals and trends compared to two stations in different regions.
The authors found that utilizing the snowfall subregions allows for improved spatial resolution to analyze snowfall trends, whilst removing the effects of missing data and inhomogeneities present at the station level.
Figure 3.1. Five Central New York snowfall subregions. Included are the geographic features of the area and average snowfall for COOP stations from 1931/32 – 2014/15.
67
Studies suggest that when utilized properly (Kunkel et al. 2009c, 2007), station data provides the best representation of the spatial variability of snowfall within the Great
Lakes region (Burnett et al. 2003; Kunkel et al. 2009a; Hartnett et al. 2014, etc.). This is because the localized patterns of snowfall can be highly dependent on both geographic factors and the specific type of storm system that affects the region. Geographic factors, such as the elevation, the exposure of an area to an approaching storm, and the distance from moisture sources, can all influence how the overlying air interacts with the surface
(Grünewald et al. 2014; Johnson and Hanson 1995; Liu et al. 2011; Perry et al. 2007; Veals and Steenburgh 2015; Niziol et al. 1995). Hill (1971) notes that a 100-meter rise in elevation to the lee of the Great Lakes leads to a 25-50 cm increase in seasonal snowfall totals. Perry et al. (2007) suggest that the exposure of an area has a greater influence than its elevation, as an area surrounded by lower elevated terrain tends toward higher snowfall totals than an area surrounded by similarly elevated or higher elevated areas. Similarly, the type of snowstorm affects the total amount of moisture available, the extent of fetch across one or more lakes, and the atmospheric lapse rate, and potentially the intensity of the precipitation (Saslo and Greybush 2017; White et al. 2010; Ware et al. 2006; Niziol et al.
1995; Liu and Moore 2004; Lawrimore et al. 2014; Changnon et al. 2008; Mullens et al.
2016; Baxter et al. 2005). Laird et al. (2003) and Laird and Kristovich (2004) note that the morphology of lake-effect snowstorms is often linked to surface wind speeds and the fetch across the lake. The complexity and multiplicity of variables that therefore contribute to the magnitude of a storm, and the spatial distribution of snowfall associated with any one storm have yet to be examined in detail. So, although general statements on the percentage
68 of the seasonal snow deriving from lake-effect storms can be made, it may in fact vary quite considerably over a small region, such as Central New York.
The purpose of this chapter then is multifold. First, the contribution of different snowstorm types to seasonal snowfall totals is examined at the regional scale for all of
Central New York, at the subregional scale using the five subregions of Central New York, and at the local scale using station data. From this, assessments of the overall significance of a snowstorm type in Central New York can be made in terms of both the magnitude of the storm as well as how much snow it is likely to contribute in any one year. Second, to interpret the spatial variation in the amount of snow received across the region it is important to understand the linkages between storm type and regional geography.
Therefore, a regression model was developed based on locational and topographic factors and their influence on the spatial patterns of snowfall that emerged from the passage of different storm types. The influence of topography on snowfall contributions provided the background necessary to interpret maps depicting the percentage of the seasonal snowfall contributed by different storm types. Ultimately, these maps were used to assess the percentage of snowfall from different storms throughout parts of Central New York.
Snowfall prognostications in Central New York are important from both the short term, forecasting perspective, and from longer term climate predictions. Therefore, knowledge of the relative amount of snowfall from an individual storm type in specific parts of the region can provide forecasters with useful information to help with emergency planning and management, and resource allocation (Rooney 1967; Zhu and Wang 2016). Secondly,
69 climate change scenarios must consider not only temperature changes, but also changes in the frequency and intensity of storm types associated with larger scale shifts in the jet stream (Suriano and Leathers 2016; Notaro et al. 2013b; Kunkel et al. 2002). Unless the specific seasonal contribution of an individual storm type to the local, subregional, and regional snowfall is known, then these longer-term predictions are likely inaccurate.
3.2 Methods
3.2.1 Data
The data used to develop the database of the 2055 storms used in these analyses are detailed in Chapter 2. Because this analysis uses multiple spatial scales to examine the impact of storms, the storm classification scheme was used in both a summarized form and an extended form. The summarized form included the five general snowstorm types
(Canadian lows, Rocky lows, non-cyclonic storms, Nor’easters, and lake-effect snowstorms), and lake snowstorms versus non-lake snowstorms (Table 3.1). The extended form included the previous seven storms with the additional eight contributing snowstorm types (Table 3.1).
The extended form of the classification was applied to the station-level modeling due to the importance of storm longevity, source region and track on how the storm interacts with the study area. Tropical cyclones were omitted from this analysis due to their low frequency of occurrence. For the assessments of storm magnitudes and frequencies, because of the similarities across some of the contributing storms, the generalized categorization was used. In addition, the division between lake snowstorms or non-lake snowstorms was used
70 to analyze the importance of the Great Lakes on seasonal snowfall totals in this region.
Lake snowstorms differ from lake-effect snowstorms as they also include snowfall from lake-enhanced snowstorms. In this context, lake-enhanced snowstorms are defined as storms that fulfill the requirements to be classified as a lake-effect storms (see Figure 2.5 for details on how storms were classified), but have a cloud structure noticeably linked to other cloud masses; have precipitation that is not distinct from other mesoscale precipitation; or precipitation is not separated by at least six hours from the precipitation of another defined snowstorm. Therefore, there are situations in which a cyclonic storm, such as a Nor’easter, may also be categorized as a lake snowstorm. Since a cyclonic storm can also be classified as a lake snowstorm, comparisons were only drawn between lake snowstorms and non-lake snowstorms and not between the other snowstorm types.
Table 3.1. Storm type classification applied in this chapter and their average seasonal (July – June) frequency and snowfall. Confidence intervals are at the 95% confidence. General Storm Type Contributing Storm Types Storms yr-1 Snowfall (cm) yr-1 Lake Snowstorms Any 27.1 2.0 659.6 58.1 Non-Lake Snowstorms Any 41.4 2.5 736.4 65.8 Clippers 7.1 0.6 115.4 13.5 G.Lakes Lows 4.0 0.4 63.2 8.2 Canadian Lows Hudson Lows 1.1 0.2 19.9 4.9 Total 12.2 0.8 198.5 17.6 Lake-Effect Snowstorms Lake-Effect Snowstorms 24.0 1.0 583.6 27.8 Frontal Storms 13.7 0.8 87.6 9.5 Non-Cyclonic Snowstorms Upper Disturbance Storms 7.8 0.6 137.9 12.4 Total 13.7 0.8 225.5 18.0 Nor'easters Nor'easters 7.7 0.5 206.1 19.1 Colorado Lows 4.8 0.4 80.8 9.4 Oklahoma Hooks 3.4 0.3 59.5 7.2 Rocky Lows Texas Hooks 2.5 0.2 41.0 5.2 Total 10.7 0.5 181.3 10.8
71
3.2.2 Analysis
The average seasonal frequency and the average snowfall produced by each of the snowstorm types were calculated for all of Central New York from the 1985/86 season to the 2014/15 season (Table 3.1). Seasonal snowfall totals were calculated by identifying the largest snowfall total received amongst stations within Central New York for each individual storm and then summing them over a winter season (1 July – 30 June 30).
Therefore, the seasonal values represent the regional maximum snowfall (see Section 2.2 for details). The percent contributions of different snowstorm types were also calculated during the study period. This was done by comparing the seasonal snowfall total for an individual storm type to the total seasonal snowfall for Central New York. Since the data were normally distributed (Kolmogorov-Smirnov test) and homoscedastic (Bartlett test), differences between storm types and their frequency were tested ( = 0.05) using parametric two-sample difference tests and ANOVAs.
Individual storm totals were subdivided into storm magnitude categories of light representing less than 10.2 cm, moderate with snowfall between 10.2 and 25.4 cm, and heavy with snowfall greater than 25.4 cm. Since data were again normally distributed and homoscedastic, differences between storm types and their magnitude were tested using two-sample difference tests and ANOVAs. A similar sequence of procedures was then used to observe the average snowfall and percent contribution of different snowstorm types to seasonal snowfall totals at the subregional scale. This was done by calculating seasonal snowfall totals for each subregion using the stations outlined in Table 3.2.
72
Subregional analyses highlight some of the spatial variation lost when examining seasonal snowfall contributions at the regional scale; however, previous studies note that snowfall, especially within the Great Lakes region, can vary from location to location (Peace and
Sykes 1966; Ellis and Leathers 1996; Ballentine et al. 1998). This is highlighted by the influence of lake-effect snowstorms which have the ability to produce whiteout conditions with clear skies only kilometers away (Niziol 1987). To account for the spatial variability of snowfall within the study region, I identify the typical patterns of snowfall distribution at the station level produced by each snowstorm type, patterns that reflect both the underlying regional geography and the storm’s characteristics. Patterns are observed using percent contributions rather than seasonal snowfall totals to compare snowfall across the region and remove some of the bias introduced through higher snowfall totals in
Regions 4 and 5.
Although analyses at the station level increase the spatial resolution, there are inherent drawbacks. First, data are less reliable and are more susceptible to missing and inaccurate observations (Kunkel et al. 2007; Wu et al. 2005; Fiebrich and Crawford 2009; Leeper et al.
2015). This includes inaccuracies in the proper measurement practices outlined by
Doesken and Judson (1996). Station data are also more influenced by the observer, as the frequency of observations, the types of observations, and the hour of the observation(s) is at the discretion of the observer. Lastly, analyses at the local level are more susceptible to uncertainty. For example, seasonal snowfall contributions are likely more accurate in areas where stations are clustered in comparison to areas with few if any stations. However, even with these drawbacks, station data allow for a better representation of any small-
73 scale variations in seasonal snowfall totals in the Great Lakes region. The use of station data also allows for a better understanding of the microclimatic influences of the environment on seasonal snowfall contributions. For example, analyses at the station level can help determine the influence of topography or absolute location on seasonal snowfall totals.
To lessen some of the biases introduced when working with station data, methods proposed by Kunkel et al. (2007) were used to filter COOP stations for inhomogeneities.
Due to the climatological focus of these analyses, stations were only used if consistent daily observations had been recorded for at least 27 of the 30 snowfall seasons (Kunkel et al.
2009c). This meant that snowfall contributions were calculated for twenty-six of the sixty original COOP stations (Table 3.2). Average seasonal snowfall contributions from these stations were then used to create interpolated surfaces using simple Kriging routines with no trend removal in ArcGIS (Eynon 1988; Daly et al. 1994; Guan et al. 2005).
Studies have suggested that environmental conditions such as topography, orography, and distance from a lake can influence the amount of snowfall produced at a location, especially in lake-effect prone regions (Alcott and Steenburgh 2013; Veals and Steenburgh 2015;
Niziol et al. 1995; Giorgi et al. 1997; Sharples et al. 2005; Perry et al. 2007; Hill 1971;
Dewey 1979b; Pease et al. 1988; Laird and Kristovich 2004; McCabe et al. 2007; Veals et al.
2018). Alcott and Steenburgh (2013) suggest that the concaved-shaped terrain to the lee of the Great Salt Lake enhances storm intensity by reinforcing the lake-breeze-induced convergence zone. Within the Great Lakes, Niziol et al. (1995), Hill (1971), and Wilson
74
(1977) note that annual snowfall totals increase with elevation to the lee of the lakes. Veals and Steenburgh (2015) suggest that there is an inland/orographic intensification of lake- effect snowbands over the western slope of the Tug Hill and a possible shadow-effect to the lee of the Tug Hill. The authors however, focused solely on the orographic effects of the
Tug Hill on lake-effect snowstorms; but as Campbell et al. (2016) note, the influences of orography depend on characteristics of the larger-scale environment, the incident flow, atmospheric stability, the topographic characteristics, and moisture availability. How these factors interact with different storm types is poorly documented and unclear.
Table 3.2. Central New York COOP stations by region. COOP stations used for seasonal snowfall contributions are in bold. Region 1 Region 2 Region 3 Region 4 Region 5 AUB BAIN DDAM BMOOS BARN AUR CHEP FRANK BOON BEAV CAY CVAL GRIFF BREW BENN BALD COOP HINK CAM HOOK CINCY GRN LFALLS CON LOW CORN MARY NLOND FUL OSW CORT MOVILLE NEW HIGH PUL FREE NBER ONCA LFALLS2 REC LOCKE NOR TRNT OFRG WTR SKAN ONY UTC PAL WELL ESF SHER UTC7 STILL SYR UN WEST WILL TULLY BERG
To assess which physical factors best predict the spatial patterns of snowfall from the different storm types seen in Figures 3.10 and 3.11, regression models were built in R using a suite of locational and geographic factors developed from the literature (Saslo and
Greybush 2017; White et al. 2010; Ware et al. 2006; Niziol et al. 1995; Hartnett et al. 2014).
The variables used in the models included a location’s subregion as a random effect, and its elevation, latitude, longitude, distance from a fixed point over Lake Ontario, and a set of 2.5
75 km, 5 km, and 10 km exposure variables as fixed effects. For justification and descriptions of each variable see Section 2.4. Variables were tested for collinearity, and although longitude and distance from Lake Ontario were correlated (r = 0.58), it was below the 0.6 threshold, so both were used in the models (Yoo et al. 2014). Normality tests indicated that linear mixed-effect models could be applied and compared using the AIC and ANCOVA. The procedures used to compare, rank, and evaluate models are described in Section 2.4.
3.3 Results and Analyses
3.3.1 Magnitude and Frequency of Storm Types at the Regional Scale
Regional analyses presented in this section examine snowfall trends for all of Central New
York. The total amount of seasonal snowfall associated with a storm type depends on both the frequency of the storm and the magnitude of the storms that occur. To assess the trend in the seasonal contribution of different storm magnitudes and their frequencies, seasonal data for Central New York were plotted and are shown on Figure 3.2. Since normality and homoscedasticity were not violated, two-sample mean difference tests were used to differentiate between classes of storm magnitude in terms of their average contributions to seasonal snowfall and their frequency. Results are show in Table 3.3.
The results show that although moderate snowfall storms occur more frequently ( < 0.01) than both light and heavy snowfall storms (Table 3.3), heavy snowfall storms contribute by far the most ( < 0.01) to seasonal snowfall (807 cm) (Figure 3.2; Table 3.3).
76
Light Moderate Heavy Light Moderate Heavy 40 2000
1800 35
1600 30 1400
25 1200
20 1000
800
15 Total Total Snowfall (cm)
600 10
400 Snowstorm Snowstorm Frequency (Number of Storms)
5 200
0 0
Figure 3.2. Seasonal snowstorm frequency and seasonal snowfall totals for light, moderate and heavy snowstorms affecting Central New York from 1985/86 – 2014/15. Line graphs represent snowstorm frequency, while stacked area charts represent total snowfall.
Table 3.3. The average frequency and snowfall for different magnitude snowstorms to influence Central New York from 1985/86 – 2014/15. Significant ( < 0.01) differences are bold and italicized. Magnitude Storms season-1 Avg. Total Snowfall Light 22.2 0.9 126 5.2 cm Moderate 28.1 1.1 463 17.4 cm Heavy 18.0 0.9 807 41.6 cm n = 2055
In a similar manner, the seasonal frequencies and magnitudes of lake snowstorms and non- lake snowstorms were plotted in Figure 3.3 and tested for differences (Table 3.4). The
77 results indicate that although there is a significant ( < 0.01) difference in their frequencies, there is no statistical difference ( = 0.09) in the average seasonal snowfall from these storms. This suggests that lake snowstorms on average produce more snowfall per storm than non-lake snowstorms. Even though lake snowstorms occur less frequently than non- lake snowstorms, they contribute 659.6 cm of the 1,396 cm of seasonal snowfall (47.3
3.1%, < 0.05) in Central New York. This corroborates previous estimates within the Great
Lakes region, and highlights its importance on seasonal snowfall totals in the region (Miner and Fritsch 1997; Liu and Moore 2004; Veals and Steenburgh 2015).
60 1200 Lake Snow Non-Lake Snow Lake Snow Non-Lake Snow 50 1000
40 800
30 600
20 400 Total Snowfall (cm) Storm Storm Frequency (Number of Storms) 10 200
0 0
Figure 3.3. Seasonal frequency and seasonal snowfall totals (cm) for lake-snowstorms (LS) and non-lake snowstorms (NLS) in Central New York from 1985/86 – 2014/15. The line graphs represent storm frequency, while the bar graphs represent total snowfall. frequency of lake snowstorms (27.1 1.0 storms season-1) versus non-lake snowstorms.
78
Table 3.4. The average frequency of occurrence, and snowfall produced by light, moderate, and heavy lake snowstorms and non-lake snowstorms from 1985/86 – 2014/15. Significant differences are bold if significant at = 0.05, and bold and italicized if significant at = 0.01. Storm Magnitude Lake Snow Non-Lake Snow
Light 7.1 ± 0.5 15.1 ± 0.7
Moderate 10.5 ± 0.6 17.8 ± 0.9
season) Heavy 9.5 ± 0.6 8.5 ± 0.6 per per
Frequency (storms (storms Frequency Total 27.1 ± 1.0 41.4 ± 1.3
Light 39.9 ± 2.7 86.4 ± 4.3
Moderate 177.7 ± 10.2 285.5 ± 14.9
Heavy 442.0 ± 26.4 364.6 ± 29.5 Snowfall (cm) Snowfall Total 659.6 ± 29.6 736.4 ± 33.6
These results are helpful for understanding the role of the Great Lakes in producing snow; however, to understand how frequency and magnitude of snowstorms may change in the future, it is important to examine the type of storm based on its area of formation. The ten contributing snowstorm types were initially examined (Table 3.1), however similarities between some of the storms allowed for the use of the five general snowstorm types. This provided an adequate dataset to evaluate relationships between storm type and their effect on seasonal snowfall totals. Comparisons across storms were conducted using ANOVA and two-sample means difference tests. Results are presented in Table 3.5, and season-to- season variation for each storm are plotted on Figures 3.4 and 3.5.
Lake-effect snowstorms are both significantly (ρ < 0.01) more frequent (24.0 1.0 storms season-1) and contribute (ρ < 0.01) more seasonal snowfall (583.5 27.8 cm) than other
79 snowstorm types. This pattern generally holds for each winter season over the study period (Figures 3.4 and 3.5). The average seasonal snowfall from lake-effect snowstorms is more than double that of any other single contributor (non-cyclonic storms contribute
225.5 18.0 cm season-1), a total that accounts for 41.8% of the region’s snowfall. In comparison to other parts of the Great Lakes region, the contribution of lake-effect snowfall to seasonal snowfall totals in Central New York is slightly lower than those documented by
Veals and Steenburgh (2015), Miner and Fritsch (1997), and Liu and Moore (2004).
Table 3.5. The average seasonal frequency and seasonal snowfall of different snowstorm types in Central New York. The -values of two-sample mean difference tests are included; significant differences are bold if = 0.05. Snowstorm Frequency (storms per season) Canadian Low LES Non-Cyclonic Nor'easter Rocky Low Average 12.2 ± 0.8 24.0 ± 1.0 13.7 ± 0.8 7.7 ± 0.5 10.7 ± 0.5 Canadian Low N/A < 0.01 0.21 < 0.01 0.13 LES < 0.01 N/A < 0.01 < 0.01 < 0.01 Non-Cyclonic 0.21 < 0.01 N/A < 0.01 < 0.01 Nor'easter < 0.01 < 0.01 < 0.01 N/A < 0.01 Rocky Low 0.13 < 0.01 < 0.01 < 0.01 N/A df = 58
Snowstorm Snowfall Totals (cm) Canadian Low LES Non-Cyclonic Nor'easter Rocky Low Average 198.5 ± 17.6 583.5 ± 27.8 225.5 ± 18.0 206.1 ± 19.1 181.4 ± 10.8 Canadian Low N/A < 0.01 0.29 0.77 0.41 LES < 0.01 N/A < 0.01 < 0.01 < 0.01 Non-Cyclonic 0.29 < 0.01 N/A 0.46 0.04 Nor'easter 0.77 < 0.01 0.46 N/A 0.27 Rocky Low 0.41 < 0.01 0.04 0.27 N/A df = 58
80
90 Canadian Lows Rocky Lows Nor'easters LES Non-cyclonic 80
70
60
50
40
30
20 Storm Storm (storms/season) Frequency 10
0
2008/09 1985/86 1986/87 1987/88 1988/89 1989/90 1990/91 1991/92 1992/93 1993/94 1994/95 1995/96 1996/97 1997/98 1998/99 1999/00 2000/01 2001/02 2002/03 2003/04 2004/05 2005/06 2006/07 2007/08 2009/10 2010/11 2011/12 2012/13 2013/14 2014/15 Season
Figure 3.4. Seasonal frequency (storms season-1) of the five snowstorm types identified to influence Central New York from 1985/86 – 2014/15.
2000 Canadian Lows Rocky Lows 1800 Nor'easters LES Non-cyclonic 1600
1400
1200
1000
800
600 Sesonal Snowfall (cm) Snowfall Sesonal 400
200
0
1992/93 2007/08 1985/86 1986/87 1987/88 1988/89 1989/90 1990/91 1991/92 1993/94 1994/95 1995/96 1996/97 1997/98 1998/99 1999/00 2000/01 2001/02 2002/03 2003/04 2004/05 2005/06 2006/07 2008/09 2009/10 2010/11 2011/12 2012/13 2013/14 2014/15 Season
Figure 3.5. Seasonal snowfall totals (cm) from the five snowstorm types identified to influence Central New York from 1985/86 – 2014/15. 81
Nor’easters were the least frequent storm (7.7 storms season-1), but Rocky lows contributed the lowest average snowfall (181.4 10.8 cm season-1). The importance in understanding the relative contribution of a storm to seasonal snowfall totals is highlighted when comparing the frequency of a storm to its seasonal snowfall. For instance, frequent storms may add little to seasonal snowfall totals if mostly light snowfall occurs. Nor’easter for example, are often considered the dominant snowfall contributor to the Northeast (e.g.
Kocin and Uccellini 2004b; Zielinski 2002); however, they are relatively rare events contributing significantly more snow per storm than other storm types.
Extending the comparisons between storm type magnitude, Figure 3.6 illustrates the frequency of different storm types and their magnitudes, and Figure 3.7 shows the percentage of the seasonal snowfall from different storm types across magnitude classes.
ANOVA and two sample difference tests were used to compare differences across storms and categories, and results are shown in Table 3.6. Heavy (8.5 0.5 storms season-1) and moderate (9.1 0.7 storms season-1) lake-effect snowstorms are significantly ( < 0.01) more frequent than any other storm type at those magnitude classes. In fact, a snowstorm with a large snowfall total was almost twice as likely to be a lake-effect snowstorm than other storm types. Nor’easters were the second most frequent (3.0 0.3 storms season-1) heavy snowstorm, which in combination with lake-effect snowstorms accounted for 16.7% of all snowstorms and over 60% of all heavy snowstorms. Moderate (2.8 0.3 storms season-1) and light (2.0 0.2 storms season-1) Nor’easters were less frequent than all other similar magnitude storm types, whereas light lake-effect snowstorms (6.5 0.5 storms season-1) and non-cyclonic storms (5.6 0.5 storms season-1) were the most frequent.
82
Canadian Lows Lake-Effect Non-Cyclonic Nor'easters Rocky Lows
45.0
40.0 6.5 35.0 4.0 30.0 5.6 8.6 25.0 3.5 2.9
4.3 20.0 8.2 3.3 15.0 13.2
10.0 9.5
12.4 Seasonal Seasonal Storm Frequency (%)
5.0 8.9 6.1 2.9 0.0 Heavy Moderate Light Storm Magnitude
Canadian Lows Lake-Effect Non-Cyclonic Nor'easters Rocky Lows
100.0 13.1 90.0 15.8 17.4
80.0 16.5 9.8 9.0 70.0 12.6 20.7 60.0 25.3
50.0
40.0 32.0 47.0 29.3 30.0
20.0 Seasonal Seasonal Storm Frequency (%)
10.0 21.6 18.8 10.9 0.0 Heavy Moderate Light Storm Magnitude
Figure 3.6. The average percent frequency for each snowstorm type based on the storm magnitude. The top figure represents the percent frequency compared to the total amount of snowstorms (2055) from 1985/86 – 2014/15. The bottom figure represents the percent frequency of heavy, moderate, and light snowstorms relative to the number of heavy (541), moderate (849), and light (665) snowstorms. 83
Canadian Lows Lake-Effect Non-Cyclonic Nor'easters Rocky Lows 60.0
6.4 50.0 10.7
40.0 7.1
30.0 5.1 3.3
6.8 20.0 28.2
10.9 1.5 10.0
Seasonal Seasonal Snowfall Contribution (%) 0.8 2.3 5.4 7.0 2.6 0.0 1.8 Heavy Moderate Light Storm Magnitude
Canadian Lows Lake-Effect Non-Cyclonic Nor'easters Rocky Lows 100.0 11.1 90.0 15.4 16.4
80.0 18.5 9.9 9.0
70.0 12.2 20.6 60.0 25.1
50.0
40.0 33.0 48.9 29.0 30.0
20.0 Seasonal Seasonal Snowfall Contribution (%) 10.0 21.0 20.3 9.4 0.0 Heavy Moderate Light Storm Magnitude Figure 3.7. The average percent snowfall contribution for each snowstorm type based on the storm magnitude. The top figure represents the percent snowfall contribution compared to the total amount of snowfall (41,880.1 cm) from 1985/86 – 2014/15. The bottom figure represents the percent contribution relative to the amount of snowfall from heavy (24,197 cm), moderate (13,895 cm), and light snowstorms (3,788 cm).
84
Table 3.6. Average seasonal frequency of occurrence, and average seasonal snowfall produced by the five snowstorm types identified to influence Central New York from 1985/86 – 2014/15. Bold and italicized values are significantly ( < 0.01) different than other storms. Average Frequency (storms per season) Heavy Moderate Light Canadian Lows 2.0 ± 0.3 6.1 ± 0.5 4.2 ± 0.5 Lake-Effect 8.5 ± 0.5 9.1 ± 0.7 6.5 ± 0.5 Non-Cyclonic 2.3 ± 0.3 5.9 ± 0.6 5.6 ± 0.5 Nor'easters 3.0 ± 0.3 2.8 ± 0.3 2.0 ± 0.2 Rocky Lows 2.4 ± 0.2 4.5 ± 0.4 3.9 ± 0.4
Average Seasonal Snowfall (cm) Heavy Moderate Light Canadian Lows 75.6 ± 12.6 97.3 ± 5.5 25.6 ± 3.1 Lake-Effect 394.2 ± 24.5 152.8 ± 11.9 36.6 ± 2.7 Non-Cyclonic 98.7 ± 13.4 95.2 ± 9.1 31.7 ± 2.7 Nor'easters 149.0 ± 18.1 45.7 ± 5.0 11.4 ± 1.3 Rocky Lows 89.2 ± 9.9 71.4 ± 6.1 20.7 ± 2.3
Based on contributions to the average seasonal snowfall, only two significant differences emerged. First, heavy lake-effect snowstorms (394.2 24.5 cm) contribute more than twice the seasonal snowfall than any other heavy storm (Table 3.6). The second largest contributor was Nor’easters, which together with heavy lake-effect snowstorms produced almost 39% of seasonal snowfall. The second significant difference to emerge is that moderate and light magnitude Nor’easters (45.7 5.0 cm and 11.4 1.3 cm, respectively) contribute the least amount of seasonal snow. This suggests that although Nor’easters occur less frequently than other storms, when they do occur, they are associated with heavy snowfall.
85
3.3.2 Magnitude and Frequency of Storm Types at the Subregional Scale
The influence of the spatial scale used to observe seasonal snowfall from the seven generalized snowstorm types was examined using the snowfall subregions of Central New
York (Hartnett et al. 2014). To assess the average seasonal snowfall (cm) and average seasonal snowfall percentage from lake snowstorms and non-lake snowstorms, and the five general snowstorm types, seasonal data for the five subregions were plotted in Figures 3.8 and 3.9, respectively. After determining the data were normally distributed with equal variance, an ANOVA and two-sample mean difference tests were used to test whether snowfall contributions within a region are significantly different between the different storm types. Results are shown in Table 3.7.
Non-lake snowstorms produced significantly ( < 0.01) more snow than lake snowstorms for Regions 1-4, and they constituted a significantly ( ≤ 0.05) higher percentage of the seasonal snowfall for all five subregions. This contrasts with the findings at the regional scale, and those from previous studies, which suggest that lake-effect snow accounts for approximately half of the seasonal snowfall within the Great Lakes region (Eichenlaub
1970; Miner and Fritsch 1997; Liu and Moore 2004; Veals and Steenburgh 2015). Thus, the importance of scale is highlighted by these results, and the spatial scale used may vary results greatly.
86
Figure 3.8. Average seasonal snowfall totals (cm) per snowstorm type for the five snowfall regions of Central New York from 1985/86 – 2014/15. Figures are divided by region: Region 1 (top left), Region 2 (top right), Region 3 (middle left), Region 4 (middle right), Region 5 (bottom).
87
Figure 3.9. Average seasonal snowfall contributions (%) per snowstorm type for the five snowfall regions of Central New York from 1985/86 – 2014/15. Figures are divided by region: Region 1 (top left), Region 2 (top right), Region 3 (middle left), Region 4 (middle right), Region 5 (bottom)
88
Table 3.7. Average percent contributions and average seasonal snowfall (cm) from each storm type for each subregion. Significant differences between lake snowstorms and non- lake snowstorms or the five general snowstorm types are bold if ≤ 0.05, and bold and italicized if ≤ 0.01.
Storm Type Region 1 Region 2 Region 3 Region 4 Region 5
Lake Snow 38.2 3.9 29.6 3.9 37.8 4.3 45.0 3.8 46.5 4.1
Non-Lake Snow 61.8 3.9 70.4 3.9 62.2 4.3 55.0 3.8 53.5 4.1 Canadian Lows 11.4 2.3 13.8 3.2 14.9 3.9 13.8 2.6 12.4 2.6 Lake-Effect Snow 33.5 3.6 23.1 3.8 33.9 5.8 39.0 3.7 41.6 3.7 Noncyclonic Storms 13.1 4.7 13.9 4.2 12.1 3.3 16.1 3.6 16.8 3.8
Percentage Nor'easters 26.7 4.4 28.9 4.8 22.8 5.2 16.6 3.2 15.6 2.9 Rocky Lows 15.3 2.8 20.3 3.7 16.3 4.4 14.5 2.4 13.7 2.7 Lake Snow 90.9 10.5 56.0 7.5 87.1 12.5 178.2 19.4 172.8 23.6
Non-Lake Snow 152.0 19.8 138.8 18.0 145.3 18.6 217.9 19.5 194.1 20.5 Canadian Lows 26.5 5.3 26.6 4.5 34.2 7.1 54.8 8.8 45.4 8.4 Lake-Effect Snow 75.8 9.4 45.4 7.0 75.5 12.3 154.3 17.3 151.5 21.3
Noncyclonic Storms 30.1 6.8 28.6 6.2 28.3 5.8 63.7 10.2 60.8 11.0 Snowfall Nor'easters 66.0 14.6 63.4 13.8 56.0 15.1 67.0 13.1 56.3 10.9 Rocky Lows 33.3 4.8 39.1 6.5 36.7 8.4 56.4 8.1 46.4 6.2
Statistical differences in the seasonal snowfall totals and the percentage of seasonal
snowfall totals were then tested for the same class of snowstorms across the five
subregions. All data were normally distributed, but for variables with equal variance,
ANOVAs were used to determine if the average contribution for at least one region was
statistically different from the rest, while a Kruskal-Wallis test was used in cases where
homoscedasticity was violated (Table 3.8). If the mean seasonal snowfall (cm) or the mean
seasonal percent snowfall for at least one region was statistically different than the others,
then Student t-tests for parametric data and Mann-Whitney tests for nonparametric data
were applied. The results are presented in Table 3.9.
89
Table 3.8. Significance of Bartlett Tests of Variances and Analysis of Variances (ANOVAs)/Kruskal-Wallis Tests between the five subregions for each snowstorm type. Statistically significant values (ρ < 0.05) are bold and italicized. Analyses conducted using the Kruskal -Wallis test of variance are marked with an ‘*’. Bartlett Test of Variances ANOVA/Kruskal-Wallis Storm Type Contributions Snowfall Contributions Snowfall Lake Snow 0.97 < 0.01 0.00 0.00* Non-Lake Snow N/A 0.92 0.00 0.00 Canadian Lows 0.46 < 0.01 0.38 0.00* Lake-Effect Snow 0.89 < 0.01 0.00 0.00* Noncyclonic Storms 0.52 0.13 0.14 0.00 Nor'easters 0.01 0.64 0.00* 0.68 Rocky Lows 0.01 0.22 0.03* 0.00
Table 3.9. ρ-values of Student t-tests and Mann-Whitney tests comparing the average percent contribution or average snowfall contribution of different snowstorms across regions. Significant (ρ < 0.05) values are bold and italicized. Percent Contributions Snowfall Contributions Lake Snow R1 R2 R3 R4 R5 R1 R2 R3 R4 R5 Region 1 N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A Region 2 0.00 N/A N/A N/A N/A 0.00 N/A N/A N/A N/A Region 3 0.89 0.01 N/A N/A N/A 0.65 0.00 N/A N/A N/A Region 4 0.02 0.00 0.02 N/A N/A 0.00 0.00 0.00 N/A N/A Region 5 0.01 0.00 0.01 0.60 N/A 0.00 0.00 0.00 0.73 N/A Non-Lake Snow R1 R2 R3 R4 R5 R1 R2 R3 R4 R5 Region 1 N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A Region 2 0.00 N/A N/A N/A N/A 0.34 N/A N/A N/A N/A Region 3 0.89 0.01 N/A N/A N/A 0.63 0.62 N/A N/A N/A Region 4 0.02 0.00 0.02 N/A N/A 0.00 0.00 0.00 N/A N/A Region 5 0.01 0.00 0.01 0.60 N/A 0.01 0.00 0.00 0.10 N/A Canadian Lows R1 R2 R3 R4 R5 R1 R2 R3 R4 R5 Region 1 N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A Region 2 N/A N/A N/A N/A N/A 0.98 N/A N/A N/A N/A Region 3 N/A N/A N/A N/A N/A 0.09 0.08 N/A N/A N/A Region 4 N/A N/A N/A N/A N/A 0.00 0.00 0.00 N/A N/A Region 5 N/A N/A N/A N/A N/A 0.00 0.00 0.05 0.14 N/A Lake Effect Snow R1 R2 R3 R4 R5 R1 R2 R3 R4 R5 Region 1 N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A Region 2 0.00 N/A N/A N/A N/A 0.00 N/A N/A N/A N/A Region 3 0.91 0.00 N/A N/A N/A 0.97 0.00 N/A N/A N/A Region 4 0.04 0.00 0.15 N/A N/A 0.00 0.00 0.00 N/A N/A Region 5 0.00 0.00 0.03 0.35 N/A 0.00 0.00 0.00 0.84 N/A TABLE CONTINUED ON NEXT PAGE
90
TABLE 3.9 CONTINUED Noncyclonic Storms R1 R2 R3 R4 R5 R1 R2 R3 R4 R5 Region 1 N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A Region 2 N/A N/A N/A N/A N/A 0.74 N/A N/A N/A N/A Region 3 N/A N/A N/A N/A N/A 0.69 0.95 N/A N/A N/A Region 4 N/A N/A N/A N/A N/A 0.00 0.00 0.00 N/A N/A Region 5 N/A N/A N/A N/A N/A 0.00 0.00 0.00 0.70 N/A Nor'easters R1 R2 R3 R4 R5 R1 R2 R3 R4 R5 Region 1 N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A Region 2 0.52 N/A N/A N/A N/A N/A N/A N/A N/A N/A Region 3 0.26 0.10 N/A N/A N/A N/A N/A N/A N/A N/A Region 4 0.00 0.00 0.05 N/A N/A N/A N/A N/A N/A N/A Region 5 0.00 0.00 0.02 0.67 N/A N/A N/A N/A N/A N/A Rocky Lows R1 R2 R3 R4 R5 R1 R2 R3 R4 R5 Region 1 N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A Region 2 0.02 N/A N/A N/A N/A 0.16 N/A N/A N/A N/A Region 3 0.67 0.12 N/A N/A N/A 0.49 0.66 N/A N/A N/A Region 4 0.91 0.01 0.60 N/A N/A 0.00 0.00 0.00 N/A N/A Region 5 0.60 0.00 0.42 0.66 N/A 0.00 0.12 0.07 0.06 N/A
Lake snowstorms produced the most snowfall per season and contributed the most to seasonal snowfall totals in Regions 4 and 5 (Figure 3.8; Table 3.7). These regions are most linked to greater seasonal snowfall totals largely due to the occurrence of lake snowstorms
(e.g. Campbell et al. 2016; Veals and Steenburgh 2015; Minder et al. 2015; Veals et al.
2018). These storms also produce significantly (ρ < 0.01) lower seasonal snowfall totals
(56.0 ± 7.5 cm) and percentages of seasonal snowfall (29.6 ± 3.9%) in Region 2 compared to Regions 1 (90.9 ± 10.5 cm; 38.2 ± 3.9%) and 3 (87.1 ± 12.5 cm; 37.8 ± 4.3%).
Average seasonal snowfall totals from non-lake snowstorms in Regions 4 (217.9 ± 19.5 cm) and 5 (194.1 ± 20.5 cm) were also significantly (ρ < 0.05) higher than those in Regions 1-3.
However, the range in the average seasonal snowfall between the five regions was smaller for non-lake snowstorms than lake snowstorms. This suggests that seasonal snowfall
91 totals are more homogeneous after non-lake snowstorms than lake-snowstorms. This is further supported by results shown in Table 3.7, indicating that average seasonal snowfall totals from non-lake snowstorms are similar (ρ > 0.05) across Regions 1-3. Even though seasonal snowfall totals are similar across the three regions, there is a significant difference
( 0.01) in the percent contribution from non-lake snowstorms in Region 2 compared to
Regions 1 and 3 (Table 3.9). Thus, understanding the percent contribution of different snowstorm types to seasonal snowfall is necessary because although a storm may average a lot of snowfall per season, if it is an anomalously snowy season, its relative influence is less.
Subregional differences between the five general snowstorm types suggest that lake-effect snowstorms produce significantly ( < 0.01) more snowfall and contribute a significantly
( < 0.01) higher percentage to seasonal snowfall in Regions 3-5, than the four other general snowstorm types (Figures 3.8 and 3.9; Table 3.7; Table 3.9). Although lake-effect snowstorms are also the largest snowfall producer in Region 1, average seasonal snowfall totals were not significantly (ρ = 0.12) higher than that from Nor’easters. Seasonal snowfall totals from these storms are significantly (ρ < 0.01) greater in Regions 4 (154.3 ±
17.3 cm) and 5 (151.5 ± 21.3 cm) than in Regions 1-3. Similarly, the percentage of seasonal snowfall from lake-effect snowstorms was significantly greater in Regions 4 and 5 than in
Regions 1 and 2, but not significantly (ρ = 0.15) different between Regions 3 (33.9 ± 5.8%) and 4 (39.0 ± 3.7%). This suggests that even though lake-effect snowstorms produce more seasonal snowfall in Region 4, they are as influential on seasonal snowfall totals in Region 3 as in Region 4. Overall, lake-effect snowstorms contribute less than 42% of the seasonal
92 snowfall for each region, and although lower than expected, they still have a strong influence on the winter climates of the subregions (Peace and Sykes 1966; Norton and
Bolsenga 1993; Clark et al. 2016; Veals and Steenburgh 2015).
Seasonal snowfall totals were most interesting in Region 2, as Nor’easters produced the most (ρ 0.01) snowfall per season (63.4 ± 13.8 cm), followed by lake-effect snowstorms and Rocky lows. Although Nor’easters also contributed the highest percentage to seasonal snowfall totals in Region 2, it was not significantly (ρ = 0.06) greater than the percentage from lake-effect snowstorms. Overall, Nor’easters contribute significantly (ρ ≤ 0.05) more to seasonal snowfall totals in Regions 1-3 compared to the two northern regions.
Interestingly, Nor’easters were the only snowstorm type failing to exhibit a significant (ρ <
0.05) difference in seasonal snowfall contributions (cm) between the five regions, suggesting that they produce a relatively uniform snowfall across the entire study area.
Seasonal snowfall totals from Canadian lows and noncyclonic snowstorms were significantly (ρ < 0.05) greater in Regions 4 and 5 than Regions 1 and 2. However, the percentages of seasonal snowfall from these storms were not significantly (ρ > 0.05) different for any of the five regions. This suggests that even though these storms average more seasonal snowfall in Regions 4 and 5 than in Regions 1-3, they do not account for a greater percentage of the seasonal snowfall. Rocky lows averaged a significantly (ρ < 0.05) higher contribution to seasonal snowfall totals in Region 2 (20.3 ± 3.7%) than in Regions 1,
4 and 5. This suggests that even though Rocky lows produce a similar amount of snowfall
93 for Region 2 as the other four regions, they have a disproportionate effect on seasonal snowfall totals.
3.3.3 Magnitude and Frequency of Storm Types at the Local Scale
Using the 26 filtered COOP stations, the percentages of the seasonal snowfall contributed from the five general snowstorm types and lake snowstorms and non-lake snowstorms were calculated from 1985/86 – 2014/15. Percent contributions were then plotted in
ArcGIS and interpolated to produce gridded surfaces representing the seasonal percentage of the different snowstorm types throughout Central New York. The results are presented in Figures 3.10 and 3.11.
Although previous results suggest that lake snowstorms account for approximately 47% of the average seasonal snowfall in Central New York, local results suggest that this percentage is not consistent throughout the study area (Figure 3.10). Depending on the location, contributions ranged from 25-50% of the seasonal snowfall total. This is notable variation in a relatively small area, where the highest contributions were observed over and near the Tug Hill. The majority (62% area) of Central New York received at least 35% of its seasonal snowfall totals from lake snowstorms including areas downwind of Lake
Ontario, the western Adirondack Mountains, and the Southern Hills. Snowfall contributions were the least in areas further from, and sub-parallel to the axis of Lake
Ontario.
94
Figure 3.10. The percent contribution of lake snowstorms (left) and non-lake snowstorms (right) to seasonal snowfall totals in Central New York from 1985/86 – 2014/15.
To examine the spatial variation of seasonal snowfall contributions from different storm types, snowfall contributions were mapped for the five generalized snowstorms (Figure
3.11). Although lake-effect snowstorms are the dominant snowfall producer in Central
New York, the maps show that across the region this is not the case (Figure 3.11a). In parts of the Southern Tier, lake-effect snow contributed only 20-25% of the seasonal total, whereas for the Tug Hill the contribution was approximately 50%.
95
96
FIGURE 3.11 CONTINUED
Figure 3.11. The percent contribution of seasonal snowfall totals associated with different snowstorm types to affect Central New York from 1985/86 – 2014/15. Figure A is the average seasonal snowfall contribution for Canadian Lows, Figure B is from Lake-Effect Snowstorms, Figure C is from Non-Cyclonic Snowstorms, Figure D is from Nor’easters, and Figure E is from Rocky Lows.
In southern and northwestern Central New York, Nor’easters are the dominant snowfall contributor (Figure 3.11b). Figure 3.11 highlights the fact that these maps represent the percentage of the seasonal snowfall contributed by a particular storm type. That means that if a storm type dominates the snowfall in an area, then its relative percentage is higher, even if its snowfall contribution is similar to that of other areas across the region. This is seen in Figure 3.11b, as Nor’easters only contribute between 5-10% of the seasonal total snow accumulation in the Tug Hill. Although Nor’easters contribute a relatively low percentage to seasonal snowfall totals in the Tug Hill, the Tug Hill averages approximately
200 more centimeters of snowfall than other areas in Central New York. Conversely, the
97 large contribution to snowfall in southern and southeastern Central New York is in large part due to their lower seasonal snowfall totals. However, the use of percentages allows for comparisons between different areas of Central New York, regardless of their seasonal snowfall totals.
In this analysis, Canadian lows were associated with the third highest seasonal snowfall totals in Central New York, contributing 198 17.6 cm compared to the lake-effect contribution of 583.5 27.8 cm (Table 3.1). However, their relative contribution to snowfall across the study region is quite homogeneous; the northern half receives about
15% of its seasonal snow from Canadian lows, and the southern half about 5-10%. Like
Canadian lows, non-cyclonic snowstorms have a regional-scale effect, contributing a consistent snowfall percent across the study area ranging between 12.5% in the south to
20% in the north. Although the greater contributions northwards are possibly linked to the storms’ abilities to produce lake-enhanced snowfall, it is more plausible that they are a function of the latitudinal and elevational effects on temperature. The direct effects of latitude and elevation are further investigated in the next section using mixed-effects regression models.
The spatial distribution of the percent contribution to snowfall from Rocky lows is almost the inverse of that of lake-effect snow (Figure 3.11e). The eastern shore of Lake Ontario and the Tug Hill show the lowest contributions from Rocky lows (5-15%), whereas the southeastern hills show the greatest contribution (~20%).
98
3.3.4 Modeling the Effects of the Physical Characteristics of a Location on Snowfall
Mixed-effects linear regression models were used to evaluate which of the seven locational and geographic factors covaried with seasonal snowfall totals from different storm types.
Comparisons between models for each snowstorm type were made using ANCOVA and the
AIC (see Chapter 2 for details). The weights were the determining factor for selecting the top models, and the strength of the models were tested using the marginal R2 and conditional R2 to determine which geographical factors helped explain the distribution best
(Table 3.10).
The locational and geographical influences on the spatial distribution of snowfall from any storms are supported by the high conditional R2 values (Table 3.10). In all cases, expect for
Hudson lows, clippers, and Texas hooks, the models explained at least 80% of the variance.
Two thirds of the top models (with a weight greater than 0.25) were explained either by the combination of latitude, longitude, elevation and distance from the lake, or the combination of latitude, longitude and 5 km exposure. Storms associated with lake-effect and lake-enhanced snowfall were best explained by the former combination.
The snowfall distribution from cyclonic systems were more influenced by the latitude, longitude, and exposure. This supports the claim from Basist et al. (1994) that the exposure of a location to prevailing winds is the dominant factor influencing the spatial distribution of precipitation in mountainous terrain. The results also support the conclusions of Perry et al. (2007) and Perry and Konrad (2006) in their analysis of the southern Appalachians, that the exposure of a location exerts a strong control on its
99
Table 3.10. The influence of a location’s environmental conditions on the snowfall contribution (cm) of different snowstorms in Central New York. The top four models using AICc weights are reported for lake snowstorms (LS), non-lake snowstorms (NLS), lake- effect snowstorms (LES), Nor’easters (NOR), Canadian lows (CND), clippers (CLIP), Hudson lows (HL), Great Lakes lows (GLL), Rocky lows (ROCK), Colorado lows (COL), Oklahoma hooks (OKH), Texas hooks (TXH), Non-cyclonic snowstorms (NCYC), frontal snowstorms (FRT), upper disturbances (UP), and total snowfall (TTL). The explained variance of each model is also reported using the marginal R2 and conditional R2. Storm Variables df AICc ⧍ AICc Weight Marg. Cond. R2 R2 ABCD 8 201.0 0.00 0.496 0.886 0.886 ABF 7 202.1 1.12 0.284 0.638 0.933 LS ABG 7 205.7 4.70 0.047 0.620 0.894 BCD 7 206.1 5.06 0.040 0.876 0.876 ABCD 8 195.6 0.00 0.795 0.886 0.886 ABF 7 200.8 5.19 0.059 0.638 0.933 NLS ABCDF 9 201.8 6.24 0.035 0.879 0.879 ACD 7 202.0 6.37 0.033 0.876 0.876 ABF 7 197.2 0.00 0.409 0.613 0.941 ABCD 8 197.6 0.40 0.335 0.885 0.885 LES ABG 7 201.0 3.79 0.061 0.605 0.905 AF 6 201.3 4.05 0.054 0.671 0.928 ABF 7 163.0 0.00 0.313 0.505 0.850 ABD 7 163.7 0.69 0.221 0.278 0.967 NOR E 6 164.7 1.62 0.140 0.254 0.911 BF 6 164.9 1.83 0.126 0.527 0.837 ABCD 8 164.1 0.00 0.293 0.864 0.864 ABF 7 164.5 0.32 0.250 0.753 0.821 CND ACD 7 164.8 0.70 0.207 0.869 0.869 AF 6 167.3 3.12 0.062 0.740 0.740 ABF 7 147.8 0.00 0.268 0.708 0.773 ACD 7 149.0 1.17 0.150 0.835 0.835 CLIP ABCD 8 149.2 1.31 0.140 0.830 0.830 AF 6 149.5 1.64 0.118 0.689 0.689 A 5 91.1 0.00 0.430 0.639 0.639 E 6 91.6 0.52 0.332 0.669 0.708 HL B 5 95.3 4.28 0.051 0.183 0.765 ABF 7 95.9 4.88 0.038 0.748 0.863 AF 6 121.2 0.00 0.285 0.801 0.801 ABF 7 122.1 0.80 0.191 0.797 0.856 GLL ACD 7 122.7 1.49 0.136 0.892 0.892 AG 6 123.0 1.71 0.121 0.777 0.816 TABLE CONTNUED ON NEXT PAGE
100
TABLE 3.10 CONTINUED ABCD 8 158.9 0.00 0.270 0.806 0.860 ABF 7 159.1 0.26 0.238 0.646 0.848 ROCK E 6 160.4 1.51 0.127 0.331 0.890 BF 7 161.6 2.75 0.068 0.533 0.852 E 6 132.7 0.00 0.419 0.347 0.905 B 5 134.5 1.78 0.173 0.333 0.912 COL ABF 7 136.1 3.39 0.077 0.577 0.870 BF 6 136.3 3.67 0.067 0.561 0.875 BF 6 117.4 0.00 0.183 0.596 0.847 ABF 7 177.8 0.42 0.149 0.630 0.875 OKH BG 6 118.1 0.71 0.129 0.411 0.932 E 6 118.3 0.94 0.115 0.239 0.930 ABG 7 109.7 0.00 0.237 0.710 0.733 E 6 110.7 1.00 0.143 0.452 0.479 TXH ABF 7 110.7 1.02 0.142 0.694 0.694 AG 6 111.3 1.60 0.107 0.629 0.732 ABCD 8 159.8 0.00 0.371 0.914 0.914 BCD 7 159.9 0.15 0.345 0.918 0.918 NCYC ABF 7 161.6 1.87 0.146 0.769 0.924 AF 6 164.5 4.69 0.036 0.777 0.903 ABF 7 126.8 0.00 0.447 0.803 0.925 ACD 7 129.5 2.71 0.115 0.906 0.906 FRT AF 6 129.7 2.88 0.106 0.769 0.912 ABG 7 130.6 3.76 0.068 0.764 0.892 ABF 7 149.5 0.00 0.225 0.709 0.917 ACD 7 150.1 0.64 0.164 0.887 0.887 UP ABCD 8 150.3 0.84 0.148 0.882 0.882 AF 6 150.6 1.12 0.128 0.733 0.897 ABCD 8 216.3 0.00 0.805 0.905 0.905 ABCDF 9 222.2 5.86 0.043 0.902 0.904 TTL ABF 7 222.2 5.86 0.043 0.703 0.928 ABCDE 9 222.9 6.58 0.030 0.902 0.902 A - Latitude, B - Longitude, C - Elevation, D - Lake Distance, E - 2.5 km Exposure, F - 5 km Exposure, G - 10 km Exposure
average seasonal snowfall total. In the case of Central New York, the elevation of the surrounding 2.5 km had less of an influence than terrain 5 km from a station, indicated by the top models. The greater influence of the 5 km exposure needs further examination, but as Campbell et al. (2016) suggest, it may be linked to a larger orographic ratio present at
101 the 5 km scale compared to the 2.5 km scale. The modeling experiment also suggests that latitude and elevation are key factors in the best-fit models for non-cyclonic storms (Table
3.10). Since these storms can be slow moving or semi-stationary (Kaspi and Schneider
2013; Grover and Sousounis 2002; Lau 1988), they have a high sensitivity to temperature.
So, although the entire region may experience precipitation, it is largely the higher elevations of the northern half of the study area that experience snow.
3.4. Discussion
A persistent question in the climatological community has been how much of the Great
Lakes snowfall derives from lake-effect snow. Previous estimates suggest that lake-effect snowstorms contribute between 30-70% of the seasonal snowfall (Eichenlaub 1970; Miner and Fritsch 1997; Liu and Moore 2004; Veals and Steenburgh 2015). However, this range is quite variable and ignores the spatiality that exists across the Great Lakes region. The research presented here shows that snowfall contributions vary across Central New York for the different snowstorm types, and depends on the defined area of study. Results show that lake snowstorms produce approximately 47% of the total snowfall in Central New
York, an amount similar to the estimates for the entire Great Lakes region. These percentages change at the subregional and local levels. Although lake snowstorms account for nearly 47% of the seasonal snowfall in parts of Central New York, most notably over the
Tug Hill, contributions are closer to 25% in the southern extent of the study area. This highlights the importance of non-lake snowstorms in this region, as these storms contribute more than half of the seasonal snowfall totals. The influence of non-lake- snowstorms are potentially due to the additional moisture supplied from the Atlantic
102
Ocean, a source less available to other parts of the Great Lakes region (Zishka and Smith
1980; Sanders and Gyakum 1980; Jacobs et al. 2005; Changnon et al. 2008). Since the scale used to observe snowfall influences seasonal contributions, caution should be given to estimates for an entire region.
From the spatial analyses conducted for this research, it is evident that different storm magnitudes and storm types have varying influences throughout the study region. The majority of seasonal snowfall occurs from heavy-magnitude storms, especially heavy- snowfall lake-effect storms and Nor’easters. Lukens et al. (2018) noted similar patterns as strong storms, those that achieve a maximum potential vorticity exceeding the mean value by one standard deviation, represent approximately 16% of all storms within North
America and contribute 30-50% of the annual precipitation. Clearly, Lake Ontario plays a considerable role in producing heavy storms. However, snowstorm magnitude can be misleading compared to the precipitable water produced as lake-effect snow also tends toward a higher snow-to-liquid ratio which can result in greater measurable snow (Baxter et al. 2005).
From the spatial snowfall pattern, lake-effect snowstorms contribute the most to seasonal snowfall totals to the lee of Lake Ontario, with greater percentages also over the Southern
Hills and western Adirondack Mountains. The higher elevations and orientations of these locations are also prime for lake-effect snow, leading to larger seasonal snowfall totals
(Notaro et al. 2013a; Hjelmfelt 1992; Wilson 1977). The lower contributions from lake- effect snow in southeastern Central New York are likely due to its distance from Lake
103
Ontario and the Atlantic Ocean. The presence of lake-effect snow in this area requires strong, organized snowbands capable of retaining moisture over high terrain. The main mechanism that can carry this out is the presence of L2L snowbands. These tend to be less- organized and less frequent (Schultz et al. 2004), resulting in a lower overall lake-effect snow contribution in southeastern Central New York.
Nor’easters contribute the most to seasonal snowfall totals in the outer areas of Central
New York. The greater contributions in southeastern Central New York are likely due to its proximity to the central low pressure of the storm. The noticeable hot spot in the St.
Lawrence area is particularly interesting. This area is often far from the storm’s center and does not favor lake-enhanced snowfall, yet Nor’easters contribute more than 30% of the seasonal snowfall. The influence of Nor’easters in this area needs further investigation but may be due to lake-enhancement as the storm moves up the coast. As this occurs, winds shift from the north-northwest to west-southwest. These winds are favorable for the development of lake-effect snowstorms, especially due to cold air troughing over the Great
Lakes. The greater snowfall contributions likely also reflect the general size of these storms. Since Nor’easters are some of the largest snowstorms (Davis and Dolan 1993), they tend to affect all of Central New York. The St. Lawrence lowlands however, are not well-known for snowfall (Wright et al. 2013), only averaging 200 cm per year. Therefore, even though Nor’easters produce less snowfall in this area than southeastern Central New
York, its relative importance is similar due to the area’s relatively low seasonal snowfall totals.
104
Snowfall contributions from Canadian lows, non-cyclonic storms, and Rocky lows are relatively homogeneous throughout the study area. Canadian lows generally contribute little to seasonal snowfall totals throughout Central New York, and despite their association with moderate seasonal snowfall totals, it is likely due to the northern formation of these storms. Since they form at northern latitudes, inland from major moisture sources, these storms are often associated with anomalously cold, yet fairly dry conditions (Hutchinson
1995; Thomas and Martin 2007). Snowfall directly generated from the low pressure from these storms are often light to moderate in magnitude, and the snow that does fall tends to have a high snow-to-liquid ratio (Rochette et al. 2017)
Non-cyclonic storms are associated with upper level disturbances, cold air advection, or quasi-stationary fronts, but no central low pressure near the study area. Although they can be accompanied by significant development of lake-effect snow, they tend to be varied in form and genesis (Lackmann 2001; Scott and Sousounis 2001; Chuang and Sousounis
2003). Upper atmospheric disturbances tend to favor ascending air and an unstable atmosphere ahead of a westerly trough (NWS 2014a). Since the trough can extend to the
Gulf of Mexico, the snowfall produced often occurs throughout Central New York, as shown by the relatively homogeneous percent contributions (Figure 3.11d). Behind the western trough winds generally shift from the northwest, which are conducive for the formation of lake-effect snowfall, and the greater snowfall totals in northern Central New York.
Stationary fronts can also produce a relatively homogenous percent snowfall contribution across Central New York when the warmer air mass contains a lot of water vapor (Neiman
105 et al. 1998; Kusunoki et al. 2005). Since the front is stationary, this can lead to prolonged periods of intense precipitation.
The relatively homogenized snowfall contributions from Rocky lows are believed to be influenced by a lack of lake-effect enhancement during these storms. Since Rocky lows tend to occur more frequently in late winter (Whittaker and Horn 1981), lake surface temperatures are cold and ice cover extent is at a maximum (Wang et al. 2012). When
Rocky lows pass across the region, there is limited access to moisture from the lake inhibiting the formation of lake-effect and lake-enhanced snow. So, the areas that typically receive lake-effect snow do not receive any additional snow from Rocky lows than areas outside of the main lake-effect snowbelt. Since Rocky lows are generally extensive also, the entire region is similarly influenced by their passage receiving an average of 181.4 10.8 cm of seasonal snowfall.
It is clear that a storm’s influence on snowfall is not just driven by the regional geography, but also by the nature of the storm. The modeling experiment showed that latitude and longitude are almost universal in their influence on snowfall totals, as might be expected.
This helps explain the regionalization of the spatial distribution of snowfall contributions from different storm types. Canadian lows and non-cyclonic storms both illustrate a north- south division in their contributions to the region (Figure 3.11c and d), where the highest contributions are to the north, and slightly less to the south.
106
Lake-effect storms, Nor’easters, and Rocky lows show a spatial pattern that reflects the role of lake-effect or lake-enhanced snow. This is reflected by the models as the “distance from
Lake Ontario” influences the seasonal snowfall from lake-effect storms and Rocky lows.
Areas with the largest seasonal contribution from lake-effect snow, further referred to as the lake-effect snowbelt, are generally located to the lee of Lake Ontario, in close proximity to the lake, and in areas with a relatively exposed terrain. For Nor’easters, however, the spatial patterns in snowfall that emerge are because of the immense size of these storms relative to the region. In general, every part of Central New York receives snow, but the percentage of seasonal snowfall from these storms is far less in the lake-effect snowbelt.
Because of the construction of the maps as percentage contributions, that means that relatively speaking, Nor’easters provide a much lower proportion of the seasonal snowfall total in the lake-effect snowbelt.
The inverse pattern of the spatial distribution of snowfall contribution from lake-effect snowstorms and Nor’easters reflects the different scale of these storms, highlighting the local impact of lake-effect, but regional impact of Nor’easters. The heavy snowfall from
Nor’easters tends to be experienced across a large area of the eastern United States as the storms track north or northeastwards (Changnon et al. 2008; Bosart 1973; Kocin and
Uccellini 2004a; Hirsch et al. 2001; Mercer and Richman 2007). Their strength is driven by the strong baroclinic conditions from the nearby Atlantic Ocean, and they can often undergo bombogenesis (Zishka and Smith 1980; Sanders and Gyakum 1980; Jacobs et al.
2005; Cione et al. 1993; Kocin and Uccellini 2004a; Hirsch et al. 2001). Central New York’s position to the west of the storm center often brings a period of significant snowfall
107 sourced in Atlantic moisture brought in by the northeasterly winds to the northwest of the storm (Figure 3.12). The northwest winds over Central New York also often cause the formation of lake-effect and lake-enhanced snowfall to the southeast of Lake Ontario
(Niziol 1987; Suriano and Leathers 2017b; Liu and Moore 2004). Lake-effect snowstorms, on the other hand, are intrinsically connected to the lake, or in some cases, two or more lakes, and are experienced as streamers of strong convection that drop snow in spatially discrete, and relatively narrow swaths on the leeward shore in Central New York. These storms can produce snowfall locally at rates exceeding 15 cm hr-1 and with a single storm total in excess of 127 cm (Niziol 1987; Ellis and Leathers 1996; Ballentine et al. 1998).
Figure 3.12. Daily weather map from the NOAA/ESRL’s 20th Century Reanalysis V2 representing a Nor’easter on 14 March 1993 at 06 Z.
The seasonal snowfall from a snowstorm type is further complicated by the frequency and magnitude of the storm, as some relatively infrequent storms tend to produce heavy snowfall across the region (Lawrimore et al. 2014; Vose et al. 2014; Kunkel et al. 2013a; 108
Groisman et al. 2012, 2005). Although infrequent, these storms contribute considerably to seasonal snowfall totals because of their regional influence. Central New York experiences most of its snow in the form of heavy magnitude snowstorms that occur less frequently than lower magnitude storms. This relationship highlights the importance of severe, often highly disruptive snowstorms in the region (Cerruti and Decker 2011; Kocin and Uccellini
2004a; Uccellini et al. 1995). It is evident that lake-effect snowstorms and Nor’easters are the most frequent heavy-magnitude storms and are potentially disruptive (Table 3.6).
Changes in either of these storm types will have significant consequences for seasonal snowfall totals and potential for social and economic impacts.
3.5. Conclusion
Lake-effect snowfall is a regional phenomenon that greatly influences the climate, hydrology, biology, and economy of the Laurentian Great Lakes. The research presented here suggests that lake-effect snowstorms produce between 25-47% of the total snowfall in lake-influenced areas, but exhibits considerable variation across the region. Results suggest that the spatiality of snowfall contributions are best represented using station data; however, these data can be unreliable with inhomogeneities such as missing data.
Therefore, further analyses in this dissertation utilize snowfall contributions at the subregional scale because the data are more reliable than that at the local scale, yet also account for some of the spatiality lost at the regional scale.
Understanding the actual snowfall contribution from different snowstorm types is needed for future climate predictions. For accurate predictions, the model needs to be able to
109 determine whether snow is from lake-effect storms or not, because the potential trajectory of future snowfall varies according to the type of storm. In the eastern United States, since the early 20th century snowfall has significantly increased in areas dominated by lake-effect snow (Norton and Bolsenga 1993; Ellis and Johnson 2004; Burnett et al. 2003; Kunkel et al.
2009a; Hartnett et al. 2014). The increase in snowfall is linked to a faster warming of the
Great Lakes than surface air temperatures, especially in the spring and summer months
(Lofgren 2004; Trumpickas et al. 2009; Dietz and Bidwell 2011; Bai et al. 2012). This affects the moisture transfer and lake dynamics (Lofgren 2004; Trumpickas et al. 2009;
Dietz and Bidwell 2011; Bai et al. 2012), which amplify the warming of the lakes and delays or in some cases prevents their freezing (Assel 2003; Dietz and Bidwell 2011; Wang et al.
2012). When a cold outbreak occurs, the open water and the increased temperature lapse rate that a warmer lake surface presents, create ideal conditions for the development of strong, persistent snowbands (Hanson et al. 1992; Wright et al. 2013).
Increases of similar magnitude are not observed in non-lake effect influenced areas (Hirsch et al. 2001; Burnett et al. 2003; Thomas and Martin 2007; Kluver and Leathers 2015).
Previous findings suggest that this is linked to changes in synoptic storms, which have been associated with both a decrease in frequency and snowfall produced during the 20th century, a negative trend that is most evident in regions with average winter air temperatures at or just below 0C (Notaro et al. 2014). Explanations for the contrasting trends include a general warming trend, and therefore an increasing ratio of precipitation falling as rain during synoptic storms (Barnett et al. 2005; Knowles et al. 2006). NOAA suggests that most locations in New York State have experienced a decline in the ratio of
110 precipitation falling as snow from 1949-2015, ranging from a 17.6% loss in snowfall to a
7.3% gain. Alternatively, the decreased snowfall may be due to changes in the characteristics of non-lake-effect snowstorms. As Lawrimore et al. (2014) note, the frequency of severe storms has increased since the early 1900s. However, even though the magnitude of snowstorms has increased, Zarzycki (2017) suggests that there are fewer days that support the conditions necessary for snowstorms to develop, thus decreasing their total frequency and seasonal snowfall. Therefore, the analyses in this chapter help to better understand how snowfall may change in the future by directly teasing out the contribution from lake snowstorms versus that from non-lake snowstorms. These analyses also emphasize the spatial variability of snowfall contributions, which suggest that future snowfall trends may vary across a region depending on whether snowfall is dominated by lake-effect snow or non-lake-effect snow.
Therefore, if current snowfall trends and predictions for future snowfall are extrapolated
(Notaro et al. 2015; Suriano and Leathers 2016; Gula and Peltier 2012; Peltier et al. 2018;
Smith 1991), and lake-effect snowfall contributions increase at the expense of snowfall from Nor’easters, then snowfall in Central New York will become increasingly localized.
Snow will fall in very intense bands, and produce large single-storm totals, similar to the
November 14-19 2014 lake-effect snowstorms to affect northern Pennsylvania, Buffalo, NY, and the Tug Hill. The results from this study provide an important baseline to track these future scenarios, and to help isolate the changes in frequency and contributions of all the storms that track across the region.
111
4.0 THE SYNOPTIC CONDITIONS ASSOCIATED WITH DIFFERENT SNOWSTORMS WITHIN CENTRAL NEW YORK
4.1 Introduction
Snowstorms are common weather phenomena that occur within the Great Lakes region and the northeastern United States. The types of snowstorms to influence these areas and their frequency are linked to the synoptic conditions within the upper and lower atmosphere (Suriano and Leathers 2017a; Liu and Moore 2004; Ellis and Leathers 1996).
Comparing the average synoptic conditions associated with specific storm types allows for a comparison of the atmospheric features associated with weather impacts on the ground.
In this chapter, I identify the synoptic patterns that are favorable for the occurrence of different snowstorm types in Central New York. Snowfall patterns for three different magnitudes of storms, NCEP/NCAR reanalyses and air trajectories from the HYSPLIT model are plotted and used to examine linkages between the synoptic conditions and the snowfall distribution. Inferences are drawn from these links to identify the impact of specific atmospheric features on snowfall distribution.
Synoptic classification is used widely in climatological studies to better understand the links between the three-dimensional, transient properties of the atmosphere, and its interaction with the ground surface. It has been successfully applied to examine a diverse array of climatological questions, including how synoptic conditions influence snowfall distribution patterns following heavy (25.4 cm) snowfall producing snowstorms (Leathers and Ellis 1996; Karmosky 2007; Suriano and Leathers 2017a; O’Hara et al. 2009; Liu and
Moore 2004; Lackmann 2001; Notaro et al. 2013b; Zielinski 2002; Mullens et al. 2016; Ellis
112 and Leathers 1996; Leathers et al. 2002). For example, Younkin (1968) found that the heaviest snowfall from snowstorms west of 100W in the United States occurred between the 5340-geopotential-meter (gpm) and the 5460 gpm 1000-500 hPa contours. O’Hara et al. (2009) note that heavy-snowfall snowstorms in the Sierra Nevada Mountains correspond to enhanced moisture from the subtropics, low static stability, and strong upper-level dynamics. Mullens et al. (2016) found that freezing precipitation in the southern Great Plains is associated with the southern propagation of Arctic anticyclones, weak or absent surface cyclone formation, and a western trough axis in the jet stream. The authors also found that during periods of stronger cyclone development, snowstorms were favored over ice storms. From 1950/51 – 1981/82, Leathers and Ellis (1996) found that there were twenty-four synoptic patterns which produce snowfall in Syracuse, NY. Fifteen of these accounted for 98% of the snowfall days, and nine produced at least 2.0 cm of snowfall throughout the study area.
Within the Great Lakes region, researchers have focused on the occurrence of lake-effect snowstorms (Ellis and Leathers 1996; Liu and Moore 2004; Suriano and Leathers 2017a,b;
Leathers and Ellis 1996; Niziol 1987; Karmosky 2007; Peace and Sykes 1966). Since lake- effect snowstorms are often initiated within 1-2 days after the passage of a synoptic-scale low pressure system (Liu and Moore 2004), they occur when weather patterns favor clear conditions (Ellis and Leathers 1996; Peace and Sykes 1966; Suriano and Leathers 2017b;
Holroyd 1971). Ellis and Leathers (1996) suggest that there are five synoptic patterns that are linked to the occurrence of lake-effect snow in upstate New York. The authors found that synoptic conditions are similar for each lake-effect snowstorm, but are distinguished
113 by slight variations in sea level pressures, 850 hPa temperatures and heights, 500 hPa heights, seasonality, and overlake fetch and strength of flow. The five synoptic patterns were classified by their prevailing winds and included WNW-S, W-S, NW-S, WNW-W and
W-W. It is estimated that the five patterns account for 50-60% of the seasonal snowfall in the region. Leathers and Ellis (1996) found that the WNW-S synoptic type produced the most snowfall per storm from 1950/51 – 1981/82 in Syracuse, NY, but was the least frequent type. In a study from 1950-2009, Suriano and Leathers (2017b) found two additional synoptic types responsible for the formation of lake-effect snowfall leeward of the eastern Great Lakes. Generally, the synoptic types had a northwesterly or southwesterly flow over the lakes with a low pressure to the north and/or east of Buffalo,
NY and a high pressure to the west and/or south of Buffalo. Although researchers have examined the synoptic conditions conducive to the occurrence of lake-effect snowfall in the
Great Lakes region, no research examines the synoptic conditions favorable for the occurrence of the other ten snowstorm types identified in Chapter 2. Recognizing the synoptic conditions which promote a certain snowstorm type will improve forecasters ability to predict snowfall totals from an impending snowstorm. Thus, in this chapter, I examine the synoptic conditions responsible for the heavy-snowfall (≥ 25.4 cm) snowstorms identified in Chapter 2 within Central New York.
Since storms are dynamic systems, storm track analysis is often applied to examine how the conditions leading up to a storm influence the storm’s properties. Storm tracks are defined as the region where the synoptic-scale transient eddy activity is most intense
(Glickman 2000). These eddies are critical for the redistribution of heat, momentum, and
114 moisture in the earth’s climate system, and therefore effect the storm’s propagation characteristics, including direction of motion and geographic influence (Lareau and Horel
2012). Storm track analysis is often applied by examining time series of synoptic conditions from the moment of cyclogenesis (Petterssen 1956; Reitan 1974; Zishka and
Smith 1980; Whittaker and Horn 1981; Hoskins and Hodges 2002; Jeglum et al. 2010) or through baroclinic eddies within the mid- to upper-troposphere (Blackmon 1976; Lefevre and Nielson-Gammon 1995; Hoskins and Hodges 2002; Hakim 2000). Differences in storm tracks are driven by variability in the synoptic conditions and have been shown to influence the amount and distribution of snowfall from a storm (Changnon et al. 2008;
O’Hara et al. 2009). Changnon et al. (2008) found that the heaviest snowfall from snowstorms in the central and eastern United States was to the left of the cyclone track.
The average distance from the cyclone track to the edge of heavy snow (> 15.2 cm) was 201 km, but decreased as the cyclone progressed. The authors also found a general southwest to northeast track for the heaviest snowfall producing storms. Storm tracks also influence the type of precipitation that falls over a region during a cyclone. Since these storms form underneath the jet stream, they are often associated with continental polar air from Canada and maritime tropical air from the Gulf of Mexico or subtropical Atlantic (Serebreny et al.
1962). Thus, storm tracks considerably influence the location of the rain-snow boundary line (Braham 1983). Slight changes in the storm track can influence the type of precipitation that falls in an area, and in return the amount of snowfall. From storm track analyses, it is possible to predict where the heaviest snowfall will occur, distribution patterns of the snow, and the properties of the snow (e.g. snow-to-liquid ratio).
115
In the Great Lakes region, air trajectories have also been shown to influence snowfall patterns of non-cyclonic snowstorms. Liu and Moore (2004) note that the parent synoptic low preconditions the atmosphere for the development of lake-effect snow, and modulates the intensity of such storms (Ballentine et al. 1998; Sousounis and Mann 2000; Schmidlin and Kosarik 1999). If the air moves across multiple Great Lakes, its moisture and heat content is altered, which in return affects snowfall totals (e.g. Mann et al. 2002; Lang et al.
2018; Rodriguez et al. 2007; Laird et al. 2017). Lang et al. (2018) found that snowbands with a L2L connection favor heavier snowfall totals, while Laird et al. (2017) suggest that these higher totals are due to greater instability over the upwind lake, the availability of more near-surface moisture, faster wind speeds, and larger surface heat fluxes over the upstream lake.
To examine how these mechanisms influence snowfall distributions in Central New York, I examine the influence of storm tracks of the different snowstorm types identified in
Chapter 2. In addition, I assess the influence of storm tracks on the snowstorm magnitude
(light, moderate, or heavy) for each snowstorm type. By combining composite analyses with storm track analyses, this study provides a better understanding of how synoptic conditions influence snowfall totals and distribution patterns for both cyclonic and non- cyclonic snowstorms. This research will help improve short-term and long-term forecasts and will provide a basis for predicting where the heaviest snowfall will occur following any of the different snowstorm types.
116
4.2 Methods
The synoptic conditions and storm trajectories for snowstorms in Central New York were observed from the 1985/86 to 2014/15 snow seasons. The synoptic conditions associated with heavy-snowfall (≥ 25.4 cm) storms were observed using NCEP-NCAR reanalysis data outlined in Chapter 2 (Table 2.1; Liu and Moore 2004). This method was used as a reduction technique to objectively identify features common to a particular population, and to provide a basis for interpreting case studies (Carleton 1999). Composite analyses have been successfully used to examine lake-effect snowstorms as Ellis and Leathers (1996) used this technique to represent the average synoptic patterns favorable for lake-effect snow downwind of Lakes Erie and Ontario. Lackmann (2001) also used composite analyses to identify similarities in the synoptic conditions prior to and during lake-effect snowstorms along southern Lake Ontario.
Composite techniques were used in this study to examine the average synoptic conditions of the different snowstorm types that are detailed in Chapter 2. Storms were grouped into five classes for evaluation in this analysis: Canadian lows (clippers, Great Lakes lows, and
Hudson lows), lake-effect snowstorms (LES-H and LES-UL), Nor’easters (east coast storms and Gulf Coast storms), Rocky lows (Colorado lows, Oklahoma Hooks, and Texas hooks), and non-cyclonic snowstorms (upper disturbance, cold front, and stationary front).
Nor’easters were subdivided into two categories based on their area of cyclogenesis (Miller
1946). East coast storms were categorized as any Nor’easter that formed east of the
Appalachian Mountains, typically over or near the Atlantic Ocean, while Gulf Coast storms are those that formed south of 35N, over or near the Gulf of Mexico. Lake-effect
117 snowstorms were also separated into multiple categories based on the primary source of their snowfall including: high-pressure induced storm (LES-H), upper atmospheric low or trough induced storm (LES-UL), clipper induced storm (LES-CL), Colorado low induced storm (LES-CO), Nor’easter induced storm (LES-NE), Great Lakes low induced storm (LES-
GL), and Panhandle hook induced storm (LES-HK). Only LES-H and LES-UL storms were examined in this study because the other events occur relatively infrequently. Frontal storms were also subclassified by the type of front (cold, warm, occluded, stationary), but were only analyzed for cold and stationary frontal storms because of their frequency (> 30 storms over the study period).
Composite analyses were constructed for twenty randomly selected snowstorms for each of the thirteen snowstorm types. In instances where storm counts were less than twenty, all available snowstorms were used. Although thirty observations are ideal, twenty are sufficient to highlight the synoptic conditions during different snowstorm types.
Composite plots were constructed for the upper and lower atmosphere and included the
850 hPa geopotential height, the 850 hPa air temperatures, 850 hPa vector wind speeds, the sea level pressure, the surface air temperatures, and the surface vector wind speeds.
The 850 hPa level was chosen as it represents conditions in the lower troposphere, and the
850 hPa geopotential height field magnifies low pressure features (Liu and Moore 2004;
Morrison and Businger 2001). In addition, a temperature difference (> 13C) between the lake surface and the 850 hPa level is necessary for the formation of lake-effect snowstorms
(Holroyd 1971; Ellenton and Danard 1979). The synoptic patterns conducive to the
118 different snowstorm types were defined subjectively from the composite plots in a manner similar to that used by Ellis and Leathers (1996).
NOAA’s Air Resources Laboratory’s Hybrid Single-Particle Lagrangian Integrated
Trajectory model (HYSPLIT) was used to observe storm trajectories (Keighton et al. 2016;
Cordeira and Laird 2008; Fuhrmann and Konrad 2013; Saslo and Greybush 2017; Draxler and Hess 1997; Draxler 1998; Draxler and Rolph 2003; Rolph 2003). HYSPLIT computes air parcel trajectories and is commonly used for back-trajectory analyses to determine the origin of air masses (Stein et al. 2015; Fleming et al. 2012). The trajectories are calculated using a hybrid of the Lagrangian and Eulerian approaches. Since air trajectories indicate the general airflow rather than the exact pathway of an air parcel, twenty random observations were used for each snowstorm magnitude (heavy, moderate, and light) and type to reduce the effects of individual errors (Harris and Kahl 1990). Back trajectories were plotted at the hour of the storm’s maturation for three locations (Syracuse, Utica, and
Watertown) (see Chapter 2 for details). The distance of each trajectory was calculated in
ArcGIS and one-way ANOVAs were used to test whether there was a significant (ρ ≤ 0.05) difference between the mean distance traveled by heavy, moderate, and light snowstorms of the same snowstorm type. If statistical differences existed, two-sample t-tests were used to determine which storm tracks were significantly different. The results were interpreted as the forward speed of extratropical cyclones or the speed of upper atmospheric winds during non-cyclonic storms (Stein et al. 2015). Larger distances travelled signify faster steering winds. Linear directional means were calculated and plotted for each storm trajectory using ArcGIS (Mardia and Jupp 2000).
119
Storm trajectories for three magnitudes of snowstorms to affect Central New York from the winter seasons 1985/86 – 2014/15 were plotted against their snowfall distribution for using COOP data (described in Chapter 2) and the cokriging interpolation method (Eynon
1988; Daly et al. 1994; Guan et al. 2005). Kriging was used in preference to other methods due to its ability to interpolate values based on statistical relationships including autocorrelation. Since Kriging is more complex than rudimentary interpolation methods, it is not deterministically based on maximum and minimum values observed in the data. Even though the spatial distribution of COOP stations provides an adequate representation of snowfall throughout Central New York, there is a higher potential for error in areas with fewer stations. Since higher elevations tend to have larger snowfall totals due to colder air temperatures and increased orographic uplift, cokriging was used to incorporate the influences of elevation on snowfall totals (Lundquist et al. 2015). In addition, due to the geostatistical techniques used by the Kriging analysis, it has the capability of producing a prediction surface and a certainty surface of the predictions. The interpolated surfaces were used to calculate zonal statistics across Central New York. They were also used to calculate the area of Central New York covered by arbitrarily chosen snowfall thresholds (10, 15, 20, and 25 cm), similar to the methods outlined by Gao and
Hartnett (2016). These were then used to calculate a proportion of coverage by comparing the area covered by the threshold to the total area of Central New York. The results provided a measure of whether the storm’s effects are relatively local or regional.
120
4.3 Results and Analyses
Zonal statistics for heavy-snowfall snowstorms in Central New York are presented in Table
4.1. The most snow in Central New York following a storm occurs after Nor’easters (east coast storms and Gulf Coast storms), with the least snow following clippers. Not only do
Nor’easters produce the most snow, but heavier snowfall totals extend throughout the study area. This is represented by the relatively small range and the large percentage of
Central New York covered by at least 10 cm of snow following these storms (Table 4.2).
Table 4.1. Zonal statistics for the average snowfall (cm) per heavy snowstorm by snowfall type. Statistics include: the minimum snow (‘Min’), maximum snow (‘Max’), range in snowfall (‘Range’), mean snowfall (‘Mean’), standard deviation (‘Std’), sum of all snowfall in the area (‘Sum’), all measured in centimeters, and the snow coverage. Snow Coverage Storm Type Min Max Range Mean Std Sum (cm km-2) Clippers 3.1 17.0 13.9 7.2 3.3 313089.3 10.6 Cold Fronts 2.4 26.9 24.5 9.8 4.9 429050.2 14.5 Colorado Lows 8.4 20.3 11.9 12.2 1.9 532550.6 18.0 East Coast Storms 14.9 23.2 8.3 18.4 1.7 800352.9 27.1 Great Lakes Lows 4.0 20.2 16.3 9.3 4.0 407013.6 13.8 Gulf Coast Storms 12.9 25.2 12.3 18.6 2.1 809432.2 27.4 Hudson Lows 1.4 27.0 25.6 8.6 5.8 375536.4 12.7 LES – H 5.5 22.5 17.0 9.8 3.9 427088.1 14.5 LES – UL 4.1 23.9 19.8 9.1 3.7 397431.6 13.4 Oklahoma Hooks 7.6 16.9 9.3 10.6 1.5 460382.0 15.6 Texas Hooks 8.0 18.5 10.5 13.1 2.1 556958.7 18.8 Stationary Fronts 3.8 22.1 18.2 10.1 3.1 439001.5 14.9 Upper Disturbances 3.8 27.2 23.5 10.2 4.4 444749.9 15.1 *Values in table are in cm *Area = 29,550 km2
Storms that are more associated with lake-effect and lake-enhanced snow (e.g. clippers, cold fronts, Hudson lows, LES-H, and LES-UL) tend toward more localized snowfall patterns, as less than 50% of Central New York was covered by at least 10 cm snow following the storm. Each storm type is examined in detail in the next section.
121
The average 72-hour backward air trajectories for the different snowstorm types are shown in Table 4.3 and are organized by storm magnitude. Generally, as the average distance of the air trajectories increases, so too does the magnitude of the snowstorm.
Non-cyclonic storms (LES - H, cold fronts, and stationary fronts) were the only storms to have a significant difference (ρ < 0.05) between the distance travelled by the storm and the amount of snowfall produced. In all three instances, as the length of the trajectory decreases, the amount of snowfall decreases.
Table 4.2. The percentage of Central New York covered by various snowfall thresholds. Storm 10 cm 15 cm 20 cm 25 cm Clippers 23.4 1.7 0.0 0.0 Cold Fronts 45.1 10.9 3.1 0.1 Colorado Lows 86.8 4.8 0.0 0.0 East Coast Storms 100.0 99.7 18.0 0.0 Great Lakes Lows 49.7 7.2 0.0 0.0 Gulf Coast Storms 100.0 96.1 18.8 0.0 Hudson Lows 43.8 11.9 0.9 0.2 LES – H 45.1 7.6 1.6 0.0 LES – UL 43.6 4.2 0.7 0.0 Oklahoma Hooks 62.5 0.5 0.0 0.0 Texas Hooks 78.2 23.3 0.0 0.0 Stationary Fronts 48.9 4.6 0.6 0.0 Upper Disturbances 54.3 7.4 2.8 0.4
Table 4.3. Average trajectory length of snowstorms 72 hours prior to maturation. Trajectories are provided for three locations (S – Syracuse; U – Utica; W – Watertown). The difference in means were tested using a one-way ANOVA, with the significance of the test reported under the ‘P-value.’ Storm Type City Heavy Moderate Light df F-Stat Ρ-value S 3529.9 2900.5 2901.7 1.035 0.379 Hudson Lows U 2783.3 2825.8 2957.3 17 0.061 0.941 W 3225.4 2784.2 2912.8 0.595 0.565 S 2575.7 2377.1 2675.9 0.718 0.492 Clippers U 2495.6 2423.4 2665.3 58 0.383 0.684 W 2449.8 2298.8 2624.1 1.128 0.331 TABLE CONTINUED ON NEXT PAGE
122
TABLE 4.3 CONTINUED S 2854.5 2483.3 2408.4 0.446 0.096 LES – H U 2962.3 2458.3 2409.5 59 3.340 0.044 W 2785.0 2397.3 2513.7 1.506 0.232 S 2701.2 2452.0 2640.2 2.450 0.096 LES – UL U 2744.2 2445.3 2687.5 58 0.686 0.508 W 2635.4 2361.1 2581.4 0.572 0.568 S 2364.4 2191.7 2187.1 0.354 0.703 East Coast Storms U 1957.0 2228.9 2101.1 58 0.978 0.383 W 2036.1 2195.6 2050.0 0.406 0.669 S 2319.5 2323.5 1981.0 0.801 0.456 Gulf Coast Storms U 2481.0 2275.1 2005.8 41 0.771 0.471 W 2383.1 2274.5 1996.8 0.671 0.519 S 2225.2 2153.6 2059.8 0.420 0.659 Colorado Lows U 2313.2 2231.8 2127.2 59 0.457 0.635 W 2328.8 2147.4 2071.7 1.084 0.346 S 2632.7 2289.1 2368.3 0.769 0.469 Texas Hooks U 2493.2 2276.1 2346.0 52 0.229 0.796 W 2422.9 2305.1 2230.4 0.187 0.830 S 2500.3 2478.2 2156.4 1.363 0.264 Oklahoma Hooks U 2303.6 2511.7 2190.3 59 1.126 0.332 W 2366.5 2422.8 2197.7 0.595 0.556 S 2290.3 2119.7 2479.0 1.312 0.278 Great Lakes Lows U 2343.1 2103.2 2472.5 54 1.227 0.302 W 2174.7 2114.6 2391.7 0.783 0.462 S 2478.8 2471.2 2463.5 0.002 0.998 Upper Disturbances U 2528.0 2576.3 2523.3 59 0.022 0.978 W 2473.2 2441.9 2437.9 0.009 0.991 S 3177.5 2308.8 2470.1 3.551 0.037 Cold Fronts U 3160.2 2370.7 2432.0 46 2.692 0.079 W 3084.0 2307.0 2460.7 2.589 0.087 S 2704.5 2800.5 2041.4 5.634 0.006 Stationary Fronts U 2826.9 2849.8 2024.7 52 7.375 0.002 W 2718.6 2756.3 1949.0 7.013 0.002 *Track distance measured in km *Bold and italicized font represents a significant (ρ ≤ 0.10) difference between means *Bold, italicized, and red font represents a significant (ρ ≤ 0.10) difference between means
123
4.3.1 Canadian Lows
The largest snowfall totals from heavy-snowfall Canadian lows (Hudson lows, clippers, and
Great Lakes lows) are concentrated over northern Central New York near the Tug Hill and are considerably smaller to the south (Figure 4.1). The largest snowfall totals after Hudson lows and Great Lakes lows are closer to Lake Ontario than those produced by clippers.
Snowfall patterns following these storms may be tied to their trajectories as there is a prominent northwest flow during Hudson lows (Figure 4.1). Since air is from the west and moves directly over at least one Great Lake, there is the potential for lake-enhancement.
This is supported by the large range (25.6 cm) and standard deviation (5.8 cm) in snowfall following heavy-snowfall Hudson lows, traits that support a localized snowfall pattern
(Table 4.1). However, even though snowfall totals vary considerably among Hudson lows, more of Central New York is covered by at least 10 cm of snowfall after a Hudson low than a clipper (Table 4.2). As trajectories shift progressively to the north, lake-effect snow is less conducive and snowfall totals are less across the study area.
Compared to other snowstorms, Great Lakes lows have a relatively high snowfall range
(16.3 cm) and a low minimum (4.0 cm) and mean (9.3 cm) snowfall. The relatively small snowfall totals may be linked to the northern trajectories of these storms. Air for heavy- snowfall Great Lakes lows tend to cross northeast through Central New York, whereas there is a pronounced easterly trajectory during moderate and light storms. Snowfall from clippers is the most localized; it has the smallest mean snowfall (7.2 cm) per storm and the smallest area (23.4%) covered by at least 10 cm of snow. As winds enter Central New York, they are mostly from the southwest, and may account for the heavier snowfall
124
A. Hudson Lows
B. Clippers
125
C. Great Lakes Lows
Figure 4.1. Average snowfall (cm) and air trajectories per heavy (left), moderate (center), and light (right) Hudson lows (top), clippers (middle) and Great Lakes low (bottom) in Central New York. Individual air trajectories are displayed in the inset.
126 concentrated over northern Central New York (Figure 4.1). There are no noticeable differences in storm trajectories and the magnitude of the snowstorm.
The average synoptic conditions which promote heavy-snowfall Canadian lows are shown in Figure 4.2. The lowest geopotential heights (< 1250 m) are over the high latitudes of eastern Canada and the Hudson Bay, likely due to the northern cyclogenesis of Hudson lows. There is also a prominent ridge in sea level pressure over the central and western
United States. The ridge is most prominent during Great Lakes lows, as it extends from the southern Rockies to the Arctic, with sea level pressures in excess of 1025 hPa north of
Alaska. The ridge is less prominent (1017-1022 hPa) during clippers and least prominent during Hudson lows as higher (> 1015 hPa) sea level pressures only extend into southern
Alberta and Saskatchewan. A secondary low pressure also exists near the western coast of
Alaska for all three storms. This secondary low is the largest and strongest (< 1000 hPa) during Hudson lows.
The coldest surface air temperatures (< 245 K) during Hudson Lows are concentrated over
Hudson Bay, with cold air (265 – 270 K) penetrating into Central New York. Surface air temperatures are mostly below 265 K for locations north of 40N in North America. The coldest air temperatures during clippers are further to the north, resulting in warmer surface air temperatures in Central New York (265 – 275 K). Overall, air temperatures are warmer during clippers than Hudson lows throughout the United States. Temperatures over Central New York range between 265 K and 270 K during Great Lakes lows, but the coldest air is not as widespread as with Hudson lows. Upper atmospheric winds are also
127 similar for these storms as wind speeds are stronger over the Great Lakes and Northeast
United States. Upper and surface winds are fastest off the coast of the Mid-Atlantic during
Great Lakes lows, and over the Great Lakes during Hudson lows and clippers. There is also a stronger meridional flow to the winds during Hudson lows than the other two storm types.
a. Hudson Lows
b. Clippers
128
FIGURE 4.2 CONTINUED
c. Great Lakes Lows
Figure 4.2. Average atmospheric conditions during heavy-snowfall Canadian lows: Hudson lows (a), clippers (b), and Great Lakes lows (c). Reanalysis composites include the: 850 hPa Geopotential Height (top left), 850 hPa air temperatures (top center), 850 hPa wind velocity (top right), surface level pressure (bottom left), surface air temperatures (bottom center), and surface wind velocities (bottom right).
The snowfall distribution and relatively large range in snowfall throughout Central New
York following the passage of Canadian lows suggests these storms produce localized heavy snowbands (Figure 4.1; Table 4.1; Table 4.2). Since these storms originate in polar air masses, cold air advects over the Great Lakes producing the conditions conducive for the formation of lake-enhanced snow in Central New York (Niziol 1987; Scott and Huff
129
1996). Thus, lake-enhanced snow may account for the localized distribution of snow from these storms and the large differences in snowfall totals between northern and southern
Central New York (Figure 4.1; Laird et al. 2001).
This argument is supported by the atmospheric trough situated over the Great Lakes during these storms (Figure 4.2). The extension of the trough over the Great Lakes leads to lower air temperatures and enhanced wind speeds across the Great Lakes region. These conditions are prime for the development of lake-effect snow (Niziol 1987; Peace and
Sykes 1966), and likely explain the localized higher snowfall totals over the Tug Hill, yet relatively low snowfall totals throughout the rest of Central New York (Table 4.1; Figure
4.1; Thomas and Martin 2007; Hutchinson 1995; Mercer and Richman 2007). The 850 hPa atmospheric trough also explains air trajectories during Hudson lows and clippers (Figure
4.1), as the meridional flow of air results in trajectories from the northwest and southwest, respectively. In return, these air trajectories influence the location of heaviest snowfall
(Figure 4.1). For example, the largest snowfall totals for all three storms are confined to areas just east and northeast of Lake Ontario. The average trajectory for heavy-snowfall
Canadian lows, especially for Great Lakes lows, are more conducive for lake-enhanced snow over the Tug Hill than that of moderate and light snowfall producing storms.
Research suggests that the Great Lakes can alter the intensity and speed of passing cyclones (Sousounis and Fritsch 1994; Boudra 1981; Danard and McMillan 1974), especially during the ice-free season from September – November (Angel and Isard 1997).
Sousounis and Fritsch (1994) note that cyclones accelerate rapidly into the Great Lakes
130 region, then slow for 12 hours while deepening over the lakes before progressing to the east. However, since most clippers influencing Central New York pass south of the Great
Lakes (Figure 4.1), there was little moisture influx into the air resulting in relatively low snowfall totals downwind of Lake Ontario. Since average air trajectories are from the south-southwest, only air over Watertown, NY passes directly over Lake Ontario (Figure
4.1). Therefore, the low mean snowfall following clippers is likely due to air trajectories favoring lake-enhanced snow over the Tug Hill and western Adirondack Mountains, but nowhere else. The largest snowfall totals after Hudson lows are displaced further to the north compared to those of clippers and Great Lakes lows (Figure 4.1). This may be linked to the northern formation of these storms, originating over Hudson Bay, and thus producing more snow at higher latitudes.
4.3.2 Lake-Effect Snowstorms
The range in average snowfall totals is considerably greater following lake-effect snowstorms (LES – H and LES – UL) than that of Nor’easters and Rocky lows (Table 4.1).
This is likely because only 40-50% of Central New York is covered by at least 10 cm of snowfall following lake-effect snowstorms, compared to more than 60% following large- scale cyclonic storms (Table 4.2). However, more of Central New York is covered by at least 20 cm of snowfall following lake-effect snowstorms compared to Colorado lows, Texas hooks, and Oklahoma hooks. Air trajectories for these storms often originate over the
Ontario and southern Manitoba Provinces, traverse over the Great Lakes and enter Central
New York (Figure 4.3). Their northerly origin results in the air traveling a greater distance
131
A. LES - H
B. LES – UL
Figure 4.3. Average snowfall (cm) and air trajectories per heavy (left), moderate (center), and light (right) LES-H (top) and LES-UL (bottom) snowstorms in Central New York. Individual air trajectories are displayed in the inset.
132 over the Great Lakes and is favorable for L2L snowbands, which may be responsible for the heavy localized snowfall patterns.
The highest snowfall totals from lake-effect snowstorms are concentrated over the Tug Hill, diminishing rapidly with distance from the Tug Hill (Figure 4.3). As the snowstorm magnitude diminishes, the heaviest snowfall is more widespread and concentrated further from the lake. The shift in the heaviest snow is likely due to changes in air trajectories.
Although air is predominately from the northwest, trajectories during heavy-snowfall storms are more from the west, travelling a longer distance over Lake Ontario, while air has a greater northerly component during moderate and light-snowfall storms.
Heavy-snowfall lake-effect snowstorms are common when there is a surface low pressure
(< 1005 hPa) within a cold air mass (240 – 250 K) located over northeast Canada, and a secondary low pressure (< 1002 hPa) over the Aleutian Islands (Figure 4.4). This is similar to previous findings (Ellis and Leathers 1996; Leathers and Ellis 1996; Liu and Moore
2004; Suriano and Leathers 2017b), and also similar to the synoptic conditions during
Canadian lows, but the low pressure (< 999 hPa) over northeast Canada is much smaller.
The smaller low creates a stronger pressure gradient, with faster 850 hPa winds (> 14 m s-
1) over the northwest Atlantic Ocean. There is also a strong (> 1023 hPa) surface high pressure over the Mississippi River Valley. This is a common pattern resulting in lake- effect snowfall leeward of the eastern Great Lakes, one that Suriano and Leathers (2017b) identified as the WSW2 and W2 patterns.
133 a. LES-H
b. LES-UL
Figure 4.4. Average atmospheric conditions during heavy-snowfall lake-effect storms: high-induced LES (LES-H) and upper atmospheric disturbance induced LES (LES-UL). Reanalyses include the: 850 hPa Geopotential Height (top left), 850 hPa air temperatures (top center), 850 hPa wind velocity (top right), surface level pressure (bottom left), surface air temperatures (bottom center), and surface wind velocities (bottom right).
134
Surface and upper air temperatures during heavy lake-effect snowstorms are most similar to Hudson lows, suggesting that air temperatures are well below freezing. This suggests that since lake-effect snow is generally confined to areas near the lake (Peace and Sykes
1966), air temperatures may strongly influence the production of snowfall during these events. This is supported by snowfall patterns following lake-effect snowstorms whereby snowfall totals are largest in regions where air temperatures are generally colder and elevations are higher. The influence of air temperatures is also supported by the timing of the WSW2 and W2 patterns, as these patterns occur most frequently during the coldest months from December – February (Suriano and Leathers 2017b).
A strong pressure gradient over the Great Lakes region and northwest Atlantic Ocean forms during lake-effect snowstorms. The faster winds are likely responsible for the significantly (two-sample t-tests, ρ < 0.05) longer (2780.4 km) air trajectories with heavy- snowfall lake-effect snowstorms than that of heavy-snowfall Canadian lows (2652.0 km).
Lighter storms travel a significantly (ρ < 0.05) shorter distance than heavy storms for lake- effect snowstorms, cold fronts, and stationary fronts. This suggest that strong upper atmospheric winds promote higher snowfall totals from these storms. The higher snowfall totals might be linked to an increase in jet streaks associated with the faster winds.
Increased jet streaks enhance atmospheric instability and result in more precipitation
(Uccellini and Kocin 1987). Faster upper atmospheric winds also allow lake-effect snowstorms to penetrate further inland, such as over Utica, NY (Veals and Steenburgh
2015; Veals et al. 2018; Strommen and Harman 1978).
135
Extratropical cyclones that exhibit longer trajectories indicate a faster forward speed, and therefore a shorter time spent over Central New York (Angel and Isard 1997). Since the speed of a snowstorm influences snowfall totals (Changnon et al. 2008), the longer a storm influences an area, the higher the likelihood for heavy snowfall. Thus, longer air trajectories during extratropical cyclones equate to less snowfall, which likely explains the nonsignificant (ρ > 0.10) difference in trajectory lengths for these storms (Table 4.3).
Results also suggest that larger snowfall totals from lake-effect snowstorms occur with longer mean air trajectories over Lake Ontario (Figure 4.3). When air travels a large distance over the Great Lakes, snowbands are more organized (Kristovich et al. 2003; Veals and Steenburgh 2015; Mann et al. 2002) and independent of other synoptic snowstorm types (Grover and Sousounis 2002; Ellis and Leathers 1996; Liu and Moore 2004; Suriano and Leathers 2017b; Leathers and Ellis 1996). Storm trajectories passing over multiple
Great Lakes enhance the likelihood for multi-lake snow enhancement (Mann et al. 2002;
Lang et al. 2018; Rodriguez et al. 2007; Laird et al. 2017). L2L connections are common during heavy-snowfall lake-effect storms (Figure 4.3), especially between Lake Ontario and
Lake Huron (Laird et al. 2017). Thus, the enhanced snowfall associated with L2L snowbands may explain why lake-effect snowstorms average more snowfall throughout
Central New York than clippers (Table 4.1; Table 4.2). In addition, since the average trajectories for lake-effect snowstorms pass directly over Lake Ontario (Figure 4.3), slight variations in the wind direction influence snowfall totals. For example, storms with a more westerly component produce the most snowfall, probably because it increases the fetch
136 over Lake Ontario. This increases the moisture content and decreases the stability of the air, leading to larger snowfall totals throughout Central New York (Niziol 1987).
4.3.3 Nor’easters
Even though Nor’easters are coastal storms, their influence extends throughout Central
New York. This is highlighted by their smaller range and significantly (two sample t-tests, df = 19, ρ < 0.05) larger mean average snowfall than other snowstorms (Table 4.1).
However, the maximum snowfall totals from these storms are not significantly (two sample t-tests, df = 19, ρ < 0.05) greater than after other storm types. This suggests that although snowfall totals are not the greatest from Nor’easters, they produced heavy snowfall throughout Central New York. This is reflected by a considerably larger percentage of
Central New York covered by at least 10 cm (100%) and 15 cm (96.1 – 99.7%) of snow per
Nor’easter (Table 4.2), and a statistically (two sample t-tests, df = 19, ρ < 0.01) higher mean minimum snowfall (12.9 – 14.9 cm) than all other snowstorms.
Between the two types of Nor’easters, east coast storms have more evenly distributed snowfall totals throughout Central New York (Table 4.2; Figure 4.5). This may be due to air trajectories directly over Lake Ontario during east coast storms compared to north- northwest air during light-snowfall Gulf Coast storms, northwest air for moderate storms, and northern air for heavy storms (Figure 4.5). Air from the northwest passes directly over
Lake Ontario and has the potential for lake-enhanced snowfall in areas less prone to snowfall directly associated with the central low. Interestingly, the trajectories of east coast storms with moderate and heavy snowfall are the reverse of the trajectories of Gulf
137
A. East Coast Storms
B. Gulf Coast Storms
Figure 4.5. Average snowfall (cm) and air trajectories per heavy (left), moderate (center), and light (right) east coast storms (top) and Gulf Coast storms (bottom) in Central New York. Individual air trajectories are displayed in the inset.
138
a. East Coast Storms
b. Gulf Coast Storms
Figure 4.6. Average atmospheric conditions during heavy-snowfall Nor’easters: east coast storms (a) and Gulf Coast storms (b). Reanalyses include the: 850 hPa Geopotential Height (top left), 850 hPa air temperatures (top center), 850 hPa wind velocity (top right), surface level pressure (bottom left), surface air temperatures (bottom center), and surface wind velocities (bottom right).
139
Coast storms with moderate and heavy snowfall. Overall, there was little cohesion in air trajectories during Nor’easters.
Heavy snowfall producing east coast storms are most common when the surface low pressure (1003 – 1008 hPa) is located around Long Island, NY (42N, 70W) (Figure 4.6).
In comparison, Gulf Coast storms produced the most snow when the central low is centered around 38N, 70W. Heavy snowfall totals also occur when a strong (> 1020 hPa) surface ridge of high pressure is located over the entire Rocky Mountains, with a secondary high (>
1020 hPa) centered around 29N, 55W. East coast storms are common when a relatively small high pressure is centered over 40N, 110W, and a secondary high near 32N, 45W.
Differences in snowfall distributions between Gulf Coast storms and east coast storms may be associated with upper and surface wind patterns. Upper atmospheric wind speeds are typically faster during east coast storms than Gulf Coast storms and show a prominent trough in the jet stream over the Great Lakes region (Figure 4.6). Since air circulates counterclockwise around cyclones in the Northern Hemisphere, storm tracks east of
Central New York advect polar air from Canada into Central New York over the Great Lakes
(Davis and Dolan 1993). This enhances the likelihood for lake-enhanced snow suggesting that snowfall can occur from a mixture of cyclonic snowfall and lake-enhanced snowfall, which is consistent with snowfall patterns from this analysis (Figure 4.5; Niziol 1987;
Suriano and Leathers 2017b; Liu and Moore 2004). For example, snowfall totals are highest southeast of Lake Ontario over Oswego and Onondaga Counties, even though these locations are relatively far from the central low pressure.
140
Since the location of the central low pressure also varies between east coast storms and
Gulf Coast storms, their influence on the location and strength of lake-enhanced snow is also different. East coast storms track closer to Central New York, with winds blowing northwest across Lake Ontario, increasing the likelihood for lake-enhanced snow (Kocin and Uccellini 2004b; Goree and Younkln 1966). In contrast, Gulf Coast storms track further to the east, with winds predominantly from the north, failing to pass over Lake Ontario
(Figure 4.5). Since elevations in Central New York are highest in the Adirondack Mountains
(Figure 2.1), as air advects over this terrain, orographic uplift promotes large snowfall totals in this region (Figure 4.5; Joly et al. 2018; Perry et al. 2007; Perry and Konrad 2006).
This suggests that Gulf Coast storms are less likely to induce lake-effect snowfall with their passage, and that the moisture source for snowfall is almost entirely from the Atlantic
Ocean. Thus, the northern air trajectories of these storms likely produce the large range in average snowfall totals between eastern and western Central New York (Table 4.1).
4.3.4 Rocky Lows
Individual air trajectories for heavy-snowfall Rocky lows primarily originate over northwest Ontario and Manitoba, Canada (Figure 4.7). These trajectories typically move across the Great Lakes prior to entering Central New York. After the storm passes, snowfall is relatively uniform across the study area. Unlike lake-effect snowstorms, these storms have a relatively high mean snowfall and a relatively small range in average snowfall per storm (Table 4.1). This is likely due to the large aerial extent of these storms (Mote et al.
1997), as 62.5 – 86.8% of Central New York averages at least 10 cm of snowfall (Table 4.2).
141
A. Colorado Lows
B. Texas Hooks
142
C. Oklahoma Hooks
Figure 4.7. Average snowfall (cm) and air trajectories per heavy (left), moderate (center), and light (right) Colorado lows (top), Texas hooks (middle) and Oklahoma hooks (bottom) in Central New York. Individual air trajectories are displayed in the inset.
143
The greatest snowfall totals after heavy, moderate, and light Colorado lows are concentrated in eastern Central New York (Figure 4.7). The snowfall distribution following heavy-snowfall Texas hooks and Oklahoma hooks is also widespread (Figure 4.7); however, the largest snowfall totals are concentrated over the Tug Hill. Larger snowfall totals following moderate and light snowstorms are generally to the north, with smaller totals to the south. The less localized snowfall patterns are likely associated with a lack of lake- effect snow, as trajectories are primarily from the south-southwest (Figure 4.7). Lighter- snowfall storms tend toward trajectories progressively to the north, suggesting a strong influx of southern air which is likely warmer and inhibits the formation of snow.
The strength of the central low pressure has less influence on the occurrence of heavy- snowfall Rocky lows than it did with the occurrence of heavy-snowfall Nor’easters (Figure
4.8). Rather, the strength and location of high-pressure systems over the northcentral
United States and southcentral Canada are most influential. Rocky lows have an omega pattern with a prominent ridge in the 850 hPa geopotential heights over the central/western United States and a trough over the Great Lakes and Northeast United
States. 850 hPa air temperatures are slightly lower west of the Great Lakes. Heavy- snowfall Rocky lows are also common when the surface low pressure passes directly over
Central New York and a prominent high pressure exists between 40-50˚N and 90-110˚W.
Although similarities exist between the Rocky lows, geopotential heights during Colorado lows are much higher throughout the southern United States, while the central low pressure is much lower (Figure 4.8). The high pressure over the northern Central Plains
144 a. Colorado Lows
b. Texas Hooks
145
FIGURE 4.8 CONTINUED c. Oklahoma Hooks
Figure 4.8. Average atmospheric conditions during heavy-snowfall Rocky lows: Colorado lows (a), Texas hooks (b), and Oklahoma hooks (c). Reanalyses include: 850 hPa Geopotential Height (top left), 850 hPa air temperatures (top center), 850 hPa wind velocity (top right), surface level pressure (bottom left), surface air temperatures (bottom center), and surface wind velocities (bottom right).
is strongest during Texas hooks and Oklahoma hooks. In comparison, the high pressures in the Atlantic and Pacific Oceans are most organized during Colorado lows with an eastern extension of the high in the Atlantic. Finally, Colorado lows and Texas hooks have considerably faster upper level winds over the Northeast United States.
Snowfall patterns following Rocky lows are generally consistent across different snowstorm magnitudes (Figure 4.7). Since different magnitude storms have similar patterns and since air temperatures associated with these storms are often near freezing
(Hartjenstein and Bleck 1991), it is suggested that the magnitude of these storms is largely influenced by surface air temperatures and the strength of the meridional flow of the jet
146 stream. Rauber et al. (2002) suggest that the prominence of cold air flow from Canada influences the production of blizzard conditions during Colorado lows. The presence of a strong omega blocking pattern associated with Rocky lows can produce the anomalously cold temperatures in the central and eastern United States (Figure 4.8). Barriopedro et al.
(2006) also proposed that omega patterns over the central United States can increase the frequency of midlatitude cyclones in the eastern United States. When Colorado lows are compared to Texas hooks and Oklahoma hooks, the high pressure in the Atlantic is displaced further east (Figure 4.8), resulting in more zonal air trajectories (Figure 4.7;
Table 4.2). This increases the frequency of Colorado lows and may be responsible for its greater amount of Central New York covered by at least 10 cm of snow.
Air trajectories for heavy-snowfall Rocky lows are from the south-southwest. This is consistent with the characteristics of these storms which typically make a sharp northerly turn into the Great Lakes or Northeast, hence their name “hooks” (Changnon et al. 2008).
Generally, a sharper turn of the storm results in higher precipitation totals due to increased energy associated with the storm system (Angel and Isard 1998; Grover and Sousounis
2002). Results here substantiate the importance of the storm’s track. Heavy snowstorms have a southerly component, whereas smaller storms have more noticeable westerly trajectories within Central New York (Figure 4.7). Western winds indicate an increased zonal flow in the jet stream, which often inhibits the strengthening of midlatitude cyclones
(Archambault et al. 2008). Compared to other snowstorms, snowfall totals are relatively high in southern Central New York following Rocky lows (Figure 4.7), likely due to their northern air flow (Figure 4.7).
147
Heavy-snowfall Colorado lows have less of a northerly track than that of moderate-snowfall
Colorado lows (Figure 4.7). The more eastern movement of the heavy-snowfall Colorado
Lows may account for their greater snowfall totals. The westerly component leads to a longer fetch across Lake Ontario, especially for air over Watertown, NY, which is conducive for the production of lake-enhanced snow (Ellis and Leathers 1996; Liu and Moore 2004;
Suriano and Leathers 2017b; Leathers and Ellis 1996; Niziol 1987; Notaro et al. 2013b;
Karmosky 2007; Peace and Sykes 1966). A northerly track however is unfavorable for lake-enhanced snow, likely resulting in the lower snowfall totals of moderate-snowfall
Colorado lows. Interestingly, the trajectories for light snowfall Rocky lows also cross the lake. However, lake-enhanced snow during these storms is unlikely because the jet stream has a more zonal flow. Since zonal flow is associated with smaller temperature gradients in the atmosphere (Hartjenstein and Bleck 1991), there is less energy during these storms and air temperatures often hover near freezing. Thus, even if temperatures are below freezing, they are still often too warm to satisfy the temperature difference needed between the lake surface and 850 hPa layer in the atmosphere to form lake-effect snow
(Laird et al. 2009a). From composite analyses, surface air temperatures in Central New
York are approximately 5C warmer during Rocky lows (-3C) than all other storm types.
Finally, these storms are most frequent during the late-winter to early-spring, when lake surface temperatures are the coldest (Wang et al. 2012). Thus, even if wind directions are conducive for lake-effect snow, other environmental conditions are often not.
148
4.3.5 Non-Cyclonic Snowstorms
There was strong spatial variability in the snowfall distributions from non-cyclonic snowstorms (Table 4.1; Figure 4.9). Like most storms, the greatest snowfall totals were concentrated over the Tug Hill. This was most notable following heavy-snowfall upper atmospheric disturbances and cold fronts, as larger snowfall totals following heavy- snowfall stationary fronts also existed in southeastern and northern Central New York and into the Southern Hills. As the storm magnitude lessened, snowfall totals were more similar throughout the region.
Most air trajectories during heavy snowfall producing upper atmospheric disturbances and frontal storms were from the west. Air typically passed over the northern Great Lakes during upper atmospheric disturbances and over the southern Great Lakes during cold fronts and stationary fronts. The average trajectories for heavy and moderate snowfall producing upper atmospheric disturbances and cold fronts were similar, with air passing directly over Lake Ontario (Figure 4.9). In both cases, the average air trajectories were from the west-northwest, passing over the long axis of the lake. Air trajectories were also from the northwest for light-snowfall cold fronts, but were predominately from the north during light-snowfall atmospheric disturbances. Comparatively, trajectories for moderate and light snowfall stationary fronts were similar, coming from the northwest and passing over Lake Ontario. However, trajectories during heavy-snowfall stationary fronts had a strong southerly component. This air did not pass over Lake Ontario, and instead entered
Central New York over the Finger Lakes region.
149
A. Upper Disturbances
B. Cold Fronts
150
C. Stationary Fronts
Figure 4.9. Average snowfall (cm) and air trajectories per heavy (left), moderate (center), and light (right) upper atmospheric disturbances (top), cold fronts (middle) and stationary fronts (bottom) in Central New York. Individual air trajectories are displayed in the inset
151
Heavy-snowfall upper atmospheric disturbances occurred when geopotential heights over the Northeast and eastern Canada were less than 1350 m, while geopotential heights over the Midwest were less than 1440 m (Figure 4.10). These storms also occurred when the pressure gradient between the high pressure over the southern Mississippi River Valley and the low pressure over New England was at least 18 hPa. Heavy-snowfall cold fronts occurred when the 1530 m geopotential height was over northeast Canada (Figure 4.10).
Surface pressure patterns were similar to those illustrated during lake-effect snowstorms, but the low pressure was located just off the coast of eastern Canada and the high pressure was centered over the coastal Carolinas. Higher geopotential heights (> 1550 m) over the subtropical Atlantic and Pacific Oceans were indicative of stationary fronts associated with heavy snowfall (Figure 4.10). These storms occurred when geopotential heights were greater than 1410 m throughout the United States, with lower (1250 m) geopotential heights over northern Quebec and the Hudson Bay. These storms also formed when a strong contrast exists between the surface low pressure over northeast Canada and the surface high over Georgia. Finally, upper atmospheric winds in excess of 10 m s-1 above
Central New York lead to heavier snowfall totals for all three snowstorm types.
The snowfall distribution following upper atmospheric disturbances and cold fronts is likely due to an influx of the cold air (Lackmann 2001; Steenburgh et al. 2000). Since cold front precipitation is often intense, but short lived (Austin and Blackmer Jr. 1956; Cox
1959), lake-enhanced snow is likely responsible for the higher snowfall totals after heavy- snowfall cold front storms. Typically, after the front’s passage, stable air advects into a region producing little to no precipitation. However, the presence of the Great Lakes
152 initiates lake-effect snow, by supplying moisture (Eichenlaub 1970; Notaro et al. 2013a).
The synoptic conditions show a lower geopotential height over the Great Lakes (Figure
4.10), which allows colder air to penetrate further south, fueling lake-enhanced snow.
a. Upper Atmospheric Disturbances
b. Cold Fronts
153
FIGURE 4.10 CONTINUED c. Stationary Fronts
Figure 4.10. Average atmospheric conditions during heavy-snowfall non-cyclonic storms: upper atmospheric disturbances (a), cold fronts (b), and stationary fronts (c). Reanalyses composites include the: 850 hPa Geopotential Height (top left), 850 hPa air temperatures (top center), 850 hPa wind velocity (top right), surface level pressure (bottom left), surface air temperatures (bottom center), and surface wind velocities (bottom right).
Heavy-snowfall snowstorms also tend to occur when the pressure gradient between the high pressure over the southern Mississippi River Valley and the low pressure over New
England is at least 18 hPa. The stronger pressure gradient enhances low-level instability and the formation of lake-effect snow in Central New York (Mann et al. 2002; Ellis and
Leathers 1996). Surface pressure patterns are similar to those that occur during lake-effect snowstorms, but the low pressure is further from Central New York and less organized, and the high pressure is centered over the coastal Carolinas instead of the interior United
States. These similarities likely result in the formation of lake-enhanced snowfall after the passage of cold fronts. However, differences in the location of the central pressure may lead to the localized heavy snowfall totals after cold fronts (Table 4.1; Table 4.2). 154
Air trajectories also suggest that lake-enhanced snow is common during non-cyclonic storms. Snowstorm trajectories for upper atmospheric disturbances and cold fronts producing heavy and moderate snowfall pass directly over Lake Ontario (Figure 4.9). Since both of these storms advect cold air into Central New York with winds from the northwest
(Lackmann 2001; Steenburgh et al. 2000), the higher snowfall totals are likely a reflection of lake-enhanced snow. Air trajectories during smaller magnitude upper atmospheric disturbances have a shorter fetch over the lake producing less lake-enhanced snow. Air trajectories during cold frontal storms were nearly identical for all three magnitudes
(Figure 4.9). This suggests that the fetch has less of influence on snowfall totals and instead are influenced by other factors (e.g. conditions of the lake and air).
Stationary fronts during the winter tend to form along the boundary of two high pressure systems or underneath the jet stream (Grover and Sousounis 2002). Thus, air temperatures are often below freezing north of the frontal boundary and above freezing south of the frontal boundary. Therefore, snowfall totals after a stationary front often fluctuate depending on the location of frontal boundary. Average synoptic conditions suggest that heavy-snowfall stationary fronts form during strong contrasts between a surface low pressure over northeast Canada and a surface high over Georgia (Figure 4.10).
This is likely because these pressure systems remain relatively stagnant (Turner and
Gyakum 2011), bringing prolonged moisture to Central New York. Since stationary fronts are not moving, air trajectories play less of a role than in other snowstorm types. Instead, snowfall totals are likely a product of the characteristics of the air masses and the specific location of the front. For example, air trajectories during heavy-snowfall stationary fronts
155 are from the southwest. The southern flow of air allows humid air to enter Central New
York, where it clashes with cold air from Canada, and can produce heavy snowfall totals if air temperatures are below freezing. However, heavy snowfall totals are less likely with wind from the northwest because even though that air is cold, it is relatively dry, producing little precipitation (Turner and Gyakum 2011). Since stationary fronts are less associated with lake-enhanced snow than cold fronts and upper atmospheric disturbances, snowfall totals are more homogeneous throughout the study area (Table 4.1; 4.2).
4.4 Discussion
In this chapter, I use the snowstorms classified in Chapter 2 to assess the snowfall patterns following different snowstorm types and the synoptic conditions leading up to the storms.
Snowfall patterns across Central New York vary depending on the type of storm, with localized patterns often following storms associated with lake-effect or lake-enhanced snow. These include lake-effect snowstorms (LES – H and LES – UL), Canadian lows
(Hudson lows, clippers, and Great Lakes lows), upper disturbances, and cold fronts. The localized nature of these storms is reflected by their relatively large range in average snowfall totals (Table 4.1). Although the average maximum snowfall for most of these storms is greater than 22 cm, some locations within the study area average less than 5 cm of snow. Heavier snowfall totals are generally concentrated over or near the Tug Hill, with lower totals often in southern and eastern Central New York. Results suggest that the exact location of the heaviest snowfall is tied to the synoptic conditions and air trajectories of the storm.
156
A common synoptic pattern associated with localized snowfall is a surface high pressure over the United States and a surface low pressure over northeastern Canada. The exact location and strength of these pressure patterns dictate the snowstorm type. For example, lake-effect snowstorms are common when a well-organized high-pressure (> 1023 hPa) is situated over the Mississippi River Valley and a well-organized low (< 1005 hPa) is centered over the Labrador Sea (Figure 4.4). This is consistent with previous findings (Ellis and Leathers 1996; Leathers and Ellis 1996; Liu and Moore 2004), as Suriano and Leathers
(2017b) classified these storms as WSW2 and W2 lake-effect storms. If the high and low pressures are slightly displaced, snowstorms can still occur, but they are often no longer
‘pure’ lake-effect storms. Canadian lows, for example, are favored if the high pressure is displaced west of the Rocky Mountains, with the low pressure located over the Hudson Bay,
Labrador Sea, or New England (Figure 4.2). The high pressure during non-cyclonic storms is also displaced further west over the North American Great Plains, with a secondary high over the southeastern United States.
Regardless of where the pressure systems are, localized heavy snowfalls are consistent with surface air temperatures below 8.5C and enhanced (≥10 m s-1) 850 hPa winds over the Great Lakes and Mid-Atlantic regions. These are two of the conditions necessary for the production of lake-effect and lake-enhanced snowfall (Holroyd 1971; Laird et al. 2009a,b;
Kristovich et al. 2018), which have been shown to cause localized snowfall patterns (Niziol
1987; Ellis and Leathers 1996; Ballentine et al. 1998). Therefore, storms that produce localized snowfall patterns likely have concurrent lake-effect or lake-enhanced snowfall.
The influence of lake snowfall is supported by air trajectory patterns, as storms with
157 localized snowfall often travel directly over Lake Ontario, with winds blowing parallel to the lake’s axis. Jiusto and Kaplan (1972) and Sousounis (2001) suggest that the development of lake-effect snow downwind of the Great Lakes is more likely with a longer fetch over the Great Lakes. Air trajectory patterns for these heavy-snowfall storms also often pass over multiple lakes prior to reaching Central New York, which Laird et al. (2017) suggests can lead to L2L snowbands and anomalously heavy (> 25.4 cm) snowfall totals.
Therefore, the air trajectories of these storms combined with the ingredients necessary for lake-effect snow are likely responsible for the localized snowfall patterns.
Snowfall patterns after large cyclonic storms (Nor’easters and Rocky lows) are more regional, with heavier snowfall totals throughout Central New York (Table 4.2). The average snowfall following these storms is significantly ( < 0.05) greater than that following storms with a more localized snowfall pattern (Table 4.1). The larger mean is likely due to the smaller range in snowfall across the study area, with average maximums between 16.9 – 25.2 cm and average minimums between 7.6 – 14.9 cm. The widespread snowfall may be tied to an omega pattern in the atmosphere, with high pressure systems located near 30N in the Atlantic and Pacific Oceans, and a third high pressure over the north-central United States (Figures 4.8 and 4.11). Barriopedro et al. (2006) show that omega patterns favor strong meridional flow in the jet stream, which increases the cyclogenesis of storms (Clark 1990; Whittaker and Horn 1981). The stronger the meridional flow, the more likely heavy snowfall producing cyclonic storms track across
Central New York. The location and strength of the central high pressure influences the trajectory of storms. When the central high is weaker, zonal flow is common bringing air
158 from the west and favoring Colorado lows over Texas hooks and Oklahoma hooks (Figure
4.8). This may explain the lower snowfall totals in southern Central New York following
Colorado lows as compared to the passage of Texas and Oklahoma hooks. When the central high is stronger and displaced further east, Nor’easters are favored over Rocky lows
(Figure 4.6). This creates a greater interaction between the central high and the cyclone resulting in northerly winds advecting cold air over the Great Lakes (Figure 5.13). This advection can result in lake-enhanced snow during the storm (Niziol 1987; Liu and Moore
2004), which likely explains the concentration of heavy snowfall to the east and southeast of Lake Ontario following Nor’easters (Figure 4.5).
Another objective of this chapter was to determine how the air trajectories influence the distribution and magnitude of snowfall following a storm. Typically, a greater fetch over
Lake Ontario means more snowfall in Central New York. This is most noticeable for storms with a localized snowfall pattern. The average trajectories for heavy-snowfall lake-effect snowstorms have more of a westerly component than that of moderate and light-snowfall producing storms (Figure 4.3). The westerly winds lead to a greater fetch, which often result in heavier snowfall totals downwind of the lake (Jiusto and Kaplan 1972). The wind direction is also shown to influence the location of the heaviest snow. For example, when air trajectories are from the west-southwest, the greatest snowfall following Great Lakes lows is over the Tug Hill (Figures 4.1 and 4.3); but as air trajectories shift to the west- northwest, the heaviest snowfall moves progressively to the south. In the case of lake- effect snowstorms, as air trajectories shift from the west-northwest to the northwest, the heaviest snowfall again moves progressively south and is also more widespread. Unlike
159 winds from the west, winds from the northwest typically only pass over Lake Ontario, which is not conducive for multi-lake interactions. Research suggests that multi-lake interactions are capable of producing lake-enhanced snow, and heavier snowfall totals downwind of the eastern Great Lakes (Mann et al. 2002; Lang et al. 2018; Rodriguez et al.
2007; Laird et al. 2017). Therefore, the occurrence of L2L snowbands may be partly responsible for heavy, localized snowfall patterns, while lighter, widespread snowfall may be lacking that multi-lake interaction.
Although less influential, air trajectories appear to affect the magnitude of large-cyclonic storms as well. Air trajectories for heavy-snowfall east coast storms typically move across
Lake Ontario, while moderate and light-snowfall storms do not (Figure 4.5). Research suggests that lake-effect and lake-enhanced snowfall is often linked to the passage of cyclonic storms, increasing snowfall totals downstream of the Great Lakes (Niziol 1987; Liu and Moore 2004; Mercer and Richman 2007). Therefore, greater snowfall totals from these storms may coincide with lake-enhancement. This conclusion is supported by snowfall patterns following heavy-snowfall east coast storms as the heaviest totals are adjacent to the southeast shore of Lake Ontario, rather than in southeastern Central New York, which is closest to the central low pressure of the storm (Figure 4.5).
4.5 Conclusion
The purpose this chapter was to identify the synoptic conditions and air trajectories that favor the occurrence of different snowstorm types in Central New York and associate these with snowstorm magnitude (heavy, moderate, or light). The results suggest that surface
160 pressures have a considerable influence on the types of snowstorms to affect Central New
York, their trajectories, and their snowfall amounts and patterns. An omega blocking pattern over North America plays an important role in producing widespread heavy snowfall across Central New York; while localized snowfall seems to be tied to a well- organized high pressure near the Mississippi River Valley and a low near the Labrador Sea.
As many researchers have suggested, upper atmospheric heights, winds, and temperatures are being modified by anthropogenic climate change (IPCC 2018). This has led to changes to the strength, location, and permanence of many high- and low-pressure systems since the early 20th century (Scaife et al. 2010; Barriopedro et al. 2006; Rohrer et al. 2018). For example, warming in the Arctic and over Alaska has increased the frequency and duration of omega blocking patterns over North America (Barriopedro et al. 2006). From the results presented, a stronger omega pattern can have multiple consequences on Central New York, such as more frequent Nor’easters and Rocky lows. If the central high is located to the south/southwest of Central New York, colder air may advect over the Great Lakes and into
Central New York for an extended time, increasing the likelihood for lake-effect snow.
Thus, this chapter helps determine the synoptic conditions that promote different storm types and magnitudes in the eastern Great Lakes region, and aids in the prediction of future changes to storm magnitude and frequency. At a broader context, this work provides a methodology for assessing the synoptic conditions which promote different snowstorms in a region, and more importantly showcases the commonalities that exist between storms and provides a foundation for assessing how future climate scenarios may alter the occurrence and characteristics of these storms.
161
5.0 SEASONAL TRENDS IN SNOWFALL CONTRIBUTIONS FROM DIFFERENT SNOWSTORM TYPES AND THE ENVIRONMENTAL FACTORS INFLUENCING THOSE TRENDS IN CENTRAL NEW YORK
5.1 Introduction
The Laurentian Great Lakes have a profound influence on the regional climate of the Great
Lakes region, especially during winter (Call 2005; Schmidlin 1993; Monmonier 2012).
However, as the climate changes, so too does the influence of the lakes. Previous research suggests that seasonal snowfall totals have increased within the Great Lakes region since the early 20th century, while snowfall trends in areas less prone to lake-effect snow were smaller, if not negative (Burnett et al. 2003; Hartnett et al. 2014; Norton and Bolsenga
1993; Kunkel et al. 2009a; Leathers and Ellis 1996). The assumption is that increasing snowfall from lake-effect snowstorms is largely responsible for the increasing snowfall within the Great Lakes region. However, this assumption was made on the basis of circumstantial evidence rather than through the examination of lake-effect snowfall trends versus snowfall trends of other snowstorm types. This chapter is a direct effort to estimate long-term trends in snowfall using observational data from the 1985/86 to the 2014/15 snow seasons, and the storm classification developed in the previous chapter. These long- term trends are further analyzed to determine if they are unidirectional or periodic. Mixed effects modeling and the exploratory application of the AIC enable the integration of changes in the ambient environmental conditions (e.g. air temperature, precipitation, lake temperature) that explain seasonal snowfall variability.
Recently, global climate models have been used to project seasonal snowfall totals within the Great Lakes region throughout the 21st century (Kunkel et al. 2002; Notaro et al. 2013b,
162
2015; Suriano and Leathers 2016). Seasonal snowfall totals are projected to decrease by the mid to late-21st century, largely due to an expected decrease in seasonal snowfall from lake-effect snowstorms. Findings suggest that seasonal snowfall totals have already begun to decrease in some parts of the Great Lakes region since the mid to late-20st century (Bard and Kristovich 2012; Hartnett et al. 2014). Bard and Kristovich (2012) found a significant trend reversal in seasonal snowfall totals to the lee of Lake Michigan. From 1920-1980 there was a positive snowfall trend, whereas from 1980-2005 there was a negative snowfall trend. Similarly, Hartnett et al. (2014) found a significant increase in seasonal snowfall totals downwind of Lake Ontario from 1931-1971, and a decrease in seasonal snowfall from 1971-2011. It is suggested that the decreasing snowfall is due to a negative trend in lake-effect snowfall, possibly because of increased temperatures over the Great
Lakes.
Seasonal snowfall patterns within the Great Lakes region are closely tied to air temperatures, lake surface temperatures, and ice cover on the lake (Tsuboki et al. 1989;
Segal and Kubesh 1996; Hanson et al. 1992; Wang et al. 2012; Notaro et al. 2015). Since the
1980s, regional warming within the Great Lakes region has led to alterations in seasonal water levels, precipitation and evaporation patterns, water temperatures, and winter ice extent and thickness (Bolsenga and Norton 1993; Dietz and Bidwell 2011; Vavrus et al.
2013). This warming, largely due to anthropogenically sourced greenhouse gases in the atmosphere (IPCC 2013), is likely responsible for some of the variability in seasonal snowfall totals within and outside of the Great Lakes region. Since the formation of lake- effect snowstorms and non-lake-effect snowstorms are fundamentally different, a warming
163 climate may have contrasting influences on these storms. Within the Great Lakes region, the lakes are warming faster than winter air temperatures (Lofgren 2004; Trumpickas et al. 2009; Dietz and Bidwell 2011). Since warmer lake temperatures have decreased ice thickness and extent, while relatively stagnant air temperatures have allowed temperatures to remain below freezing (Wang et al. 2012), there has been an observed increase in the transfer of heat and moisture from the lakes to overlying air masses throughout winter (Cordeira and Laird 2008; Zulauf 2003; Wright et al. 2013). Research suggests that this increased transfer is potentially responsible for increases in lake-effect snowfall (Tsuboki et al. 1989; Segal and Kubesh 1996; Wright et al. 2013), leading to the increasing seasonal snowfall totals in the Great Lakes region (Burnett et al. 2003; Hartnett et al. 2014; Norton and Bolsenga 1993; Kunkel et al. 2009a; Leathers and Ellis 1996).
The purpose of this chapter is to examine changes in snowfall from different storm types in an effort to resolve whether the increase in snowfall since the early-20th century within the
Great Lakes region is attributable to increasing snowfall from lake-effect snowstorms or from another snowstorm type. This study better resolves seasonal snowfall trends because it directly analyzes snowfall from different snowstorm types outlined in previous chapters, instead of estimating changes based on circumstantial evidence. Variability in seasonal snowfall trends are also examined for each snowstorm type within Central New York, to determine whether prominent trend reversals similar to those presented by previous studies occur for storms linked to lake-effect and lake-enhanced snowfall. Finally, I observe the influences of air and lake conditions on seasonal snowfall totals from different
164 snowstorm types to determine which variables most influence snowfall totals, so that it can be determined how snowfall may change based on future climate scenarios.
5.2 Methods
To address the research questions, seasonal snowfall trends are examined for the five snowfall subregions referenced in Chapter 3 (see Figure 3.1). Seasonal snowfall data are analyzed for lake snowstorms, non-lake snowstorms, and the five general snowstorm types from 1985/86 – 2014/15 (Table 3.1). Seasonal snowfall data are also analyzed for the contributing storms (Table 3.1), with details of these data presented in Chapters 2 and 3.
To address whether seasonal snowfall totals for each snowstorm type within the five subregions have changed over time, and since data were normally distributed with equal variance, temporal snowfall trends were calculated using simple linear regressions. To detect non-linearity in the seasonal snowfall trendlines seven-year trends with a 95% confidence were calculated with a one-year moving window. Trends were calculated from the 1985/86 season through the 2008/09 season and did not include the entire study period because after the 2008/09 season, there were no longer seven years of data to conduct the trend analyses. Periods in which trends significantly ( < 0.05) changed from positive (negative) values to negative (positive) values were identified as trend reversals
(Bard and Kristovich 2012; Hartnett et al. 2014). A trend reversal in the data suggests that long-term (30-year) snowfall trends may poorly reflect snowfall variability and should be further investigated.
165
Environmental forcings shown to influence seasonal snowfall totals within the Great Lakes region include air temperatures (Groisman and Easterling 1994; Mote et al. 2005), precipitation totals (Hanson et al. 1992; Sellinger et al. 2008; Trumpickas et al. 2009), and lake surface temperatures and ice cover (Bolsenga and Norton 1993; Dietz and Bidwell
2011; Vavrus et al. 2013). Therefore, in this study I examine the influence of environmental forcings on seasonal snowfall totals for each snowstorm type in Central
New York using fixed-effects models. The environmental forcings examined are described in Chapter 2 and include Lake Ontario winter and seasonal surface temperatures, Lake Erie winter and seasonal surface temperatures, the Great Lakes winter and seasonal surface temperatures, days with a minimum temperate 0C, days with a maximum temperature
0C, days with a minimum temperature of at least -17.8C, average winter and seasonal temperatures, and average maximum temperature. To remove any potential trends or bias in the data, variables were detrended and tested for collinearity using Pearson correlations
(Appendix 9.2). If two or more explanatory variables were highly correlated (r > 0.60), then the most significant variable was used in the model development (Yoo et al. 2014).
Since data were normally distributed (Kolmogorov-Smirnov test) with equal variance
(Bartlett test), linear fixed-effects models were used to model the influence of the environmental factors on seasonal snowfall totals and contributions. F-tests were used to determine which environmental variables explained significant (ρ ≤ 0.10) variance within the models. If more than one variable significantly influenced seasonal snowfall totals for a storm type, then predictor models were created for each combination of the significant variables. The relative importance of each model was compared using AIC (Chowdhury and Sharma 2009; Kharin and Zwiers 2002; Woolhiser 2008), which is described in
166
Chapter 2. The significance () and fit (R2) of the top model was then extracted for each snowstorm type. This type of model selection has been frequently employed in biological studies (e.g. Posada and Buckley 2004; Arnold 2010), but seldomly used in atmospheric studies (e.g. Armal et al. 2018; Reeves et al. 2007). Wong et al. (2014) suggest that the AIC can help reduce bias introduced by random small-scale variability introduced when working with station data. Therefore, this study not only examines the influence of different environmental forcings on seasonal snowfall totals, but the applicability of the AIC to examine these influences.
5.3 Results and Analyses
5.3.1 Modeling the Secular Trends in Seasonal Snowfall Contributions
Seasonal snowfall trends for different snowstorm types occurring during the period from
1985/86 – 2014/15 are shown in Table 5.1. Findings suggest that total seasonal snowfall in Central New York did not significantly (ρ > 0.05) change, corroborating previous findings that snowfall in the Great Lakes region has remained relatively steady since the late-1970s
(Bard and Kristovich 2012; Hartnett et al. 2014). Although total seasonal snowfall did not significantly change in any of the five subregions, there were significant changes in the seasonal snowfall from some of the individual snowstorm types.
Overall, storms with significant (ρ < 0.05) trends in seasonal snowfall totals were those most associated with lake-enhanced or lake-effect snow (Tables 5.1). In fact, there were no significant (ρ < 0.05) trends in seasonal snowfall totals from non-lake snowstorms for any of the five subregions. In addition, there were no significant (ρ < 0.05) snowfall trends for
167
any of the cyclonic storms forming south of Canada. This is similar to previous findings
which suggests that although seasonal snowfall totals significantly changed for areas
influenced by lake-effect snow, there was no significant trends in snowfall for areas not
influenced by lake-effect snow (Burnett et al. 2003; Kunkel et al. 2009a; Krasting et al.
2013).
Table 5.1. Linear regression results for seasonal snowfall totals (cm yr-1) for different snowstorm types within Central New York from 1985/86 – 2014/15 for the five snowfall subregions. Region 1 Region 2 Region 3 Region 4 Region 5 Storm Type Trend Error Trend Error Trend Error Trend Error Trend Error Total Snow -0.07 1.45 0.65 1.26 1.65 1.38 -0.73 1.66 2.01 2.02 Lake Snowstorms 0.34 0.63 0.29 0.45 1.36* 0.70 0.29 1.16 1.00 1.41 Non-Lake Snowstorms -0.41 1.18 0.36 1.08 0.29 1.12 -1.02 1.15 1.02 1.22 Lake-Effect Storms 0.27 0.56 1.20* 0.58 1.62* 0.02 0.73 1.03 1.64 1.24 Nor'easters -0.53 0.87 -0.18 0.83 -0.18 0.90 0.57 0.78 0.08 0.66 Canadian Lows -0.49 0.30 -0.35 0.26 -1.05** 0.38 -1.05* 0.49 -0.51 0.50 Clippers -0.57* 0.26 -0.51* 0.23 -0.94** 0.30 -1.03** 0.10 -0.72* 0.30 G.Lakes Lows 0.12 0.10 0.14 0.11 -0.01 0.17 0.12 0.28 0.39 0.28 Hudson Lows -0.04 0.04 0.02 0.08 -0.10 0.08 -0.15 0.15 -0.18 0.19 Rocky Lows 0.19 0.28 0.24 0.39 0.21 0.50 0.24 0.48 0.38 0.37 Colorado Lows 0.18 0.26 0.01 0.39 0.24 0.32 0.16 0.34 0.27 0.28 Oklahoma Hooks -0.10 0.19 0.09 0.21 -0.01 0.32 0.01 0.30 0.00 0.22 Texas Hooks 0.10 0.18 0.15 0.23 -0.02 0.22 0.07 0.25 0.10 0.22 Non-Cyclonic 0.14 0.26 0.32 0.24 0.11 0.25 0.23 0.38 0.40 0.38 Frontal Storms -0.18 0.22 0.03 0.24 -0.07 0.18 -0.07 0.31 0.10 0.29 Upper Disturbances 0.32 0.30 0.29 0.24 0.18 0.31 0.30 0.45 0.30 0.47 *ρ < 0.05; **ρ < 0.01; df = 29
The most noticeable decrease in seasonal snowfall totals was from Canadian lows,
particularly clippers (Table 5.1). Seasonal snowfall totals from clippers significantly (ρ <
0.05) decreased in every region, with the largest decreases in Regions 4 (-1.03 ± 0.10 cm yr-
1) and 5 (-0.72 ± 0.30 cm yr-1). Most regions also experienced a decrease in seasonal
168 snowfall from Hudson lows; however, this decrease was not (ρ > 0.05) significant. There was an increase in snowfall from Great Lakes lows in most regions, however none of the trends were significant (ρ < 0.05).
The expectation that a warming climate will lead to more lake-effect snowfall during the first half of the 21st century was not reflected in the regression analyses for the lake-effect snowbelt, Regions 4 and 5 (Tables 5.1). There was a significant (ρ < 0.05) increase in seasonal snowfall totals from lake snowstorms in Region 3 (1.36 ± 0.70 cm yr-1), and from lake-effect snowstorms in Region 2 (1.20 ± 0.58 cm yr-1) and Region 3 (1.62 ± 0.02 cm yr-1).
Interestingly, these two regions are outside of the lake-effect snowbelt due to their distance from and orientation to Lake Ontario. Although this supports the conclusions of previous research that lake-effect snow is increasing in lake-effect dominated areas due to a greater heat and moisture transfer between abnormally warm lakes and overlying air masses
(Notaro et al. 2015; Suriano and Leathers 2016), it also suggests that snowfall is not increasing equally in all areas that experience lake-effect snow. In fact, snowfall in areas less associated with lake-effect snowstorms is increasing more than snowfall in the traditional lake-effect snowbelt.
5.3.2 Trend Reversals
The purpose of this section is to examine the thirty-year seasonal snowfall trends calculated in the previous section for potential trend reversals. The results from the seven- year trends with a one-year moving window are shown in Figures 5.1-5.6. Trends in seasonal snowfall totals from lake snowstorms suggest that trend reversals occurred in
169
Central New York from 1985/86 – 2008/09 (Figure 5.1). The most prominent trend reversals were in Region 3, as trends were significantly ( < 0.05) positive during the late-
1980s, then significantly ( < 0.05) decreased in the early-1990s, followed by a significant
( < 0.05) increase in the late-1990s. Similar trends for lake snowstorms were also shown in Regions 2, 4, and 5; however, the trend reversals for these regions were not significant
( > 0.05). All three regions observed a significant ( < 0.05) increase in snowfall from
1990/91 – 1996/97, and non-significant ( > 0.05) snowfall trends from 1993/94 –
1999/00. Seasonal snowfall trends in Region 1 significantly ( > 0.05) decreased from
1993/94 – 1999/00.
There were no significant ( > 0.05) trend reversals in the seven-year snowfall trends of non-lake snowstorms for any of the five subregions (Figure 5.1). Although there were no significant trend reversals, there was a significant ( < 0.05) negative trend in snowfall from non-lake snowstorms in Region 4 from 2000/01 – 2006/07. Snowfall trends for all other years were not statistically significant. There was also a significant ( < 0.05) positive trend in seasonal snowfall for non-lake snowstorms in Region 5 from 1996/97 –
2002/03.
Five different snowstorm types affecting Central New York exhibited a significant ( <
0.05) trend reversal in seasonal snowfall totals from 1985/86 – 2008/09 (Figure 5.2-5.6).
Significant trend reversals for lake-effect snowstorms occurred in Regions 3 and 5 (Figures
5.4b and 5.6b). In both subregions, snowfall trends went from significantly positive in the early 1990s to significantly negative from 1993/94 – 1998/99, and then positive again in 170 the late-1990s. Snowfall trends in Regions 2 and 4 were significantly positive from the late-1980s to the early 1990s, but were not significant ( > 0.05) afterwards (Figures 5.3b and 5.5b). Region 1 lacked any notable trend reversals, similar to the results for lake snowstorms (Figure 5.2b).
Region 1
Region 2
171
FIGURE 5.1 CONTINUED Region 3
Region 4
Region 5
Figure 5.1. Seven-year seasonal snowfall trends from 1985/86 – 2008/09 for lake snowstorms and non-lake snowstorms.
172
173
FIGURE 5.2 CONTINUED
174
FIGURE 5.2 CONTINUED
Figure 5.2. Region 1 seven-year seasonal snowfall trends from 1985/86 – 2008/09 for Canadian lows (a), clippers (aa), Great Lakes lows (ab), Hudson lows (ac), lake-effect snowstorms (b), non-cyclonic snowstorms (c), frontal storms (ca), upper disturbance storms (cb), Nor’easters (d), Rocky lows (e), Colorado lows (ea), Oklahoma Hooks (eb), and Texas Hooks (ec).
175
FIGURE 5.3 CONTINUED
176
FIGURE 5.3 CONTINUED
Figure 5.3. Region 2 seven-year seasonal snowfall trends from 1985/86 – 2008/09 for Canadian lows (a), clippers (aa), Great Lakes lows (ab), Hudson lows (ac), lake-effect snowstorms (b), non-cyclonic snowstorms (c), frontal storms (ca), upper disturbance storms (cb), Nor’easters (d), Rocky lows (e), Colorado lows (ea), Oklahoma Hooks (eb), and Texas Hooks (ec).
177
FIGURE 5.4 CONTINUED
178
FIGURE 5.4 CONTINUED
Figure 5.4. Region 3 seven-year seasonal snowfall trends from 1985/86 – 2008/09 for Canadian lows (a), clippers (aa), Great Lakes lows (ab), Hudson lows (ac), lake-effect snowstorms (b), non-cyclonic snowstorms (c), frontal storms (ca), upper disturbance storms (cb), Nor’easters (d), Rocky lows (e), Colorado lows (ea), Oklahoma Hooks (eb), and Texas Hooks (ec).
179
180
FIGURE 5.5 CONTINUED
181
FIGURE 5.5 CONTINUED
Figure 5.5. Region 4 seven-year seasonal snowfall trends from 1985/86 – 2008/09 for Canadian lows (a), clippers (aa), Great Lakes lows (ab), Hudson lows (ac), lake-effect snowstorms (b), non-cyclonic snowstorms (c), frontal storms (ca), upper disturbance storms (cb), Nor’easters (d), Rocky lows (e), Colorado lows (ea), Oklahoma Hooks (eb), and Texas Hooks (ec).
182
FIGURE 5.6 CONTINUED
183
FIGURE 5.6 CONTINUED
Figure 5.6. Region 5 seven-year seasonal snowfall trends from 1985/86 – 2008/09 for Canadian lows (a), clippers (aa), Great Lakes lows (ab), Hudson lows (ac), lake-effect snowstorms (b), non-cyclonic snowstorms (c), frontal storms (ca), upper disturbance storms (cb), Nor’easters (d), Rocky lows (e), Colorado lows (ea), Oklahoma Hooks (eb), and Texas Hooks (ec).
Significant trend reversals occurred for Canadian Lows in Regions 1, 4 and 5 (Figures 5.2a,
5.5a, and 5.6a). Snowfall from Canadian lows significantly increased from 1986/87 –
1992/93 and decreased from 1991/92 – 1997/98 in Region 1. Snowfall also significantly decreased from 1991/92 – 1997/98 in Regions 4 and 5; however, the trend reversal occurred later in the study period, as snowfall trends were significantly positive during the
1996/97 season and significantly negative in the early and mid-2000s. Although the
184
1991/92 snowfall trend was significant in Regions 2 and 3, no other trends were significant
(Figures 5.3a and 5.4a).
Significant trend reversals in the snowfall from upper atmospheric disturbances occurred in Regions 2 and 3 (Figures 5.3cb and 5.4cb), and in the snowfall from frontal storms for
Regions 1 and 5 (Figure 5.2ca and 5.6ca). Snowfall trends from upper atmospheric disturbances in both regions were positive in the early 1990s, then became negative in the mid-1990s, followed by a positive snowfall trend in the mid-2000s. Seasonal snowfall trends in snowfall from frontal storms in Regions 1 and 5 were significantly negative during the 1991/92 season and significantly positive in the 1996/97 season. An additional trend reversal in the snowfall from frontal storms occurred in Region 5, as trends were significantly positive during the 1986/87 season. Great Lakes lows also exhibited a significant trend reversal in Region 2 (Figure 5.3ab). Snowfall trends were negative during the 1993/94 season, then became positive in 1997/98, followed by a negative trend in
2001/02.
Although trend reversals did not occur for the remaining snowstorm types, multiple snowstorms had at least one season with a significant seven-year trend in seasonal snowfall. At least one significant trend occurred for clippers, Oklahoma hooks, and upper atmospheric disturbances in Region 1 (Figure 5.2). In Region 2, significant trends occurred for Canadian lows, clippers, lake-effect snowstorms, and non-cyclonic snowstorms (Figure
5.3). In Region 3, trends were significant for Canadian lows, Colorado lows, frontal storms,
Great Lakes lows, and Texas hooks (Figure 5.4). Significant trends were observed for
185 clippers, Colorado lows, frontal storms, Great Lakes lows, lake-effect snowstorms, non- cyclonic snowstorms, and upper atmospheric disturbances in Region 4 (Figure 5.5).
Finally, significant trends occurred in Region 5 for clippers, Colorado lows, Great Lakes lows, Oklahoma hooks, and upper atmospheric disturbances (Figure 5.6). The only snowstorms not to experience a significant seven-year trend in snowfall in any subregion were Hudson lows, Nor’easters, and Rocky lows. To determine the factors possibly responsible for the significant trends or lack thereof, found in this section, the influence of environmental variables on seasonal snowfall totals are examined.
5.3.3 Modeling the Environmental Effects on Snowfall Contributions
Results from the regression modeling show that lake temperatures explain significant variance in the seasonal snowfall totals for about half of the different snowstorm types to affect Central New York (Appendix 9.3). Lake surface temperatures have the greatest influence on seasonal snowfall totals from storms typically originating just to the lee of the
Rocky Mountains. For example, seasonal snowfall totals from Rocky lows in Regions 3-5 are significantly explained by surface temperatures of the Great Lakes. Seasonal snowfall totals for the majority of snowstorm types to affect Central New York are also significantly explained by the percentage of ice cover on the Great Lakes, with the greatest influence on snowfall totals from Nor’easters (Appendix 9.4). Seasonal air temperatures and precipitation patterns are also shown to influence seasonal snowfall totals for the majority of snowstorm types. The only seasonal snowfall totals not influenced by either air temperatures nor precipitation patterns are those from Oklahoma hooks, Texas hooks, and non-cyclonic storms (Appendix 9.5).
186
The results of the linear fixed-effects models for snowfall totals from storms explained by only a single variable are presented in Table 5.2. The results suggest that a single environmental variable significantly ( ≤ 0.05) explains the seasonal snowfall totals for four snowstorm types in Region 2, three storm types in Region 5, and only one storm type in Regions 1, 3, and 4. Seasonal snowfall from Colorado lows and non-lake snowstorms are significantly ( ≤ 0.05) explained by a single environmental factor in two different subregions, but only a single subregion for all other snowstorms. The only variable shown to significantly influence snowfall either in multiple regions or for multiple snowstorm types is the average winter (October – April) lake surface temperatures of Lake Ontario (E).
Winter lake surface temperatures of Lake Ontario are shown to significantly influence the seasonal snowfall totals from non-lake snowstorms in Region 2, and the total seasonal snowfall of Regions 1 and 2. Lake surface temperatures have the greatest effect on the seasonal snowfall totals in Region 2 (R2 = 0.38) (Table 5.2), as snowfall totals increase as winter lake surface temperatures decrease (r = -0.48) (Table 5.3). Winter lake surface temperatures also significantly influence seasonal snowfall totals in Region 1 (R2 = 0.25) in a manner similar to Region 2 (r = -0.44). In Region 2, winter lake surface temperatures significantly influence the amount of snowfall from non-lake snowstorms (R2 = 0.23), as snowfall totals increase as lake temperatures decrease (r = -0.31).
Seasonal snowfall totals from Colorado lows are significantly ( < 0.05) influenced by the average percentage of ice cover on Lake Ontario and the number of days the minimum temperature is below -17.8C (Table 5.2). There is an increase in snowfall from Colorado
187
Table 5.2. The modeled results for the influence of a single environmental variable on seasonal snowfall totals from different snowstorm types for the five subregions of Central New York. Significant models are denoted with an ‘*’ if ≤ 0.05 and ‘**’ if ≤ 0.01. Region 1 Region 4 Storm -value F-stat df R2 Variable Storm -value F-stat df R2 Variable CO Low 0.002** 38.51 5 0.86 H Frontal 0.016* 12.73 5 0.66 I Lake Snow 0.110 2.82 18 0.09 E Nor'easter 0.135 2.37 28 0.13 M LES 0.073 3.62 18 0.12 E OK Hook 0.352 0.91 18 0.00 E Non-Cyclonic 0.918 0.01 5 -0.20 I Rocky Low 0.488 0.50 18 0.49 E Nor'easter 0.301 1.33 5 0.05 G Upper Dist. 0.691 0.16 28 0.69 J Total Snow 0.015* 7.21 18 0.25 E Region 5 Region2 Storm -value F-stat df R2 Variable Storm -value F-stat df R2 Variable Canadian Low 0.377 0.81 28 -0.01 J Canadian Low 0.026* 5.57 28 0.14 O CO Low 0.065 3.68 28 0.08 Q Clipper 0.050* 4.21 28 0.10 M Lake Snow 0.896 0.02 28 -0.04 K CO Low 0.728 0.14 5 -0.17 H LES 0.001** 15.01 28 0.33 L Hudson Low 0.415 0.68 28 -0.01 U Non-Lake Snow 0.010** 7.59 28 0.19 Q Non-Cyclonic 0.887 0.02 5 -0.19 I Nor'easter 0.008** 8.28 28 0.20 R Non-Lake Snow 0.018* 6.79 18 0.23 E OK Hook 0.121 3.48 5 0.29 I Nor'easter 0.431 0.57 5 -0.08 G Total Snow 0.424 0.66 28 -0.01 K Rocky Low 0.900 1.80 28 0.03 K B - L.Ontario Avg. Temp O - Avg. Max Temp Total Snow 0.002** 12.60 18 0.38 E E - L.Ontario Avg. WTemp Q - Avg. Min Temp Upper Dist. 0.446 0.60 28 -0.01 J G - L.Erie Avg. % Ice Cover R - Avg. Min Winter Temp Region 3 H - L.Ontario Avg. % Ice Cover U - Seasonal Precip Storm -value F-stat df R2 Variable I - G.Lakes Avg. % Ice Cover J - Days Min. Temp < 0⁰C CO Low 0.017* 6.49 28 0.16 K K - Days Min. Temp < -17.8⁰C Nor'easter 0.269 1.55 5 0.08 G M - Avg. Temp OK Hook 0.541 0.39 18 -0.03 B
188
Table 5.3. Correlations between the seasonal snowfall totals from different snowstorm types (Storm) and the environmental parameters (Env. Variable) for models significantly explained by a single variable.
Region 1 Region 3 Storm – Env. Variable Correlation Storm –Env. Variable Correlation Colorado Low – Colorado Low – 0.87 0.44 L.Ontario Avg. Percent Ice Cover Days Min Temp. < -17.8C Total Snow – -0.44 Region 4 L.Ontario Avg. Winter Temp. Frontal – Region 2 -0.72 Great Lakes Avg. Ice Cover Canadian Low – -0.45 Region 5 Avg. Max Air Temp. Clipper – Lake-Effect Snow – -0.45 0.63 Avg. Seasonal Air Temp. Day Max Temp. < 0C Non-Lake Snow – Non-Lake Snow – -0.31 -0.38 L.Ontario Avg. Winter Temp. Avg. Min Temp. Total Snow – Nor'easter – -0.48 -0.56 L.Ontario Avg. Winter Temp. Avg. Winter Temp.
lows in Region 1 and Region 3, as the percent of Lake Ontario ice cover increases (r = 0.87) and the number of days with a minimum temperature below -17.8C increases (r = 0.44), respectively (Table 5.3). In Region 2, average maximum air temperatures (R2 = 0.14) and average air temperatures (R2 = 0.10) significantly influence snowfall totals from snowstorms. As average maximum air temperatures and average air temperatures decrease, the amount of snowfall from Canadian lows (r = -0.45) and clippers (r = -0.45) increases, respectively. Seasonal snowfall totals from frontal storms are significantly ( =
0.016) influenced by the average percent ice cover on the Great Lakes, as snowfall totals increase as ice cover decreases (r = -0.72). The seasonal snowfall totals of lake-effect snowstorms, non-lake snowstorms, and Nor’easters in Region 5 are all significantly influenced by environmental factors. Snowfall from non-lake snowstorms and Nor’easters increases as the average minimum air temperature (r = -0.38) and the average winter air
189 temperature (r = -0.56) decreases, respectively. Snowfall totals from lake-effect snowstorms increase as the number of days the maximum temperature is less than or equal to 0C increases (r = 0.63).
Results of the mixed-effects models for snowfall totals from storms explained by at least two environmental variables are presented in Table 5.4. The results suggest that seven different snowstorm types are significantly ( ≤ 0.05) influenced by at least two different environmental variables. Environmental variables significantly influence seasonal snowfall totals from clippers in three subregions, lake-effect snowstorms and upper disturbances in two subregions, and Canadian lows, Hudson lows, lake snowstorms, and non-lake snowstorms in one subregion. Seasonal snowfall totals for four different snowstorm types are significantly influenced by environmental variables in Region 3, three storm types in
Region 4, and only one storm type in Regions 1, 2, and 5.
The average seasonal (July – June) air temperature is found most often in significant models, as nine different top models contain this parameter (Table 5.4). The second most common parameter is the number of days the minimum air temperature drops below
-17.8˚C, which is found in six top models. The number of winter precipitation days is in five top models, the number of days the minimum air temperature drops below 0˚C is in four models, the number of days the maximum temperature is less than 0˚C and the average winter air temperature is in three models, Lake Erie winter lake surface temperatures, the winter surface temperatures of the Great Lakes, and average winter air temperatures are in
190
Table 5.4. AIC table for the influence of multiple environmental variables on seasonal snowfall totals from different snowstorm types within the five subregions of Central New York. Significant models are denoted with an ‘*’ if ≤ 0.05 and ‘**’ if ≤ 0.01. Region 1 Region4 Storm ⧍AICc Weight R2 p-value Variables Storm ⧍AICc Weight R2 p-value Variables 0.00 0.70 0.24 0.02* D 0.00 0.25 0.10 0.05* M Upper Disturbance 3.13 0.15 0.20 0.06 DJ Clipper 0.30 0.20 0.14 0.05* KM Region 2 1.80 0.10 null Storm ⧍AICc Weight R2 p-value Variables 0.00 0.49 0.25 0.01** MT Non-Lake Snow 0.00 0.69 0.22 0.00** J 0.36 0.41 0.20 0.01** M Lake-Effect Snow 1.90 0.27 0.22 0.01** JU 0.00 0.76 0.35 0.00** RT Total Region 3 2.71 0.20 0.25 0.00** T Storm ⧍AICc Weight R2 p-value Variables 0.00 0.71 0.22 0.00** T Canadian Low Clipper 0.00 1.00 1.00 0.00** KMT 2.14 0.24 0.22 0.00** KT 0.00 0.31 0.06 0.10 R Region 5 G.Lakes Low 0.22 0.28 0.10 0.09 RU Storm ⧍AICc Weight R2 p-value Variables 0.00 0.55 0.11 0.04* R 0.00 0.21 0.20 0.02* MU Hudson Low 0.20 0.20 null 0.04 0.21 0.16 0.02* M Clipper 0.00 0.61 0.55 0.00** CF 0.90 0.14 0.22 0.02* KMU Upper Disturbance 1.04 0.36 0.48 0.00** F 1.45 0.10 0.16 0.03* KM 0.00 0.58 0.13 0.03* J 0.00 0.41 0.09 0.10 E Lake Snow Texas Hooks 2.16 0.20 0.11 0.08 JV 0.26 0.36 null 0.00 0.75 0.31 0.00** L 0.00 0.50 null Total 2.39 0.23 0.29 0.01** KL Rocky Low 2.41 0.15 -0.04 0.56 F 0.00 0.44 0.11 0.04* L 2.79 0.12 -0.06 0.95 Q Lake-Effect Snow 1.96 0.16 null C - G.Lakes Avg. Temp M - Avg. Temp 0.00 0.50 Null D - L.Erie Avg. WTemp Q - Avg. Min Temp Rocky Low 1.86 0.21 0.00 0.34 C E - L.Ontario Avg. WTemp R - Avg. Min Winter Temp F - G.Lakes Avg. WTemp T - No. of WPrecip Days J - Days Min. Temp < 0⁰C U - Seasonal Precip K - Days Min. Temp < -17.8⁰C V - Winter Precip (cm) L - Days Max Temp < 0⁰C
191 two models, and average lake surface temperatures of the Great Lakes are in a single top model. Generally, in Regions 1, 2 and 3, snowfall totals are significantly ( ≤ 0.05) influenced by lake surface temperatures (C, D, and F) and air temperatures (K, M, R), especially winter temperatures below 0˚C (J and L). Only two models in these regions are significantly influenced by precipitation, clippers in Region 3 and lake-effect snowstorms in
Region 2. Five variables significantly influence snowfall totals in Regions 4 and 5: number of days the minimum temperature is less than -17.8˚C (K), seasonal average air temperatures (M), average minimum winter air temperatures (R), the number of winter precipitation days (T), and the total seasonal precipitation (U). Therefore, snowfall totals in Regions 4 and 5 seem to be more influenced by precipitation, average seasonal air temperatures, and air temperatures below -17.8˚C than Regions 1-3.
Correlations were used to determine the strength and direction of relationships between snowfall totals and environmental variables in significant ( ≤ 0.05) models (Table 5.4). It was determined that snowfall totals from upper disturbances in Region 1 increase when winter lake surface temperatures on Lake Erie decrease (-0.45) and the number of days the minimum temperature is below 0C increase (r = 0.16) (Table 5.5). Lake-effect snowfall totals in Region 2 also increase when there is an increase in the number of days the minimum temperature is below 0C (r = 0.45), but decreases when seasonal precipitation totals increase (r = -0.09). Snowfall totals from clippers increase in Region 3 when the number of winter precipitation days increase (r = 0.42), but the average seasonal air temperature decreases (r = -0.49). Snowfall totals decrease in Region 3 for Hudson lows and upper atmospheric disturbances when average winter air temperatures increase
192
(r = -0.31) and when the average seasonal temperatures (r = -0.46) and average winter temperatures (r = -0.62) of the Great Lakes increases, respectively. Snowfall from lake- effect snowstorms increase as the number of days the maximum temperature is less than or equal to 0C increase (r = 0.45).
Table 5.5. Correlations between the seasonal snowfall totals from different snowstorm types (Storm) and the environmental parameters (Env. Variable) for models significantly explained by at least two variables. Region 1 Region 4 Storm – Env. Variable Correlation Storm –Env. Variable Correlation Upper Disturbance – Clipper – -0.45 -0.47 L.Erie Avg. Winter Temp. Avg. Seasonal Air Temp. Upper Disturbance – Clipper – 0.16 0.01 Days Min Temp. < 0C Days Min Temp. < -17.8C Non-Lake Snow – Region 2 -0.35 Avg. Seasonal Air Temp. Lake-Effect Snow – Non-Lake Snow – 0.45 0.47 Days Min. Temp. < 0C Winter Days with Precip. Lake-Effect Snow – Total Snow – -0.09 -0.51 Seasonal Precip. Avg. Winter Air Temp. Total Snow – Region 3 0.53 Winter Days with Precip. Clipper – Canadian Low – 0.01 0.53 Days Min Temp. < -17.8C Winter Days with Precip. Clipper – Canadian Low – -0.49 0.19 Avg. Seasonal Air Temp. Days Min Temp. < -17.8C Clipper – 0.42 Region 5 Winter Days with Precip. Hudson Low – Clipper – -0.31 -0.53 Avg. Winter Air Temp. Avg. Seasonal Air Temp. Upper Disturbance – Clipper – -0.46 0.15 G.Lakes Avg. Temp. Seasonal Precip. Upper Disturbance – Clipper – -0.62 0.10 G.Lakes Avg. Winter Temp. Days Min Temp. < -17.8C Total Snow – 0.61 Days Max Temp. < 0C Total Snow – 0.38 Days Min Temp. < -17.8C Lake-Effect Snow – 0.45 Days Max Temp. < 0C
193
Correlations for Regions 4 and 5 suggest that snowfall from clippers increases as average seasonal air temperatures decrease in Region 4 (r = -0.47) and Region 5 (r = -0.53), and as seasonal precipitation totals (r = 0.15) and the number of days the minimum temperature is below -17.8C (r = 0.10) increases in Region 5. In Region 4, snowfall totals from non-lake snowstorms increase when average temperatures decrease (r = -0.35) and the number of winter precipitation days increases (r = 0.47); while totals increase from Canadian lows when the number of winter precipitation days (r = 0.53) and the number of days the minimum temperature drops below -17.8C increases (r = 0.19).
5.4 Discussion
Assessing trends in snowfall in order to project future scenarios is notoriously difficult
(Norton and Bolsenga 1993; Burnett et al. 2003; Ellis and Johnson 2004; Kunkel et al.
2009a,b; Knowles et al. 2006; Krasting et al. 2013; Janoski et al. 2018; Changnon 2018;
Groisman and Easterling 1994; Notaro et al. 2015). There are many interacting variables, often operating at different temporal and spatial scales, and there are data limitations that often determine the time scale over which trends are assessed (Kunkel et al. 2007, 2009c;
Bard and Kristovich 2012; Hartnett et al. 2014; Clark et al. 2016). The results presented in this chapter shed some light on these difficulties by assessing snowfall trends for each of the storm types that influence the five subregions of Central New York, and through the examination of a suite of environmental variables that affect snowfall.
Much of the research on snowfall trends in the Great Lakes and Eastern United States has applied long-term historic datasets to assess overall snowfall changes (Norton and
194
Bolsenga 1993; Burnett et al. 2003; Ellis and Johnson 2004; Kunkel et al. 2009a; Changnon
2018). Proxy measures are often used to assess changes in snowfall from lake-effect snowstorms versus non-lake-effect snowstorms by examining stations deemed influenced by lake-effect and those considered non-lake influenced. In this chapter, I use the storms classified in Chapter 2 to directly examine how seasonal snowfall totals have changed over time and space for each individual storm type.
During the study period there is a significant (ρ ≤ 0.05) increase in snowfall from lake snowstorms in Region 3, while there is no significant change in snowfall from non-lake snowstorms for any of the five subregions. This is consistent with previous findings that points to any changes being the result of lake-effect increases (Burnett et al. 2003; Kunkel et al. 2009a). However, the lake snowstorm changes observed in this study are unexpectedly constrained to a subregion not highly associated with lake-effect or lake- enhanced snowfall (Table 3.; Figure 3.8).
Comparing snowfall trends of the five general snowstorm types, lake-effect snowstorms were the only storm with a significant (ρ ≤ 0.05) positive trend during the study period.
Snowfall significantly increased in both Regions 2 and 3, again in subregions least associated with lake-effect snowfall. The non-significant trends in Regions 1, 4, and 5 are supported by Bard and Kristovich (2012) and Hartnett et al. (2014), both of which suggest that lake-effect snowfall in the Great Lakes region stopped significantly increasing following a trend reversal in the late-1970s to early 1980s. The conclusions of Lang et al.
(2018) may help explain the increase in lake-effect snowfall in these subregions, as multi-
195 lake interactions often produce heavier, more widespread snowbands capable of moving further inland. Therefore, if these events are occurring more frequently, they can produce more lake-effect snow further from the lake. However, the frequency of these events has yet to be examined, and therefore the snowfall increases in Regions 2 and 3 are subject to future research.
Snowfall from none of the other snowstorms significantly increased over the study period, but there was a decrease (ρ ≤ 0.05) in snowfall across all five subregions from clippers and from Canadian lows in Regions 3 and 4. This may be on account of a reduction in their frequency due to shifts in the jet stream driven by a warming Arctic, a hypothesis proposed by Changnon et al. (2006). However, the trend reversal analysis shows that although there was a decrease in the snowfall from Canadian lows and clippers, they also experienced trend reversals whereby a notable positive trend occurred in the mid-1980s followed by a reversal and steep negative trend in the early 1990s. In Canadian lows specifically, the positive trend continued into the late 1990s in Regions 4 and 5.
Why these trends are being observed, however, is not clear. Trumpickas et al. (2009) and
Lofgren (2004) suggest that Lake Ontario surface temperatures have increased since the mid-20th century, a factor that Notaro et al. (2015) and Suriano and Leathers (2016) suggest will lead to more lake-effect snowfall throughout the first half of the 21st century.
Typically, air temperatures are the controlling factor influencing global snowfall trends
(Kapnick and Delworth 2013; Krasting et al. 2013); however, the Great Lakes have an undeniable influence on its regional climate (Notaro et al. 2013a). As lake surface
196 temperatures warm, polar air masses advecting over the lake become more unstable, leading to higher clouds (Notaro et al. 2015), and a greater ability to move further inland as lower topographic features have less influence in producing orographic uplift (Veals and
Steenburgh 2015). This will enhance lake-effect snowfall further inland, in areas typically less prone to lake-effect snow (Minder et al. 2015). Therefore, in this chapter I examined how different storms are influenced by a suite of environmental variables.
The modeling results suggest that temperatures influence seasonal snowfall totals the most in Central New York similar to previous findings (Kapnick and Delworth 2013; Krasting et al. 2013), as it is incorporated in 21 of the top 35 significant models (Tables 5.2 and 5.4).
Conditions within the Great Lakes themselves significantly influence snowstorms the most in Regions 1 and 3, with lake parameters in three of the top four models in Region 1 and two of the top ten models in Region 3. Overall, snowfall in Regions 1-3 is more influenced by lake conditions and air temperatures at or below 0C, while snowfall in Regions 4 and 5 are mostly influenced by average air temperatures and precipitation totals. Average seasonal air temperatures and average winter air temperatures have an inverse relationship with snowfall totals in Central New York. This was expected as colder air temperatures are necessary for the production of snow, as highlighted by the mid-1990s, which were anomalously cold and snowy in the Northeast United States (Kocin and
Uccellini 2004b). Freezing air temperatures are often only part of the conditions necessary for increased snowfall, as snowstorms not associated with lake-effect snow were often linked to precipitation totals. Intuitively, as precipitation days and totals increase, seasonal snowfall totals also increase. However, increased precipitation does not necessarily equate
197 to more snow, as exemplified by snowfall in January 2019, in which the Northeast was fairly wet, yet little snowfall occurred (NRCC 2019). This is because average air temperatures were above normal, producing more rain than snow.
There is a positive relationship between snowfall and the number of days the minimum temperature drops below -17.8C. These cold days are especially influential in Regions 1-3, for storms linked to lake-effect or lake-enhanced snowfall. Results suggest that even though lake-effect storms tend to occur during relatively cold seasons, for this snow to occur, air temperatures must be considerably colder than 0C. Previous studies have suggested this connection, as a greater temperature difference between the air and lake surface is more conducive for larger and more-organized lake-effect snowbands (e.g. Niziol
1987; Perry et al. 2007; Laird et al. 2009). However, this is the first study to examine how temperatures influence snowfall from different storm types, while accounting for spatial variability. The Great Lakes also significantly influence snowfall in Regions 1 and 3.
Interestingly, warmer lake surface temperatures signify less snowfall, but mostly from storms not highly linked to lake-effect snow. It is believed that the warmer lakes have less influence on the lake-air temperature difference, and instead indicates a warmer pattern that inhibits snowfall from cyclonic storms.
It is not well understood why there appears to be a regional discrepancy in the influence of environmental conditions. However, the environmental parameters shown to significantly influence snowfall totals may account for the trend reversals shown in this study. For example, the positive seven-year snowfall trends in lake-effect snowfall during the
198
1990/91 season in Regions 2 and 3 may be due to an increased number of days air temperatures dropped below 0C (r= 0.45 and r = 0.45, respectively) (Tables 5.3 and 5.5).
This is consistent with previous findings which suggest the mid to late 1990s were relatively cooler (Joyce 2002) and snowier (Kocin and Uccellini 2004b) than previous years in the eastern United States. The 2000s were a relatively warm period, and thus the positive snowfall trends during the 1997/98 season in Regions 3 and 5 were likely a product of warmer lake temperatures rather than cooler air temperatures, as suggested by
Notaro et al. (2015) and Suriano and Leathers (2016). This is further supported by the increase in lake-effect snowfall in Regions 2 and 3, as warming lake temperatures will favor well-developed, high intensity snowbands (Laird et al. 2009a; Veals and Steenburgh 2015).
These snowbands have the potential to extend across multiple lakes (Laird et al. 2017), which can extend their influences further from the lakes (Lang et al. 2018; Rodriguez et al.
2007).
5.5 Conclusion
The results presented in this chapter shed some light on previous observations of seasonal snowfall trends in the Great Lakes region (Norton and Bolsenga 1993; Ellis and Johnson
2004; Burnett et al. 2003; Kunkel et al. 2009a; Bard and Kristovich 2012; Hartnett et al.
2014). As previous research suggests, there has been an increase in lake-effect snow, but this increase is not uniform across the study area, and instead is disproportionately effecting areas further from the lake. The increase in lake-effect snowfall in these areas needs further investigation, but may be linked to warmer lake surface temperatures and an increase in L2L snowbands. These findings extend beyond Central New York as they
199 showcase the spatial variability of snowfall trends, and that the greatest increases in lake- effect snow in the future may be in areas on the outer edges of the traditional ‘lake-effect basin.’
As suggested by the literature (Kapnick and Delworth 2013; Krasting et al. 2013), air temperature appears to be the driving force behind changes in seasonal snowfall totals in
Central New York. However, freezing temperatures do not necessarily equate to more lake-effect snow, as snowfall in Regions 1-3 are closely linked to air temperatures dropping below -17.8⁰C. Therefore, as the climate changes there is a delicate balance for areas that experience lake-effect snow as to whether snow will increase or decrease. This balance depends on whether air temperatures warm beyond the threshold necessary to create the instability needed to form lake-effect snowbands, or if the warming disproportionately effects the lakes resulting in more lake-effect snow.
By examining snowfall trends associated with individual snowstorm types, a clearer picture emerges of how each individual storm has changed in both its snow contribution over time, and how it is linked to changes in other environmental conditions in the region.
In combination, these analyses help to refine seasonal snowfall estimates and provide better future projections of snowfall change in the Great Lakes region. There are other external variables, not examined in this chapter, that play a role in determining Great Lakes environmental conditions and the frequency of different storm types. An understanding of these may help elucidate some of the unexplained variance observed in this chapter,
200 therefore in Chapter 6 I will focus specifically on the interactions between teleconnection patterns and snowstorm contributions in Central New York.
201
6.0 THE INFLUENCE OF ATMOSPHERIC LOW-FREQUENCY VARIABILITY ON THE SEASONAL SNOWFALL CONTRIBUTIONS FROM DIFFERENT SNOWSTORM TYPES AFFECTING CENTRAL NEW YORK
6.1 Introduction
The IPCC (2013) considers increased anthropogenic greenhouse gases as the most significant external forcing on snowfall today. Warmer global temperatures and increased precipitation totals in the Great Lakes region and Northeast United States (Groisman and
Easterling 1994; Bolsenga and Norton 1993; Dietz and Bidwell 2011; Vavrus et al. 2013), are expected to have mixed influences on seasonal snowfall totals in the northern United
States (Suriano and Leathers 2016; Notaro et al. 2013b; Kunkel et al. 2002). In the previous chapter, I examined how seasonal snowfall totals have changed for different snowstorm types affecting Central New York from 1985/86 – 2014/15. However, natural variability in the climate has also been shown to influence seasonal snowfall totals in the
Great Lakes region and Northeast United States (Vavrus et al. 2013; Serreze et al. 1998; Ge and Gong 2009; Kunkel et al. 2009a; Mote et al. 2005). Therefore, I also examined the impact of several environmental variables on seasonal snowfall totals in the study area.
This did not consider the low-frequency atmospheric variability often associated with teleconnections, a factor that might account for the unexplained variance in the models of snowfall variability. In this chapter, I use mixed-effects modeling and the AIC to incorporate the influence of Atlantic, Pacific and Arctic teleconnection patterns on seasonal snowfall contributions. This research differs from previous research through its application of snowfall from specific storm types rather than proxy measures of lake-effect snow, and it applies the combination of mixed effect modeling and the AIC. This novel
202 technique is capable of evaluating the interaction and therefore combined effects of teleconnections.
Previous research suggests that the unaccounted variability in seasonal snowfall totals may be linked to low-frequency variability in the atmosphere and oceans, such as teleconnection patterns (e.g. Serreze et al. 1998; Ge and Gong 2009). Teleconnections reflect large-scale changes in the atmospheric wave and jet stream patterns and have been shown to influence seasonal snowfall totals in the northern United States through their influence on air temperatures and precipitation patterns, storm tracks, the jet stream intensity and location, and the characteristics of the Great Lakes (Serreze et al. 1998; Ge and Gong 2009; Groisman and Easterling 1994; Ghatak et al. 2010; Grise et al. 2013; Wise et al. 2015). Teleconnection patterns are an important part of the interannual and interdecadal variability of snowfall because they can persist for weeks to years and can span across the globe (Barnston and Livezey 1987). Several teleconnection patterns have been linked to seasonal snowfall variability within the Great Lakes region and Northeast
United States including the El Niño Southern Oscillation (ENSO), the North Atlantic
Oscillation (NAO), the Pacific North American (PNA) pattern, the Pacific Decadal Oscillation
(PDO), the Arctic Oscillation (AO), the East Atlantic (EA) pattern, and the West Pacific (WP) pattern.
Most studies have examined the influence of teleconnection patterns on seasonal snowfall totals using principal component analyses (e.g. McCabe and Dettinger 2002; Clark et al.
2016; Hawkins and Ellis 2002). Studies have shown that the El Niño (La Niña) phase of
203
ENSO favors anomalously low (high) seasonal snowfall totals in the Northeast and Great
Lakes regions of the United States (Serreze et al. 1998; Ge and Gong 2009; Ghatak et al.
2010; Allen and Zender 2011; Grise et al. 2013; Baxter et al. 2014; Yu et al. 2014; Gan and
Wu 2015; Grimaldi 2008). An inverse relationship has been illustrated between the NAO and snowfall in the Northeast United States, when positive (negative) index years often produce below (above) average snowfall totals due to above (below) normal air temperatures (Ghatak et al. 2010; Wise et al. 2015; Seager et al. 2010a). The positive
(negative) phase of the EA pattern is believed to cause below (above) normal temperatures in the eastern United States, increasing the likelihood for above (below) normal snowfall totals (Wise et al. 2015; Davis and Benkovic 1994; Seierstad et al. 2007). In the Pacific
Ocean, the positive (negative) phases of the PNA and WP have been linked to below
(above) average temperatures in the central and eastern United States and above-normal
(below-normal) seasonal snowfall totals (Ghatak et al. 2010; Leathers et al. 1991; Barnston and Livezey 1987; Wise et al. 2015; Tanaka et al. 2016). Research suggests that the positive phase of the PDO leads to above (below) average snowfall totals in the Northeast
United States due to larger troughs in the jet stream (Ge and Gong 2009; Gutzler et al. 2002;
McCabe and Dettinger 2002; Goodrich and Walker 2011). Finally, the negative (positive) phase of the Arctic Oscillation (AO) has been linked to anomalously cold (warm) air over
North American and above (below) average seasonal snowfall (Bai et al. 2012; Rohli and
Vega 2011; Zhu and Wang 2016).
The purpose of this chapter is to examine how teleconnection patterns influence the seasonal snowfall variability examined in Chapter 5. Although the influences of
204 teleconnection patterns on seasonal snowfall totals in North America have been widely studied (e.g. Patten et al. 2003; Hirsch et al. 2001; McCabe and Dettinger 2002; Groisman and Easterling 1994; Wise et al. 2015; Seager et al. 2010a; Kunkel and Angel 1999; Ghatak et al. 2010; Notaro et al. 2006; Grimaldi 2008), they have not been examined for their influence on seasonal snowfall totals from individual snowstorm types within the Great
Lakes region. In this study, I utilize linear fixed effect modeling to examine the influence of teleconnections on seasonal snowfall totals, a methodology that has not been used before in this context. The rationale for its application is that previous methods have produced varying and sometimes conflicting results. Since large-scale changes in the atmosphere and oceans can have global consequences, understanding the influences of teleconnections on seasonal snowfall totals produced by different snowstorm types may provide a better understanding of seasonal snowfall variability within the Laurentian Great Lakes region.
6.2 Methods
To determine the influence of teleconnections patterns on seasonal snowfall totals in
Central New York, snowfall totals for different snowstorm types affecting the five subregions were examined from 1985/86 – 2014/15. Seasonal snowfall totals were examined for lake snowstorms, non-lake snowstorms, and the five general snowstorm types (see Table 3.1). To remove potential trends or bias in the data, seasonal snowfall totals were detrended and the residuals were used for analysis. Since only one COOP station reported seasonal snowfall totals during the 1999/00 and 2004/05 seasons in
Region 3, data for these years were removed for all regions.
205
The teleconnection patterns used in this analysis are the AO, EA, ENSO, NAO, PDO, PNA and
WP, all of which are fully described in Chapter 2. Since the indices were normally distributed (Kolmogorov-Smirnov tests) and homoscedastic (Bartlett tests), linear fixed- effects models were applied to examine their relations with seasonal snowfall totals (see
Chapter 2). Although this type of model creation was used in the previous chapter and is commonly used in biological studies (e.g. Arnold 2010), few studies have used it to examine teleconnections (Giannini et al. 2001; Risbey et al. 2015; Shimura et al. 2013). Instead, principal component analyses have typically been used to examine the effects of teleconnections on different weather phenomena (e.g. McCabe and Dettinger 2002; Clark et al. 2016; Hawkins and Ellis 2002). However, the results from the principal components have not provided definitive links between teleconnections and seasonal snowfall totals in the United States, as some findings are contradictory (Ge and Gong 2009). Therefore, in this study I use linear fixed-effects models to examine whether they are able to provide more definitive relationships between different teleconnection patterns and seasonal snowfall totals from different snowstorm types.
Prior to developing the models, teleconnections were tested for collinearity using Pearson correlations (Appendix 9.6). If two or more teleconnection patterns had a significant (ρ <
0.05) correlation (> 0.60), then the variables were considered collinear and only the most significant variable was used in the model development (Yoo et al. 2014). Linear regression models incorporating every teleconnection were tested for the different snowstorm types in the five subregions, and F-tests were used to determine which teleconnections explained significant (ρ ≤ 0.05) variance within the models. Linear fixed-
206 effects models were then developed for each combination of the significant teleconnection patterns, with seasonal snowfall totals as the response variables. The relative importance of each model was compared using the AIC (further described in Section 2.6.1), and has been used to examine the influence of teleconnections on local climates (Chowdhury and
Sharma 2009; Kharin and Zwiers 2002; Woolhiser 2008). The top five models or the top models with a cumulative weight greater than 0.80 were recorded. A top model was subjectively chosen based on its weight, the number of explanatory variables, its fit (R2), and its significance (ρ) (Geyer 2003). For significant (ρ 0.10) models with an R2 greater than 0.15, the relationships (strength and direction) between different explanatory variables were examined using Pearson correlations (Baigorria and Jones 2010). The residuals of the correlations were examined, and if a pattern existed in the residuals, then the relationship was assumed non-linear and the appropriate transformation (e.g. square root, logarithm, reciprocal) was used to linearize the data. Once linearized, the correlations were then tested.
6.3 Results and Analysis
6.3.1 Linear Fixed-Effects Models
Results for the linear fixed-effects models are presented by subregion in Tables 6.1 – 6.5.
Model results are first presented for Region 1, where teleconnections were most effective
(R2 = 0.36, < 0.01) at explaining seasonal snowfall totals from Nor’easters (Table 6.1).
The relatively low fit of these models suggest that other variables significantly influence seasonal snowfall variability in this region. The AO (A), NIÑO 3 (C), NAO (F), and the PDO
(G) significantly influenced seasonal snowfall totals from Nor’easters in Region 1. The
207
Table 6.1. AIC table for the teleconnection predictors of seasonal snowfall totals for the different snowstorm types in Region 1. The table shows AIC values (AICc), differences in AIC values between the models (ΔAICc), Akaike weights (Weights), and the teleconnection models included in the analysis (Model Variables). Models in red signify the top model for that storm based on weight, p-value, and adjusted R2. Lake Snow Non-Lake Snow AICc ΔAICc Weights Model Variables AICc ΔAICc Weights Model Variables 469.7 0 0.455 Null 464.7 0 0.351 G 471.4 1.78 0.187 AI 464.9 0.16 0.323 CG 471.8 2.16 0.155 AC 466.0 1.27 0.186 FG 473.2 3.52 0.078 ACI 466.6 1.83 0.140 CFG 474.3 4.65 0.045 ABCI Adjusted R2 = 0.06 Adjusted R2 = 0.19 -value = 0.175 -value = 0.040 Canadian Lows Lake-Effect Snowstorms AICc ΔAICc Weights Model Variables AICc ΔAICc Weights Model Variables 182.9 0 0.288 F 422.6 0 0.278 Null 185.0 2.17 0.097 EF 423.8 1.26 0.148 I 185.3 2.44 0.085 ABF 424.1 1.54 0.129 A 185.4 2.55 0.080 FG 424.5 1.88 0.109 C 185.5 2.64 0.077 FH 424.6 2.05 0.1 F Adjusted R2 = 0.19 Adjusted R2 = 0.01 -value = 0.04 -value = 0.29 Non-Cyclonic Storms Nor'easters AICc ΔAICc Weights Model Variables AICc ΔAICc Weights Model Variables 207.9 0 0.200 G 440 0 0.411 FG 208.2 0.27 0.175 FG 442 2.01 0.15 AG 209.5 1.52 0.094 BG 442.9 2.89 0.097 AFG 210.0 2.06 0.071 AFG 443.1 3.08 0.088 G 210.1 2.16 0.068 EG 443.3 3.28 0.08 ACFG Adjusted R2 = 0.08 Adjusted R2 = 0.36 -value = 0.12 -value = 0.01 Rocky Lows AICc ΔAICc Weights Model Variables 212.4 0 0.282 Null 213.6 1.2 0.155 E 214 1.52 0.132 G 214.6 2.18 0.095 H 214.8 2.41 0.084 B Adjusted R2 = 0.01 -value = 0.28 A – AO; B – EA; C – NIÑO 3; D – NIÑO 4; E – NIÑO 3.4; F – NAO; G – PDO; H – PNA; I – WP
significant influence of these teleconnections on seasonal snowfall totals from Nor’easters is likely due to changes in location and speed of the polar jet stream (Baigorria and Jones
2010). Previous research suggests that frequent troughs (ridges) in the jet stream due to anomalously warm (cold) surface waters in the western Atlantic Ocean and Gulf of Mexico
208 increase (decrease) seasonal snowfall totals from Nor’easters in the eastern United States
(Hirsch et al. 2001; Mercer and Richman 2007). Therefore, the influence of the previous teleconnections on troughing in the jet stream may be responsible for the effects of these teleconnections on seasonal snowfall totals from Nor’easters (Zhang et al. 2000; Changnon et al. 2008; Bosart 1973; Kocin and Uccellini 2004b).
Teleconnections had less (R2 < 0.20) influence on the remaining snowstorm types in
Central New York. The only other seasonal snowfall totals that were significantly influenced by teleconnection patterns were from non-lake snowstorms and Canadian lows.
Seasonal snowfall totals from non-lake snowstorms were significantly influenced by the
ENSO (NIÑ0 3), PNA, and PDO, while snowfall totals from Canadian lows were influenced by the AO, EA, and NAO. Interestingly, snowfall from Canadian lows is most influenced by conditions in the Atlantic and Arctic, with little influence from conditions in the Pacific
Ocean. This is maybe due to the link between warm temperatures in the North Atlantic and
Arctic and troughs in the jet stream over the eastern United States which can lead to anomalously cold air temperatures and abnormally large snowfall totals (Ghatak et al.
2010; Wise et al. 2015; Seager et al. 2010a; Davis and Benkovic 1994; Seierstad et al. 2007).
Teleconnections had less of an influence on the seasonal snowfall totals from snowstorms in Region 2 compared to those in Region 1 (Table 6.2), as highlighted by the low modeled fits (R2 ≤ 0.30). Similar to Region 1, teleconnections had the greatest influence (R2 = 0.30) on the seasonal snowfall totals from Nor’easters. Totals were significantly influenced by
209
Table 6.2. AIC table for the teleconnection predictors of seasonal snowfall totals for the different snowstorm types in Region 2. The table shows AIC values (AICc), differences in AIC values between the models (ΔAICc), Akaike weights (Weights), and the teleconnection models included in the analysis (Model Variables). Models in red signify the top model for that storm based on weight, p-value, and adjusted R2. Lake Snow Non-Lake Snow AICc ΔAICc Weights Model Variables AICc ΔAICc Weights Model Variables 408.4 0 0.178 A 460.3 0 0.274 G 409.2 0.79 0.12 AI 460.7 0.47 0.216 FG 409.3 0.94 0.111 Null 461.3 1.08 0.16 AG 409.6 1.23 0.096 AC 461.7 1.47 0.131 Null 409.7 1.33 0.091 AG 462.1 1.82 0.11 BG Adjusted R2 = 0.12 Adjusted R2 = 0.13 p-value = 0.01 p-value = 0.05 Canadian Lows Lake-Effect Snowstorms AICc ΔAICc Weights Model Variables AICc ΔAICc Weights Model Variables 380.0 0 0.316 Null 403.8 0 0.336 I 381.1 0.34 0.266 G 404.7 0.88 0.217 AI 381.2 0.44 0.253 F 404.8 0.99 0.205 Null 382.1 1.30 0.165 FG 405.5 1.68 0.145 GI 406.3 2.48 0.097 BFI Adjusted R2 = 0.06 Adjusted R2 = 0.10 p-value = 0.18 p-value = 0.12 Non-Cyclonic Storms Nor'easters AICc ΔAICc Weights Model Variables AICc ΔAICc Weights Model Variables 229.8 0 0.329 Null 438.5 0 0.494 A 231.6 1.76 0.137 H 439.8 1.29 0.259 AFG 231.8 1.96 0.124 E 442.2 3.67 0.079 AFGH 232 2.16 0.112 F 442.4 3.9 0.07 AFGI 232.3 2.47 0.096 B 442.9 4.34 0.056 F Adjusted R2 = 0.01 Adjusted R2 = 0.30 p-value = 0.42 p-value = 0.01 Rocky Lows AICc ΔAICc Weights Model Variables 229.0 0 0.376 Null 229.2 0.13 0.353 AG 230.8 1.80 0.152 GHI 231.3 2.29 0.119 AGHI
Adjusted R2 = 0.14 p-value = 0.10 A – AO; B – EA; C – NIÑO 3; D – NIÑO 4; E – NIÑO 3.4; F – NAO; G – PDO; H – PNA; I – WP
the AO, NAO, and PDO, which was similar to Region 1, but without a significant influence from NIÑO 3. Seasonal snowfall totals from lake snowstorms and non-lake snowstorms were also significantly ( ≤ 0.05) influenced by teleconnection patterns. The NAO and PDO were significant influencers in Region 2, while the AO and WP influenced snowfall from
210 lake snowstorms. Seasonal snowfall totals from Rocky lows were influenced by the AO,
PDO, PNA and WP, but the model was not significant ( = 0.10). Interestingly, teleconnection patterns had a larger influence on the seasonal snowfall totals from lake snowstorms in Region 2 than Region 1. The WP significantly influences seasonal snowfall totals from lake snowstorms in Region 2, which few studies have noted. The influence of the WP may be due to the southern displacement of the jet stream during its positive phase
(Wise et al. 2015; Tanaka et al. 2016), as polar air is advected into Central New York.
Teleconnections had a prominent influence on seasonal snowfall totals from different snowstorm types in Region 3, as six out of the seven top models were significant ( ≤ 0.05) and five of them had a modeled fit (R2) greater than 0.25 (Table 6.3). Again, teleconnections had the greatest influence on seasonal snowfall totals for Nor’easters (R2 =
0.41); however, they also considerably influenced seasonal snowfall totals from non-lake snowstorms (R2 = 0.40). Every teleconnection observed in this study had a significant influence on seasonal snowfall totals from Nor’easters, while seasonal snowfall totals from non-lake snowstorms were most influenced by the PDO, PNA, and WP. Since Nor’easters are also sometimes classified as non-lake snowstorms, the influence of the PDO, PNA, and
WP on seasonal snowfall totals from non-lake snowstorms may derive from their influence on Nor’easters. However, this suggests that the Pacific Ocean has a greater influence on seasonal snowfall totals in Region 3 from non-lake snowstorms than the Atlantic and
Arctic. This corroborates previous findings suggesting that the frequency and severity of cyclonic storms originating over western North America is influenced by teleconnections
211
Table 6.3. AIC table for the teleconnection predictors of seasonal snowfall totals for the different snowstorm types in Region 3. The table shows AIC values (AICc), differences in AIC values between the models (ΔAICc), Akaike weights (Weights), and the teleconnection models included in the analysis (Model Variables). Models in red signify the top model for that storm based on weight, p-value, and adjusted R2. Lake Snow Non-Lake Snow AICc ΔAICc Weights Model Variables AICc ΔAICc Weights Model Variables 408.4 0 0.531 GI 441.9 0 0.309 G 409.3 0.92 0.335 CGHI 442.2 0.27 0.27 BG 412.9 4.53 0.055 I 443.5 1.59 0.14 GHI 413 4.6 0.053 Null 443.9 1.98 0.115 FG 414.5 6.12 0.025 BFI 444.2 2.27 0.099 AG Adjusted R2 = 0.31 Adjusted R2 = 0.40 p-value = 0.01 p-value = 0.00 Canadian Lows Lake-Effect Snowstorms AICc ΔAICc Weights Model Variables AICc ΔAICc Weights Model Variables 377.1 0 0.282 F 404.3 0 0.384 GI 377.3 0.24 0.25 G 405.5 1.23 0.208 I 377.9 0.84 0.185 AF 405.9 1.53 0.178 Null 379.1 2.08 0.099 EGHI 406.7 2.34 0.119 CGHI 408.2 3.85 0.056 CI Adjusted R2 = 0.29 Adjusted R2 = 0.15 p-value = 0.01 p-value = 0.05 Non-Cyclonic Storms Nor'easters AICc ΔAICc Weights Model Variables AICc ΔAICc Weights Model Variables 188 0 0.298 Null 421.9 0 0.55 G 189.4 1.37 0.15 B 424.8 2.95 0.126 ABCFGHI 189.5 1.5 0.141 F 425.4 3.52 0.095 AFG 190.3 2.26 0.096 E 426 4.15 0.069 AFGH 190.4 2.35 0.092 H 426 4.15 0.069 AFGI Adjusted R2 = 0.00 Adjusted R2 = 0.41 p-value = 0.31 p-value = 0.01 Rocky Lows AICc ΔAICc Weights Model Variables 228.4 0 0.441 FG 229.8 1.36 0.224 Null 229.9 1.44 0.215 AFG 231 2.6 0.12 AG
Adjusted R2 = 0.28 p-value = 0.02 A – AO; B – EA; C – NIÑO 3; D – NIÑO 4; E – NIÑO 3.4; F – NAO; G – PDO; H – PNA; I – WP
in the Pacific Ocean (Serreze et al. 1998; Ge and Gong 2009; Ghatak et al. 2010; Allen and
Zender 2011; Grise et al. 2013; Baxter et al. 2014; Yu et al. 2014; Gan and Wu 2015;
Grimaldi 2008). Model results also suggest that the AO, NAO, and PDO significantly ( ≤
0.01) influence seasonal snowfall totals from Rocky lows, again highlighting the importance
212 of the Pacific and Atlantic Oceans on snowstorms originating in western North America.
Seasonal snowfall totals from Canadian lows were most influenced by teleconnections in the Pacific Ocean, including the ENSO (EÑSO 3.4), PDO, PNA, and WP.
Finally, teleconnections patterns considerably influenced seasonal snowfall totals from lake snowstorms (R2 = 0.31) and lake-effect snowstorms (R2 = 0.41) within Region 3.
Interestingly, seasonal snowfall totals from lake snowstorms and lake-effect snowstorms were most influenced by teleconnection patterns in the Pacific Ocean, since seasonal snowfall totals for both snowstorm types were significantly explained by the PDO and WP.
In addition, seasonal snowfall totals from lake snowstorms were also significantly influenced by ENSO (NIÑO 3) and the PNA. This suggests that even though Region 3 is closer to the Atlantic, the Pacific’s influence on lake-enhanced and lake-effect snowfall is greater.
Teleconnections had a significant ( ≤ 0.05) influence on seasonal snowfall totals in Region
4 for five of the seven snowstorm types (Table 6.4). Unlike Regions 1-3, teleconnections in
Region 4 had the greatest influence on seasonal snowfall totals from lake snowstorms (R2 =
0.31), followed by Nor’easters (R2 = 0.26). The prominent influence of teleconnections on seasonal snowfall totals from lake snowstorms in Region 4 is especially important due to the frequent occurrence of lake-effect and lake-enhanced snowstorms affecting this region
(see Chapter 3). Model results suggest that the EA, PDO, and PNA all significantly influence seasonal snowfall totals from lake snowstorms and lake-effect snowstorms in Region 4.
Seasonal snowfall totals have been linked to variability in teleconnection patterns (e.g.
213
Grimaldi 2008; Serreze et al. 1998; Ge and Gong 2009; Ghatak et al. 2010; Allen and Zender
2011; Grise et al. 2013; Baxter et al. 2014; Yu et al. 2014; Gan and Wu 2015); however, this is the first direct evidence of a link between teleconnections and specifically lake-effect and lake-enhanced snowstorms. Interestingly, seasonal snowfall totals from these storms were most influenced by conditions in the Arctic and Pacific Ocean, which may be due to the advection of cold air into Central New York, or lack thereof, associated with meridional patterns in the jet stream (Wise et al. 2015).
Seasonal snowfall totals from Nor’easters, non-lake snowstorms, and Canadian lows in
Region 4 were also significantly influenced by teleconnections. Model results suggest that teleconnections were less influential on seasonal snowfall totals from Nor’easters in Region
4 compared to Regions 1-3, as the R2 value was considerably lower. In addition, the only teleconnections shown to influence snowfall from Nor’easters were the AO and WP. The greater influence of teleconnections on Nor’easters in Regions 1-3 may be due to the proximity of these regions to the storm’s center. Since these regions are further to the south, they tend to experience more direct influences from the central low pressure. In
Region 4 however, a considerable amount of snowfall from a Nor’easter is from lake- enhanced snow associated with the storm, as shown in Chapter 3. Therefore, it is suggested that teleconnections have a greater influence on the position or strength of the central low pressure of a Nor’easter than on the amount of lake-enhanced snow that it produces. The ENSO (NIÑO 3.4), PDO, and PNA had the greatest effect on seasonal snowfall totals from non-lake snowstorms in Region 4, while the NAO significantly ( ≤ 0.01) influenced totals from Canadian lows. Similar to other regions, seasonal snowfall totals
214 from non-lake snowstorms were most influenced by teleconnections in the Pacific, while the Atlantic had a greater influence on snowfall from Canadian lows.
Table 6.4. AIC table for the teleconnection predictors of seasonal snowfall totals for the different snowstorm types in Region 4. The table shows AIC values (AICc), differences in AIC values between the models (ΔAICc), Akaike weights (Weights), and the teleconnection models included in the analysis (Model Variables). Models in red signify the top model for that storm based on weight, p-value, and adjusted R2. Lake Snow Non-Lake Snow AICc ΔAICc Weights Model Variables AICc ΔAICc Weights Model Variables 459.3 0 0.252 Null 463.6 0 0.354 EGH 460.1 0.81 0.168 BGI 464.8 1.15 0.199 GH 460.5 1.16 0.141 I 465.1 1.52 0.166 H 461.6 2.26 0.081 GHI 466.6 3.02 0.078 Null 462.2 2.85 0.061 CGI 466.7 3.04 0.077 EH Adjusted R2 = 0.31 Adjusted R2 = 0.23 p-value = 0.01 p-value = 0.02 Canadian Lows Lake-Effect Snowstorms AICc ΔAICc Weights Model Variables AICc ΔAICc Weights Model Variables 418.1 0 0.498 F 455.2 0 0.226 I 420.1 2.04 0.179 A 456.2 1.02 0.135 BI 420.6 2.50 0.142 AF 456.5 1.29 0.119 GI 422.2 4.09 0.065 AFG 456.6 1.43 0.11 DI 422.3 4.22 0.060 Null 457.3 2.06 0.081 BGI Adjusted R2 = 0.17 Adjusted R2 = 0.20 p-value = 0.01 p-value = 0.03 Non-Cyclonic Storms Nor'easters AICc ΔAICc Weights Model Variables AICc ΔAICc Weights Model Variables 195 0 0.128 G 441.3 0 0.26 AI 195 0.01 0.127 Null 442.2 0.87 0.168 A 195.3 0.3 0.11 BG 442.2 0.88 0.168 I 195.7 0.76 0.087 B 442.8 1.51 0.122 Null 196.5 1.51 0.06 I 443.3 1.99 0.096 AGI Adjusted R2 = 0.09 Adjusted R2 = 0.26 p-value = 0.11 p-value = 0.05 Rocky Lows AICc ΔAICc Weights Model Variables 412.4 0 0.201 I 413.1 0.69 0.143 Null 413.4 0.95 0.125 BI 414.4 1.97 0.075 B 414.4 2.02 0.073 AI 412.4 0 0.201 I Adjusted R2 = 0.09 p-value = 0.11 A – AO; B – EA; C – NIÑO 3; D – NIÑO 4; E – NIÑO 3.4; F – NAO; G – PDO; H – PNA; I – WP
215
Teleconnections significantly influenced seasonal snowfall totals for four of the seven snowstorm types in Region 5 (Table 6.5). The greatest influence was on seasonal snowfall totals from Canadian lows (R2 = 0.31), whereas modeled fits were relatively low for the remaining snowstorms (R2 < 0.15). Seasonal snowfall totals from Canadian lows were most influenced by the EA and NAO. The influence of the Artic and Atlantic Ocean on seasonal snowfall totals from Canadian lows in Region 5 is similar to their influence in
Regions 1-4. The influence of the EA and NAO on seasonal snowfall totals from Canadian lows is greater in Region 5 than the other subregions, possibly due to the northern track of these storms and the northern position of Region 5. Since Canadian lows form and move at high latitudes, they often advect cold air into Central New York with western winds over
Lake Ontario (Changnon 1969; Zielinski 2002; Branick 1997; Changnon et al. 2008). Since western winds favor the formation of lake-effect and lake-enhanced snow over Region 5, the larger influence of teleconnections on these storms may be linked to lake-enhanced snow produced by Canadian lows.
Teleconnections did not significantly ( = 0.09) influence seasonal snowfall totals from
Nor’easters in Region 5. The smaller influence in this region may be due to the lower percentage of the seasonal snowfall that Nor’easters account for here (Table 3.7; Figure 3.9;
Figure 3.11b). Model results suggest that the influence of teleconnections on east coast storms varies throughout Central New York. This is consistent with the general patterns across the United States as Ghatak et al. (2010) notes that the positive phase of the PNA has an influence throughout North America, while the NAO’s influence is restricted to eastern
North America. Wise et al. (2015) also show that precipitation patterns across the
216 continent vary when different phases of climate modes interact. For example, if the positive phases of the Southern Oscillation Index (SOI) and NAO are matched with the negative phase of the EA or the positive phase of the PNA or WP, it leads to anomalously high cool-season precipitation totals everywhere throughout the United States except the
Table 6.5. AIC table for the teleconnection predictors of seasonal snowfall totals for the different snowstorm types in Region 5. The table shows AIC values (AICc), differences in AIC values between the models (ΔAICc), Akaike weights (Weights), and the teleconnection models included in the analysis (Model Variables). Models in red signify the top model for that storm based on weight, p-value, and adjusted R2. Lake Snow Non-Lake Snow AICc ΔAICc Weights Model Variables AICc ΔAICc Weights Model Variables 473.8 0 0.35 I 469.7 0 0.36 Null 475.8 2.07 0.124 GI 471.4 1.78 0.148 I 476.2 2.42 0.104 BI 471.8 2.16 0.122 C 476.2 2.5 0.1 HI 473.2 3.52 0.062 CI 476.3 2.51 0.1 CI 474.3 4.65 0.035 BCI Adjusted R2 = 0.13 Adjusted R2 = 0.00 p-value = 0.03 p-value = 0.39 Canadian Lows Lake-Effect Snowstorms AICc ΔAICc Weights Model Variables AICc ΔAICc Weights Model Variables 416.8 0 0.506 BF 469.7 0 0.36 I 419.6 2.81 0.124 ABF 472 2.28 0.115 Null 420.3 3.41 0.092 ABFG 472.2 2.4 0.109 HI 421.3 4.46 0.055 AB 472.2 2.42 0.107 BI 421.5 4.69 0.049 F 472.3 2.52 0.102 GI Adjusted R2 = 0.31 Adjusted R2 = 0.12 p-value = 0.00 p-value = 0.04 Non-Cyclonic Storms Nor'easters AICc ΔAICc Weights Model Variables AICc ΔAICc Weights Model Variables 204.8 0 0.231 B 431.3 0 0.197 A 205.1 0.33 0.196 BG 431.7 0.38 0.163 AD 205.5 0.74 0.16 Null 431.9 0.64 0.144 Null 206.4 1.61 0.103 G 433.1 1.76 0.082 AB 207.1 2.38 0.07 BI 433.3 1.95 0.075 AI Adjusted R2 = 0.17 Adjusted R2 = 0.10 p-value = 0.13 p-value = 0.09 Rocky Lows AICc ΔAICc Weights Model Variables 397.6 0 0.256 Null 397.9 0.23 0.229 E 398.2 0.55 0.195 B 399.3 1.65 0.113 EI 399.3 1.68 0.111 I Adjusted R2 = 0.04 p-value = 0.15 A – AO; B – EA; C – NIÑO 3; D – NIÑO 4; E – NIÑO 3.4; F – NAO; G – PDO; H – PNA; I – WP
217 west coast. However, matched with the positive phase of the EA or the negative phase of the PNA or WP, cool-season precipitation totals are greater in the Northeast, but lower around the Great Lakes. The WP also significantly influenced seasonal snowfall totals from lake snowstorms and lake-effect snowstorms in Region 5. The influence of the WP on lake- effect and lake-enhanced snowfall is something that should be further investigated, especially due to its apparent importance in terms of its association with lake-effect snow in the traditional lake-effect snowbelt (see Chapter 3).
6.3.2 Variable Correlations
The relationships (strength and direction) between seasonal snowfall totals from different snowstorms and the teleconnection patterns shown to influence those storms are presented in Table 6.6. Relationships were tested using Pearson correlations for each variable in the top models identified in Section 6.3.1.
Results from the correlations suggest that the PDO and the WP have the greatest influence on seasonal snowfall totals in Central New York, as both were predictor variables in sixteen of the thirty-five top snowfall models, and fourteen and twelve of the twenty-two top significant ( ≤ 0.05) snowfall models, respectively (Table 6.6). The next most influential teleconnections were the AO and the NAO, as both were in nine top significant snowfall models. The EA had the least influence on seasonal snowfall totals in Central New York, as it was only in five top significant snowfall models. ENSO was a predictor variable in seven of the top significant models, with the greatest influence from NIÑO 3 (four top models).
218
Table 6.6. The correlation of all teleconnection variable against seasonal snowfall totals for the top models identified in Section 6.3.1. Correlations are bolded and italicized if the top model was significant (ρ 0.05). AO EA NIÑO 3 NIÑO 4 NIÑO 3.4 NAO PDO PNA WP Region 1 Lake Snow -0.29 0.17 Non-Lake Snow -0.20 -0.15 0.42 Canadian Lows 0.20 0.08 -0.30 Lake-Effect Snow -0.20 Non-Cyclonic -0.11 0.20 Nor'easters 0.20 0.03 -0.30 0.27 Rocky Lows -0.27 Region 2 Lake Snow -0.33 -0.24 Non-Lake Snow -0.21 0.35 Canadian Lows -0.26 0.26 Lake-Effect Snow -0.18 -0.17 -0.33 Non-Cyclonic -0.08 Nor'easters 0.18 -0.26 0.26 Rocky Lows 0.07 -0.04 0.16 0.05 Region 3 Lake Snow -0.01 0.37 -0.26 -0.31 Non-Lake Snow 0.55 -0.11 0.11 Canadian Lows 0.01 0.27 -0.28 -0.32 Lake-Effect Snow 0.47 0.23 Non-Cyclonic 0.22 Nor'easters 0.23 -0.25 -0.07 -0.47 0.44 -0.19 -0.10 Rocky Lows 0.12 -0.34 0.20 Region 4 Lake Snow 0.24 0.25 -0.49 Non-Lake Snow -0.14 0.19 -0.11 Canadian Lows 0.08 Lake-Effect Snow -0.06 0.25 0.31 Non-Cyclonic 0.32 0.11 Nor'easters 0.38 -0.03 Rocky Lows -0.20 -0.32 Region 5 Lake Snow -0.40 Non-Lake Snow 0.21 -0.10 -0.15 Canadian Lows 0.18 -0.54 Lake-Effect Snow 0.14 Non-Cyclonic 0.42 -0.02 Nor'easters -0.43 -0.12 Rocky Lows -0.32
219
The PDO had the greatest influence on seasonal snowfall totals from non-lake snowstorms, as it was an explanatory variable in the top model for four of the five snowfall regions.
Results suggest that a positive (negative) phase of the PDO leads to greater (lesser) seasonal snowfall totals from non-lake snowstorms. A similar influence was shown on seasonal snowfall totals from Nor’easters as totals were greater (smaller) in three of the five subregions during the positive phase of the PDO. This is consistent with previous findings suggesting that a positive phase of the PDO is linked to above average snowfall totals in the Northeast United States (Ge and Gong 2009; Kunkel et al. 2009b; Hidalgo and
Dracup 2003; Gutzler et al. 2002; McCabe and Dettinger 2002; Goodrich and Walker 2011).
Interestingly, the PDO was also a predictor variable for seasonal snowfall totals from lake snowstorms and lake-effect snowstorms in Regions 3 and 4. Results suggest that the positive (negative) phase enhances (reduces) seasonal snowfall totals from lake snowstorms and lake-effect snowstorms in these regions.
The WP significantly influenced seasonal snowfall totals for every snowstorm type, except those from non-cyclonic snowstorms. The WP had the greatest influence on lake snowstorms and lake-effect snowstorms, as snowfall from lake snowstorms was generally lower (higher) and snowfall from lake-effect snowstorms was higher (lower) during the positive (negative) phase of the WP. The greater snowfall from lake-effect snowstorms during the positive phase of the WP may be linked to colder air temperatures in the eastern
United States (Barnston and Livezey 1987; Wise et al. 2015). The colder air likely advects over the Great Lakes, and barring an ice cover, creates additional precipitation downwind of the lakes, including Central New York. Interestingly, the positive phase and its
220 associated colder air, leads to lower seasonal snowfall totals from Nor’easters and
Canadian lows. The conclusions in Chapter 5 suggest that seasonal snowfall totals from
Nor’easters and Canadian lows are most tied to average air temperatures and precipitation totals. Since air temperatures are conducive for snowfall during the positive phase of the
WP, it is suggested that the air is relatively dry. This is consistent with the formation of lake-effect snow as a high moisture content of the air is not necessary for the formation of snow from these storms, but cold air temperatures are required. The lower snowfall totals from lake snowstorms during the positive phase of the WP may be tied to the reduced snowfall from Nor’easters and Canadian lows. Since these storms frequently produce lake- enhanced snow, if their frequency decreases, the snowfall from these storms also decreases, including their lake-enhancement.
The AO influenced the seasonal snowfall totals of four different snowstorms. The greatest influence was on seasonal snowfall totals from Nor’easters, as totals were greater (lower) during the positive (negative) phase of the AO in Regions 1-4, and lower in Region 5. Model results also suggest that the AO significantly influences seasonal snowfall from Rocky Lows in Regions 2 and 3, as a positive phase tends to produce abnormally high seasonal snowfall from these storms. However, in Regions 1 and 2, seasonal snowfall totals from lake snowstorms are lower (higher) during the positive (negative) phase, which is consistent with previous findings that suggest that the negative phase of the AO leads to anomalously cold temperatures in the midlatitudes of North America due to frequent troughing in the jet stream (Bai et al. 2012; Rohli and Vega 2011; Zhu and Wang 2016).
221
The NAO significantly influences seasonal snowfall totals for four snowstorm types affecting Central New York, with the greatest influence on totals from Canadian lows and
Nor’easters. Models suggest that the negative (positive) phase leads to above (below) average snowfall from Canadian Lows in Regions 1, 2 and 5, and above (below) average snowfall from Nor’easters in Regions 1-3. The negative phase is also linked to above average seasonal snowfall from Rocky Lows in Region 3 and above average snowfall from non-lake snowstorms in Regions 1 and 2. These results are consistent with previous findings which suggest that the negative phase of the NAO is associated with lower air temperatures in the eastern United States and above average seasonal snowfall totals in the region (Ghatak et al. 2010; Hurrell 1995; Notaro et al. 2006; Bai et al. 2012; Barnston and
Livezey 1987; Archambault et al. 2008; Seager et al. 2010a; Osborn 2011; Coleman and
Budikova 2013; Roller et al. 2016).
The influences of ENSO on seasonal snowfall totals throughout the United States have been widely examined (Serreze et al. 1998; Ge and Gong 2009; Ghatak et al. 2010; Allen and
Zender 2011; Grise et al. 2013; Baxter et al. 2014; Yu et al. 2014; Gan and Wu 2015).
Although research suggests that the El Niño (La Niña) phase favors anomalously low seasonal snowfall totals in the Northeast United States, the results from this study suggest that ENSO has a relatively weak influence on seasonal snowfall totals in Central New York.
Model results suggest that ENSO’s greatest influence is on seasonal snowfall totals from cyclonic snowstorms, including non-lake snowstorms, Nor’easters, and Rocky lows. This is consistent with previous findings which suggest that ENSO has a greater effect on altering seasonal snowfall totals from cyclonic snowstorms than snowfall totals from non-cyclonic
222 snowstorms (Ge and Gong 2009; Ghatak et al. 2010; Gan and Wu 2015). Modeled results also suggest that the El Niño phase favors abnormally low seasonal snowfall totals from these storms, which is consistent with previous findings (Serreze et al. 1998; Ge and Gong
2009; Ghatak et al. 2010; Allen and Zender 2011; Grise et al. 2013; Baxter et al. 2014; Yu et al. 2014; Gan and Wu 2015). Although ENSO had a relatively small influence on seasonal snowfall totals in Central New York, results are consistent with Grimaldi (2008) as its intraseasonal influences may need to be examined. The weaker influence of ENSO illustrated in this study may be due to the linear relationship assumed between ENSO and snowfall in this region, since previous research have suggested that there is a quadratic relationship between snowfall and ENSO in the northern United States (e.g. Grimaldi 2008;
Kunkel and Angel 1999; Kunkel et al. 2013b,a).
The PNA and the EA were the least influential in affecting seasonal snowfall totals in
Central New York. Although the PNA influences seasonal snowfall totals in five subregions, its effects are mostly confined to Regions 3 and 4. The PNA has the greatest effect on seasonal snowfall totals from non-lake snowstorms, where the positive (negative) phase leads to below (above) normal snowfall totals in Regions 3 and 4. However, seasonal snowfall totals from Rocky lows are higher during the positive phase. The negative response of snowfall during the positive phase is unexpected since this phase is associated with anomalously cold temperatures over the eastern United States (Ge and Gong 2009;
Ghatak et al. 2010; Barnston and Livezey 1987; Leathers et al. 1991; Wise et al. 2015).
Since seasonal snowfall totals in Region 3 are most influenced by the PNA, the negative relationship between snowfall and the PNA may be region specific.
223
Seasonal snowfall totals were only significantly affected by the EA for four snowstorms types. The EA has the greatest influence on seasonal snowfall totals from Canadian lows as the positive (negative) phase leads to above (below) average snowfall totals. However, during the positive phase, seasonal snowfall totals from Nor’easters and lake-effect snowstorms are generally below average. Although the models were not significant (p >
0.10), the positive phase also leads to above average seasonal snowfall totals from non- cyclonic snowstorms in Regions 3-5. The general increase in snowfall during the positive phase is consistent with previous findings which suggest above average snowfall typically occurs in the United States due to anomalously cold surface temperatures (Wise et al. 2015;
Barnston and Livezey 1987; Davis and Benkovic 1994; Seierstad et al. 2007; Woollings and
Blackburn 2012; Moore et al. 2013).
6.4 Discussion
Results suggest that seasonal snowfall totals in Central New York are significantly influenced by the AO, EA, ENSO, NAO, PDO, PNA and WP; however, these influences vary depending on the subregion and the type of snowstorm. Teleconnections appear to have the greatest (R2 > 0.15) influence on seasonal snowfall totals from Nor’easters, especially in
Regions 1-3. Above normal snowfall totals were linked to the positive phase of the AO and
PDO, and the negative phase of the NAO, suggesting that the AO, NAO, and PDO had the greatest influence on seasonal snowfall totals from Nor’easters. This could be due to a displacement of the jet stream to the south of Central New York, which results in abnormally cold air temperatures and favors the cyclogenesis of storms forming south of
35N (Zhu and Wang 2016; Rohli and Vega 2011; Ghatak et al. 2010; Notaro et al. 2006;
224
McCabe and Dettinger 2002). As shown in Chapter 4, storms forming south of 35N tend to produce heavier snowfall throughout Central New York (Table 4.1), with the greatest totals extending from the eastern shores of Oneida Lake to eastern Otsego County (Figure 4.5).
Seasonal snowfall totals from lake snowstorms and lake-effect snowstorms were significantly influenced by the WP. Although previous research has provided little evidence of its influence, the results here suggest that it is the most influential teleconnection on seasonal snowfall totals from these storms. Above average seasonal snowfall totals from lake snowstorms in Regions 2-5 are linked to the negative phase, whereas above average snowfall totals from lake-effect snowstorms in Regions 3-5 are linked to the positive phase.
Although the link between the WP and lake-effect snow may be a spurious correlation, possible physical connections are via changes to the number of days air temperatures drop below 0C and the average seasonal and winter air temperatures as shown in Chapter 5.
According to the CPC (2005), the positive phase of the WP tends to result in warmer surface air temperatures over the Great Lakes region. This likely leads to less freezing days and is responsible for the greater snowfall totals from lake snowstorms. Interestingly, lake snowstorms and lake-effect snowstorms were mostly influenced by teleconnections in the
Pacific Ocean, suggesting that it has a greater influence on the variability of seasonal snowfall totals from these storms than the Atlantic Ocean.
Although seasonal snowfall totals from lake snowstorms and lake-effect snowstorms were largely affected by teleconnections in the Pacific Ocean, totals from Canadian lows were mostly influenced by teleconnections in the Atlantic. The EA had a significant influence on
225 seasonal snowfall totals from Canadian lows in Regions 1 and 5, where the positive phase favors above normal snowfall. The anomalously high snowfall is potentially linked to cooler air temperatures and more days below 0C, as suggested by the conclusions of
Chapter 5, and corroborating the results of Wise et al. (2015), Davis and Benkovic (1994) and Seierstad et al. (2007). The NAO had a broader influence, affecting seasonal snowfall totals in Regions 1, 2, 4, and 5. Generally, seasonal snowfall totals from Canadian lows were higher during the negative phase of the NAO, potentially due to more frequent cold- air outbreaks as suggested in Chapter 5 and corroborated by Ghatak et al. (2010), Wise et al. (2015), and Seager et al. (2010a). The influence of both the Atlantic and Pacific Oceans on seasonal snowfall totals in Central New York highlight the importance of teleconnection patterns in this region. Patterns in the Atlantic can influence snowstorms that originate in northwest Canada, while conditions in the Pacific can affect seasonal snowfall totals from lake snowstorms and lake-effect snowstorms.
6.5 Conclusion
The purpose of this chapter was to examine the influence of teleconnection patterns on the seasonal snowfall totals from different snowstorms in Central New York from 1985/86 –
2014/15. The influence of the teleconnections was examined using linear fixed-effect models and AIC techniques commonly used in the biological sciences, but seldom used to understand the influence of teleconnections on the oceans and atmosphere. Therefore, this study also assessed the applicability of these methods.
226
Although previous research has linked teleconnection patterns to changes in seasonal snowfall totals in North America (Wise et al. 2015; Grimaldi 2008; Smith and O’Brien 2001;
Ghatak et al. 2010; Ge and Gong 2009; Barnston and Livezey 1987), none of these studies directly observe the influence of teleconnections on seasonal snowfall totals from individual snowstorm types. By examining individual storm types this chapter provides a more nuanced scientific understanding of snowfall in the Great Lakes region and builds upon the local variables responsible for changes in snowfall observed in Chapter 5.
Results suggests that the AO, NAO, and PDO explain the most variance in snowfall from
Nor’easters, while lake-effect snowfall is most influenced by the WP. However, for each scenario, the amount of variance explained was less than 50%. Although this type of modeling helps account for the potential additive (reductive) properties of teleconnections as they interact, it does not significantly improve upon techniques traditionally used in the atmospheric sciences, such as principal component analyses. Also, although model results indicate associations between different teleconnections and seasonal snowfall totals, these associations are statistical and not physical. Therefore, the use of linear-fixed effects models and the AIC to observe the influence of teleconnections on seasonal snowfall totals has potential, but needs to be developed further and confirmed by physical modeling to corroborate environmental conditions driving relations.
227
7.0 CONCLUSIONS
The Laurentian Great Lakes region experiences several different snowstorm types throughout winter. These snowstorms originate in many parts of North America that can lend to the storms’ characteristics in terms of their snowfall totals, snow densities, air temperatures, and wind speeds (Whittaker and Horn 1981; Zishka and Smith 1980; Jones and Davis 1995; Zielinski 2002; Changnon et al. 2008). Snowfall from these storms is an important component of the winter environment, climate system, and hydrologic cycle
(Rohr et al. 2012; Ziska et al. 2011; Cortinas and Kitron 2006; Mastin et al. 2011; Andersen and Shepherd 2013; Changnon and Changnon 2006; Mote 2008; Francis and Vavrus 2015).
Numerous studies have examined how seasonal snowfall totals have changed in this region since the early-20th century and attempted to understand the potential causes behind any changes. Many have suggested that changes in lake-effect snowstorms are largely responsible for the increasing seasonal snowfall totals in the Great Lakes region throughout the 20th century (Hartnett et al. 2014; Burnett et al. 2003; Baxter et al. 2005;
Suriano and Leathers 2017a; Norton and Bolsenga 1993; Kunkel et al. 2009a). However, these conclusions are derived from proxy evidence rather than directly examining specific storm contributions to seasonal snowfall totals. Therefore, this research is the first comprehensive study to examine the contributions of different snowstorm types to seasonal snowfall totals within the Laurentian Great Lakes region, and to directly determine how these totals are changing over time.
The research is presented in seven chapters. Chapter 1 provides the framework for the research, by examining past studies which have observed snowfall within the Laurentian
228
Great Lakes region and Northeast United States. Previous estimates suggest that lake-effect snowstorms account for at least 30% of the seasonal snowfall in the Great Lakes region, with some estimates suggesting that lake-effect snowstorms account for more than 75% of the seasonal snowfall (Eichenlaub 1970; Veals and Steenburgh 2015; Miner and Fritsch
1997; Liu and Moore 2004). However, these claims do not consider spatial variations in seasonal snowfall from lake-effect snowstorms and therefore likely do not accurately reflect the contribution of lake-effect snowfall to seasonal snowfall totals throughout the entire Great Lakes region. Understanding the contribution of different snowstorms to seasonal snowfall totals is necessary to better predict how snowfall may change in the future.
Seasonal snowfall trends differ between areas with a considerable influence from lake- effect snowstorms and those areas without a considerable influence from lake-effect snowstorms (Burnett et al. 2003; Kunkel et al. 2009a; Hartnett et al. 2014). An increase in snowfall in areas affected by lake-effect snowstorms is expected to continue throughout the first half of the 21st century, while at the same time a decrease in snowfall is expected in areas outside the Great Lakes region (Notaro et al. 2013b; Suriano and Leathers 2016).
Separating out individual storm types and their contributions to seasonal snowfall provides a better picture of snowfall changes. From an overview of the current literature presented in Chapter 1, a detailed analysis of the influence of individual snowstorm types on seasonal snowfall totals both within and outside of the Great Lakes region is lacking.
Studies have examined the influence of different environmental forcings (e.g. air temperatures, precipitation totals, Great Lakes characteristics, and teleconnections) on
229 seasonal snowfall totals within the Great Lakes region (Tsuboki et al. 1989; Segal and
Kubesh 1996; Hanson et al. 1992; Notaro et al. 2015). However, these studies did not examine the influences of environmental conditions on different snowstorm types, and therefore, uncertainty remains as to their effects on different snowstorm types.
Chapter 2 explains the data and methods used to address the research questions. The contribution of different snowstorm types to seasonal snowfall totals was examined for
Central New York, a subregion of the Great Lakes region, from 1985/86 – 2014/15. To determine the contribution of different snowstorm types to seasonal snowfall totals within
Central New York, snowstorms were identified using data from the COOP. Snowstorms were classified into different categories using NCEP reanalysis data, GOES infrared imagery, NEXRAD data, and a combination of methods proposed by previous studies
(Figure 2.5; Perry et al. 2007; Kocin and Uccellini 2004b; Kelly et al. 2012; Niziol et al.
1995; Suriano and Leathers 2017a,b; Sobash et al. 2000; Laird et al. 2009a; Kelly 1986;
Whittaker and Horn 1981; Jones and Davis 1995). Results suggest that twelve different snowstorm types affected Central New York during the study period, including: clippers,
Colorado lows, frontal storms, Great Lakes lows, Hudson lows, lake-effect snowstorms,
Nor’easters, Oklahoma hooks, Oklahoma hooks, Texas hooks, tropical cyclones, and upper atmospheric disturbances. Due to similarities in some of these storms, storms were generalized into Canadian lows (clippers, Great Lakes lows, and Hudson lows), non-lake snowstorms (frontal storms and upper atmospheric disturbances), and Rocky Lows
(Colorado lows, Oklahoma hooks, and Texas hooks). Also, since cyclonic storms have been shown to initiate the formation of lake-effect snowfall in the Great Lakes region (Tardy
230
2000; Liu and Moore 2004), snowstorms were either classified as lake snowstorms or non- lake snowstorms. Lake snowstorms were defined as any snowstorm that met the criteria of a lake-effect snowstorm, but the storm’s precipitation was not separated by at least six hours from the precipitation of another system or was noticeably linked to another system.
Snowstorms that did not meet the criteria of a lake-effect snowstorm were classified as non-lake snowstorms. The chapter goes on to detail the procedures for calculating the average seasonal snowfall from these different storms and the methods used to analyze their contributions. Finally, this chapter discusses the data collected to observe the influences of the environment on seasonal snowfall totals from the different snowstorms to affect Central New York. Data included elevation, exposure, and distance from Lake
Ontario calculated using ArcGIS, composite NCEP reanalysis data from the Earth Systems
Research Laboratory (Kalnay et al. 1996), backward air trajectories from the Air Resource
Laboratory’s HYSPLIT model (Stein et al. 2015b; Rolph et al. 2017), atmospheric and lake data obtained from Syracuse Hancock International and the Great Lakes Environmental
Research Laboratory, and teleconnection data from the National Center for Environmental
Prediction and the National Center for Environmental Information (Table 2.1).
Chapter 3 examines the seasonal snowfall contributions of the snowstorms identified in
Chapter 2 at the regional, subregional, and local scale between 1985/86 and 2014/15.
Results suggest that although non-lake snowstorms occur more frequently ( < 0.05) than lake snowstorms in Central New York, they do not produce significantly ( = 0.09) more snowfall. This is consistent with previous findings which suggest that lake-effect and lake- enhanced snowstorms account for approximately half of the seasonal snowfall in the Great
231
Lakes region (Miner and Fritsch 1997; Liu and Moore 2004; Veals and Steenburgh 2015).
Spatial patterns suggest that lake snowstorms have the greatest influence on seasonal snowfall totals in Regions 4 and 5, which include areas to the lee of Lake Ontario, over the
Tug Hill. Non-lake snowstorms contribute the most to seasonal snowfall totals in Region 2, which includes southern Central New York, and along the St. Lawrence River. Lake-effect snowstorms occur more frequently and average more snowfall than Canadian lows, non- cyclonic snowstorms, Nor’easters, and Rocky lows in Central New York.
Similar to lake snowstorms, lake-effect snowstorms have the greatest influence on seasonal snowfall totals in Regions 4 and 5, just north of Oneida Lake. Nor’easters have the second greatest influence on seasonal snowfall totals, and are the dominant snowstorm in southeastern Central New York, the Finger Lakes region, and along the St. Lawrence River, contributing more than 35% of the seasonal snowfall total. Even though percentages vary,
Nor’easters contribute similar seasonal snowfall totals across Central New York, which is likely due to the northwest winds from these storms which often result in lake-effect and lake-enhanced snowfall (Niziol 1987; Suriano and Leathers 2017a; Liu and Moore 2004).
However, these storms have a disproportionate effect on seasonal snowfall totals in Region
2, potentially due to its position relative to Lake Ontario and the Atlantic Ocean (Changnon et al. 2008; Kocin and Uccellini 2004b). Seasonal snowfall from Canadian lows and non- cyclonic snowstorms were fairly homogenized throughout the study area, while larger contributions from Rocky lows were concentrated in southern and eastern Central New
York. Results from linear mixed-effects models suggest that the latitude, longitude, elevation, distance from Lake Ontario, and exposure of a location influence its seasonal
232 snowfall totals from different snowstorms types. Lake-effect and lake-enhanced snowfall totals at a location are most influenced by the latitude, longitude, elevation and distance from Lake Ontario. These results corroborate previous findings which suggest that seasonal snowfall totals within the Great Lakes region are greatest in higher elevations and closer to the lakes (Hill 1971; Dewey 1979b,a; Niziol 1987; Pease et al. 1988; Laird and
Kristovich 2004). Subregional results from this chapter were used to create a dataset used to analyze how different snowstorms have changed over time and the potential causes for such changes in Chapters 5 and 6.
In Chapter 4, I use NCEP/NCAR reanalysis data and the HYSPLIT model to examine the specific atmospheric conditions associated with different snowstorm types and how differences in the synoptic conditions influence the magnitude of storms affecting Central
New York. The synoptic conditions do have a considerable influence on both the type and magnitude of snowstorms to influence Central New York, and help explain the snowfall distribution patterns found in Chapter 3. Heavy-snowfall (> 25.4 cm) lake-effect snowstorms typically occur with the presence of a well-organized high pressure (> 1023 hPa) over the Mississippi River Valley and a strong (< 1005 hPa) surface low over northeastern Canada. This is consistent with conclusions of Leathers and Ellis (1996) and
Suriano and Leathers (2017a) who noted that similar synoptic patterns were responsible for the occurrence of lake-effect snowstorms in Syracuse, NY and the eastern Great Lakes.
The distribution of snowfall following lake-effect snowstorms was similar to that following
Canadian lows and upper atmospheric disturbances. This suggests that lake-effect and lake-enhanced snowfall is common following Canadian lows and upper atmospheric
233 disturbances, again consistent with previous findings (Niziol 1987; Liu and Moore 2004;
Mercer and Richman 2007). Snowfall distributions following Gulf Coast storms, Colorado lows, Texas hooks, and Oklahoma hooks is fairly homogeneous throughout Central New
York. These storms typically occur during omega patterns with high pressures near 30N in the Atlantic and Pacific, and over the north-central United States. This is consistent with previous findings which suggest that a stronger meridional pattern in the jet stream often increases the cyclogenesis and strength of these storms (Liu and Moore 2004; Barriopedro et al. 2006; Clark 1990; Whittaker and Horn 1981). The patterns in the jet stream can also influence storm tracks (Barriopedro et al. 2006), and snowfall totals from different snowstorms. For example, results suggest that Gulf Coast storms produce larger snowfall totals in eastern Central New York than east coast storms because of the more easterly tracks of the Gulf Lows.
Chapter 5 builds upon Chapter 3 to examine how seasonal snowfall contributions have changed over time for the different snowstorm types. This is the first comprehensive study to examine how seasonal snowfall totals have changed for different snowstorm types within the Great Lakes region. Results show that seasonal snowfall totals from clippers significantly decreased in every subregion between 1985/86 and 2014/15, and significantly increased from lake-effect snowstorms in Regions 2 and 3. The increase in snowfall from lake-effect snowstorms is consistent with previous findings (Norton and
Bolsenga 1993; Ellis and Johnson 2004; Burnett et al. 2003; Kunkel et al. 2009a), but the increase was observed outside of the typical lake-effect snowbelt (Hartnett et al. 2014).
The increased snowfall is potentially linked to a greater lake-air temperature difference
234
(Notaro et al. 2015), which has been shown to form more organized snowbands with lake- to-lake connections (Laird et al. 2017). Results also show that many of the long-term trends exhibit trend reversals, a finding identified previously by Bard and Kristovich
(2012) and Hartnett et al. (2014). Lake Ontario surface temperatures and freezing air temperatures are shown to have the greatest influence on seasonal snowfall trends in
Central New York, corroborating previous findings (Kapnick and Delworth 2013; Krasting et al. 2013). This was especially apparent in Regions 1-3, where precipitation totals and average air temperatures were more influential on seasonal snowfall totals in Regions 4 and 5.
Chapter 6 also examines the environmental conditions that influence seasonal snowfall totals from different snowstorms in Central New York. Results suggest that the AO, EA,
ENSO, NAO, PDO, and PNA significantly influence seasonal snowfall totals in Central New
York. Teleconnections have the greatest influence on seasonal snowfall totals from
Nor’easters, as seen by the above normal snowfall from these storms in association with the positive phases of the AO and PDO and the negative phase of the NAO. Teleconnections in the Atlantic Ocean and the Arctic have a significant influence on seasonal snowfall totals from Canadian lows, whereas the WP influences lake-effect and lake-enhanced snow.
Although studies have not previously linked the influence of the WP to lake-effect snowstorms in the Great Lakes region, the negative phase of the WP is linked to above average seasonal snowfall totals in the central and eastern United States (Wise et al. 2015;
Davis and Benkovic 1994; Seierstad et al. 2007). This study highlights the fact that teleconnections can have an influence at the subregional scale, producing different effects
235 on seasonal snowfall totals across the five subregions. Although analyzed separately, conclusions from this chapter suggest that global-scale patterns can directly influence regional snowfall patterns via changes to the regional variables identified in Chapter 5.
Snowfall has a considerable influence on the Great Lakes region; however, the influence varies over both time and space. The results from this study will help to better understand the different types of snowstorms that affect the eastern Great Lakes region, and their relative importance to seasonal snowfall totals within the region. Central New York is a particularly appropriate study area because of the complexities borne by the combination of multiple storm systems to influence the region. This complexity has here-to-fore not been specifically examined in studies observing spatial and temporal snowfall trends. The methodologies used in this study can also be applied to other areas that regularly experience seasonal snowfall, especially within the Great Lakes region, such as areas downwind of Lakes Erie, Michigan, Huron, and Superior.
Understanding the types of snowstorms to affect the Great Lakes region and their relative contribution to seasonal snowfall totals is vital to understanding how snowfall may change in the future, particularly due to climate change (Collins et al. 2013; Hartmann et al. 2013).
As the climate warms, increased temperatures threaten temperature sensitive processes such as the formation of snow. Changes in snowfall are further complicated by alterations to the polar jet stream and a weakening of the circumpolar vortex leading to more frequent cold-air damming over North America (Barnes and Simpson 2017; Delcambre et al. 2013).
Recent climate change has had varying effects on snowfall, where snowfall in lake-effect
236 dominated regions has increased since the early-20th century and snowfall in non-lake- effect dominated areas has remained relatively unchanged or significantly decreased
(Braham and Dungey 1984; Norton and Bolsenga 1993; Burnett et al. 2003; Ellis and
Johnson 2004; Kunkel et al. 2009a; Hartnett et al. 2014). Similar snowfall trends are expected to continue throughout the first half of the 21st century due to warming lake surface temperatures and more frequent cold-air (< 0⁰C) events advecting over the Great
Lakes from the Arctic (Kunkel et al. 2002; Notaro et al. 2013b; Suriano and Leathers 2016).
Currently, there is a high degree of uncertainty in seasonal snowfall predictions for a region, because estimates rely on future conditions in the atmosphere and oceans to make these predictions. Seasonal snowfall predictions are aggregated into estimates of the seasonal snowfall totals, and do not separate out snowfall totals from individual snowstorms. Since a location’s seasonal snowfall total is the summation of snowfall from multiple storms throughout the season, and storms are responding differently to climate change, these predictions could be enhanced by estimating them at smaller regional scales and for the individual snowstorm types. Therefore, the results from this study will help better resolve the percent contribution of different snowstorm types to seasonal snowfall totals within the Great Lakes region, so that improved assessments of future snowfall change can be made.
Improved seasonal snowfall predictions are necessary for the high latitudes (> 35⁰) of
North America because many environmental, ecological, and human processes rely on snowfall. For example, the results from this research may help botanists and ecologists understand the vulnerability of plants and animals to disease due to changes in snowfall
237 cover (Kreyling and Henry 2011; Campbell et al. 2014; Cortinas and Kitron 2006; Rempel
2011; Notaro and Liu 2008). A better understanding of seasonal snowfall totals can also provide a valuable tool for agriculture, water resource management, and forestry, through better drought and wildfire forecasting (Pederson et al. 2006; Bumbaco and Mote 2010;
Mishra et al. 2010; Groisman et al. 2004; Westerling et al. 2006). Improved seasonal snowfall predictions can also help improve societal resilience and preparation for the health effects of snowfall and the impacts on businesses and transportation that rely on snowfall or are negatively affected by it (Persinger et al. 1993; Andreescu and Frost 1998;
Graham and Diaz 2001; McCabe et al. 2001; Falk 2010; Hopkins and Maclean 2014).
Therefore, although snowfall is a seasonal phenomenon, it has a considerable influence on the climate, hydrology, ecology, biology, economy, and society of an area throughout the year.
Snowfall’s critical role in northern latitudes demands that further research should continue to investigate it at the scale of the individual snowstorm type. Additional questions have emerged that could be examined in future studies. For example, studies might examine whether the amount of snowfall produced by different snowstorms is consistent throughout the winter season, or if certain snowstorms dominate during different periods of the snowfall season. This may help better understand why snowfall totals from different snowstorms are changing and how they may change in the future because regional warming is not uniform throughout the year for most areas, especially in the Great Lakes region (Bolsenga and Norton 1993; Dietz and Bidwell 2011; Vavrus et al. 2013). Future research might also examine the timing of different snowstorms to determine if there is a
238 general pattern in the accumulation of seasonal snowfall totals. Previous findings suggest that there has been a general delay in the occurrence of lake-effect snowstorms to later in the season and an earlier occurrence of heavy-snowfall producing cyclonic snowstorms
(Whittaker and Horn 1981; Jones and Davis 1995; Hirsch et al. 2001; Vavrus et al. 2013).
Since this research examines individual snowstorms and their associated snowfall totals, future research could examine the timing of accumulations and if the date of the accumulations has changed over time.
With the individual snowstorm data provided from this research, future studies could also investigate whether the period between successive snowstorms of the same type has changed over time. Previous findings suggest that atmospheric patterns are becoming increasingly more stagnant (e.g. omega blocking patterns lasting days to weeks) as the climate changes (Barriopedro et al. 2006). These stagnant patterns have led to extreme events such as Hurricane Harvey, the 2010 Russian heat wave, the 2010 Pakistan floods, and the extreme cold of the 2014/15 Northeast United States’ winter (Carrera et al. 2004;
Whan et al. 2016; Brunner et al. 2017; Sillmann et al. 2011). Therefore, future research could examine whether the time between successive snowstorms has changed due to these stagnant atmospheric patterns.
A final future research question that emerged is whether other environmental conditions influence seasonal snowfall totals from different snowstorms to affect Central New York.
Although results from the linear fixed-effects models did not considerably improve the understanding of the environmental conditions on seasonal snowfall totals compared to
239 traditional methods (e.g. principal component analyses), results did suggest that additional parameters influence seasonal snowfall totals in Central New York. Therefore, future research could examine additional parameters which may influence seasonal snowfall totals so that seasonal predictions can be further enhanced. All of these future research projects will build upon this study, which has provided a basis for improving seasonal snowfall predictions in Central New York.
240
8.0 Appendices
Appendix 8.1. Quality COOP stations used for analysis and their latitude (Lat.), longitude (Long.), elevation (Elev.) in meters, start and end dates, hour(s) that measurements are recorded (recording time), and number of years the station has been active (years). Station ID Region Network ID County Lat. Long. Auburn AUB 1 USC00300321 Cayuga 42.9327 -76.5447 Aurora Research AUR 1 USC00300331 Cayuga 42.73333 -76.65 Farm Bainbridge 2 E BAIN 2 USC00300360 Chenango 42.28333 -75.45 Baldwinsville BVIL 1 USC00300379 Onondaga 43.15 -76.33333 Barnes Corners BARN 5 USC00300424 Lewis 43.81667 -75.8 Beaver Falls BEAV 5 USC00300500 Lewis 43.887 -75.4349 Bennetts Bridge BEN 5 USC00300608 Oswego 43.5318 -75.95296 Big Moose 3 SE BIG 4 USC00300668 Herkimer 43.8 -74.86667 Boonville 4 SSW BOON 4 USC00300785 Oneida 43.45 -75.35 Brewerton Lock BREW 4 USC00300870 Onondaga 43.2386 -76.1964 23 Camden NY CAM 4 USC00301110 Oneida 43.31889 -75.84722 Cayuga Lock CAY 1 USC00301265 Cayuga 42.948 -76.7342 Number 1 Chepachet CHEP 2 USC00301424 Herkimer 42.9096 -75.1109 Cherry Valley 2 CHRY 2 USC00301436 Otsego 42.8238 -74.7386 NNE Cincinnatus CINCY 1 USC00301492 Cortland 42.53333 -75.9 Constantia 6 N CON 4 USC00301732 Oswego 43.3426 -76.0006 Cooperstown COOP 2 USC00301752 Otsego 42.7166 -74.9266 Cortland CORT 1 USC00301799 Cortland 42.6 -76.18333 Delta Dam NY DD 3 USC00302047 Oneida 43.2735 -75.4271 Frankfort Lock FRANK 3 USC00303010 Herkimer 43.06667 -75.11667 19 Freeville 1 NE FREE 1 USC00303050 Tompkins 42.51667 -76.33333 Fulton FUL 4 USC00303087 Oswego 43.3049 -76.3938 Greene GRN 2 USC00303444 Chenango 42.3239 -75.7711 Griffiss AFB GRIF 3 USW00014717 Oneida 43.23333 -75.4 Highmarket HI 4 USC00303851 Lewis 43.5752 -75.5207 Hinckley 2 SW HINK 3 USC00303889 Herkimer 43.3 -75.15 Hooker 12 NNW HOOK 5 USC00303961 Lewis 43.8524 -75.7158 Ithaca Cornell U. CORN 1 USC00304174 Tompkins 42.4491 -76.4491 Little Falls City LFALL 3 USC00304791 Herkimer 43.0603 -74.8686 Reservoir S Locke 2 W LOC 1 USC00304836 Cayuga 42.6702 -76.4722 Lowville LOW 5 USC00304912 Lewis 43.7929 -75.4829
241
APPENDIX 8.1 CONTINUED Lyons Falls LYF 4 USC00304944 Lewis 43.61667 -75.36667 Maryland 6 SW MD 2 USC00305113 Otsego 42.4694 -75.0105 Montague MONT 5 USC00305438 Lewis 43.76111 -75.68028 Morrisville 6 SW MOV 1 USC00305512 Madison 42.85 -75.65 New Berlin NBLN 2 USC00305687 Chenango 42.61667 -75.33333 Newport 7 NE NEW 3 USC00305769 Herkimer 43.2 -74.91667 Norwich NOR 2 USC00306085 Chenango 42.5117 -75.5197 Old Forge OF 4 USC00306184 Herkimer 43.7026 -74.9838 Oneonta ONT 2 USC00306217 Otsego 42.4604 -75.0643 Oswego East OS 5 USC00306314 Oswego 43.4622 -76.4934 Palermo 2 SSE PAL 4 USC00306376 Oswego 43.3319 -76.2691 Pulaski PUL 5 USC00306867 Oswego 43.5696 -76.1163 Rectors Corners RTC 5 USC00306965 Lewis 43.75 -75.68333 Sherburne SHER 2 USC00307705 Chenango 42.6773 -75.5066 Skaneateles SKAN 1 USC00307780 Onondaga 42.95 -76.43333 Stillwater STILL 4 USC00308248 Herkimer 43.8999 -75.0367 Reservoir SUNY ESF ESF 1 USC00308386 Onondaga 43.0344 -76.1344 Syracuse Syracuse Hancock Int’l SYR 1 USW00014771 Onondaga 43.11667 -76.11667 Airport Trenton Falls TNT 3 USC00308578 Herkimer 43.2761 -75.1566 Tully 4 NE TUL 1 USC00308625 Onondaga 42.83333 -76.03333 Tully Heiberg BERG 1 USC00308627 Cortland 42.76667 -76.08333 Forest Unadilla 2 N UN 2 USC00308670 Otsego 45.3541 -75.3241 Utica UT 3 USC00308739 Herkimer 43.08333 -75.2 Utica 7 SSW UT7 3 USC00308742 Oneida 43 -75.26667 Utica Oneida CO ONCA 3 USW00094794 Oneida 43.145 -75.38389 Airport Watertown WTR 5 USC00309000 Jefferson 43.9761 -75.8753 Wellesley Island WELL 5 USC00309055 Jefferson 44.3565 -75.9285 Westmoreland 4 WEST 3 USC00309248 Oneida 43.1595 -75.2611 N Williamstown WIL 4 USC00309480 Oswego 43.37694 -75.92056
242
Appendix 8.1 Continued Part II Station Elev. Start Date End Date Recording Time Years 1800; 2-1995 Auburn 234.7 8/22/1897 4/9/17 118 (0700) Aurora Research 253 11/1/56 4/12/17 0800 59 Farm Bainbridge 2 E 303 12/1/07 12/6/11 0700 104 Baldwinsville 115.5 1/1/1893 4/13/17 0800 122 Barnes Corners 463.3 11/1/79 3/31/90 1100 11 Beaver Falls 225.6 4/1/34 2/28/17 0700 81 Bennetts Bridge 201.2 5/1/41 4/7/17 0700 74 1700; 10-2010 Big Moose 3 SE 536.4 3/1/31 4/13/17 84 (0700) Boonville 4 SSW 481.6 10/1/49 4/14/17 0700 66 Brewerton Lock 23 114.9 1/1/32 4/7/17 0700 83 1000; 0800 (4- Camden NY 176.8 8/1/46 2/28/17 1990); 0700 (9- 69 1998) Cayuga Lock No. 1 115.8 1/1/32 4/14/17 0800 83 Chepachet 402.3 9/2/57 7/31/01 0700 44 Cherry Valley 2 NNE 414.5 11/1/44 5/31/11 1800 67 Cincinnatus 320 5/1/28 12/31/10 0700 82 Constantia 6 N 177.1 1/1/03 12/18/07 0700 4 0800; 4-2009 Cooperstown 383.1 1/1/1893 4/8/17 122 (2400) Cortland 344.1 7/1/1895 12/31/00 0700 105 Delta Dam NY 167.6 4/1/00 4/9/17 0700 15 Frankfort Lock 19 125 1/1/32 10/31/97 0800 65 Freeville 1 NE 320 6/1/48 4/14/17 0800 67 1800; 11-1996 Fulton 109.7 3/1/00 4/7/17 115 0800/0700 Greene 280.4 2/24/09 4/9/17 0700 106 Griffiss AFB 158.2 1/1/1893 8/31/95 2400 102 Highmarket 537.4 5/20/24 4/9/17 0600 91 Hinckley 2 SW 347.8 5/1/26 11/18/08 0700 82 1800; 9-1986 Hooker 12 NNW 450.5 5/1/11 4/6/17 104 (0800) Ithaca Cornell U. 292.6 1/1/1893 4/14/17 0700 122 Little Falls City 1900; 7-1987 272.2 1/1/1897 8/31/15 118 Reservoir (0700) Locke 2 W 365.8 1/1/32 2/29/12 0800 80 1800; 6-2010 Lowville 263 11/1/1891 4/9/17 124 (0700) Lyons Falls 243.8 8/1/26 12/31/00 0700 74
243
APPENDIX 8.1 PART II CONTINUED Maryland 6 SW 373.4 11/1/83 4/9/17 0700 32 Montague 542.5 2/1/98 12/31/99 0700 1 Morrisville 6 SW 423.7 8/1/11 4/13/17 0700 104 New Berlin 329.2 10/1/07 2/28/97 0700 90 Newport 7 NE 516.3 3/1/85 1/31/95 0700 10 Norwich 301.4 8/1/06 4/9/17 0700 9 1900; 1-1988 Old Forge 532.8 12/1/07 4/9/17 108 (0900) Oneonta 350.5 2/1/10 4/4/16 0700 5 Oswego East 106.7 1/1/26 12/31/16 1700 89 Palermo 2 SSE 135.6 3/1/11 4/7/17 0600 4 Pulaski 121.6 5/2/48 4/8/17 0700-0900 67 Rectors Corners 550.5 1/1/87 12/26/90 1800 3 Sherburne 333.8 12/14/07 4/9/17 0700 8 Skaneateles 266.7 2/1/00 3/16/17 0800 115 Stillwater Reservoir 515.1 6/1/25 4/9/17 0800 90 SUNY ESF Syracuse 173.4 9/1/01 4/1/06 0700 5 Syracuse Hancock 125 5/1/38 4/14/17 2400 77 Int’l Airport Trenton Falls 243.8 8/1/09 4/7/17 0800 106 Tully 4 NE 396.2 8/1/79 11/30/94 0800 15 Tully Heiberg Forest 579.1 1/1/67 9/26/07 0700 40 Unadilla 2 N 451.1 6/1/75 12/12/14 0700 39 1800 / 10-1989 Utica 176.8 8/1/48 11/14/91 43 (1930) Utica 7 SSW 318.8 2/1/92 5/31/94 0900 2 Utica Oneida CO 216.7 12/1/50 1/17/07 2400 57 Airport Watertown 151.5 1/1/1893 2/28/17 0800 122 1700; 9-1998 Wellesley Island 86.9 7/1/74 6/30/05 31 (0700) Westmoreland 4 N 172.2 1/1/09 4/9/17 0700 6 Williamstown 214.9 1/1/32 2/28/05 0700 73
244
Appendix 8.2. Pearson correlations between the different environmental variables. Correlations are used to test for collinearity within the variables. Collinearity was considered for all variables with a correlation greater than 0.60. Variables include Lake Erie surface temperatures (EL.slt), Lake Ontario surface temperatures (OL.slt), Great Lakes surface temperatures (GL.slt), Lake Erie winter surface temperatures (EL.wlt), Lake Ontario winter surface temperatures (OL.wlt), Great Lakes winter surface temperatures (GL.wlt), Lake Erie ice cover (EL.ice), Lake Ontario ice cover (OL.ice), Great Lakes ice cover (GL.ice), minimum temperature ≤ 32˚C, minimum temperature ≤ 0˚C, maximum temperature ≤ 32˚C, average temperature (AvgT), average winter temperature (AvgWT), average maximum temperature (MaxT), average maximum winter temperature (MaxWT), average minimum temperature (MinT), average minimum winter temperature (MinWT), number of precipitation days (PcpDay), number of winter precipitation days (PcpDayW), seasonal precipitation (Sprecip), winter precipitation (Wprecip). EL.slt OL.slt GL.slt El.wlt OL.wlt GL.wlt EL.ice OL.ice GL.ice MinT.32 MinT.0 MaxT.32
EL.slt N/A
OL.slt 0.94 N/A
GL.slt 0.94 0.99 N/A
El.wlt 0.44 0.39 0.42 N/A
OL.wlt 0.50 0.44 0.45 0.79 N/A
GL.wlt 0.51 0.46 0.48 0.89 0.96 N/A
EL.ice -0.36 -0.25 -0.28 -0.85 -0.83 -0.80 N/A
OL.ice -0.54 -0.47 -0.49 -0.77 -0.87 -0.82 0.90 N/A GL.ice -0.26 -0.10 -0.13 -0.74 -0.80 -0.79 0.89 0.90 N/A MinT.32 -0.45 -0.42 -0.44 -0.42 -0.65 -0.57 0.79 0.72 0.77 N/A MinT.0 -0.23 -0.21 -0.23 -0.47 -0.63 -0.56 0.43 0.54 0.33 0.35 N/A MaxT.32 -0.47 -0.37 -0.40 -0.51 -0.76 -0.64 0.93 0.87 0.73 0.69 0.53 N/A AvgT 0.44 0.42 0.44 0.57 0.86 -0.74 -0.81 -0.88 -0.70 -0.62 -0.63 -0.87 AvgWT 0.50 0.44 0.46 0.58 0.83 -0.73 -0.88 -0.84 -0.73 -0.82 -0.65 -0.90 MaxT 0.43 0.41 0.44 0.54 0.83 0.73 -0.83 -0.87 -0.67 -0.60 -0.56 -0.86 MaxWT 0.48 0.42 0.45 0.55 0.79 0.70 -0.88 -0.81 -0.69 -0.81 -0.55 -0.89 MinT 0.42 0.41 0.42 0.56 0.86 0.77 -0.78 -0.89 -0.73 -0.63 -0.67 -0.84 MinWT 0.48 0.43 0.44 0.59 0.83 0.74 -0.87 -0.86 -0.76 -0.79 -0.72 -0.87 PcpDay -0.06 -0.01 0.00 -0.16 -0.22 -0.24 0.28 -0.05 0.31 0.22 0.07 0.08 PcpDayW 0.00 0.06 0.05 -0.31 -0.34 -0.36 0.25 0.01 0.16 0.15 0.20 0.19 Sprecip -0.03 -0.01 0.00 -0.15 0.04 -0.03 0.21 0.06 0.00 -0.03 0.11 0.07 Wprecip 0.14 0.20 0.18 -0.13 0.06 -0.03 0.30 0.01 0.22 0.12 -0.02 0.03
APPENDIX 8.2 CONTNUED AvgT AvgWT MaxT MaxWT MinT MinWT PcpDay PcpDayW Sprecip Wprecip
EL.slt
OL.slt
GL.slt
El.wlt
OL.wlt
GL.wlt
EL.ice
OL.ice
GL.ice
245
APPENDIX 8.2 CONTINUED MinT.32
MinT.0
MaxT.32
AvgT N/A
AvgWT 0.88 N/A
MaxT 0.98 0.86 N/A
MaxWT 0.88 0.98 0.89 N/A
MinT 0.98 0.88 0.94 0.85 N/A
MinWT 0.84 0.98 0.78 0.92 0.87 N/A
PcpDay -0.10 -0.17 -0.10 -0.18 -0.07 -0.16 N/A
PcpDayW -0.23 -0.21 -0.21 -0.22 -0.20 -0.19 0.85 N/A
Sprecip -0.02 -0.09 -0.03 -0.09 0.01 -0.12 0.58 0.55 N/A Wprecip -0.05 -0.15 -0.09 -0.20 0.01 -0.11 0.51 0.48 0.70 N/A
Appendix 8.3. Lake temperature variables that significantly explain variance within the snowfall contributions of different snowstorm types within Central New York. The variables include: average Lake Erie surface temperature (LE Temp), average Lake Ontario surface temperature (LO Temp), average Great Lakes surface temperature (GL Temp), average Lake Erie winter surface temperature (LE WTemp), average Lake Ontario winter surface temperature (LO WTemp), average Great Lakes winter surface temperature (GL WTemp). Storm Region LE Temp LO Temp GL Temp LE WTemp LO Wtemp GL Wtemp 1 2 Clipper 3 None 4 5 1 2 Colorado 3 None lows 4 5 1 2 Frontal 3 None 4 5 1 2 G.Lakes 3 None Low 4 5 1 2 Hudson low 3 None 4 5
246
APPENDIX 8.3 CONTINUED 1 2 Oklahoma 3 0.046 hook 4 0.069 0.065 5 1 2 Texas hook 3 4 5 0.084 0.044 0.022 0.007 0.009 1 0.092 2 Upper 3 0.075 0.032 0.060 Disturbance 4 5 1 0.012 0.064 2 Lake Snow 3 4 5 1 2 0.079 Non-Lake 3 Snow 4 5 1 0.016 0.065 2 0.038 Total 3 4 5 1 0.028 2 Lake-effect 3 Snow 4 5 1 2 Nor'easter 3 None 4 5 1 2 Canadian 3 None lows 4 5
247
APPENDIX 8.3 CONTINUE D 1 2 Rocky lows 3 0.024 0.013 0.073 4 0.097 0.057 5 0.057 0.081 1 Non- 2 Cyclonic 3 None Storms 4 5 df = 13
Appendix 8.4. Lake ice cover variables that significantly explain variance within the snowfall contributions of different snowstorm types within Central New York. The variables include average Lake Erie percent ice cover (LE ice), average Lake Ontario percent ice cover (LO Ice), and average Great Lakes percent ice cover (GL Ice). Storm Region LE Ice LO Ice GL Ice 1 2 Clipper 3 None 4 5 1 0.079 2 0.073 0.039 Colorado lows 3 4 5 1 2 Frontal 3 4 0.087 5 1 2 G.Lakes Low 3 None 4 5 1 2 Hudson low 3 None 4 5 1 2 Oklahoma hook 3 4 5 0.054
248
APPENDIX 8.4 CONTINUED 1 2 Texas hook 3 None 4 5 1 0.062 2 0.044 0.044 Upper 3 Disturbance 4 5 1 0.093 0.091 2 Lake Snow 3 4 5 1 2 0.022 0.068 Non-Lake Snow 3 4 5 1 0.007 0.033 2 0.003 0.014 Total 3 4 5 1 2 Lake-effect Snow 3 None 4 5 1 0.009 0.037 2 0.007 0.030 Nor'easter 3 0.079 4 5 1 2 Canadian lows 3 None 4 5 1 2 Rocky lows 3 4 5 0.023
249
APPENDIX 8.4 CONTINUED 1 0.032 2 0.027 0.018 Non-Cyclonic 3 Storms 4 5 df = 3
Appendix 8.5. Environmental variables that significantly explain variance within the snowfall contributions of different snowstorm types within Central New York. The variables include the: number of days the minimum temperature is 0⁰C (MinT.0), the number of the days the minimum temperature is -17.8⁰ (MinT.-17.8), number of the days the maximum temperature was at most 0⁰C (MaxT.0), average temperature (AvgT), average winter temperature (AvgWT), average maximum temperature (MaxT), average maximum winter temperature (MaxWT), number of precipitation days (PcpDy), number of winter precipitation days (WPcpDy), seasonal precipitation (Precip), and winter precipitation (WPrec). Storm Region MinT.0 MinT.-17.8 MaxT.0 AvgT AvgWT 1 2 0.082 0.050 Clipper 3 0.003 0.000 0.021 4 0.001 0.000 0.080 5 0.006 0.002 0.012 1 2 Colorado 3 0.012 lows 4 5 1 2 Frontal 3 None 4 5 1 2 G.Lakes 3 Low 4 5 1 2 Hudson low 3 0.089 4 5 1 2 Oklahoma 3 None hook 4 5
250
APPENDIX 8.5 CONTINUED 1 2 Texas hook 3 None 4 5 1 0.071 0.058 2 0.042 Upper 3 Disturbance 4 0.047 5 1 2 Lake Snow 3 0.013 0.014 4 5 0.092 1 2 Non-Lake 3 Snow 4 0.048 5 1 2 Total 3 0.058 0.051 4 0.048 5 0.082 1 2 0.082 0.090 Lake-effect 3 0.011 0.007 Snow 4 5 0.036 1 2 Nor'easter 3 4 0.022 5 0.045 1 2 0.085 Canadian 3 lows 4 0.065 5 0.016 0.073 1 2 0.075 Rocky lows 3 4 5 0.079
251
APPENDIX 8.5 CONTINUED 1 Non- 2 Cyclonic 3 None Storms 4 5
Appendix 8.5 Continued Part II WPcp Storm Region MaxT MaxWT MinT MinWT PcpDy Prec WPrec Dy 1 2 0.084 Clipper 3 0.014 0.053 0.051 4 0.062 0.003 0.005 0.081 5 0.044 0.022 0.046 0.058 1 2 Colorado 3 lows 4 5 0.058 1 2 Frontal 3 None 4 5 1 2 G.Lakes 3 0.061 0.016 Low 4 5 1 2 0.025 Hudson low 3 0.055 0.052 4 5 1 2 Oklahoma 3 None hook 4 5 1 2 Texas hook 3 None 4 5
252
APPENDIX 8.5 PART II CONTINUED 1 0.075 0.098 2 Upper 3 Disturbance 4 5 1 2 Lake Snow 3 0.010 4 5 1 2 Non-Lake 3 Snow 4 0.034 5 0.076 1 2 Total 3 4 0.063 0.041 0.031 5 1 2 0.077 Lake-effect 3 0.094 0.028 Snow 4 5 1 2 Nor'easter 3 4 0.028 0.087 0.072 5 0.032 1 2 0.075 Canadian 3 lows 4 0.049 5 1 2 Rocky lows 3 4 5 0.040 1 Non- 2 Cyclonic 3 None Storms 4 5
253
Appendix 8.6. Pearson correlations between the different teleconnection patterns. Correlations are used to test for collinearity within the variables. Collinearity was considered for all variables with a correlation greater than 0.60. AO EA ENSO 3 ENSO 3.4 ENSO 4 NAO PDO PNA WP AO 1.00 EA -0.14 1.00 ENSO 3 0.08 0.18 1.00 ENSO 3.4 -0.01 0.22 0.85** 1.00 ENSO 4 -0.01 0.25 0.90* 0.99** 1.00 NAO 0.75** -0.32 0.09 0.08 0.04 1.00 PDO -0.23 -0.10 0.14 0.41** 0.35 0.11 1.00 PNA -0.41* -0.01 -0.17 -0.06 -0.07 -0.23 0.27 1.00 WP 0.01 -0.08 -0.38* -0.12 -0.17 0.10 0.07 0.23 1.00 *ρ < 0.05 **ρ < 0.01
254
9.0 REFERENCES Agee, E. M., and S. R. Gilbert, 1989: An aircraft investigation of mesoscale convection over Lake Michigan during the 10 January 1984 cold air outbreak. J. Atmos. Sci., 46, 1877– 1897, doi:10.1175/1520-0469(1989)046%3C1877:AAIOMC%3E2.0.CO;2. http://cat.inist.fr/?aModele=afficheN&cpsidt=19401815. Akaike, H., 1974: A new look at the statistical model identification. IEEE Trans. Automat. Contr., 19, 716–723. Alcott, T. I., and W. J. Steenburgh, 2013: Orographic Influences on a Great Salt Lake–Effect Snowstorm. Mon. Weather Rev., 141, 2432–2450, doi:10.1175/MWR-D-12-00328.1. http://journals.ametsoc.org/doi/abs/10.1175/MWR-D-12-00328.1. Allen, R. J., and C. S. Zender, 2011: Forcing of the Arctic Oscillation by Eurasian snow cover. J. Clim., 24, 6528–6539, doi:10.1175/2011JCLI4157.1. Andersen, T. K., and M. J. Shepherd, 2013: Floods in a changing climate. Geogr. Compass, 7, 95–115, doi:10.1111/gec3.12025. Andreescu, M. P., and D. B. Frost, 1998: Weather and traffic accidents in Montreal, Canada. Clim. Res., 9, 225–230, doi:10.3354/cr009225. Angel, J. R., and S. A. Isard, 1997: An Observational Study of the Influence of the Great Lakes on the Speed and Intensity of Passing Cyclones. Mon. Weather Rev., 125, 2228–2237, doi:10.1175/1520-0493(1997)125<2228:AOSOTI>2.0.CO;2. ——, and ——, 1998: The frequency and intensity of Great Lake cyclones. J. Clim., 11, 61– 71, doi:10.1175/1520-0442(1998)011<0061:TFAIOG>2.0.CO;2. Archambault, H. M., L. F. Bosart, D. Keyser, and A. R. Aiyyer, 2008: Influence of Large-Scale Flow Regimes on Cool-Season Precipitation in the Northeastern United States. Mon. Weather Rev., 136, 2945–2963, doi:10.1175/2007MWR2308.1. http://journals.ametsoc.org/doi/abs/10.1175/2007MWR2308.1. Armal, S., N. Devineni, and R. Khanbilvardi, 2018: Trends in extreme rainfall frequency in the contiguous United States: Attribution to climate change and climate variability modes. J. Clim., 31, 369–385, doi:10.1175/JCLI-D-17-0106.1. Arnold, T. W., 2010: Uninformative Parameters and Model Selection Using Akaike’s Information Criterion. J. Wildl. Manage., 74, 1175–1178, doi:10.2193/2009-367. http://www.bioone.org/doi/abs/10.2193/2009-367. Assel, R. A., K. Cronk, and D. Norton, 2003: Recent Trends in Laurentian Great Lakes Ice Cover. Clim. Chang., 57, 185–204. ——, 2003: Great Lakes ice cover, first ice, last ice, and ice duration: Winters 1973-2002. Ann Arbor, MI, 1–49 pp. Athanasiadis, P. J., and Coauthors, 2017: A multisystem view of wintertime NAO seasonal predictions. J. Clim., 30, 1461–1475, doi:10.1175/JCLI-D-16-0153.1. Austin, J. M., and R. H. Blackmer Jr., 1956: The variability of cold front precipitation. Bull. Am. Meteorol. Soc., 37, 447–453, doi:10.1126/science.58.1504.335. http://www.sciencemag.org/cgi/doi/10.1126/science.58.1504.335. Bai, X., J. Wang, C. Sellinger, A. Clites, and R. Assel, 2012: Interannual variability of Great Lakes ice cover and its relationship to NAO and ENSO. J. Geophys. Res. Ocean., 117, C03002, doi:10.1029/2010JC006932. Baigorria, G. A., and J. W. Jones, 2010: GiST: A stochastic model for generating spatially and temporally correlated daily rainfall data. J. Clim., 23, 5990–6008, doi:10.1175/2010JCLI3537.1. 255
Ballentine, R. J., A. J. Stamm, E. E. Chermack, G. P. Byrd, and D. Schleede, 1998: Mesoscale Model Simulation of the 4–5 January 1995 Lake-Effect Snowstorm. Weather Forecast., 13, 893–920, doi:10.1175/1520-0434(1998)013<0893:MMSOTJ>2.0.CO;2. Bard, L., and D. A. R. Kristovich, 2012: Trend reversal in Lake Michigan contribution to snowfall. J. Appl. Meteorol. Climatol., 51, 2038–2046, doi:10.1175/JAMC-D-12-064.1. Barnes, E. A., and I. R. Simpson, 2017: Seasonal sensitivity of the Northern Hemisphere jet streams to Arctic temperatures on subseasonal time scales. J. Clim., 30, 10117–10137, doi:10.1175/JCLI-D-17-0299.1. Barnett, T. P., J. C. Adam, and D. P. Lettenmaier, 2005: Potential impacts of a warming climate on water availability in snow-dominated regions. Nature, 438, 303–309, doi:10.1038/nature04141. http://www.ncbi.nlm.nih.gov/pubmed/16292301. Barnston, A. G., 2015: Evolution of ENSO Prediction over the Past 40 Years. Sci. Technol. Infus. Clim. Bull., 26–29. ——, and R. E. Livezey, 1987: Classification, Seasonality and Persistence of Low-Frequency Atmospheric Circulation Patterns. Mon. Weather Rev., 115, 1083–1126, doi:10.1175/1520-0493(1987)115<1083:CSAPOL>2.0.CO;2. http://journals.ametsoc.org/doi/abs/10.1175/1520- 0493%281987%29115%3C1083%3ACSAPOL%3E2.0.CO%3B2. Barriopedro, D., R. García-Herrera, A. R. Lupo, and E. Hernández, 2006: A climatology of Northern Hemisphere blocking. J. Clim., 19, 1042–1063, doi:10.1175/JCLI3678.1. Barros, A. P., and R. J. Kuligowski, 1998: Orographic Effects during a Severe Wintertime Rainstorm in the Appalachian Mountains. Mon. Weather Rev., 126, 2648–2672, doi:10.1175/1520-0493(1998)126<2648:OEDASW>2.0.CO;2. Barthold, F. E., and D. a. R. Kristovich, 2011: Observations of the Cross-Lake Cloud and Snow Evolution in a Lake-Effect Snow Event. Mon. Weather Rev., 139, 2386–2398, doi:10.1175/MWR-D-10-05001.1. Basist, A., G. D. Bell, and V. Meentemeyer, 1994: Statistical relationships between topography and precipitaiton patterns. J. Clim., 7, 1305–1315. Bates, G. T., F. Giorgi, and S. W. Hostetler, 1993: Toward the Simulation of the Effects of the Great Lakes on Regional Climate. Mon. Weather Rev., 121, 1373–1387, doi:10.1175/1520-0493(1993)121<1373:TTSOTE>2.0.CO;2. http://journals.ametsoc.org/doi/abs/10.1175/1520- 0493%281993%29121%3C1373%3ATTSOTE%3E2.0.CO%3B2. Baxter, M. A., C. E. Graves, and J. T. Moore, 2005: A Climatology of Snow-to-Liquid Ratio for the Contiguous United States. Weather Forecast., 20, 729–744, doi:10.1175/WAF856.1. Baxter, S., and S. Nigam, 2015: Key role of the North Pacific oscillation-West Pacific pattern in generating the extreme 2013/14 North American Winter. J. Clim., 28, 8109–8117, doi:10.1175/JCLI-D-14-00726.1. ——, S. Weaver, J. Gottschalck, and Y. Xue, 2014: Pentad evolution of wintertime impacts of the Madden-Julian oscillation over the contiguous United States. J. Clim., 27, 7356– 7367, doi:10.1175/JCLI-D-14-00105.1. Beaulieu, C., and R. Killick, 2018: Distinguishing trends and shifts from memory in climate data. J. Clim., 31, 9519–9543, doi:10.1175/JCLI-D-17-0863.1. Beckman, S. K., 1987: Use of enhanced IR/visible satellite imagery to determine heavy snow areas. Mon. Weather Rev., 115, 2060–2087. Belmecheri, S., F. Babst, A. R. Hudson, J. Betancourt, and V. Trouet, 2017: Northern 256
Hemisphere jet stream position indices as diagnostic tools for climate and ecosystem dynamics. Earth Interact., 21, 1–23, doi:10.1175/EI-D-16-0023.1. Bentley, M., and S. Horstmeyer, 1998: The Witch of November. Weatherwise, 51, 29–35, doi:10.1080/00431672.1998.9926174. Berger, C., A. Lupo, P. Browning, M. Bodner, M. Chambers, and C. Rayburn, 2002: A Climatology of Northwest Missouri Snowfall Events: Long-Term Trends and Interannual Variability. Phys. Geogr., 23, 427–448, doi:10.2747/0272-3646.23.6.427. Bjerknes, J., 1969: Atmospheric Teleconnections From the Equatorial Pacific. Mon. Weather Rev., 97, 163–172, doi:10.1175/1520-0493(1969)097<0163:ATFTEP>2.3.CO;2. Blackmon, M. L., 1976: A climatological spectral study of the 500 mb geopotential height of the Northern Hemisphere. J. Atmos. Sci., 33, 1607–1623. Bolsenga, S. J., and D. C. Norton, 1993: Great Lakes Air Temperature Trends For Land Stations, 1901-1987. J. Great Lakes Res., 19, 379–388. Bonan, G. B., 1995: Sensitivity of a GCM simulation to inclusion of inland water surfaces. J. Clim., 8, 2691–2704, doi:10.1175/1520-0442(1995)008<2691:SOAGST>2.0.CO;2. Bosart, L. F., 1973: Detailed Analyses of Precipitation Patterns Associated With Mesoscale Features Accompanying United States East Coast Cyclogenesis. Mon. Weather Rev., 101, 1–12. Boudra, D. B., 1981: A study of the early winter effects of the Great Lakes. I: Comparison of very fine scale numerical simulations with observed data. Mon. Weather Rev., 109, 2507–2526. Bradley, B. A., D. S. Wilcove, and M. Oppenheimer, 2010: Climate change increases risk of plant invasion in the Eastern United States. Biol. Invasions, 12, 1855–1872, doi:10.1007/s10530-009-9597-y. Braham, R. R., 1983: The Midwest snow storm of 8-11 December 1977. Mon. Weather Rev., 111, 253–272, doi:10.1175/1520-0493(1983)111<0253:TMSSOD>2.0.CO;2. ——, and M. J. Dungey, 1984: Quantitative estimates of the effect of Lake Michigan on snowfall. J. Clim. Appl. Meteorol., 23, 940–949, doi:10.1175/1520- 0450(1984)023<0940:QEOTEO>2.0.CO;2. ——, and ——, 1995: Lake-effect snowfall over Lake Michigan. J. Appl. Meteorol., 34, 1009– 1019, doi:10.1175/1520-0450(1995)034<1009:LESOLM>2.0.CO;2. http://www.scopus.com/inward/record.url?eid=2-s2.0- 0028975107&partnerID=tZOtx3y1. Branick, M. L., 1997: A Climatology of Significant Winter-Type Weather Events in the Contiguous United States, 1982–94. Weather Forecast., 12, 193–207, doi:10.1175/1520-0434(1998)013<0884:COACOS>2.0.CO;2. van den Brink, H. W., G. P. Koennen, and J. D. Opsteegh, 2004: Statistics of synoptic-scale wind speeds in ensemble simulations of current and future climate. J. Clim., 17, 4564– 4574. Brown, M. J., and E. L. Peck, 1962: Reliability of precipitation measurements as related to exposure. J. Appl. Meteorol., 1, 203–207. Brown, P. J., and A. T. DeGaetano, 2011: A paradox of cooling winter soil surface temperatures in a warming northeastern United States. Agric. For. Meteorol., 151, 947–956, doi:10.1016/j.agrformet.2011.02.014. http://dx.doi.org/10.1016/j.agrformet.2011.02.014. Brown, R. D., 2000: Northern Hemisphere snow cover variability and change, 1915-97. J. 257
Clim., 13, 2339–2355, doi:10.1175/1520-0442(2000)013<2339:NHSCVA>2.0.CO;2. Brunner, L., G. C. Hegerl, and A. K. Steiner, 2017: Connecting atmospheric blocking to European temperature extremes in spring. J. Clim., 30, 585–594, doi:10.1175/JCLI-D- 16-0518.1. Bumbaco, K. A., and P. W. Mote, 2010: Three recent flavors of drought in the Pacific Northwest. J. Appl. Meteorol. Climatol., 49, 2058–2068, doi:10.1175/2010JAMC2423.1. Burakowski, E., and M. Magnusson, 2012: Climate Impacts on the Winter Tourism Economy in the United States. New York, NY,. Burnett, A. W., M. E. Kirby, H. T. Mullins, and W. P. Patterson, 2003: Increasing Great Lake- effect snowfall during the twentieth century: A regional response to global warming? J. Clim., 16, 3535–3542, doi:10.1175/1520-0442(2003)016<3535:IGLSDT>2.0.CO;2. Burnham, K. P., and D. R. Anderson, 2002: Model selection and multimodel inference: a practical information-theoretic approach. 2nd ed. Springer, New York, NY,. Cai, M., and H. M. van den Dool, 1992: Frequency waves and traveling storm tracks. Part II: Three-dimensional structure. J. Atmos. Sci., 49, 2506–2524. Call, D. A., 2005: Rethinking snowstorms as snow events: A regional case study from upstate New York. Bull. Am. Meteorol. Soc., 86, 1783–1793, doi:10.1175/BAMS-86-12- 1783. Campbell, J. L., A. M. Socci, and P. H. Templer, 2014: Increased nitrogen leaching following soil freezing is due to decreased root uptake in a northern hardwood forest. Glob. Chang. Biol., 20, 2663–2673, doi:10.1111/gcb.12532. Campbell, L. S., W. J. Steenburgh, P. G. Veals, T. W. Letcher, and J. R. Minder, 2016: Lake- Effect Mode and Precipitation Enhancement over the Tug Hill Plateau during OWLeS IOP2b. Mon. Weather Rev., 144, 1729–1748, doi:10.1175/MWR-D-15-0412.1. http://journals.ametsoc.org/doi/10.1175/MWR-D-15-0412.1. Carleton, A. M., 1999: Methodology in Climatology. Ann. Assoc. Am. Geogr., 89, 713–735. Carrera, M. L., R. W. Higgins, and V. E. Kousky, 2004: Downstream weather impacts associated with atmospheric blocking over the Northeast Pacific. J. Clim., 17, 4823– 4839, doi:10.1175/JCLI-3237.1. Cayan, D. R., 1996: Interannual climate variability and snowpack in the western United States. J. Clim., 9, 928–948, doi:10.1175/1520-0442(1996)009<0928:ICVASI>2.0.CO;2. Cerruti, B. J., and S. G. Decker, 2011: The local winter storm scale: A measure of the intrinsic ability of winter storms to disrupt society. Bull. Am. Meteorol. Soc., 92, 721–737, doi:10.1175/2010BAMS3191.1. Chang, E. K. M., 2006: An idealized nonlinear model of the Northern Hemisphere winter storm tracks. J. Atmos. Sci., 63, 1818–1839, doi:10.1175/JAS3726.1. http://journals.ametsoc.org/doi/abs/10.1175/JAS3726.1. Changnon, D., 2018: A Spatial and Temporal Analysis of 30-Day Heavy Snowfall Amounts in the Eastern United States , 1900 – 2016. J. Appl. Meteorol. Climatol., 57, 319–331, doi:10.1175/JAMC-D-17-0217.1. ——, C. Merinsky, and M. Lawson, 2008: Climatology of Surface Cyclone Tracks Associated with Large Central and Eastern U.S. Snowstorms, 1950–2000. Mon. Weather Rev., 136, 3193–3202, doi:10.1175/2008MWR2324.1. Changnon, S. A., 1969: Climatology of Severe Winter Storms in Illinois. Champaign, IL, 1–42 pp. http://www.isws.uiuc.edu/pubdoc/B/ISWSB-53.pdf. ——, and D. M. A. Jones, 1972: Review the Influences of the Great Lakes on Weather. Water 258
Resour. Res., 8, 360–371. ——, and D. Changnon, 2006: A spatial and temporal analysis of damaging snowstorms in the United States. Nat. Hazards, 37, 373–389, doi:10.1007/s11069-005-6581-4. ——, ——, and T. R. Karl, 2006: Temporal and Spatial Variations of Freezing Rain in the Contiguous United States. J. Appl. Meteorol. Climatol., 45, 1141–1155. ——, and A. Sharma, 2009: Long-range Niño-3.4 predictions using pairwise dynamic combinations of multiple models. J. Clim., 22, 793–805, doi:10.1175/2008JCLI2210.1. Christiansen, D. E., S. L. Markstrom, and L. E. Hay, 2011: Impacts of climate change on the growing season in the United States. Earth Interact., 15, 1–17, doi:10.1175/2011EI376.1. Chuang, H.-Y., and P. J. Sousounis, 2003: The Impact of the Prevailing Synoptic Situation on the Lake-Aggregate Effect. Mon. Weather Rev., 131, 990–1010, doi:10.1175/1520- 0493(2003)131<0990:tiotps>2.0.co;2. Chung, Y.-S., K. D. Hage, and E. R. Reinelt, 1976: On Lee Cyclogenesis and Airflow in the Canadian Rocky Mountains and the East Asian Mountains. Mon. Weather Rev., 104, 879–891, doi:10.1175/1520-0493(1976)104<0879:OLCAAI>2.0.CO;2. Cione, J. J., S. Raman, and L. J. Pietrafesa, 1993: The Effect of Gulf Stream-induced Baroclinicity on U.S. East Coast Winter Cyclones. Mon. Weather Rev., 121, 421–430, doi:10.1175/1520-0493(1993)121<0421:TEOGSI>2.0.CO;2. Clark, C. A., and Coauthors, 2016: Spatiotemporal snowfall variability in the Lake Michigan Region: How is warming affecting wintertime snowfall? J. Appl. Meteorol. Climatol., 55, 1813–1830, doi:10.1175/JAMC-D-15-0285.1. Clark, J. H. E., 1990: An observational and theoretical study of Colorado lee cyclogenesis. J. Atmos. Sci., 47, 1541–1561. Coleman, J. S. M., and J. C. Rogers, 2003: Ohio River Valley winter moisture conditions associated with the Pacific-North American teleconnection pattern. J. Clim., 16, 969– 981, doi:10.1175/1520-0442(2003)016<0969:ORVWMC>2.0.CO;2. ——, and D. Budikova, 2013: Eastern U.S. summer streamflow during extreme phases of the North Atlantic oscillation. J. Geophys. Res. Atmos., 118. Collins, M., and Coauthors, Long-term climate change: Projections, commitments and irreversibility. Climate Change 2013: The Physical Science Basis. IPCC Working Group I Contribution to AR5, IPCC, Ed., Cambridge University Press, Cambridge, UK, 1031– 1106. Colucci, S. J., 1976: Winter Cyclone Frequencies over the Eastern United States and Adjacent Western Atlantic, 1964-1973. Bull. Am. Meteorol. Soc., 57, 548–553, doi:10.1175/1520-0477(1976)057<0548:WCFOTE>2.0.CO;2. Cordeira, J. M., and N. F. Laird, 2008: The Influence of Ice Cover on Two Lake-Effect Snow Events over Lake Erie. Mon. Weather Rev., 136, 2747–2763, doi:10.1175/2007MWR2310.1. http://journals.ametsoc.org/doi/abs/10.1175/2007MWR2310.1. Cortinas, M. R., and U. Kitron, 2006: County-level surveillance of white-tailed deer infestation by Ixodes scapularis and Dermacentor albipictus (Acari: Ixodidae) along the Illinois River. J. Med. Entomol., 43, 810–819, doi:10.1603/0022-2585(2006)43. Coumou, D., J. Lehmann, and J. Beckmann, 2015: The weakening summer circulation in the Northern Hemisphere mid-latitudes. Science (80-. )., 348, 324–327. Cox, M. K., 1959: The Distribution and Variability of Cold-Front Precipitation. Bull. Am. 259
Meteorol. Soc., 40, 477–480. CPC, 2005: West Pacific. National Weather Service. Accessed 10 October 2018. http://www.cpc.ncep.noaa.gov/data/teledoc/wp.shtml ——, 2012: Northern Hemisphere Teleconnection Patterns. https://www.cpc.ncep.noaa.gov/data/teledoc/telecontents.shtml. ——, 2018: Teleconnection Introduction. http://www.cpc.ncep.noaa.gov/data/teledoc/teleintro.shtml. Croci-Maspoli, M., C. Schwierz, and H. C. Davies, 2007: A multifaceted climatology of atmospheric blocking and its recent linear trend. J. Clim., 20, 633–649, doi:10.1175/JCLI4029.1. CRU, 2007: CRUTEM3v Dataset. http://www.cru.uea.ac.uk/cru/data/temperature. Curry, J., 1983: On the Formation of Continental Polar Air. J. Atmos. Sci., 40, 2278–2292, doi:10.1175/1520-0469(1983)040<2278:OTFOCP>2.0.CO;2. http://journals.ametsoc.org/doi/abs/10.1175/1520- 0469%281983%29040%3C2278%3AOTFOCP%3E2.0.CO%3B2. Dai, A., and T. M. L. Wigley, 2000: Global patterns of ENSO-induced precipitation. Geophys. Res. Lett., 27, 1283–1286. ——, T. M. . Wigley, B. A. Boville, J. T. Kiehl, and L. E. Buja, 2001: Climates of the Ttwentieth and twenty-first centuries simulated by the NCAR Climate System Model. J. Clim., 14, 485–519, doi:10.1175/1520-0442(2001)014<0485:COTTAT>2.0.CO;2. papers2://publication/uuid/178C38D9-4B3F-4EAA-B16B-CECA542317E8. Daly, C., R. P. Neilson, and D. L. Phillips, 1994: A statistical-topographic model for mapping climatological precipitation over mountainous terrain. J. Appl. Meteorol., 33, 140–158. Danard, M. B., and A. C. McMillan, 1974: Further numberical studies of the effects of the Great Lakes on winter cyclones. Mon. Weather Rev., 102, 166–175. Davis, C. A., M. T. Stoelinga, and Y.-H. Kuo, 1993: The integrated effect of condensation in numerical simulations of extratropical cyclogenesis. Mon. Weather Rev., 121, 2309– 2330. Davis, R. E., and R. Dolan, 1993: Nor’easters. Am. Sci., 81, 428–439. ——, and S. Benkovic, 1994: Spatial and temporal variations of the January circumpolar vortex over the Northern Hemisphere. Int. J. Climatol., 14, 415–428. Delcambre, S. C., D. J. Lorenz, D. J. Vimont, and J. E. Martin, 2013: Diagnosing northern hemisphere jet portrayal in 17 CMIP3 global climate models: Twenty-first-century projections. J. Clim., 26, 4930–4946, doi:10.1175/JCLI-D-12-00359.1. Demaria, E. M. C., J. K. Roundy, S. Wi, and R. N. Palmer, 2016: The effects of climate change on seasonal snowpack and the hydrology of the Northeastern and Upper Midwest United States. J. Clim., 29, 6527–6541, doi:10.1175/JCLI-D-15-0632.1. Dewey, K. F., 1979a: Lake Erie induced mesosystems - an operational forecast model. Mon. Weather Rev., 107, 421–425. ——, 1979b: An Objective Forecast Method Developed for Lake Ontario Induced Snowfall Systems. J. Appl. Meteorol., 18, 787–793, doi:10.1175/1520- 0450(1979)018<0787:AOFMDF>2.0.CO;2. Diaz, H. F., R. S. Bradley, and J. K. Eischeid, 1989: Precipitation fluctuations over global land areas since the late 1800’s. J. Geophys. Res., 94, 1195–1210, doi:10.1029/JD094iD01p01195. Dickson, R. R., and J. Namias, 1976: North American Influences on the Circulation and 260
Climate of the North Atlantic Sector. Mon. Weather Rev., 104, 1255–1265, doi:10.1175/1520-0493(1976)104<1255:NAIOTC>2.0.CO;2. Dietz, T., and D. Bidwell, 2011: Climate Change in the Great Lakes Region. T. Dietz and D. Bidwell, Eds. Michigan State University Press, Lansing, MI, 278 pp. http://www.great- lakes.net/envt/refs/cchange.html. Dobson, A., and R. Carper, 1992: Global Warming and Potential Changes in Host-Parasite and Disease-Vector Relationships. R.L. Peters and T.E. Lovejoy, Eds. Yale University Press, New Haven, CT,. Doesken, N. J., and A. Judson, 1996: The snow booklet: a guide to the science, climatology and measurement of snow in the United States. Colorado State University, 85 pp. Douglas, A. V., D. R. Cayan, and J. Namias, 1982: Large-Scale Changes in North Pacific and North American Weather Patterns in Recent Decades. Mon. Weather Rev., 110, 1851– 1862, doi:10.1175/1520-0493(1982)110<1851:LSCINP>2.0.CO;2. Draxler, R. R., 1998: An overview of the HYSPLIT 4 modeling system for trajectories, dispersion, and deposition. Aust. Meteorol. Mag., 47, 295–308. ——, and G. D. Hess, 1997: Description of the HYSPLIT 4 modeling system. 24. ——, and G. D. Rolph, 2003: HYSPLIT (HYbrid Single-Particle Lagrangian Integrated Trajectory) Model. Durre, I., M. F. Squires, R. S. Vose, X. Yin, A. Arguez, and S. Applequist, 2013: NOAA’s 1981- 2010 U.S. climate normals: Monthly precipitation, snowfall, and snow depth. J. Appl. Meteorol. Climatol., 52, 2377–2395, doi:10.1175/JAMC-D-13-051.1. Dyer, J. L., and T. L. Mote, 2006: Spatial variability and trends in observed snow depth over North America. Geophys. Res. Lett., 33, 1–6, doi:10.1029/2006GL027258. Eichenlaub, V. L., 1970: Lake effect snowfall to the lee of the Great Lakes: Its role in Michigan. Bull. Am. Meteorol. Soc., 51, 403–412, doi:10.1175/1520- 0477(1970)051<0403:LESTTL>2.0.CO;2. http://journals.ametsoc.org/doi/abs/10.1175/1520- 0477%281970%29051%3C0403%3ALESTTL%3E2.0.CO%3B2. ——, and T. W. Hodler, 1979: Weather and Climate of the Great Lakes Region. University of Notre Dame Press, South Bend, IN, 335 pp. Eichler, T., and W. Higgins, 2006: Climatology and ENSO-related variability of North American extratropical cyclone activity. J. Clim., 19, 2076–2093, doi:10.1175/JCLI3725.1. Ekstrand, S., and P. Wallenberg, 2010: Climate change impact on nutrient loss in regions with pronounced winter seasons. J. Water Clim. Chang., 1, 181–192. Ellenton, G. E., and M. B. Danard, 1979: Inclusion of sensible heating in convective parameterization applied to lake-effect snow. Mon. Weather Rev., 107, 551–565. Ellis, A. W., and D. J. Leathers, 1996: A Synoptic Climatological Approach to the Analysis of Lake-Effect Snowfall: Potential Forecasting Applications. Weather Forecast., 11, 216– 229, doi:10.1175/1520-0434(1996)011<0216:ASCATT>2.0.CO;2. ——, and D. J. Leathers, 1998: A Quantitative Approach to Evaluating the Effects of Snow Cover on Cold Airmass Temperatures across the U.S. Great Plains. Weather Forecast., 13, 688–701, doi:10.1175/1520-0434(1998)013<0688:AQATET>2.0.CO;2. http://journals.ametsoc.org/doi/abs/10.1175/1520- 0434%281998%29013%3C0688%3AAQATET%3E2.0.CO%3B2. ——, and J. J. Johnson, 2004: Hydroclimatic Analysis of Snowfall Trends Associated with the 261
North American Great Lakes. J. Hydrometeorol., 5, 471–486, doi:10.1175/1525- 7541(2004)005<0471:HAOSTA>2.0.CO;2. Evans, C., and Coauthors, 2017: The Extratropical Transition of Tropical Cyclones. Part I: Cyclone Evolution and Direct Impacts. Mon. Weather Rev., 145, 4317–4344, doi:10.1175/MWR-D-17-0027.1. http://journals.ametsoc.org/doi/10.1175/MWR-D- 17-0027.1. Ewen, T., S. Brönnimann, and J. Annis, 2008: An extended Pacific-North American index from upper-air historical data back to 1922. J. Clim., 21, 1295–1308, doi:10.1175/2007JCLI1951.1. Eynon, B., 1988: Statistical analysis of precipitation chemistry measurements over the eastern United States. Part II: Kriging analysis of regional patterns and trends. J. Appl. Meteorol., 27, 1334–1343. Falk, M., 2010: A dynamic panel data analysis of snow depth and winter tourism. Tour. Manag., 31, 912–924, doi:10.1016/j.tourman.2009.11.010. http://dx.doi.org/10.1016/j.tourman.2009.11.010. Fereday, D. R., A. Maidens, A. Arribas, A. A. Scaife, and J. R. Knight, 2012: Seasonal forecasts of northern hemisphere winter 2009/10. Environ. Res. Lett., 7, 1–7, doi:10.1088/1748- 9326/7/3/034031. Fiebrich, C. A., and K. C. Crawford, 2009: Automation: A step toward improving the quality of daily temperature data produced by climate observing networks. J. Atmos. Ocean. Technol., 26, 1246–1260, doi:10.1175/2009JTECHA1241.1. Fleming, Z. L., P. S. Monks, and A. J. Manning, 2012: Review: Untangling the influence of air- mass history in interpreting observed atmospheric composition. Atmos. Res., 104, 1– 39. Fortier, C., A. A. Assani, M. Mesfioui, and A. G. Roy, 2011: Comparison of the interannual and interdecadal variability of heavy flood characteristics upstream and downstream from dams in inversed hydrologic regime: case study of Matawin River (Quebec, Canada). River Res. Appl., 27, 1277–1289, doi:10.1002/rra. Francis, J. A., and S. J. Vavrus, 2015: Evidence for a wavier jet stream in response to rapid Arctic warming. Environ. Res. Lett., 10, 14005, doi:10.1088/1748-9326/10/1/014005. http://dx.doi.org/10.1088/1748-9326/10/1/014005. Fuhrmann, C. M., and C. E. Konrad, 2013: A Trajectory Approach to Analyzing the Ingredients Associated with Heavy Winter Storms in Central North Carolina. Weather Forecast., 28, 647–667, doi:10.1175/WAF-D-12-00079.1. http://journals.ametsoc.org/doi/abs/10.1175/WAF-D-12-00079.1. Galway, J. G., and A. Pearson, 1981: Winter tornado outbreaks. Mon. Weather Rev., 109, 1072–1080. Gan, B., and L. Wu, 2015: Feedbacks of sea surface temperature to wintertime storm tracks in the North Atlantic. J. Clim., 28, 306–323, doi:10.1175/JCLI-D-13-00719.1. http://journals.ametsoc.org/doi/abs/10.1175/JCLI-D-13-00719.1. Gao, P., and J. J. Hartnett, 2016: Exploring the causes of an extreme flood event in Central New York, USA. Phys. Geogr., 37, 38–55, doi:10.1080/02723646.2016.1153332. Ge, Y., and G. Gong, 2009: North American snow depth and climate teleconnection patterns. J. Clim., 22, 217–233, doi:10.1175/2008JCLI2124.1. Geyer, C. J., 2003: Model Selection in R. 1–25. Ghatak, D., G. Gong, and A. Frei, 2010: North American temperature, snowfall, and snow- 262
Depth response to winter climate modes. J. Clim., 23, 2320–2332, doi:10.1175/2009JCLI3050.1. Giannini, A., M. A. Cane, and Y. Kushnir, 2001: Interdecadal changes in the ENSO Teleconnection to the Caribbean Region and the North Atlantic Oscillation. J. Clim., 14, 2867–2879, doi:10.1175/1520-0442(2001)014<2867:ICITET>2.0.CO;2. Giorgi, F., J. W. Hurrell, M. R. Marinucci, and M. Beniston, 1997: Elevation dependency of the surface climate change signal: A model study. J. Clim., 10, 288–296, doi:10.1175/1520- 0442(1997)010<0288:EDOTSC>2.0.CO;2. Glickman, T. S., ed., 2000: Glossary of meteorology. 2nd ed. American Meteorological Society, 855 pp. Goodrich, G. B., and J. M. Walker, 2011: The Influence of the PDO on Winter Precipitation During High- and Low-Index ENSO Conditions in the Eastern United States. Phys. Geogr., 32, 295–312, doi:10.2747/0272-3646.32.4.295. http://bellwether.metapress.com/openurl.asp?genre=article&id=doi:10.2747/0272- 3646.32.4.295. Goree, P. A., and R. J. Younkln, 1966: Synoptic climatology of heavy snowfall over the central and eastern United States. Mon. Weather Rev., 94, 663–668. Graham, N. E., and H. F. Diaz, 2001: Evidence for intensification of North Pacific winter cyclones since 1948. Bull. Am. Meteorol. Soc., 82, 1869–1893, doi:10.1175/1520- 0477(2001)082<1869:EFIONP>2.3.CO;2. Grimaldi, R., 2008: Climate teleconnections related to El Niño winters in a lake-effect region of west-central New York. Atmos. Sci. Lett., 9, 18–25, doi:10.1002/asl.166. Grippa, M., L. Kergoat, T. Le Toan, N. M. Mognard, N. Delbart, J. L’Hermitte, and S. M. Vicente-Serrano, 2005: The impact of snow depth and snowmelt on the vegetation variability over central Siberia. Geophys. Res. Lett., 32, 1–4, doi:10.1029/2005GL024286. Grise, K. M., S.-W. Son, and J. R. Gyakum, 2013: Intraseasonal and Interannual Variability in North American Storm Tracks and Its Relationship to Equatorial Pacific Variability. Mon. Weather Rev., 141, 3610–3625, doi:10.1175/MWR-D-12-00322.1. http://journals.ametsoc.org/doi/abs/10.1175/MWR-D-12-00322.1. Groisman, P. Y., and D. R. Easterling, 1994: Variability and trends of total precipitation and snowfall over the United States and Canada. J. Clim., 7, 184–205, doi:10.1175/1520- 0442(1994)007<0184:VATOTP>2.0.CO;2. ——, and Coauthors, 2004: Contemporary Changes of the Hydrological Cycle over the Contiguous United States: Trends Derived from In Situ Observations. J. Hydrometeorol., 5, 64–85, doi:10.1175/1525-7541(2004)005<0064:CCOTHC>2.0.CO;2. http://journals.ametsoc.org/doi/abs/10.1175/1525- 7541%282004%29005%3C0064%3ACCOTHC%3E2.0.CO%3B2. ——, R. W. Knight, D. R. Easterling, T. R. Karl, G. C. Hegerl, and V. N. Razuvaev, 2005: Trends in intense precipitation in the climate record. J. Clim., 18, 1326–1350, doi:10.5631/jibirin.98.367. http://joi.jlc.jst.go.jp/JST.Journalarchive/jibirin1925/98.367?from=CrossRef. ——, ——, and T. R. Karl, 2012: Changes in Intense Precipitation over the Central United States. J. Hydrometeorol., 13, 47–66, doi:10.1175/JHM-D-11-039.1. http://journals.ametsoc.org/doi/abs/10.1175/JHM-D-11-039.1. Grover, E. K., and P. J. Sousounis, 2002: The influence of large-scale flow on fall 263
precipitation systems in the Great Lakes Basin. J. Clim., 15, 1943–1956, doi:10.1175/1520-0442(2002)015<1943:TIOLSF>2.0.CO;2. Grünewald, T., Y. Bühler, and M. Lehning, 2014: Elevation dependency of mountain snow depth. Cryosphere, 8, 2381–2394, doi:10.5194/tc-8-2381-2014. Guan, H., J. L. Wilson, and O. Makhnin, 2005: Geostatistical Mapping of Mountain Precipitation Incorporating Autosearched Effects of Terrain and Climatic Characteristics. J. Hydrometeorol., 6, 1018–1031, doi:10.1175/JHM448.1. http://journals.ametsoc.org/doi/abs/10.1175/JHM448.1. Gula, J., and W. R. Peltier, 2012: Dynamical downscaling over the Great Lakes Basin of North America using the WRF regional climate model: The impact of the great lakes system on regional greenhouse warming. J. Clim., 25, 7723–7742, doi:10.1175/JCLI-D-11- 00388.1. Gutzler, D. S., and R. D. Rosen, 2002: Interannual Variability of Wintertime Snow Cover across the Northern Hemisphere. J. Clim., 5, 1441–1447, doi:10.1175/1520- 0442(1992)005<1441:ivowsc>2.0.co;2. ——, D. M. Kann, and C. Thornbrugh, 2002: Modulation of ENSO-Based Long-Lead Outlooks of Southwestern U.S. Winter Precipitation by the Pacific Decadal Oscillation. Weather Forecast., 17, 1163–1172, doi:10.1175/1520- 0434(2002)017<1163:MOEBLL>2.0.CO;2. http://journals.ametsoc.org/doi/full/10.1175/1520- 0434(2002)017%3C1163:MOEBLL%3E2.0.CO;2. Hagenstad, M., E. Burakowski, and R. Hill, 2018: The Economic Contributions of Winter Sports in a Changing Climate. Prot. Our Winters, 1–69. Hakim, G. J., 2000: Climatology of Coherent Structures on the Extratropical Tropopause. Mon. Weather Rev., 128, 385–406, doi:10.1175/1520- 0493(2000)128<0385:COCSOT>2.0.CO;2. Hansen, J., M. Sato, R. Ruedy, K. Lo, D. W. Lea, and M. Medina-elizade, 2006: Global temperature change. PNAS, 103, 14288–14293, doi:10.1073/pnas.0606291103. Hanson, H. P., C. S. Hanson, and B. H. Yoo, 1992: Recent Great Lakes Ice Trends. Bull. Am. Meteorol. Soc., 73, 577–584, doi:10.1175/1520- 0477(1992)073<0577:RGLIT>2.0.CO;2. http://link.springer.com/referenceworkentry/10.1007/978-90-481-2642- 2_316%5Cnhttp://link.springer.com.libproxy.ucl.ac.uk/referenceworkentry/10.1007/ 978-90-481-2642-2_316. Harms, R. W., 1973: Snow forecasting for southwestern Wisconsin. Weatherwise, 26, 250– 255. Harrington, J. A., Jr., R. S. Cerveny, and K. F. Dewey, 1987: A climatology of mean monthly snowfall for the conterminous United States: Temporal and spatial patterns. J. Clim. Appl. Meteorol., 26, 897–912. Harris, J. M., and J. D. Kahl, 1990: A descriptive atmospheric transport climatology for the Mauna Loa Observatory using clustered trajectories. J. Geophys. Res., 13651–13667. Hart, R. E., and J. L. Evans, 2001: A climatology of the extratropical transition of Atlantic tropical cyclones. J. Clim., 14, 546–564, doi:10.1175/1520- 0442(2001)014<0546:ACOTET>2.0.CO;2. Hartjenstein, G., and R. Bleck, 1991: Factors Affecting Cold-Air Outbreaks East of the Rocky Mountains. Mon. Weather Rev., 119, 2280–2292, doi:10.1175/1520- 264
0493(1991)119<2280:FACAOE>2.0.CO;2. Hartley, S., and M. J. Keables, 1998: Synoptic associations of winter climate and snowfall variability in New England, USA. Int. J. Climatol., 18, 281–298. Hartmann, D. L., and Coauthors, 2013: Observations: atmosphere and surface. Climate Change 2013: The Physical Science Basis. IPCC Working Group I Contribution to AR5, T.F. Stocker et al., Eds., Cambridge University Press, Cambridge, UK. Hartnett, J. J., J. M. Collins, M. A. Baxter, and D. P. Chambers, 2014: Spatiotemporal snowfall trends in Central New York. J. Appl. Meteorol. Climatol., 53, 2685–2697, doi:10.1175/JAMC-D-14-0084.1. Harvell, C. D., C. E. Mitchell, J. R. Ward, S. Altizer, A. P. Dobson, R. S. Ostfeld, and M. D. Samuel, 2002: Climate Warming and Disease Risks for Terrestrial and Marine Biota. Science (80-. )., 296, 2158–2163, doi:10.1126/science.1063699. Hawkins, T. W., and A. W. Ellis, 2002: Intra-annual analysis of the North American snow cover-monsoon teleconnection: Seasonal forecasting utility. J. Clim., 15, 1743–1753, doi:10.1175/1520-0442(2002)015<1743:IAAOTN>2.0.CO;2. Hayden, B. P., 1981: Secular Variation in Atlantic Coast Extratropical Cyclones. Mon. Weather Rev., 109, 159–167, doi:10.1175/1520- 0493(1981)109<0159:SVIACE>2.0.CO;2. Henderson, G. R., and D. J. Leathers, 2010: European snow cover extent variability and associations with atmospheric forcings. Int. J. Climatol., 30, 1440–1451, doi:10.1002/joc.1990. Hidalgo, H. G., and J. A. Dracup, 2003: ENSO and PDO Effects on Hydroclimatic Variations of the Upper Colorado River Basin. J. Hydrometeorol., 4, 5–23, doi:10.1175/1525- 7541(2003)004<0005:EAPEOH>2.0.CO;2. http://journals.ametsoc.org/doi/abs/10.1175/1525- 7541%282003%29004%3C0005%3AEAPEOH%3E2.0.CO%3B2. Hill, J. D., 1971: Snow Squalls in the Lee of Lake Erie and Lake Ontario: A Review of the Literature. Volume 43. NOAA - National Weather Service Eastern Region, 20 pg. pp. Hirsch, M. E., A. T. DeGaetano, and S. J. Colucci, 2001: An East Coast winter storm climatology. J. Clim., 14, 882–899, doi:10.1175/1520- 0442(2001)014<0882:AECWSC>2.0.CO;2. Hjelmfelt, M. R., 1990: Numerical Study of the Influence of Environmental Conditions on Lake-Effect Snowstorms over Lake Michigan. Mon. Weather Rev., 118, 138–150, doi:10.1175/1520-0493(1990)118<0138:NSOTIO>2.0.CO;2. http://journals.ametsoc.org/doi/abs/10.1175/1520- 0493%281990%29118%3C0138%3ANSOTIO%3E2.0.CO%3B2. ——, 1992: Orographic effects in simulated lake-effect snowstorms over Lake Michigan. Mon. Weather Rev., 120, 373–377, doi:10.1175/1520- 0493(1992)120<0373:OEISLE>2.0.CO;2. http://www.scopus.com/inward/record.url?eid=2-s2.0- 0026447469&partnerID=40&md5=1d5639e16312c84bbeb9046e91877613. Holroyd, E. W., 1971: Lake-Effect Cloud Bands as Seen From Weather Satellites. J. Atmos. Sci., 28, 1165–1170, doi:10.1175/1520-0469(1971)028<1165:LECBAS>2.0.CO;2. http://journals.ametsoc.org/doi/abs/10.1175/1520- 0469(1971)028%3C1165:LECBAS%3E2.0.CO;2. Hopkins, D., and K. Maclean, 2014: Climate change perceptions and responses in Scotland’s 265
ski industry. Tour. Geogr., 16, 400–414, doi:10.1080/14616688.2013.823457. Hoskins, B. J., and K. I. Hodges, 2002: New perspectives on the Northern Hemisphere winter storm tracks. J. Atmos. Sci., 59, 1041–1061, doi:10.1175/1520- 0469(2002)059<1041:NPOTNH>2.0.CO;2. http://ams.allenpress.com/perlserv/?request=get-toc&issn=1520- 0469&volume=59&issue=6. Hurrell, J. W., 1995: Decadal trends in the north atlantic oscillation: regional temperatures and precipitation. Science, 269, 676–679, doi:10.1126/science.269.5224.676. Hutchinson, T. A., 1995: An Analysis of NCM’s Nested Grid Models of Alberta Clippers. Weather Forecast., 10, 632–641. IPCC, 2013: Climate Change 2013: The Physical Science Basis. Contributi. T.F. Stocker et al., Eds. Cambridge University Press, Cambridge, United Kingdom, 1535 pp. ——, 2018: Global warming of 1.5˚C. An IPCC special report on the impacts of global warming of 1.5˚C above pre-industrial levels and related global greenhouse gas emission pathways, in the context of strengthening the global response to the threat of climate change,. V. Masson-Delmotte et al., Eds. Jacobs, N. A., G. M. Lackmann, and S. Raman, 2005: The combined effects of Gulf Stream- induced baroclinicity and upper-level vorticity on US East Coast extratropical cyclogenesis. Mon. Weather Rev., 133, 2494–2501, doi:10.1175/MWR2969.1. http://journals.ametsoc.org/doi/abs/10.1175/MWR2969.1. Janoski, T. P., A. J. Broccoli, S. B. Kapnick, and N. C. Johnson, 2018: Effects of climate change on wind-driven heavy-snowfall events over eastern North America. J. Clim., 31, 9037– 9054. Jeglum, M. E., W. J. Steenburgh, T. P. Lee, and L. F. Bosart, 2010: Multi-Reanalysis Climatology of Intermountain Cyclones. Mon. Weather Rev., 138, 4035–4053, doi:10.1175/2010MWR3432.1. http://journals.ametsoc.org/doi/abs/10.1175/2010MWR3432.1. Jiusto, J. E., and M. L. Kaplan, 1972: Snowfall From Lake-Effect Storms. Mon. Weather Rev., 100, 62–66, doi:10.1175/1520-0493(1972)100<0062:SFLS>2.3.CO;2. Johnson, G. L., and C. L. Hanson, 1995: Topographic and atmospheric influences on precipitation variability over a mountainous watershed. J. Appl. Meteorol., 34, 68–87. Joly, D., T. Castel, B. Pohl, and Y. Richard, 2018: Influence of spatial information resolution on the relation between elevation and temperature. Int. J. Climatol., 1–12, doi:10.1002/joc.5771. Jones, G. V., and R. E. Davis, 1995: Climatology of Nor’easters and the 30 kPa jet. J. Coast. Res., 11, 1210–1220. Joyce, T. M., 2002: One Hundred Plus Years of Wintertime Climate Variability in the Eastern United States. J. Clim., 15, 1076–1086. Jurewicz, M. L., and M. S. Evans, 2004: A comparison of two banded, heavy snowstorms with very different synoptic settings. Weather Forecast., 19, 1011–1028, doi:10.1175/WAF-823.1. http://journals.ametsoc.org/doi/abs/10.1175/WAF-823.1. Kaczor, G., and P. Bugajski, 2012: Impact of snowmelt inflow on temperature of sewage discharged to treatment plants. Polish J. Environ. Stud., 21, 381–386. Kahya, E., and J. A. Dracup, 1993: U.S. streamflow patterns in relation to the El Niño/Southern Oscillation. Water Resour. Res., 29, 2491–2503, doi:10.1029/93WR00744. 266
Kalkstein, L., and P. Corrigan, 1986: Average Annual U.S. Snowfall. Ann. Assoc. Am. Geogr., 76, 381–395. Kalnay, E., and Coauthors, 1996: The NCEP/NCAR 40-Year Reanalysis Project. Bull. Am. Meteorol. Soc., 77, 437–471, doi:http://dx.doi.org/10.1175/1520- 0477(1996)077<0437:TNYRP>2.0.CO;2. http://dx.doi.org/10.1175/1520- 0477(1996)077%3C0437:TNYRP%3E2.0.CO;2. Kalra, A., and S. Ahmad, 2012: Estimating annual precipitation for the Colorado River Basin using oceanic-atmospheric oscillations. Water Resour. Res., 48, 1–24, doi:10.1029/2011WR010667. Kapnick, S. B., and T. L. Delworth, 2013: Controls of global snow under a changed climate. J. Clim., 26, 5537–5562, doi:10.1175/JCLI-D-12-00528.1. Karl, T. R., P. Y. Groisman, R. W. Knight, and R. R. Heim, 1993: Recent variations of snow cover and snowfall in North America and their relation to precipitation and temperature variations. J. Clim., 6, 1327–1344. Karmosky, C., 2007: Synoptic climatology of snowfall in the Northeastern United States: an analysis of snowfall amounts from diverse synoptic weather types. Delaware, . Kaspi, Y., and T. Schneider, 2013: The Role of Stationary Eddies in Shaping Midlatitude Storm Tracks. J. Atmos. Sci., 70, 2596–2613, doi:10.1175/JAS-D-12-082.1. http://journals.ametsoc.org/doi/abs/10.1175/JAS-D-12-082.1. Katurji, M., and S. Zhong, 2012: The influence of topography and ambient stability on the characteristics of cold-air pools: A numerical investigation. J. Appl. Meteorol. Climatol., 51, 1740–1749, doi:10.1175/JAMC-D-11-0169.1. Keighton, S., D. K. Miller, D. Hotz, P. D. Moore, L. Baker Perry, L. G. Lee, and D. T. Martin, 2016: Northwest Flow Snow Aspects of Hurricane Sandy. Weather Forecast., 31, 173– 195, doi:10.1175/WAF-D-15-0069.1. http://journals.ametsoc.org/doi/abs/10.1175/WAF-D-15-0069.1. Kelly, G. M., L. B. Perry, B. F. Taubman, and P. T. Soulé, 2012: Synoptic classification of 2009- 2010 precipitation events in the southern Appalachian Mountains, USA. Clim. Res., 55, 1–15, doi:10.3354/cr01116. Kelly, R. D., 1986: Mesoscale Frequencies and Seasonal Snowfalls for Different Types of Lake Michigan Snow Storms. J. Clim. Appl. Meteorol., 25, 308–312. Kharin, V. V, and F. W. Zwiers, 2002: Climate Predictions with Multimodel Ensembles. J. Clim., 15, 793–799. Klein, W. H., 1957: Principal Tracks and Mean Frequencies of Cyclones and Anticyclones in the Northern Hemisphere. US Weather Bureau, Washington, D.C., 1–60 pp. ——, 1958: The Frequency of Cyclones and Anticyclones in Relation to the Mean Circulation. J. Meteorol., 15, 98–102, doi:10.1175/1520- 0469(1958)015<0098:TFOCAA>2.0.CO;2. Kluver, D., and D. Leathers, 2015: Regionalization of snowfall frequency and trends over the contiguous United States. Int. J. Climatol., 35, 4348–4358, doi:10.1002/joc.4292. Knowles, N., 2015: Trends in snow cover and related quantities at weather stations in the conterminous United States. J. Clim., 28, 7518–7528, doi:10.1175/JCLI-D-15-0051.1. ——, M. D. Dettinger, and D. R. Cayan, 2006: Trends in snowfall versus rainfall in the western United States. J. Clim., 19, 4545–4559, doi:10.1175/JCLI3850.1. Kocin, P. J., and L. W. Uccellini, 1990: Snowstorms along the northeastern coast of the United States: 1955 to 1985. Meterologi. American Meteorological Society, Boston, MA,. 267
——, and ——, 2004a: A snowfall impact scale derived from Northeast storm snowfall distributions. Bull. Am. Meteorol. Soc., 85, 177–194, doi:10.1175/BAMS-85-2-177. ——, and ——, 2004b: Northeast snowstorms. No. 54. American Meteorological Society,. Krasting, J. P., A. J. Broccoli, K. W. Dixon, and J. R. Lanzante, 2013: Future changes in northern hemisphere snowfall. J. Clim., 26, 7813–7828, doi:10.1175/JCLI-D-12- 00832.1. Kreitzberg, C. W., and H. A. Brown, 1970: Mesoscale Weather Systems within an Occlusion. J. Appl. Meteorol., 9, 417–432, doi:10.1175/1520- 0450(1970)009<0417:MWSWAO>2.0.CO;2. Kretschmer, M., D. Coumou, J. F. Donges, and J. Runge, 2016: Using causal effect networks to analyze different arctic drivers of midlatitude winter circulation. J. Clim., 29, 4069– 4081, doi:10.1175/JCLI-D-15-0654.1. Kreyling, J., and H. A. L. Henry, 2011: Vanishing winters in Germany: Soil frost dynamics and snow cover trends, and ecological implications. Clim. Res., 46, 269–276, doi:10.3354/cr00996. Kristovich, D. A. R., L. Bard, L. Stoecker, and B. Geerts, 2018: Influence of Lake Erie on a Lake Ontario lake-effect snowstorm. J. Appl. Meteorol. Climatol., 57, 2019–2033. ——, and R. A. Steve III, 1995: A satellite stidy of cloud-band frequencies over the Great Lakes. J. Appl. Meteorol., 34, 2083–2090. ——, N. F. Laird, and M. R. Hjelmfelt, 2003: Convective Evolution across Lake Michigan during a Widespread Lake-Effect Snow Event. Mon. Weather Rev., 131, 643–655, doi:10.1175/1520-0493(2003)131<0643:CEALMD>2.0.CO;2. Kunkel, K. E., and J. R. Angel, 1999: Relationship of ENSO to snowfall and related cyclone activity in the contiguous United States. J. Geophys. Res., 104, 19425–19434, doi:10.1029/1999JD900010. ——, N. E. Westcott, and D. A. R. Kristovich, 2000: Climate Change and Lake-Effect Snow. Prep. a Chang. Clim. Potential Consequences Clim. Var. Chang., 25–28. ——, ——, and D. A. R. Kristovich, 2002: Assessment of potential effects of climate change on heavy lake-effect snowstorms near Lake Erie. J. Great Lakes Res., 28, 521–536, doi:10.1016/S0380-1330(02)70603-5. ——, M. A. Palecki, K. G. Hubbard, D. a. Robinson, K. T. Redmond, and D. R. Easterling, 2007: Trend identification in twentieth-century U.S. snowfall: The challenges. J. Atmos. Ocean. Technol., 24, 64–73, doi:10.1175/JTECH2017.1. ——, L. Ensor, M. Palecki, D. Easterling, D. Robinson, K. G. Hubbard, and K. Redmond, 2009a: A new look at lake-effect snowfall trends in the Laurentian Great Lakes using a temporally homogeneous data set. J. Great Lakes Res., 35, 23–29, doi:10.1016/j.jglr.2008.11.003. ——, M. a. Palecki, L. Ensor, D. Easterling, K. G. Hubbard, D. Robinson, and K. Redmond, 2009b: Trends in twentieth-century U.S. extreme snowfall seasons. J. Clim., 22, 6204– 6216, doi:10.1175/2009JCLI2631.1. ——, M. Palecki, L. Ensor, K. G. Hubbard, D. Robinson, K. Redmond, and D. Easterling, 2009c: Trends in twentieth-century U.S. snowfall using a quality-controlled dataset. J. Atmos. Ocean. Technol., 26, 33–44, doi:10.1175/2008JTECHA1138.1. ——, and Coauthors, 2013a: Monitoring and understanding trends in extreme storms: State of knowledge. Bull. Am. Meteorol. Soc., 94, 499–514, doi:10.1175/BAMS-D-11-00262.1. ——, L. E. Stevens, S. E. Stevens, L. Sun, E. Janssen, D. Wuebbles, and J. G. Dobson, 2013b: 268
Regional climate trends and scenarios for the U.S. National climate assessment. Washington, D.C.,. ——, D. A. Robinson, S. Champion, X. Yin, T. Estilow, and R. M. Frankson, 2016: Trends and Extremes in Northern Hemisphere Snow Characteristics. Curr. Clim. Chang. Reports, 2, 65–73, doi:10.1007/s40641-016-0036-8. http://link.springer.com/10.1007/s40641- 016-0036-8. Kusunoki, K., and Coauthors, 2005: Observations of Quasi-Stationary and Shallow Orographic Snow Clouds: Spatial Distributions of Supercooled Liquid Water and Snow Particles. Mon. Weather Rev., 133, 743–751, doi:10.1175/mwr2874.1. Lackmann, G. M., 2001: Analysis of a Surprise Western New York Snowstorm. Weather Forecast., 16, 99–116, doi:10.1175/1520-0434(2001)016<0099:AOASWN>2.0.CO;2. Lahey, J. F., R. A. Bryson, H. A. Corzine, and C. W. Hutchings, 1960: Atlas of 300 mb Wind Char- acteristics for the Northern Hemisphere. 10 pp. Laird, N. F., L. J. Miller, and D. A. R. Kristovich, 2001: Synthetic Dual-Doppler Analysis of a Winter Mesoscale Vortex. Mon. Weather Rev., 129, 312–331, doi:10.1175/1520- 0493(2001)129<0312:sddaoa>2.0.co;2. ——, J. E. Walsh, and D. A. R. Kristovich, 2003: Model Simulations Examining the Relationship of Lake-Effect Morphology to Lake Shape, Wind Direction, and Wind Speed. Mon. Weather Rev., 131, 2102–2111, doi:10.1175/1520- 0493(2003)131<2102:MSETRO>2.0.CO;2. ——, and D. A. R. Kristovich, 2004: Comparison of Observations with Idealized Model Results for a Method to Resolve Winter Lake-Effect Mesoscale Morphology. Mon. Weather Rev., 132, 1093–1103, doi:10.1175/1520- 0493(2004)132<1093:COOWIM>2.0.CO;2. ——, R. Sobash, and N. Hodas, 2009a: The Frequency and Characteristics of Lake-Effect Precipitation Events Associated with the New York State Finger Lakes. J. Appl. Meteorol. Climatol., 48, 873–886, doi:10.1175/2008JAMC2054.1. ——, J. Desrochers, and M. Payer, 2009b: Climatology of lake-effect precipitation events over Lake Champlain. J. Appl. Meteorol. Climatol., 48, 232–250, doi:10.1175/2008JAMC1923.1. http://journals.ametsoc.org/doi/abs/10.1175/2008JAMC1923.1. ——, R. Sobash, and N. Hodas, 2010: Climatological conditions of lake-effect precipitation events associated with the New York state Finger Lakes. J. Appl. Meteorol. Climatol., 49, 1052–1062, doi:10.1175/2010JAMC2312.1. ——, N. D. Metz, L. Gaudet, C. Grasmick, L. Higgins, C. Loeser, and D. A. Zelinsky, 2017: Climatology of cold season lake-effect cloud bands for the North American Great Lakes. Int. J. Climatol., 37, 2111–2121, doi:10.1002/joc.4838. Lang, C. E., J. M. McDonald, L. Gaudet, D. Doeblin, E. A. Jones, and N. F. Larid, 2018: The influence of a lake-to-lake connection from Lake Huron on the lake-effect snowfall in the vicinity of Lake Ontario. J. Appl. Meteorol. Climatol., 57, 1423–1439. Lareau, N. P., and J. D. Horel, 2012: The Climatology of Synoptic-Scale Ascent over Western North America: A Perspective on Storm Tracks. Mon. Weather Rev., 140, 1761–1778, doi:10.1175/MWR-D-11-00203.1. http://journals.ametsoc.org/doi/abs/10.1175/MWR-D-11-00203.1. Lau, N.-C., 1988: Variability of the Observed Midlatitude Storm Tracks in Relation to Low- Frequency Changes in the Circulation Pattern. J. Atmos. Sci., 45, 2718–2743, 269
doi:10.1175/1520-0469(1988)045<2718:VOTOMS>2.0.CO;2. http://journals.ametsoc.org/doi/abs/10.1175/1520- 0469%281988%29045%3C2718%3AVOTOMS%3E2.0.CO%3B2. Lawrimore, J., T. R. Karl, M. Squires, D. A. Robinson, and K. E. Kunkel, 2014: Trends and Variability in Severe Snowstorms East of the Rocky Mountains. J. Hydrometeorol., 15, 1762–1777, doi:10.1175/JHM-D-13-068.1. http://journals.ametsoc.org/doi/abs/10.1175/JHM-D-13-068.1. Lazo, J. K., M. Lawson, P. H. Larsen, and D. M. Waldman, 2011: U.S. economic sensitivity to weather variability. Bull. Am. Meteorol. Soc., 92, 709–720, doi:10.1175/2011BAMS2928.1. Leathers, D. J., and A. W. Ellis, 1996: Synoptic Mechanisms Associated with Snowfall Increases to the Lee of Lakes Erie and Ontario. Int. J. Climatol., 16, 1117–1135, doi:10.1002/(SICI)1097-0088(199610)16:10<1117::AID-JOC80>3.0.CO;2-4. ——, B. Yarnal, and M. A. Palecki, 1991: The Pacific/North American Teleconnection Pattern and United States Climate. Part I: Regional Temperature and Precipitation Associations. J. Clim., 4, 517–528. ——, T. L. Mote, A. J. Grundstein, D. A. Robinson, K. Felter, K. Conrad, and L. Sedywitz, 2002: Associations between continental-scale snow cover anomalies and air mass frequencies across eastern North America. Int. J. Climatol., 22, 1473–1494, doi:10.1002/joc.807. Leeper, R. D., J. Rennie, and M. A. Palecki, 2015: Observational perspectives from U.S. Climate Reference Network (USCRN) and Cooperative Observer Program (COOP) Network: Temperature and precipitation comparison. J. Atmos. Ocean. Technol., 32, 703–721, doi:10.1175/JTECH-D-14-00172.1. Lefevre, R. J., and J. W. Nielson-Gammon, 1995: An objective climatology of mobile troughs in the northern hemisphere. Tellus A, 47, 638–655, doi:10.1034/j.1600- 0870.1995.00110.x. Linkin, M. E., and S. Nigam, 2008: The North Pacific Oscillation-West Pacific teleconnection pattern: Mature-phase structure and winter impacts. J. Clim., 21, 1979–1997, doi:10.1175/2007JCLI2048.1. Liu, A. Q., and G. W. K. Moore, 2004: Lake-Effect Snowstorms over Southern Ontario, Canada, and Their Associated Synoptic-Scale Environment. Mon. Weather Rev., 132, 2595–2609, doi:10.1175/MWR2796.1. Liu, C., K. Ikeda, G. Thompson, R. Rasmussen, and J. Dudhia, 2011: High-resolution simulations of wintertime precipitation in the Colorado headwaters region: sensitivity to physics parameters. Mon. Weather Rev., 139, 3533–3553. Lofgren, B. M., 1997: Simulated effects of idealized Laurentian Great Lakes on regional and large-scale climate. J. Clim., 10, 2847–2858, doi:10.1175/1520- 0442(1997)010<2847:SEOILG>2.0.CO;2. ——, 2004: Global warming effects on Great Lakes water: More precipitation but less water? Bulletin of the American Meteorological Society, 4743–4745. Long, Z., W. Perrie, J. Gyakum, D. Caya, and R. Laprise, 2007: Northern Lake Impacts on Local Seasonal Climate. J. Hydrometeorol., 8, 881–896, doi:10.1175/JHM591.1. Lukens, K. E., E. H. Berbery, and K. I. Hodges, 2018: The imprint of strong-storm tracks on winter weather in North America. J. Clim., 31, 2057–2074, doi:10.1175/JCLI-D-17- 0420.1. 270
Lundquist, J. D., M. Hughes, B. Henn, E. D. Gutmann, B. Livneh, J. Dozier, and P. Neiman, 2015: High-Elevation Precipitation Patterns: Using Snow Measurements to Assess Daily Gridded Datasets across the Sierra Nevada, California*. J. Hydrometeorol., 16, 1773–1792, doi:10.1175/jhm-d-15-0019.1. Mann, G. E., R. B. Wagenmaker, and P. J. Sousounis, 2002: The Influence of Multiple Lake Interactions upon Lake-Effect Storms. Mon. Weather Rev., 130, 1510–1530, doi:10.1175/1520-0493(2002)130<1510:TIOMLI>2.0.CO;2. Mardia, K. V., and P. E. Jupp, 2000: Directional Statistics. Wiley, Chichester,. Mason, S. J., and L. Goddard, 2001: Probabilistic precipitation anomalies associated with ENSO. Bull. Am. Meteorol. Soc., 82, 619–638, doi:10.1175/1520- 0477(2001)082<0619:PPAAWE>2.3.CO;2. Mastin, M. C., K. J. Chase, and R. W. Dudley, 2011: Changes in spring snowpack for selected basins in the United States for different climate-change scenarios. Earth Interact., 15, 1–18, doi:10.1175/2010EI368.1. Mather, J. R., H. Adams, and G. A. Yoshioka, 1964: Coastal Storms of the Eastern United States. J. Appl. Meteorol., 3, 693–706, doi:10.1175/1520- 0450(1964)003<0693:CSOTEU>2.0.CO;2. ——, R. T. Field, and G. A. Yoshioka, 1967: Storm damage hazards along the east coast of the United States. J. Appl. Meteorol., 6, 20–30. papers://2061853f-ee45-498b-ad44- 72732d1032f6/Paper/p241. McCabe, G. J., and M. D. Dettinger, 2002: Primary Modes and Predictability of Year-to-Year Snowpack Variations in the Western United States from Teleconnections with Pacific Ocean Climate. J. Hydrometeorol., 3, 13–25, doi:10.1175/1525- 7541(2002)003<0013:PMAPOY>2.0.CO;2. http://journals.ametsoc.org/doi/full/10.1175/1525- 7541(2002)003%3C0013:PMAPOY%3E2.0.CO;2%5Cnhttp://journals.ametsoc.org/do i/pdf/10.1175/1525-7541(2002)003%3C0013:PMAPOY%3E2.0.CO;2. ——, M. P. Clark, and M. C. Serreze, 2001: Trends in Northern Hemisphere surface cyclone frequency and intensity. J. Clim., 14, 2763–2768, doi:10.1175/1520- 0442(2001)014<2763:TINHSC>2.0.CO;2. ——, ——, and L. E. Hay, 2007: Rain-on-snow events in the western United States. Bull. Am. Meteorol. Soc., 88, 319–328, doi:10.1175/BAMS-88-3-319. Mercer, A. E., and M. B. Richman, 2007: Statistical Differences of Quasigeostrophic Variables, Stability, and Moisture Profiles in North American Storm Tracks. Mon. Weather Rev., 135, 2312–2338, doi:10.1175/MWR3395.1. http://journals.ametsoc.org/doi/abs/10.1175/MWR3395.1. Miller, J. E., 1946: Cyclogenesis in the Atlantic coastal region of the United States. J. Atmos. Sci., 3, 31–44. http://journals.allenpress.com/jrnlserv/?request=get- abstract&issn=1520-0469&volume=3&page=31. Millerd, F., 2011: The potential impact of climate change on Great Lakes international shipping. Clim. Change, 104, 629–652, doi:10.1007/s10584-010-9872-z. Minder, J. R., T. W. Letcher, L. S. Campbell, P. G. Veals, and W. J. Steenburgh, 2015: The Evolution of Lake-Effect Convection during Landfall and Orographic Uplift as Observed by Profiling Radars. Mon. Weather Rev., 143, 4422–4442, doi:10.1175/MWR-D-15- 0117.1. http://journals.ametsoc.org/doi/10.1175/MWR-D-15-0117.1. Miner, T. J., and J. M. Fritsch, 1997: Lake-Effect Rain Events. Mon. Weather Rev., 125, 3231– 271
3248, doi:10.1175/1520-0493(1997)125<3231:LERE>2.0.CO;2. Mishra, V., K. a. Cherkauer, and S. Shukla, 2010: Assessment of Drought due to Historic Climate Variability and Projected Future Climate Change in the Midwestern United States. J. Hydrometeorol., 11, 46–68, doi:10.1175/2009JHM1156.1. http://journals.ametsoc.org/doi/abs/10.1175/2009JHM1156.1. Mock, C. J., 1996: Climatic controls and spatial variations of precipitation in the western United States. J. Clim., 9, 1111–1124, doi:10.1175/1520- 0442(1996)009<1111:CCASVO>2.0.CO;2. Monmonier, M. S., 2012: Lake Effect: Tales of Large Lakes, Arctic Winds, and Recurrent Snow. Syracuse University Press, Syracuse, NY,. Moore, G. W. K., I. A. Renfrew, and R. S. Pickart, 2013: Multidecadal mobility of the north atlantic oscillation. J. Clim., 26, 2453–2466, doi:10.1175/JCLI-D-12-00023.1. Morrison, I., and S. Businger, 2001: Synoptic structure and evolution of a Kona low. Weather Forecast., 16, 81–98, doi:10.1175/1520- 0434(2001)016<0081:SSAEOA>2.0.CO;2. http://journals.ametsoc.org/doi/pdf/10.1175/1520- 0434(2001)016%3C0081:SSAEOA%3E2.0.CO;2. Mote, P. W., 2006: Climate-Driven Variability and Trends in Mountain Snowpack in Western North America. J. Clim., 19, 6209–6220, doi:10.1175/jcli3971.1. ——, A. F. Hamlet, M. P. Clark, and D. P. Lettermaier, 2005: Declining mountain snowpack in Western North America. Bull. Am. Meteorol. Soc., 86, 39–49, doi:10.1130/2008.2435(05). Mote, T. L., 2008: On the role of snow cover in depressing air temperature. J. Appl. Meteorol. Climatol., 47, 2008–2022, doi:10.1175/2007JAMC1823.1. ——, D. W. Gamble, S. J. Underwood, and M. L. Bentley, 1997: Synoptic-Scale Features Common to Heavy Snowstorms in the Southeast United States. Weather Forecast., 12, 5–23, doi:10.1175/1520-0434(1997)012<0005:SSFCTH>2.0.CO;2. Mullens, E. D., L. M. Leslie, and P. J. Lamb, 2016: Synoptic Pattern Analysis and Climatology of Ice and Snowstorms in the Southern Great Plains, 1993–2011. Weather Forecast., 31, 1109–1136, doi:10.1175/WAF-D-15-0172.1. http://journals.ametsoc.org/doi/10.1175/WAF-D-15-0172.1. Najafi, M. R., F. W. Zwiers, and N. P. Gillett, 2016: Attribution of the spring snow cover extent decline in the Northern Hemisphere, Eurasia and North America to anthropogenic influence. Clim. Change, 136, 571–586, doi:10.1007/s10584-016-1632- 2. Nakagawa, S., and H. Schielzeth, 2013: A general and simple method for obtaining R2 from generalized linear mixed-effects models. Methods Ecol. Evol., 4, 133–142, doi:10.1111/j.2041-210x.2012.00261.x. Namias, J., 1963: Interactions of circulation and weather between hemispheres. Mon. Weather Rev., 482–486. Neiman, P. J., and Coauthors, 1998: An Observational Study of Fronts and Frontal Mergers over the Continental United States. Mon. Weather Rev., 126, 2521–2554, doi:10.1155/2014/816729. http://journals.ametsoc.org/doi/abs/10.1175/1520- 0493%281998%29126%3C2521%3AAOSOFA%3E2.0.CO%3B2. Newman, M., and Coauthors, 2016: The Pacific Decadal Oscillation, Revisited. J. Oceanogr., 29, 35–44. 272
Nielsen, J. W., and R. M. Dole, 1992: A Survey of Extratropical Cyclone Characteristics during GALE. Mon. Weather Rev., 120, 1156–1168, doi:10.1175/1520- 0493(1992)120<1156:ASOECC>2.0.CO;2. http://journals.ametsoc.org/doi/abs/10.1175/1520- 0493(1992)120%3C1156:ASOECC%3E2.0.CO;2. Niziol, T. A., 1987: Operational Forecasting of Lake Effect Snowfall in Western and Central New York. Weather Forecast., 2, 310–321, doi:10.1175/1520- 0434(1987)002<0310:OFOLES>2.0.CO;2. ——, W. R. Snyder, and J. S. Waldstreicher, 1995: Winter Weather Forecasting throughout the Eastern United States. Part IV: Lake Effect Snow. Weather Forecast., 10, 61–77, doi:10.1175/1520-0434(1995)010<0061:WWFTTE>2.0.CO;2. Norton, D. C., and S. J. Bolsenga, 1993: Spatiotemporal trends in lake effect and continental snowfall in the Laurentian Great Lakes, 1951-1980. J. Clim., 6, 1943–1956, doi:10.1175/1520-0442(1993)006<1943:STILEA>2.0.CO;2. Notaro, M., and Z. Liu, 2008: Statistical and dynamical assessment of vegetation feedbacks on climate over the boreal forest. Clim. Dyn., 31, 691–712. ——, W.-C. Wang, and W. Gong, 2006: Model and Observational Analysis of the Northeast U.S. Regional Climate and Its Relationship to the PNA and NAO Patterns during Early Winter. Mon. Weather Rev., 134, 3479–3505, doi:10.1175/MWR3234.1. ——, K. Holman, A. Zarrin, E. Fluck, S. Vavrus, and V. Bennington, 2013a: Influence of the Laurentian Great Lakes on regional climate. J. Clim., 26, 789–804, doi:10.1175/JCLI-D- 12-00140.1. ——, A. Zarrin, S. Vavrus, and V. Bennington, 2013b: Simulation of Heavy Lake-Effect Snowstorms across the Great Lakes Basin by RegCM4: Synoptic Climatology and Variability. Mon. Weather Rev., 141, 1990–2014, doi:10.1175/MWR-D-11-00369.1. http://journals.ametsoc.org/doi/abs/10.1175/MWR-D-11-00369.1. ——, D. Lorenz, C. Hoving, and M. Schummer, 2014: Twenty-first-century projections of snowfall and winter severity across central-eastern North America. J. Clim., 27, 6526– 6550, doi:10.1175/JCLI-D-13-00520.1. http://journals.ametsoc.org/doi/abs/10.1175/JCLI-D-13-00520.1. ——, V. Bennington, S. Vavrus, and A. Zarrin, 2015: Dynamically Downscaled Projections of Lake-Effect Snow in the Great Lakes Basin. J. Clim., 28, 1661–1684. NRC, 2012: Climate Change: Evidence, Impacts, and Choices. National Academy of Sciences, 1–36 pp. NRCC, 2019: January - Wet but not necessarily snowy. Ithaca, NY, http://www.nrcc.cornell.edu/services/blog/2019/02/01/index.html. NWS, 2004: National Weather Service Glossary. https://w1.weather.gov/glossary/. ——, 2012: Snow Measurement Guidelines for National Weather Service Snow Spotters. Peachtree City, Georgia, 1–5 pp. ——, 2014a: Winter Weather Basics: Cold Air Outbreaks. ——, 2014b: Lake Effect Summary: November 17-19, 2014. O’Handley, C., and L. F. Bosart, 1996: The Impact of the Appalachian Mountains on Cyclonic Weather Systems. Part I: A Climatology. Mon. Weather Rev., 124, 1353–1373, doi:10.1175/1520-0493(1996)124<1353:TIOTAM>2.0.CO;2. O’Hara, B. F., M. L. Kaplan, and S. J. Underwood, 2009: Synoptic Climatological Analyses of Extreme Snowfalls in the Sierra Nevada. Weather Forecast., 24, 1610–1624, 273
doi:10.1175/2009WAF2222249.1. http://journals.ametsoc.org/doi/abs/10.1175/2009WAF2222249.1. Olson, K. R., and L. W. Morton, 2012: The impacts of 2011 induced levee breaches on agricultural lands of Mississippi River Valley. J. Soil Water Conserv., 67, 5A-10A, doi:10.2489/jswc.67.1.5A. http://www.jswconline.org/cgi/doi/10.2489/jswc.67.1.5A. Onton, D. J., and W. J. Steenburgh, 2001: Diagnostic and Sensitivity Studies of the 7 December 1998 Great Salt Lake–Effect Snowstorm. Mon. Weather Rev., 129, 1318– 1338. Osborn, T. J., 2011: Winter 2009/2010 temperatures and a record-breking North Atlantic Oscillation Index. Weather, 66, 19–21. Overland, J. E., J. E. Walsh, and M. Wang, 2007: Why are Snow and Ice Changing? Global Outlook for Snow & Ice, J. Eamer, Ed., Division of Early Warning and Assessment United Nations Environment Programme, Nairobi, Kenya, 28–38. Patten, J. M., S. R. Smith, and J. J. O’Brien, 2003: Impacts of ENSO on Snowfall Frequencies in the United States. Weather Forecast., 18, 965–980, doi:10.1175/1520- 0434(2003)018<0965:IOEOSF>2.0.CO;2. %5C%5CIBSERVER%5CID_enl%5CWeather_and_forecasting%5C2003%5C018-05- 0965.pdf. Peace, R. L., and R. B. Sykes, 1966: Mesoscale Study of a Lake Effect Snow Storm. Mon. Weather Rev., 94, 495–507, doi:10.1175/1520- 0493(1966)094<0495:MSOALE>2.3.CO;2. Pease, S. R., W. A. Lyons, C. S. Keen, and M. Hjelmfelt, 1988: Mesoscale spiral vortex embedded within a Lake Michigan snow squall band - High resolution satellite observations and numerical model simulations. Mon. Weather Rev., 116, 1374–1380, doi:10.1175/1520-0493(1988)116<1374:MSVEWA>2.0.CO;2. Pederson, G. T., S. T. Gray, D. B. Fagre, and L. J. Graumlich, 2006: Long-duration drought variability and impacts on ecosystem services: A case study from Glacier National Park, Montana. Earth Interact., 10, 1–28, doi:10.1175/EI153.1. Pelletier, J. D., and C. Rasmussen, 2009: Geomorphically based predictive mapping of soil thickness in upland watersheds. Water Resour. Res., 45, 1–15, doi:10.1029/2008WR007319. Peltier, W. R., M. D’Orgeville, A. R. Erler, and F. Xie, 2018: Uncertainty in future summer precipitation in the Laurentian Great Lakes basin: dynamical downscaling and the influence of continental-scale processes on regional climate change. J. Clim., 31, 2651– 2673. Perry, L. B., and C. E. Konrad, 2006: Relationships between NW flow snowfall and topography in the Southern Appalachians, USA. Clim. Res., 32, 35–47, doi:10.3354/cr032035. ——, C. E. Konrad, and T. W. Schmidlin, 2007: Antecedent Upstream Air Trajectories Associated with Northwest Flow Snowfall in the Southern Appalachians. Weather Forecast., 22, 334–352, doi:10.1175/WAF978.1. http://journals.ametsoc.org/doi/abs/10.1175/WAF978.1. Persinger, M. A., S. E. Ballance, and M. Moland, 1993: Snow fall and heart attacks. J. Psychol., 127, 243–252. Petterssen, S., 1941: Introduction to Meteorology. McGraw-Hill, 236 pp. ——, 1956: Weather Analysis and Forecasting. Vol. I. McGraw-Hill, 428 pp. 274
Piechota, T. C., and J. A. Dracup, 1996: Drought and regional hydrologic variation in the United States: Associations with the El Nino-Southern oscillation. Water Resour. Res., 32, 1359–1373, doi:10.1029/96WR00353. Posada, D., and T. R. Buckley, 2004: Model selection and model averaging in phylogenetics: Advantages of akaike information criterion and bayesian approaches over likelihood ratio tests. Syst. Biol., 53, 793–808, doi:10.1080/10635150490522304. Pütz, M., D. Gallati, S. Kytzia, H. Elsasser, C. Lardelli, M. Teich, F. Waltert, and C. Rixen, 2011: Winter Tourism, Climate Change, and Snowmaking in the Swiss Alps: Tourists’ Attitudes and Regional Economic Impacts. Mt. Res. Dev., 31, 357–362, doi:10.1659/MRD-JOURNAL-D-11-00039.1. http://www.bioone.org/doi/10.1659/MRD-JOURNAL-D-11-00039.1. Qu, X., and A. Hall, 2007: What controls the strength of snow-albedo feedback? J. Clim., 20, 3971–3981, doi:10.1175/JCLI4186.1. Rasmussen, R., and Coauthors, 2012: How well are we measuring snow? Bull. Am. Meteorol. Soc., 93, 811–829, doi:10.1175/BAMS-D-11-00052.1. Rasmusson, E. M., and T. H. Carpenter, 1982: Variations in Tropical Sea Surface Temperature and Surface Wind Fields Associated with the Southern Oscillation/El Niño. Mon. Weather Rev., 110, 354–384, doi:10.1175/1520- 0493(1982)110<0354:VITSST>2.0.CO;2. http://journals.ametsoc.org/doi/abs/10.1175/1520- 0493%281982%29110%3C0354%3AVITSST%3E2.0.CO%3B2. Rauber, R. M., J. E. Walsh, and D. J. Charlevoix, 2002: Severe and Hazardous Weather. Kimball/Hunt Publishers, 616 pp. Reeves, J., J. Chen, X. L. Wang, R. Lund, and Q. Q. Lu, 2007: A review and comparison of changepoint detection techniques for climate data. J. Appl. Meteorol. Climatol., 46, 900– 915, doi:10.1175/JAM2493.1. Reinking, R. F., and Coauthors, 1993: The Lake-Ontario Winter Storms (LOWS) Project. Bull. Am. Meteorol. Soc., 74, 1828–1849, doi:10.1175/1520-0477-74-10-1828. Reitan, C. H., 1974: Frequencies of Cyclones and Cyclogenesis for North America, 1951– 1970. Mon. Weather Rev., 102, 861–868, doi:10.1175/1520- 0493(1974)102<0861:FOCACF>2.0.CO;2. Rempel, R. S., 2011: Effects of climate change on moose populations: Exploring the response horizon through biometric and systems models. Ecol. Modell., 222, 3355– 3365, doi:10.1016/j.ecolmodel.2011.07.012. Risbey, J. S., T. J. O’Kane, D. P. Monselesan, C. Franzke, and I. Horenko, 2015: Metastability of Northern Hemisphere Teleconnection Modes. J. Atmos. Sci., 72, 35–54, doi:10.1175/JAS-D-14-0020.1. http://journals.ametsoc.org/doi/10.1175/JAS-D-14- 0020.1. Ritchie, E. A., and R. L. Elsberry, 2007: Simulations of the Extratropical Transition of Tropical Cyclones: Phasing between the Upper-Level Trough and Tropical Cyclones. Mon. Weather Rev., 135, 862–876, doi:10.1175/MWR3303.1. http://journals.ametsoc.org/doi/abs/10.1175/MWR3303.1. Robinson, P. J., and A. Henderson-Sellers, 1999: Contemporary Climatology. 2nd ed. Routledge, University of Michigan,. Robock, A., and A. Robock, 1980: The Seasonal Cycle of Snow Cover, Sea Ice and Surface Albedo. Mon. Weather Rev., 108, 267–285, doi:10.1175/1520- 275
0493(1980)108<0267:TSCOSC>2.0.CO;2. http://journals.ametsoc.org/doi/abs/10.1175/1520- 0493%281980%29108%3C0267%3ATSCOSC%3E2.0.CO%3B2. Rochette, S. M., P. S. Market, C. M. Gravelle, and T. A. Niziol, 2017: A Case Study of Anomalous Snowfall with an Alberta Clipper. Adv. Meteorol., 1–14. Rodriguez, Y., D. A. R. Kristovich, and M. R. Hjelmfelt, 2007: Lake-to-Lake Cloud Bands: Frequencies and Locations. Mon. Weather Rev., 135, 4202–4213, doi:10.1175/2007MWR1960.1. http://journals.ametsoc.org/doi/abs/10.1175/2007MWR1960.1. Rohli, R. V, and A. J. Vega, 2011: Climatology Related books. 2nd ed. Jones & Bartlett Learning,. Rohr, J. R., A. P. Dobson, P. T. J. Johnson, A. M. Kilpatrick, S. H. Paull, T. R. Raffel, D. Ruiz- Moreno, and M. B. Thomas, 2012: Frontiers in climate change-disease research. Trends Ecol. Evol., 2011, 270–277, doi:10.1016/j.tree.2011.03.002.Frontiers. Rohrer, M., S. Bronnimann, O. Martius, C. C. Raible, M. Wild, and G. P. Compo, 2018: Representation of extratropical cyclones, blocking anticyclones, and alpine circulation types in multiple reanalysis and model simulations. J. Clim., 31, 3009–3031. Roller, C. D., J.-H. Qian, L. Agel, M. Barlow, and V. Moron, 2016: Winter weather regimes in the Northeast United States. J. Clim., 29, 2963–2980, doi:10.1175/JCLI-D-15-0274.1. Rolph, G. D., 2003: Real-time Environmental Applocations and Display System (READY). ——, A. Stein, and B. Stunder, 2017: Real-time Environmental Applications and Display sYstem: READY. Environ. Model. Softw., 95, 210–228, doi:10.1016/j.envsoft.2017.06.025. http://dx.doi.org/10.1016/j.envsoft.2017.06.025. Rooney, J. F., 1967: The urban snow hazard in the United States: an appraisal of disruption. Geogr. Rev., 57, 538–559. Ropelewski, C. F., and M. S. Halpert, 1986: North American precipitation and temperature patterns associated with the El Niño/Southern Oscillation (ENSO). Mon. Weather Rev., 114, 2352–2362, doi:10.1175/1520-0493(1986)114<2352:napatp>2.0.co;2. http://dx.doi.org/10.1175/1520-0493(1986)114%3C2352:NAPATP%3E2.0.CO;2. Ross, B., and J. E. Walsh, 1986: Synoptic-Scale Influences of Snow Cover and Sea Ice. Mon. Weather Rev., 114, 1795–1810, doi:10.1175/1520- 0493(1986)114<1795:SSIOSC>2.0.CO;2. Rutty, M., D. Scott, P. Johnson, E. Jover, M. Pons, and R. Steiger, 2015: Behavioural adaptation of skiers to climatic variability and change in Ontario, Canada. J. Outdoor Recreat. Tour., 11, 13–21, doi:10.1016/j.jort.2015.07.002. http://dx.doi.org/10.1016/j.jort.2015.07.002. Rydzik, M., and A. R. Desai, 2014: Relationship between snow extent and midlatitude disturbance centers. J. Clim., 27, 2971–2982, doi:10.1175/JCLI-D-12-00841.1. http://journals.ametsoc.org/doi/abs/10.1175/JCLI-D-12-00841.1. Sanders, F., and J. R. Gyakum, 1980: Synoptic-Dynamic Climatology of the “Bomb.” Mon. Weather Rev., 108, 1589–1606, doi:10.1175/1520- 0493(1980)108<1589:SDCOT>2.0.CO;2. Saslo, S., and S. J. Greybush, 2017: Prediction of Lake-Effect Snow Using Convection- Allowing Ensemble Forecasts and Regional Data Assimilation. Weather Forecast., 32, 1727–1744, doi:10.1175/WAF-D-16-0206.1. http://journals.ametsoc.org/doi/10.1175/WAF-D-16-0206.1. 276
Saylor, K., and E. B. Fawcett, 1965: A study of the distribution of weather accompanying Colorado cyclogenesis. Mon. Weather Rev., 93, 359–367. Scaife, A. A., T. Woollings, J. Knight, G. Martin, and T. Hinton, 2010: Atmospheric blocking and mean biases in climate models. J. Clim., 23, 6143–6152, doi:10.1175/2010JCLI3728.1. Schmidlin, T.W., and B. E. Dethier, 1985: Freezing Degree-Days in New York State. Cold Reg. Sci. Technol., 11, 37–43. ——., 1993: Impacts of severe winter weather during December 1989 in the Lake Erie snowbelt. J. Clim., 6, 759–767, doi:10.1175/1520- 0442(1993)006<0759:IOSWWD>2.0.CO;2. ——, and J. Kosarik, 1999: A record Ohio snowfall during 9-14 November 1996. Bull. Am. Meteorol. Soc., 80, 1107–1116, doi:10.1175/1520- 0477(1999)080<1107:AROSDN>2.0.CO;2. Schultz, D. M., D. S. Arndt, D. J. Stensrud, and J. W. Hanna, 2004: Snowbands during the cold- air outbreak of 23 January 2003. Mon. Weather …, 132, 827–842, doi:10.1175/1520- 0493(2004)132<0827:SDTCOO>2.0.CO;2. http://search.ebscohost.com/login.aspx?direct=true&profile=ehost&scope=site&auth type=crawler&jrnl=00270644&AN=12480751&h=sfUz447X4l9fcC61yFbl84deHDwOb IjqZaBNMn6a4JsmL3%2FRGTvNexQ5LhUm1tUdVkxVjWXXB6%2FWQrmAPLJ6qg%3 D%3D&crl=c. Schumacher, P. N., D. J. Knight, and L. F. Bosart, 1996: Frontal interaction with the Appalachian Mountains. Part I: A climatology. Mon. Weather Rev., 124, 2453–2468. http://journals.ametsoc.org/doi/abs/10.1175/1520- 0493(1996)124%3C2453:FIWTAM%3E2.0.CO;2. Scott, C. P. J., and P. J. Sousounis, 2001: The Utility of Additional Soundings for Forecasting Lake-Effect Snow in the Great Lakes Region. Weather Forecast., 16, 448–462, doi:10.1175/1520-0434(2001)016<0448:TUOASF>2.0.CO;2. Scott, D., and D. Kaiser, 2004: Variability and Trends in United States Snowfall Over the Last Half Century. Proceedings of the 15th Symposium on Global Climate Change, 85th Annual AMS Meeting, Ed., Seattle, WA. ——, J. Dawson, and B. Jones, 2008: Climate change vulnerability of the US Northeast winter recreation- tourism sector. Mitig. Adapt. Strateg. Glob. Chang., 13, 577–596, doi:10.1007/s11027-007-9136-z. Scott, R. W., and F. A. Huff, 1996: Impacts of the Great Lakes on regional climate conditions. J. Great Lakes Res., 22, 845–863, doi:10.1016/S0380-1330(96)71006-7. http://dx.doi.org/10.1016/S0380-1330(96)71006-7. Screen, J. A., and I. Simmonds, 2014: Amplified mid-latitude planetary waves favour particular regional weather extremes. Nat. Clim. Chang., 4, 704–709. Seager, R., Y. Kushnir, J. Nakamura, M. Ting, and N. Naik, 2010a: Northern Hemisphere winter snow anomalies: ENSO, NAO, and the winter of 2009/10. Geophys. Res. Lett., 37. ——, N. Naik, and G. A. Vecchi, 2010b: Thermodynamic and dynamic mechanisms for large- scale changes in the hydrological cycle in response to global warming. J. Clim., 23, 4651–4668, doi:10.1175/2010JCLI3655.1. Segal, M., and R. Kubesh, 1996: Inferring snow-breeze characteristics from frozen-lake breezes. J. Appl. Meteorol., 35, 1033–1039, doi:http://dx.doi.org/10.1175/1520- 0450(1996)035<1033:ISBCFF>2.0.CO;2. 277
Seierstad, I. A., D. B. Stephenson, and N. G. Kvamstø, 2007: How useful are teleconnection patterns for explaining variability in extratropical storminess? Tellus, Ser. A Dyn. Meteorol. Oceanogr., 59, 170–181, doi:10.1111/j.1600-0870.2007.00226.x. Sellinger, C. E., C. A. Stow, E. C. Lamon, and S. S. Qian, 2008: Recent water level declines in the Lake Michigan-Huron system. Environ. Sci. Technol., 42, 367–373, doi:10.1021/es070664+. Serebreny, S. M., E. J. Wiegman, and R. G. Hadfield, 1962: Some characterisitc features of the jet stream complex during selected synoptic conditions. J. Appl. Meteorol., 1, 137–153. Serreze, M. C., M. P. Clark, D. L. Mcginnis, and D. a. Robinson, 1998: Characteristics of snowfall over the eastern half of the United States and relationships with principal modes of low-frequency atmospheric variability. J. Clim., 11, 234–250, doi:10.1175/1520-0442(1998)011<0234:COSOTE>2.0.CO;2. Sharples, J. J., M. F. Hutchinson, and D. R. Jellett, 2005: On the Horizontal Scale of Elevation Dependence of Australian Monthly Precipitation. J. Appl. Meteorol., 44, 1850–1865, doi:10.1175/JAM2289.1. Sharratt, B. S., D. G. Baker, D. B. Wall, R. H. Skaggs, and D. L. Ruschy, 1992: Snow depth required for near steady-state soil temperatures. Agric. For. Meteorol., 57, 243–251, doi:10.1016/0168-1923(92)90121-J. Shimura, T., N. Mori, and H. Mase, 2013: Ocean waves and teleconnection patterns in the Northern Hemisphere. J. Clim., 26, 8654–8670, doi:10.1175/JCLI-D-12-00397.1. Siegert, S., D. B. Stephenson, P. G. Sansom, A. A. Scaife, R. Eade, and A. Arribas, 2016: A Bayesian framework for verification and recalibration of ensemble forecasts: How uncertain is NAO predictability? J. Clim., 29, 995–1012, doi:10.1175/JCLI-D-15-0196.1. Silberberg, S. R., 1990: The role of mesoscale features in a wintertime Great Lakes cyclone. Weather Forecast., 5, 89–114. Sillmann, J., C. M. Mischa, M. Kallache, and R. W. Katz, 2011: Extreme cold winter temperatures in Europe under the influence of North Atlantic atmospheric blocking. J. Clim., 24, 5899–5913, doi:10.1175/2011JCLI4075.1. Sittel, M., 1994: Differences in the means of ENSO extremes for maximum temperature and precipitation in the United States. The Florida State University, Tallahasse, FL, 50 pp. Small, E. E., L. C. Sloan, S. Hostetler, and F. Giorgi, 1999: Simulating the water balance of the Aral Sea with a coupled regional climate-lake model. J. Geophys. Res., 104, 6583–6602, doi:10.1029/98JD02348. Smith, J. B., 1991: The Potential Impacts of Climate Change on the Norfolk Broads. Bull. Am. Meteorol. Soc., 72, 21–28. Smith, S. R., and J. J. O’Brien, 2001: Regional snowfall distributions associated with ENSO: Implications for seasonal forecasting. Bull. Am. Meteorol. Soc., 82, 1179–1191, doi:10.1175/1520-0477(2001)082<1179:RSDAWE>2.3.CO;2. Sobash, R., H. Carr, and N. F. Laird, 2000: An investigation of New York State Finger Lakes snow band events. 11th Conference on Mesoscale Processes, P3M.3. Sobolowski, S., and A. Frei, 2007: Lagged relationships between North American snow mass and atmospheric teleconnection indices. Int. J. Climatol., 27, 221–231. Sousounis, P. J., 2001: Lake-effect storms. Encyclopedia of Atmospheric Sciences, J. Holton, J. Pyle, and J. Curry, Eds., Academic Press, 1104–1115 http://www.sciencedirect.com/science/article/B77C7-4B0C797- 4R/2/3753a4e619ece81924fc03fcd60ae986. 278
——, and J. M. Fritsch, 1994: Lake-aggregaate mesoscale disturbances: Part II: A case study of the effects on regional and synoptic-scale weather systems. Bull. Am. Meteorol. Soc., 75, 1793–1811. ——, and G. E. Mann, 2000: Lake-Aggregate Mesoscale Disturbances. Part V: Impacts on Lake-Effect Precipitation. Mon. Weather Rev., 128, 728–745, doi:10.1175/1520- 0493(2000)128<0728:LAMDPV>2.0.CO;2. %5C%5CIBSERVER%5CID_enl%5CMonthly_weather_review%5C2000%5C128-03- 0728.pdf. Spreen, W., 1947: A determination of the effect of topography upon precipitation. Trans. Am. Geophys. Union, 28, 285–290. Stachowicz, J. J., J. R. Terwin, R. B. Whitlatch, and R. W. Osman, 2002: Linking climate change and biological invasions: Ocean warming facilitates nonindigenous species invasions. Proc. Natl. Acad. Sci. U. S. A., 99, 15497–15500, doi:10.1073/pnas.242437499. http://www.pubmedcentral.nih.gov/articlerender.fcgi?artid=137745&tool=pmcentre z&rendertype=abstract. Steenburgh, W. J., S. F. Halvorson, and D. J. Onton, 2000: Climatology of Lake-Effect Snowstorms of the Great Salt Lake. Mon. Weather Rev., 128, 709–727, doi:10.1175/1520-0493(2000)128<0709:COLESO>2.0.CO;2. Steiger, R., and J. Stötter, 2013: Climate Change Impact Assessment of Ski Tourism in Tyrol. Tour. Geogr., 15, 577–600, doi:10.1080/14616688.2012.762539. Stein, A. F., R. R. Draxler, G. D. Rolph, B. J. B. Stunder, M. D. Cohen, and F. Ngan, 2015a: NOAA’s HYPSLIT atmospheric transport and dispersion modeling system. Bull. Am. Meteorol. Soc., 96, 2059–2077, doi:10.1175/BAMS-D-14-00110.1. ——, ——, ——, ——, ——, and ——, 2015b: NOAA’s HYSPLIT atmospheric transport and dispersion modeling system. Bull. Am. Meteorol. Soc., 96, 2059–2077, doi:10.1175/BAMS-D-14-00110.1. Strommen, N. D., and J. R. Harman, 1978: Seasonally Changing Patterns of Lake-Effect Snowfall in Western Lower Michigan. Mon. Weather Rev., 106, 503–509. Strong, A. E., 1972: The Influence of a Great Lake Anticyclone on the Atmospheric Circulation. J. Appl. Meteorol., 11, 598–612. Strong, C., and R. E. Davis, 2008: Variability in the position and strength of winter jet stream cores related to northern hemisphere teleconnections. J. Clim., 21, 584–592, doi:10.1175/2007JCLI1723.1. Sui, C.-H., and K.-M. Lau, 1992: Multiscale Phenomena in the Tropical Atmosphere over the Western Pacific. Mon. Weather Rev., 120, 407–430, doi:10.1175/1520- 0493(1992)120<0407:MPITTA>2.0.CO;2. http://journals.ametsoc.org/doi/abs/10.1175/1520- 0493(1992)120%3C0407:MPITTA%3E2.0.CO;2%5Cnhttp://journals.ametsoc.org/doi /abs/10.1175/1520-0493(1992)120%3C0407:MPITTA%3E2.0.CO;2. Suriano, Z. J., and D. J. Leathers, 2016: Twenty-first century snowfall projections within the eastern Great Lakes region: Detecting the presence of a lake-induced snowfall signal in GCMs. Int. J. Climatol., 36, 2200–2209, doi:10.1002/joc.4488. ——, and ——, 2017a: Synoptic climatology of lake-effect snowfall conditions in the eastern Great Lakes region. Int. J. Climatol., 37. ——, and ——, 2017b: Synoptically classified lake-effect snowfall trends to the lee of Lakes 279
Erie and Ontario. Clim. Res., 74, 1–13. Symonds, M. R. E., and A. Moussalli, 2011: A brief guide to model selection, multimodel inference and model averaging in behavioural ecology using Akaike’s information criterion. Behav. Ecol. Sociobiol., 65, 13–21. Tanaka, S., K. Nishii, and H. Nakamura, 2016: Vertical structure and energetics of the Western Pacific teleconnection pattern. J. Clim., 29, 6597–6616, doi:10.1175/JCLI-D- 15-0549.1. Tardy, A., 2000: Lake-effect and lake-enhanced snow in the Champlain Valley of Vermont. East. Reg. Tech. Attach., 1–27. http://www.erh.noaa.gov/er/btv/research/Tardy- ta2000-05.pdf. Thomas, B. C., and J. E. Martin, 2007: A Synoptic Climatology and Composite Analysis of the Alberta Clipper. Weather Forecast., 22, 315–333, doi:10.1175/WAF982.1. Trumpickas, J., B. J. Shuter, and C. K. Minns, 2009: Forecasting impacts of climate change on Great Lakes surface water temperatures. J. Great Lakes Res., 35, 454–463, doi:10.1016/j.jglr.2009.04.005. Tsuboki, K., Y. Fujiyoshi, G. Wakahama, K. Tsuboki, Y. Fujiyoshi, and G. Wakahama, 1989: Structure of a Land Breeze and Snowfall Enhancement at the Leading Edge. J. Meteorol. Soc. Japan, 67, 757–770. Turner, J. K., and J. R. Gyakum, 2011: The development of arctic air masses in northwest Canada and their behavior in a warming climate. J. Clim., 24, 4618–4633, doi:10.1175/2011JCLI3855.1. Uccellini, L. W., and P. J. Kocin, 1987: The Interaction of Jet Streak Circulations during Heavy Snow Events along the East Coast of the United States. Weather Forecast., 2, 289–308, doi:10.1175/1520-0434(1987)002<0289:TIOJSC>2.0.CO;2. ——, ——, R. S. Schneider, P. M. Stokols, and R. a. Dorr, 1995: Forecasting the 12–14 March 1993 Superstorm. Bull. Am. Meteorol. Soc., 76, 183–199, doi:10.1175/1520- 0477(1995)076<0183:FTMS>2.0.CO;2. Vavrus, S., M. Notaro, and A. Zarrin, 2013: The Role of Ice Cover in Heavy Lake-Effect Snowstorms over the Great Lakes Basin as Simulated by RegCM4. Mon. Weather Rev., 141, 148–165, doi:10.1175/MWR-D-12-00107.1. Veals, P. G., and W. J. Steenburgh, 2015: Climatological Characteristics and Orographic Enhancement of Lake-Effect Precipitation East of Lake Ontario and over the Tug Hill Plateau. Mon. Weather Rev., 143, 3591–3609, doi:10.1175/MWR-D-15-0009.1. http://journals.ametsoc.org/doi/10.1175/MWR-D-15-0009.1. ——, W. J. Steenburgh, and L. S. Campbell, 2018: Factors affecting the inland and orographic enhancement of lake-effect precipitation over the Tug Hill Plateau. Mon. Weather Rev., 146, 1745–1762. Villarini, G., G. A. Vecchi, and J. A. Smith, 2010: Modeling the Dependence of Tropical Storm Counts in the North Atlantic Basin on Climate Indices. Mon. Weather Rev., 138, 2681– 2705, doi:10.1175/2010MWR3315.1. Vinzani, P. G., and S. A. Changnon, 1981: A case study: 1980’s surprise long-track snowstorm. Weatherwise, 34, 74–76. Vitasse, Y., A. J. Porté, A. Kremer, R. Michalet, and S. Delzon, 2009: Responses of canopy duration to temperature changes in four temperate tree species: Relative contributions of spring and autumn leaf phenology. Oecologia, 161, 187–198, doi:10.1007/s00442-009-1363-4. 280
Vose, R. S., and Coauthors, 2014: Monitoring and understanding changes in extremes: Extratropical storms, winds, and waves. Bull. Am. Meteorol. Soc., 95, 377–386, doi:10.1175/BAMS-D-12-00162.1. Wagenmakers, E., and S. Farrell, 2004: AIC model selection using Akaike weights. Psychon. Bul. Rev., 11, 192–196, doi:10.3758/BF03206482. Walker, G. T., and E. W. Bliss, 1932: World Weather V. Mem. R. Meteorol. Soc., IV, 54–84, doi:10.1002/qj.49705422601. Wallace, J. M., and D. S. Gutzler, 1981: Teleconnections in the Geopotential Height Field during the Northern Hemisphere Winter. Mon. Weather Rev., 109, 784–812, doi:10.1175/1520-0493(1981)109<0784:TITGHF>2.0.CO;2. http://journals.ametsoc.org/doi/abs/10.1175/1520- 0493%281981%29109%3C0784%3ATITGHF%3E2.0.CO%3B2. Walsh, J. E., D. R. Tucek, and M. R. Peterson, 1982: Seasonal snow cover and short-term climatic fluctuations over the United States. Mon. Weather Rev., 110, 1474–1485, doi:10.1175/1520-0493(1982)110<1474:SSCAST>2.0.CO;2. http://journals.ametsoc.org/doi/abs/10.1175/1520- 0493(1982)110%3C1474:SSCAST%3E2.0.CO%3B2. Walther, G.-R., and Coauthors, 2002: Ecological responses to recent climate change. Nature, 416, 389–395, doi:10.1038/416389a. http://www.nature.com/doifinder/10.1038/416389a. Wang, J., X. Bai, H. Hu, A. Clites, M. Colton, and B. Lofgren, 2012: Temporal and spatial variability of Great Lakes ice cover, 1973-2010. J. Clim., 25, 1318–1329, doi:10.1175/2011JCLI4066.1. Wang, M., J. E. Overland, V. Kattsov, J. E. Walsh, X. Zhang, and T. Pavlova, 2007: Intrinsic versus forced variation in coupled climate model simulations over the Arctic during the twentieth century. J. Clim., 20, 1093–1107, doi:10.1175/JCLI4043.1. Ware, E. C., D. M. Schultz, H. E. Brooks, P. J. Roebber, and S. L. Bruening, 2006: Improving Snowfall Forecasting by Accounting for the Climatological Variability of Snow Density. Weather Forecast., 21, 94–103, doi:10.1175/WAF903.1. Warren, S. G., and W. J. Wiscombe, 1980: A Model for the Spectral Albedo of Snow. II: Snow Containing Atmospheric Aerosols. J. Atmos. Sci., 37, 2734–2745. Weiss, C. C., and P. J. Sousounis, 1999: A Climatology of Collective Lake Disturbances. Mon. Weather Rev., 127, 565–574, doi:10.1175/1520- 0493(1999)127<0565:ACOCLD>2.0.CO;2. Westerling, A. L., H. G. Hidalgo, D. R. Cayan, and T. W. Swetnam, 2006: Warming and earlier spring increase western U.S. forest wildfire activity. Science (80-. )., 313, 940–943. Wettstein, J. J., and L. O. Mearns, 2002: The influence of the North Atlantic-Arctic Oscillation on mean, variance, and extremes of temperature in the Northeastern United States and Canada. J. Clim., 15, 3586–3600, doi:10.1175/1520- 0442(2002)015<3586:TIOTNA>2.0.CO;2. Whan, K., F. Zwiers, and J. Sillmann, 2016: The influence of atmospheric blocking on extreme winter minimum temperatures in North America. J. Clim., 29, 4361–4381, doi:10.1175/JCLI-D-15-0493.1. White, A. B., D. J. Gottas, A. F. Henkel, P. J. Neiman, F. M. Ralph, and S. I. Gutman, 2010: Developing a Performance Measure for Snow-Level Forecasts. J. Hydrometeorol., 11, 739–753, doi:10.1175/2009JHM1181.1. 281
http://journals.ametsoc.org/doi/abs/10.1175/2009JHM1181.1. Whittaker, L. M., and L. H. Horn, 1981: Geographical and Seasonal Distribution of North American Cyclogenesis, 1958–1977. Mon. Weather Rev., 109, 2312–2322, doi:10.1175/1520-0493(1981)109<2312:GASDON>2.0.CO;2. Wilson, J. W., 1977: Effect of Lake-Ontario on Precipitation. Mon. Weather Rev., 105, 207– 214, doi:10.1175/1520-0493(1977)105<0207:EOLOOP>2.0.CO;2. Wise, E. K., M. L. Wrzesien, M. P. Dannenberg, and D. L. McGinnis, 2015: Cool-season precipitation patterns associated with teleconnection interactions in the United States. J. Appl. Meteorol. Climatol., 54, 494–505, doi:10.1175/JAMC-D-14-0040.1. Wobus, C., and Coauthors, 2017: Projected climate change impacts on skiing and snowmobiling: A case study of the United States. Glob. Environ. Chang., 45, 1–14, doi:10.1016/j.gloenvcha.2017.04.006. https://doi.org/10.1016/j.gloenvcha.2017.04.006. Wong, G., D. Maraun, M. Vrac, M. Widmann, J. M. Eden, and T. Kent, 2014: Stochastic Model Output Statistics for Bias Correcting and Downscaling Precipitation Including Extremes. J. Clim., 27, 6940–6959, doi:10.1175/JCLI-D-13-00604.1. Woolhiser, D. A., 2008: Combined effects of the Southern Oscillation index and the Pacific decadal oscillation on a stochastic daily precipitation model. J. Clim., 21, 1139–1152, doi:10.1175/2007JCLI1862.1. Woollings, T., and M. Blackburn, 2012: The north Atlantic jet stream under climate change and its relation to the NAO and EA patterns. J. Clim., 25, 886–902, doi:10.1175/JCLI-D- 11-00087.1. Wright, D. M., D. J. Posselt, and A. L. Steiner, 2013: Sensitivity of Lake-Effect Snowfall to Lake Ice Cover and Temperature in the Great Lakes Region. Mon. Weather Rev., 141, 670–689, doi:10.1175/MWR-D-12-00038.1. Wu, H., K. G. Hubbard, and J. You, 2005: Some concerns when using data from the cooperative weather station networks: A Nebraska case study. J. Atmos. Ocean. Technol., 22, 592–602, doi:10.1175/JTECH1733.1. Wyrtki, K., 1985: Water displacements in the Pacific and the genesis of El Niño cycles. J. Geophys. Res., 90, 7129–7132, doi:10.1029/JC090iC04p07129. Yarnal, B., and H. F. Diaz, 1986: Relationships between extremes of the Southern oscillation and the winter climate of the Anglo‐American Pacific Coast. J. Climatol., 6, 197–219, doi:10.1002/joc.3370060208. Yoo, W., R. Mayberry, S. Bae, K. Singh, Q. He, and J. W. Lillard, 2014: A study of effect of multicollinearity in the multivariable analysis. J. Appl. Sci. Technol., 4, 9–19, doi:10.1016/j.dcn.2011.01.002.The. Younkin, R. J., 1968: Circulation patterns associated with heavy snowfall over the western United States. Mon. Weather Rev., 96, 851–853. Yu, B., X. L. Wang, X. B. Zhang, J. Cole, and Y. Feng, 2014: Decadal covariability of the northern wintertime land surface temperature and atmospheric circulation. J. Clim., 27, 633–651. Yu, Z., S. Liu, J. Wang, P. Sun, W. Liu, and D. S. Hartley, 2013: Effects of seasonal snow on the growing season of temperate vegetation in China. Glob. Chang. Biol., 19, 2182–2195, doi:10.1111/gcb.12206. Yuen, C.-W., and J. A. Young, 1986: Dynamical adjustment theory for boundary layer flow in cold surges. J. Atmos. Sci., 43, 3089–3108, doi:10.1175/1520- 282
0469(1986)043<3089:DATFBL>2.0.CO;2. Zarzycki, C. M., 2017: Finding Snowmageddon: Detecting and quantifying northeastern U.S. snowstorms in a multi-decadal global climate ensemble. American Geophysical Union, New Orleans, LA, A51K-07. Zhang, K., B. C. Douglas, and S. P. Leatherman, 2000: Twentieth-century storm activity along the U.S. East Coast. J. Clim., 13, 1748–1761, doi:10.1175/1520- 0442(2000)013<1748:TCSAAT>2.0.CO;2. Zhang, T., 2005: Influence of seasonal snow cover on the ground thermal regime: an overview. Rev. Geophys., 43, RG4002, doi:10.1029/2004RG000157.1.INTRODUCTION. Zhu, C., D. P. Lettenmaier, and T. Cavazos, 2005: Role of Antecedent Land Surface Conditions on North American Monsoon Rainfall Variability. J. Clim., 18, 3104–3121, doi:10.1175/JCLI3387.1. http://journals.ametsoc.org/doi/abs/10.1175/JCLI3387.1. Zhu, Y., and T. Wang, 2016: The relationship between the Arctic Oscillation and ENSO as simulated by CCSM4. Atmos. Ocean. Sci. Lett., 9, 198–203, doi:10.1080/16742834.2016.1149287. http://www.tandfonline.com/doi/full/10.1080/16742834.2016.1149287. Zielinski, G. A., 2002: A classification scheme for winter storms in the Eastern and Central United States with an emphasis on Nor’easters. Bull. Am. Meteorol. Soc., 83, 37–51, doi:10.1175/1520-0477(2002)083<0037:ACSFWS>2.3.CO;2. Zishka, K. M., and P. J. Smith, 1980: The Climatology of Cyclones and Anticyclones over North America and Surrounding Ocean Environs for January and July 1950–77. Mon. Weather Rev., 108, 387–401, doi:10.1175/1520- 0493(1981)109<1356:COCOCA>2.0.CO;2. Ziska, L. H., D. M. Blumenthal, G. B. Runion, E. R. Hunt, and H. Diaz-Soltero, 2011: Invasive species and climate change: An agronomic perspective. Clim. Change, 105, 13–42, doi:10.1007/s10584-010-9879-5. Zulauf, M. A., 2003: Two-dimensional cloud-resolving modeling of the atmospheric effects of Arctic leads based upon midwinter conditions at the Surface Heat Budget of the Arctic Ocean ice camp. J. Geophys. Res., 108, 4312, doi:10.1029/2002JD002643. http://doi.wiley.com/10.1029/2002JD002643.
283
Justin J. Hartnett, Ph.D. Email: [email protected] | Phone: 607-591-9341 7098 State Route 91, Tully, NY 13159
EDUCATION Syracuse University, Syracuse, NY Ph.D. Geography, successfully defended thesis on August 20, 2019 Dissertation: “Snowstorms in Upstate New York: synoptics, spatial modeling and temporal variability.” Committee: Susan W.S. Millar (Advisor), Peng Gao, Jane Read, Jacob Bendix, Adam W. Burnett, and Melissa L. Chipman
University of South Florida, Tampa, FL M.S. Environmental Science and Policy, successfully defended thesis on March 18, 2013 Thesis: “Spatial and temporal trends of snowfall in Central New York – a lake-effect dominated region.” Committee: Jennifer M. Collins (Advisor), Martin A. Baxter, and Don P. Chambers
Coastal Carolina University, Conway, SC B.S. Marine Science with a minor in Physics, May 2011 Summa Cum Laude Alumnus of the Honors College Honors project advisor: Dr. Paul Gayes
ACADEMIC EMPLOYMENT The State University of New York College at Oneonta, Visiting Professor, Department of Geography and Environmental Sustainability Introduction to Geography (GEOG 100) Fall 2017 – Fall 2019 Physical Geography of the Global Environment (GEOG 201) Fall 2017 – Fall 2019 New Orleans Disaster Relief (GEOG 397) May 2019 Coastal Zone Management (GEOG 232) Spring 2018 & 2019 Environmental Applications of GIS (GEOG 244) Spring 2019 Puerto Rico Disaster Relief (GEOG 397) July 2018 Geography of a Changing Climate (GEOG 304) Fall 2017
Syracuse University, Adjunct Professor, Geography Department The Natural Environment (GEO 155) Summer 2016 Global Environmental Change (GEO 215) Fall 2016
Syracuse University, Teaching Assistant, Geography Department The Natural Environment (GEO 155) Fall 2014 – Spring 2016 Instructor of record: Jacob Bendix, Ph.D. Global Environmental Change (GEO 215) Fall 2013 Instructor of record: Jane Read, Ph.D.
Syracuse University, Research Assistant, Geography Department Spring 2014 Supervisor: Peng Gao, Ph.D.
University of South Florida, Teaching Assistant, Geography, Environment, and Policy Department Weather Studies (MET 4010) Fall 2011 – Summer 2013 Instructor of record: Jennifer Collins, Ph.D.
Page 1 of 7
REFEREED PUBLICATIONS 2018 Bendix, J., and J.J. Hartnett: Asynchronous lightning and Santa Ana winds highlight human role in southern California fire regimes. Environmental Research Letters, 13 (7), 074024.
2016 Gao, P., and J.J. Hartnett: Exploring the causes of an extreme flood event in Central New York, USA. Physical Geography, 37, 38-55.
2014 Hartnett, J.J., J.M. Collins, M.A. Baxter, & D.P. Chambers: Spatiotemporal snowfall trends in Central New York. Journal of Applied Meteorology and Climatology, 53, 2685-2697.
MANUSCRIPTS IN PREPARATION Hartnett, J.J.: An updated classification scheme for identifying snowstorm types in the Great Lakes region. To be submitted to the Journal of Climate.
Hartnett, J.J.: Teasing out the seasonal contribution of different snowstorm types to seasonal snowfall totals to the lee of Lake Ontario. To be submitted to the Journal of Applied Meteorology and Climatology.
Hartnett, J.J.: Synoptic conditions associated with different snowstorms and snowfall magnitudes in central New York State. To be submitted to the Journal of Great Lakes Research.
Hartnett, J.J.: Identifying new techniques to examine forcings behind spatiotemporal snowfall variability in the Great Lakes region. To be submitted to Weather and Forecasting.
NON-REFEREED PUBLICATIONS 2014 Hartnett, J.J., J.M. Collins, M.A. Baxter, & D.P. Chambers: The spatial and temporal variability of snowfall trends in Central New York. 71st Eastern Snow Conference Proceedings. Appalachian State University, Boone, NC.
2013 Hartnett, J.J. & J.M. Collins: WCFLAMS Annual Banquet and Presentation on “Hurricane Probability in Tampa.” American Meteorological Society, National Chapter Newsletter, April. 2013 Hartnett, J.J., K. Roberts, & J.M. Collins: Towards building and implementing a Regional Coastal Observing System (RCOOS) for the southeast region of the U.S. American Meteorological Society, National Chapter News Letter, March.
2013 Hartnett, J.J. & J.M. Collins: Panel of Broadcasters. American Meteorological Society, National Chapter News Letter, February.
2012 Collins, J.M., J.J. Hartnett, & C. Gauthier: Web quizzing (15 chapters) for Meteorology Today by C. Donald Ahrens. Cengage Learning.
2012 Roache, D.R., J.J. Hartnett, & J.M. Collins: Communicating storm surge risk presentation by Dr. Betty Morrows. American Meteorological Society, National Chapter News Letter, February.
Page 2 of 7
ACADEMIC LECTURES & PRESENTATIONS Conference Presentations Paper Presentations American Association of Geographers 2018 Annual Meeting, New Orleans, LA “Variations in snowfall distributions due to upper-atmospheric air patterns in Upstate New York.” American Association of Geographers 2017 Annual Meeting, Boston, MA “The Contribution of Upper-Level Flow Regimes to Seasonal Snowfall Totals in Upstate New York” American Association of Geographers 2017 Annual Meeting, Boston, MA Coauthor with Jacob Bendix on “Santa Ana Winds and Lightning: the Role of Timing in Southern California Wildfires.” American Meteorological Society’s 2016 Annual Meeting, New Orleans, LA 22nd Applied Climatology Conference “The Contribution of Upper-Level Flow Regimes to Seasonal Snowfall Totals in Upstate New York.” American Association of Geographers 2014 Annual Meeting, Tampa, FL “Spatiotemporal Snowfall Trends in Central New York.” American Association of Geographers 2014 Annual Meeting, Tampa, FL Presented work with Susan Millar titled, “Updating the Permafrost Realm at the Time of the Last Glacial Maximum (LGM) in North America.”
Poster Presentations American Association of Geographers 2019 Annual Meeting, Washington D.C. “Exploring the effects of different snowstorm types on school closings in central New York State” American Geophysical Union Annual Conference 2016, San Francisco, CA Coauthor with Jacob Bendix on “Asynchronous Timing of Lightning Strikes and Santa Ana Winds in Southern California.” National Center for Atmospheric Research’s 2015 Triannual Unidata Workshop, Boulder, CO “The Contribution of Lake Effect Snowfall to Seasonal Snowfall Totals in Central New York.” American Meteorological Society’s Annual Conference on Broadcast Meteorology/Conference on Weather Warnings and Communication 2015, Raleigh, NC “Communicating Snowfall Changes within Central New York.” Eastern Snow Conference 2014, Appalachian State University, Boone, NC “Spatiotemporal Snowfall Trends of Central New York.”
Invited Lectures 2017 “Plate Tectonics.” Geography 155 – the Natural Environment Lecture, Spring 2017.
2016 “Hydrology: Modelling Water Flow.” Geography 155 – the Natural Environment Lecture, Spring 2016.
2015 “Hydrology: The Water Cycle.” Geography 155 – the Natural Environment Lecture, Fall 2015.
2015 “Plate Tectonics.” Geography 155 – the Natural Environment Lecture, Spring 2015.
2014 “Plate Tectonics.” Geography 155 – the Natural Environment Lecture, Fall 2014.
Page 3 of 7
2014 Eastern Snow Conference, Appalachian State University, Boone, NC “Spatiotemporal Snowfall Trends in Central New York” 2013 University of South Florida, “Weather, Climate, and Society” Guest Lecturer, Tampa, FL “The Development and Implementation of Research Methods in a Climate Based Study”
2012 National Center for Atmospheric Research’s Triannual Unidata Workshop, Boulder, CO “An Introduction to Integrated Data Viewer (IDV) Through an Examination of Recent Atlantic Hurricanes.”
2011 West Central Florida American Meteorological Society’s Teach for Teachers Workshop, Tampa, FL “The Deadly Outbreak – Spring 2011 Tornado Season” “Enjoy Sunny Florida – the Formation of Snow in the Great Lakes Basin”
2009-11 Hartnett Elementary Guest Lecturer, Truxton, NY Multiple presentations for grades 4-6 on subject matter ranging from marine science to atmospheric science.
AWARDS & FELLOWSHIPS 2019 Graduate Student Organization Travel Award, Syracuse University ($300) 2018 SUNY Oneonta Discretionary Lump Sum Payment Award ($1500) 2018 SUNY Oneonta Faculty Development Grant ($1125) 2018 Roscoe-Martin Graduate Award, Maxwell School of Syracuse University ($1200) 2018 Geography Department Student Travel Award, Syracuse University ($500) 2017 Graduate Student Organization Travel Award, Syracuse University ($300) 2017 Geography Department Summer Research Award, Syracuse University ($1500) 2016 Best Student Presentation, 22nd Applied Climatology Conference, 2nd Place ($50) 2016 Maxwell Dean’s Summer Research Award, Syracuse University ($2000) 2016 Graduate Student Organization Travel Award, Syracuse University ($550) 2015 Roscoe-Martin Graduate Award, Maxwell School of Syracuse University ($1200) 2015 Student Travel Award, University Corporation for Atmospheric Research’s Unidata Workshop 2015 Maxwell Dean’s Summer Research Award, Syracuse University ($2000) 2014 Weisnet Medal, best student paper at the Eastern Snow Conference ($750) 2014 Graduate Student Organization Travel Award, Syracuse University ($400) 2014 Maxwell Dean’s Summer Research Award, Maxwell School of Syracuse University ($1000) 2013 Roscoe-Martin Graduate Award, Maxwell School of Syracuse University ($1100) 2013 Tharpe Fellowship, Geography, Environment, and Planning Department at the University of South Florida ($400) 2012 Dewey Stower’s Merit Award, West Central Florida American Meteorological Society 2012 University Corporation for Atmospheric Research Presentation Award 2011 Summa Cum Laude, Coastal Carolina University 2011 Honors College Student Excellence Award, Coastal Carolina University 2010 Honors College Student Excellence Award, Coastal Carolina University 2009 Omicron Delta Kappa Honors Society, Coastal Carolina University 2009 Golden Key Honors Society, Coastal Carolina University
Page 4 of 7
PROFESSIONAL SERVICE Interview for the Daily Orange, Syracuse University’s Newspaper October 4, 2019 Task: I was interviewed about the ways climate change has impacted Central New York Board of Directors for Truxton Alumni and Community Supporters Non-for-profit Organization October 2015 – Present Task: Helping to maintain, run, and develop a community organization in support of the residents of the Town of Truxton. Duties include developing programs, making financial decisions, and enhancing community engagement. Faculty sponsor for the Geography Club, SUNY Oneonta August 2019 – Present Geofest at Syracuse University September 2018 Task: I helped organize a group of undergraduate students from SUNY Oneonta to attend an information fair on graduate school in geography at Syracuse University Journal Article Reviewer Journal of Applied Meteorology and Climatology Climate Change Physical Geography Journal of Hydrometeorology SUNY Oneonta Climate Change Certificate Committee Member November 2017 – Present Task: Aiding in the development of implementation of a Climate Change Certificate in the SUNY Oneonta curriculum. SUNY Oneonta President’s Conversation on Diversity with Dr. Peggy McIntosh April 2018 Task: Attended this workshop in diversity, understanding white privilege, and how to be more mindful of unintentional biases in the classroom.
PROFESSIONAL MEMBERSHIP American Meteorological Society Member Fall 2011 – Present American Association of Geographers Member Fall 2013 – Present Future Professoriate Program, Syracuse University Fall 2013-Spring 2016 West Central Florida American Meteorological Society Fall 2011 – Present Corresponding Secretary (June 2012 – June 2013) Recording Secretary (June 2011 – June 2012) Member of the Organizing Committee (June 2011 – June 2013)
STUDENT ADVISING 2019 Mentor for an undergraduate semester project. Students: Ian Devlin, Nick Lindovski, Ema Serra, and Louis Hellers. Affiliation: Department of Geography and Environmental Sustainability. Topic: Commonalities in school-closing snowstorms in Central New York. 2018 Mentor for an undergraduate research project. Student: Ceili Getman. Affiliation: Department of Geography and Environmental Sustainability. Topic: Sustainability food practices.
Page 5 of 7
SOFTWARE KNOWLEDGE ESRI’s Geographic Information System (ArcGIS) Exelis Visual Information Solutions Inc.’s Environment for Visualizing Images (ENVI) GRLevel 3 & GR2Analyst IBM’s Statistical Package for the Social Sciences (SPSS) The R Project for Statistical Computing (R) Scilab Enterprise’s Scilab Unidata’s Integrated Data Viewer (IDV) Wolfram’s Mathematica
RESEARCH & TEACHING INTERESTS Climatology & Meteorology Climatology of severe and extreme weather in North America Natural disasters – resilience and risk Ocean/atmosphere interactions and dynamics (teleconnection patterns) Snowstorms and snowfall Geospatial Techniques and Analysis Applied classical statistics Geographic Information Sciences Geospatial statistics Physical Oceanography Coastal erosion Sea level rise Storm surge
RESEARCH EXPERIENCE & FIELD WORK Current Collaboration with Adam Burnett (Colgate U.), Art Samel (Bowling Green U.), and Chris Karmosky (SUNY Oneonta) examining the use of a classification scheme capable of predicting when extreme snowfall events will occur in the Tug Hill of New York State. Current Collaboration with Mark Welford (Georgia Southern U.) and Jennifer Collins (U. of South Florida) examining the influence of climate variability on bird migration pattern along the Chilean coastline. 2019 Standardization of “snow days” in Central New York. The objective of this research is to develop a way to standardize snow days in Upstate New York using regular school closures and snowstorms identified in my dissertation work. 2017 Collaboration with Jacob Bendix (Syracuse U.) analyzing the atmospheric conditions associated with the onset of California wildfires from lightning strikes during Santa Ana winds. Research published in the Environmental Research Letters. 2017 Literature review on the historical development of lake-effect snow research with Dr. Susan Millar (Syracuse U.). The purpose was to develop a comprehensive review of research papers on lake-effect snow, and to provide an overview of the main findings of past, current, and future expected trends for lake-effect snow. 2016 NSF submitted proposal with Jacob Bendix (Syracuse U.) and John Stella (SUNY ESF). This study examined the historical changes and projected changes to riparian fires in Southern California and the Sierra Nevada Mountains. 2014 Hydrological studies with Peng Gao (Syracuse U.). This research utilized the Dynamic Watershed Simulation Model to simulate the causes of an extreme (100-year event) flood in the Oneida Creek Watershed in upstate New York. Work published in Physical Geography.
Page 6 of 7
2014 Field work with Susan Millar (Syracuse U.) retrieving, updating, and analyzing soil and temperature loggers throughout Central New York. The purpose of this study was to determine the importance of snow cover (extent and depth) on insulating the soil from the overlying air. 2011 Hurricane Evacuation Study under Jennifer Collins (U. of South Florida). This study was funded by the National Science Foundation and examined the plans and knowledge of individuals in Florida during a hurricane evacuation. 2011 Rip Current study collaborated with the National Weather Service, Ruskin, FL. This study examined the small- and large-scale atmospheric conditions leading up to rip current deaths throughout the United States. 2010 “Storm Surge Inundation from a Hurricane as Sea Level Rises” research project under Paul Gayes (Coastal Carolina U.) and Leonard Pietrafesa (North Carolina St. U.). This study observed coastal inundation expected along the North and South Carolina coastlines from tropical storms as sea levels rise. 2009 Paleotempestology and paleoclimatology study with Jenna Hill (Coastal Carolina U.). The purpose of this study was to use sediment cores to examine the historical climate and environment of a coastal marsh in northern Southern Carolina.
Page 7 of 7