• Branick, M. L., 1997: A climatology of significant winter-type weather events in the contiguous United States, 1982–94. Wea. Forecasting, 12, 193207, https://doi.org/10.1175/1520-0434(1997)012<0193:ACOSWT>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Changnon, S. A., 1967: A method of evaluating substation records of hail and thunder. Mon. Wea. Rev., 95, 209212, https://doi.org/10.1175/1520-0493(1967)095<0209:MOESRO>2.3.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Changnon, S. A., and T. Creech, 2003: Sources of data on freezing rain and resulting damages. J. Appl. Meteor., 42, 15141518, https://doi.org/10.1175/1520-0450(2003)042<1514:SODOFR>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Changnon, S. A., and D. Changnon, 2005: Damaging snowstorms in the U.S. Nat. Hazards, 31, 117.

  • Changnon, S. A., and K. E. Kunkel, 2006: Changes in instruments and sites affecting historical weather records: A case study. J. Atmos. Oceanic Technol., 23, 825828, https://doi.org/10.1175/JTECH1888.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Changnon, S. A., D. Changnon, and T. Karl, 2006: Temporal and spatial characteristics of snowstorms in the contiguous United States. J. Appl. Meteor. Climatol., 45, 11411155, https://doi.org/10.1175/JAM2395.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Changnon, S. A., D. Changnon, T. Karl, and T. Houston, 2008: Snowstorms across the nation: An atlas about storms and their damages. National Climatic Data Center, 96 pp.

  • Compo, G. P., and Coauthors, 2011: The Twentieth Century Reanalysis Project. Quart. J. Roy. Meteor. Soc., 137, 128, https://doi.org/10.1002/qj.776.

  • Doesken, N. J., and A. Judson, 1996: The snow booklet: A guide to the science, climatology, and measurement of snow in the United States. Dept. of Atmospheric Science, Colorado State University, 87 pp., http://ccc.atmos.colostate.edu/pdfs/snowbook.pdf.

  • Durre, I., M. P. 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. Meteor. Climatol., 52, 23772395, https://doi.org/10.1175/JAMC-D-13-051.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Eichenlaub, V. L., 1970: Lake-effect snowfall to the lee of the Great Lakes: Its role in Michigan. Bull. Amer. Meteor. Soc., 51, 403412, https://doi.org/10.1175/1520-0477(1970)051<0403:LESTTL>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Harrington, J. A., Jr., R. S. Cerveny, and K. F. Dewey, 1987: A climatology of mean monthly snowfall for conterminous United States: Temporal and spatial patterns. J. Climate Appl. Meteor., 26, 897912, https://doi.org/10.1175/1520-0450(1987)026<0897:ACOMMS>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Kalnay, E., and Coauthors, 1996: The NCEP/NCAR Reanalysis 40-Year Project. Bull. Amer. Meteor. Soc., 77, 437471, https://doi.org/10.1175/1520-0477(1996)077<0437:TNYRP>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Kocin, P. J., and L. W. Uccellini, 2005: Northeast Snowstorms. Vols. 1 and 2, Meteor. Monogr., No. 54, Amer. Meteor. Soc., 818 pp.

    • Crossref
    • Export Citation
  • Kunkel, K. E., 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. Oceanic Technol., 24, 6473, https://doi.org/10.1175/JTECH2017.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Kunkel, K. E., M. A. Palecki, L. Ensor, D. R. Easterling, K. G. Hubbard, D. A. Robinson, and K. T. Redmond, 2009: Trends in twentieth-century U.S. extreme snowfall seasons. J. Climate, 22, 62046216, https://doi.org/10.1175/2009JCLI2631.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Lawrimore, J., T. Karl, M. Squires, D. A. Robinson, and K. E. Kunkel, 2014: Trends and variability in severe snowstorms east of the Rocky Mountains. J. Hydrometeor., 15, 17621777, https://doi.org/10.1175/JHM-D-13-068.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Marinaro, A., S. Hilberg, D. Changnon, and J. R. Angel, 2015: The North Pacific–driven severe Midwest winter of 2013/14. J. Appl. Meteor. Climatol., 54, 21412151, https://doi.org/10.1175/JAMC-D-15-0084.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Menne, M. J., I. Durre, R. S. Vose, B. E. Gleason, and T. G. Houston, 2012: An overview of the Global Historical Climatology Network-Daily database. J. Atmos. Oceanic Technol., 29, 897910, https://doi.org/10.1175/JTECH-D-11-00103.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Moftakhari, H. R., A. AghaKouchak, B. F. Sanders, and R. A. Mathews, 2017: Cumulative hazard: The case of nuisance flooding. Earth’s Future, 5, 214223, https://doi.org/10.1002/2016EF000494.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Mook, C. P., and K. S. Norquest, 1956: The heavy snowstorm of March 18-19, 1956: The climax of a record late-season snow accumulation in southern New England. Mon. Wea. Rev., 84, 116125, https://doi.org/10.1175/1520-0493(1956)084<0116:THSOM>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • MRCC, 2017: Cli-MATE: MRCC Application Tools Environment. Midwestern Regional Climate Center, http://mrcc.isws.illinois.edu.

  • National Assessment Synthesis Team, 2001: Climate change impacts on the United States: The potential consequences of climate variability and change. National Assessment Synthesis Team Rep., 541 pp., https://data.globalchange.gov/assets/e9/97/436129058f2107f4925aeec13ed8/nca-2000-foundation-report.pdf

  • NOAA/NCEI, 2017: U.S. billion-dollar weather and climate disasters 1980-2017. NOAA/National Centers for Environmental Information, https://www.ncdc.noaa.gov/billions/events.pdf.

  • Norton, D. C., and S. J. Bolsenga, 1993: Spatiotemporal trends in lake effect and continental snowfall in the Laurentian Great Lakes, 1951–1980. J. Climate, 6, 19431956, https://doi.org/10.1175/1520-0442(1993)006<1943:STILEA>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Robinson, D. A., 1989: Construction of a United States historical snow data base. Proc. 45th Eastern Snow Conf., Lake Placid, NY, Eastern Snow Conference, 50–59.

  • Schwartz, R. M., and T. W. Schmidlin, 2002: Climatology of blizzards in the conterminous United States, 1959–2000. J. Climate, 15, 17651772, https://doi.org/10.1175/1520-0442(2002)015<1765:COBITC>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • U.S. Weather Bureau, 1922: Measurement of precipitation: Instructions. U.S. Dept. of Agriculture Circular E, 36 pp.

  • U.S. Weather Bureau, 1970: Instructions for Cooperative (Climatological) Observers. U.S. Department of Commerce Observing Handbook 2, 77 pp.

  • View in gallery

    Top 30-day heavy snowfall amounts for NWS stations in western Illinois, southeastern Iowa, and northeastern Missouri. Macomb, identified by the star, was removed from the study because of its unusually high value.

  • View in gallery

    The distribution of NWS stations with quality snowfall data during the 1900–2016 period. Florida was not included in this study.

  • View in gallery

    The greatest 30-day snowfall amount (cm) during the 1900–2016 period.

  • View in gallery

    The average of the top five 30-day snowfall amounts (cm) during the 1900–2016 period.

  • View in gallery

    The ratio of the greatest 30-day snowfall amount to the average winter snowfall (1981–2010).

  • View in gallery

    The ratio of the average of the top five 30-day snowfall amounts to the average winter snowfall (1981–2010).

  • View in gallery

    Winter number of stations experiencing the greatest 30-day snowfall amount, from 1900/01 to 2015/16.

  • View in gallery

    Winter number of stations experiencing one of the top five 30-day heavy snowfall amounts, 1900/01–2015/16.

  • View in gallery

    Number of stations experiencing one of the top five 30-day heavy snowfall amounts in independent 4-yr periods (1901–04, 1905–08, …, 2013–16).

  • View in gallery

    General frequency of top-five 30-day snowfall amounts by state through the 1900–2016 period. The word early indicates a greater frequency of heavy snowfall amounts early (before 1959), mix indicates a balanced occurrence of events through the study period, and late indicates a greater frequency of heavy snowfall amounts in the latter half of the period (in or after 1959).

  • View in gallery

    Snowfall episodes for which 10% or more (>50) stations experienced one of their top five 30-day snowfall amounts: (a) 28 Jan–25 Mar 1912, (b) 20 Dec 1917–2 Feb 1918, (c) 7 Feb–28 Mar 1960, (d) 31 Dec 1977–24 Feb 1978, (e) 29 Dec 1978–25 Feb 1979, and (f) 17 Dec 1995–4 Feb 1996.

  • View in gallery

    Composite 500-hPa geopotential height (m) anomalies for the six episodes shown in Figs. 11a–f. Anomalies were calculated using the 1981–2010 average. Solid lines represent positive height anomalies, and dashed lines represent negative height anomalies.

All Time Past Year Past 30 Days
Abstract Views 0 0 0
Full Text Views 37 37 2
PDF Downloads 26 26 2

A Spatial and Temporal Analysis of 30-Day Heavy Snowfall Amounts in the Eastern United States, 1900–2016

View More View Less
  • 1 Department of Geographic and Atmospheric Sciences, Northern Illinois University, DeKalb, Illinois
© Get Permissions
Full access

Abstract

Heavy 30-day snowfall amounts were evaluated to identify spatial and temporal characteristics east of the Rocky Mountains in the United States during the period 1900–2016. An extensive data assessment identified 507 stations for use in this long-term climate study. The top 30-day heavy snowfall amount and the average of the top five 30-day heavy snowfall amounts were examined. Both amounts generally increased with latitude; however, much higher amounts were found downwind of the Great Lakes, at higher elevations, or in locations impacted by topographic features (e.g., Rockies, Black Hills, and Appalachians). When compared with the 1981–2010 average winter snowfall, the top 30-day amount was found to be greater than the winter average in most areas of the eastern United States. The number of stations experiencing a top-five 30-day heavy snowfall period in a winter ranged from 1 to 128 (1959/60), with a greater overall occurrence in the second half of the 117-yr period. Six episodes had 10% or more stations experiencing one of the top five 30-day snowfall amounts, with the February–March 1960 episode impacting 124 stations, and these episodes were associated with large negative 500-hPa height anomalies. The northern Great Plains, Great Lakes, Midwest, and Northeast experienced more top-five periods in the second half of the 117-yr period, whereas most of the southern states experienced top-five periods throughout the study’s time frame. Examining extremes at periods beyond the daily event and less than the season contributes to our knowledge of climate and provides useful information to snow-sensitive sectors.

© 2018 American Meteorological Society. For information regarding reuse of this content and general copyright information, consult the AMS Copyright Policy (www.ametsoc.org/PUBSReuseLicenses).

E-mail: David Changnon, dchangnon@niu.edu

Abstract

Heavy 30-day snowfall amounts were evaluated to identify spatial and temporal characteristics east of the Rocky Mountains in the United States during the period 1900–2016. An extensive data assessment identified 507 stations for use in this long-term climate study. The top 30-day heavy snowfall amount and the average of the top five 30-day heavy snowfall amounts were examined. Both amounts generally increased with latitude; however, much higher amounts were found downwind of the Great Lakes, at higher elevations, or in locations impacted by topographic features (e.g., Rockies, Black Hills, and Appalachians). When compared with the 1981–2010 average winter snowfall, the top 30-day amount was found to be greater than the winter average in most areas of the eastern United States. The number of stations experiencing a top-five 30-day heavy snowfall period in a winter ranged from 1 to 128 (1959/60), with a greater overall occurrence in the second half of the 117-yr period. Six episodes had 10% or more stations experiencing one of the top five 30-day snowfall amounts, with the February–March 1960 episode impacting 124 stations, and these episodes were associated with large negative 500-hPa height anomalies. The northern Great Plains, Great Lakes, Midwest, and Northeast experienced more top-five periods in the second half of the 117-yr period, whereas most of the southern states experienced top-five periods throughout the study’s time frame. Examining extremes at periods beyond the daily event and less than the season contributes to our knowledge of climate and provides useful information to snow-sensitive sectors.

© 2018 American Meteorological Society. For information regarding reuse of this content and general copyright information, consult the AMS Copyright Policy (www.ametsoc.org/PUBSReuseLicenses).

E-mail: David Changnon, dchangnon@niu.edu

1. Introduction

On 24 January 2015, Boston, Massachusetts, received 13.0 cm of snowfall. Measurable snowfall was observed on 22 of the next 29 days (Table 1)! When summed over that 30-day period, Boston received 239.8 cm (94.4 in.), a record amount for that duration, and more than 2 times the annual average. Impacts related to this unusual amount of snowfall were felt across many sectors. The transportation sector (e.g., highways, railroads, and airports) was severely impacted as the ability to remove snow in a timely fashion became increasingly difficult. Insurance-related property losses from this snowfall period were reported through summer 2015 (J. Darr 2017, personal communication). Losses in Massachusetts exceeded $1 billion, while regional losses from the mid-Atlantic through New England were estimated at $3 billion (NOAA/NCEI 2017). For many in eastern New England, this would be an unforgettable snowfall period. However, this unusually snowy period raised a broader set of questions regarding our understanding of snowfall periods that exceed the 1–2-day events but are shorter than the winter or annual totals. Moftakhari et al. (2017) implied in a recent study that the cumulative effects of more frequent snowfall events may create greater impacts than one extreme event (i.e., blizzard). Analysis results of 30-day heavy snowfall amounts should expand our knowledge of snowfall variability across time and space and be useful as the atmospheric science community tries to define changes in twenty-first-century climate parameters (National Assessment Synthesis Team 2001).

Table 1.

Daily snowfall at Boston (station identifier 190770). The T means trace.

Table 1.

In the United States, extensive snowfall research has focused at understanding the climatology of damaging snowstorms and blizzards (Branick 1997; Schwartz and Schmidlin 2002; Kocin and Uccellini 2005; Changnon and Changnon 2005; Changnon et al. 2006), and trends in monthly and winter snowfall totals (Harrington et al. 1987; Norton and Bolsenga 1993; Kunkel et al. 2009). This study examines the spatial and temporal characteristics of 30-day heavy snowfall amounts for areas east of the Rocky Mountains in the United States for the period 1900–2016. This study builds off a similar research theme by Mook and Norquest (1956), who evaluated a record late-season snowfall period that occurred in southern New England during March–April 1956. In this study we address several questions regarding these unusual heavy snowfall periods: How does the record 30-day snowfall amount vary across the eastern United States? What is the relationship of this heavy 30-day snowfall amount to the average annual snowfall total? Has the frequency of these extreme snowfall periods changed over time across the United States? What 30-day periods are associated with a large number of stations experiencing a heavy snowfall amount?

Daily snowfall observations do not produce exact values and are impacted by measurement errors made by observers, wind, other precipitation types, and station exposure issues that impact data homogeneity (Robinson 1989; Doesken and Judson 1996; Changnon et al. 2006; Kunkel et al. 2007). Snowfall can be the most subjective measurement made by a weather observer. Thus, identifying stations with good long-term homogeneous daily snowfall measurements is difficult and requires time-consuming objective and subjective assessments. For this study, good long-term daily snowfall records for National Weather Service (NWS) stations identified in two recent studies, one focused on snowstorms (Changnon et al. 2006) and the other on winter snowfall totals (Kunkel et al. 2009), were examined and used. However, further assessment of a station’s 30-day heavy snowfall amounts was completed, applying comparative evaluation analyses among adjacent stations, noting when these periods occurred during the 117-yr period and whether the greatest amount or the average of the top five snowfall amounts was too high or too low relative to other nearby stations. These analyses were critical to developing a new snowfall climatology.

The contributions of this research are twofold: 1) development of a new heavy snowfall climatology and 2) delivery of information that can be used to design and/or prepare for future events to weather-sensitive decision-makers impacted by heavy snowfall periods. Although these events occur rarely, developing plans for a worst-case scenario such as what Boston experienced in early 2015 would be useful to decision-makers, as pointed out by Moftakhari et al. (2017).

2. Data and approach

Identifying good long-term NWS first-order and cooperative station records to examine 30-day heavy snowfall periods required an intensive and primarily subjective approach. Although snow depth and meltwater observational procedures changed during the period of interest, daily snowfall measurements had generally consistent instructions (U.S. Weather Bureau 1922, 1970; Changnon et al. 2008). Some practices, such as permitting NWS cooperative observers to add four 6-h observations of snowfall to come up with a daily total, were applied during a few years in some U.S. locations, but were found to have no impact on the daily snowfall record. Thus, there should be little or no systematic shift in snowfall amounts as a result of observational practices. In many daily temperature and precipitation studies, a frequent reason to remove a station is related to the number of missing daily observations. In this study, that was the initial station selection criterion. Similar to the Kunkel et al. (2007) study, those stations that were missing more than 10% of the daily snowfall measurements during 1900–2016 were removed from further consideration.

The selection of stations for this study required further careful examination of the snowfall data due to a plethora of known issues that can negatively impact snowfall observations and cause data inhomogeneity (Robinson 1989; Doesken and Judson 1996). Two earlier studies by Changnon et al. (2006) and Kunkel et al. (2009) developed and used different techniques to identify and remove stations with poor snowfall records. Changnon et al. (2006) primarily applied a subjective approach, initially developed and tested to identify good daily hail and thunder records (Changnon 1967), to develop a final list of good stations for a snowstorm analysis. This procedure, used in the Changnon et al. (2006) study, required that, after each station was analyzed, the temporal frequency of snowstorms for a given station was compared with those at adjacent stations using a double-mass curve approach. This process identified data discontinuities within the historical record. If the calculated frequency was different (too high or too low), then the station was removed from the study. The number of Changnon et al. (2006) stations initially considered for this study was 488. In the Kunkel et al. (2007, 2009) research, a more objective approach was applied to evaluate daily snowfall data through each winter of their study. This process initially identified stations with inhomogeneous records for potential removal, using a number of criteria involving the number of missing daily data per month and winter, winter seasonal snowfall totals that were either too high or low, and the number of winters per decade with near-perfect data. The final decision to keep a station on the “good” list in the Kunkel et al. (2009) study was manually determined by those examining the daily snowfall records. The number of high quality stations identified by Kunkel et al. (2007) and considered for this study, beyond those identified earlier by Changnon et al. (2006), was 26. Upon further review of stations with snowfall records dating back to 1900 (Changnon and Creech 2003; Menne et al. 2012), 210 stations not used in either previous study were identified for further consideration. All 724 long-term stations identified in earlier studies (Changnon et al. 2006; Kunkel et al. 2009) and in this study were then subjected to further scrutiny using the 30-day heavy snowfall amounts.

The top five independent 30-day heavy snowfall periods for each of the 724 stations were acquired from a Midwestern Regional Climate Center (MRCC) tool [the MRCC's Application Tools Environment (cli-MATE; MRCC 2017)]. The snowfall values used in this study were units of depth. The top five periods were chosen to provide an adequate sample of these extreme events and to identify and eliminate stations with a greatest value that was unusually different than the values for the other four periods. The top event averaged 17% more snowfall than the second highest event. Because of how the program in cli-MATE worked, one might have to evaluate 50 or more 30-day periods for a given station to identify the top five independent 30-day heavy snowfall periods. For one 30-day period to be independent from a second period in the same winter, a minimum 10-day break between the two periods had to exist. In rare cases (21 of 724 stations), two 30-day heavy snowfall periods fell in the same winter.

Once the top five independent 30-day heavy snowfall events were identified, three different techniques, similar to those used in Changnon et al. (2006), were applied to further eliminate stations from the study. For each station, the number of 30-day heavy snowfall periods falling in the period prior to 1960 and from 1960 to 2017 was noted. If a station had all five 30-day periods before or after 1960, while adjacent stations generally experienced most if not all their 30-day periods in the other part of the record, then that station was removed. For example, all five 30-day periods experienced at Hoopeston, Illinois, occurred prior to 1960 (all prior to 1915), whereas those Illinois and Indiana stations near Hoopeston experienced four of five or all five 30-day periods after 1960. In cases when the top 30-day snowfall amount was much greater or less than adjacent stations, that station was removed from the study. For example, Macomb, Illinois, reported 136 cm during its greatest 30-day heavy snowfall period. Those western Illinois, northeastern Missouri, and southeastern Iowa stations near Macomb generally observed greatest values that were 35–50 cm less than that for Macomb (Fig. 1). Finally, the average of the top five 30-day snowfall amounts was compared with that value for adjacent stations to determine whether that amount was much greater or less than the adjacent stations. For example, Paris, Illinois, experienced an average value of 83.1 cm for its top five 30-day periods, a value that was 8–13 cm greater than the average of the top-five-period amounts at adjacent stations in Illinois and Indiana. This time-consuming interstation evaluation reduced the number of stations for this study to 507 (Fig. 2). The station distribution is uneven across the study region and this could have some influence on the statistical results; however, other snowfall studies that have incorporated station data have not encountered such issues (Changnon et al. 2008).

Fig. 1.
Fig. 1.

Top 30-day heavy snowfall amounts for NWS stations in western Illinois, southeastern Iowa, and northeastern Missouri. Macomb, identified by the star, was removed from the study because of its unusually high value.

Citation: Journal of Applied Meteorology and Climatology 57, 2; 10.1175/JAMC-D-17-0217.1

Fig. 2.
Fig. 2.

The distribution of NWS stations with quality snowfall data during the 1900–2016 period. Florida was not included in this study.

Citation: Journal of Applied Meteorology and Climatology 57, 2; 10.1175/JAMC-D-17-0217.1

Of these selected NWS stations, 80 (16%) had combined station records (i.e., two station numbers used) to cover the 117-yr period, of which 58 of those 80 (73%) were stations that moved from a downtown to an airport location and are NWS first-order stations. Whereas station moves often influence long-term temperature records, Changnon and Kunkel (2006) noted that when the Urbana, Illinois, cooperative station was moved, no effects on the snowfall amounts were noted. For those combined station records with overlapping snowfall data, 30-day heavy snowfall amounts for similar periods were examined and found to be within 7% of each other in most cases. A further comparison of these combined station records with an adjacent station using a double-mass curve approach indicated that the station move resulted in little impact on the long-term snowfall record.

Although all 507 selected station records had less than 10% missing data, 95 stations had between 5% and 10% missing data. The decades with the greatest number of stations with notable missing data included 1901–10 (34 stations) and 2010–16 (28 stations). In the remaining decades, the maximum number of stations with 5% and 10% missing data was 10.

The top 30-day heavy snowfall amount and the average of the top-five-period amounts were mapped for the 507 stations and the pattern of values evaluated. These two amounts were then divided by the 1981–2010, 30-yr winter average snowfall (Durre et al. 2013) and mapped to determine how unusual the 30-day amounts were compared with the winter average. Temporal variability during the 1900–2016 period was assessed in three ways: 1) the time distribution of the annual number of stations experiencing a 30-day heavy snowfall period in a given winter (1900/01–2015/16); 2) the time distribution of the number of stations experiencing a 30-day period in independent 4-yr periods (1900/01–1903/04, ...., 2012/13–2015/16); and 3) an assessment at the state level by examining the five 30-day events for each station determining whether the station is categorized as early (four or five periods prior to 1960), mixed (two or three prior to 1960), or late (one or no periods prior to 1960).

Major 30-day heavy snowfall episodes were identified and mapped. In each of these episodes, 10% or more (>50) of the stations experienced one of the top five 30-day heavy snowfall periods. When identifying 30-day heavy snowfall episodes based on all state station records, it was expected that 30-day periods would be slightly different (e.g., 4 January–2 February vs 6 January–4 February) based on geography and movement and timing of extratropical cyclones. As long as the 30-day periods overlapped with each other, they were considered to be part of the same heavy snowfall episode. The spatial characteristics of these episodes will be examined. Furthermore, composite 500-hPa geopotential height anomaly maps were developed for each episode and atmospheric features are discussed.

3. Results

The greatest 30-day heavy snowfall period pattern generally increased from less than 25 cm along the Gulf Coast to values exceeding 125 cm in a swath from western South Dakota to Lake Superior and in the northeast United States (Fig. 3). Values were generally much higher to the lee of the Great Lakes, at higher elevations, or across areas impacted by topographic features associated with the Rockies, the Black Hills, and the Appalachians. These features were found to impact the frequency and magnitude of snowstorms in earlier studies (Eichenlaub 1970; Norton and Bolsenga 1993; Changnon et al. 2006). When examining the pattern of the average of the top five 30-day snowfall amounts, a similar spatial distribution was found, with somewhat smaller values (Fig. 4). Awareness of these values can assist local and regional governments, emergency management personnel, and building designers as they plan or design for snow removal, snow loads, and potential transportation issues that could arise if one of these 30-day heavy snowfalls were to occur.

Fig. 3.
Fig. 3.

The greatest 30-day snowfall amount (cm) during the 1900–2016 period.

Citation: Journal of Applied Meteorology and Climatology 57, 2; 10.1175/JAMC-D-17-0217.1

Fig. 4.
Fig. 4.

The average of the top five 30-day snowfall amounts (cm) during the 1900–2016 period.

Citation: Journal of Applied Meteorology and Climatology 57, 2; 10.1175/JAMC-D-17-0217.1

When station values were divided by the station’s average winter snowfall (based on a 30-yr average, 1981–2010), the magnitude of snowfall associated with these 30-day heavy snowfall periods relative to normal winter snowfall could be evaluated (Figs. 5 and 6). Ratios that exceed one indicate that the 30-day heavy snowfall amount is more than the average winter amount. This line generally runs along the northern tier of states and suggests only those regions near Canada experience a 30-day heavy snowfall period amount that is similar to or less than the average winter amount. Although still causing snowfall-related impacts, these 30-day heavy snowfall amounts are not considered extreme. Moving south of the 1.0 line, the potential impacts to local governments and those in emergency management related to experiencing one of the 30-day heavy snowfall amounts increases. The 2.0 ratio line for the top 30-day heavy snowfall amount (Fig. 5) runs west to east across northern Kansas and Missouri, before crossing into central Illinois and Indiana, and is then located northeast from the Appalachians to the New Jersey coast. This boundary was slightly farther south for the average of the top five 30-day snowfall amounts (Fig. 6). Those areas where the ratio is greater than five are regions where the 30-day amounts are at least 5 times greater than the annual average. In most of these locations the annual average snowfall is less than 12 cm.

Fig. 5.
Fig. 5.

The ratio of the greatest 30-day snowfall amount to the average winter snowfall (1981–2010).

Citation: Journal of Applied Meteorology and Climatology 57, 2; 10.1175/JAMC-D-17-0217.1

Fig. 6.
Fig. 6.

The ratio of the average of the top five 30-day snowfall amounts to the average winter snowfall (1981–2010).

Citation: Journal of Applied Meteorology and Climatology 57, 2; 10.1175/JAMC-D-17-0217.1

The time distribution of the number of stations in the eastern United States experiencing their top 30-day snowfall amount in a winter showed a great deal of interannual variability and little trend through the 117-yr period (Fig. 7). Three winters (1911/12, 1917/18, and 1959/60) had ≥4% (≥20 stations) of the 507 stations experiencing their greatest 30-day snowfall, with 45 (9%) stations in 1959/60. The mean winter number of stations increased from three stations during the first 58 yr (1901–58) to five stations in the later 58 yr. When examined as four consecutive 29-yr periods (i.e., 1901–29, 1930–58, 1959–87, and 1988–2016), the values were four, two, seven, and four.

Fig. 7.
Fig. 7.

Winter number of stations experiencing the greatest 30-day snowfall amount, from 1900/01 to 2015/16.

Citation: Journal of Applied Meteorology and Climatology 57, 2; 10.1175/JAMC-D-17-0217.1

A similar time distribution of the number of stations experiencing one of their top five 30-day heavy snowfall amounts in a winter is shown in Fig. 8. The annual number of stations ranged from 1 (1910/11, 1990/91, 2005/06, and 2011/12) to 128 stations (1959/60). Although the time series appears similar to that for the top 30-day snowfall station frequency shown in Fig. 7, there are differences when examining the top 30-day amount occurrence to that of the top five 30-day amount occurrences in the latter half of the 117-yr period. Winters when the number of stations exceeds 50 (≥2% of 2535 station periods) include 1909/10, 1911/12, 1917/18, 1947/48, 1959/60, 1977/78, 1978/79, 1995/96, and 2009/10. When considering the 117-yr period, this indicates winters with a large number of stations experiencing a 30-day heavy snowfall amount should occur about once every 13 yr. However, three winters occurred within a 10-yr period early in the study period and two occurred back to back in the late 1970s. These results are similar to those found by Changnon et al. (2006) with high snowstorm values during the 1911–20 and 1970–80 periods. The average number of stations experiencing one of their top five 30-day snowfall amounts in a winter increased from 17 in the early period (1901–58) to 27 in the later period, a significant increase at the 5% level using the Student’s t test. However, when this dataset is examined in four equal 29-yr periods, the lower value in the first half of the 117-yr period is related to the 1930–58 average (15) and the higher value in the second half is related to the 1959–87 average (34). When examining the number of stations experiencing one of their top five 30-day heavy snowfall periods in independent 4-yr periods, 101 or more stations (≥4% of 2535 station periods) were found to occur during nine 4-yr periods: 1909–12, 1917–20, 1957–60, 1965–68, 1969–72, 1977–80, 1985–88, 1993–96, and 2009–12 (Fig. 9). Similar to the annual distribution, more of the high 4-yr totals occurred in the last half of the study period, primarily in the third quarter (1959–87).

Fig. 8.
Fig. 8.

Winter number of stations experiencing one of the top five 30-day heavy snowfall amounts, 1900/01–2015/16.

Citation: Journal of Applied Meteorology and Climatology 57, 2; 10.1175/JAMC-D-17-0217.1

Fig. 9.
Fig. 9.

Number of stations experiencing one of the top five 30-day heavy snowfall amounts in independent 4-yr periods (1901–04, 1905–08, …, 2013–16).

Citation: Journal of Applied Meteorology and Climatology 57, 2; 10.1175/JAMC-D-17-0217.1

The study examined how the frequency of occurrence varied among the states in the study region (Fig. 10). As noted earlier those states identified by the word “late” have a much greater percentage of stations with four or five of their top five 30-day heavy snowfall periods occurring from 1959 on (>25% more than that for the early period, prior to 1959). Those with the word “mix” are states with 60% or more stations with two or three of the top five 30-day heavy snowfall periods in the late period and, thus, two or three in the early period. States with more stations experiencing late (≥1959) 30-day heavy snowfall periods are located along the northern tier from the Great Plains eastward, including the Great Lakes, the Midwest, the Atlantic Coast and New England. Those with a mix are located across the southern tier of states. Only Georgia and Louisiana experienced more stations with “early” occurrence of heavy snowfall periods. These regional trends are similar to what Lawrimore et al. (2014) found when examining severe snowstorms in the eastern United States.

Fig. 10.
Fig. 10.

General frequency of top-five 30-day snowfall amounts by state through the 1900–2016 period. The word early indicates a greater frequency of heavy snowfall amounts early (before 1959), mix indicates a balanced occurrence of events through the study period, and late indicates a greater frequency of heavy snowfall amounts in the latter half of the period (in or after 1959).

Citation: Journal of Applied Meteorology and Climatology 57, 2; 10.1175/JAMC-D-17-0217.1

A final analysis examined those episodes when more than 10% of stations experienced one of their top five 30-day heavy snowfall amounts (Table 2). The episodes listed in Table 2 extend beyond 30 days because several stations’ 30-day period were somewhat different. Although the timing of the 30-day periods varied among stations (e.g., 11 February 1960–11 March 1960, 20 February 1960–20 March 1960), they overlapped with each other and were considered to be part of the same weather pattern. As the number of stations in these episodes count toward the winter total, there is little surprise that these periods occurred during the peak winters shown in Fig. 8. However, to better understand their spatial characteristics, the six episodes were mapped (Figs. 11a–f). All six episodes impacted some part of the Midwest; however, their patterns were distinct and in two of the episodes there were two separate heavy snowfall areas. The 1960 and the 1978 episodes were the only episodes when more than 20% of all stations (>101 stations) experienced a top-five 30-day heavy snowfall amount at one time. The 1960 episode impacted stations in a broad west-to-east swath from eastern Colorado into the Southeast and mid-Atlantic states, while the 1978 episode extended east-northeastward from northern Texas to the New England coast. These episodes demonstrate that large areas of the eastern United States can be impacted by one of these extended heavy snowfall periods. Composite 500-hPa geopotential height anomaly maps were constructed for these episodes (Figs. 12a–f) using data from the NOAA–CIRES Twentieth Century Reanalysis Project (Kalnay et al. 1996; Compo et al. 2011). These analyses indicate large 500-hPa geopotential negative height anomalies across the impacted regions, with the 1960 episode showing the greatest departure from the 1981–2010 average. Three of the six episodes—1960, 1977/78, and 1978/79—had extreme negative height anomalies, ranging from 80 to 140 m below average over some portion of the study area. These long-term height anomalies were likely associated with numerous extratropical cyclones and infusions of cold air. Awareness of these patterns could assist those forecasting extended unusual snowy patterns for parts of the eastern United States.

Table 2.

The six heavy snowfall episodes impacting the greatest number of stations.

Table 2.
Fig. 11.
Fig. 11.

Snowfall episodes for which 10% or more (>50) stations experienced one of their top five 30-day snowfall amounts: (a) 28 Jan–25 Mar 1912, (b) 20 Dec 1917–2 Feb 1918, (c) 7 Feb–28 Mar 1960, (d) 31 Dec 1977–24 Feb 1978, (e) 29 Dec 1978–25 Feb 1979, and (f) 17 Dec 1995–4 Feb 1996.

Citation: Journal of Applied Meteorology and Climatology 57, 2; 10.1175/JAMC-D-17-0217.1

Fig. 12.
Fig. 12.

Composite 500-hPa geopotential height (m) anomalies for the six episodes shown in Figs. 11a–f. Anomalies were calculated using the 1981–2010 average. Solid lines represent positive height anomalies, and dashed lines represent negative height anomalies.

Citation: Journal of Applied Meteorology and Climatology 57, 2; 10.1175/JAMC-D-17-0217.1

4. Conclusions

An assessment of 30-day heavy snowfall periods for the eastern United States provides valuable spatiotemporal information for climatologists and those impacted by these damaging and costly weather episodes, including government officials, emergency managers, and building designers (Changnon et al. 2008; Lawrimore et al. 2014). A new dataset that identified the greatest five 30-day heavy snowfall amounts, for 507 U.S. weather stations located east of the Rockies, was developed after an extensive snowfall data evaluation.

Similar to spatial patterns of snowstorms (Changnon et al. 2006) and winter seasonal snowfall totals (Kunkel et al. 2009), both the greatest 30-day snowfall amount and the average of the top five 30-day snowfall amounts exhibited a general increase northward from the Gulf of Mexico to the Canadian border (Figs. 3 and 4). Areas with anomalously higher values occurred to the lee of the Great Lakes, at higher elevations and those areas impacted by broader topographic features including the Rockies, Black Hills, and Appalachians. In most U.S. locations the greatest 30-day snowfall amount exceeded the annual snowfall average (Fig. 5), suggesting that regional impacts related to these unusual snowfall periods could be more extreme than what occurs when considering an individual snowstorm or the magnitude of snowfall in a winter (Moftakhari et al. 2017).

Temporal analysis of the number of U.S. stations experiencing one of its top five 30-day snowfall periods by winter identified that there was great interannual variability through the 117-yr period (Fig. 8). All analyses pointed toward an increased frequency in 30-day heavy snowfall periods beginning with the winter of 1959/60, with the highest occurrence in the third quarter (1959–87). Those states from the northern Great Plains eastward through the Great Lakes, Midwest, Atlantic coast, and Northeast experienced more 30-day heavy snowfall periods in the latter half of the study period, while most states across the southern tier of the country experienced these periods throughout the 117-yr period (Fig. 10). It should be noted that the uneven distribution of stations in the study region could have some influence on these results. These spatiotemporal findings are similar to those noted in Lawrimore et al. (2014) and may be related to broader, long-term variations in atmospheric characteristics such as winter temperatures, changes in the frequency or location of winter cyclone tracks, or the location or strength of mid-atmospheric height anomalies.

Six episodes (Table 2) were identified where more than 10% of the stations experienced one of their top-five 30-day heavy snowfall amounts. Two episodes occurred in 1910–20 while the other four episodes occurred during the latter half of the 117-yr period, with two in back-to-back winters in the late 1970s. These heavy snowfall episodes impacted large regions of the eastern United States. The February–March 1960 episode impacted 25% of the stations and occurred in a broad swath from the eastern plains of Colorado to the Atlantic coast (Fig. 11c). An examination of the 500-hPa geopotential height anomalies for these six episodes identified large areas of negative height anomalies over the United States, with the maximum negative departure from average east of the Rockies.

Further study of these episodes will examine other atmospheric components associated with these extreme 30-day heavy snowfall amounts. Potential analyses include examining the frequency and location of cyclone tracks, the location and strength of the upper-level jet stream, and the relationship of oceanic–atmospheric teleconnections to these events, as others have done (Lawrimore et al. 2014; Marinaro et al. 2015).

Acknowledgments

Lauren Haas and Dan O’Sullivan helped digitize the station data and identify station-related data issues. Amanda Carew of NIU’s Geovisual Mapping Lab and Mark Howland of NIU’s Geology Department created the maps for this paper. Victor Gensini developed the 500-hPa geopotential height anomaly maps. Suzy Changnon provided an editorial review. Last, I acknowledge my father Stan, who fostered my interest in atmospheric sciences during the severe Midwestern winters in the late 1970s.

REFERENCES

  • Branick, M. L., 1997: A climatology of significant winter-type weather events in the contiguous United States, 1982–94. Wea. Forecasting, 12, 193207, https://doi.org/10.1175/1520-0434(1997)012<0193:ACOSWT>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Changnon, S. A., 1967: A method of evaluating substation records of hail and thunder. Mon. Wea. Rev., 95, 209212, https://doi.org/10.1175/1520-0493(1967)095<0209:MOESRO>2.3.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Changnon, S. A., and T. Creech, 2003: Sources of data on freezing rain and resulting damages. J. Appl. Meteor., 42, 15141518, https://doi.org/10.1175/1520-0450(2003)042<1514:SODOFR>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Changnon, S. A., and D. Changnon, 2005: Damaging snowstorms in the U.S. Nat. Hazards, 31, 117.

  • Changnon, S. A., and K. E. Kunkel, 2006: Changes in instruments and sites affecting historical weather records: A case study. J. Atmos. Oceanic Technol., 23, 825828, https://doi.org/10.1175/JTECH1888.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Changnon, S. A., D. Changnon, and T. Karl, 2006: Temporal and spatial characteristics of snowstorms in the contiguous United States. J. Appl. Meteor. Climatol., 45, 11411155, https://doi.org/10.1175/JAM2395.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Changnon, S. A., D. Changnon, T. Karl, and T. Houston, 2008: Snowstorms across the nation: An atlas about storms and their damages. National Climatic Data Center, 96 pp.

  • Compo, G. P., and Coauthors, 2011: The Twentieth Century Reanalysis Project. Quart. J. Roy. Meteor. Soc., 137, 128, https://doi.org/10.1002/qj.776.

  • Doesken, N. J., and A. Judson, 1996: The snow booklet: A guide to the science, climatology, and measurement of snow in the United States. Dept. of Atmospheric Science, Colorado State University, 87 pp., http://ccc.atmos.colostate.edu/pdfs/snowbook.pdf.

  • Durre, I., M. P. 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. Meteor. Climatol., 52, 23772395, https://doi.org/10.1175/JAMC-D-13-051.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Eichenlaub, V. L., 1970: Lake-effect snowfall to the lee of the Great Lakes: Its role in Michigan. Bull. Amer. Meteor. Soc., 51, 403412, https://doi.org/10.1175/1520-0477(1970)051<0403:LESTTL>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Harrington, J. A., Jr., R. S. Cerveny, and K. F. Dewey, 1987: A climatology of mean monthly snowfall for conterminous United States: Temporal and spatial patterns. J. Climate Appl. Meteor., 26, 897912, https://doi.org/10.1175/1520-0450(1987)026<0897:ACOMMS>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Kalnay, E., and Coauthors, 1996: The NCEP/NCAR Reanalysis 40-Year Project. Bull. Amer. Meteor. Soc., 77, 437471, https://doi.org/10.1175/1520-0477(1996)077<0437:TNYRP>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Kocin, P. J., and L. W. Uccellini, 2005: Northeast Snowstorms. Vols. 1 and 2, Meteor. Monogr., No. 54, Amer. Meteor. Soc., 818 pp.

    • Crossref
    • Export Citation
  • Kunkel, K. E., 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. Oceanic Technol., 24, 6473, https://doi.org/10.1175/JTECH2017.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Kunkel, K. E., M. A. Palecki, L. Ensor, D. R. Easterling, K. G. Hubbard, D. A. Robinson, and K. T. Redmond, 2009: Trends in twentieth-century U.S. extreme snowfall seasons. J. Climate, 22, 62046216, https://doi.org/10.1175/2009JCLI2631.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Lawrimore, J., T. Karl, M. Squires, D. A. Robinson, and K. E. Kunkel, 2014: Trends and variability in severe snowstorms east of the Rocky Mountains. J. Hydrometeor., 15, 17621777, https://doi.org/10.1175/JHM-D-13-068.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Marinaro, A., S. Hilberg, D. Changnon, and J. R. Angel, 2015: The North Pacific–driven severe Midwest winter of 2013/14. J. Appl. Meteor. Climatol., 54, 21412151, https://doi.org/10.1175/JAMC-D-15-0084.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Menne, M. J., I. Durre, R. S. Vose, B. E. Gleason, and T. G. Houston, 2012: An overview of the Global Historical Climatology Network-Daily database. J. Atmos. Oceanic Technol., 29, 897910, https://doi.org/10.1175/JTECH-D-11-00103.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Moftakhari, H. R., A. AghaKouchak, B. F. Sanders, and R. A. Mathews, 2017: Cumulative hazard: The case of nuisance flooding. Earth’s Future, 5, 214223, https://doi.org/10.1002/2016EF000494.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Mook, C. P., and K. S. Norquest, 1956: The heavy snowstorm of March 18-19, 1956: The climax of a record late-season snow accumulation in southern New England. Mon. Wea. Rev., 84, 116125, https://doi.org/10.1175/1520-0493(1956)084<0116:THSOM>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • MRCC, 2017: Cli-MATE: MRCC Application Tools Environment. Midwestern Regional Climate Center, http://mrcc.isws.illinois.edu.

  • National Assessment Synthesis Team, 2001: Climate change impacts on the United States: The potential consequences of climate variability and change. National Assessment Synthesis Team Rep., 541 pp., https://data.globalchange.gov/assets/e9/97/436129058f2107f4925aeec13ed8/nca-2000-foundation-report.pdf

  • NOAA/NCEI, 2017: U.S. billion-dollar weather and climate disasters 1980-2017. NOAA/National Centers for Environmental Information, https://www.ncdc.noaa.gov/billions/events.pdf.

  • Norton, D. C., and S. J. Bolsenga, 1993: Spatiotemporal trends in lake effect and continental snowfall in the Laurentian Great Lakes, 1951–1980. J. Climate, 6, 19431956, https://doi.org/10.1175/1520-0442(1993)006<1943:STILEA>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Robinson, D. A., 1989: Construction of a United States historical snow data base. Proc. 45th Eastern Snow Conf., Lake Placid, NY, Eastern Snow Conference, 50–59.

  • Schwartz, R. M., and T. W. Schmidlin, 2002: Climatology of blizzards in the conterminous United States, 1959–2000. J. Climate, 15, 17651772, https://doi.org/10.1175/1520-0442(2002)015<1765:COBITC>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • U.S. Weather Bureau, 1922: Measurement of precipitation: Instructions. U.S. Dept. of Agriculture Circular E, 36 pp.

  • U.S. Weather Bureau, 1970: Instructions for Cooperative (Climatological) Observers. U.S. Department of Commerce Observing Handbook 2, 77 pp.

Save