An Updated Severe Hail and Tornado Climatology for Eastern Colorado

Samuel J. Childs Department of Atmospheric Science, Colorado State University, Fort Collins, Colorado

Search for other papers by Samuel J. Childs in
Current site
Google Scholar
PubMed
Close
and
Russ S. Schumacher Department of Atmospheric Science, Colorado State University, Fort Collins, Colorado

Search for other papers by Russ S. Schumacher in
Current site
Google Scholar
PubMed
Close
Full access

Abstract

A localized tornado and severe hail climatology is updated and enhanced for eastern Colorado. This region is one of the most active severe weather areas in the United States because of its location immediately east of the Rocky Mountains, intrusions of Gulf of Mexico moisture into a dry climate, and various small-scale topographically forced features such as the “Denver Cyclone.” Since the 1950s, both annual tornado and severe (≥1.0 in.; 1 in. = 25.4 mm) hail reports and days have been increasing across the area, but several nonmeteorological factors distort the record. Of note is a large population bias in the severe hail data, with reports aligned along major roadways and in cities, and several field projects contributing to an absence of (E)F0 tornado reports [on the (enhanced) Fujita scale] in the 1980s. In the more consistently observed period since 1997, tornado reports and days show a slight decreasing trend while severe hail reports and days show an increasing trend, although large variability exists on the county level. Eastern Colorado tornadoes are predominantly weak, rarely above (E)F1 intensity, and with a maximum just east of the northern urban corridor. Severe hail has a maximum along the foothills and shows a trend toward a larger ratio of significant (≥2.0 in.; ≥50.8 mm) hail to severe hail reports over time. Both tornadoes and severe hail have trended toward shorter seasons since 1997, mostly attributable to an earlier end to the season. By assessing current and historical trends from a more localized perspective, small-scale climatological features and local societal impacts are exposed—features that national climatologies can miss.

© 2019 American Meteorological Society. For information regarding reuse of this content and general copyright information, consult the AMS Copyright Policy (www.ametsoc.org/PUBSReuseLicenses).

Corresponding author: Samuel J. Childs, sjchilds@rams.colostate.edu

Abstract

A localized tornado and severe hail climatology is updated and enhanced for eastern Colorado. This region is one of the most active severe weather areas in the United States because of its location immediately east of the Rocky Mountains, intrusions of Gulf of Mexico moisture into a dry climate, and various small-scale topographically forced features such as the “Denver Cyclone.” Since the 1950s, both annual tornado and severe (≥1.0 in.; 1 in. = 25.4 mm) hail reports and days have been increasing across the area, but several nonmeteorological factors distort the record. Of note is a large population bias in the severe hail data, with reports aligned along major roadways and in cities, and several field projects contributing to an absence of (E)F0 tornado reports [on the (enhanced) Fujita scale] in the 1980s. In the more consistently observed period since 1997, tornado reports and days show a slight decreasing trend while severe hail reports and days show an increasing trend, although large variability exists on the county level. Eastern Colorado tornadoes are predominantly weak, rarely above (E)F1 intensity, and with a maximum just east of the northern urban corridor. Severe hail has a maximum along the foothills and shows a trend toward a larger ratio of significant (≥2.0 in.; ≥50.8 mm) hail to severe hail reports over time. Both tornadoes and severe hail have trended toward shorter seasons since 1997, mostly attributable to an earlier end to the season. By assessing current and historical trends from a more localized perspective, small-scale climatological features and local societal impacts are exposed—features that national climatologies can miss.

© 2019 American Meteorological Society. For information regarding reuse of this content and general copyright information, consult the AMS Copyright Policy (www.ametsoc.org/PUBSReuseLicenses).

Corresponding author: Samuel J. Childs, sjchilds@rams.colostate.edu

1. Introduction

Severe thunderstorms, particularly those that produce tornadoes and severe hail, pose a serious societal risk throughout much of the United States. An average of 1200 tornadoes occur across the United States each year, some of which cause human casualties and destruction (NSSL 2017). Hail is also an increasingly concerning hazard; as population and the built environment continue to expand, there are more structures and vehicles for large hail to damage. Indeed, losses due to hailstorms are increasing (Changnon 2009; Sander et al. 2013; Prein and Holland 2018), and even a single hail event can easily amass more than $1 billion worth of property damage. Human injuries can also occur, as the Cheyenne Mountain Zoo incident in Colorado Springs, Colorado, in August 2018 proved (Childs and Schumacher 2018). “Severe weather” is defined by the Storm Prediction Center (SPC) as an occurrence of any of the following: 1) a wind gust in excess of 58 mi h−1 (93 km h−1), 2) hail of at least 1-in. (1 in. = 2.54 cm) diameter, or 3) a tornado of any kind. Establishing a climatological framework is a helpful way to begin any severe weather research endeavor, as it provides a base on which to perform analysis and also gives a glimpse as to past and present characteristics of the hazard. Climatological studies that analyze frequency, strength, variability, and spatial patterns of tornadoes in a broad sense have existed for many decades (e.g., Kelly et al. 1978; Doswell and Burgess 1988; Brooks et al. 2003b; Verbout et al. 2006; Farney and Dixon 2015; Agee et al. 2016; Gensini and Brooks 2018). In addition, filtered climatologies have been established for subsets of tornadoes, such as violent tornadoes (Concannon et al. 2000), nocturnal tornadoes (Kis and Straka 2010), cold-season tornadoes (Childs et al. 2018), and tornado deaths (Ashley 2007; Agee and Taylor 2019). The influence of teleconnections on the variability of the tornado record has also been analyzed (Gensini and Marinaro 2016; Lee et al. 2016; Cook et al. 2017; Allen et al. 2018; Molina et al. 2018; Tippett 2018). Additionally, severe hail and wind climatological studies have become more frequent in recent years (e.g., Cintineo et al. 2012; Allen and Tippett 2015; Gensini and Allen 2018; Edwards et al. 2018).

The aforementioned climatological studies of severe weather, most of which cover a broad region of the United States, are advantageous for establishing long-term trends and providing a general idea of where and how severe weather hazards are distributed. Often lacking, however, are the impacts of small-scale features that can lead to unique climatological patterns over a much smaller area. Moreover, nonmeteorological biases inherent in the data records can be exposed more fully on smaller scales and must be acknowledged in analyses. Localized climatologies can also appeal to local residents in ways that national-level analyses cannot by specifically focusing on severe weather patterns where someone lives. The eastern Colorado domain considered here features both intricate topographical influences on convection and a stark population distribution, yielding small-scale climatological intricacies. A localized approach has also been applied to other areas across the United States (e.g., Oklahoma; Guo et al. 2016).

While severe weather can happen anywhere in the United States, the majority of severe events occur east of the Rocky Mountains. Here, the roles of topography, upper-level flow patterns, and moisture transport via the low-level jet into a relatively cool and dry continental air mass help provide vertical instability and other ingredients needed for convection (Doswell 1980; Brooks et al. 2003b). While some of the most extreme severe weather events have occurred well downstream of the Rocky Mountains, the area in its immediate lee across eastern Colorado has experienced several high-impact events in recent years (e.g., Schumacher et al. 2010; Hamill 2014; Gochis et al. 2015) and is a worthy domain across which to analyze severe weather events and their trends. For reference, Fig. 1 displays a state map of Colorado showing counties and select cities and roadways, many of which will be referenced throughout the following analysis and discussion. The influence of the Rocky Mountains on Colorado severe weather is undeniable and has been the subject of myriad field projects and research studies. The 1980s produced several seminal studies of severe weather environments for Colorado and the adjacent “High Plains” region. Whereas severe weather across much of the eastern United States is dependent upon high values of convective available potential energy (CAPE) and storm-relative vertical wind speed shear, eastern Colorado severe weather environments typically have lower CAPE and speed shear and are characterized by an anomalously moist boundary layer and very strong storm-relative directional wind shear (Doswell 1980; Maddox et al. 1981; Szoke et al. 1984). Two mesoscale features that aid in the formation of intense convection despite relatively low CAPE and shear were also discovered and formally defined during this period as a result of multiple field projects. First, the Denver convergence vorticity zone (DCVZ) is an area of converging winds that forms when two different air masses meet near metropolitan Denver, Colorado. Specifically, warm and moist flow from the southeast ascends the Palmer Divide topographical feature, an area of locally higher elevation between Colorado Springs and Denver, and converges with northwesterly flow moving downslope from the foothills to the north and west of Denver. This convergence leads to a localized area of vorticity, commonly known as the Denver Cyclone (Szoke et al. 1984). As low-level vorticity is a key ingredient for tornadogenesis, it is not surprising that the Denver Cyclone was observed on many severe weather days across the Front Range during the Joint Airport Weather Studies (JAWS) and Convective Initiation Downburst Experiment (CINDE) field campaigns in the 1980s (Szoke et al. 1984; Wilson et al. 1988; Brady and Szoke 1989), and continues to be a frequent feature of Front Range summertime severe weather today. The DCVZ also explains why many tornadoes that form across eastern Colorado are of the nonmesocyclonic variety (Brady and Szoke 1989; Wakimoto and Wilson 1989); that is, they form along surface boundaries rather than stemming from supercell storms with strong midlevel rotation.

Fig. 1.
Fig. 1.

State map of Colorado showing county names (uppercase), select cities (light-gray shading and boldface italic labels), interstate highways (thick blue), other primary roadways (thin blue), and Denver International Airport (black airplane symbol).

Citation: Journal of Applied Meteorology and Climatology 58, 10; 10.1175/JAMC-D-19-0098.1

Despite not being located in the traditional “Tornado Alley,” eastern Colorado is actually a hot spot for tornadoes. In fact, when tornadoes over the period 1950–2016 are considered, Weld County in northern Colorado ranks first of all U.S. counties for most tornado segments passing through it (Erdman 2017). These results must be taken in context, however, noting that Colorado sees a much higher proportion of weak tornadoes and landspouts compared to other parts of the country, and its tornadoes rarely result in human fatalities (Ashley 2007). Nevertheless, several studies have shown a localized maximum in tornado frequency across eastern Colorado (e.g., Brooks et al. 2003a; Ashley 2007; Allen et al. 2015; Farney and Dixon 2015). Similarly, eastern Colorado is also prone to large hail events, and some particularly damaging events have occurred in recent years. The costliest hailstorm in Colorado state history, and one of the costliest in U.S. history, impacted the Denver metropolitan area on 8 May 2017, causing an estimated $2.3 billion in damage at the time of its occurrence (RMIIA 2018). Other destructive hailstorms impacted the populated areas along the Front Range in 2018, including the 6 August 2018 event that resulted in the rare occurrence of at least a dozen human injuries and five animal deaths at the Cheyenne Mountain Zoo in Colorado Springs. In fact, 2018 broke multiple state records for the number and proportion of severe hail reports that were of the ≥2.0-in. (≥50.8 mm) and ≥3.0-in. (≥76.2 mm) varieties (Childs and Schumacher 2018), and Colorado even bested Texas for the most hail losses in 2018 (State Farm 2019). Then on 13 August 2019, the state record for largest hailstone was eclipsed when a 4.83-in. (122.7 mm) hailstone was measured in Kit Carson County. Aside from very large hailstones, the state of Colorado also sees its fair share of so-called plowable hailstorms, that is, those in which a large amount of usually small hailstones accumulate to the point of requiring snow plows to remove them (Schlatter and Doesken 2010; Kalina et al. 2016; Friedrich et al. 2019; Kumjian et al. 2019). As with tornadoes, a localized maximum in hail reports has been noted across eastern Colorado in previous work (Changnon 1999; Cintineo et al. 2012; Allen and Tippett 2015; Allen et al. 2015).

There have been several case study analyses of Colorado severe weather events, but few attempts have been made at developing robust climatologies for tornadoes and severe hail. The Colorado Climate Center houses general statistics about tornadoes, including an analysis by Spears (2017) that presents state- and county-level statistics of Colorado tornadoes, partitioned by intensity, for the period 1950–2012. However, the study does not account for biases in the data record, nor does it explore potential reasons for some of the observed trends. Former Colorado State Climatologist N. Doesken presented a unique cultural history of hailstorms in Colorado as well as general climatological statistics in his 1994 overview (Doesken 1994), but since then no comprehensive hail climatology has been published. The tornado and severe hail climatologies and associated statistics herein thus encapsulate a comprehensive analysis of these hazards for an eastern Colorado domain and serves as a foundation for future work that will explore projected changes in the tornado and severe hail landscape due to climate change and population growth.

The rest of this article is outlined as follows: section 2 will discuss the SPC Storm Data archives for tornadoes and severe hail and the methods taken to overcome the limitations and inhomogeneities in the data, largely through a synthesis of previous research. Sections 3 and 4 present the eastern Colorado tornado and severe hail climatologies, respectively, for the period 1997–2017, and the various associated trends in frequency and variability. Section 5 concludes with a summary of key points and advancement toward future work.

2. Data and methods

Any climatological study that makes use of the U.S. severe weather data records must acknowledge numerous deficiencies in the existing data and make carefully considered choices when determining the extent of data used. Despite the limitations, the SPC maintains the largest and most complete severe weather database in the world (Schaefer and Edwards 1999). Severe weather data are available in tabular format, which gives numeric values of, for example, tornado intensity rating, pathlength and width, and hail size, as well as geographical coordinates of the event location, loss amounts, and casualty statistics. In addition, each severe weather hazard comes in a shapefile for use in a graphical information system (GIS) framework. (The tabular files and a link to the shapefiles are publicly available online at https://www.spc.noaa.gov/wcm/.) This study makes use of the tornado and severe hail datasets, both of which contain noteworthy biases and inhomogeneities.

a. Tornadoes

The SPC tornado database currently contains tornado events from 1950 to the present. Over such a long period, many different changes to reporting methodology, rating systems, and definition changes have led to biases in the data record. While a thorough treatment of the issues has been given in previous work (Doswell and Burgess 1988; Schaefer and Edwards 1999; Verbout et al. 2006; Agee and Childs 2014), an overview of the key biases present is worthwhile for establishing the eastern Colorado tornado climatology.

While the middle of the twentieth century saw a building interest in severe weather and tornado forecasting, with numerous field campaigns and programs, it was not until the early 1950s and the establishment of the Severe Local Storms unit (SELS) that procedures for reporting, documenting, and assigning intensity ratings to tornadoes began to be standardized (Doswell 2007). As a result, almost all modern tornado analyses begin with either the year 1953 or 1954, and this study will also start with 1953. The next major milestone in tornado documentation came in 1974 with the introduction of the Fujita (F) scale for rating tornadoes. The F scale was applied to tornadoes that occurred before 1974 but again relied primarily upon newspaper reports and photographs. Not surprisingly, pre-1974 tornadoes were sometimes given incorrect F-scale ratings (Grazulis et al. 1993), and Agee and Childs (2014) exposed enhanced F2 counts at the expense of F1 counts in the period 1953–74; the enhancement does not stem from meteorological factors. A more prominent feature in the tornado data related to F-scale ratings is due to the advent of NEXRAD WSR-88D network, which was installed in weather forecast offices across the country over a period during the 1990s (Crum et al. 1998). The WSR-88Ds allowed meteorologists to more easily identify tornado vortex signatures on radar, which in turn led to greater observation of tornadoes, especially weak ones, by people on the ground (Bieringer and Ray 1996). Time series plots of F0 tornado reports show a huge jump in the mid-1990s, explainable only by the advent of Doppler radar (Verbout et al. 2006; Agee and Childs 2014). It therefore behooves the researcher to use caution when performing analysis with F0 tornadoes prior to the late 1990s. Crum et al. (1998) reported that the last WSR-88D was installed in 1997, and all with coverage in the eastern Colorado domain were installed by early 1996; therefore a start year of 1997 has been selected for this study. Yet another change was instituted in 2007, as the Fujita scale was upgraded to the enhanced Fujita (EF) scale, which introduced a tornado damage rating scale based on numerous damage indicators (DI) and degrees of damage (DoD; McDonald and Mehta 2006). The switch to the EF scale has produced some subtle changes in national tornado trends but more significantly has revolutionized tornado storm surveys and damage analysis (Edwards and Brooks 2010; Edwards et al. 2013). Further amendments to the EF scale are forthcoming, including modification of DIs and a greater emphasis on wind speed as a rating metric (LaDue et al. 2018). An increasing trend exists in all [i.e., (E)F0–(E)F5] tornado reports across the contiguous United States (CONUS) simply as a result of increases in population (Potvin et al. 2019) as well as public interest and technological advancements. However, when (E)F0 tornadoes are excluded from analysis, there is in fact no appreciable trend in (E)F1+ tornadoes, at least on the national level (Verbout et al. 2006; Gensini and Brooks 2018). Because of the high fraction of (E)F0 tornadoes in the Colorado record, they are included in the analysis presented herein; however, the time frame of the climatology is restricted to 1997–2017 to provide a more accurate representation of trends.

b. Hail

The current SPC hail data record extends from 1955 to the present day, nearly as long as the tornado record. In perhaps the most thorough synthesis of the characteristics of severe hail data, Allen and Tippett (2015) note that advancements in technology, procedures, and awareness have led to a nonmeteorological increase in severe hail reports over the years. Particularly in the 1980s and 1990s, the formation of the SPC, growth of the storm spotter network, and advent of Doppler radar led to increased hail reporting. In addition, the Community Collaborative Rain, Hail, and Snow Network (CoCoRaHS; Reges et al. 2016) was formed in the late 1990s and has since become an additional source of hail reporting through public outreach. Additional efforts such as the Severe Hazards Analysis and Verification Experiment (SHAVE; Ortega et al. 2009) and Colorado Hail Accumulation from Thunderstorms (CHAT; Friedrich et al. 2019) projects have also increased interest in and reporting of hail. Given this increasing interest, population growth, and technological advancements, severe hail data prior to the late-1990s are suspect and likely suffer from many more missing reports in comparison with the more recent decades. As such, this study will follow Allen and Tippett (2015) by focusing on severe hail reports since 1997. A change in the threshold of severe hail was made in 2010, with an upward shift from 0.75- to 1.0-in. hail diameter (NCEI 2009). This minor change has led to a nontrivial decrease in 0.75-in. hail reports since 2010 (Allen and Tippett 2015), and thus the Colorado severe hail climatology presented here will consider only those hail reports that meet the current severe threshold of 1.0 in. (25.4 mm).

There is also an overwhelming tendency for hailstone sizes to be reported on the basis of the sizes of reference objects used by the NWS in forecasts and severe thunderstorm warning texts (Allen et al. 2017). For example, hailstone size reports are clustered around 1.0 in. (25.4 mm; quarter sized), 1.75 in. (44.5 mm; golf ball sized), and 2.75 in. (69.9 mm; baseball sized). There is no reason to believe that hailstones preferentially hit the surface at these sizes, but the data record reveals very few reports of other equally likely sizes such as 1.1 in. (27.9 mm) or 2.25 in. (57.2 mm). Blair et al. (2017) found that hail sizes recorded in the SPC database are actually underestimates of actual hail size. This can be due to the shattering of hail upon impact, melting that occurs between impact and measurement, and the tendency for the largest hailstones in a storm to fall in a narrow swath (Changnon 1977; Blair et al. 2017). Further, while instructions for measuring hailstone size are available to trained spotters and the public, there is still human subjectivity and error in taking hail measurements.

Arguably the greatest issue with the severe hail database is population bias, namely, the clustering of severe hail reports near population centers and roadways, as shown by Allen and Tippett (2015). Population bias was a particular concern in the early decades of severe hail reporting given the lack of standardized reporting procedures and the absence of Doppler radar. However, even today severe hail is more likely to be reported where people are around to measure them and underreported in more rural areas. This is in contrast to tornadoes in rural areas: although given a weak intensity rating because of a lack of appreciable damage, tornadoes are easier to see and can be confirmed with storm surveys, but confirming that hailstones actually fell and of what size in rural areas is very difficult. Population bias has a unique influence on the eastern Colorado hail distribution (as will be shown in section 4), which has a stark population contrast between the urban Front Range corridor and the rural eastern plains. It is also important to note that, in addition to severe hail being reported more readily in populated areas, property risk from severe hail is also enhanced in areas where people live, which is a topic to be explored in future work.

c. Summary and domain

According to the SPC severe weather data archives, 2050 tornadoes were reported in the state of Colorado between 1953 and 2017, for an average of 31 tornadoes per year. Approximately 96% of these tornadoes occurred in the eastern half of the state, which motivates a western border of the domain to be set at 105.3°W longitude, roughly coinciding with the eastern edge of the foothills. This is not to diminish significance or societal impacts of tornado and hail events that occur in mountainous terrain and western Colorado, but these events are too anomalous, are likely underreported, and affect far fewer people to justify inclusion in this study. The eastern half of Colorado, however, presents two very different societal regimes, with the developed business and technology hubs along the urban corridor abutting the foothills and the traditional agriculture and ranching communities farther east. Moreover, the selected domain lies in the lee of the Rocky Mountains, which have unique synoptic and mesoscale effects on severe weather. The tornado climatology presented in section 3 will investigate all (E)F-scale intensity categories, and in section 4 “severe hail” will refer to hailstones documented as at least 1.0 in. (25.4 mm) in diameter.

3. Eastern Colorado tornado climatology

Over the full record (1953–2017), an increasing trend is noted for both (E)F0 and (E)F0+ tornado count time series (Fig. 2a), and these two time series are nearly identical in the more recent period when (E)F0 counts are more reliable (Fig. 2b). Statistical significance of both tornado and severe hail trends in the following two sections is measured in two ways: first, a Student’s t test determines if the slope of the linear regression line is statistically different from zero, and second, a Mann–Kendall nonparametric test (Mann 1945; Kendall 1975) assesses whether a monotonic trend exists in the variable of interest over time. The Mann–Kendall test is especially useful in confirming linear trend analysis for time series with outliers or large variability. As expected, most tornadoes across eastern Colorado of the (E)F0 variety that are weak and often nonmesocyclonic. In Colorado, (E)F0 tornadoes account for 85% of total tornadoes in the more recent period (Table 1). In fact, 96% of tornadoes in the more recent period across eastern Colorado were of either (E)F0 or (E)F1 intensity, with only 3% of tornadoes rated as “significant” [i.e., (E)F2+]. Even when considering tornado data back to 1953, only 6% of eastern Colorado tornadoes have been significant, and only one tornado in the modern record has been rated above an (E)F3 (an F4 in Baca County on 18 May 1977 that actually began its track in Oklahoma). This is not to say that the domain has been spared tornado destruction, but most of the tornadoes across the eastern plains are weak and/or are given low ratings because of the lack of structures to damage. Tornadoes impacting major metropolitan areas along the Front Range are a rare occurrence, although an EF3 tornado that tore a path through more populated areas between the cities of Greeley and Fort Collins, Colorado, resulted in dozens of injuries and one fatality in the town of Windsor, Colorado (NWS Boulder 2008). Aside from the Windsor tornado of 2008, the other five (E)F3 tornadoes in the domain since 1997 occurred in 1999, 2000, 2001, 2007, and 2015. The most catastrophic of these (E)F3 tornadoes was the 28 March 2007 tornado that produced major damage in the town of Holly in Prowers County, Colorado, killing two people and injuring several others (Aguilera 2007). The only other killer tornado in the modern data record occurred in 1960 near Holyoke, Colorado. This tornado, which killed two people, was also posteriorly given an EF3 rating. Since 1950, no tornado rated weaker than (E)F3 has resulted in fatalities, which is consistent with national statistics that show almost all killer tornadoes to be rated (E)F2 or greater (Ashley 2007). However, prior to 1950, when warning methods were much more rudimentary or were nonexistent, Grazulis (1993) reports 11 killer tornadoes in Colorado, most of which occurred in rural areas. The highest fatality count from a single tornado was recorded on 10 August 1924 near the town of Thurman, Colorado, when nine children and one adult woman died when a tornado ripped through their farmhouse.

Fig. 2.
Fig. 2.

Time series of eastern Colorado tornado counts for (a) 1953–2017 and (b) 1997–2017, partitioned by (E)F-scale intensity. Linear trends are shown as dashed lines.

Citation: Journal of Applied Meteorology and Climatology 58, 10; 10.1175/JAMC-D-19-0098.1

Table 1.

Eastern Colorado tornadoes partitioned by (E)F-scale rating.

Table 1.

The next thing to note about eastern Colorado tornadoes is the behavior of the (E)F0 and (E)F1 time series in the 1980s. This entire decade shows an absence of (E)F0 tornado reports, with very few or zero (E)F0 tornadoes reported each year. This is coupled by an associated spike in (E)F1 tornadoes. At the dawn of the 1990s, (E)F0 tornado counts suddenly jump upward to levels not seen in the prior decades, while (E)F1 tornado counts fall back to a level consistent with the 1970s. The major uptick in (E)F0 tornadoes in the 1990s can be explained in part by the advent of Doppler radar, but there is no reason to believe why (E)F0 tornadoes should disappear entirely from the data record for several years in the 1980s. From a meteorological perspective, there were at least three tornado outbreaks during this decade: a 3 June 1981 outbreak that produced seven (E)F1+ tornadoes, including two (E)F2 tornadoes in the Denver metropolitan area that caused (at that time) $15 million in damage (Szoke et al. 1984); a 15 June 1988 outbreak that included a damaging (E)F3 tornado in southern Denver (Szoke et al. 2006); and 11 tornadoes over the course of 7–9 July 1988, of which 9 were given an (E)F1 rating. These outbreaks may help to explain the jump in (E)F1 tornado counts, but even in tornado outbreaks one would expect at least some tornadoes to be assigned an (E)F0 rating. The population of the Front Range urban corridor increased in the 1980s as well, but expansion had been ongoing before this time. Field campaigns, including JAWS in 1982 and CINDE in 1987, and at least two NOAA exercises in which “forecasters issu[ed] warnings and chase crews tried to verify” (E. Szoke 2017, personal communication), occurred in the 1980s yet do not explain why (E)F1 tornadoes would be necessarily favored over (E)F0 tornadoes. Reporting standards and local office overreporting are other potential influences, but in any case, the (E)F0 counts rebounded and increased substantially in the 1990s.

A closer look at Fig. 2b reveals that for the more recent period of 1997–2017, there is a slight decreasing trend in all tornado reports, influenced by the concurrent slight decrease in (E)F0 tornado reports. There is little to no appreciable trend for the (E)F1 and (E)F2+ records over this time (slopes of linear fit line = −0.19 and 0.02, respectively), which is consistent with recent research on a national level (Verbout et al. 2006; Brooks et al. 2014; Coleman and Dixon 2014; Clark 2017). However, while overall changes in U.S. tornado counts are small, there have been reported changes in the annual count variability (Brooks et al. 2014), annual count volatility (Tippett 2014), and efficiency of tornado days, that is, how many tornadoes occur on a day in which at least one tornado occurs (Elsner et al. 2015). For the local eastern Colorado domain, Fig. 3 shows the number of tornado days over the two periods of interest. The resulting linear trends are very similar to those of tornado counts, with statistically significant increasing trends in (E)F0 and (E)F0+ tornado days when the entire record is considered, but slight decreasing trends in all but (E)F2+ tornado days in the more recent period. This further validates that reports of tornadoes across eastern Colorado since 1997 have not become more numerous despite the increasing population. Tippett (2014) defines volatility of an annual report time series as the standard deviation Di of a differenced count, where Di is the difference between the number of reports in year i and the number of reports in year i − 1, that is Di = NiNi−1. This method transforms a supposed nonstationary dataset into one that has stationary statistics, and thus the volatility “measures the expected range of the change in reports from one year to the next (Tippett 2014).” Tippett (2014) found that the volatility in the F0+ and F1+ tornado count time series both more than doubled from the period 1980–2004 to 2005–13, with values in the more recent period of 439.7 for F0+ and 268.8 for F1+, respectively. This method is employed for both tornadoes and severe hail in the Colorado domain for two recent decades (1998–2007 and 2008–17), with results presented in Table 2. In contrast to the national-level findings of Tippett (2014), the Colorado tornado count time series show a slight decrease in volatility in the 2008–17 decade across all (E)F-scale intensity bins. The magnitude of the volatility is much lower in Colorado than for the entire United States because of the smaller sample size and use of more recent time periods; therefore, it is more useful to examine the relative changes in volatility between these two decades rather than focus on actual values. For example, Tippett (2014) notes reporting practices in earlier years can influence volatility, but changes in the 2000s seem to be linked to changes in the environment, which may be hard to capture on such a local scale. Colorado tornado days show only a slight increase in volatility in the more recent period. The difference between the means of the two decadal time series of differenced counts Di were tested for statistical significance using the nonparametric Wilcoxon signed-rank test (Wilcoxon 1945) and showed no significant difference. In other words, the variability of tornado counts and days across eastern Colorado over the two most recent decades is very small throughout all size bins, in contrast to what Brooks et al. (2014) found over a much larger domain. Testing the sensitivity of the two selected periods did not yield any appreciable difference in volatility.

Fig. 3.
Fig. 3.

As in Fig. 2, but for tornado days, i.e., the number of days per year on which at least one tornado of given intensity was reported.

Citation: Journal of Applied Meteorology and Climatology 58, 10; 10.1175/JAMC-D-19-0098.1

Table 2.

Tornado and hail volatility, computed by taking the standard deviation of the differenced tornado and hail count time series, for two recent decades.

Table 2.

Tornadoes have occurred throughout most of the eastern Colorado domain (Fig. 4), but a preference is clearly seen abutting the foothills along the north-central urban corridor. Many tornado reports are also scattered throughout the eastern plains, which has much less population. To assess the level of population bias in the tornado data, Fig. 5 presents plots of starting latitude–longitude points for all tornadoes across the domain since 1953 (Fig. 5a), as well as a smoothed version (Fig. 5b) in which the tornado data has been upscaled to a larger grid with cells of 0.25° latitude by 0.275° longitude. This upscaling process is often done with SPC tornado data to account for erroneous reporting of starting latitude–longitude locations of tornado tracks (e.g., Brooks et al. 2003a; Hoogewind et al. 2017). Last, to account for the inhomogeneities in the data record described in section 2, similar plots of points and smoothed grids of tornadoes for the more recent period of 1997–2017 are presented in Figs. 5c and 5d. The same general pattern is noted among both periods, with a maximum in tornado reports near Denver International Airport (DIA). The northern half of the domain is climatologically more tornadic than the southern half, aside from a relative maximum near the town of Lamar at the intersection of Prowers, Bent, and Kiowa Counties in southeastern Colorado. A local bulls-eye of reports is also found just north of the Palmer Divide in southwestern Elbert County, which likely corresponds to the many weak tornadoes that form along the DCVZ. The domain space south and west of Pueblo, Colorado, where relatively few people live, sees the least tornado reports. Given the relatively small sample size, tornado outbreaks, in which multiple tornadoes are documented in the same general location on the same day, may be a source of bias to the spatial distribution, particularly east of the urban corridor.

Fig. 4.
Fig. 4.

All tornado tracks for the eastern Colorado domain for the period 1953–2017, partitioned by (E)F-scale intensity rating (the data are taken from https://www.spc.noaa.gov/svrgis/).

Citation: Journal of Applied Meteorology and Climatology 58, 10; 10.1175/JAMC-D-19-0098.1

Fig. 5.
Fig. 5.

(left) Starting latitude–longitude tornado points and (right) gridded annual tornado reports for the periods (a),(b) 1953–2017 and (c),(d) 1997–2017. The gridded reports are smoothed from a 0.25° latitude × 0.275° longitude grid.

Citation: Journal of Applied Meteorology and Climatology 58, 10; 10.1175/JAMC-D-19-0098.1

Figure 6 further highlights the spatial variability that exists in tornado reporting by examining four county-level tornado trends. The four counties include Weld County in north-central Colorado, which has seen the most tornado segments (265) of any county in the United States since 1950; Adams County, a growing county that houses the northern Denver suburbs and many rural areas and that has seen the second-most tornado segments in the state (172) in the modern record; Elbert County, a rural county southeast of Denver that has the greatest increasing trend in tornado counts since 1997; and Logan County, another rural county east of Weld County that has the greatest decreasing trend in tornado counts since 1997. The two familiar time periods are represented, namely, 1953–2017 and 1997–2017. Clearly, counties that have urban components (e.g., Adams) and those that are predominantly rural (e.g., Elbert) can both have increasing trends in tornado reports. Interestingly, although it is the top tornado county in the United States, Weld County shows a slightly decreasing trend in tornado counts and no trend in tornado days since 1997. Nearby Logan County has a sharp decreasing trend in tornado counts and days over this period, with several years of zero or one tornado report in the most recent decade. While there are too few reports to assess statistical significance in the respective trends, Fig. 6 at least reveals the intricacies of localized tornado trends as opposed to a composited trend over a much larger domain.

Fig. 6.
Fig. 6.

Time series of tornado reports (solid dots) and tornado days (open triangles) for the periods 1953–2017 and 1997–2017 for four eastern Colorado counties. Linear trends for tornado reports or days are respectively shown as solid or dashed lines.

Citation: Journal of Applied Meteorology and Climatology 58, 10; 10.1175/JAMC-D-19-0098.1

The NCEI Storm Events Database (available at https://www.ncdc.noaa.gov/stormevents/), provides the severe weather reports that are then archived in SPC’s Storm Data but also contains information on the source of severe weather reports since 1998. It is therefore intriguing to assess the distribution of tornado and hail reports among various sources in the eastern Colorado domain. More than 30 different storm report sources exist in the data records, which can be sorted into more general categories, as was the approach taken by Allen and Tippett (2015) in their assessment of sources of severe hail reports over the CONUS. Table 3 shows the sources of eastern Colorado tornado reports for the period 1998–2017. Unsurprisingly, the majority of tornadoes are reported by trained spotters who are in the area where the tornado occurs. The second highest category is storm chasers, who account for 21% of tornado reports. The lack of obstructions to sight also allows tornadoes to be observed and reported much more easily across eastern Colorado relative to cities and forested areas. Thus, despite the low number of residents across eastern Colorado (which leads to only 6.4% tornado reports contributed from the general public), the population here effectively booms during days in which there is a tornado risk, as storm chasers flock to the area and are able to document tornadoes that occur. As noted by an anonymous reviewer, the effect of storm chasers may also be influencing the relative maximum of tornado reports near the town of Lamar in southeastern Colorado (Figs. 5c,d), because this town, being the largest for many miles, caters well to storm chasers who descend on the area during a threat of severe weather. Note that storm chasers can be documented in other categories as well, such as public and trained spotters (Allen and Tippett 2015), and therefore the actual percent contribution of storm chasers may be higher. Other nontrivial sources of tornado reports across eastern Colorado include law enforcement, NWS meteorologists, and emergency managers. It is interesting to note that, despite the rising popularity of promoting weather on social media in recent years, only one tornado report in the database is attributed to social media. This tornado, which occurred near the Flagler Airport in July 2016, was determined to be an EF0 landspout tornado from a picture posted on Twitter that was obtained by the NWS. This is not to say that social media does not play a role in tornado reporting, and reference is even made to a social media post in reports attributed to other sources. Further, if a social media post is seen by a spotter or meteorologist before it is received at a local weather forecast office, it may get cataloged as coming from a different source, or alternatively it may get cataloged as “general public” rather than “social media” (R. Cox 2018, personal communication). Nevertheless, it is safe to say that most tornado reports across eastern Colorado are gathered from trained spotters and storm chasers. As will be seen in section 4, the ranking of sources contributing to hail reports is not the same.

Table 3.

Sources of eastern Colorado tornado reports for the period 1998–2017. General categories are presented in the first column, and their included subcategories are given in the second column.

Table 3.

As a final piece of the eastern Colorado tornado climatology, analysis is presented on the length of the tornado season, that is, the number of days in a calendar year between first and last reported tornadoes. Table 4 lists these statistics for tornadoes (top section) and severe hail (bottom section, to be discussed in section 4) for the period 1997–2018 and also the two halves of that period separately (1997–2007 and 2008–18). The year 2018 has been included in this analysis to provide two periods of equal length and to show a continuation of the overall trend. Figure 7a shows a time series (with linear regression line) of the length of season, and Fig. 7b shows time series (with linear regression lines) of the earliest and latest annual tornado reports for the period 1997–2018. It is seen that the (E)F0+ (i.e., all tornadoes) season is on average trending shorter in the past 20 years, due almost entirely to an earlier end to the season of roughly 15 days when comparing the trend line value in 1997 to that in 2018. In fact, the most recent tornado season at the time of writing (2018) was only 79 days long, which was the second shortest (E)F0+ season in the modern record and the shortest since 2001. A similar shortening is noted when only (E)F1+ tornadoes are considered, again due mostly to an earlier end to the season. The average start of the (E)F1+ tornado season in the 2008–18 period, however, is earlier than that of the 1997–2007 period by about 12 days. Note that there were no (E)F1+ tornadoes in the domain in 2011, and 2018 (E)F1+ values are not included because the tornado reports were preliminary at the time of writing. Student’s t tests and nonparametric Mann–Kendall tests were performed on these time series, but none of the trends were found to be significant at the 95% confidence level. While various reasons could be surmised as to the reason for the patterns seen in the length of the eastern Colorado tornado season, this will be left as future work.

Table 4.

Length of the eastern Colorado tornado and hail seasons. Tornado seasons are categorized by (E)F0+ and (E)F1+ tornado reports; hail seasons are categorized according to severe (≥1.0 in.) and significant (≥2.0 in.) hail reports. Three different periods are tabulated, as well as the average dates of the first and last respective reports during each period.

Table 4.
Fig. 7.
Fig. 7.

(a) Time series of length of tornado season for the eastern Colorado domain for the period 1997–2018. All tornadoes [i.e., (E)F0+] are shown in green, and (E)F1+ tornadoes are shown in blue, along with their respective linear trend lines. (b) Time series of the dates of the first (triangles) and last (circles) (E)F0+ (filled orange) and (E)F1+ (open blue) tornado reports, along with respective linear trend lines, for the same period and domain.

Citation: Journal of Applied Meteorology and Climatology 58, 10; 10.1175/JAMC-D-19-0098.1

To summarize, eastern Colorado has consistently experienced tornadoes since the modern data record began. Most of these tornadoes are weak, owing to both the local meteorological environment and the lack of structures to damage over the eastern plains, but significant and deadly (E)F3 tornadoes have occurred in the state’s history. Tornado counts and days trends are generally slightly downward since 1997, though not statistically significant, show little change in volatility, and have large variations between counties. The other major severe weather hazard in Colorado has substantially different climatological patterns.

4. Eastern Colorado severe hail climatology

Severe hail reports are the most frequent among the severe weather hazards in eastern Colorado. Figure 8 shows the time series of severe hail counts and days for the two periods 1955–2017 and 1997–2017, partitioned by hailstone diameter, where a severe hail day is defined as a day in which at least one hailstone that exceeds the particular size threshold is reported. The linear trends are also shown in Fig. 8 and reveal that severe hail reports have been increasing across the domain over time, independent of size threshold. In fact, each of the size threshold trends for the period 1955–2017, as well as the 1.0-in. (25.4 mm) time series for the period 1997–2017, are statistically significant at the 95% confidence level, according to both Student’s t tests and Mann–Kendall tests. As with tornadoes, this upward trend is influenced by nonmeteorological factors, such as the increasing public awareness and interest in measuring hail, better documentation practices, and increasing population across the Front Range. While not as pronounced as in the tornado time series, a jump in hail reports is seen in the mid-1990s as a result of the expanding Doppler radar network. A greater inhomogeneity is seen after 2010, when the SPC threshold for severe hail was raised from 0.75 in. (19.1 mm) to 1.0 in. (25.4 mm). The effect of this change is clearly seen in Fig. 8b, as almost all ≥0.75-in. (≥19.1 mm) hailstones reported since 2010 are also of the ≥1.0-in. (≥25.4 mm) variety. That is not to say that 0.75-in. (19.1 mm) hail is suddenly disappearing, but as reported by Allen and Tippett (2015), people are now more likely to report a hailstone as 1.0 in. (25.4 mm) to align with the SPC severe threshold. The upward trends also hold across all size thresholds for severe hail days (Fig. 9), and all of the 1955–2017 trends (Fig. 9a) are statistically significant at the 95% confidence level according to both Student’s t tests and Mann–Kendall tests. In the more recent period (Fig. 9b), the numbers of ≥1.0-in. (≥25.4 mm), ≥2.0-in. (≥50.8 mm), and ≥3.0-in. (≥76.2 mm) severe hail days across eastern Colorado are also increasing, although none of the trends are significant at the 95% confidence level. Nevertheless, the upward trend in ≥1.0-in. (≥25.4 mm) hail days is notable because the national record shows no trend in severe hail day thresholds since 1997 (Allen and Tippett 2015). In 2018, which is not represented in the trends in Figs. 8 and 9, a record of ten ≥3.0-in. (≥76.2 mm) hailstones was reported over seven different days, adding to the recent trends of increasing hail reports and hail days of these larger sizes.

Fig. 8.
Fig. 8.

Trends in eastern Colorado severe hail reports for the periods (a) 1955–2017 and (b) 1997–2017, partitioned by hailstone diameter. Linear trends are shown as dashed lines.

Citation: Journal of Applied Meteorology and Climatology 58, 10; 10.1175/JAMC-D-19-0098.1

Fig. 9.
Fig. 9.

As in Fig. 8, but for severe hail days; that is, days in which at least one hailstone of the given diameter was reported.

Citation: Journal of Applied Meteorology and Climatology 58, 10; 10.1175/JAMC-D-19-0098.1

Another metric by which to assess hail reports and days is the fraction of severe hail that is either significant (≥2.0 in.; ≥50.8 mm) or in excess of 3.0 in. (76.2 mm). Figure 10 shows these ≥2.0-in. (≥50.8 mm) and ≥3.0-in. (≥76.2 mm) fractions as percent contributions for reports (Fig. 10a) and days (Fig. 10b) over the period 1997–2018, along with linear regression lines. The percent contribution of ≥2.0-in. (≥50.8 mm) reports and ≥3.0-in. (≥76.2 mm) days shows increasing trends over time, and the percent contribution of ≥3.0-in. (≥76.2 mm) reports has no trend. The percent contribution of ≥2.0-in. (≥50.8 mm) days has a slight decreasing trend, but the 43% value in 2002, which contributes to the downward slope, came in the year of least number of severe hail days over this period (28), and the twelve ≥2.0-in. (≥50.8 mm) hail days in 2002 was not abnormally high. Large variability is present in these metrics, which precludes any statistical significance in the trends, although it is interesting to note that the highest values of every fractional metric except ≥2.0-in. hail days were seen in 2018. In fact, 20.1% of all severe hail reports in 2018 were at least 2.0 in. (50.8 mm), and 3.0% were at least 3.0 in. (76.2 mm), both of which either tied or eclipsed state records for the period since 1997. In addition, 27% of all severe hail days in 2018 included at least one ≥2.0-in. (≥50.8 mm) report, and a new state record of 14% of severe hail days included at least one ≥3.0-in. (≥76.2 mm) report (Childs and Schumacher 2018). Increasing human exposure to hailstorms, particularly along the Front Range, could be contributing to increasing fractional percentages over time, although approximately one-half of the sixty-six ≥2.0-in. (≥50.8 mm) hail reports in 2018 were located in rural areas.

Fig. 10.
Fig. 10.

Time series of percent contributions of ≥2- and ≥3-in. eastern Colorado hail (a) reports and (b) days for the period 2009–18. Linear trends are shown as dashed lines.

Citation: Journal of Applied Meteorology and Climatology 58, 10; 10.1175/JAMC-D-19-0098.1

Severe hail volatility is computed in the same way as tornado volatility in section 3, with results presented in Table 2 comparing two recent decades. Unlike tornadoes, ≥1.0-in. (≥25.4 mm) and ≥2.0-in. (≥50.8 mm) severe hail report volatilities show a sizeable increase in the 2008–17 decade. Increasing volatility is also noted for ≥1.0-in. (≥25.4 mm) and ≥2.0-in. (≥50.8 mm) severe hail days. While the ≥3.0-in. (≥76.2 mm) category shows muted volatility changes, there are very few reports of this diameter to begin with. Increasing volatility does not say much about changes in numbers of hail reports and days, but rather indicates a greater expected range of the annual number of hail report counts and days in the 2008–17 decade relative to the 1998–2007 decade, albeit the variability of the record precludes any statistical significance in the year-to-year differences for each period.

As already emphasized, hail reporting is subject to population bias, and this is no different for the eastern Colorado domain. In the same format as Fig. 5 for tornadoes, Fig. 11 presents the spatial distribution of severe hail reports, sorted by diameter, for the two periods of interest [1955–2017 (Figs. 11a,b) and 1997–2017 (Figs. 11c,d)]. The point coordinates of the hail reports are shown in Figs. 11a and 11c, and smoothed versions are shown in Figs. 11b and 11d. It is immediately apparent via inspection of Figs. 11a and 11c that population bias is contributing to hail reporting. For example, cities along the Front Range such as Fort Collins, Denver, Boulder, Colorado Springs, and Pueblo can be overlain atop hail report clusters. Across the eastern plains, hail reports form more linear patterns corresponding to interstate highways and other major roadways. With relatively few paved roads across eastern Colorado, hail is more likely to be reported along major thoroughfares on which people drive. The smoothed plots in Figs. 11b and 11d are similar and affirm the higher number of hail reports along the Front Range cities, with local maxima in both periods near Denver, Colorado Springs, and Pueblo. As was the case for tornado reports, the least amount of hail reports is found in extreme southern Colorado, where very few people live; however, the relative maximum in tornado counts near Lamar in southeastern Colorado does not show up in severe hail reports, potentially reflecting the storm chasing preference for tornadoes in this area. The greatest concentration of hail reports is also shifted a bit farther west than that of tornadoes, which has its maximum east of downtown Denver. Given the high population bias leading to the concentration of severe hail reports along the Front Range urban corridor, there are likely many severe hailstones that are going unreported because they are falling in places that are sparsely populated. Unlike tornadoes, which can be rated posteriorly by matching damage indicators to a wind speed range, accurately measuring and reporting hailstones before they melt requires a physical presence, which is often not possible because of either a lack of people in the vicinity or dangerous weather conditions that prevent people from venturing outside to measure hailstones. Despite the nonmeteorological influences on severe hail reports across eastern Colorado, it is interesting to note that the maximum in Colorado lightning activity from 1996 to 2016 situated along and near the Palmer Divide (Hodanish et al. 2019) coincides with a local maximum in severe hail reports just north of Colorado Springs over roughly the same time period (Fig. 11d).

Fig. 11.
Fig. 11.

(left) Latitude–longitude severe hail reports and (right) gridded annual severe hail reports for the periods (a),(b) 1955–2017 and (c),(d) 1997–2017. The gridded reports are smoothed from a 0.25° latitude × 0.275° longitude grid.

Citation: Journal of Applied Meteorology and Climatology 58, 10; 10.1175/JAMC-D-19-0098.1

The influence of population dynamics on the eastern Colorado hail time series can be seen at the county level by computing correlations between county population trends since 2000 and county hail report and days trends since 1997 for 26 eastern Colorado counties within the study domain. County populations are taken from the U.S. Census Bureau and have been normalized by county land area. Pre-2000 county population is only available in decadal increments, but from 2000 onward annual values are available and therefore used here. The correlations between county population trend and county hail reports and days trends are 0.425 and 0.441, respectively, both of which are statistically significant. That is, counties in which population has been growing since 2000 also generally have a growing number of annual severe hail reports and severe hail days. Figure 12 displays this result graphically, with each county’s population trend, normalized by area, and hail trend shown. Fourteen of the 26 counties had a positive population trend over this time period, of which 12 had a positive hail report trend and 8 had a positive hail days trend. Of the 12 counties that had declining or stable population, all of which are in the rural eastern plains, 8 had decreasing trends in hail reports and 10 had decreasing trends in hail days. Individual county intricacies can be noted in Fig. 12 as well that do not fit with the general correlation pattern. For example, El Paso County, home to Colorado Springs, has seen a growing population and by far the greatest increasing trend in severe hail reports, at just over two additional hail reports per year since 1997. However, the trend in hail days is slightly negative; that is, El Paso County is reporting more severe hail on fewer days. This of course does not hold in every year or for every size threshold, as the 2018 calendar year contributed four ≥3.0-in. (≥76.2 mm) hail reports and three ≥3.0-in. (≥76.2 mm) hail days in El Paso County, both of which were county records. Douglas County, which has the second-highest population trend, also had the greatest increase in severe hail days over the period. Yuma County, on the other hand, has the greatest decreasing trend in both hail reports and hail days of all counties analyzed, despite having a slight increase in population since 2000. Weld and Adams Counties, the top two counties in the state for tornado segments, show decreasing tornado counts since 1997 yet increasing severe hail counts over the same period. Elbert County, which has the greatest increasing tornado report trend since 1997, also has modest increases in severe hail counts and hail days over the same period while increasing in population. Also interesting to note are Cheyenne and Kiowa Counties: as reported in section 3, these counties are two of the most active for tornado segments when counties are normalized by area, but both counties also report decreasing trends in both severe hail reports and days. Again, population growth and smaller sample sizes preclude any statistical significance of the trends, but a stronger signal of population influence does exist on the eastern Colorado severe hail record relative to the tornado record.

Fig. 12.
Fig. 12.

Bar chart of county population trends (normalized by county area) for the period 2000–17 along with county severe hail report and days trends for the period 1997–2017 for 26 eastern Colorado counties within the study domain. The correlations between county population and severe hail reports and severe hail days are 0.425 and 0.441, respectively.

Citation: Journal of Applied Meteorology and Climatology 58, 10; 10.1175/JAMC-D-19-0098.1

The sources of severe hail reports shed further light on the population bias. Table 5 showcases the sources of severe hail reports for the period 1998–2017 across the eastern Colorado domain, using the NCEI Storm Events database. Two-thirds of all reports originate from trained spotters, a much higher proportion than for tornadoes (Table 3). This is largely due to a smaller contribution from storm chasers and law enforcement, as these sources only make up 2.8% and 2.9% of reports, respectively, as opposed to 21.4% and 12.8% of tornado reports. On the other hand, the general public reports a higher proportion of severe hail (12.4%) than for tornadoes (6.4%) across eastern Colorado. The linear nature of hail reports across the eastern plains is most likely from commuters and/or storm chasers who are limited in their choice of paved roadways. Although trained spotters make up the bulk of all severe hail reports in eastern Colorado, Allen et al. (2015) also found that storm chasers are apt to report hail in the Great Plains in spring because of a lack of navigable roads and sparse population. Social media is again near the bottom of the hail report sources. Despite the wealth of hailstone pictures posted on social media outlets, more often than not, the official report for that particular storm comes from a trained spotter in the same area, or, alternatively, the source may get recorded as “general public” even though that person has posted a picture of the hailstone online. However, note that social media reports of hailstones have become very advantageous in rural areas that are hard to reach by car. It is interesting that some 238 severe hail reports (4.5%) are missing source information as compared with only nine tornado reports. This simply means that a hailstone was reported but the meteorologist documenting the report either failed to enter the source or had conflicting information as to where and/or how the report originated (R. Cox 2018, personal communication).

Table 5.

Sources of eastern Colorado severe hail reports for the period 1998–2017. General categories are presented in the first column, and their included subcategories are given in the second column.

Table 5.

Finally, a glimpse at the changes in the length of the eastern Colorado severe hail season is given in Table 4 (bottom) and Fig. 13. Table 4 (bottom) lists the total length as well as average calendar-day first and last severe and significant hail report for the period 1997–2018, and the two halves of that period. Figure 13 then presents time series of the length, starting date, and ending date, of the eastern Colorado severe hail season since 1997. There is an appreciable decreasing trend in both severe and significant hail season length over this time period, though not statistically significant according to a Mann–Kendall nonparametric test. Specifically, the 11-yr average severe hail season length has declined from 175 days to 157 days from the period 1997–2007 to 2008–18, and significant hail season length has declined from 111 days to 101 days over the same two periods. Most of the decrease comes on the tail end of the hail season. In comparing 2018 with 1997, it is seen that the average last severe hail report of the year was approximately 25 days earlier, with significant hail reports ending approximately 14 days earlier (Fig. 13). As was the case for tornadoes, the year 2018 had the shortest severe hail season in the modern data record for Colorado at only 113 days, as a result of a particularly early end to the season on 21 August. In fact, not since 1995 had the last severe hail report in Colorado occurred in August. Figure 13 also reveals a trend toward an earlier first severe hail date since 1997, whereas the date of the first significant hail report has not changed much over time. Although it is beyond the scope of this paper to speculate on the reasons for the shorter severe hail season, and specifically the trend toward an earlier end of the season, one major contributor to rainfall and convection in late summer across eastern Colorado is the southwest monsoon (Mock 1996; Colorado Climate Center 2019). Therefore, a change or shunting of the monsoon annual cycle may serve to affect hailfall across the domain. Further investigation into this trend is warranted. Nevertheless, this finding is insightful when coupled with the previously reported climatological trends, because it shows that even though the severe hail season across eastern Colorado is, on average, shorter in the more recent period, this period has also seen a slightly higher proportion of significant hail reports. In addition to the increasing trend in hail reports and days across the domain, the other critical attribute of the eastern Colorado hail climatology is the population bias that, although improving over time, continues to distort the hail record and influence both the temporal and spatial hail report patterns.

Fig. 13.
Fig. 13.

As in Fig. 7, but for severe (orange) and significant (blue) hail reports, showing time series of (a) total length of season in days and (b) date of first and last annual reports, along with respective linear trend lines.

Citation: Journal of Applied Meteorology and Climatology 58, 10; 10.1175/JAMC-D-19-0098.1

5. Summary and conclusions

Tornadoes and severe hailstorms are the two major severe weather phenomena that occur routinely across Colorado each year. In fact, the eastern half of the state is a local maximum in the frequency of both of these hazards across the United States. Much work has been done to identify unique features of severe weather over eastern Colorado and the rest of the High Plains region, both in case study and broader dynamical contexts, but relatively little work has been done to analyze climatological trends in tornadoes and severe hail. This study has presented the most comprehensive look at the frequency and geography of eastern Colorado tornadoes and severe hail, with analysis starting from the mid-1950s, but focused on 1997 to present, when tornado and severe hail data are more reliable. Documented problems with the quality of tornado and severe hail data are exposed in this localized study, including but not limited to a spike in (E)F0 tornadoes with the advent of Doppler radar in the mid-1990s and a population bias contributing to the upward trend in severe hail reports. In fact, severe hail reports are closely aligned with major population centers and roadways across eastern Colorado. In a mean sense, topography certainly plays a role in the spatial distribution of severe weather in Colorado, with tornado reports showing a maximum in north-central Colorado and severe hail reports concentrated right along the foothills. This study also finds that in general, the length of the tornado season across eastern Colorado is slowly declining along with their frequency of occurrence. On the other hand, although the length of the severe hail season is also decreasing over time (mostly due to an earlier end to the season), the number of severe hail reports over the period 1997–2017 is increasing at a statistically significant pace. In addition, the proportion of hailstones observed as significant and in excess of 3.0 in. (76.2 mm) has an increasing trend since 1997, with 2018 seeing record fractional percentages.

The aim of this study is not to attribute any tornado or severe hail trends to any changes in meteorological or climate patterns. However, given eastern Colorado’s unique location near the Rocky Mountains, dependence on moisture for severe weather, and even some of the analysis presented in this study, investigation into meteorological influences on severe weather trends past and future is a worthy venture and the topic of ongoing work. Of particular interest is the increase in days of severe hail per year that is not consistent with the national trend, and the increasing number of observed hailstones that are greater than 2.0 in. (50.8 mm) across eastern Colorado. However, a limited domain and population influence must also be considered. Also worthy of future investigation and ongoing work by the first author is the human risk from these hazards. As mentioned, the Front Range of Colorado is one of the fastest-growing areas in the United States, and as more people move into the region there are more people and objects that can potentially be damaged by tornadoes and hailstorms, as evidenced by several damaging events in past few years. In addition, the perceptions of risk and vulnerability of the residents of eastern Colorado is important in determining how one prepares for and responds to these hazards. As a motivation for these future endeavors, the comprehensive climatology presented here for one of the most active tornado and severe hail areas in the country can be a worthy and helpful reference for both the local Colorado and national meteorological communities.

Acknowledgments

The authors thank Ed Szoke for sharing his thoughts on Colorado tornado trends and Rob Cox for insights into the documentation of severe weather reports. The authors are also grateful for the helpful comments from three anonymous reviewers. This work is made possible through funding from the National Science Foundation (NSF) Graduate Research Fellowship Program Grant DGE-1321845 as well as NSF Grant AGS-1637244.

REFERENCES

  • Aguilera, E., 2007: Holly tornado: Total destruction. Denver Post, 29 March, https://www.denverpost.com/2007/03/29/holly-tornado-total-destruction/.

    • Search Google Scholar
    • Export Citation
  • Agee, E., and S. Childs, 2014: Adjustments in tornado counts, F-scale intensity, and path width for assessing significant tornado destruction. J. Appl. Meteor. Climatol., 53, 14941505, https://doi.org/10.1175/JAMC-D-13-0235.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Agee, E., and L. Taylor, 2019: Historical analysis of U.S. tornado fatalities (1808–2017): Population, science, and technology. Wea. Climate Soc., 11, 355368, https://doi.org/10.1175/WCAS-D-18-0078.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Agee, E., J. Larson, S. Childs, and A. Marmo, 2016: Spatial redistribution of U.S. tornado activity between 1954 and 2013. J. Appl. Meteor. Climatol., 55, 16811697, https://doi.org/10.1175/JAMC-D-15-0342.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Allen, J. T., and M. K. Tippett, 2015: The characteristics of United States hail reports: 1955–2014. Electron. J. Severe Storms Meteor., 10 (3), http://www.ejssm.org/ojs/index.php/ejssm/article/viewArticle/149.

    • Search Google Scholar
    • Export Citation
  • Allen, J. T., M. K. Tippett, and A. H. Sobel, 2015: An empirical model relating U.S. monthly hail occurrence to large-scale meteorological environment. J. Adv. Model. Earth Syst., 7, 226243, https://doi.org/10.1002/2014MS000397.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Allen, J. T., M. K. Tippett, Y. Kaheil, A. H. Sobel, C. Lepore, S. Nong, and A. Muehlbauer, 2017: An extreme value model for U.S. hail size. Mon. Wea. Rev., 145, 45014519, https://doi.org/10.1175/MWR-D-17-0119.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Allen, J. T., M. J. Molina, and V. A. Gensini, 2018: Modulation of annual cycle of tornadoes by El Niño–Southern Oscillation. Geophys. Res. Lett., 45, 57085717, https://doi.org/10.1029/2018GL077482.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Ashley, W. S., 2007: Spatial and temporal analysis of tornado fatalities in the United States: 1880–2005. Wea. Forecasting, 22, 12141228, https://doi.org/10.1175/2007WAF2007004.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Bieringer, P., and P. S. Ray, 1996: A comparison of tornado warning lead times with and without NEXRAD Doppler radar. Wea. Forecasting, 11, 4752, https://doi.org/10.1175/1520-0434(1996)011<0047:ACOTWL>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Blair, S. F., and Coauthors, 2017: High-resolution hail observations: Implications for NWS warning operations. Wea. Forecasting, 32, 11011119, https://doi.org/10.1175/WAF-D-16-0203.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Brady, R. H., and E. J. Szoke, 1989: A case study of nonmesocyclone tornado development in northeast Colorado: Similarities to waterspout formation. Mon. Wea. Rev., 117, 843856, https://doi.org/10.1175/1520-0493(1989)117<0843:ACSONT>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Brooks, H. E., C. A. Doswell III, and M. P. Kay, 2003a: Climatological estimates of local daily tornado probability for the United States. Wea. Forecasting, 18, 626640, https://doi.org/10.1175/1520-0434(2003)018<0626:CEOLDT>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Brooks, H. E., J. W. Lee, and J. P. Craven, 2003b: The spatial distribution of severe thunderstorm and tornado environments from global reanalysis data. Atmos. Res., 67, 7394, https://doi.org/10.1016/S0169-8095(03)00045-0.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Brooks, H. E., G. W. Carbin, and P. T. Marsh, 2014: Increased variability of tornado occurrence in the United States. Science, 346, 349352, https://doi.org/10.1126/science.1257460.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Changnon, S. A., 1977: The scales of hail. J. Appl. Meteor., 16, 626648, https://doi.org/10.1175/1520-0450(1977)016<0626:TSOH>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Changnon, S. A., 1999: Data and approaches for determining hail risk in the contiguous United States. J. Appl. Meteor. Climatol., 38, 17301739, https://doi.org/10.1175/1520-0450(1999)038<1730:DAAFDH>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Changnon, S. A., 2009: Increasing major hail losses in the US. Climatic Change, 96, 161166, https://doi.org/10.1007/s10584-009-9597-z.

  • Childs, S. J., and R. S. Schumacher, 2018: The anomalously large 2018 hail season of eastern Colorado: A local perspective on a national trend. 29th Conf. on Severe Local Storms, Stowe, VT, Amer. Meteor. Soc., 33, https://ams.confex.com/ams/29SLS/webprogram/Paper348238.html.

  • Childs, S. J., R. S. Schumacher, and J. T. Allen, 2018: Cold-season tornadoes: Climatological and meteorological insights. Wea. Forecasting, 33, 671691, https://doi.org/10.1175/WAF-D-17-0120.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Cintineo, J. L., T. M. Smith, V. Lakshmanan, H. E. Brooks, and K. L. Ortega, 2012: An objective high-resolution hail climatology of the contiguous United States. Wea. Forecasting, 27, 12351248, https://doi.org/10.1175/WAF-D-11-00151.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Clark, A. J., 2017: Observed trends in US tornado frequency. Climate Extremes: Patterns and Mechanisms, Geophys. Monogr., Vol. 226, Amer. Geophys. Union, 237–247.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Coleman, T. A., and P. G. Dixon, 2014: An objective analysis of tornado risk in the United States. Wea. Forecasting, 29, 366376, https://doi.org/10.1175/WAF-D-13-00057.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Colorado Climate Center, 2019: North American monsoon. Colorado State University, accessed 23 July 2019, https://climate.colostate.edu/co_nam.html.

  • Concannon, P., H. Brooks, and C. Doswell, 2000: Climatological risk of strong and violent tornadoes in the United States. Preprints, Second Conf. on Environmental Applications, Long Beach, CA, Amer. Meteor. Soc., 9.4, https://ams.confex.com/ams/annual2000/techprogram/paper_6471.htm.

    • Search Google Scholar
    • Export Citation
  • Cook, A. R., L. M. Leslie, D. B. Parsons, and J. T. Schaefer, 2017: The impact of El Niño–Southern Oscillation (ENSO) on winter and early spring US tornado outbreaks. J. Appl. Meteor. Climatol., 56, 24552478, https://doi.org/10.1175/JAMC-D-16-0249.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Crum, T. D., R. E. Saffle, and J. W. Wilson, 1998: An update on the NEXRAD program and future WSR-88D support to operations. Wea. Forecasting, 13, 253262, https://doi.org/10.1175/1520-0434(1998)013<0253:AUOTNP>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Doesken, N. J., 1994: Hail, hail, hail! The summertime hazard of eastern Colorado. Colorado State University Colorado Climate Publ. Vol. 17 No. 7, 11 pp., https://www.cocorahs.org/media/docs/hail_1994.pdf.

  • Doswell, C. A., III, 1980: Synoptic-scale environments associated with High Plains severe thunderstorms. Bull. Amer. Meteor. Soc., 61, 13881400, https://doi.org/10.1175/1520-0477(1980)061<1374:MCC>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Doswell, C. A., III, 2007: Historical overview of severe convective storms research. Electron. J. Severe Storms Meteor., 2 (1), http://www.ejssm.org/ojs/index.php/ejssm/article/viewArticle/13/15.

    • Search Google Scholar
    • Export Citation
  • Doswell, C. A., III, and D. W. Burgess, 1988: On some issues of United States tornado climatology. Mon. Wea. Rev., 116, 495501, https://doi.org/10.1175/1520-0493(1988)116<0495:OSIOUS>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Edwards, R., and H. Brooks, 2010: Possible impacts of the enhanced Fujita scale on United States tornado data. 25th Conf. on Severe Local Storms, Denver, CO, Amer. Meteor. Soc., P8.28, https://ams.confex.com/ams/pdfpapers/175398.pdf.

    • Search Google Scholar
    • Export Citation
  • Edwards, R., J. G. LaDue, J. T. Ferree, K. Scharfenberg, C. Maier, and W. L. Coulbourne, 2013: Tornado intensity estimation: Past, present, and future. Bull. Amer. Meteor. Soc., 94, 641653, https://doi.org/10.1175/BAMS-D-11-00006.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Edwards, R., J. T. Allen, and G. W. Carbin, 2018: Reliability and climatological impacts of convective wind estimations. J. Appl. Meteor. Climatol., 57, 18251845, https://doi.org/10.1175/JAMC-D-17-0306.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Elsner, J. B., S. C. Elsner, and T. H. Jagger, 2015: The increasing efficiency of tornado days in the United States. Climate Dyn., 45, 651659, https://doi.org/10.1007/s00382-014-2277-3.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Erdman, J., 2017: The most tornado-prone counties in the U.S. The Weather Channel, https://weather.com/storms/tornado/news/most-tornado-prone-us-counties.

  • Farney, T. J., and P. G. Dixon, 2015: Variability of tornado climatology across the continental United States. Int. J. Climatol., 35, 29933006, https://doi.org/10.1002/joc.4188.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Friedrich, K., and Coauthors, 2019: CHAT: The Colorado Hail Accumulation from Thunderstorms project. Bull. Amer. Meteor. Soc., 100, 459472, https://doi.org/10.1175/BAMS-D-16-0277.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Gensini, V. A., and J. T. Allen, 2018: U.S. hail frequency and the Global Wind Oscillation. Geophys. Res. Lett., 45, 16111620, https://doi.org/10.1002/2017GL076822.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Gensini, V. A., and H. E. Brooks, 2018: Spatial trends in United States tornado frequency. Nat. Climate Atmos. Sci., 1, 38, https://doi.org/10.1038/s41612-018-0048-2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Gensini, V. A., and A. Marinaro, 2016: Tornado frequency in the United States related to global relative angular momentum. Mon. Wea. Rev., 144, 801810, https://doi.org/10.1175/MWR-D-15-0289.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Gochis, D., and Coauthors, 2015: The great Colorado flood of September 2013. Bull. Amer. Meteor. Soc., 96, 14611487, https://doi.org/10.1175/BAMS-D-13-00241.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Grazulis, T. P., 1993: Significant Tornadoes: 1680-1991: A Chronology and Analysis of Events. The Tornado Project of Environmental Films, 1340 pp.

  • Grazulis, T. P., J. T. Schaefer, and R. F. Abbey, 1993: Advances in tornado climatology, hazards, and risk assessment since Tornado Symposium II. The Tornado: Its Structure, Dynamics, Prediction, and Hazards, Geophys. Monogr., Vol. 79, Amer. Geophys. Union, 409–426.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Guo, L., K. Wang, and H. B. Bluestein, 2016: Variability of tornado occurrence over the continental United States since 1950. Geophys. Res. Atmos., 121, 69436953, https://doi.org/10.1002/2015JD024465.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Hamill, T. M., 2014: Performance of operational model precipitation forecast guidance during the 2013 Colorado Front-Range floods. Mon. Wea. Rev., 142, 26092618, https://doi.org/10.1175/MWR-D-14-00007.1; Corrigendum, 145, 403–404, https://doi.org/10.1175/MWR-D-16-0006.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Hodanish, S. J., B. J. Vogt, and P. Wolyn, 2019: Colorado lightning climatology. J. Oper. Meteor., 7, 4560, https://doi.org/10.15191/NWAJOM.2019.0704.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Hoogewind, K. A., M. E. Baldwin, and R. J. Trapp, 2017: The impact of climate change on hazardous convective weather in the United States: Insight from high-resolution dynamical downscaling. J. Climate, 30, 10 08110 100, https://doi.org/10.1175/JCLI-D-16-0885.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Kalina, E. A., K. Friedrich, B. C. Motta, W. Deierling, G. T. Stano, and N. N. Rydell, 2016: Colorado plowable hailstorms: Synoptic weather, radar, and lightning characteristics. Wea. Forecasting, 31, 663693, https://doi.org/10.1175/WAF-D-15-0037.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Kelly, D., J. Schaefer, R. McNulty, C. Doswell III, and R. Abbey Jr., 1978: An augmented tornado climatology. Mon. Wea. Rev., 106, 11721183, https://doi.org/10.1175/1520-0493(1978)106<1172:AATC>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Kendall, M. G., 1975: Rank Correlation Measures. Charles Griffin, 202 pp.

  • Kis, A. K., and J. M. Straka, 2010: Nocturnal tornado climatology. Wea. Forecasting, 25, 545561, https://doi.org/10.1175/2009WAF2222294.1.

  • Kumjian, M. R., Z. J. Lebo, and A. M. Ward, 2019: Storms producing large accumulations of small hail. J. Appl. Meteor. Climatol., 58, 341364, https://doi.org/10.1175/JAMC-D-18-0073.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • LaDue, J. G., J. Wurman, M. Levitan, F. Lombardo, C. Karstens, J. Robinson, and W. Coulbourne, 2018: Advances in development of the ASCE/SEI/AMS standard for wind speed estimation in tornadoes and other windstorms. 29th Conf. on Severe Local Storms, Stowe, VT, Amer. Meteor. Soc., 29, https://ams.confex.com/ams/29SLS/webprogram/Paper348726.html.

    • Search Google Scholar
    • Export Citation
  • Lee, S.-K., A. T. Wittenberg, D. B. Enfield, S. J. Weaver, C. Wang, and R. Atlas, 2016: U.S. regional tornado outbreaks and their links to spring ENSO phases and North Atlantic SST variability. Environ. Res. Lett., 11, 044008, https://doi.org/10.1088/1748-9326/11/4/044008.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Maddox, R., D. Rodgers, W. Deitrich, and D. Bartels, 1981: Meteorological settings associated with significant convective storms in Colorado. NOAA ERL OWRM Tech. Memo., 75 pp.

  • Mann, H. B., 1945: Nonparametric tests against trend. Econometrica, 13, 245259, https://doi.org/10.2307/1907187.

  • McDonald, J. R., and K. C. Mehta, 2006: A recommendation for an enhanced Fujita scale (EF-scale). Texas Tech University Wind Science and Engineering Center, 95 pp., https://www.spc.noaa.gov/faq/tornado/ef-ttu.pdf.

  • Mock, C. J., 1996: Climatic controls and spatial variations of precipitation in the western United States. J. Climate, 9, 11111125, https://doi.org/10.1175/1520-0442(1996)009<1111:CCASVO>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Molina, M. J., J. T. Allen, and V. A. Gensini, 2018: The Gulf of Mexico and ENSO influence on subseasonal and seasonal CONUS winter tornado variability. J. Appl. Meteor. Climatol., 57, 24392463, https://doi.org/10.1175/JAMC-D-18-0046.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • NCEI, 2009: State Climate Extremes Committee: Proposed standards for the collection, storage, and measurement of hailstones. NOAA Tech. Rep., 6 pp., https://ncdc.noaa.gov/monitoring-content/extremes/scec/reports/SCEC-Hail-Guide.pdf.

  • NSSL, 2017: NSSL Research: Tornadoes. NOAA, https://www.nssl.noaa.gov/research/tornadoes/.

  • NWS Boulder, 2008: May 22, 2008 Windsor Tornado. National Weather Service, https://www.weather.gov/bou/TornadoMay22nd2008.

  • Ortega, K. L., T. M. Smith, K. L. Manross, K. A. Scharfenberg, A. Witt, A. G. Kolodziej, and J. J. Gourley, 2009: The Severe Hazards and Verification Experiment. Bull. Amer. Meteor. Soc., 90, 15191530, https://doi.org/10.1175/2009BAMS2815.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Potvin, C. K., C. Broyles, P. S. Skinner, H. E. Brooks, and E. Rasmussen, 2019: A Bayesian hierarchical modeling framework for correcting reporting bias in the U.S. tornado database. Wea. Forecasting, 34, 1530, https://doi.org/10.1175/WAF-D-18-0137.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Prein, A. F., and G. J. Holland, 2018: Global estimates of damaging hail hazard. Wea. Climate Extremes, 22, 1023, https://doi.org/10.1016/j.wace.2018.10.004.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Reges, H. W., N. Doesken, J. Turner, N. Newman, A. Bergantino, and Z. Schwalbe, 2016: COCORAHS: The evolution and accomplishments of a volunteer rain gauge network. Bull. Amer. Meteor. Soc., 97 , 18311846, https://doi.org/10.1175/BAMS-D-14-00213.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Rocky Mountain Insurance Information Association, 2018: Hail. RMIIA, http://www.rmiia.org/catastrophes_and_statistics/Hail.asp.

  • Sander, J., J. F. Eichner, E. Faust, and M. Steuer, 2013: Rising variability in thunderstorm-related U.S. losses as a reflection of changes in large-scale thunderstorm forcing. Wea. Climate Soc., 5, 317331, https://doi.org/10.1175/WCAS-D-12-00023.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Schaefer, J. T., and R. Edwards, 1999: The SPC tornado/severe thunderstorm database. Preprints, 11th Conf. on Applied Climatology, Dallas, TX, Amer. Meteor. Soc., 215–220.

  • Schlatter, T. W., and N. Doesken, 2010: Deep hail: Tracking an elusive phenomenon. Weatherwise, 63 (5), 3541, https://doi.org/10.1080/00431672.2010.503841.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Schumacher, R. S., D. T. Lindsey, A. B. Schumacher, J. Braun, S. D. Miller, and J. L. Demuth, 2010: Multidisciplinary analysis of an unusual tornado: Meteorology, climatology, and the communication and interpretation of warnings. Wea. Forecasting, 25, 14121429, https://doi.org/10.1175/2010WAF2222396.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Spears, C. K., 2017: Climatology of Colorado tornadoes: Are Colorado tornadoes growing longer and stronger in time?, Colorado State University Colorado Climate Publ., 16 pp., http://ccc.atmos.colostate.edu/pdfs/Climatology_of_Colorado_Tornadoes.pdf.

  • State Farm, 2019: Top 10 states for hail damage. State Farm, accessed 16 July 2019, https://newsroom.statefarm.com/top-10-states-for-hail-damage-claims-2019?cmpid=PArel031819Top10HailStates&utm_source=Direct.

  • Szoke, E. J., M. Weisman, J. Brown, F. Caracena, and T. Schlatter, 1984: A subsynoptic analysis of the Denver tornadoes of 3 June 1981. Mon. Wea. Rev., 112, 790808, https://doi.org/10.1175/1520-0493(1984)112<0790:ASAOTD>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Szoke, E. J., D. Barjenbruch, R. Glancy, and R. Kleyla, 2006: The Denver Cyclone and tornadoes 25 years later: The continued challenge of predicting non-supercell tornadoes. 23rd Conf. on Severe Local Storms, St. Louis, MO, Amer. Meteor. Soc., P8.7, https://ams.confex.com/ams/pdfpapers/115160.pdf.

    • Search Google Scholar
    • Export Citation
  • Tippett, M. K., 2014: Changing volatility of U.S. annual tornado reports. Geophys. Res. Lett., 41, 69566961, https://doi.org/10.1002/2014GL061347.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Tippett, M. K., 2018: Robustness of relations between the MJO and U.S. tornado occurrence. Mon. Wea. Rev., 146, 38733884, https://doi.org/10.1175/MWR-D-18-0207.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Verbout, S. M., H. E. Brooks, L. M. Leslie, and D. M. Schultz, 2006: Evolution of the U.S. tornado database: 1954–2003. Wea. Forecasting, 21, 8693, https://doi.org/10.1175/WAF910.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Wakimoto, R. M., and J. W. Wilson, 1989: Non-supercell tornadoes. Mon. Wea. Rev., 117, 11131140, https://doi.org/10.1175/1520-0493(1989)117<1113:NST>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Wilcoxon, F., 1945: Individual comparisons by ranking methods. Biom. Bull., 1, 8083, https://doi.org/10.2307/3001968.

  • Wilson, J. W., J. A. Moore, G. B. Foote, B. Martner, A. R. Rodi, T. Uttal, and J. M. Wilczak, 1988: Convection Initiation and Downburst Experiment (CINDE). Bull. Amer. Meteor. Soc., 69, 13281347, https://doi.org/10.1175/1520-0477(1988)069<1328:CIADE>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
Save
  • Aguilera, E., 2007: Holly tornado: Total destruction. Denver Post, 29 March, https://www.denverpost.com/2007/03/29/holly-tornado-total-destruction/.

    • Search Google Scholar
    • Export Citation
  • Agee, E., and S. Childs, 2014: Adjustments in tornado counts, F-scale intensity, and path width for assessing significant tornado destruction. J. Appl. Meteor. Climatol., 53, 14941505, https://doi.org/10.1175/JAMC-D-13-0235.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Agee, E., and L. Taylor, 2019: Historical analysis of U.S. tornado fatalities (1808–2017): Population, science, and technology. Wea. Climate Soc., 11, 355368, https://doi.org/10.1175/WCAS-D-18-0078.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Agee, E., J. Larson, S. Childs, and A. Marmo, 2016: Spatial redistribution of U.S. tornado activity between 1954 and 2013. J. Appl. Meteor. Climatol., 55, 16811697, https://doi.org/10.1175/JAMC-D-15-0342.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Allen, J. T., and M. K. Tippett, 2015: The characteristics of United States hail reports: 1955–2014. Electron. J. Severe Storms Meteor., 10 (3), http://www.ejssm.org/ojs/index.php/ejssm/article/viewArticle/149.

    • Search Google Scholar
    • Export Citation
  • Allen, J. T., M. K. Tippett, and A. H. Sobel, 2015: An empirical model relating U.S. monthly hail occurrence to large-scale meteorological environment. J. Adv. Model. Earth Syst., 7, 226243, https://doi.org/10.1002/2014MS000397.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Allen, J. T., M. K. Tippett, Y. Kaheil, A. H. Sobel, C. Lepore, S. Nong, and A. Muehlbauer, 2017: An extreme value model for U.S. hail size. Mon. Wea. Rev., 145, 45014519, https://doi.org/10.1175/MWR-D-17-0119.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Allen, J. T., M. J. Molina, and V. A. Gensini, 2018: Modulation of annual cycle of tornadoes by El Niño–Southern Oscillation. Geophys. Res. Lett., 45, 57085717, https://doi.org/10.1029/2018GL077482.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Ashley, W. S., 2007: Spatial and temporal analysis of tornado fatalities in the United States: 1880–2005. Wea. Forecasting, 22, 12141228, https://doi.org/10.1175/2007WAF2007004.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Bieringer, P., and P. S. Ray, 1996: A comparison of tornado warning lead times with and without NEXRAD Doppler radar. Wea. Forecasting, 11, 4752, https://doi.org/10.1175/1520-0434(1996)011<0047:ACOTWL>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Blair, S. F., and Coauthors, 2017: High-resolution hail observations: Implications for NWS warning operations. Wea. Forecasting, 32, 11011119, https://doi.org/10.1175/WAF-D-16-0203.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Brady, R. H., and E. J. Szoke, 1989: A case study of nonmesocyclone tornado development in northeast Colorado: Similarities to waterspout formation. Mon. Wea. Rev., 117, 843856, https://doi.org/10.1175/1520-0493(1989)117<0843:ACSONT>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Brooks, H. E., C. A. Doswell III, and M. P. Kay, 2003a: Climatological estimates of local daily tornado probability for the United States. Wea. Forecasting, 18, 626640, https://doi.org/10.1175/1520-0434(2003)018<0626:CEOLDT>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Brooks, H. E., J. W. Lee, and J. P. Craven, 2003b: The spatial distribution of severe thunderstorm and tornado environments from global reanalysis data. Atmos. Res., 67, 7394, https://doi.org/10.1016/S0169-8095(03)00045-0.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Brooks, H. E., G. W. Carbin, and P. T. Marsh, 2014: Increased variability of tornado occurrence in the United States. Science, 346, 349352, https://doi.org/10.1126/science.1257460.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Changnon, S. A., 1977: The scales of hail. J. Appl. Meteor., 16, 626648, https://doi.org/10.1175/1520-0450(1977)016<0626:TSOH>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Changnon, S. A., 1999: Data and approaches for determining hail risk in the contiguous United States. J. Appl. Meteor. Climatol., 38, 17301739, https://doi.org/10.1175/1520-0450(1999)038<1730:DAAFDH>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Changnon, S. A., 2009: Increasing major hail losses in the US. Climatic Change, 96, 161166, https://doi.org/10.1007/s10584-009-9597-z.

  • Childs, S. J., and R. S. Schumacher, 2018: The anomalously large 2018 hail season of eastern Colorado: A local perspective on a national trend. 29th Conf. on Severe Local Storms, Stowe, VT, Amer. Meteor. Soc., 33, https://ams.confex.com/ams/29SLS/webprogram/Paper348238.html.

  • Childs, S. J., R. S. Schumacher, and J. T. Allen, 2018: Cold-season tornadoes: Climatological and meteorological insights. Wea. Forecasting, 33, 671691, https://doi.org/10.1175/WAF-D-17-0120.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Cintineo, J. L., T. M. Smith, V. Lakshmanan, H. E. Brooks, and K. L. Ortega, 2012: An objective high-resolution hail climatology of the contiguous United States. Wea. Forecasting, 27, 12351248, https://doi.org/10.1175/WAF-D-11-00151.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Clark, A. J., 2017: Observed trends in US tornado frequency. Climate Extremes: Patterns and Mechanisms, Geophys. Monogr., Vol. 226, Amer. Geophys. Union, 237–247.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Coleman, T. A., and P. G. Dixon, 2014: An objective analysis of tornado risk in the United States. Wea. Forecasting, 29, 366376, https://doi.org/10.1175/WAF-D-13-00057.1.