Abstract

Wind chill temperature (WCT) is a measure of the human sensation of cold and also is a parameter used to represent the severity of winter weather. This study provides a unique investigation to quantify the spatial patterns of monthly mean, extreme, and severe WCTs across Canada and the United States. WCT was examined across 45 winters (December–February) spanning 1969/70–2013/14 using 156 surface locations reporting hourly meteorological conditions. Intraseasonal analyses of WCT showed that January had 1) the coldest mean WCTs, 2) the most extreme WCTs as statistically represented by the coldest 1% of the monthly WCT frequency distribution at each surface location, and 3) the greatest frequency of severe WCT hours that were ≤ −32°C. The most extreme WCTs were most often located in the Hudson Bay region of Canada, and north-central and northeastern North America experienced the largest monthly changes in WCT during the winter season. Results suggest that intraseasonal changes of air temperature are the primary influence on variations of WCT and that changes of wind speed are a secondary factor.

1. Introduction

The severity of winter weather is often characterized using wind chill temperature (WCT) since the parameter was originally developed as a measure of the human sensation to cold and windy conditions. The WCT quantifies the process of conductive heat loss from the skin as wind transports the warmed air in contact with the skin away and replaces it with cold air. The original development in the 1930s of the wind chill index (WCI; i.e., the predecessor of the WCT that quantified the rate of heat loss from a cylinder of water under varying freezing temperature and wind conditions) was founded with application to assessing impacts on researchers in Antarctica. Since then several formulations of the WCI and WCT have been used for studies at locations in the Arctic, Europe, Asia, South America, and North America. Although many are familiar with WCT from its use in operational meteorology to describe outdoor conditions in winter environments, only a small handful of studies have examined the spatial and temporal variations of WCT—often focused on a specific region or location across varied time periods (e.g., Baldwin and Smithson 1979; Balafoutis 1989; Coronato 1995; Keimig and Bradley 2002; Toros et al. 2005; Rieck and Binau 2008; Lussenden et al. 2014; Mekis et al. 2015).

A recent report (National Academies of Sciences, Engineering, and Medicine 2016) noted that the exploration of WCT is just starting to be used in cold-event attribution studies to investigate the likely impact of changing extreme cold conditions (e.g., Gao et al. 2015). Understanding the spatial, intraseasonal, and interannual variations of cold-air events and the severity of winter weather across a range of time scales can be enhanced through examining WCT because of its derivation from both wind speed and air temperature conditions, as well as its direct relation to human comfort and health (e.g., severe frostbite). The current study provides a unique investigation quantifying and presenting the winter variation of mean, extreme, and severe WCTs, and examines the influence of temperature and wind speed conditions on these variations across Canada and the United States using 45 winter seasons. For this study, extreme WCTs are statistically defined as being in the coldest 1% of hourly WCTs at a location and severe WCTs are defined as −32°C or colder. This study of monthly mean and extreme WCT spatial distributions and the frequency of severe WCT conditions also serves as a baseline for further study of long-term temporal changes of WCT to advance understanding previously provided by Keimig and Bradley (2002) and Mekis et al. (2015). Subsequent research should be able to straightforwardly compare these results with other expansive continental or oceanic regions around the globe to explore a variety of topics ranging from the impacts of winter weather to differences resulting from changing climate regimes.

The historical development of the WCT relation, the variety of studies that have explored variants of the WCT relation for different locations, and studies using WCT to examine its relation to biological and health aspects are described in sections 2a, 2b, and 2c, respectively. Section 3 presents the data and methods used for this study. The results from analyses of monthly mean, extreme, and severe WCTs and their spatial variation are described in section 4. A summary of results is provided in section 5.

2. Background

a. History of WCT

The WCT is a function of both wind speed and air temperature and was developed to straightforwardly represent the experienced decrease in air temperature on exposed skin of the human body that is due to the flow of cold air (Osczevski and Bluestein 2005). As temperatures decrease and/or wind speeds increase, the WCT accordingly cools. Siple and Passel (1945) reviewed much of the historical work to understand atmospheric cooling power attributed to the WCI. In the 1960s, scientists began to report the WCI in terms of an equivalent temperature, although it is not thoroughly documented as to who was the first to develop and utilize this technique. In 1964, Antarctic explorer C. Eagan theorized that people rarely remain still when outside and that this could affect the WCI. Instead of using “calm” conditions as no wind, calm was newly defined as walking speed, 4 mi h−1 (1.79 m s−1), which lessened the severity of the earlier-developed WCI (Eagan 1964). In the development of a WCI for clothed persons, Steadman (1971) also suggested adjusting the WCI to the standard height at which wind speed was measured (i.e., 10 m).

During continued investigation of the WCI, Osczevski (1995) developed a mathematical model of the WCI relation specific to the human face, which is most often exposed during cold weather activity. Bluestein and Zecher (1999) and Osczevski (2000) continued investigation and later developed a new WCT relation (Osczevski and Bluestein 2005) that was implemented by the weather services of both Canada and the United States. The WCT relation takes into account the assumptions of convective and radiative heat loss described in modern heat transfer theory, a still airspeed of 1.34 m s−1 (average walking speed), and the wind speed at face level (i.e., assumed to be two-thirds of the 10-m wind speed measured at weather locations). In addition, a relation connecting WCT to estimates of frostbite risk was developed by Tikuisis and Osczevski (2002).

The WCT calculation was simplified for ease of use in an operational setting by developing the multiple-linear-regression relation

 
formula

where T is the temperature (°C), V is the wind speed at 10-m height (km h−1), and WCT is reported in degrees Celsius. Equation (1) is valid for air temperatures from −50° to 10°C and for wind speeds from 10 to 80 km h−1 (Osczevski and Bluestein 2005). A complementary relation was developed for slower wind speed conditions (i.e., <10 km h−1) and has been used in past studies (e.g., Mekis et al. 2015) to expand the examination of WCT in weak wind regimes:

 
formula

These regression WCT equations and the associated frostbite-risk relation are utilized in the research that is presented here.

b. WCT climatological studies

Studies have examined WCI or WCT in several locations around the world, but these studies provide a fairly limited scope for understanding widespread geographic variations of WCT and the influence of spatial and temporal differences in both temperatures and wind speeds on WCT. For example, WCT has been investigated in Greece (Balafoutis 1989), Patagonia (Coronato 1993, 1995), the United Kingdom (Smithson and Baldwin 1978; Baldwin and Smithson 1979), Turkey (Toros et al. 2005), and China (Yan 2009).

Only a small number of studies have investigated WCTs at specific sites or across portions of Canada and the United States. For example, WCTs in the north-central United States were examined by Rieck and Binau (2008) for the winters from 1973 to 2000 with a specific focus on the frequency of WCTs ≤ −20°F (−29°C). They found a greater frequency of these WCTs in the northern portion of the region and found that heavily forested landscapes typically had warmer WCTs. In addition, Lussenden et al. (2014) investigated winter WCTs for the U.S. high plains region using 57 surface locations and found that January had the highest percentage of WCTs cooling to −10°F (−23°C) or colder and that the coldest WCTs occurred more frequently in the northern portion of the region. Last, Soulé et al. (1992) explored the spatial and temporal variability of mean and extreme WCTs for a 30-yr time period (1961–90) across Georgia in the southeastern United States. Soulé et al. (1992) found that cold WCTs did occur in this more equatorward region and that, although both the interannual and spatial variabilities of WCTs were substantial, there was a large covariance among the six stations spread across the geographic area.

Two studies have examined the temporal trends of WCT across high-latitude regions. Keimig and Bradley (2002) examined trends of afternoon WCTs at 15 Alaskan and northern Canadian locations during October–April for 1953–93. They found that 76% of monthly WCTs warmed in most months over the 40-yr time period, with the most robust warming occurring in Alaska and western Canada. Increases were most substantial in December, January, March, and April, whereas October and November had the largest number of locations experiencing cooling trends. Keimig and Bradley (2002) also examined the associated changes in temperatures and wind speeds and found that 66% of monthly temperature trends warmed and 64% of wind speed trends showed decreasing wind speeds during the time period. Mekis et al. (2015) more recently used hourly observations from 1953 to 2012 at 126 Canadian surface stations to find that significant decreasing trends in the number of days with WCT < −30°C have occurred in every region of Canada with the exception of a few locations along the Pacific and Atlantic coasts.

c. WCT impact studies

Extreme WCTs have been shown to have a severe impact on people, livestock, and wildlife. Thomas (2014) found that, even though livestock species typically have a thick winter coat, high wind speeds can degrade that coat’s insulating qualities and increase vulnerability to the cold. When an animal gets cold, it expends more energy and therefore needs more food and nutrients, which increases the costs for farmers. This situation also increases the likelihood of the animal getting sick, which again increases costs for farmers. Caribou and deer have also been observed to experience increased stress and elevated metabolic rate when exposed to extreme WCTs (Hart et al. 1961). For example, a winter with a record number of hours having dangerous WCTs (e.g., ≤−50°F) led to a crash die-off of several thousand reindeer on Saint Matthew Island in the Bering Sea during the winter of 1964 (Klein et al. 2009). Like mammals, snow crab, which is typically fished for during the winter months off the west coast of Alaska, has shown increased mortality after being exposed to extreme WCTs when brought onto the deck (Warrenchuk and Shirley 2002).

The few studies that have examined the relation of WCT to human health provide evidence that cold WCTs are linked to a wide variety of significant health conditions. Conditions associated with cold WCTs can drastically decrease the time for hypothermia to occur, even at above-freezing ambient air temperatures. The reported annual number of deaths in the United States resulting from cold varies widely. Nixdorf-Miller et al. (2006) reported that hypothermia, resulting from prolonged exposure to cold temperatures, causes approximately 350 deaths in the United States per year, with the elderly and the very young being especially susceptible and the highest mortality rates occurring in Alaska, Montana, and North Dakota (CDCP 2005). In contrast, the U.S. National Weather Service reports much lower numbers of deaths per year related to cold—an average of 35—as compiled by the Office of Services and the National Centers for Environmental Information (NCEI) from information contained in Storm Data (NWS 2007). The effects of wind chill can also increase the risk of frostbite (Danielsson 1996; Tikuisis and Osczevski 2003), and, although clothing can aid in slowing the effects of WCT, no amount of clothing is sufficient to prevent cooling (Steadman 1971). In addition, WCT has been connected to cardiovascular deaths (Kunst et al. 1994) and cardiorespiratory mortality (Carder et al. 2005). A study conducted by Kunst et al. (1994) found that, in the Netherlands, daily variation in mortality from heart disease was more strongly related to WCI than to air temperature. Last, Gill et al. (1988) found that seasonal variation in hospital admission rates for subarachnoid hemorrhage, thromboembolic brain infarction, and ill-defined cerebrovascular disease was more highly correlated with WCT than exclusively with air temperature, humidity, or wind speed.

3. Data and methods

Integrated Surface Global Hourly Data (DSI-3505) archived at NCEI were used as the primary dataset for this research. From the archived dataset that includes 35 000 stations worldwide, 156 stations across Canada and the United States were selected to create a thorough spatial distribution of locations that had continuous data records at least back to 1970 (Fig. 1). Available parameters in the hourly dataset include wind speed and direction, temperature, dewpoint temperature, sea level pressure, station pressure, present weather, and visibility, among others, and all data were quality controlled by NCEI for extremes, consistency, and continuity (Smith et al. 2011). The locations selected were required to have at least 70% of data recorded for 1970–2014, an approach that is similar to that used by Keimig and Bradley (2002). Station selection also required that the location of a station had not moved by more than 5 km during the 1970–2014 time period to reduce the possibility of inhomogeneity in the archived data.

Fig. 1.

Locations of hourly reporting stations used for analyses.

Fig. 1.

Locations of hourly reporting stations used for analyses.

Station histories were examined to determine whether wind speeds had been measured at a height of 10 m above the ground and whether anemometer height had changed. Following the methods of Peterson and Hennessey (1978), Klink (1999a,b), and Pryor et al. (2009), wind speeds measured at nonstandard heights were adjusted to the standard 10-m height using the wind power law with λ = 1/7, assuming near-neutral stability and a flat and relatively smooth surface. These adjustments were not required for all stations or time periods from 1970 to 2014 (e.g., most anemometer height changes occurred prior to 1970; although numerous automated stations in the United States measure at a height other than 10 m). When station data required them, the modifications to wind speed for most hourly observations were found to be small (e.g., generally <0.5 m s−1) since height adjustments were typically on the order of 2 m and therefore resulted in only a very small change in WCT [see Eq. (1)]. This is consistent with an approach used by Pryor et al. (2009) to examine wind speed trends over the contiguous United States, as well as the findings of Mekis et al. (2015) in their study of trends in WCT across Canada. Additional quality-control analyses of the individual station data files were performed to remove duplicate hourly reports.

The data were organized by the meteorological winter months [December–February (DJF)] to examine intraseasonal variations in WCT across Canada and the United States. The mean WCT at each surface location, as well as the most extreme WCTs as statistically represented by the coldest 1%, were determined from hourly observations at each station taking into account times having conditions within the valid limits for Eqs. (1) and (2). Severe WCT conditions were designated as WCTs colder than −25°F (−32°C); this threshold WCT was determined on the basis of frostbite risk (Tikuisis and Osczevski 2002). The frequency of the severe WCT threshold was determined by comparing the number of hours satisfying the threshold with the total number of hourly observations each month for the given station. Monthly mean temperatures and wind speeds were calculated from all hourly observations at each station for the winters from 1969/70 to 2013/14.

Maps developed for this research were constructed within a geographic information system (GIS) framework using ArcGIS software. Derived parameters from station data were objectively analyzed with an inverse-distance-weighting scheme in ArcMap using a variable search radius incorporating 15 data points and an output cell size of 60 km. Maps were developed using a North America Lambert conformal conic projection masked to Canada and the United States. Although the analysis approach resulted in a grid with relatively high spatial resolution, the station coverage is not fully adequate to represent the complex spatial variations for mountainous regions, such as the Rocky Mountains, or across Alaska. An investigation of these types of local- and regional-scale spatial variations would benefit from use of a mesoscale modeling system and falls outside the scope of the current investigation.

4. Results

a. Mean WCT, temperatures, and wind speeds

In examining monthly mean WCTs for December, January, and February, it is observed that colder WCTs were located in north-central Canada and warmed equatorward (Figs. 2a–c). The eastern half of the continent displayed a zonal pattern, whereas the western half had a more meridional pattern oriented from northwest to southeast. For each winter month, the coldest mean WCTs occurred north and west of Hudson Bay with cold monthly mean WCTs positioned over central and eastern continental locations and extending farthest equatorward in the continental interior. These features are likely due to the temperature-modifying influence of oceans in coastal areas to the east and topographic influences associated with the Rocky Mountains, which restrict westward movement of the shallow cold-air masses that are present over the interior of the continent. In each month, there was a localized region of relatively warm mean WCTs around the Great Lakes, demonstrating the temperature-modifying influence of the lakes, both individually and collectively, on equatorward-moving polar air masses (e.g., Sousounis and Shirer 1992).

Fig. 2.

Monthly mean WCT for (a) December, (b) January, and (c) February. Also shown are the change in monthly mean WCT (d) from December to January and (e) from January to February. The changes were calculated as January minus December and February minus January. Positive values indicate a warming of the monthly mean WCT, and negative values indicate a cooling of the monthly mean WCT. All values are in degrees Celsius for the winters of 1969/70–2013/14.

Fig. 2.

Monthly mean WCT for (a) December, (b) January, and (c) February. Also shown are the change in monthly mean WCT (d) from December to January and (e) from January to February. The changes were calculated as January minus December and February minus January. Positive values indicate a warming of the monthly mean WCT, and negative values indicate a cooling of the monthly mean WCT. All values are in degrees Celsius for the winters of 1969/70–2013/14.

Changes from December to January of monthly mean WCT showed cooling across nearly the entire continent with a magnitude of ≤8°C (Fig. 2d). The most significant cooling occurred across the northeastern United States and much of eastern Canada, most notably in the vicinity east of Hudson Bay. This may be due to a seasonal increase of ice and snow cover in the region and the consequent intensification of polar air masses that form in that region before moving equatorward. Much of the United States and western Canada cooled only slightly (≤2°C), whereas areas in the southwestern United States warmed by ≤2°C in monthly mean WCT from December to January. Changes from January to February of monthly mean WCT showed a warming of ≤6°C over almost the entire area (Fig. 2e). The largest amount of warming (2°–6°C) occurred over southern Canada and the north-central United States. Much of the United States warmed by ≤2°C, with very isolated areas of Alaska and near Hudson Bay in Canada experiencing a cooling of mean WCT by ≤2°C.

Monthly mean temperatures (Figs. 3a–c) and monthly mean wind speeds (Figs. 4a–c) were also examined using all hourly observations during winters of 1969/70–2013/14. For DJF, the eastern half of North America had a zonal pattern of monthly mean temperature while the western half had a northwest-to-southeast-oriented pattern with colder temperatures to the north and east. January had the coldest monthly mean temperatures of the winter months located in northern Canada and Alaska (Fig. 3b). To arrive at these cold temperatures, the monthly mean temperatures cooled across the eastern and central portions of the continent by ≤6°C from December to January, with the exception of the southwestern coast of the United States, which warmed by ≤2°C (Fig. 3d). The most pronounced cooling (4°–6°C) was located in far eastern Canada and the Hudson Bay area. This pattern of cooling paralleled the changes in monthly mean WCT from December to January (Fig. 2d).

Fig. 3.

As in Fig. 2, but for temperature.

Fig. 3.

As in Fig. 2, but for temperature.

Fig. 4.

As in Fig. 2, but for wind speed (km h−1). For the changes, positive and negative values indicate an increase and a decrease of monthly wind speed, respectively.

Fig. 4.

As in Fig. 2, but for wind speed (km h−1). For the changes, positive and negative values indicate an increase and a decrease of monthly wind speed, respectively.

Most of North America experienced warming of ≤4°C in monthly mean temperature from January to February (Fig. 3e). The most robust warming (2°–4°C) occurred in western Canada. Less robust warming (≤2°C) occurred in northern Canada, north of Hudson Bay, and in the eastern United States. This pattern differed slightly from that of the monthly mean WCT changes from January to February (Fig. 2e), for which greater warming was located in eastern Canada rather than over the western portion of the continent.

Monthly mean wind speeds had greater heterogeneity to their spatial pattern. In general, the monthly mean wind speeds were faster over the central and eastern portion of the continent, with the exceptions of the southeastern United States and north of the Great Lakes, and slower west of the Rocky Mountains with an extension into Alaska (Figs. 4a–c). Monthly mean wind speeds increased slightly over much of the continent from December to January, with larger increases in isolated regions of the eastern United States (Fig. 4d). Monthly mean wind speeds slightly decreased from December to January in the western United States, southwestern Canada, and other localized regions of North America. From January to February, monthly mean wind speeds slightly decreased across the continental interior, with the largest decreases occurring in the eastern Great Lakes region (Fig. 4e). Monthly mean wind speeds increased in northwestern Canada and Alaska, as well as in the southwestern United States from January to February.

b. Extreme WCTs

The coldest 1% of hourly WCTs were also examined to determine the magnitude and spatial distribution of extreme WCTs across Canada and the United States (Figs. 5a–c). As might be expected because of the high latitude, the most extreme WCTs were located in northern Canada and Alaska. The extreme WCTs resembled the spatial distribution of the mean WCT, with the coldest extreme WCTs extending equatorward through the continental interior, a northwest–southeast-oriented pattern along the Rocky Mountains, and a zonal pattern over the eastern part of the continent. The coldest extreme WCT was −62°C and occurred during January near Hudson Bay in northern Canada, a region known for frequent high winds. This region is called the wind alley of Hudson Bay by locals (Mekis and Brown 2010).

Fig. 5.

Similar to Fig. 2, but for extreme WCTs, which are statistically represented by the coldest 1% of hourly WCTs.

Fig. 5.

Similar to Fig. 2, but for extreme WCTs, which are statistically represented by the coldest 1% of hourly WCTs.

The extreme WCTs cooled in the eastern half of the United States, Alaska, and in the majority of Canada by ≤6°C from December to January (Fig. 5d) while a warming of generally ≤2°C occurred in the western and south-central United States, with an area of weak cooling (≤2°C) around Utah. The regions of cooling in the extreme WCTs had contributions from both cooling air temperatures (Fig. 3d) and increasing wind speeds (Fig. 4d). The region over the western United States of slight warming in the extreme WCTs resulted primarily from weakening winds offsetting steady or slightly cooling air temperatures.

The extreme WCTs warmed by ≤6°C through Canada and much of the eastern United States, as well as most of the western United States from January to February (Fig. 5e). The most notable warming of extreme WCTs (4°–6°C) occurred in the eastern United States, with substantial warming also occurring along the western coast of Canada and Alaska. Given that air temperatures over most of North America warmed by ≤4°C from January to February (Fig. 3e), the more rapid warming of extreme WCTs in the eastern United States resulted from the additional contribution of decreasing wind speeds, especially over the upper Midwest and Ohio River valley (Fig. 4e). From January to February, cooling of extreme WCTs was generally ≤ 2°C, occurred in eastern Colorado and western Texas, and seemed to result from strengthening winds in the region (Fig. 4e) while temperatures warmed (Fig. 3e).

An examination of several stations demonstrates the varied contributions of wind speed and temperature to extreme WCTs in different geographic regions during the winter months. While extreme WCTs were generally the coldest during January, the timing of the coldest extreme WCTs does vary by location, especially in high-latitude regions of Canada. For example, monthly thresholds for extreme WCTs during December, January, and February were −37.5°, −39.2°, and −36.7°C, respectively, at Huron, South Dakota, and were −55.1°, −56.6°, and −57.6°C, respectively, at Coral Harbour, Nunavut, Canada. For Huron, the coldest WCTs were primarily during January when there was a greater frequency of stronger winds than in December and February (Figs. 6a,b). Colder extreme WCTs at Coral Harbour primarily occurred during February in association with cold air temperatures and wind speeds of generally ≤20 km h−1 as compared with extreme WCTs during both December and January that generally occurred with warmer air temperatures and wind speeds of 15–25 km h−1 (Figs. 6a,b). Although Coral Harbour is a high-latitude location that is often in the area of developing polar air masses, a large frequency of extreme winter WCTs at this site (80.5%) occurred with wind speeds > 15 km h−1 rather than weak or calm wind conditions. In contrast, extreme WCTs during winters at Huron occurred frequently (40.8%) with wind speeds of ≤15 km h−1 and cold temperatures. This finding demonstrates that times of extreme WCTs in the northern Great Plains of the United States are less likely to be associated with blizzards that often occur in this region (Schwartz and Schmidlin 2002; Coleman and Schwartz 2017).

Fig. 6.

Monthly analyses of (a) WCTs at or below the extreme WCT threshold as a function of wind speed for Huron (top grouping) and Coral Harbour (bottom grouping) and (b) air temperature as a function of wind speed during extreme wind chill hours at Huron (top grouping) and Coral Harbour (bottom grouping). The frequency of binned data points is represented by the size scale of the marker for data from the winters of 1969/70–2013/14. Monthly extreme WCTs during December, January, and February are −37.5°, −39.2°, and −36.7°C, respectively, at Huron and −55.1°, −56.6°, and −57.6°C at Coral Harbour.

Fig. 6.

Monthly analyses of (a) WCTs at or below the extreme WCT threshold as a function of wind speed for Huron (top grouping) and Coral Harbour (bottom grouping) and (b) air temperature as a function of wind speed during extreme wind chill hours at Huron (top grouping) and Coral Harbour (bottom grouping). The frequency of binned data points is represented by the size scale of the marker for data from the winters of 1969/70–2013/14. Monthly extreme WCTs during December, January, and February are −37.5°, −39.2°, and −36.7°C, respectively, at Huron and −55.1°, −56.6°, and −57.6°C at Coral Harbour.

c. Frequency of severe WCTs

Severe WCTs occurred most frequently in the vicinity of Hudson Bay and in northern Alaska (Figs. 7a–c). During December, the frequency of severe WCT hours neared 60% of the time in northern Canada, although the frequency sharply decreased south of that region (Fig. 7a). Severe WCTs seldom occurred in southeastern and southwestern Canada (<5%) or in the United States with the exception of Montana, North Dakota, South Dakota, and Minnesota (5%–10%). In January, the severe WCTs became more frequent at equatorward locations across the continent and also increased in frequency in both eastern and central Canada (up to 20%–40%; Fig. 7b). In February, the largest frequency of severe WCTs was in the vicinity of Hudson Bay and northern Alaska (Fig. 7c). In addition, their equatorward occurrence lessened relative to that of January (Fig. 7b) and was more similar to December (Fig. 7a).

Fig. 7.

Similar to Fig. 2, but for monthly percent of hours of severe WCTs, −32°C or colder. For the changes, positive values indicate an increase in the percent of hours of severe WCTs, and negative values represent a decrease in the percent of hours of severe WCTs.

Fig. 7.

Similar to Fig. 2, but for monthly percent of hours of severe WCTs, −32°C or colder. For the changes, positive values indicate an increase in the percent of hours of severe WCTs, and negative values represent a decrease in the percent of hours of severe WCTs.

The frequency of severe WCT hours increased across all of Canada by ≤24% from December to January, with the most notable increase occurring east of Hudson Bay (Fig. 7d). Small changes in frequency occurred in most of the United States, with the exception of a 4%–8% increase in the north-central United States. The frequency of severe WCT hours decreased across most of Canada and the United States by ≤12% from January to February (Fig. 7e), with the largest decrease in frequency (6%–12%) over south-central Canada and the north-central United States.

Although most of the United States experiences severe WCTs on average for less than 5% of the hours during the winter months, many location often have many hours of subfreezing WCTs. For example, Binghamton (New York), Helena (Montana), and Chicago (Illinois) experienced more than 48 h with WCTs between 0° and −10°C during each winter month (Fig. 8). The largest frequency typically occurred for WCTs between 0° and −15°C for each of the stations shown in Fig. 8. Binghamton experienced 63% of observed winter WCTs in this range, with Helena at 61%, Huron at 51.7%, and Chicago at 64.4%. Huron has a broader distribution of winter WCTs with greater frequency of WCTs colder than −25°C because the site location is in an area of faster mean wind speeds (Fig. 4) and is more prominently in the path of equatorward moving polar and Arctic air masses associated with cold-air outbreaks over the United States (e.g., Walsh et al. 2001; Cellitti et al. 2006).

Fig. 8.

Monthly frequency (h month−1) of WCT for (a) Binghamton, (b) Helena, (c) Huron, and Chicago for the winters of 1969/19–2013/14.

Fig. 8.

Monthly frequency (h month−1) of WCT for (a) Binghamton, (b) Helena, (c) Huron, and Chicago for the winters of 1969/19–2013/14.

5. Summary

An investigation of the spatial distribution and intraseasonal variation of monthly mean, severe, and extreme WCTs across Canada and the United States was completed for the winter months from 1969/70 to 2013/14. The variables analyzed were monthly mean WCT, the coldest 1% of hourly WCTs to represent extreme WCTs, the frequency of severe WCT hours ≤ −32°C, monthly mean temperatures, and monthly mean wind speeds.

January was typically the coldest month for most of Canada and the United States for both the mean and extreme WCTs and had the largest area of most frequently occurring severe WCTs. North-central Canada was the location that experienced the coldest mean WCTs, the most extreme WCTs, and the highest frequency of severe WCT hours. The frequency of severe WCT hours was ≤ 5% of hours during December, January, and February for nearly all of the United States and over southeastern and southwestern Canada. This suggests that on average ≤ 108 h each winter experience these severely cold WCTs. The number of hours with severe WCTs during an individual winter can vary greatly depending on location and the large-scale atmospheric pattern, leading to differing average winter temperatures. For example, at Grand Island, Nebraska, during the cold winter of 1978/79, 174 h of severe WCTs were spread across 24 days and during the warm winter of 2001/02 no hours with severe WCTs occurred.

The region of coldest WCTs extended farthest equatorward through the interior of Canada and the United States and was hindered from expanding westward by the Rocky Mountains, as well as being modified over a portion of eastern North America by the Great Lakes. The changes in these variables from December to January show an intensification in the polar air masses associated with extreme WCTs in north-central Canada, an extension equatorward, and movement eastward. The region of the United States that experiences the coldest WCTs in January included Montana, North Dakota, South Dakota, and Minnesota where mean WCTs existed from −16° to −24°C and the threshold-defining extreme WCTs were observed from −36° to −44°C.

As inferred from the analyses in the current study, over most regions of Canada and the United States the changes of the monthly temperature appeared to have a greater influence on intraseasonal variations of WCT than did changes in wind speed. The changes in mean and extreme WCTs during the winter season are consistent with the typical strengthening and/or increasing frequency of polar air masses forming and moving equatorward during cold-air outbreaks over continental North America (Cellitti et al. 2006; Walsh et al. 2001; Colle and Mass 1995; Konrad and Colucci 1989). Since WCTs are derived from both air temperatures and wind speeds, it is an important parameter to investigate further with respect to long-term trends so as to advance understanding of changing winter climates. An examination of the spatial and temporal tendencies of mean and extreme WCTs and the frequency of severe WCT hours over intraseasonal, interannual, and decadal time scales can provide insight into both meteorological and climate variability across North America from a different and impact-relevant perspective.

Acknowledgments

This research began during the 2014 Undergraduate Summer Research Program at Hobart and William Smith (HWS) Colleges, and we extend our gratitude to Michael Brackett of North Carolina State University for his assistance. Additional research was completed within the HWS Honors program through collaboration of the authors. This research was supported by National Science Foundation Research Grant AGS-1258548 and the HWS Provost’s Office. We gratefully acknowledge contributions from beneficial conversations with Dr. Nicholas D. Metz of HWS and Dr. John E. Walsh of the International Arctic Research Center at the University of Alaska Fairbanks and from the IT/GIS support of Robert Beutner at HWS.

REFERENCES

REFERENCES
Balafoutis
,
C.
,
1989
:
Diurnal variation of wind-chill at Thessaloniki, Greece
.
Int. J. Biometeor.
,
33
,
266
271
, doi:.
Baldwin
,
H.
, and
P.
Smithson
,
1979
:
Wind chill in upland Britain
.
Weather
,
34
,
294
308
, doi:.
Bluestein
,
M.
, and
J.
Zecher
,
1999
:
A new approach to an accurate wind chill factor
.
Bull. Amer. Meteor. Soc.
,
80
,
1893
1899
, doi:.
Carder
,
M.
,
R.
McNamee
,
I.
Beverland
,
R.
Elton
,
G. R.
Cohen
,
J.
Boyd
, and
R. M.
Agius
,
2005
:
The lagged effect of cold temperature and wind chill on cardiorespiratory mortality in Scotland
.
Occup. Environ. Med.
,
62
,
702
710
, doi:.
Cellitti
,
M. P.
,
J. E.
Walsh
,
R. M.
Rauber
, and
D. H.
Portis
,
2006
:
Extreme cold air outbreaks over the United States, the polar vortex, and the large‐scale circulation
.
J. Geophys. Res.
,
111
,
D02114
, doi:.
Centers for Disease Control and Prevention
,
2005
:
Hypothermia-related deaths: United States, 2003–2004
.
Morbidity Mortal. Wkly. Rep.
,
54
,
173
175
.
Coleman
,
J. S.
, and
R. M.
Schwartz
,
2017
:
An updated blizzard climatology of the contiguous United States (1959–2014): An examination of spatiotemporal trends
.
J. Appl. Meteor. Climatol.
,
56
,
173
187
, doi:.
Colle
,
B.
, and
C.
Mass
,
1995
:
The structure and evolution of cold surges east of the Rocky Mountains
.
Mon. Wea. Rev.
,
123
,
2577
2610
, doi:.
Coronato
,
F.
,
1993
:
Wind chill factor applied to Patagonian climatology
.
Int. J. Biometeor.
,
37
,
1
6
, doi:.
Coronato
,
F.
,
1995
:
Wind chill influence on thermal conditions in north Patagonia
.
Int. J. Biometeor.
,
39
,
87
93
, doi:.
Danielsson
,
U.
,
1996
:
Windchill and the risk of tissue freezing
.
J. Appl. Physiol.
,
81
,
2666
2673
.
Eagan
,
C.
,
1964
: The effect of air movement on atmospheric cooling power. In Review of research on military problems in cold regions. C. Kolb and F. Holstrom, Eds., Arctic Aeromedical Laboratory Tech. Rep. TDR-64-28, 147–156.
Gao
,
Y.
,
L. R.
Leung
,
J.
Lu
, and
G.
Masato
,
2015
:
Persistent cold air outbreaks over North America in a warming climate
.
Environ. Res. Lett.
,
10
,
044001
, doi:.
Gill
,
J.
,
P.
Davies
,
S.
Gill
, and
D.
Beevers
,
1988
:
Wind-chill and the seasonal variation of cerebrovascular disease
.
J. Clin. Epidemiol.
,
41
,
225
230
, doi:.
Hart
,
J.
,
O.
Heroux
,
W.
Cottle
, and
C.
Mills
,
1961
:
The influence of climate on metabolic and thermal responses of infant caribou
.
Can. J. Zool.
,
39
,
845
856
, doi:.
Keimig
,
F.
, and
R.
Bradley
,
2002
:
Recent changes in wind chill temperatures at high latitudes in North America
.
Geophys. Res. Lett.
,
29
,
4-1
4-4
, doi:.
Klein
,
D.
,
J.
Walsh
, and
M.
Shulski
,
2009
:
What killed the reindeer of Saint Matthew Island?
Weatherwise
,
62
(
6
),
32
38
, doi:.
Klink
,
K.
,
1999a
:
Climatological mean and interannual variance of United States surface wind speed, direction and velocity
.
Int. J. Climatol.
,
19
,
471
488
, doi:.
Klink
,
K.
,
1999b
:
Trends in mean monthly maximum and minimum surface wind speeds in the coterminous United States, 1961 to 1990
.
Climate Res.
,
13
,
193
205
, doi:.
Konrad
,
C. E.
, and
S.
Colucci
,
1989
:
An examination of extreme cold air outbreaks over eastern North America
.
Mon. Wea. Rev.
,
117
,
2687
2700
, doi:.
Kunst
,
A.
,
F.
Groenhof
, and
J.
Mackenbach
,
1994
:
The association between two wind chill indices and daily mortality variation in the Netherlands
.
Amer. J. Public Health
,
84
,
1738
1742
, doi:.
Lussenden
,
H. B.
,
N.
Umphlett
,
M. D.
Shulski
, and
D.
Ebert
,
2014
:
Winter wind chill climatology for the high plains region
.
J. Service Climatol.
,
7
,
1
21
.
Mekis
,
É.
, and
R.
Brown
,
2010
:
Derivation of an adjustment factor map for the estimation of the water equivalent of snowfall from ruler measurements in Canada
.
Atmos.–Ocean
,
48
,
284
293
, doi:.
Mekis
,
É.
,
L. A.
Vincent
,
M. W.
Shephard
, and
X.
Zhang
,
2015
:
Observed trends in severe weather conditions based on humidex, wind chill, and heavy rainfall events in Canada for 1953–2012
.
Atmos.–Ocean
,
53
,
383
397
, doi:.
National Academies of Sciences, Engineering, and Medicine
,
2016
: Attribution of Extreme Weather Events in the Context of Climate Change. The National Academies Press, 186 pp., doi:.
Nixdorf-Miller
,
A.
,
D. M.
Hunsaker
, and
J. C.
Hunsaker
,
2006
:
Hypothermia and hyperthermia medicolegal investigation of morbidity and mortality from exposure to environmental temperature extremes
.
Arch. Pathol. Lab. Med.
,
130
,
1297
1304
, doi:.
NWS
,
2007
: Storm Data preparation. National Weather Service Instruction (NWSI) 10-1605, 110 pp., https://www.ncdc.noaa.gov/stormevents/pd01016005curr.pdf.
Osczevski
,
R. J.
,
1995
:
The basis of wind chill
.
Arctic
,
48
,
372
382
, doi:.
Osczevski
,
R. J.
,
2000
:
Windward cooling: An overlooked factor in the calculation of wind chill
.
Bull. Amer. Meteor. Soc.
,
81
,
2975
2978
, doi:.
Osczevski
,
R. J.
, and
M.
Bluestein
,
2005
:
The new wind chill temperature equivalent chart
.
Bull. Amer. Meteor. Soc.
,
86
,
1453
1458
, doi:.
Peterson
,
E. W.
, and
J. P.
Hennessey
Jr.
,
1978
:
On the use of power laws for estimates of wind power potential
.
J. Appl. Meteor.
,
17
,
390
394
, doi:.
Pryor
,
S. C.
, and Coauthors
,
2009
:
Wind speed trends over the contiguous United States
.
J. Geophys. Res.
,
114
,
D14105
, doi:.
Rieck
,
T. C.
, and
S. E.
Binau
,
2008
: Wind chill climatology for the north-central United States. NOAA Tech. Service Publ. TSP-15, 11 pp., https://www.weather.gov/media/crh/publications/TSP/TSP-15.pdf.
Schwartz
,
R. M.
, and
T. W.
Schmidlin
,
2002
:
Climatology of blizzards in the conterminous United States, 1959–2000
.
J. Climate
,
15
,
1765
1772
, doi:.
Siple
,
P.
, and
C.
Passel
,
1945
:
Measurements of dry atmospheric cooling in subfreezing temperatures
.
Proc. Amer. Philos. Soc.
,
89
,
177
199
.
Smith
,
A.
,
N.
Lott
, and
R.
Vose
,
2011
:
The integrated surface database: Recent developments and partnerships
.
Bull. Amer. Meteor. Soc.
,
92
,
704
708
, doi:.
Smithson
,
P.
, and
H.
Baldwin
,
1978
:
Wind chill in lowland Britain
.
Weather
,
33
,
463
476
, doi:.
Soulé
,
P.
,
V.
Meentemeyer
, and
W.
Bland
,
1992
:
Temporal trends of mean and extreme wind chill in Georgia, 1961–1990
.
Southeast. Geogr.
,
32
,
97
112
, doi:.
Sousounis
,
P. J.
, and
H. N.
Shirer
,
1992
:
Lake aggregate mesoscale disturbances. Part I: Linear analysis
.
J. Atmos. Sci.
,
49
,
80
96
, doi:.
Steadman
,
R. G.
,
1971
:
Indices of wind chill of clothed persons
.
J. Appl. Meteor.
,
10
,
674
683
, doi:.
Thomas
,
H.
,
2014
: Wind chill affects cattle. American Angus Association, http://www.angusbeefbulletin.com/extra/2014/01jan14/0114hn_windchill.html.
Tikuisis
,
P.
, and
R.
Osczevski
,
2002
:
Dynamic model of facial cooling
.
J. Appl. Meteor.
,
41
,
1241
1246
, doi:.
Tikuisis
,
P.
, and
R.
Osczevski
,
2003
:
Facial cooling during cold air exposure
.
Bull. Amer. Meteor. Soc.
,
84
,
927
933
, doi:.
Toros
,
H.
,
A.
Deniz
,
L.
Şaylan
,
M.
Şen
, and
O.
Baloğlu
,
2005
:
Spatial variability of chilling temperature in Turkey and its effect on human comfort
.
Meteor. Atmos. Phys.
,
88
,
107
118
, doi:.
Walsh
,
J.
,
A.
Phillips
,
D.
Portis
, and
W.
Chapman
,
2001
:
Extreme cold outbreaks in the United States and Europe, 1948–99
.
J. Climate
,
14
,
2642
2658
, doi:.
Warrenchuk
,
J.
, and
T.
Shirley
,
2002
. Effects of wind chill on the snow crab (Chionoecetes opilio). Crabs in Cold Water Regions: Biology, Management, and Economics, A. J. Paul et al., Eds., University of Alaska Sea Grant College Program, 81–96.
Yan
,
Y.
,
2009
:
Spatial variability of wind chill temperatures in China
.
Int. J. Meteor.
,
34
,
219
228
.

Footnotes

© 2017 American Meteorological Society. For information regarding reuse of this content and general copyright information, consult the AMS Copyright Policy (www.ametsoc.org/PUBSReuseLicenses).