Southeastern Naturalist
.A. Vogel and R.J. Leffler
2015 Vol. 14, Special Issue 7
18
Canaan Valley & Environs
2015 Southeastern Naturalist 13(Special Issue 7):18–32
Climate of Canaan Valley
Christoph A. Vogel1,2,* and Robert J. Leffler3
Abstract - In this paper, we present and examine climate data from 1944 to 2002 for
Canaan Valley, WV, including average, extreme, and monthly and seasonal temperature,
and precipitation and snowfall amounts. The data, collected over decades by
several dedicated National Weather Service cooperative observers, indicate that Canaan
Valley’s “cash crop” may indeed be its climate. The Canaan Valley has summer
temperatures similar to those found in northern New England, an average seasonal
snowfall higher than any large city in the US, and a shorter growing season than that
of Fairbanks, AK. We highlight the area’s exceptional climate and compare it to other
well known locations. We also present and assess climate trends, including some relationships
to the El Niño Southern Oscillation state, in Canaan Valley’s 57-year record.
Physical Geography of Canaan Valley
Canaan Valley (hereafter, the Valley) lies in north-central West Virginia
in a physiographic region called the Allegheny Plateau. Canaan Valley is the
highest valley of its size east of the Mississippi River in North America and
has an average elevation of 3200 ft (974 m) above sea level. The oval-shaped
valley resembles a huge bathtub with the axis orientated northeast–southwest.
The Valley floor is about 5 mi (8 km) wide and 10 mi (16.1 km) long,
covering about 50 mi2 (80.5 km2). The Valley and the erosion-softened slopes
surrounding it are drained by the Blackwater River, a tributary of the Ohio and
Mississippi river drainages.
The lowest elevation is 3100 ft (944.8 m) on the northwest side of the Valley
in a notch between Canaan and Brown mountains where the Blackwater River
exits Canaan Valley on its journey westward. The highest point is about 4450 ft
(1356 m) at the summit of Weiss Knob on the southeast rim of the Valley.
The Valley’s northeastern rim—the Eastern Continental Divide—forms an
important weather, climate, and drainage boundary. Precipitation falling to the
west of the Divide drains into the Gulf of Mexico. Precipitation falling to the east
drains into the Atlantic Ocean.
Observation and Data Sources
In this study, we used weather and climate observations made by volunteer
weather observers from the National Oceanic and Atmospheric Administration
1NOAA Liaison to Canaan Valley Institute, NOAA Atmospheric Turbulence and Diffusion
Division, Oak Ridge Associated Universities, Oak Ridge, TN 37830. 2Current
address - C.A. Vogel Consulting, Mt. Pleasant, SC 29466.3NOAA National Weather
Service, Office of Climate, Water, and Weather Services, Silver Spring, MD 20910 [now
retired]. *Corresponding author - chascav100@gmail.com.
Southeastern Naturalist
C.A. Vogel and R.J. Leffler
2015 Vol. 14, Special Issue 7
19
(NOAA)/National Weather Service (NWS) Cooperative Observer Program
(COOP) in Canaan Valley, WV. We considered the period July 1944–July
2002. The Thompson Family gathered data from 1944 to 1995 at their farm
located near the center of the Valley at an elevation of 3250 ft (990.6 m). After
1995, a new site about 1 mi (1.6 m) northwest of the Thompson’s farm, near
State Route 32, called Canaan Valley 2, was established. The two stations ran
concurrently between 1993 and 1996; the Canaan Valley 2 station continues to
operate today.
NWS COOP daily data are quality controlled, archived, and then published
on a monthly basis by the National Climatic Data Center in the NOAA Climatological
Data publication. Observations are taken using instrumentation installed
and maintained by the NWS. COOP observers are trained and supervised by
NWS field technicians, and site visitations occur at least once per year. Daily
observations taken include 24-hour maximum and minimum temperatures, liquid-
precipitation equivalent, snowfall and snow depth, and special phenomena
(e.g., days with thunder, hail, damaging winds, fog, etc.). Thirty-year decadal
averages are based on the 1961–1990 period. Extremes are based on the period
July 1944–July 2002.
General Climatic Conditions
According to Thornthwaite’s climate classifications (Thornthwaite 1948), the
Valley has a cold, humid-type climate, just two steps warmer than the tundra classification.
The Valley’s climate is primarily influenced by three factors: location,
high elevation (3200 ft/ 975 m), and topographic setting (bowl-like frost hollow).
Location
The Valley is in the mid-latitudes on the east side of a northern hemisphere
continent, on the crest of the Allegheny Mountains, on the windward side of the
eastern Continental Divide, and in close proximity to the Atlantic Ocean. This
location has broad implications. Westerly winds dominate; thus, relatively warm,
moist air masses from the Gulf of Mexico and cold, dry masses from Canada
alternately affect the area, creating large day-to-day weather variability and unsettled
conditions as the air masses mix.
Summer is characterized by light winds and moist air flowing northeastward
from the southeast and Gulf of Mexico, and localized thundershowers are
common. Winter is dominated by strong winds and alternating pulses of cold, dry
air from Canada and warm, moist air from the Gulf of Mexico. Unsettled conditions
are common during the colder months and can bring unrelenting days of
blizzards or rain, depending on the temperature.
The Valley’s location at the crest of the Alleghenies is a primary factor in
enhancing the abundant precipitation—both rain and snow—that normally falls
over much of the eastern US. Moisture-laden air moving into the Valley from the
north, south, east, and west is moved by the prevailing westerlies and forced to
rise and cool, enhancing cloud cover and precipitation. The high elevation also
Southeastern Naturalist
.A. Vogel and R.J. Leffler
2015 Vol. 14, Special Issue 7
20
means that temperatures are frequently below freezing; thus, more of the precipitation
falls as snow here than at lower-elevation sites.
The Allegheny Mountains crest in north–central West Virginia, forming a
significant barrier about 100 mi (160.9 km) long and 15 mi (24.1 km), wide with
elevations of 3000 ft (914 m) to nearly 5000 ft (1524 m). The elevated plateau surface
averages about 3000 ft (914 m) above the lowlands to the west and east. This
topographic barrier causes broad orographic uplift of air masses traveling across
it. Being near the crest of the rise also results in easterly winds laden with Atlantic
moisture having an upslope path. Upslope causes water vapor to cool and condense
which produces increased clouds and precipitation (Leffler 1974, 1977)
High elevation
The Valley’s 3200-ft (975-m) elevation means that temperatures are typically
10–15 °F (5.6–8.3 °C) cooler than in surrounding lowlands. Elevational differences
in temperature are greater in the summer than in the winter. During winter
in the Valley, temperatures are more frequently below freezing, causing more
of the precipitation to fall as snow there than at nearby sites. During the winter
months, temperatures occasionally increase with elevation. This phenomenon
occurs when strong, high-level winds force warm air up and over the shallow,
dense, cold air at lower elevations. Under these conditions, a temperature inversion
occurs, with warmer conditions occurring in the Valley and the surrounding
highlands than in the lowlands.
Topographic setting
The combination of the Valley’s position on the crest of the Alleghenies and
its high-elevation floor are an ideal configuration for creating low temperatures.
This topography creates conditions which efficiently contain the build-up of cold
air drainage under clear, calm weather conditions, resulting in low minimum
temperatures during all months.
Topographic settings such as the Valley’s are sometimes referred to as frost
hollows. The term is usually reserved for low spots, which exhibit an increased
frequency of frosts and are topographically ideal for trapping cold air under clear
skies and windless conditions. The Valley is a textbook example of a large-scale
elevated frost hollow.
Specific Climatic Conditions
The Valley has a high-altitude climate, a feature that translates into cool
shade and warm sun on bright, sunny days. Another high-elevation feature is
the potential for large day–night temperature fluctuations, especially when the
air is dry.
The Valley’s average annual temperature is 45 °F (7.2 oC; Table 1). Summers
(June–August) are cool—the average afternoon maximum is 75 oF (23.9
°C) and the average morning minimum is 51 °F (10.5 °C). About half of
summer mornings experience temperatures ~40 oF (4.4 °C) or below. Summer
Southeastern Naturalist
C.A. Vogel and R.J. Leffler
2015 Vol. 14, Special Issue 7
21
temperatures average 5 degrees cooler than Burlington, VT, 400 mi (644 km) to
the northeast, and about 15 °F (9.4 °C) cooler than big cities such as Baltimore,
MD and Washington, DC, which are only 200 mi (322 km) and 150 mi (241
km) to the east, respectively. On average, days with a high temperature of 90 °F
(32.2 °C) occur only once every 15 years in the Valley. For comparison, Washington,
DC sizzles with an average of 38 days per year when the temperature
reaches 90o F (32.2 °C).
Freezes can occur in the Valley during any month; on average, there is a freeze
in July once every six years. A minimum temperature of 27 °F (-2.8 °C) or lower
has been recorded in all summer months. The Valley’s average summer growing
season (consecutive days without freezing temperatures) is only 89 days, with
the average last frost occurring 1 June and the first on 30 August, and a considerably
shorter frost-free period in some years. The average growing season in the
Valley is 99 days, 16 days shorter than the 115 day average for the Fairbanks
International Airport, AK (NRCC 2015), which is located only 100 mi (160 km)
south of the Arctic Circle.
Winters are considered moderate to severe; freezing or sub-freezing temperatures
occur on about 160 days. Zero-degree (-17.8 °C) minimum temperatures
occur about eight times a year. However, in most winters, northward incursions
of mild Gulf of Mexico or Atlantic Ocean air provide frequent breaks in the cold
weather. These breaks sometimes do not reach lower elevations to the east due to
Table 1. The climate of Blackwater Falls and Canaan Valley, WV, elevation = 3250 ft (991 m)
above sea level. T = trace, less than 0.1 inch; means are for the period 1961–1990; temperatures adjusted
to midnight–midnight observation time; extremes = all except rainfall 1945–1994; rainfall
1945–1964; mean number of days is estimated from published values for 1945–1964; data compiled
from official National Weather Service cooperative station data from Thomas and Canaan
Valley, WV. [Table is continued on next page.]
Temperature (ºF)
Extremes
Means Record Record Heating
Month Max Min Avg max Year min Year degree days
Jan 33.0 13.7 23.4 70 1950 -27 1984 1237
Feb 36.2 15.7 26.0 69 1985 -26 1963 1047
Mar 45.4 24.3 35.4 80 1954 -23 1960 877
Apr 55.9 32.2 44.0 85 1976 -1 1985 579
May 66.2 41.4 53.8 86 1979 14 1947 312
Jun 73.2 48.4 60.8 91 1952 23 1977 116
Jul 76.5 52.7 64.6 96 1988 27 1988 51
Aug 75.1 51.3 63.2 93 1948 25 1957 81
Sep 69.3 45.2 57.2 84 1953 18 1964 210
Oct 59.3 35.0 47.2 82 1951 5 1952 521
Nov 47.9 27.4 37.6 75 1958 -14 1956 780
Dec 37.9 18.9 28.4 76 1951 -20 1983 1101
Year 56.4 33.8 45.1 96 1988 -27 1984 6912
Southeastern Naturalist
.A. Vogel and R.J. Leffler
2015 Vol. 14, Special Issue 7
22
the shallow cold-air damming which occurs when northeast winds trap cold air
in the lower elevations east of the Eastern Continental Divide.
Precipitation
Precipitation averages about 55 in (139.7 cm) per year and is evenly distributed
throughout the seasons (Table 1). Drought conditions are seldom experienced
Table 1, continued.
Precipitation (inches)
Snow
Total precipitation Max on
Month Avg Max day Year Avg Max month Year ground
Jan 4.23 1.55 1955 31.3 69.0 1985 68
Feb 3.80 2.00 1948 29.5 80.5 1958 83
Mar 5.06 2.18 1954 21.8 73.0 1960 60
Apr 5.18 1.54 1952 10.7 30.0 1961 16
May 5.27 2.25 1960 0.3 5.0 1960 2
Jun 5.38 2.33 1949 0.0 T 1945 T
Jul 5.46 2.44 1958 0.0 0.0 - 0
Aug 4.91 3.50 1955 0.0 0.0 - 0
Sep 4.17 2.08 1964 0.0 T 1990 T
Oct 3.87 4.00 1954 2.5 14.0 1979 8
Nov 4.32 1.98 1962 12.1 37.0 1976 24
Dec 4.74 2.11 1948 25.3 67.0 1969 52
Year 56.39 4.00 1954 133.5 80.5 1958 83
Mean number of days
Temperatures
Max Min
Month Snowfall >1" Precip. ≥ 0.10" ≥90º ≤32º ≤32º ≤0º
Jan 12 11 0 14 28 5
Feb 11 12 0 8 25 3
Mar 6 13 0 5 25 1
Apr 2 12 0 1 17 0
May less than 1 12 0 0 6 0
Jun 0 10 0 0 1 0
Jul 0 11 less than 1 0 less than 1 0
Aug 0 8 0 0 1 0
Sep 0 7 0 0 4 0
Oct 1 7 0 1 14 0
Nov 4 9 0 3 21 less than 1
Dec 8 10 0 9 28 3
Year 44 122 less than 1 41 170 12
Southeastern Naturalist
C.A. Vogel and R.J. Leffler
2015 Vol. 14, Special Issue 7
23
but when drought occurs, the threat of forest fires increases, and stream flow
and well levels are low. Precipitation records from the Thomas station may be
more representative than the Canaan Valley Station because we believe the more
protected exposure of the rain gauge at the Thomas station may result in a
more accurate catch of both rain and snow. The Canaan Valley station, although
protected somewhat from high winds in the immediate vicinity of the rain gauge,
is located in the middle of the Valley where large open meadows allow the full
force of the strong winds that frequent the Highlands to blow unimpeded and
reduce the catch in the rain gauge.
Snowfall
The Valley is referred to by some as a snow bowl because it has long been noted
for its exceptionally heavy snowfalls. Snowfall on the Valley’s floor averages
11.2 ft (3.4 m) for the 30 years between 1961 and 1990. Snow has been observed
as early as September and as late as June. The lowest seasonal snowfall total for
that period was 5.5 ft (1.67 m) and the highest was 21.4 ft. (6.53 m) during the
winters of 1948–1949 and 1995–1996, respectively. The average maximum snow
depth reached is about 2.5 ft (0.76 m) in late February. However, due to the high
variability of conditions, depth and times of maximum snow-cover vary considerably.
Occasionally, a winter month occurs with a daily measurable snowfall of
≥0.1 in (0.25 cm).
Due to the frequent thaws that result from warm-air intrusions from the east
and south, snow cover is usually intermittent. Occasionally, in cold winters
that feature a persistent and strong northwest-wind flow, a snow pack up to 5 ft
(1.5 m) deep can develop and persist. Under such conditions, north-facing slopes
above 4000 ft (1219 m) in elevation can hold the pack into mid-April, and drifts
in high, windblown meadows can persist well into May.
Significant snowfalls of 4–8 in (10.1–20.3 cm) or more resulting from
northwest winds are common. These snowfalls, sometimes referred to by local
residents as lake flakes, are commonly the result of upslope movement and
cooling of air masses as they cross over the broad, elevated Allegheny Plateau.
Sometimes the northwest snows are enhanced by shallow-level moisture inputs
and streamers from the Great Lakes, but this is the exception rather than the rule.
With elevations on the Valley rim averaging about 4100 ft (1250 m), air is
forced to rise 3500 ft (1067 m) from the western Ohio River Valley bottomlands
130 mi (209.2 km) to the west. The ascent forces expansion, cooling, and increased
precipitation. Thus, when many areas at lower elevations have clearing
skies after cold frontal passages in northwest-wind flow, the Valley, the immediate
windward crest of the Alleghenies, and surrounding highlands can experience
persistent and heavy snowfall and blizzard conditions (winds greater than 35
mph. [56.3 km/h] for 3 hours or more, combined with snow and blowing snow
that reduces visibility to less than 1/4 mile [0.4 km]; Leffler 2005) .
Heavy snows can also result from Atlantic coastal storms (nor’easters) passing
along the coast. Under these conditions, moist easterly winds flowing off the
Southeastern Naturalist
.A. Vogel and R.J. Leffler
2015 Vol. 14, Special Issue 7
24
ocean are further lifted by the Allegheny Front (Eastern Continental divide). With
the forced ascent, the lift squeezes more precipitation from an already moisturesaturated
air mass. Under this scenario, the heaviest snowfalls occur over and just
to the east of the Allegheny Front (Dolly Sods, Roaring Plains, Spruce Mountain)
and towns just east of the front—Petersburg and Keyser. Some of the enhanced
snowfall from these northeasters spills over into the Valley on the stiff easterly
winds. When coastal storms move to the north of the Valley, cold air usually
begins flowing in and winds turn around to the northwest, adding more highelevation
upslope snow to the totals.
All of these factors combine to give the Valley an average snowfall of 134 inches
(340 cm), over one-foot greater than Rangeley, ME (121 inches [307 cm]), the
snowiest reporting station in Maine (Nexus 2015). The northwest wind flow and
upslope-enhanced lift also combine to produce an unusually high number of snow
days: 44 days with 1.0 in (2.54 cm) or more of snowfall (Table 1). By contrast,
Washington, DC experiences only four snow days in an average winter.
Severe weather
Thunderstorms occur in the Valley on about 50 days per year. Violent localized
winds from any direction may accompany intense summer thunderstorms.
Damaging tornadoes have been recorded only twice—in June 1944 and May
1948—though damaging winds can occur in any month. The high elevation and
preponderance of open meadows contribute to the increased frequency of damaging
winds and howling blizzards in the winter. In general, the strongest winds
precede the passage of low-pressure storms in the fall, winter, and spring, and
come from the westerly quadrants.
Dense fog with near-zero visibility is a frequent occurrence in the Valley,
especially during clear, calm conditions in the fall and during the moist summer
months. Dense fog is also common on the ridgetops surrounding the Valley,
when moist air is cooled to below the condensation point as it ascends the higher
elevations.
During the winter months, the Valley’s surrounding ridges are frequently
cloaked with a frost line. This beautiful line, above which all trees are coated
pure white, actually corresponds to the cloud base and is the result of rime icing.
Rime ice forms when super-cooled water vapor in clouds freezes upon contact
with below-freezing surfaces, forming a buildup of ice crystals.
Snowfall and Temperature Analyses
We examined the data presented above to determine long-term trends. The
questions we examined were: (1) Are there any significant statistical trends in
the climate records, or have they remained stable? and (2) Are there any lowfrequency
atmospheric phenomena which serve as predictors of seasonal climate
variabililty, especially in the winter? Here we summarize the results of our brief
analyses of temperature and precipitation records.
Southeastern Naturalist
C.A. Vogel and R.J. Leffler
2015 Vol. 14, Special Issue 7
25
The Valley area has seen some population decline during the period of analyses
with little anthropogenically induced change likely (Wikelopedia 2015).
Population growth and land-use patterns are expected to remain stable as nearly 94
percent of the land is is now owned by the federal and state governments (70 percent
of the Valley's lands are in national Canaan Valley Wildlife Refuge while the
WV state government owns an additional 24 percent in the Canaan Valley Resort
state park).Thus, the Valley is an excellent candidate for studying naturally driven
long-term climate variability and trends at higher elevations of the eastern US.
One situation that introduces data discontinuities and resulting biases into
long-term climate-variability/change studies is the relocation of measurement
sites. Even changes of tens of feet can produce apparent changes in climate
analyses. In the case of the Canaan Valley stations, a comparison of monthly
averaged temperatures during a period of overlapping data (January 1993–March
1996) showed a systematic difference between the two sites. A linear regression
fit through data representing COOP station 1 (S1) as a function of COOP station
2 (S2) gave the relation,
S1 = 5.63 + 0.933*S2,
with a correlation coefficient R2 = 0.997. Thus, in order to establish a continuous
temperature record through July of 2002, we used station 1 records through
March 1996; after March 1996, we adjusted station 2 records according to the
above linear relation to estimate their value as if they had been measured at
station 1.
A similar analysis on snowfall data yielded a non-linear relationship. However,
because we observed that a limited number of extraneous points significantly
altered the correlation, and because one would not expect to see a bias (assuming
the same observer) in precipitation values between two stations within 2 km of
each other and at the same elevation, we derived a continuous data set by using
station 2 data when station 1 data was missing.
Snowfall
Figure 1 shows seasonal (from 1 July to 30 June) snowfall totals for Canaan
Valley, as measured at the NWS COOP station. Over the 54 winter seasons available
from the digitized data record, the average snowfall was 133 in (338 cm)
with a standard deviation of 39 in (99 cm), not unlike the 1961–1990 averages.
However, snowfall amounts in this extended record show extremes of 69 in (175
cm) in the 1948–1949 season to 257 in (653 cm) during 1995–1996.
When observing time-series records such as these, it is useful to assess trends
by fitting an average curve to the data, keeping in mind the relatively short period
of record compared to the longer time intervals often associated with climate
change. Figure 1 shows a 30-y running-average fit to the data. Any point indicated
along the 30-y running-mean curve is an average of 15 years ahead and 15
years behind that point. The curve has a standard deviation of 2.4 in (6.1 cm; less
than 2% of the mean) and a slope of only 0.15 in (0.38 cm) per year. These can
Southeastern Naturalist
.A. Vogel and R.J. Leffler
2015 Vol. 14, Special Issue 7
26
be considered small deviations, and we found no significant trend in the snowfall
time series over the measurement period.
A further question we evaluated was whether large-scale phenomena such
as the El Niño Southern Oscillation (ENSO) correlate with climate variability
in the Valley. ENSO is known to affect global weather and climate (NOAA
2015). The ENSO index is defined as a 3-month average of sea surface-temperature
departures from normal for a critical region of the equatorial Pacific
(Nino 3.4 region: 120°W–170°W longitude, 5°N–5°S latitude). Any case in
which this region is characterized by a positive departure of >0.5 °C from normal
is termed El Niño. A case with a negative departure of >0.5 °C is called
La Niña. The index is termed neutral when sea surface temperature is within ±
0.5 °C of average. The particular state in which ENSO resides (La Niña, neutral,
or El Niño) and its intensity have implications for seasonal climates
across the globe, and can have significant impacts on weather and climate in
particular regions of the US (Patton et al. 2003). This situation suggests the
question: Can we use the phase of ENSO as a predictor for snowfall amounts
in the Valley?
Researchers have found significant correlations between ENSO phases and
seasonal snowfall amounts in some parts of the mid-Atlantic region. One consideration
is the length of the time lag between the establishment of an ENSO
state and the resulting effect (if any) on the Valley’s climate. In this analysis, we
Figure 1. Annual snowfall totals for Canaan Valley, WV
Southeastern Naturalist
C.A. Vogel and R.J. Leffler
2015 Vol. 14, Special Issue 7
27
decided to observe both the ENSO state during January–March of the observed
winter season, and the ENSO state during the previous summer.
Figure 2 illustrates the Valley’s snowfall record categorized according to the
phase of ENSO during January–March. The results of our analysis indicate that
when ENSO was in its neutral phase, the Valley experienced higher snowfall
amounts, but during the El Niño phase, slightly less snow was recorded. During
the La Niña phase, we would expect the least amount of snowfall with >10% below
average amounts. However, there is considerable variation in the data. Some
of the lowest snowfall amounts occurred during the neutral and El Niño phases,
while high amounts occurred during La Niña events.
Figure 3 shows the same data categorized according to ENSO state during
the previous summer; the differences in snowfall amounts are more pronounced,
with the ranking of ENSO state as it relates to snowfall amounts. These plots
seem to indicate that over the longer term, ENSO correlates with snowfall
amount. However, the relatively large amount of scatter in the data suggests that
other unidentified phenomena also influence climate variability.
Results of statistical tests (here P signifies a level of significance for a twotailed
Student’s-t analysis for differences in mean values; the lower the value,
the higher the significance) to determine whether the snowfall totals categorized
according to ENSO state are statistically different showed that when we used the
January–March ENSO state as a predictor (Fig. 2), only the difference between
Figure 2. Seasonal snowfall amounts categorized according to the ENSO state during
January, February, and March.
Southeastern Naturalist
.A. Vogel and R.J. Leffler
2015 Vol. 14, Special Issue 7
28
Figure 3. Snowfall amounts categorized according to the ENSO state during the previous
summer.
neutral and La Niña groups was significant (P ≤ 0.05). When we used ENSO state
during the previous summer as a predictor (Fig. 3), the difference between neutral
and La Niña totals was significant (P ≤ 0.0125), and the difference between El
Niño and La Niña snowfall amounts was significant (P ≤ 0.05). Although the correlations
are relatively weak, these findings suggest that the ENSO state recorded
during the summer prior to a winter season is a better predictor of snowfall
amounts than the ENSO state of the current season, especially when considering
differences between neutral and La Niña conditions. Differences between
snowfall amounts categorized according to neutral and El Niño states were not
significantly different for either method.
Temperature
We followed the same methodology used for snowfall to analyze temperature
records. Figure 4 shows the average annual temperature over the 54-year
period of record. The average is 46.5 °F (8.1 °C) with a standard deviation of
1.4 °F (0.8 °C). The highest average annual temperature was 49.5 °F (9.7 °C)
in 1991, and the lowest was 43.4 °F (6.3 °C ) in 1977. Unlike snowfall, which
shows no significant trend in its 30-year running mean over the past 15 years,
the 30-year running mean of temperature increases by almost a full degree
over the same period.
An ENSO analysis similar to that performed with the temperature records is
shown in Figure 5. Here, January–March average temperatures are associated
Southeastern Naturalist
C.A. Vogel and R.J. Leffler
2015 Vol. 14, Special Issue 7
29
Figure 4. Average annual temperature for the Canaan Valley, WV.
Figure 5. Average 3-month daily temperature during January, February, and March categorized
according to the ENSO state during the previous summer.
Southeastern Naturalist
.A. Vogel and R.J. Leffler
2015 Vol. 14, Special Issue 7
30
Figure 6. Average 3-month daily temperature during January, February, and March.
with ENSO states of the previous summer. Although variability is similar to what
we observed in our analyses of snowfall amounts, there is a weak dependence on
the ENSO state. Neutral phases correlate with the coldest periods, El Niño phases
with warmer periods, and La Niña phases with the warmest winters.
We conducted statistical tests similar to those performed for snowfall amounts.
The results for temperature using the ENSO state of the previous summer as a
predictor were similar to those found for snowfall using the same states, albeit
at a significantly higher level of confidence. The differences between neutral and
La Niña-grouped temperatures were significant with P ≤ 0.005, while the differences
between El Niño and La Niña were significant with P ≤ 0.025. We found
no significant difference between neutral and El Niño cases ( P ≥ 0.05).
It is also useful to compare summer and winter average temperatures. Figure 6
shows average temperatures during the January–March periods smoothed with
a 30-year running mean. The average 3-month temperature for the 54-year record
was 29.8 °F (-1.2 °C) with a standard deviation of 3.3 °F (1.8 °C). Figure 7
depicts average 3-month temperatures for June–August with a 30-year running
mean. The average temperature over the entire record was 63.9 °F (17.7 °C) with
a standard deviation of 1.5 °F (0.8 °C). As expected, variability in average temperatures
was greater in the winter months than in the summer. Both plots show
increases in 30-year mean temperatures over the last 15 years, although it is more
pronounced in the summer data. Also, notable in these data are the warm summer
of 1952 and the cold winter of 1978.
Southeastern Naturalist
C.A. Vogel and R.J. Leffler
2015 Vol. 14, Special Issue 7
31
Figure 7. Average 3-month daily temperature during June, July, and August.
Summary
In this limited analysis, the Valley’s snowfall and temperature records exhibit
a weak correlation to the ENSO phenomenon. The fact that extremes in snowfall
and temperature simultaneously run counter to the expected effects of the ENSO
state suggests other unknown factors play a dominant role.
In conclusion, we summarize the unique climate of the Valley. In more populated,
urban areas, hourly updates of wind speed, temperature, and humidity are
available and the general public has a greater awareness of weather events. In
the Valley, there are relatively few available real-time displays of local weather
data; only those who have witnessed first-hand the large, rapid swings in weather
and climate have a true appreciation for the uniqueness of the area. One can
often travel less than 5 mi (8 km) in any direction and experience a climate very
different from the Valley’s. From a climate perspective, the Valley remains a
particularly distinctive location, with its climate considered its “cash crop”
(Weedfall and Dickerson 1965) and its location indeed a “bit of Canada gone
astray” (Fortney 1977).
Acknowledgments
Grateful acknowledgement is made to the Thompson family for their decades of dedicated
volunteer service in all kinds of weather, some not fit for man nor beast. Thanks
are also extended to Ken Sturm for his recent years of service. This study could not have
Southeastern Naturalist
.A. Vogel and R.J. Leffler
2015 Vol. 14, Special Issue 7
32
been conducted without their outstanding volunteer public service in the field of weather
observation. We also acknowledge the National Climactic Data Center in Asheville,
NC, for providing the edited, published data used in our analyses. We extend thanks to
Robert O. Weedfall and W.H. Dickerson, authors of The Climate of the Canaan Valley
and Blackwater Falls State Park, West Virginia (1965). Their report provided a critical,
rich source of historical information on the Thompson family observations in the Valley.
Finally, thanks to Jocelyn Smith, Laura Brake, and Denise Webb of the Canaan Valley
Institute for their assistance in preparing the manuscript.
Literature Cited
Fortney, R.H., 1977. A bit of Canada gone astray. Wonderful West Virginia 41:24–31.
Leffler, R.J. 1974. Snowfall on the Allegheny Plateau of Maryland and West Virginia.
Weatherwise 10/1974; 27:199–201. DOI:10.1080/00431672.1974.9931708.
Leffler, R.J. 1977. A distinct precipitation shadow in the valley of the South Branch of the
Potomac River, West Virginia, National Weather Digest 2(4):21–24.
Leffler, R.J. 2005. Snowfall on the Allegheny Plateau of Maryland, Virginia, and West
Virginia. NWS Baltimore/Washington Forecast Office Winter Weather Workshop
Seminar.
National Oceanic and Atmospheric Administration (NOAA). 2015. NOAA's El Niño
portal. Available online at http://www.elnino.noaa.gov.
Nexus. 2015. Current results: Average annual snowfall in Maine. Available online at
http://www.currentresults.com/Weather/Maine/annual-snowfall.php.
Northeast Regional Climate Center (NRCC). 2015. CLIMOD dataset. Available online at
http://climod.nrcc.cornell.edu/climod/frost/.
Patton, J.M., S.R. Smith, and J.J. O’Brien, 2003. Impacts of ENSO on snowfall frequencies
in the United States. Monthly Weather Review 18:965–980.
Thornthwaite, C.W. 1948. An approach toward a rational classification of climate. Geographical
Review 38:55–94.
Weedfall, R.O., and W.H. Dickerson 1965. The climate of the Canaan Valley and Blackwater
Falls State Park, West Virginia, WV University Agricultural Experiment Station
Current Report 43.
Wikipedia. 2015. Davis, West Virginia. Available online at https://en.wikipedia.org/wiki/
Davis,_West_Virginia.