1957 report by Bilello. The equations, which do
temperature, wind speed and precipitation are
not account for wind speed and which are appli-
characterized as high or low.
cable only to locations north of 6200′N at eleva-
tions less than 1500 ft (or approximately 450 m),
CENTRAL GERMANY
are as follows:
The snowfall data for Germany are relatively
sparse in the most readily available data sources
Nov: R (Nov)
= 0.0098Tavg (Nov) + 0.090
(at least for those commonly used by U.S.-based
DecFeb: R (month) = 0.010Tavg (month) + 0.016 (2)
researchers); for example, only 10 stations in the
Airfield Weather Summary database have com-
March: R(Mar)
= 0.0048Tavg (Mar) + 0.197
plete snow data (U.S. Navy, U.S. Air Force, Dept.
of Commerce 1992). However, there are numer-
where R (month) is the average monthly snow
ous stations where air temperature data are
density (g/cm3), and Tavg (month) is the average
available but no snow data. Bates and Bilello
monthly air temperature (C). These equations
(1988) compensated for this shortage of data by
yield the monthly estimated snow density values
matching data for the northeastern United States
shown in Table 5.
with that for Germany. Snow and temperature
Table 5. Estimated snow density (g/cm3) at
conditions are quite comparable in these two ar-
various Alaska sites calculated using the
eas, if the analysis is limited to those areas with a
mean January temperature ≥ 8C and an eleva-
equations of Billelo (1957).
tion < 600 m.
Nov
Dec
Jan
Feb
Mar
A linear regression of 45 U.S. stations and 10
Anchorage
0.144
0.105
0.110
0.094
0.216
German stations (Bates and Bilello 1988) yields
Barrow
0.264
0.249
0.260
0.283
0.317
the following relationship between average Jan-
Bethel
0.139
0.116
0.122
0.122
0.226
uary temperature and annual total snowfall:
Big Delta
0.226
0.210
0.216
0.183
0.248
Fairbanks
0.242
0.227
0.244
0.210
0.253
Kotzebue
0.215
0.194
0.194
0.210
0.280
S(Total) = 54.0 17.6 T(Jan)
(3)
A later study by Bilello (1984) examined snow
where S(Total) is the total annual snowfall (cm),
cover properties in the former Soviet Union by
and T(Jan) is the average January temperature
analyzing published data from 41 locations
(C).
where both snow density and wind speed were
Snow depth data were available only for the
measured. The Soviet data proved to be consis-
U.S. cities (Bates and Bilello 1988). When the val-
tent with the North American data in Bilello's
ues for the average maximum snow depth (dur-
earlier studies, once an adjustment was made for
ing an average winter) were regressed against
the Soviet densitometers, which systematically
the January temperature, the following relation-
gave readings about 25% lower than those ob-
ship was derived:
tained by the equivalent measurement instru-
ments used in the U.S.
D(max) = 14.1 + 5.34 T(Jan)
(4)
A recent study by Sturm et al. (in press) pro-
posed a scheme for classifying snow cover into
where D(max) is the average maximum snow
seven categories. Data collected in the field in
depth (cm).
Alaska showed that snow density was the most
Bates and Bilello's (1988) work was applied to
powerful statistical criterion for distinguishing
the Hunfelds quadrangle, which is a roughly 22-
22-km area just west of the border between the
between various types of snow. The goal of
Sturm's work is to develop techniques for distin-
former East Germany and West Germany. The
guishing between these classes of snow without
Hunfelds area has been intensively studied by
an extensive field measurement program, using
WES and other Corps of Engineers laboratories
readily available information such as satellite
to determine the environmental factors that af-
imagery and climatic data. Five of these classes
fect Army materiel. The estimated January tem-
(excluding ephemeral snow, which is found in
peratures and other climatic parameters were de-
areas with intermittent snow cover, and moun-
termined for every grid square within the quad-
tain snow, which is found in mountainous areas
rangle (at 30-m resolution), and these estimated
and other cold areas with highly variable snow
parameters were used to estimate the snowfall
cover) are defined by a binary scheme whereby
and snow depth (Fig. 1).
3
Previous Page
