its, thickness is taken as the measured or estimated
Using a modified approach, he computed 0.004 <
thickness of the ice cover, whereas for accumula-
ni < 0.013 and 0.015 < nc < 0.022 for measurements
tions made up of broken sheet ice, thickness is that
made during 1966.
of the parent ice material. He specifically states
Michel (1980) states that the roughness of
that these data are applicable to initial ice cover
breakup jams will be primarily a function of the
formation and not to breakup jams. Because
shape and size of the floes making up the jam. For
the data underlying this figure were carefully
example, he points out that jams made up largely
of small floes will have smaller roughness than
collected and the sample size was large (n = 368),
those made up of large floes. Although Michel dis-
they provide a valuable resource when selecting
putes the use of the relationship suggested by
roughness values at the time of initial ice cover
Nezhikhovskiy, in fact his reliance on floe shape
formation.
Beltaos (1981) calculated ni for a January 1980
and size is related to Nezhikhovskiy's reliance on
frazil ice jam on the Thames River, New
parent ice thickness. Michel suggests that in the
Brunswick, using estimated discharge and thick-
case of breakup jams that are decaying, the rough-
ness. Like Nezhikhovskiy, he observed that the
ness will decrease with decay. Using his own for-
initial roughness was the highest and that smooth-
mulation, he determined values of nc for ten loca-
ing occurred over time, also three days in this case.
tions on three Canadian Rivers ranging from 0.053
Although four methods of estimating thickness
to 0.142.
and roughness were presented, two were consid-
Using an approach combining elements of the
ered plausible. The values of ni ranged from 0.040
Nezhikhovskiy approach, Beltaos (1978) deter-
to 0.016 using equilibrium ice jam criteria and from
mined nc and ni for two breakup ice jams on the
0.038 to 0.013 using Nezhikhovskiy's relationship
Smoky and Wapiti Rivers, Alberta. The ice rough-
between ni and thickness. At the time of initial
ness, ni, was found to be constant and equal to
freezeup in 1981, Beltaos (1983) estimated that
0.10, while nc varied from 0.090 to 0.109. Using a
roughness was between 0.033 and 0.041, depend-
similar approach, Andres (1980) estimated 0.057
< ni < 0.065 (ni = 0.060) and 0.041 < nc < 0.046 for
ing on thickness.
Measurements on the Vistula River, Poland,
the April 1978 Athabasca River (Alberta) breakup
during January 1982 allowed Majewski and Grzes
ice jam based on measured discharge, slope, and
(1986) to determine roughness values for a frazil
stage. When data from this jam were combined
ice jam. Over the course of 20 days, they found
with 1977 and 1978 data, Andres and Doyle (1984)
0.020 < ni < 0.15, with the roughest sections asso-
calculated an average ni = 0.072.
ciated with shoved areas. Estimates based on field
The shear stress approach was used by Knowles
and Hodgins (1980) to estimate ni and the Belokon-
measurements made in 1984 and 1985 ranged from
ni = 0.012 for sheet ice to ni = 0.080 for frozen ice
Sabaneev equation to calculate nc for two breakup
floes mixed with frazil. They also noted the de-
jams on the Thames River, New Brunswick. They
crease in roughness over time.
found that ni ranged from 0.01 to 0.015 and de-
Carey (1966, 1967) made careful observations
pended on the bed roughness and the ratio of jam
of the sheet ice cover near a USGS gage on the St.
thickness to depth.
Croix River, Wisconsin, during three winters
(196466). He noted dunes and ripples on the un-
Roughness estimated during numerical
derside of the ice, and observed that the wave-
model calibration
length and amplitude of the dunes and ripples
Vogel and Root (1981) modeled breakup jams
on the Missisquoi River (Vermont) using ni = 0.057.
near the bank were greater than for ice taken from
Rivard et al. (1984) used HEC-2 to model a May
midstream. In contrast to Nezhikhovskiy's work
with frazil deposits and accumulations of frazil
1983 breakup ice jam on the Mackenzie River. The
and sheet ice, Carey found that roughness for sheet
roughness (ni) of the intact ice sheet was assumed
ice covers tends to increase after ice cover forma-
to be 0.020. The best fit with measured water sur-
tion. Using his own formulation for calculating ni
face profiles was found with ni = 0.045 within the
and nc, he reported ni between 0.010 and 0.028 and
jam. Similarly, roughness values between 0.044
nc between 0.018 and 0.027 during 1964 and 1965,
and 0.090 provided a good fit in modeling ob-
with lower values reported earlier in the winter
served jam stages on the Aroostook River (White
(Carey 1966). Later, Carey (1967) indicated that
and Acone 1998). Tuthill and White (1997) used ni
these values may have been underreported be-
= 0.060 to model a breakup jam on the Salmon
cause of an incorrect assumption in his method.
River, Connecticut. The same value was report-
11
Previous Page
