miles long. Further, if discharge continued to in-
As shown in Table 4, factors examined included
crease after a jam was formed, a discharge magni-
freezing degree-days, snowfall, breakup period
tude would eventually be reached where a stable
water discharge Qb and Garrison stage. Freezing
jam would no longer be possible, the jam would
degree-days and snowfall are used to reflect the
release and the ice would pass downstream. Stag-
thickness and strength of ice on the river. Freezing
es would then return to the open-water curve. As
degree-days can be used in a relatively simple
a result, extrapolation of ice-induced stages to
equation to predict ice thickness:
more extreme events is generally not reliable,
h = c (FDD)0.5
particularly if such limiting factors as floodplain
flow and jam release cannot be defined.
The jam release discharge remains unclear, as
where h = calculated solid ice thickness
it depends not only on water discharge but also
FDD = accumulated degree-days of freezing
(F)
on the quantity of ice and the strength of the ice
accumulation. In 1986 we know that two jams re-
c = empirical constant to account for wind
leased with discharges no higher than about
exposure and snow cover.
60,000 cfs. Five of the six known jams occurred
This constant was calibrated for the Williston area
with discharges between 30,000 and 78,000 cfs.
during the essentially snowless winter of 1991-92
The 1952 event had a mean daily flow of 120,000
to a value of 0.60. Using this value for the winter of
cfs and a peak discharge of 170,000 cfs in the same
1952 resulted in a predicted ice thickness of 31 in.
time frame as the jam event, but it is unclear
The measured ice thickness in 1952 was 32 in., and
whether the jam was still in place or had (more
there was about 10 in. of snow on the ground one
likely) released prior to that peak flow. Experi-
week prior to breakup (USDOC 1953). For winters
ence and data from other rivers indicate that the
with greater snowfall this value might overpredict
maximum discharge during an ice jam event is
ice thickness and thus may be considered conser-
usually no more than a two-year open-water
vative.
flow, which in this area would be 90,000 cfs. A
The next term in the table (FDDmax) is the water
ten-year breakup period discharge is approxi-
year Julian day (days since October 1st) when the
mately 92,000 cfs. It is suggested here that the
FDD term began to decrease, taken as an indicator
maximum ice jam discharge should lie at or be-
of net melting of the snowice cover. Low values
low this ten-year breakup period flow. Beyond
of this term often indicate that the weather
that level, ice jams should become unstable, and
warmed gradually in the spring and that signifi-
the water levels would return to nearly open-wa-
cant ice deterioration and melting may have oc-
curred prior to any large increase in runoff. The
ter levels.
term in the next column (Qmax) indicates the esti-
mated date of the spring runoff causing breakup
Ice jam potential
in a given year. Lacking direct field observations
Although the prediction of ice jam occurrence
of ice breakup for the period of record, this term
and severity is still beyond the state of the art, it is
had to be estimated from discharge records, an
sometimes possible to rate the likelihood of ice
approach that leaves some uncertainty. The date
jams based on historical data. Such a prediction
of the maximum river discharge, Qmax, when
mechanism could prove useful in estimating the
compared to FDDmax, can be used to reflect the ar-
potential for ice jam formation in a given year,
rival of significant spring runoff relative to warm
both for early warning of potential flooding and
weather in the District.
for determining whether advance measures to
In the Williston area the local weather that gov-
limit ice-related flood damage are advisable.
erns the thickness and strength of the river ice is
Based on interviews with local residents and a re-
relatively uncoupled from the weather in the
view of literature, six historic ice jam events were
mountains that governs the snowmelt runoff that
identified in the BufordTrenton area (1952, 1972,
produces most breakup discharges. If the weather
1975, 1976, 1978 and 1986). To review the winter
warms up in the BufordTrenton area well in ad-
characteristics leading up to significant ice jam
vance of snowmelt runoff in the mountains, the
events, weather and hydrologic data from 1970
ice is likely to be too thin or rotten to pose a signif-
through the present were reviewed. The year pre-
icant flood threat. On the other hand, an increase
ceding the closure of Garrison Reservoir, the ice
in runoff from upstream early in the winter may
jam year of 1952 and the randomly selected year
encounter a strong, resistant ice cover requiring a
of 1960 were also included.
14
Previous Page
