term average values can be determined. For Zhuk-
method using these four indices was developed
ova's sites a plot of the long-term average values
that rates the relative potential for significant (i.e.
of the normalized stage parameters exhibits a gen-
damaging) ice jams.
eral relationship which can then be used to devel-
Recently CRREL has been involved in several
op an expression useful in predicting ice-jam stag-
attempts to develop prediction schemes for ice
es. Within each basin or other subarea, the data
events based on historical ice data. Wuebben and
could be further analyzed to determine the prob-
Gagnon (1995), for example, developed a model
ability that a given increase in stage due to ice
based on an evaluation of historical discharge,
will occur, assuming that an ice jam occurs at a
stage and meteorological information for a reach
given discharge.
of the Missouri River near Williston, North Dako-
Galbraith (1981) developed a method for pre-
ta. Ranges of parameters associated with low and
dicting the occurrence of significant ice jams on
high potential for ice jams were identified, and a
the Saint John River in New Brunswick. His meth-
method to predict ice-jam occurrence within this
od was based on a review of data associated with
reach was proposed.
historical ice-jam formation. Three factors were
Winter parameters examined were AFDD, the
assumed to be important in the formation of ice
Julian date of the maximum AFDD, the maximum
jams: the characteristics of the ice (e.g. strength,
discharge, the Julian date of the maximum dis-
piece size, quantity), its movement and river chan-
charge, the difference in time between the maxi-
nel properties that restrict the movement of the
mum discharge and the maximum AFDD, the
ice. Meteorological factors were thought to have a
stage of the downstream reservoir (which con-
major impact on both the characteristics and move-
trols backwater levels affecting ice jamming), the
ment of ice, particularly the rate of snowmelt and
total snowfall, and the timing of the snowfall (i.e.
rainfall during the ice breakup period.
early or late in the season). Each parameter was
Galbraith's method requires the determination
weighted and assigned a positive or negative rat-
of indices for several meteorological parameters
ing factor according to its correlation with ice-
following the start of the breakup period. These
jam events.
include the number of days to reach an ATDD of
Wuebben's prediction method involves cal-
50, the rate of heat transfer over the breakup peri-
culating each parameter, determining the weight
od, the number of days to reach a snowmelt index
and rating factor, and summing positive and neg-
of 90 mm, and the accumulated precipitation in
ative weighting factors. The ratio of positive to
mm water equivalent. The rate of heat transfer
negative factors is used to predict the occurrence
can be calculated from the incident solar radia-
of a breakup ice jam. During the winter of 1993-
tion or hours of sunlight, incident atmospheric
94, when the method predicted a high ice-jam
radiation, air temperature, cloud cover, long-wave
potential, the most damaging ice-jam flood on
radiation emitted from the surface and its emis-
record occurred.
sivity, surface temperature, estimated conductive
heat transfer and wind speed. The snowmelt in-
Lower Platte River
dex is estimated from the mean daily wind speed,
predictive model
air temperature (if greater than 0C) and water
The approach taken in developing an initial
vapor pressure.
predictive model of ice-jam occurrence within the
From past records the meteorological indices
Lower Platte River basin study reach is similar to
were calculated for the years available and then
the methods above in that it relies on an analysis
ranked in ascending (length of breakup period,
of hydrological and meteorological data associat-
number of days to reach ATDD = 50 and snow-
ed with historical ice jams. The data presented in
melt index = 90) and descending (rate of heat
the previous section, which encompasses the en-
transfer and accumulated precipitation) order. In
tire study area, do provide some information on
general, high-ranking or extreme values of the
general trends and interrelationships between pa-
indices were associated with damaging ice jams.
rameters. However, for various reasons it is not
The time to ATDD = 50 and the snowmelt indices
possible to develop a model applicable to the en-
were the best predictors of damaging ice jams.
tire study area at this time. These reasons include
Because these terms are related, Galbraith sug-
uncertainties in the data, which might be rectified
gested that the snowmelt index might be the bet-
over time as archived USGS records and old news-
ter predictor. Accumulated precipitation was not
papers are searched. Also, the causes of ice-jam
useful in predicting the occurrence of ice jams. A
formation at the different locations vary, along
24
Previous Page
