Figure 16. Volume of ice in the jam retained by the cylindrical-pier
ICS as a function of river discharge. Transport losses reduce 10,000,000
ft3 pre-breakup ice supply to an initial jam volume of 7,000,000 ft3 at 2000
cfs. Ice melting, plus washouts through the ICS above 8000 cfs, reduce
ice-jam volume as discharge increases throughout an event.
as one-dimensional, steady, and gradually varied (spa-
at Mill Road and Leydecker Road. We input the 5-ft-
diameter 10-ft-tall 12-ft-gap cylindrical-pier ICS
tially), as did its predecessor HEC-2. At each cross sec-
tion, it can include an ice cover of known thickness or
using the HEC-RAS feature "multiple blocked obstruc-
solve for the thickness of an ice jam according to stan-
tions."
dard theory (U.S. Army 1998b). The latter is a steady-
All model runs used subcritical flow. The down-
state theory that treats the ice as a granular material with
stream rating curve derived from the output of the HEC-
no cohesion. The program solves the one-dimensional
2 deck (RS 308.00, U.S. Army 1986b). Several cali-
force balance for the jam, where the under-ice water
bration flows were used to compare computed water-
shear and jam self-weight are resisted by the shear
surface elevations (WSE's) with physical-model and
strength of ice at the banks. The ice jam floats on the
field data. Final runs were made using the 100-year (1%-
water and adds a rough top surface, resulting in water
exceedence) open-water flow of 15,600 cfs and ice-jam
levels that are much higher than for open water at the
flows of 400013,000 cfs.
same discharge. The user can constrain the jam to the
main channel (appropriate for treed floodplains) and
Model calibration
can set hydraulic roughness (Manning's n) of the ice to
Open-water WSE data without the ICS were avail-
a fixed value or allow it to vary with jam thickness.
able from three sources: 1) field measurements along
The program calculates the volume of ice in a jam,
the ICS site in 1984 at discharges of 783 and 3170 cfs,
allowing us to simulate a jam of known volume at each
2) a HEC-2 computed water-surface profile at 2000 cfs
discharge (upstream of which is open water).
(U.S. Army 1986b), and 3) a water-surface profile at
Input to the numerical model included 50 cross sec-
15,600 cfs from a recent flood insurance study (FEMA
tions along 3.6 miles of Cazenovia Creek, from 600 ft
1992). As shown in Figure 17, minor adjustment of
downstream of Mill Road to 100 ft downstream of Tran-
channel and over-bank roughness values along the reach
sit Road (see Fig. 15). The cross sections came from
gave good agreement between the calibration data and
three sources: 1) an HEC-2 deck called Cazpfe that used
the HEC-RAS model results at these four discharges.
surveyed cross sections from 1984, 2) closely spaced
Open-water WSE data with the ICS were available
surveyed cross sections obtained in 1998 to construct
from the physical model at discharges of 1550, 2900,
the physical model, and 3) supplemental cross sections
and 5800 cfs. We used these to adjust the contraction
determined from topographic maps (1 in. = 200 ft)
and expansion coefficients in the HEC-RAS model to
obtained from the Town of West Seneca. As far as pos-
calibrate the water-level change across the ICS (Fig.
sible, we checked these data for consistency in terms
18).
of vertical and horizontal alignment. The HEC-2 deck
No WSE field data were available for ice-jam con-
included the geometry of the two bridges in the reach,
ditions upstream of Mill Road, probably because this
16
TO CONTENTS
Previous Page
