CALIBRATION OF A SANDUSKY RIVER HYDRODYNAMIC MODEL

1
Monitoring Sediment Transport and Geomorphology
Change Associated with Dam Removal Fang Cheng,
Timothy Granata Department of Civil and
Environmental Engineering and Geodetic Science,
The Ohio State University
Introduction Sediment released after dam removal
can cause significant changes in river
morphology, flow, aquatic habitats and even in
river ecosystem processes both downstream and
upstream. However, quantitative study of sediment
transport associated with dam removal is
surprisingly insufficient. Knowing the dynamics
of sediment transport in reservoirs is the key to
understanding the influences of dam removal on
river ecosystems.

• Immediately after dam removal, reservoir was
  dewatered and water depth drop about 1m (Figure
  13-14), while downstream water level rose 0.6m.
• As we expected, downstream water turbidity
  suddenly increased 3 times within the first two
  hours after the removal and hit the peak 3 hours
  later (Figure 15).
• The base level of turbidity increased 15 after
  the dam was removed. The other peak as shown in
  Figure 15 on Nov. 21st was because of a small
  rain storm. Compared with the rainfall event,
  dam removal seems did not have significant
  short-term influence on downstream suspended
  sediment concentration.

• The pebble count method (Wolman, 1954) was
  employed for cross sections with coarse materials
  (diametergt128mm).
• Bedload traps were installed at downstream and
  upstream to measure bedload transport rates
  (Figure 8).

• Objectives
• Quantitatively monitor geomorphologic change in
  river channel associated with dam removal
• Quantify bed load transport rate caused by dam
  removal
• Develop a numerical model to simulate graded
  sediment transport

• Time series of turbidity were measured at two
  locations, one at the west bank 100 meter
  upstream and the other one at the west bank 200
  meter downstream from the dam using YSI 6600
  Sonde sensors
• Mean velocity was measured in shallow (lt 0.6 m)
  depths using a handheld Sontek FlowTracker ADV
• In deeper areas (depth gt0.6 m) a BT-ADP (3 MHz)
  was used (Figure 9).

Site Description Sandusky River (Figure 1, 2),
drainage area of 3637 km2, flows north into
Sandusky Bay at Lake Erie. St. Johns dam, 46
meter long and 2.2 meter high, was located on
Sandusky River at river mile 50.
Preliminary Results Processed cross sections are
plotted as in Figure 10. The Purple dots are GPS
points surveyed on Nov. 11th, 2003, a week before
the dam was removed, and the green dots are GPS
points surveyed on Apr. 23rd, 2004. The river
bed surface is generated by creating TIN and
Kriging interpolation.

• Downstream and upstream water turbidity were
  measured from March 17, 2003 to March 28, 2003.
  (Figure 16).
• Notching the dam did not have influence on
  short-term turbidity.

• As we expected, reservoir water turbidity was
  lower than downstream turbidity within the first
  three days after complete notching west side of
  the dam. However, this relation dramatically
  changed three days later when a small rainfall
  event occurred.
• After a small rain storm, reservoir suspended
  sediment concentration increased faster than
  downstream and became 60 higher than downstream
  on the 5th day.

Figure 10 River bed surface at downstream 75m
from the dam

• Reservoir Length 10km, slope0.01
• Reservoir impoundment 0.59 km2, and the storage
  is 561, 234 m3
• The sediment accumulated in the reservoir is
  composed of gravel, cobble, and sand
• On March 18th, 2003, the west side breach was
  notched down to the bed (Figure 3,4)
• Dam was removed on Nov. 17th, 2003 (Figure 5-6)

• Simulation
• Sediment transport models, Van Rijn, and Meyer
  Peter and Muller, will be used to calculate bed
  load and suspended load.
• These models will be compared for the event of
  St. John's dam removal using software MIKE 11
  (see right figure).
• A two-fraction model will be applied to calculate
  bed load as a comparison of the MIKE 11
  simulation results.
• The two-fraction model, sand and gravel,
  considers sand and gravel as a mixture instead of
  individual and predicts more accurate results.

Dam Breach Mar., 03
Figure 11 River bed surface change at downstream
75m from the dam
Figure 11 River bed surface change at downstream
75m from the dam

• Conclusion
• Study sediment transport of St. John's dam
  removal helps us to fully understand how river
  channel response for disturbance.

• The elevation change at 75m downstream from the
  dam was calculated by differentiate the averaged
  elevation within 1 meter grid (Figure 11).
• The maximum, minimum elevation change is 0.155m,
  and -0.23m, respectively.
• The mean elevation change is -0.02m, and standard
  deviation is 0.1m
• It indicates elevation did not have significant
  change at 75m downstream of the dam within the
  half year.

Dam Removal Nov., 03

• Detailed pre- and post-dam monitoring is
  necessary to determine the geomorphologic changes
  in river channel, and the consequential effects
  on river ecosystem.
• Quantitatively monitoring changes in river-bed
  elevation and sediment distribution provides
  foundation for numerical modeling channel
  response to dam removal.

Future Work River cross sections will be
surveyed at least twice at downstream and
upstream to avoid survey error. Bed load and
suspended load will be measured at the same
locations for comparison. Same survey will be
repeated at same sites after dam removal. Some
empirical sediment transport model will be
compared , and we will develop a new numerical
model to simulate dam removal event.

• Methodology
• Transects were each surveyed at the reservoir,
  and 75 meters, 545 meters and 3km downstream from
  the dam.
• Each series of transects has at least five cross
  sections with about two meters apart. This
  provided a baseline for the changes caused by dam
  removal.
• Elevation data were collected at 1Hz using a
  Trimble 5700 Receiver (Figure 7)
• Changes in river channel width and elevation
  resulting from dam removal is determined by
  differencing data of pre- and post- removal.
• For each series of cross sections, bed material
  distributions at the surface were collected using
  bulk sampling approach.

• Downstream and upstream surface sediment size
  distribution are plotted in the figure above.
• We can see that the D50 of downstream bed
  material is 1.2mm, 29mm, 50mm, 70mm and 82mm at
  stations of 24 km, 12.5 km upstream from the dam
  and 75 m, 545 m and 3 km downstream from the dam.
  
• It is interesting to notice the surface material
  size coarsening pattern. Most previous studies
  had shown downstream fining pattern.
• At least 60 of the transported particles are
  smaller than 2mm, which indicates that fine sands
  are mostly entrained rather than gravels.

Reference Ferguson, R.T., and C. Paola, 1997,
Bias and precision of percentiles of bulk grain
size distributions, Earth Surface Processes and
Landforms, 22 1061-1077 Wolman, M.G., 1954, A
method of sampling coarse river-bed material, EOS
Trans. AGU, 35(6) 951-956 Wilcock, P.R., and
Kenworthy, S.T., 2002, A two-fraction model for
the transport of sand/gravel mixtures, Water
Resources Research, 38(10) 1194-1205
Acknowledgements We thank Matthew Nechvatal,
Dan Gillenwater, Bryan Arvai, and Lauren Glockner
at OSU and Ryan Murphy and Constance Livchak at
the ODNR Lake Erie Geology Group for helping
field work. We also appreciate Yudan Yi for
helping GPS data processing.
Figure 7. Surveying with Trimble 5700 receiver

• The size distribution of transported sediments at
  station 75m and 200m downstream and 12.5 km
  upstream was plotted in the figure above.

For questions, please contact Fang Cheng,
cheng.206_at_osu.edu