(1)
Yanqing Lian1, Tawei Soong2, and Misganaw Demissie1,
Gary Clark1, and Robert Dalton3
1 Illinois State Water Survey, 2204 Griffith Drive, Champaign, IL 61820, USA
2 US Geological Survey, Illinois District, 221 North Broadway Avenue, Urbana, Illinois 61801 USA
3 Illinois Department of Natural Resources, Lincoln Tower Plaza, 524 South Second Street, Springfield, IL 62701, USA
Abstract: The propagation of flood waves in the Lower Illinois River from Peoria to Grafton has been studied with the 1-D unsteady flow model, UNET. After being calibrated and verified, the model was applied to simulate water surface elevations for the existing conditions and for managing the LDD as storage options with design floods.
Spillways have been placed on top of the levees to simulate managed flow into selected Levee and Drainage District (LDD) storage areas along the Lower Part of the Illinois River. Three Levee and Drainage Districts of difference sizes and locations were selected to study the relationship between the flood stage reduction and the sizes of the opening inflow section of the spillways, namely the flood reduction versus the width of the spillway with fixed depth and the reduction of flood stage versus the depth of the section when the width is fixed. The optimal width and depth of the opening sections have been examined for the three selected LDDs. The study was based on the maximum water surface elevation of 100-year design flood. A total of 24 individual Levee and Drainage Districts and combinations of two or a few LDD’s have been studied for the reduction of flood stage on the Illinois River.
The lower section of the Illinois River has experienced increased flooding and frequent levee overtopping since the early part of this century. This increase in flooding is partially due to the construction of levee and drainage districts (LDD) on the floodplains, which resulted in the loss of approximately 57 percent of the floodplain for flood conveyance (Alvord and Burdick, 1919). Analysis have shown that flooding in the Lower Illinois River is governed by the backwater effects of the Mississippi River, floods from upstream of the Peoria Lock and Dam, floods from tributaries, and the combinations among them (Singh, 1996). The effects of LDD storage on flood peaks were observed during the 1993 Midwestern flood on the Upper Mississippi River. During the 1993 Flood, flood stages at Quincy, Illinois and Hannibal, Missouri on the Upper Mississippi River showed clear drops after levee breaches upstream breached (Bhowmik et al. 1995). Such drops meant significant flood protection for towns, cities, and LDDs downstream. However for that 1993 Flood event, due to its immense magnitude, the river stages came back after the LDDs were filled.
Existing levees in the Alton and La Grange Pools can protect most of the present levee and drainage districts against a 50-year flood; only a few levees in the La Grange Pool can provide protection against a 100-year flood. The conventional approach to increase the level of protection of the levee and drainage districts in the Alton Pool and the La Grange Pool is to raise the levees. But the drawback to this approach is the prohibitive cost and the danger in flood fighting. Even if a levee were raised, the additional cost of pumping to keep the water table low in the levee drainage district and the consequent reduction in crop yields would make farming in the area behind the levee marginally profitable.
An alternative approach is to reduce the flood elevations by admitting water into a few selected levee drainage districts during flooding events. By opening a limited section on the selected levee, these selected levee drainage districts can store part of the flood volumes when river stages exceed the elevation of the openings and contribute to a reduction in flood stages in the lower reach of the river. The effects from managing flood storage depend on the location and acreage of selected individual and combination of levee drainage districts as well as the opening section on each selected levee. When the flood stages in the Illinois River drop below the bottom of the opening, the floodwaters in the storage area can return gradually to the Illinois River. The impact of a levee and drainage district can provide to reduce the flood peak is limited by its size and location and as well as management practices. Studies have shown some levee and drainage districts would have only slight reduction to the flood stage rather this type of levee drainage district is more valuable for ecosystem restoration and habitats. The one-dimensional unsteady full dynamic flow model (UNET) was applied for this practice.
The UNET model is a one-dimensional unsteady flow model developed by Barkau (1995). It solves the one-dimensional full dynamic wave Saint-Venant equations.
(1)
and
(2)
where x is the distance along the channel, t is the time, Q is the flow, A is the cross-sectional area, h is the water depth, S is the storage volume per unit length in the direction of flow, Sf is the frictional slope, vl is the lateral inflow velocity, g is the gravitational acceleration, and ql is the lateral inflow per unit distance. Continuity and momentum equations were derived based on equations (1) and (2) for the flood plain and river channel. The flows in the river channel and floodplain were solved separately by assuming equal momentum exchange.
The UNET model can simulate one-dimensional flow through single, dendritic, or looped systems of open channels. It can also simulate the interaction between channel and floodplain flows, channel and storage areas; levee failures; and flow through navigation dams, gated spillways, weir overflow structures, bridges and culverts, and pumped diversions. The UNET model allows either stage or flow hydrographs to be the boundary conditions.
The model requires a stage or discharge condition at the upstream boundaries of the main river and tributaries that have existing cross-sectional data. Other input requirements include tributary inflows, cross-sectional geometry, and hydraulic roughness parameters. The cross-sectional geometry and the stage or discharge boundary condition are prepared in two separate files. The input time-series stage and discharge data are read from the Data Storage System (DSS) database that developed by the USCOE. The output of the UNET model includes the time-series of stage and discharge at prescribed locations and plots of water surface elevation profiles. The Lower Illinois River UNET model was setup for the portion from Peoria Lock and Dam to Grafton.
In the case of the lower reach of the Illinois River (Peoria Lock and Dam to Grafton), five gaged tributaries contribute significant flow to the Illinois River. The Sangamon River is the only tributary that has existing cross-sectional data. Because of the lack of cross-section geometry data on the other major tributaries, the lower Illinois River UNET model has been set up as a system of seven river reaches as shown in Figure 1 with three reaches for the Lower Illinois River system. In this figure, Reach 1 is the segment of the Illinois River from Peoria Dam to the river section immediately upstream of the Sangamon River junction. Reach 2 is the segment of the Sangamon River from the gage at Oakford to the river mouth, and Reach 3 is the segment of the Illinois River from the Sangamon River junction to the Illinois River mouth at Grafton.
The Mackinaw and Spoon Rivers are the major tributaries in Reach 1 while the La Moine River and Macoupin Creek are the major tributaries in Reach 3. These tributaries and the smaller ones are taken as lateral inflow at their confluence with the Illinois River. The Sangamon River (downstream of Oakford) was assumed to have no point lateral inflow from its tributaries. Instead, the contributions from its tributaries were assumed to be uniformly distributed lateral inflow along the entire length of the Sangamon River. Figure 1 also shows the location of the discharge/stage gages on the Illinois River at Peoria L&D, Kingston Mines, Havana and Beardstown in Reach 1; and La Grange L&D, Meredosia, Valley City, Florence, Pearl, Hardin, and Grafton in Reach 3. The model calibration of the model is to have simulated stage or discharge hydrographs math the observed stage and discharge hydrographs at those gaging stations.
The UNET model can be applied to predict existing conditions, historical floods, statistically significant flood events, and combinations of feasible scenarios. In order to ensure that the simulation results are reliable and closely represent actual events, the model has to be calibrated and verified with historic data. The water surface elevation and the flow computed by the model are adjusted to fit closely to observed data by gradually varying the channel roughness coefficient until the difference between computed and observed water surface elevation is below the specified level of error. The Manning roughness coefficient is expressed in the model in terms of the channel conveyance. Initial roughness coefficient values are defined in the cross-section input file and are then converted to conveyances in the computation by UNET model. The conveyances are updated during the model calibration by multiplying them by an adjustment factor. Since sub-critical flow is assumed in the UNET model, the calibration of the model will start from the downstream end of the study reach and progress in the upstream direction.
In order for the model to be applied for flood protection, highly ranked flood events were selected for calibration and verification. The May 1979 flood ranked as the fourth highest flow at Meredosia and the sixth highest flow at Kingston Mines and the March 1985 flood ranked as second and third at Meredosia and Kingston Mines, respectively. The model was then verified with the December 1982, June 1974, April 1973, and July 1993 flood events, which are ranked as third, fifth, seventh, and twelfth at Meredosia and as first, seventh, fourteenth, and eighteenth at Kingston Mines, respectively (Abi and Singh, 1997).
The model was applied to evaluate the impact on water surface elevation of the floods due to the conversion of selected Levee and Drainage Districts in the La Grange Pool and the Alton Pool for managed flood storage areas. The investigations were for 100-year design flood. The model was first run with design hydrographs to define the existing 100-year flood water surface elevation profile. The representative stage and flow design hydrographs were derived from the top six historic flood events of the gaged streams. The duration of each flood was selected as 20 days, 10 days before and 10 days after the flood peak occurrence. The existing maximum peak stage of 100-year design flood has been simulated from the Lower Illinois River UNET model.
A spillway with opening ranging from 150 to 600 meters in width and 0.6 to 2 meters in depth was placed on top of the levee at prescribed river locations to simulate the flow interaction between the river and the flood storage areas. It has been realized the optimal reduction of peak stage at the levee can only achieved if the flow into the levee is controlled or managed so that it occurs near the peak of the flood hydrograph. The Spring Lake, McGee and Lacey Levee and Drainage Districts (figure 2) were used to study the relation of peak stage reduction at the levee with the width and the depth of inflow section. Figure 3 illustrates the variation of flood peak stage with the depth of the inflow section of 300-meters in width. This figure indicates that the optimal opening depths vary for different Levee and Drainage Districts and at different locations; specifically the optimal depths are 0.6 meter, 1.8 meters and about 1.5 meters respectively for Spring Lake, Lacey Creek and McGee Creek LDDs respectively. Studies have also been carried out on investigating the impact of width of opening inflow section of fixed depth at 1.2 meter to the flood reduction and the optimal widths of inflow sections are 100 meters, 500 meters and 600 meters at Spring Lake, Lacey, and McGee LDDs respective; and 150 meters, 900 and 600 meters respectively for a fixed depth of 1.8 meters for those three LDDs (figure 4). Using multiple LDDs as temporal flood storage is expected to reduce the flood stage by (1) increasing the flood storage areas and (2) reduce the same flood peak stage at several selected location when the inflow sections are well designed. Figure 5 is a comparison of the maximum water surface elevation of 100-year design flood under existing condition with one of the simulated WSE profiles. The net gain and net loss of the flood storage areas were computed for 100-year design flood, the largest gain to loss ratio is obtained by admitting flood water into both Lacey Creek and McGee Creek Levee and Drainage Districts at their respective optimal inflow sections (figure 6). About 35% more of the drainage areas would have been protected by sacrificing a 12% of the Levee and Drainage District areas at Lacey and McGee.
The unsteady flow model, UNET, has been tested and validated for the lower Illinois River reach between Peoria L&D and Grafton. The model was applied to simulate the maximum water surface profiles of 100-year design flood under existing conditions, i.e. no managed storage options and under conditions with individual and combinations of two or a few levee and drainage districts converted into temporal flood storage areas. It has been found that the optimal size of inflow sections vary with the location, bottom elevation of the inflow section and the storage area of the selected LDD(s). The percentage of net gain and net loss of Levee Drainage District with manage temporal storage options have studied and was found that the largest gain to loss ratio of LDD against 100-year design flood would have been achieved by admitting flood water into both Lacey and McGee with optimal inflow section depths at 1.2 and 1.8 meters respectively while the width fixed at 300 meters.
On going investigations on the hydraulic modeling of the Lower Illinois River include simulations of real time and forecast flood events with levee breaching or raising option to provide strategies for flood fighting, study of re-connection of LDDs with the river main stem for ecosystem restorations, and etc.
References
Akanbi, A.A., Lian, Y.Q., and Soong, T.W. 1999. An Analysis on Managed Flood Storage Options for Selected Levees along the Lower Illinois River for Enhancing Flood Protection. Report No 4: Flood Storage Reservoirs and Flooding on the Lower Illinois River. Illinois State Water Survey, Contract Report 645: 86p.
Akanbi, A.A. and Singh, K.P. 1997. Managed Flood Storage Option for Selected Levees along the Lower Illinois River for Enhancing Flood Protection, Agriculture, Wetlands, and Recreation. Second Report: Validation of the UNET Model for the Lower Illinois River. Illinois State Water Survey, Contract Report 608: 110p.
Barkau, W. 1995. UNET, Unsteady Flow through a Network of Open Channels and Lakes. Short course at Rock Island District, Corps of Engineers, April 29-May 3, 1995.
Bhowmik, N.G. editor. 1995. The 1993 Flood on the Mississippi River in Illinois. Illinois State Water Survey, Miscellaneous Publication 151.
Federal Emergency Management Agency. 1987. National Flood Insurance Program and Related Regulations. Washington, DC.
Foster, J.E. 1977. Flowline Study, Mississippi and Illinois Rivers. U.S. Army Crops of Engineers, Waterway Experiment Station, Vicksburg, Mississippi. WES-Miss-Basin Model-31-5:67p.
Hall, B. R. 1991. Impact of Agricultural Levees on Flood Hazards. Department of the Army, Waterway Experiment Station, Corps of Engineers, 309 Halls Ferry Road, Vicksburg, Mississippi, 39180-6199.
Hydrologic Engineering Center. 1997. UNET Version 3.2. One-dimensional Unsteady Flow Through a Full Network of Open Channels. Water Resources Support Center, US Army Corps of Engineers, Hydrologic Engineering Center. 609 Second Street, Davis, CA 95616-4687.
Singh, K.P. 1996. Managed Flood Storage Option for Selected Levees along the Lower Illinois River for Enhancing Flood Protection, Agriculture, Wetlands, and Recreation. First Report: Stage and Flood Frequencies and the Mississippi Backwater Effects. Illinois State Water Survey, Contract Report 590: 138p.
Singh, K.P., and Ramamurthy, G.S. 1990. Changes in Climate and Hydrology of the Illinois River Basin and What They Portend for Levee Districts and Peoria Lake. Proceedings, American Water Resources Association Conference held in Peoria, Illinois:29-38.
Soong, D., and Lian, Y.Q., 2000, Management Strategies for Flood Protection in the Lower Illinois River, Phase I: Development of the Lower Illinois River-Pool 26 UNET Model (in press).
Thompson, J. 1989. Case Studies in Drainage and Levee District Formation and Development on the Floodplain of the Lower Illinois River, 1890s to 1930s. University of Illinois Water Resources Center Special Report 17. Urbana, IL. 61801.
U.S. Army Corps of Engineers. 1994. The Great Flood of 1993, Post-Flood Report. Upper Mississippi River and Lower Missouri River Basins; Appendix C. U.S. Army Corps of Engineers, St. Louis District.
Fig. 1 Schematics of the three reaches for the Lower Illinois River UNET model
Fig. 2 Levee and drainage districts of the lower Illinois River
Fig.3 Changes in water surface elevation with depth of
inflow section at
Spring Lake, Lacey, and McGee Creek levee and
drainage districts, width of opening is 300 meters
Fig.4 Reduction in water surface elevation with varying
width of
inflow section at Spring Lake, Lacey, and
McGee Creek levee and drainage districts
Fig.5 Peak stage for 100-year return period computed
from
UNET simulations and stage frequency analysis
Fig.6 Percent area of managed storage levee and
drainage district or LDD (Loss) and
additional LDD areas (Gain) protected against the 100-year flood