In cooperation with the Ohio River Valley Water Sanitation Commission
Calibration and Validation of a Two Dimensional Hydrodynamic Model of the Ohio River, Jefferson County, Kentucky
Water-Resources Investigations Report 01-4091
U.S. Department of the Interior U.S. Geological Survey U.S. Department of the Interior U.S. Geological Survey
Calibration and Validation of a Two Dimensional Hydrodynamic Model of the Ohio River, Jefferson County, Kentucky
By Chad R. Wagner and David S. Mueller
Water-Resources Investigations Report 01-4091
In cooperation with the Ohio River Valley Water Sanitation Commission
Louisville, Kentucky 2001 U.S. DEPARTMENT OF THE INTERIOR GALE A. NORTON, Secretary
U.S. GEOLOGICAL SURVEY Charles G. Groat, Director
For additional information write to: Copies of this report can be purchased from:
District Chief, Kentucky District U.S. Geological Survey U.S. Geological Survey Information Services 9818 Bluegrass Parkway Box 25286 Louisville, KY 40299-1906 Denver, CO 80225-0286 http://ky. water.usgs.gov CONTENTS
Abstract...... 1 Introduction...... 1 Background...... 1 Purpose and scope ...... 2 Study area ...... ~...... 2 Model calibration and validation ...... 4 Model description...... 4 Field data collection and interpretation ...... 4 Water-surface elevations...... 8 Velocity and discharge...... 8 Bathymetry ...... 8 Upstream river reach...... 9 Computational-mesh configuration ...... · 9 Boundary conditions...... 14 Calibration and validation results ...... 14 Downstream river reach...... 15 Computational-mesh configuration ...... 15 Boundary conditions...... 21 Calibration and validation results ...... 21 Summary and conclusions...... 32 References cited...... 33
FIGURES
1-4. Maps showing: 1. Location of study area near Louisville, Kentucky ...... 3 2. McAlpine Locks and Dam configuration and U.S. Geological Survey (USGS) gaging station locations near Louisville, Kentucky ...... 5 3. Location of hydrographic-survey cross sections, surveyed water-surface elevation stations, and USGS gaging stations in the upstream study reach near Louisville, Kentucky...... 6 4. Location of hydrographic-survey cross sections, surveyed water-surface elevation stations, and USGS gaging stations in the downstream study reach near Louisville, Kentucky...... 7 5-21. Graphs showing: 5. Surface Water Modeling System (SMS) software triangulation of raw and straight-line bathymetry data for a section of the upstream study reach in the Ohio River model simulation near Louisville, Kentucky ...... 10 6. Mesh configuration around the earthen dike adjacent to Louisville, Kentucky, and upstream of McAlpine Locks and Dam in the Ohio River model simulation ...... 11 7. Mesh configurations around Six- and Twelve-Mile Islands in the upstream Ohio River study reach ...... 12
Contents Ill 8. Upstream reach mesh configuration around McAlpine Locks and Dam near Louisville, Kentucky ...... 13 9. Field-measured and model-simulated velocity profiles for cross-section 5, . 13 miles upstream from McAlpine Locks and Dam ...... 16 10. Field-measured and model-simulated low-flow, distance-weighted average cross-sectional velocities in the upstream study reach from the Ohio River model simulation ...... 17 11. Field-measured and model-simulated high-flow, distance-weighted average cross-sectional velocities in the upstream study reach from the Ohio River model simulation ...... 18 12. Mesh configuration in the downstream study reach around the tailwater area of McAlpine Locks and Dam ...... 20 13. Stage-discharge curve for McAlpine tailwater gaging station near Louisville, Kentucky, 1990-2000 ...... 22 14. Stage-discharge curve at Kosmosdale gaging station near Louisville, Kentucky, 1995-2000...... 23 15. Discharge-slope curve between McAlpine tailwater and Kosmosdale gaging stations near Louisville, Kentucky, 1995-2000 ...... 24 16. Field-measured and model-simulated low-flow velocity profiles for cross-sections 19 and 25 in the Ohio River model simulation...... 26 17. Scatter plot of the cross-sectional area and average velocity differences between field measurements and model-simulated results for the 18 downstream Ohio River cross sections near Louisville, Kentucky ...... 28 18. Comparison of cross-section 24 bathymetry in the downstream Ohio River study reach ...... 29 19. Comparison of cross-section 20 bathymetry in the downstream Ohio River study reach ...... 29 20. Low-flow distance-weighted average cross-sectional velocity profile for the downstream study reach in the Ohio River model simulation ...... 30 21. High-flow distance-weighted average cross-sectional velocity profile for the downstream study reach in the Ohio River model simulation...... 31
TABLES
1. Summary of water-surface elevation calibration and validation for the upstream study reach in the Ohio River model simulation near Louisville, Kentucky ...... 14 2. Summary of flow-split calibration and validation for islands in the upstream study reach in the Ohio River model simulation ...... 19 3. Summary of continuity checks in the upstream Ohio River model simulation...... 19 4. Evaluation of the Research Management Associates-2 (RMA-2) simulation representation of the Ohio River rating developed for the downstream study reach near Louisville, Kentucky ...... 25 5. Summary of water-surface elevation calibration and validation for the downstream study reach in the Ohio River model simulation ...... 25 6. Summary of continuity checks in the downstream Ohio River model simulation...... 32
lv Contents CONVERSION FACTORS AND VERTICAL DATUM
CONVERSION FACTORS
Multiply By To obtain
foot (ft) 0.3048 meter
foot per second (ftls) 0.3048 meter per second
cubic foot per second (ft3/s) 0.02832 cubic meter per second
mile (mi) 1.609 kilometer
VERTICAL DATUM
Sea level: In this report "sea level" refers to the National Geodetic Vertical Datum of 1929 (NGVD of 1929)-a geodetic datum derived from a general adjustment of the first-order level nets of both the United States and Canada, formerly called Sea Level Datum of 1929.
Elevation, as used in this report, refers to the distance above or below sea level.
Contents v
Calibration and Validation of a Two-Dimensional Hydrodynamic Model of the Ohio River, Jefferson County, Kentucky
By Chad R. Wagner and David S. Mueller
Abstract INTRODUCTION
The quantification of current patterns is Combined sewer overflows (CSO's), sanitary an essential component of a Water Quality sewer overflows (SSO's), nonpoint sources, and Analysis Simulation Program (WASP) storm-water runoff are all sources of wet-weather application in a riverine environment. The pollution that contribute to the degradation of our U.S. Geological Survey (USGS) provided a Nation's vital water resources. In situations where more than one of these sources contributes to the field validated two-dimensional Resource impairment of water quality, a holistic approach Management Associates-2 (RMA-2) defining the relative importance of each source's hydrodynamic model capable of quantifying the contribution to the problem is the most efficient way steady-flow patterns in the Ohio River extending to arrive at an economical control plan. In order to from river mile 590 to 630 for the Ohio River determine the relative significance of the various Valley Water Sanitation Commission sources of wet-weather pollution on major (ORSANCO) water-quality modeling efforts on waterways, an understanding of the local that reach. Because of the hydrodynamic hydrodynamics,pollutantlandloading,and . complexities induced by McAlpine Locks and pollutant-transport characteristics is essential. Dam (Ohio River mile 607), the model was split into two segments: an upstream reach, which extended from the dam upstream to the upper Background terminus of the study reach at Ohio River mile 590; and a downstream reach, which The Ohio River Valley Water Sanitation extended from the dam downstream to a lower Commission (ORSANCO) is nearing completion of a national demonstration study to develop a water terminus at Ohio River mile 636. quality model that is capable of simulating pollutant The model was calibrated to a low-flow concentrations in the river during wet-weather hydraulic survey (approximately 35,000 cubic periods and is able to quantify improvements to the feet per second (ft3/s)) and verified with data water quality resulting from pollution control best collected during a high-flow survey management practices. The initial project was (approximately 390,000 ft3/s). The model completed on a reach of the Ohio River, which calibration and validation process included included the Cincinnati/northern Kentucky matching water-surface elevations at metropolitan area. The Water Quality Analysis Simulation Program (WASP) was utilized to 10 locations and velocity profiles at 30 cross simulate the water-quality constituent transport and sections throughout the study reach. Based on transformation during low-flow periods spanning the calibration and validation results, the model the recreational-contact period May-September. The is a representative simulation of the Ohio River flow field required by WASP was simulated by use steady-flow patterns below discharges of of a two-dimensional Resource Management approximately 400,000 f~/s. Associates-2 (RMA-2) hydrodynamic model.
INTRODUCTION 1 In order to demonstrate the transferability of used to calibrate and validate the model with field this project to other riverfront metropolitan areas, a data as well as the results of the model simulations similar study was initiated in the Louisville, are discussed in this report. Kentucky, metropolitan area; this study involved both hydrodynamic and water-quality modeling and field data collection. The U.S. Geological Survey Study Area (USGS), in cooperation with ORSANCO, developed a field-validated hydrodynamic model to The Ohio River study area begins at the support the water-quality model being developed by northeast comer of Jefferson County, Ky., and a consultant contracted by ORSANCO. extends southwestward through the county and ends 6 river miles downstream from the mouth of the Salt River (fig. 1). The upstream reach is in the Purpose and Scope McAlpine Locks and Dam pool and the water level is kept at an elevation of approximately 420 ft above This report describes the calibration and sea level during low-flow periods. The .downstream validation of the two-dimensional RMA-2 model to reach of the study area is in the tailwater of simulate the complex hydrodynamics of a 40-mi McAlpine Locks and Dam. The water level at the study reach of the Ohio River near Louisville, Ky. dam is held at a normal pool level of 383 ft above sea level by the Cannelton Locks and Dam, which is (Ohio River mile 590 to 630; fig. 1). The field data located 150 river miles downstream. The channel used to calibrate and validate the model also are has a generally trapezoidal geometry with steep described. Because of the hydrodynamic banks rising at slopes greater than 22 percent complexities induced by McAlpine Locks and Dam (0.22 ft/ft). The banks along the upstream reach (Ohio River mile 607), the model was split into two extend to an approximate elevation of 440 ft, segments: an upstream reach, which extended from whereas the banks of the downstream reach rise to the dam upstream to the upper terminus of the study approximately 420ft. reach at Ohio River mile 590; and a downstream The following river characteristics were reach, which extended from the dam downstream to determined from the data collected during the a lower terminus at Ohio River mile 636. hydrograJ?hiC surveys. At a discharge of Floodplains were not included in the model 36,000 ~Is, the average depth of the river thalweg domain because of the project emphasis on low-flow was approximately 30 ft in the upstream reach and periods spanning the recreational-contact period 20 ft in the downstream reach. The average width of May-September. The final model was used to the river at a discharge of 36,000 ft3 Is was simulate the steady-flow patterns of the Ohio River approximately 3,000 ft in the upstream reach and 3 for flows ranging from 6,500 to 197,700 ft Is. The 1,600 ft in the downstream reach. A discharge of lower-limit of the simulated flows (6,500 f~ls) is 390,000 f~ Is produced an average depth in the consistent with the discharge during a dye-tracing thalweg of approximately 45 ft in both the upstream survey done by ORSANCO in September 1999. The and downstream reaches. The width of the river at a upper limit of the range of simulated flows discharge of 390,000 ft31s was approximately 3 (197,700 ft 1s) corresponds to the 90-percentile 4,000 ft in the upstream reach and 2,500 ft in the flow during the recreational-contact period based on downstream reach. The bank and bed material of the historical discharge data during 1987-98. Although study reach varies between cohesive and more historical discharge data are available, noncohesive sediment. The bank vegetation ORSANCO and the private consultant chose to use predominately consists of small shrubbery and only recent data (1987-98) to determine the vegetation, void of large trees. 90-percentile flow. The process and methodology
2 Calibration and Validation of a Two-Dimensional Hydrodynamic Model of the Ohio River, Jefferson County, Kentucky 86° 8& 45' 8&30 3~30~------~------,.~------~~~ Upstream model terminus
' -t
620 I )easure Rod ge Parl<
j"lalley Sletoon Bethany Downstream model terminus
ron I 0 1 2 4 EXPLANATION 6 8 10 MILES • River mile marker 615 Ohio River Mile 01 2 4 6 8 10 K ILOMETERS
Figure 1. Location of study area near Louisville, Kentucky.
INTRODUCTION 3 The structures that compose McAlpine Locks horizontal-velocity components for subcritical, free and Dam are located on the northwestern side of surface flow in two-dimensional flow fields Louisville, Ky., and extend from Ohio River (U.S. Army Corps of Engineers, 1997). The model mile 604.4 to 607 .4. The current locks-and-dam is designed for problems in which vertical configuration (fig. 2) consists of a hydroelectric accelerations are negligible, and velocity vectors powerhouse; two lock chambers (one 600 by 110 ft generally point in the same direction over the entire and the other 1,200 by 110 ft in length and width, respectively); nine tainter gates, four near the depth of the water column at any discrete period in powerhouse and five just upstream of the Conrail time. Railroad Bridge; and a fixed weir with a crest at an Typical applications of the RMA-2 numerical elevation of 422ft at the downstream gates, model include calculating water-surface elevations incrementally raised to an elevation of 423 ft at the and flow distribution around islands; flow patterns at upper set of gates. An earthen dike was constructed bridges having one or more relief openings, in at the upstream end of Shippingport Island to reduce the crosscurrents affecting vessels navigating the contracting and expanding reaches, into and out of lock canal. The shoal area that exists downstream off-channel hydropower plants, and at river from the upper gates and fixed weir is referred to as junctions; circulation and transport in water bodies the Falls of the Ohio and is an area of archeological with wetlands; and water levels and flow patterns in importance. To facilitate public use of the fossil rivers, reservoirs, and estuaries. beds at the Falls of the Ohio, the U.S. Army Corps One of the predominating operating of Engineers (COE) has agreed to a 2-month assumptions in RMA-2 is that acceleration of flow operational period (typically August-October) each year when the fixed weir and upper gates are not in the vertical direction is negligible. The model is used to pass flow, except for extreme hydrological not intended for applications in which vortexes, events. Under low flow or when the upper gates of vibrations, or vertical accelerations are the primary McAlpine Locks and Dam are not in operation, the interests; therefore, the simulation of vertically Falls of the Ohio is an area of slack flow. stratified flow fields is beyond the capabilities of RMA-2 (U.S. Army Corps of Engineers, 1997). MODEL CALIBRATION AND VALIDATION Field Data Collection and
At least two data sets are required to Interpretation adequately calibrate and validate a numerical model. The general procedure used to calibrate and . Water-surface elevations, channel validate the RMA-2 model was to first collect field bathymetry, and detailed water-velocity data in order to develop the computational mesh. measurements were collected at two different flow The model then was calibrated to the water-surface conditions (36,000 and 390,000 ft3/s) and used in elevations and velocities observed in the field for the the model for calibration and validation, initial flow. A second flow condition then was respectively. Water-sUrface elevations were simulated without changing the computational mesh measured at 10 locations (4 upstream and or model parameters, and the simulated water 6 downstream) along the study reach concurrent surface elevations and velocities were compared with those observed in the field for this second flow with both hydraulic surveys. Detailed water-velocity condition. measurements and channel-bathymetry data were collected at 30 cross sections ( 12 upstream and 18 downstream)--spaced approximately 1.5 mi Model Description apart--during each of the hydraulic surveys (figs. 3 and 4). A separate data-collection trip from the two RMA-2 is a two-dimensional depth-averaged hydrographic surveys was used to collect channel finite-element hydrodynamic numerical model bathymetry data in sufficient detail for the capable of computing water-surface elevations and development of a computational mesh.
4 Calibration and Validation of a Two-Dimensional Hydrodynamic Model of the Ohio River, Jefferson County, Kentucky ~ Upper gates
3: 0 c m r ..""- Earthen 0 )> dike r a; JJ ~ 0 z 0 .25 .50 .75 1 MILE )> z c 0 .25 .50 .75 1 KILOMETER < EXPLANATION )> cr (%) USGS gaging station location • • Continuity check line )> USGS gaging station number CC3 Continuity line name ~ 03294600 0 z Figure 2. McAlpine Locks and Dam configuration and U.S. Geological Survey (USGS) gaging station locations near Louisville, Kentucky. 8E?45 85l35 I I 3ff25 - / - Goshen Twelve-Mile \ Island
'- \
'
~ ,,) Six- Mile -- Island ' -..... /
,.-> (ill) :r
0 2 3 4 MILES 0 2 3 4 KILOMETERS
EXPLANATION <1) USGS gaging station location - Survey cross section 03294600 USGS gaging station number 1 0 Cross-section number * Surveyed water-surface e River mile marker elevation location 600 Ohio River Mile
Figure 3. Location of hydrographic-survey cross sections, surveyed water-surface elevation stations, and U.S. Geological Survey (USGS) gaging stations in the upstream study reach near Louisville, Kentucky.
6 Calibration and Validation of a Two-Dimensional Hydrodynamic Model of the Ohio River, Jefferson County, Kentucky 8&01' 8&53 8&45 I ...___- - ...... ~ ..... _____ ,// I ... I ---~~-_, ( r 3s016' - L - .-'\
I I ...- .. -_'r··· ) l ______./' L
(
L -+-
.1 '• (
./L__ \ ...--~ ,-· I
_...
...... -
_}
·"'" -----·"'"
--...... __ ,.,... ___ _ 3ff - f ~aft- - Aiver .r•n-.:.. .1 lil'ldl' 1te . t~ . aaemen '' e 0 2 3 4 5 MILES 0 2 3 4 5 KILOMETERS EXPLANATION USGS gaging station location .._. Survey cross section 03294600 USGS gaging station number 1 0 Cross-section number Surveyed water-surface e River mile marker elevation location * 600 Ohio River Mile Figure 4. Location of hydrographic-survey cross sections, surveyed water-surface elevation stations, and U.S. Geological Survey (USGS) gaging stations in the downstream study reach near Louisville, Kentucky.
MODEL CALIBRATION AND VALIDATION 7 Water-Surface Elevations at two standard deviations; tests and prior use of this unit indicate that typically about 80 percent of the Water-surface elevations at the seven data are within 3.3 ft of the true location. locations throughout the reach (figs. 3 and 4) were Recent advances in velocity-measurement surveyed with a total station. To document the technology allow three-dimensional velocities to be changes in river stage during a hydraulic survey, the collected from a moving boat using an ~coustic water-surface elevations were sl:rrveyed in the Doppler current profiler (ADCP) (Oberg and morning and then again in the afternoon. The Mueller, 1994; Mueller, 1996). All velocities were average water-surface elevation was used to measured with an ADCP. The ADCP allows three determine a water-surface slope corresponding to dimensional velocities to be measured from the average discharge measured during the survey. approximately 4 ft beneath the water surface to The 40-mi study section of the Ohio River within 6 percent of the depth to the bottom. includes a total of four USGS stream gages-two Established methods were used to estimate the each in both the upstream and downstream reaches. discharge in the unmeasured top and bottom Both upstream gages--one located on the Second portions of the profile (Simpson and Oltmann, Street Bridge (number 03293548) and the other at 1991). Cross-sectional average velocities were Indiana Pass (number 03293550)-are located near computed by dividing the measured discharge by river mile 604 and provided to assist the COE, the measured cross-sectional area. In addition, Louisville District in maintaining the McAlpine depth-averaged velocities were computed for Locks and Dam normal pool elevation. The Second subsections of the flow in each cross section; Street Bridge gage is located on a pier near the however, these discrete depth-averaged velocities center of the channel about 4,000 ft upstream from were computed as an average of the measured McAlpine Locks and Dam; the Indiana Pass gage is velocity and did not account for the velocity in the located on the Indiana shore about 500 ft downstream from the Second Street Bridge gage unmeasured portions of the water column. (fig. 3). The McAlpine tailwater and Kosmosdale In order to compensate for the slight changes gaging stations (numbers 03294500 and 03294600, in discharge of the river during the survey, all the respectively) are located within the downstream discharge measurements collected were averaged to reach (fig. 4). The McAlpine tailwater gage is produce a flow rate that was representative of the located near the Kentucky shore on the downstream entire survey period. The time necessary to end of the lock guide wall (river mile 607). The complete a hydraulic survey on each of the Kosmosdale gage is located on the Kentucky shore,. simulated sections did not permit both the upstream 19.8 mi downstream from the tailwater gage (river and downstream reaches to be surveyed on the same mile 628). day; therefore, minor differences are present in the discharge measurements used to calibrate and validate the upper and lower models. Velocity and Discharge
Water-velocity and discharge data were Bathymetry collected from a moving boat. The horizontal position of the boat was measured using a . Bathymetry data also were collected from a differentially corrected global positioning system moving boat. The horizontal position of the boat (DGPS) receiver. The DGPS system used receives was measured using the DGPS receiver. During the its differential corrections from a commercial initial low-water survey, a 200-kHz echo sounder service's communications satellite. The unit is was used to measure the channel bathymetry at the specified by the manufacturer to be accurate to 3.3 ft 30 predefined cross sections.
8 Calibration and Validation ~fa Two-Dimensional Hydrodynamic Model of the Ohio River, Jefferson County, Kentucky The channel bathymetry obtained from the Although forcing the data points onto a straight line initial bathymetric survey at the 30 velocity does introduce some error, the straight-line cross measurement cross sections was compared to the sections produced a better representation of the 1963 COE hydrographic survey of the area. The upstream channel bathymetry than the unadjusted surveyed data points of the 1963 hydrographic data. survey were digitized along straight cross sections spaced at 114-mi intervals and oriented perpendicular to flow. All data points were digitized from paper maps. A triangulated irregular network Upstream River Reach (TIN) of the 1963 data was generated to extract cross sections from the 1963 data for the The simulated upstream reach begins 3 mi 30 locations surveyed for this project. The shapes of upstream of the Jefferson/Oldham County line the cross sections measured in the downstream (Ohio River mile 590) in Kentucky, and extends reach were similar to cross sections extracted from downstream to McAlpine Locks and Dam (Ohio the 1963 hydrographic-survey data. Differences in River mile 607). Two islands are present that add elevation were within the errors expected from the hydraulic complexities to this reach: Twelve-Mile survey technologies of 1963 and those used in this Island located at Ohio River mile 593, and Six-Mile survey; therefore, data from the 1963 hydrographic Island located at Ohio River mile 598. The survey were used for the channel bathymetry in the configuration of McAlpine Locks and Dam includes downstream hydrodynamic model. The shape of the upper and lower sets of tainter gates and a cross sections measured in the upstream reach hydropower plant that served as downstream displayed differences from the cross sections boundary conditions for the model. extracted from the 1963 hydrographic-survey data; therefore, a second bathymetric survey of the upstream reach was completed with cross sections Computational-Mesh Configuration spaced approximately 1,000 ft apart. The data from the second bathymetry survey was used for the The finite-element network for the upstream channel bathymetry in the upstream hydrodynamic river reach consisted of 6,920 elements. The islands model. and dike were simulated by installing gaps in the Bathymetric data surveyed with the echo mesh because the simulated flows did not overtop sounder in the upstream reach did not fall on these features (figs. 6 and 7). The boundary nodes at perfectly straight cross sections because of the upstream and downstream points of the islands inconsistencies in the boat course across the river. were specified as "stagnation points" (locations of The internal triangulation routine done by the zero velocity). These specifications generally are RMA-2 interface software package--Surface-water located in comers of grids or along boundaries with Modeling Systems (SMS), version 7.0 (Brigham relatively negligible flow velocities and are applied Young University, 1999)-did not properly interpret to reduce artificial loss of discharge out of the the collected bathymetry data, resulting in channel boundaries. The resolution of the grid was triangulation within the same cross section and a increased in the area around McAlpine Locks and large portion of the area between two cross sections Dam in order to reproduce the hydraulic estimated with only three data points. An example of the raw-data triangulation is presented in complexities induced by these structures (fig. 8). In figure 5a. The raw data collected along each cross order to maintain numerical stability, the sloping section was mathematically forced onto a straight banks were truncated at an elevation of 418 ft. line andre-triangulated; figure 5b shows the Above this elevation, the banks were assumed to be re-triangulation of the section shown in figure 5a. vertical walls.
MODEL CALIBRATION AND VALIDATION 9 A. Triangulation of raw data
B. Triangulation of straight-line data
Figure 5. Surface Water Modeling System (SMS) software triangulation of raw and straight-line bathymetry data for a section of the upstream study reach in the Ohio River model simulation near Louisville, Kentucky.
10 Calibration and Validation of a Two-Dimensional Hydrodynamic Model of the Ohio River, Jefferson County, Kentucky Shippingport Island
Earthen dike
3: 0 0 m r (') )> 0 500 1,000 FEET r a; I I JJ )> I I I -i 0 0 150 300 METERS z )> z 0 < )> r 6 )> Figure 6. Mesh configuration around the earthen dike adjacent to Louisville, Kentucky, and upstream of McAlpine Locks and Dam in the Ohio River model -i 5 simulation nea r Louisville, Kentucky. z ...... A. Mesh configuration around Six-Mile Island
Six -Mile Island
B. Mesh configuration around Twelve-Mile Island
Twelve-Mile Island
0 .25 .50 .75 1 MILE I I I I I I I I I 0 .25 .50 .75 1 KILOMETER
Figure 7. Mesh configurations around Six- and Twelve-Mile Islands in the upstream Ohio River study reach near Louisville, Kentucky.
12 Calibration and Validation of a Two-Dimensional Hydrodynamic Model of the Ohio River, Jefferson County, Kentucky Hydropower plant 0 .25 .50 .75 1 MlLE I I I I I I I 0 .25 .50 .75 1 KILOMETER
Upper gates
Shippingport Island
3:: 0 c m r- ~ a;r- ::tJ ~ Earthen dike z0 )> z c )>< r- 6 ~ 0 Figure 8. Upstream reach mesh configuration around McAlpine Locks and Dam near Louisville, Kentucky. z Boundary Conditions were used to validate the model. The calibration and validation process consisted of comparing the Steady-state discharges of 34,000 and 383,000 f0 Is were used to calibrate and validate the simulated water-surface elevations at the model. These discharges were used as the upstream 4 upstream water-surface-elevation stations and boundary condition for the model and the lateral 12 cross-sectional velocity profiles with those flow distribution was assumed to be uniform across surveyed in the field. A Manning's roughness the inflow boundary for both discharges. coefficient (n) was assigned to each element and The downstream-boundary condition consisted of outflows through the lower gates and iteratively adjusted until the model adequately hydropower plant (low-flow only) and a head simulated the surveyed water-surface elevations and condition assigned to the upper g,ates. The lower set velocity profiles. of tainter gates carried 14,000 ft' Is, whereas the Inspection of the velocity profiles collected in hydropower plant passed the remalning 20,000 f~ Is the field revealed no-slip conditions along the under the low-flow condition. At 383,000 f2 Is, the riverbanks, indicating that the shear stress along the flow is divided between the upper and lower gates, with 126,350 ft31s passing through the lower gates banks is great enough to cause the tangential and the remaining 256,650 tt3 Is passing through the velocity to approach zero. To simulate this upper gates. The McAlpine Locks and Dam pool characteristic with RMA-2, the Manning's n value elevations for both discharges were determined was increased to 0.035 for one row of elements from the USGS gaging station records at Indian along the outer boundary of the mesh. The Pass (number 03293550) and Second Street Bridge calibrated Manning's n in the remainder of the (number 03293548). Normal pool elevation (419.6 ft) was used for 34,000 ft31s flow, and a pool channel was 0.024. This combination produced the elevation of 423.9 ft was used for 383,000 ft31s best simulation of water-surface elevation (table 1), flow. velocity magnitudes, and lateral-velocity distribution for both low- and high-flow conditions. Calibration and Validation Results Matching the high-flow water-surface elevations Data from the low-flow (34,000 ft31s) verified that the area lost by truncating the banks hydraulic survey were used to calibrate the model, was negligible in the model simulation. and data from the high-flow (383,000 ft31s) survey
Table 1. Summary of water-surface elevation calibration and validation for the upstream study reach in the Ohio River model simulation near Louisville, Kentucky
(8/13/98) Low-flow condition (2/16/00) High-flow condition
Field Model Field Model water-surface water-surface water-surface water-surface elevation elevation elevation elevation (feet above (feet above (feet above (feet above Station sea level) sea level) Difference 1 sea level) sea level) Difference 1
Harmony Landing 419.99 419.66 -0.33 427.59 427.68 0.09
Louisville Water Company 419.91 419.64 -.27 426.51 426.46 -.05
Cox's Park Well 419.86 419.62 -.24 425.19 425.04 -.15
Indiana Pass 419.60 419.60 0 423.88 423.90 .02
1Differences are determined by subtracting field from model-simulated water-surface elevation.
14 Calibration and Validation of a Two-Dimensional Hydrodynamic Model of the Ohio River, Jefferson County, Kentucky The simulated velocity magnitudes and that allows flow beneath the wall, and Sand Island distributions compared well with the field . all contribute to the hydraulic complexities of the measurements. A comparison of the model and area located immediately downstream of the field-velocity profiles for cross-section number 5 McAlpine Locks and Dam. (13 mi upstream from McAlpine Locks and Dam) is shown in figure 9. The shape of the field- and Computational-Mesh Configuration model-velocity distributions were similar, whereas The finite-element network for the on average, the velocity magnitudes were within downstream river reach consisted of 0.1 ft/s. The average cross-sectional velocities for 10,803 elements. In order to maintain numerical the 12 cross sections also were compared. The stability, the sloping banks were truncated at an model adequately reproduced the average field elevation of 380 ft. Above this elevation, the banks velocities (figs. 10 and 11). The model also were assumed to be vertical walls. accurately reproduced the measured-flow Sand Island and an unnamed island in the distributions around Twelve- and Six-Mile Islands Falls of the Ohio region were simulated by (table 2). installing gaps in the mesh at the location of these Continuity was checked throughout the features. The resolution of the grid is increased in downstream model to assure that mass was being the area around Sand Island, the hydropower plant, conserved. The model conserved mass throughout the lock wall, and both sets of tainter gates to the reach under high flow, but a 3 percent loss improve simulation of the hydraulic complexities resulted under low flow in the approach to the lower induced by these structures (fig. 12). In order to tainter gates and hydropower plant (table 3). A simulate the passage of flow under the lock wall, an tolerance of +1- 3 percent in mass-conservation increased Manning's roughness value (n =.50) was discrepancy is typically acceptable for most models assigned to the elements representing the lock wall; (U.S. Army Corps of Engineers, 1997). Algorithms this value allowed flow through the elements under . used to import RMA-2 output into the WASP water a much greater resistance. Stagnation points were quality model can correct for mass-conservation created at sharp break points along Sand Island and discrepancies. the mesh boundary in the area of the lower tainter gates to reduce artificial loss of discharge through the boundaries. Downstream River Reach
The simulated reach· begins at the McAlpine Locks and Dam (Ohio River mile 607) and extends downstream to Ohio River mile 636, 6 mi downstream from the mouth of the Salt River. This section of the model includes part of the Cannelton Locks and Dam pool, which is held at a normal pool elevation of 383 ft. The two sets of tainter gates, the hydropower plant, a lock wall supported on piles
MODEL CALIBRATION AND VALIDATION 15 A. Low-flow velocity profile comparison
--Field ...... _Model
0 0z 0.6 (.)w Cl) a:: w D.. tu 0.4 w LL ~ ~ (.) 0 0.2 ..Jw >
0.0~--~----~----~----~----~----~----~----~--~ 0 200 400 600 800 1,000 1,200 1,400 1,600 1,800
DISTANCE FROM LEFT BANK, IN FEET
B. High-flow velocity profile comparison
6r---~.---~.---~----~----~----~-----r----~----~ _.Field ...... _Model
5 0z 0 (.) w UJ 4 c::: w D.. ~ 3 LL ~ ~- 2 (.) 0 ..J ~
0~--~----~----~----~----~----~----~----~----~ 0 200 400 600 800 1,000 1,200 1,400 1,600 1,800 DISTANCE FROM LEFT BANK, IN FEET
Figure 9. Field-measured and model-simulated velocity profiles for cross-section 5, 13 miles upstream from McAlpine Locks and Dam near Louisville, Kentucky.
16 Calibration and Validation of a Two-Dimensional Hydrodynamic Model ofthe Ohio River, Jefferson County, Kentucky O c:::::::J Field 7 r .67 · -Model
~~
() 0.6 0 _Jw > _J 0.5 ~ 0 0 z - 0 0.4 ()~ ()w ~ (/) I 0:: (j) w (j) a_ 0.3 0 ~ 0::: w () w w u.. 0.2 (!) z 3:: 0c ~ m w r- 0.1 ~ ~ r-a; ~ 0 0 z 1 2 3R 3L 4 5 6 ?R ?L 8 9 10 11 12 )>z c CROSS SECTION, IN DOWNSTREAM ORDER r-~ 6 ~ 0 Figure 10. Field-measured and model-simulated low-flow, distance-weighted average cross-sectional velocities in the upstream z study reach from the Ohio River model simulation near Louisville, Kentucky...... -..1 ..6 o:> s=~ iii :::t 0 ::::s C» ::::s Q. 1 Field < I r=- !. 5 5 s: ~~ · Model C» :::to ::::s () 5.0 a 0 C» _J ~ w 4.5 0 _J <( 0 4.0 ~::::s 0 z z 0 ::::s !. 0 8 3.5 :::r:: '< b w Q. w (/) 3.0 j (/) 0:.:: ::::s C» ch w 3 (/) 0.. 2.5 c; 0 J- &.== 0:.:: w !. () w 2.0 a LL ~ w z Cl) (!) - 1.5 0 ::::J" 0 ~ :::0 w 1.0 ~ .:"' c.. ~ Cl) 0.5 ;: (3 0 ::::s 0 1 2 3R 3L 4 7R 7L 10 c~ 5 6 8 9 11 12 ::::s ~ CROSS SECTION, IN DOWNSTREAM ORDER ~ ~n Figure 11. Field-measured and model-simulated high-flow, distance-weighted average cross-sectional velocities in the upstream study ~ reach from the Ohio River model simulation near Louisville, Kentucky. Table 2. Summary of flow-split calibration and validation for islands in the upstream study reach in the Ohio River model simulation near Louisville, Kentucky
Low flow High flow
Field Model Field Model flow split flow split flow split flow split Location (percent) (percent) (percent) (percent)
Twelve-Mile I land - Right 53.6 51.7 51.6 49.9
Twelve-Mile Island- Left 46.4 48.3 48.4 49.5
Six-Mile Island - Ri ght 8.9 7.8 1 1.2 10.0
Six-Mile Island - Left 91.1 92.0 88.8 90.0
Table 3. Summary of continuity checks in the upstream Ohio River model simulation near Louisville, Kentucky
Continuity check Percent of total discharge line description (fig. 3) Low flow High flow
Inflow 100 100
Cross ection 2 99.5 99.8
Cross section 3R 51.8 50.1
Cross section 3L 48.2 49.5
Cross section 5 100 100
Cross section 7R 8.0 10.2
Cross section 7L 92 .0 89.6
Cross section I I 100 99.7
Upper gate 0 66.5
Lower gates 45 .2 32.8
Hydropower plant 52.0 0
Total outflow 97 .2 99.3
MODEL CALIBRATION AND VALIDATION 19 N 0
0 !!. c: ~ 0 :::s A) :::s c. < a:!!. ! Unnamed Island 0 :::s a A) ~ 9 c
~:::s en 0 :::s !!. \ ::1: Lower \ '< \ c. \ gates \ \ j \ :::s \ A) \ 3 \ cr \ \ Lock wall \ == \ 8. \ !. \ a \ :;t ' ' (I) ' 0 :::J' 0 Fixed weir :::D ~ :" c:.. 0 2000 4,000 FEET i I I UJ 0 I I :::s I g 0 600 1,200 METERS c :::s ~ ~ :::s 2' Figure 12. Mesh configuration in the downstream study reach around the tailwater area of McAlpine Locks and Dam near Louisville, Kentucky. () ~ Boundary Conditions Calibration and Validation Results
Steady-state discharges of 36,000 and Similar to the upstream reach, the low-flow 3 397,000 ft Is were used to calibrate and validate the (36,000 ft3 Is) hydraulic survey was used to calibrate model, respectively. At 36,000 ft3 Is, the the model, and the high-flow (397,000 tt31s) survey hydropower plant passed 22,000 ft3 Is, whereas the was used to validate the simulation. The calibration lower gates passed 14,000 tt31s. At 397,000 ft3/s, and validation process consisted of comparing the the inflow is divided between the upper and lower simulated water-surface elevations at the gates, with 266,030 ft3Is from the upper gates and 6 downstream water-surface elevation stations and 130,970 f21s from the lower gates. 18 cross-sectional velocity profiles (fig. 4) with those surveyed in the field. A Manning's roughness The downstream head-boundary conditions coefficient (n) was assigned to each element and for both discharges were determined from iteratively adjusted until the model adequately stage-discharge relations developed from USGS simulated the surveyed water-surface elevations and gaging-station records at the McAlpine tailwater velocity profiles. (number 032946500) and Kosmosdale Inspection of the velocity profiles collected in (number 03294600) gages. The relation between the field revealed no-slip conditions along the stage and discharge was determined by plotting the riverbanks, which were similar to the upstream discharge computed for the McAlpine gage with the reach. To simulate this characteristic with RMA-2, stage measured at the McAlpine tail water and the Manning's n value was increased to 0.035 for Kosmosdale gaging stations (figs. 13 and 14). A one row of elements along the outer boundary of the relation between discharge and water-surface slope mesh. The calibrated Manning's n in the remainder between the McAlpine tail water and Kosmosdale of the channel was 0.024. This combination gages also was developed (fig. 15). The downstream produced the best simulation of water-surface head-boundary conditions were determined by elevation (table 5), velocity magnitude, and lateral translating the observed water-surface elevations velocity distribution for both low- and high-flow from the Kosmosdale gage 5 mi downstream to the conditions. Matching the high-flow water-surface elevations verified that the area lost by truncating model boundary by use of the discharge water the banks was negligible. surface slope relation. For simulated flows other Comparison of the simulated velocity than those measured for calibration and validation, magnitudes and distributions with field the stage at Kosmosdale can be determined from a measurements showed good agreement for the low known discharge and the stage-discharge relation flow simulation but less favorable agreement for the and then translated to the downstream end of the high-flow simulation. Examples of the general model based on the discharge water-surface slope agreement between the low-flow simulated relation (fig. 15). The difference between the velocities and the field measured velocities are McAlpine tail water rating and the water-surface shown in figure 16. The maximum difference at low elevation (simulated by RMA-2 based on the flow was about 0.25 ft/s. For the high-flow downstream boundary condition estimated from the condition, the simulated velocities consistently were technique described above) ranges from 0.16 to greater than the measured velocities, despite 0.4 ft (table 4). excellent agreement in the water-surface elevations.
MODEL CALIBRATION AND VALIDATION 21 403 ~~~--~~--~~~--~~--~~--~~~--~~--~~~~ 402 _J 401 w + >w 400 _J 399 + + + 0:: <( + w w 398 + ~ ~ 397 ~ 6 396 + + ~ ~ 395 + w ..... + z ttl 394 + a.. u.. 393 _J Best-fit line <( z (.) z 392 ~ 0 391 ~ ..... 390 z ~ o en 389 + ~ ~ 388 W (!) 387 u1 ~ 386 + 385 384 383 0 50,000 100,000 150,000 200,000 DISCHARGE, IN CUBIC FEET PER SECOND
Figure 13. Stage-discharge curve for McAlpine tailwater gaging station near Louisville, Kentucky, 1990-2000. 4oo~~~~~~--~~--~~--~~--~~--~~--~~--~+--~~~ 399 z 398 0 I- 397 ~ Best-fit line + (/) 396 + (!) z 395 (!) 394 (3 393 w .....1 .....1 w 392 + c:x: > 0 w (/) c:x: 391 + 0 .....1 390 (/)~ c:x:w 0 (/) 389 ~ + w 388 + ~ 6 + z m 387 0 c:x: :!!: 386 0 c + m ~ ~ 385 r w u.. (') .....1 z )> w 384 r a; 383 ~ 382~~~~~~--~~--~~--~~--~~--~~--~~~--~~~ 0z 0 50,000 100,000 150,000 200,000 z)> c DISCHARGE, IN CUBIC FEET PER SECOND <)> r 8 ~ Figure 14. Stage-discharge curve at Kosmosdale gaging station near Louisville, Kentucky, 1995-2000. 5 z flo) ~
(') !. w c;: ....J 6.0e-5 ;... <( + 0 0 :::J (/) + II) + :::J + + + c. 0 < ~ + + + !. (/) a: 0 5.0e-5 + ++ II) + ct ~ +++ 0 :::J 0 + 0 z -II) <( * ~ 0::: 9 w 4.0e-5 + + c ..... 0 + + + + + ~ ~ + \ :::J 0 *+ (I) ~ LL + 0 ....J :::J 0::: * !. w + + + :::t ~ a. 3.0e-5 + '
Rating Model water-surface elevation, water-surface elevation, Boundary condition McAlpine tallwater McAlpine tallwater Downstream discharge (elevation (elevation (elevation (cubic feet per second) above sea level} above sea level) above sea level}
16,000 382.74 383.60 383.22
23,000 382.98 384.20 383.80
43,000 383.86 386.10 385.94
97,000 387.25 391.80 392.23
200,000 395.54 402.60 402.27
Table 5. Summary of water-surface elevation calibration and validation for the downstream study reach in the Ohio River model simulation near Louisville, Kentucky
(5119/00} Low-flow condition (2/17/00) High-flow condition
Field Model Aeld Model water-surface water-surface water-surface water-surface elevation elevation elevation elevation (feet above (feet above (feet above (feet above Station sea level) sea level) Difference1 sea level) sea level) Difference1
McAlpine tailwater 385.20 384.90 -0.30 416.18 416.13 -0.05
Shawnee well 384.75 384.50 -.25 415.10 415.21 .11
RR-22 well 384.68 384.20 -.48 413.65 413.83 .18
Kosmosdale 383.85 383.50 -.35 410.10 410.12 .02
WestPoint 383.54 383.40 -.14 409.70 409.50 -.20
1Differences are determined by subtracting field from model-simulated water-surface elevation.
MODEL CALIBRATION AND VALIDATION 25 A. Low-flow velocity profile for cross-section 19
-Field 1.75 _._ Model c z 0 1.50 0w en ffi 1.25 D.
t:iw 1.00 u. ~ ~ 0.75 0 g 0.50 ~ 0.25
o.ooL-----~-----L----~----~~----~----~----~----~ 0 200 400 600 800 1,000 1,200 1,400 1,600 DISTANCE FROM LEFT BANK, IN FEET
B. Low-flow velocity profile for cross-section 25
1.75
-Field ...... ,_. Model 1.50 c z 0 0w 1.25 en w D." 1.00 t:iw u. ~ 0.75 ~ 0 g 0.50 w > 0.25
0.00 0 200 400 600 800 1,000 1,200 1,400 1,600 DISTANCE FROM LEFT BANK, IN FEET
Figure 16. Field-measured and model-simulated low-flow velocity profiles for cross-sections 19 and 25 in the Ohio River model simulation near Louisville, Kentucky.
26 Calibration and Validation of a Two-Dimensional Hydrodynamic Model of the Ohio River, Jefferson County, Kentucky To identify the cause of the disagreement to simulate the scouring of the channel bed (figs. 18 between the simulated and measured velocities for and 19). A comparison of the average cross the high-flow condition, the cross-sectional areas of sectional velocities for the field and two model the model and the field were evaluated. The meshes are shown in figures 20 and 21. difference between the magnitudes of the model and Although extending the banks to an elevation field average cross-sectional velocities is correlated of 390 ft improved the overall accuracy of the high closely with differences in cross-sectional area flow simulation (3.4 percent in average cross (fig. 17). This difference in areas for the high-flow sectional velocity), the primary reason for the condition indicates that truncating the channel bathymetry at an elevation of 380 ft may not be difference between the model and field cross acceptable for high flows. Extending the bathymetry sectional areas and velocities is that the model has a to an elevation of 390 ft improved the agreement fixed bed and the river adjusted its cross section in between field and model velocities but failed to the downstream reach during high flow. To explain all of the error. For low-flow conditions, the minimize errors of future simulations, the errors between the field data and the model are bathymetry was truncated at an elevation of 380ft somewhat randomly distributed around zero. The for flows less than 200,000 ft3/s, and bathymetry errors for the high-flow condition with banks was truncated at an elevation of 390 ft for flows truncated at an elevation of 380 ft show a linear ranging from 200,000 to 400,000 tt3 Is. pattern, while the errors for the high-flow condition Continuity was checked throughout the model with banks truncated at an elevation of 390ft show a to ensure that mass was being conserved. The less distinct pattern but still well below zero. The location of the continuity-check lines near Sand patterns indicated by the high-flow data indicate that Island and the Falls of the Ohio is shown in figure 2; something outside the scope of the model may be responsible for the apparent errors. Scouring of the mass is conserved around Sand Island and the Falls channel bottom occurred during high flow and is of the Ohio (table 6). Based on the calibration and partially responsible for the large difference validation results, the model is a representative between the model and field cross-sectional areas simulation of the Ohio River steady-flow patterns and average velocities because the model is not able below discharges of approximately 400,000 ft3Is.
MODEL CALIBRATION AND VALIDATION 27 15 I- z 0 0 Low Flow w 1:.:. High Flow 380 feet (.) 10 - 0 0:: 0 w High Flow 390 feet a.. z 5- 0 ~ I- 0 (.) 0 0 0 0 0 0 ...J 0 0 w 0 > 0 0 w -5- 1:.:. (9 0 0 ClJ:J f:t:>. 0 0 0 0 0 ~ o~:.:. w I'.:.!!KJ -10 - 0 ><( ~ 0 0 do z 0 1:.:. 0 1:.:. w A (.) -15 - & z 1:.:. w 0:: 1:.:. w -20 - LL LL 0
-25 I I I I I I I I -15 -10 -5 0 5 10 15 20 25 30 DIFFERENCE IN CROSS-SECTIONAL AREA, IN PERCENT
EXPLANATION High Flow 380 feet- All bathymetry truncated at elevation 380 feet above sea level High Flow 390 feet - All bathymetry truncated at elevation 390 feet above sea level
Figure 17. Scatter plot of the cross-sectional area and average velocity differences between field measurements and model-simulated results for the 18 downstream Ohio River cross sections near Louisville, Kentucky. 420 --- Model ..Jw Field (high flow) >w -- 1963 survey ..J 410
0 200 400 600 800 1,000 1,200 1,400 1,600 1,800
DISTANCE FROM LEFT BANK, IN FEET Figure 18. Comparison of cross-section 24 bathymetry in the downstream Ohio River study reach near Louisville, Kentucky.
420 ___._ Model ..J Field (high flow) w (ij --e-- 1963 SUI\eY ..J 410 ~ (/) w 400 > 0 co < tii 390 w LL ~ 380 z 0 ~ 370 w ..Jw 360
0 200 400 600 800 1,000 1,200 1,400 1,600 1,800
DISTANCE FROM LEFT BANK, IN FEET
Figure 19. Comparison of cross-section 20 bathymetry in the downstream Ohio River study reach near Louisville, Kentucky.
MODEL CALIBRATION AND VALIDATION 29 2.25 ___._ Field ___.___ Model ~ 2.00 (.) 0 ....J 1.75 w > ....J <{ 1.50 z 0 1- (.) 1.25 w 0 C/) z I 0 C/) (.) C/) w 1.00 0 C/) 0::: (.) 0:::w w a.. 0.75 <.9 1-w ~ w w u.. 0.50 ~ z 0.25
0.00 0 5 10 .15 20 25 DISTANCE FROM MCALPINE LOCKS AND DAM, IN MILES
Figure 20. Low-flow distance-weighted average cross-sectional velocity profile for the downstream study reach in the Ohio River model simulation near Louisville, Kentucky. 6.5 ~ () 0 6.0 _J w > _J <( 5.5 z 0 f- () w Cl 5.0 (/) z I 0 (/) () (/) w 0 (/) 4.5 0:: 0:: () w w a.. • Field (!) f- <( w 4.0 Model 380 feet 0:: w Ill w LL Model 390 feet z :!:: ~ 0 3.5 c m r-
~r- a; 3.0 ~--~------~------~------~------~------~----~ ~ 0 5 10 15 20 25 0z DISTANCE FROM MCALPINE LOCKS AND DAM, IN MILES )>z c EXPLANATION Model 380 feet - All bathymetry truncated at elevation 380 feet above sea level r-~ Model 390 feet - All bathymetry truncated at elevation 390 feet above sea level ~ 0 z Figure 21. High-flow distance-weighted average cross-sectional velocity profile for the downstream study reach in the Ohio River model simulation near Louisville, Kentucky. Table 6. Summary of continuity checks in the downstream Ohio River model simulation near Louisville, Kentucky
Continuity check Percent of total discharge line description (figs. 2, 4) Low flow High flow Upper gates - inflow 0.0 67.0 Hydropower - inflow 60.4 0 Lower gates - inflow 39.4 33.0 Total inflow 99.8 100.0
CCI .I 65.6 CC2 84.0 101 CC3 17 .6 -.4 CC4 79.4 100.7 CC5 19.2 -.5 Cross section 15 102 98.4 Cross section 22 100 99.9 Outflow 100 100
SUMMARY AND CONCLUSIONS conducted on the upstream reach to determine the channel geometry. Data from the Ohio River survey The determination of current patterns is an (1963) done by the U.S. Army Corps of Engineers essential component of a Water Quality Analysis was used to determine bathymetry on the Simulation Program (WASP) in a riverine downstream reach. Data collected during the environment. The U.S. Geological Survey (USGS), upstream bathymetric survey was mathematically in cooperation with the Ohio River Valley Water forced onto straight cro s-section lines providing a Sanitation Commission (ORSANCO), developed a more accurate triangulation of the data in the field-validated two-dimensional hydrodynamic modeling oftware package. Bathymetry of both model capable of quantifying the steady flow reaches was truncated along the banks to eliminate patterns in the reach of the Ohio River extending Resource Management Associates-2 (RMA-2) from Ohio River mile 590 to 630 for the convergence problems associated with the steeply ORSANCO water-quality modeling efforts on that sloping banks of the Ohio River. The upper reach reach. The model was calibrated to a low-flow was truncated at an elevation of 418 ft for both flow hydraulic survey (approximately 35,000 cubic feet conditions, whereas the lower reach was truncated per second (ft3/s)) and validated with data collected at an elevation of 380 ft for low flows (less than 3 during a hif,h-flow survey (approximately 200,000 ft /s) and an elevation of 390ft for high 3 390,000 ft /s). The model calibration and flows (from 200,000 to 400,000 ft /s). verification proce s included matching water Based on historical di charge data during surface elevations at 10 locations and velocity 1987-98, the 90-percenti le flow during this profiles at 30 eros section in the study reach. The recreational-contact period (May-September) is tudy area was separated into an upper and lower 197,700 ft3/s; therefore, only the model with river reach separated at McAlpine Lock and Dam bathymetry truncated at an elevation of 380 ft is (Ohio River mile 607). A bathymetric survey was needed to imulate these flows.
32 Calibration and Validation of a Two-Dimensional Hydrodynamic Model of the Ohio River, Jefferson County, Kentucky The model was calibrated and validated by Mueller, D.S., 1996, Scour at bridges-Detailed data use of water-s~ace elevations and average cross collection during floods, in Federal Interagency sectional velocities to achieve the minimum error Sedimentation Conference, 6th, Las Vegas, Nev., for both high- and low-flow conditions. The 1996, Proceedings: Subcommittee on Sedimentation, Interagency Advisory Committee simulated low-flow water-surface elevations on Water Data, p. 41-48. typically were biased between 0.2 and 0.3 ft low, Oberg, K.A., and Mueller, D.S., 1994, Recent whereas the simulated high-flow water-surface applications of Acoustic Doppler Current Profiles, elevations were within 0.1 ft of the field conditions. in Fundamentals and Advancements in Hydraulic Simulated average cross-sectional velocities were Measurements and Experimentation, Buffalo, typically within 0.1 ftls for low flow and 0.3 ftls for New York, 1994, Proceedings: Hydraulics high flow when compared with field data. Division!American Society of Civil Engineers (ASCE), p. 341-350. Simpson, M.R., and Oltmann, R.N., 1991, Discharge measurement system using an Acoustic Doppler REFERENCES CITED Current Profiler with applications to large rivers and estuaries: U.S. Geological Survey Water-Resources Brigham Young University, 1999, Surface-Water Investigations Report 91-487, 49 p. Modeling System (SMS) Reference Manual U.S. Army Corps of Engineers, 1997, Users Guide to (ver. 7.0): Provo, Utah, Environmental Modeling RMA2 WES (ver. 4.3): Vicksburg, Miss., Waterways Research Laboratory, [variously paged]. Experiment Station, 227 p.
REFERENCES CITED 33 . .. Wagner and Mueller-CALIBRATIO~ AND VALIDATION OF A TWO-DIMENSIONAL HYDRODYNAMIC MODEL OF THE OHIO RIVER, JEFFERSON COUNTY, KENTUCKY-U.S. Geological Survey Water-Resources Investigations Report 01-4091
U.S. GEOLOGICAL SURVEY 9818 BLUEGRASS PARKWAY LOUISVILLE, KY 40299-1906
Library Rate
@ Prinled on rec:yclecl paper