Study Area

The study reach includes a 9.9-mile (mi) reach of Cuyahoga River and a 6.5-mi reach of the Ohio Canal, also known as the Ohio and Erie Canal, in south-central Cuyahoga County. The area borders the communities of Independence, Valley View, Garfield Heights, Brooklyn Heights, Cuyahoga Heights, Newburgh Heights, and Cleveland, Ohio (fig. 1). The geographic limits of the hydraulic analyses of the Cuyahoga River are the Newburgh Heights streamgage at the downstream end and Hillside Road (which is about 2.4 mi upstream from the Independence streamgage) at the upstream end. During times of flooding, it is possible for streamflow to pass either from the Cuyahoga River to the Ohio Canal or in the reverse direction in various areas of low-lying terrain along the study reach. The contributing drainage areas of the Independence and Newburgh Heights streamgages are 707 and 788 square miles, respectively (table 1). The two streamgages have different periods of record; the Independence streamgage was established in September 1921. The area closest to the Independence streamgage is primarily urban, but about 500 feet (ft) upstream from the streamgage the river flows through the Cuyahoga Valley National Park (50.9 square miles of mostly forested and undeveloped area; National Park Service, 2024). The Newburgh Heights streamgage was established in October 1991. It is surrounded by urban and industrial areas and is subject to backwater from Lake Erie; it was used in some calculations, but it was not useful for establishing a stage-streamflow relation as any particular stage could be associated with multiple streamflows.

Previous Studies

The most recent flood insurance study for Cuyahoga County, Ohio was published in August of 2019 (FEMA, 2019). That study re-delineates the areas prone to large floods corresponding to 10-, 2-, 1-, and 0.2-percent AEP floods from work that originated in 1998. The 2019 flood insurance study detailed flooding analyses for the Cuyahoga River that include, and extend beyond, the upstream and downstream limits of the study reach. Selected AEPs for the Independence and Newburgh Heights streamgage locations are shown in table 3.

Standard Procedures for Creating a Flood Map

The USGS has standardized procedures for creating flood-inundation maps for flood-prone communities to ensure that the methods used and products produced are consistent (USGS, 2018). Tasks specific to development of the flood maps were to (1) collect topographic, bathymetric, and geometric data for selected cross sections and structures (such as bridges or culverts) along the study reach, (2) estimate energy-loss factors (roughness coefficients) in the stream channel and floodplain, (3) determine streamflows for each stage to be modeled, (4) compute and calibrate water-surface profiles using HEC–RAS (USACE, 1995, 2024), (5) produce estimated flood-inundation maps for selected stages by using the RAS Mapper (a feature within HEC–RAS; USACE, undated) and a geographic information system, and (6) prepare maps as shapefile polygons that depict the extent of flood inundation and as depth grids that provide the depth of floodwaters for display on a USGS flood inundation mapping application. These methods follow procedures described in Bales and others (2007) and Whitehead and Ostheimer (2009). Techniques that were modified significantly from previously documented methods to accommodate local hydrologic conditions or availability of data are described in detail in this report.

Topographic Data

All the topographic data used in this study are referenced vertically to the North American Vertical Datum of 1988 (NAVD 88) and horizontally to the North American Datum of 1983 (NAD 83). Cross-section elevation data were obtained from a digital elevation model (DEM) that was derived from lidar data collected from November 2019 to March 2020 for the Ohio Geographically Referenced Information Program (undated) as part of the USGS 3D Elevation Program (USGS, undated a). The DEM is USGS Quality Level 3 (USGS, undated a) with a cell size of 1.25 ft, a nominal pulse spacing of 1.4 m (0.43 ft), and a vertical root mean square error of 20 centimeters (0.60 ft). Lidar is a technology comprised of a global positioning system, an inertial navigation system, and a laser scanner (typically mounted in a small aircraft) that transmits brief pulses of light to the ground surface (USGS, 2024e). Those pulses are reflected, or scattered back, and their travel time is used to calculate the distance between the laser scanner and the ground.

RAS Mapper within HEC–RAS was used to extract distance and elevation data from the DEM for all 211 cross sections (table 5) for use in HEC–RAS. The DEM-derived cross sections were co-located with the field-surveyed cross sections where possible. In those cases, DEM-based elevations were combined with the survey elevations to form composite cross sections. In-channel elevations and dimensions for DEM-derived cross sections that did not have surveyed elevations were estimated by interpolating between the closest field-surveyed cross sections as a function of distance along the hydraulic baseline. The average distance between cross sections (both conventionally surveyed and DEM-derived) was 246 ft.

Bathymetric and Structure Data

Channel cross sections were surveyed to collect elevations and dimensions below the water surface that are not provided by conventional light detection and ranging (lidar). Structures that had the potential to affect water-surface elevations during floods along the streams were also surveyed.

The USGS used a differential global positioning system and differential-leveling (hereafter referred to as conventional) surveys for this study. Differential global positioning system surveys were completed using Level III real-time network surveying techniques (Rydlund and Densmore, 2012) and were used to establish an elevation control network at selected locations along the study reach. Elevations determined using the differential global positioning system at five benchmark locations (table 6) had a root mean square error of 0.10 ft when compared with National Geodetic Survey published elevations (National Oceanic and Atmospheric Administration, undated a).

The USGS used the elevation control network to complete conventional surveys at each structure to obtain its geometry. All conventional surveys were done to third-order accuracy criteria in both horizontal and vertical directions (Federal Geodetic Control Committee, 1984). USGS field crews surveyed a total of four hydraulic structures along the Cuyahoga River from February to May of 2023. Geometric data for an additional 12 structures were obtained from as-built plans from the Cuyahoga County Engineers Office (Cuyahoga County, Ohio, 2024).

Bathymetry data for cross sections were obtained from the U.S. Fish and Wildlife Service that were collected as part of their Cuyahoga River Lake Sturgeon Reintroduction Plan (Fischer and others, undated; Fischer, 2021; Lebson, 2024). The U.S. Fish and Wildlife Service used an acoustic Doppler current profiler equipped with a differential global positioning system receiver to obtain bathymetry at 87 cross-section locations within the study reach during April and May of 2022. In the spring of 2023, the USGS used conventional surveys to obtain data for seven additional cross sections to ensure that no reach length between cross sections was greater than 1 mi. The maximum distance between cross sections was 1,163 ft, with an average distance of 551 ft. In total, field data were collected at 94 cross sections and included 16 structures (table 5).

Energy-Loss Factors

Hydraulic analyses require the estimation of energy losses that result from frictional resistance between the streamflow and the channel. The amount of frictional resistance may be quantified by the Manning’s roughness coefficient (“n” value). Initial (precalibration) n values were selected based on field observations and high-resolution aerial photographs (Esri and others, 2009). As part of the water-surface elevation calibration process, the initial n values were adjusted until the differences between computed and observed water-surface elevations at the Independence streamgage were minimized. For the Cuyahoga River, the final n values ranged from 0.035 to 0.041 for the main channel and from 0.030 to 0.076 for the overbank (floodplain) areas (table 5). As noted in the “Study Area” section, the floodplains in the Cuyahoga Valley National Park are mostly open grassy fields or forested areas, whereas the floodplains from near the Independence streamgage to the downstream limit are primarily urban. The stream channel throughout the study reach is mainly composed of sand, gravel, and cobble; tall grasses and some light underbrush grow along the channel banks.

Hydrologic Data

The study reach includes two streamgages (fig. 1; tables 1 and 2). The stage is measured every 15 minutes, transmitted hourly by a satellite radio in the streamgage, and made available on the internet through the USGS National Water Information System (USGS, 2024d). Stage data for the Independence and Newburgh Heights streamgages are referenced to a local datum but can be converted to water-surface elevations referenced to NAVD 88 by adding 582.66 ft and 502.41 ft, respectively (USGS, 2024a, 2024b). Vertical datum surveys for the Independence and Newburgh streamgages were completed in April of 2013 using Level III real-time network surveying techniques (Rydlund and Densmore, 2012) and have a vertical accuracy of 0.09 foot. Continuous records of streamflow (USGS, 2024d) are computed from a stage-streamflow relation (USGS, 2024c) which has been developed for the streamgages, and are also available through the USGS National Water Information System website.

For the profiles up to a stage of 22.00 ft (table 7), the streamflows used in the model simulations were obtained from the current stage-streamflow relation (rating curve 43.0) for the Independence streamgage. For profiles above a stage of 22.00 ft, streamflows were obtained from a theoretical rating curve developed as part of this study (table 7). The theoretical rating curve is based on the same hydraulic model used in this study model whereby streamflows in the model were increased until the resulting water-surface elevation at the Independence streamgage matched the desired targeted stages (stages 23.00 ft, 24.00 ft, 25.00 ft, and 26.00 ft).

There are several small tributaries within the study reach (fig. 1). The largest of the tributaries is Big Creek (more than twice the size of the next biggest tributary in the study area), with a drainage area of 37.6 square miles at the confluence with the Cuyahoga River. These tributary flows need to be factored into the hydrologic model of the Cuyahoga River. To ensure that the Cuyahoga River streamflows were not underestimated and to simplify the adjustment calculation, the target streamflows on the Cuyahoga River downstream from Big Creek were increased (table 7) using a simple ratio of drainage areas at the Newburgh Heights and the Independence streamgages (table 8). The adjusted streamflow in the Cuyahoga River downstream from Big Creek was found by multiplying the Independence streamgage by the drainage area at Newburgh Heights streamgage and dividing by the drainage area at the Independence streamgage. The drainage area of the upstream end of the reach has only a small difference from the drainage area of the downstream end of the reach (1.1 percent and 11.5 percent respectively; table 8).

Computation of Water-Surface Profiles

The water-surface profiles used to produce the 13 flood-inundation maps in this study were computed by using HEC–RAS version 6.3.1 (USACE, 1995). HEC–RAS is a one- or two-dimensional step-backwater model used to simulate water-surface profiles with steady-state (gradually varied) or unsteady-state flow computation options. The one-dimensional steady-state option was used for this study.

Hydraulic Model

The HEC–RAS analysis for this study was completed using the steady-state flow computation option. Steady-state flow inputs were the flow regime, boundary conditions, and streamflow values that produced water-surface elevations at the streamgage cross section that matched target water-surface elevations. These target elevations coincided with 1-ft increments of stage referenced to the local gage datum for the Independence streamgage. A subcritical (tranquil) flow regime was assumed for the simulations.

The Newburgh Heights streamgage is subject to backwater from Lake Erie. This precludes the establishment or use of a traditional stage-streamflow relation because a particular stage at Newburgh Heights streamgage can be associated with multiple streamflows. Still-water elevations are defined by FEMA (2019) as the surface of the water resulting from astronomical tides, storm surge, and freshwater inputs, but excluding wave setup contribution or the effects of waves. As a result of backwater, the hydraulic model downstream boundary conditions (table 9) for the various profiles were based on information for the 10-percent to 0.2-percent AEP floods (that is, the corresponding water-surface elevations and streamflows) at the Newburgh Heights streamgage as given in the 2019 flood insurance study for Cuyahoga County (fig. 2; FEMA, 2019).

Flood profiles associated with stages of 22.00 to 26.00 ft at the Independence streamgage have corresponding drainage-area adjusted streamflows (table 7; see discussion in “Hydrologic Data”) that are between the 10-percent and 0.2-percent AEP floods at the Newburgh Heights streamgage (tables 3 and 7); therefore, the starting boundary conditions for these five profiles were set to known water-surface elevations (table 9) based on interpolation between the 10-percent to 0.2-percent AEP water-surface elevations in the 2019 flood insurance study (FEMA, 2019) at the Newburgh Heights streamgage.

Normal depth is defined as the depth when the streamflow is uniform, steady, one-dimensional, and unaffected by downstream obstructions. Streamflow is considered uniform if the energy line, water surface, and channel bottom all are parallel (Chow, 1959). For stages 14.00 to 21.00 ft at the Independence streamgage, the corresponding drainage-area adjusted streamflows are all less than the 10-percent AEP flood (16,505 ft³/s) at the Newburgh streamgage (tables 3 and 7). To avoid a straight-line extrapolation of the water-surface elevations below the 10-percent AEP flood at the Newburgh Heights streamgage, the starting water-surface elevations for stages 14.00 to 21.00 ft at the Newburgh Heights streamgage were all based upon a single normal-depth slope of 0.000645 ft/ft (fig. 2). This normal-depth slope was found by calibrating the model’s starting slope to reproduce the 10-percent AEP flood water-surface elevation (580.30 ft; FEMA, 2019) associated with the corresponding streamflow at the Newburgh Heights streamgage.

The resulting starting water-surface elevations for stages 14.00 to 16.00 ft could be affected by backwater from Lake Erie for a near 10-percent AEP storm surge flood event or a less frequent AEP storm surge flood event (fig. 2). The chances are unknown, but likely small, for a simultaneous Lake Erie storm surge event and a flood event on the Cuyahoga River for stages 14.00 to 16.00 ft. Even with a 0.2 percent AEP storm surge event on Lake Erie, the flooding for stages 16.00 ft and lower are contained within the channel of the Cuyahoga River from the Newburgh Heights streamgage upstream beyond the Independence streamgage.

A comparison of the modeled and target water-surface elevations at the Independence streamgage is given in table 10. As described in the "Hydrologic Data" section, the target water-surface elevations for stages 23.00 to 26.00 ft in the current (2024) stage-streamflow relation (rating number 43.0) are based on the theoretical rating developed from this modeling effort. As a result, the differences in the modeled to target water-surface elevations for stages 23.00 to 26.00 ft were excluded from any root mean square error calculations of the differences between modeled to target water-surface elevations. The root mean square error of the differences between the modeled to target water-surface elevations for stages 14.00 to 22.00 ft is 0.34 ft.

The error between the modeled and target water-surface elevations is larger for the lower water-surface stages. One potential reason for this difference could be the assumption of a constant normal-depth starting water-surface slope for stages of 14.00 to 21.00 ft. The normal depth slope likely varies with streamflow, but due to the variable backwater conditions at the Newburgh Heights streamgage it became very difficult to estimate the slope. All the flooding in stages 17.00 ft and lower have modeled water-surface elevation errors greater than 0.20 ft; almost all of the flooding extent of these stages is contained within the channel of the Cuyahoga River. The only flooding outside the channel for these stages is within the undeveloped Cuyahoga Valley National Park area immediately upstream from Rockside Road.

Development of Flood-Inundation Maps

Flood-inundation maps for the 13 water-surface profiles were created using geographic information system software by combining the water-surface profiles and DEM data. Estimated flood-inundation boundaries and depth grids were developed for each simulated water-surface profile with RAS Mapper—a feature within HEC–RAS (USACE, undated) that allows the preparation of geometric data for import into HEC–RAS and processes simulation results exported from HEC–RAS. Shapefile polygons and depth grids of the inundated areas for each water-surface profile were manually edited using geographic information system software to ensure a hydraulically reasonable transition of the flood boundaries among modeled cross sections.

Any inundated areas that were detached from the main channel were examined to identify subsurface connections with the Cuyahoga River or the Ohio Canal, such as culverts or similar engineering structures under roadways. Where such connections existed, the mapped inundated areas were retained in their respective flood maps; otherwise, the erroneously delineated parts of the flood extent were deleted. The flood-inundation areas were overlain on high-resolution, georeferenced aerial photographs (Esri and others, 2009) of the study area. One example of a profile (stage 26.00 ft) overlain on orthorectified imagery can be seen in figure 3. Bridge surfaces are shown as not inundated up to the lowest flood stage that completely inundates one or both road approaches to the bridge. Estimates of water depths can be determined from the depth-grid data that are included with the maps on an interactive mapping application (USGS, undated b) described in the following section, “Data Dissemination.”