Hydrologic and Unsteady-Flow Data

The study reach includes two USGS streamgages (fig. 1). Stage is measured every 15 minutes, transmitted hourly by a satellite radio in the streamgage, and made available on the internet through the National Water Information System (NWIS; USGS, 2021b, c, e). Stage data from these streamgages are referenced to a local datum but can be converted to water-surface elevations, which are referenced to the North American Vertical Datum of 1988 (NAVD 88), by adding an offset: 793.92 ft at the Schoharie Creek at North Blenheim (01350180) streamgage and 766.54 ft at the Schoharie Creek near North Blenheim (01350212) streamgage. Continuous records of discharge are computed from stage-discharge ratings, which are developed for the streamgages and are available through NWIS.

Ideally, all tributaries and other contributing flows would be included individually in a hydraulic model, though typically those contributing a small percentage of the total flow are lumped with the main inflows. A few tributaries join the Schoharie Creek within the 2.4-mi mapped study reach, most of which are minor, unnamed tributaries with small drainage areas (less than 2 mi²). The only substantial tributary to the Schoharie Creek in the mapped study reach is the West Kill, which has a drainage area of 45.4 mi², or about 11 percent of the total drainage area for the mapped reach (407 mi²), and joins the Schoharie Creek approximately midway through the reach (fig. 1).

The peak flows used in the model simulations (table 2) correspond to the target stages for the inundation maps referenced to the Schoharie Creek near North Blenheim (01350212) streamgage, and to flows from flood events in 2011, 1996, and 1987. The peak flows for the inundation maps were extracted from the active stage-discharge rating (number 3.0, effective beginning in October 2019 and updated April 2021; USGS, 2022b); the stages for the three flood events were determined from high-water marks, and peak discharge at the Schoharie Creek near North Blenheim (01350212) streamgage was computed from observed peak flows at upstream and downstream streamgages (Schoharie Creek at North Blenheim, N.Y., station number 01350180, and Schoharie Creek at Breakabeen, N.Y., station number 0130355) by using drainage-area ratios and regional flood equations (Lumia and others, 2006).

Two inflow time series were used for each simulation in the model: one at the upstream end of the modeled reach of Schoharie Creek and one at the mouth of West Kill (fig. 1). Peak flows at the upstream boundary condition of Schoharie Creek were calculated by using a flow ratio of 92 percent of the peak flow at the Schoharie Creek near North Blenheim (01350212) streamgage; this ratio was computed by using the regional flood equations (Lumia and others, 2006) and by comparing observed peak flows at streamgages upstream and downstream from the Schoharie Creek near North Blenheim (01350212) streamgage, including the Schoharie Creek at North Blenheim (01350180) and Schoharie Creek at Breakabeen (0130355) streamgages. Target water-surface elevations for model calibration at the upstream streamgage, Schoharie Creek at North Blenheim (01350180), were extracted from that station’s stage-discharge rating, corresponding to the inflow discharge for each inundation map. For model simulations used only for calibration and not for inundation mapping (the 2011, 1996, and 1987 high-flow events), peak flows at the upstream boundary condition on the Schoharie Creek were taken from observed peak flow values at the Schoharie Creek at North Blenheim (01350180) streamgage. Peak flow at the upstream boundary condition on West Kill was calculated by applying a drainage-area ratio of 11.06 percent of the peak flow at the Schoharie Creek near North Blenheim (01350212) streamgage. The sum of the peak flows at the two upstream boundary conditions listed in table 2 are not expected to equal the peak flow listed for the Schoharie Creek near North Blenheim (01350212) streamgage because the peak flow for a high-flow event at each of the locations is not expected to occur at the same time (the peak inflow from West Kill generally occurs before the peak flow on the mainstem of Schoharie Creek).

Because a 2D HEC–RAS hydraulic model was used (see the “Hydraulic Model” section of this report), unsteady-flow hydrographs were required at flow input boundary conditions; these hydrographs, which are included in the files associated with the HEC–RAS model published in Nystrom (2022), were generated on the basis of an observed storm on August 13–14, 2018. This event was chosen because it was the largest storm with time series of observed flows available at both streamgages in the study reach (fig. 2A; streamgage information in table 1). The time series from the storm were normalized so that the peak of each time series equals 1 (fig. 2B); to create each boundary condition hydrograph, the normalized time series was multiplied by the peak flow for the boundary condition at each modeled stage (fig. 2C; table 2). Because there are no continuous-record streamgages on West Kill, the normalized time series for that inflow boundary condition was created by using the normalized time series at the Schoharie Creek at North Blenheim (01350180) streamgage, time-shifted 2 hours prior to the peak at 01350180. The time shift value was approximated by visually comparing hydrographs at the two North Blenheim streamgages (01350180 and 01350212).

Topographic and Bathymetric Data

All topographic data used in this study are referenced vertically to NAVD 88 and horizontally to the North American Datum of 1983 (NAD 83). A hydraulic terrain was created in HEC–RAS Mapper (USACE, 2020a) by using elevation data that were obtained from a digital elevation model (DEM; fig. 3), which was derived from light detection and ranging (lidar) data collected in May and June 2014 by The Atlantic Group for the USGS (USGS, 2015). Postprocessing of these data was completed by Dewberry in October 2014 (Dewberry, 2015). The lidar DEM has a horizontal resolution of 3.3 ft (1 meter) and an overall vertical accuracy of 0.52 ft. By these criteria, the lidar data support production of 2-ft elevation contours (Dewberry, 2012). Elevation data were extracted from the DEM for the hydraulic model in HEC–RAS.

Because the lidar data did not provide ground elevations below the water surface of the stream, channel cross sections were surveyed by USGS field crews during August and October 2019. Wadeable cross sections were measured in August 2019 by using a survey-grade Trimble Inc. R8 global navigation satellite system (GNSS) receiver. Real-time kinematic (RTK) observations of location and elevation were made at 661 points in 39 cross sections (fig. 3) by using a virtual reference station from the NYSNet continuously operating real-time network (RTN) operated by the New York State Department of Transportation (2021). The median horizontal precision of points was 0.17 ft, and the median horizontal precision was 0.19 ft. Areas that were too deep to wade were measured by using hydroacoustic instrumentation in October 2019. Depths at 6,059 points in two areas were measured by using the vertical beam of a SonTek M9 acoustic Doppler current profiler with a 0.5-megahertz frequency, equipped with an RTK GNSS receiver (SonTek, 2021). Elevations of these points were established by referencing the measured depth to a water-surface elevation that was measured with the survey-grade Trimble Inc. R8 GNSS receiver with virtual reference station RTK corrections. Bathymetric data in the stream channel were imported into HEC–RAS; 1D cross sections were established at bathymetry survey locations, and elevations within the channel were updated with the surveyed values. HEC–RAS Mapper was used to create interpolated channel bathymetry, which was then merged with the lidar DEM to create a seamless combined topography and bathymetry terrain for use in the 2D hydraulic model.

Hydraulic Structures

Several structures in and near the Schoharie Creek, including a highway crossing (State Route 30), a covered bridge (fig. 1), and rock weirs and vanes, have the potential to affect water-surface elevations during floods along the stream. Hydraulic-structure-geometry data were collected from USGS RTK–GNSS field surveys in 2019.

Energy-Loss Factors

Hydraulic analyses require the estimation of energy losses that result from frictional resistance exerted by a channel on flow. These energy losses are quantified by the Manning’s roughness coefficient (n value). Flood plain n values were set by using mapped land cover from the 2016 National Land Cover Database (NLCD; Dewitz, 2019) and published tables of n values (Chow, 1959; Coon, 1998). Initial (precalibration) channel n values were selected on the basis of field observations and high-resolution aerial photographs.

The HEC–RAS model was calibrated to observed water-surface elevations at the two streamgages in the study reach and to observed high-water marks from high-flow events in 2011, 1996, and 1987 (see the “Hydraulic Model” section of this report). As part of the calibration process, the initial n values were adjusted until the differences between simulated and observed water-surface elevations were minimized. A calibration region was added to the model for the upstream part of the study reach (from the upstream boundary condition to where floodplain flow rejoins the main channel downstream from the Blenheim Town Hall) to allow independent control of n values in this region. The final n values ranged from 0.025 to 0.038 in the channel and 0.035 to 0.12 in the overbank areas.

Hydraulic Model

The study area is hydraulically complex, with multiple, often disconnected, flow paths in the flood plain. During high flows, including at all mapped stages in this study, water leaves the channel at a bend at the upstream end of the study reach, flowing over the channel bank and across the floodplain, before rejoining the channel downstream from the Town of Blenheim offices. This flow across the floodplain during high flow bypasses the channel at the Schoharie Creek at North Blenheim (01350180) streamgage. In this section of the study reach, water-surface elevations in the channel do not necessarily match water-surface elevations in the floodplain along drawn 1D cross-sections; a 1D HEC–RAS model assumes a single water-surface elevation across an entire cross-section, and proportions flow within any subchannels independently from cross-section to cross-section, in this case leading to unrealistic water-surface elevation profiles along the reach and discontinuous flow across the floodplain. Therefore, the best approach for the study area was determined to be a 2D model (fig. 3). The 2D model uses a base cell size of 50 ft, with breaklines added along linear features (such as roads), resulting in a total of 12,596 cells in the 2D flow area.

The hydraulic model was constructed by using the unsteady-flow option in HEC–RAS. Boundary conditions for the unsteady-flow data consisted of flow hydrographs for inflow to the study reach—one for Schoharie Creek at the upstream end of the study reach and one for West Kill—and “normal depth” for the outflow at the downstream end of the modeled study reach. A friction slope of 0.0018 was used, determined from the water-surface slope in the lidar dataset at the downstream end of the modeled study reach. Peak flows (table 2) and flow hydrographs (fig. 2) used in the model are described in the “Hydrologic and Unsteady-Flow Data” section of this report. Input time series were at 15-minute intervals, which were interpolated by HEC–RAS to match the model computation time steps. The model was run at 1-second computational time steps, with a 5-minute output time step and 2-hour initial conditions ramp-up time. The shallow-water equations (“stricter momentum” equation set in the HEC–RAS computation options) were used with the conservative turbulence model.

The HEC–RAS model was calibrated primarily to the active stage-discharge rating at the Schoharie Creek near North Blenheim (01350212) streamgage (number 3.0, effective October 2019 and updated April 2021; USGS, 2022b) and secondarily to the stage-discharge rating at the Schoharie Creek at North Blenheim (01350180) streamgage (number 19.0, updated April 2021; USGS, 2022a) as well as to high-water marks from high-flow events on August 28, 2011 (Lumia and others, 2014), January 19, 1996 (Lumia, 1998), and April 4, 1987 (Zembrzuski and Evans, 1989). Based on the flood-frequency statistics for 01350180, the August 2011 high-flow event (peak discharge of 119,000 ft³/s) had an AEP of about 0.4 percent, or a recurrence interval of 250 years (Wall and others, 2014), and the 1996 and 1987 events (peak discharges of 75,600 ft³/s and 64,200 ft³/s, respectively) had AEPs between 4 and 2 percent (recurrence intervals between 50 and 25 years, respectively; Zembrzuski and Evans, 1989; Lumia, 1998; Wall and others, 2014).

The model was calibrated by adjusting Manning’s n values (as described in the “Energy-Loss Factors” section of this report) until the results of the hydraulic computations closely agreed with the stage-discharge ratings (table 3) and with the observed high-water marks for given flows (table 4). Differences between the stage-discharge rating at the Schoharie Creek near North Blenheim (01350212) streamgage and simulated water-surface elevations ranged from −0.15 ft to 0.21 ft over the mapped stages, with a root mean square error (RMSE) of 0.13 ft. At the Schoharie Creek at North Blenheim (01350180) streamgage, the secondary calibration location, the difference between the stage-discharge rating and simulated water-surface elevations ranged from −0.33 ft to 0.96 ft, with a RMSE of 0.56 ft; the flow over the floodplain that bypasses the streamgage in this location may contribute to the larger differences between rated and simulated elevations. Differences between observed and simulated elevations of high-water marks in the study reaches ranged from −0.56 ft to 0.79 ft. The calibrated hydraulic model was used to generate maps of maximum water-surface elevation for each of the 13 peak flows (Nystrom, 2022).