Streamflow Measurements During Floods Using Video Imaging
Project Personnel (University of Iowa): Dr. Marian Muste (Research Engineer), Hao-Che Ho (Graduate Student), Dr. Dongsu Kim (Postdoctoral Research Associate)
Background: Accurate and reliable flood forecasting is a data-driven process dependent upon accurate real-time streamflow data for model initialization and historical data for calibration. River forecasts are issued by the National Weather Service using about 3,800 US Geological Survey (USGS) stream gages across the country. However the accuracy of conventional methods currently used to measure discharge at USGS gauging stations during high flow (flooding) is uncertain due to a lack of historical data.
Figure 1. Deviation of the unsteady rating curve from the conventional curve obtained through extrapolation of data obtained in steady flow (non-flood conditions)
Streamflow at the nation’s gauging stations is provided by a rating curve, which relates discharge to river stage (the river water surface elevation). The rating curve (labeled SFRC in Figure 1) is obtained through repeated measurements conducted when the flow is in normal flow conditions. A one-to-one relationship between stage and discharge obtained here is assumed to be valid for the entire range of flows at the gauge site.
For high flow, the rating curve is extrapolated from data collected during lower flow conditions, so it is prone to errors from several sources. For example, the curve does not account for flood wave propagation. A storm passing through a river site actually produces a double-sided loop rating curve, as illustrated in Figure 1 by the UFRC curve. The important practical implications of this curve are that: 1) the maximum water discharge and maximum water stage do not arrive at the same time at a given river location; and 2) for the same stage, the discharge is higher during rising stage than during falling stage. The present research attempts to provide new field observations on the effect of flow unsteadiness on rating curves using an image-based technique.
Figure 2. Location of LSPIV measurements during the 2008 flood
2008 Flood Measurement Campaign: Measurements were conducted on the Iowa River in Iowa City several days before and after the flood wave peaked on June 15, 2008. The measurement site was about 500 m upstream from the location of USGS gauging station #05454500 on the Iowa River in Iowa City. Measurements were acquired with Large-Scale Particle Image Velocimetry (LSPIV). An aerial photo of the measurement site and the imaged area used for acquiring the LSPIV measurements are shown in Figure 2.
LSPIV is a technique pioneered at IIHR-Hydroscience & Engineering to measure free-surface velocities and discharges in streams without deploying personnel or equipment into the water. Thus, LSPIV can be quickly and safely deployed in adverse conditions. The results obtained with the LSPIV around the 2008 flood peak are shown in Figure 3. The plot contains measurements taken while water overtopped the spillway of the Coralville Reservoir Dam located 8.2 miles from the measurement site. The plots displayed in Figure 3 indicate that the seven LSPIV measurements did not coincide with the rating curve obtained with conventional methods. The points are grouped on separate curves, suggesting graphically the rising and falling limbs of flood wave propagation. The loop curve displays stage differences of up to 0.5m (1.5 ft) for the same discharge value, depending on whether the measurements were taken on the rising or falling stage of the flood wave propagation. Such differences are critical when the data is used for monitoring and warning systems or as input for the models providing the flood forecast.
Figure 3. USGS Rating curve and the LSPIV measurements at the USGS gaging station 05454500 during the Iowa River flood of 2008. The peak of the flood is considered as reference for time scale to facilitate the identification of the raising (“-“) and falling (”+”) limbs of the rating curve.
This study highlights the importance of acquiring high-temporal resolution discharge measurements to accurately capture the maximum stage and discharge during flood events. Techniques such as LSPIV can contribute to enhancing the accuracy of flood monitoring and forecasting by providing better quality data during extreme flows when this information is vital for protecting lives and property.
Funding Source: The National Science Foundation
