Water Resources Research Act Program

Details for Project ID 2020MD182B

Application of Particle Image Velocimetry and Structure from Motion to reconstruct discharge hydrographs for an extreme urban flood

Institute: Maryland
Year Established: 2020 Start Date: 2020-03-01 End Date: 2021-02-28
Total Federal Funds: $21,962 Total Non-Federal Funds: $43,923

Principal Investigators: Andrew J. Miller

Abstract: The Ellicott City floods of July 30, 2016 and May 27, 2018 have attracted national and regional attention as two events occurring in the same small suburban watershed with estimated rainfall accumulations for 2-3 hour duration that exceeded the amount expected with exceedance probability p = 0.001 (i.e. recurrence interval > 1000 years). The USGS MD-DE-DC office surveyed high-water marks and generated indirect peak discharge estimates for both events. Both events exceed any previous estimates for flood peak magnitude in a watershed of comparable size in the mid-Atlantic U.S. Like many small watersheds affected by flash floods, there are no stream gages with rating curves in the Tiber River watershed and therefore there are no recorded flood hydrographs. In the aftermath of the July 2016 flood, a local property owner, Mr. Ron Peters, installed a network of closed-circuit television cameras in Ellicott City. Most of these cameras recorded a video of 4-5 hours duration covering most of the flood event of May 27, 2018, and several captured information along the channels of the Tiber River and its tributaries. With newly available software tools and high-resolution topographic data, it is now feasible to construct discharge hydrographs using video footage. We propose to use CCTV video together with Particle Image Velocimetry (PIV) software, currently being tested by the USGS Office of Surface Water, in order to reconstruct hydrographs for the May 27, 2018 flood at several locations in the Tiber River watershed. The proposed project involves collaboration between the PI and Dr. Matthew Cashman of the USGS MD-DE-DC Water Sciences Center. In order to carry out this work we will use existing topographic data from 2018 LiDAR collected for Howard County and 2018 digital elevation data derived from Structure-from-Motion (SfM) analysis of photographs captured by a drone flight over portions of the Tiber River drainage network shortly after the flood in June 2018. These data sources will be supplemented with a new high resolution (2-5 cm pixel) topographic data set to be collected during leaf-off conditions in March 2020. We also plan to collect a topographic data set at the confluence of the Tiber River and New Cut Branch using a terrestrial laser scanner to capture information that cannot be seen from above owing to canopy cover. The resulting topographic data sets can be combined with vector fields derived from 10-second video clips using PIV software for multiple time steps during the flood in order to calculate discharge at those time steps. We already have high-resolution bias-corrected radar rainfall fields for the May 2018 event. This research will yield important results for understanding the relationship between rainfall and runoff in an extremely rare event, and will also offer support for USGS efforts to use new technical tools at gaging stations to allow reconstruction of flood hydrographs that cannot be recorded in any other way.