Terrestrial Water Cycle: Hassan Mashriqui, James Halgren and Seann Reed

NOAA's National Weather Service, Office of Hydrologic Development


In coastal rivers, there is a river-estuary transition zone where water level may be influenced by freshwater flows, tides, storm tides, wind or a combination of these factors. The National Oceanic and Atmospheric Administration (NOAA) is working towards eliminating service gaps and improving the accuracy of water level forecasts in the transition zone. One method to improve transition zone forecasts is extending a one-dimensional river hydraulic model downstream well into the tidal estuarine environment. Recently we have evaluated the effectiveness of this approach for the tidal Potomac River by developing an unsteady, one-dimensional Hydrologic Engineering Center River Analysis System (HEC-RAS) model. The model extends from Little Falls, Washington, D.C. to Lewisetta, Virginia. We calibrated the model for harmonic tides and major historic freshwater flood events and validated the model on recent freshwater and storm surge events, focusing on water level gauges near Washington, D.C. The model is now implemented for operational forecasting at the NWS Mid-Atlantic River Forecast Center.

From the HEC-RAS model, root mean squared error for tide simulation at the Washington Waterfront gauge was 0.05 m (0.16 ft) with a phase error of about 2 hours. For historic flood events, simulated peak water level error varied from -0.41 m (-1.34 ft) to 0.41 m (1.34 ft) with a mean absolute error of 0.25 m (0.82 ft) at the Wisconsin Avenue gauge. HEC-RAS shows considerable potential for improvement over existing forecast techniques, particularly for events where the combined influence of freshwater flow and tides are important. However, HEC-RAS simulated peak storm tide during Hurricane Isabel at the Washington Waterfront gauge was about 0.65 m (2.13 ft) lower than the observed peak. The lower performance in predicting peak surge during Hurricane Isabel is due to the fact that HEC-RAS does not include an explicit wind forcing term. The HEC-RAS model only indirectly accounts for wind forces by propagating wind induced surge from the downstream boundary.

To further quantify the influence of wind in a 1D model, we developed and ran a 1D SOBEK model with several wind scenarios. A SOBEK model run with a calibrated wind reduction factor matched the observed peak surge from Hurricane Isabel at the Washington Waterfront. In our SOBEK runs, a single wind time series was applied to entire model domain. For additional comparisons, we ran a 2D ADCIRC model for Hurricane Isabel. This model accurately predicted the peak surge level at Washington Waterfront after spatially variable bottom friction factors were calibrated. While wind driven simulations from SOBEK and ADCIRC improved peak water level simulation for Hurricane Isabel, little improvement was seen over HEC-RAS in peak time prediction. The use of both observed and simulated wind scenarios to force SOBEK highlight the sensitivity of model results to wind reduction factors and drag coefficient parameterization. To reduce reliance on calibration, we recommend the use of spatially variable wind data and/or physically based estimates of wind reduction factors in future 1D model implementations.

In parallel to the efforts described here to implement loosely coupled river-estuary-ocean models within existing NWS operational frameworks, we are evaluating more complex hydrologic and hydrodynamic modeling techniques capable of producing new forecast products in the future. As an example, we will describe initial experiments with a dynamically coupled 1D-2D model for the Chesapeake-Delaware Bay System using MikeFlood software.

Short CV:

Dr. Hassan Mashriqui is a licensed professional engineer and civil/water resources engineer. He has a Ph.D. in Civil Engineering from Louisiana State University. He is currently developing new hydrologic/hydraulic river-estuary-ocean modeling capabilities for use throughout the NWS. Prior to becoming an OHD/NWS employee in 2008, he worked for the LSU Center for the Study of Public Health Impacts of Hurricanes utilizing storm surge forecasting. His research interests include coastal and inland flooding due to hurricanes; hydrodynamic and sediment transport modeling; wetland restoration; river management; Geographic Information Systems and Light Detection And Ranging (LIDAR) technology-based environmental modeling. He was actively involved with the Natural Systems Modeling Group and Laboratory at the LSU School of the Coast and Environment in support of numerous coastal restoration research efforts.

James Halgren is a scientist at Riverside Technology, inc. with expertise in hydrologic and hydraulic modeling and model development. James provides contract services to the Hydrologic Science and Modeling Branch within the National Weather Service Office of Hydrologic Development and was most recently part of the team developing the geoSMPDBK software for rapid development of quantitative dam break flood forecasts. James received his undergraduate and Master's degrees in Civil Engineering at BYU and is completing a Ph.D. combining modeling, GIS, and rapid visualization at Colorado State University.

Dr. Seann Reed is the Hydraulics Group Leader within the National Weather Service Office of Hydrologic Development (OHD) Hydrologic Science and Modeling Branch (HSMB). The Hydraulics Group researches and develops new hydraulic modeling techniques to improve NWS hydrologic forecasts and assists NWS forecast offices in implementing new models. Active projects include evaluating and improving NWS dam break forecasting procedures, transitioning NWS operational models from FLDWAV to HEC-RAS, developing and evaluating new techniques to link river and estuary models, and developing an operational water temperature forecasting capability. Prior to becoming Hydraulics Group Leader in 2008, Seann worked for ten years on research to improve hydrologic rainfall-runoff and routing models for operational forecasting. Specific interests include the use of geo-spatial data to improve hydrologic and hydraulic models and model evaluation techniques. Seann received his Ph.D. from The University of Texas at Austin in 1998.