Key points are not available for this paper at this time.
The National Water Center (NWC), operated by the National Oceanic and Atmospheric Administration's (NOAA) Office of Weather Prediction and located on the campus of the University of Alabama, is the hub for the new National Water Model (NWM) of the United States (U.S.). In 2015, the NWC Innovators' Program and the Consortium of Universities for the Advancement of Hydrologic Science, Inc. launched the inaugural NWC Summer Institute to engage the academic community in developing applications of the NWM. Held annually since then, the Summer Institute brings a group of graduate students to the NWC and the University of Alabama to work with faculty advisors and NWC staff on group projects, with the goal of rapidly prototyping new ideas. During the first Summer Institute (June–July 2015), the overarching goal was to demonstrate a prototype NWM, exploring whether streamflow on 2.7 million stream reaches of the U.S. could be simulated and forecast in real time using NOAA weather products and the National Center for Atmospheric Research Weather Research and Forecasting (WRF)-Hydro model. These results were summarized in a featured collection in the Journal of the American Water Resources Association in 2017 (Volume 53, Issue 2; see Nelson 2017). The second Summer Institute (June 6–July 20, 2016) included 34 graduate students from 21 U.S. universities, supported by two Student Coordinators (Adnan Rajib and Peirong Lin), five Research Theme Leaders (Sagy Cohen, Ibrahim Demir, Alfonso Mejia, Sarah Praskievicz, and Albert Van Dijk), and the students' academic advisors at their home institutions. The 2016 Summer Institute was led by David Maidment and Ed Clark. By June 2016, a first version of the NWM was in the process of being made operational on NOAA computational facilities, and so the focus of the 2016 Summer Institute was on translating the NWM forecasts of discharge into flood-inundation mapping and flood emergency response (Maidment et al. 2016), broadly grouped into four research themes: Inundation Mapping, Flood Modeling, Forecast Errors, and Emergency Response. This featured collection includes papers representative of these research themes, both by student participants in the 2016 Summer Institute and by researchers who facilitated it. In support of the Summer Institute activities, a 10-m National Elevation Dataset of the U.S. was analyzed on the CyberGIS facility in the National Center for Supercomputer Applications of the University of Illinois at Urbana-Champaign and transformed into a raster called Height Above Nearest Drainage (HAND), in which each cell contains the height difference between its elevation and the elevation of the cell in the National Hydrography Dataset Plus (NHDPlus) stream reach to which it drains. The paper by Liu et al. (2018) describes the geoprocessing methodology and advanced computational and cyberinfrastructure approaches they developed in order to calculate this national-scale 10-m HAND database. The HAND dataset can be used to map flood-inundation extent by identifying the cells whose HAND values are less than the water depth at their nearest stream. Coupled with flood forecasting from the NWM, the HAND approach establishes, for the first time, a foundation for locally informative, real-time flood-inundation mapping continuously across the continental U.S. The paper by Zheng et al. (2018) addresses a key challenge in this approach, converting NWM water discharge predictions into water depth (flow stage) data in each of the NWM's 2.7 million stream reaches. They proposed and tested a methodology for deriving channel geometry and synthetic rating curves from the 10-m HAND dataset for each stream reach. Traditionally, rating curves between discharge and stage are based on surveyed channel geometry and long-term measurement of streamflow at a gaging site. Since such information is not available for the vast majority of the NWM stream reaches, the approach proposed by Zheng et al. (2018) provides an alternative that enables the coupling of NWM predictions and HAND inundation estimates. Several papers in this featured collection focused on the use of hydraulic modeling for flood-inundation predictions. The paper by Zarzar et al. (2018) proposed an approach for quantification and visualization of uncertainty in flood-inundation simulations. They used the two-dimensional International Rivers Interface Cooperative — Flow and Sediment Transport with Morphological Evolution of Channels (iRIC-FaSTMECH) and Hydrologic Engineering Center River Analysis System (HEC-RAS) models to compile an ensemble of flood-inundation predictions based on NWM forecasts for an urban stream. They developed a web-based decision support tool (using the Tethys platform, Swain 2015) to disseminate the uncertainty in inundation forecasts through a series of probability maps. This framework is envisioned to enhance the communication capabilities of NWM predictions. The paper by Zhang et al. (2018) presents the results of a comparative analysis of inundation mapping approaches for the 2016 flood on the Brazos River in Texas. They compared the simple terrain-based HAND approach against the iRIC-FaSTMECH model. Flood-inundation extents estimated using both methods were compared to observed extents based on satellite remote sensing. The results of the intercomparison indicate, although the iRIC-FaSTMECH model produced more accurate inundation extents, the HAND method still performed reasonably well, especially when modified to explicitly account for intercatchment flows between the mainstem and tributaries. This finding is important because the minimal data requirements and computational cost of the HAND method mean it is much more feasible to couple to NWM forecasts for near-real-time flood-mapping than physically based hydraulic models. Zhang et al. (2018) demonstrated the value of using observed (i.e., remote sensing-based) flood inundation for model comparison. The paper by Munasinghe et al. (2018) compared several floodwater classification techniques, based on Landsat 8 satellite imagery, to a reference inundation map based on manual digitization and ancillary data for the 2016 Brazos flood. The compared techniques were supervised classification, unsupervised classification, delta cue change detection, normalized difference water index, and modified normalized difference water index. The authors found supervised classification resulted in the highest accuracies. These results are important for producing high-quality observed flood-inundation maps that can be used for validating models and assessing flood impacts. Cohen et al. (2017) developed a novel floodwater depth detection algorithm based on remotely sensed flood-inundation extents and topography from digital elevation models (DEMs). Their algorithm identifies boundary grid cells for the inundated area, extracts their elevations from the DEM, and calculates the flood depth by subtracting the elevation of each inundated grid cell from its nearest boundary cell. Test applications of the method to floods on the Brazos River in Texas and St. Vrain Creek in Colorado indicated depths estimated by the new algorithm correspond well with depths simulated by iRIC-FaSTMECH. The algorithm contributes to the utility of satellite remote sensing in flood emergency response and damage assessment, because such applications require information on the depth as well as extent of flooding. The paper by Souffront Alcantara et al. (2018) presents the development of cyberinfrastructure and web apps designed to manage and disseminate NWM results. The web apps were designed using the Tethys platform, a framework for developing water resources web applications, which student participants have used in every Summer Institute. In this paper, the authors used Tethys to develop two web apps for accessing NWM output: the NWM Explorer App, which is designed to act as a catalog to search NWM forecast files; and the NWM Viewer app, which was developed to provide visualization and extraction of output from the NWM. In addition, the authors used Tethys to develop three additional apps designed to demonstrate the usefulness of NWM forecasts: a Gauge Viewer app that allows for comparisons between NWM and observed streamflow, and two flood-mapping apps for Tuscaloosa County, Alabama, and Greenbrier County, West Virginia. The development of these publicly available web apps contributes to the dissemination and visualization of NWM results, and is a useful approach for moving modeling results to intended users. The Technical Note by Johnson et al. (2018) presents the development and application of "FloodHippo," a flood alert system that integrates operational model outputs, cloud-based geographic information systems, and expanded communication channels. The FloodHippo prototype is designed to address several shortcomings of current flood alert systems, namely that they are spatially vague, temporally imprecise, and lack actionable information. FloodHippo includes an Awareness Portal, with a web map displaying near-real-time flood information — NWM forecasts, NOAA warnings, Federal Emergency Management Agency flood zones, and social media data — at the street-address level. In order to make this information actionable, FloodHippo also includes a Response Portal, which focuses on navigation, allowing users to plot routes to avoid flooded areas or to evacuate to the nearest accessible shelter. Such innovative web applications are designed to engage users with more actionable intelligence, which has the potential to improve awareness and weather-readiness among the public. The final paper in this featured collection (Omranian and Sharif 2018) evaluates the accuracy of different satellite-derived precipitation products over the Lower Colorado River in Texas. The accuracy of any hydrological model is dependent upon the robustness of its precipitation input. In forecasting modeling frameworks (e.g., NWM), spatial and temporal scaling of recent and forecasted precipitation data add a layer of complexity which is challenging to evaluate. Omranian and Sharif (2018) evaluated the impact of spatial and temporal downscaling of the Integrated Multi-satellitE Retrievals for Global Precipitation Measurement mission (IMERG) products by comparing them to in situ observations at 214 rain gauges. They found IMERG products better correspond to observed precipitation when downscaled, but more improvement is needed. The papers in this featured collection illustrate the diversity of topics addressed by student teams and associated researchers in the 2016 NWC Summer Institute. These projects include demonstrations and prototypes of innovative concepts such as near-real-time flood-inundation mapping, integration of remote sensing into flood assessment, and development cyberinfrastructure and web apps for water resources. Research like this is critical to advancing the mission of the NWC to address 21st-Century water challenges.
Cohen et al. (Wed,) studied this question.