Using Digital Elevation Model Derived Height Above the Nearest Drainage for flood inundation mapping and determining river hydraulic geometry

River hydraulic geometry is an important input to hydraulic and hydrologic models that route flow along streams, determine the relationship between stage and discharge, and map the potential for flood inundation give the flow in a stream reach. Traditional approaches to quantify river geometry have involved river cross-sections, such as are required for input to the HEC-RAS model. Extending such cross-section based models to large scales has proven complex, and, in this presentation, an alternative approach, the Height Above Nearest Drainage, or HAND, is described. As we have implemented it, HAND uses multi-directional flow directions derived from a digital elevation model (DEM) using the Dinifinity method in TauDEM software (http://hydrology.usu.edu/taudem) to determine the height of each grid cell above the nearest stream along the flow path from that cell to the stream. With this information, and the depth of flow in the stream, the potential for and depth of flood inundation can be determined. Furthermore, by dividing streams into reaches or segments, the area draining to each reach can be isolated and a series of threshold depths applied to the grid of HAND values in that isolated reach catchment, to determine inundation volume, surface area and wetted bed area. Dividing these by length yields reach average cross section area, width, and wetted perimeter. Together with slope (also determined from the DEM) and roughness (Manning's n) these provide all the inputs needed for establishing a Manning's equation uniform flow assumption stage-discharge rating curve and for mapping potential inundation from discharge. This presentation will describe the application of this approach across the continental US in conjunction with NOAA’s National Water Model for prediction of stage and flood inundation potential in each of the 2.7 million reaches of the National Hydrography Plus (NHDPlus) dataset, the vast majority of which are ungauged. The continental US scale application has been enabled through the use of high performance parallel computing at the National Center for Supercomputing Applications (NCSA) and the CyberGIS Center at the University of Illinois.

Presentation at 2018 AWRA Spring Specialty Conference: Geographic Information Systems (GIS) and Water Resources X, Orlando, Florida, April 23-25, http://awra.org/meetings/Orlando2018/.

河流水力几何特征是水力学与水文学模型的重要输入项，此类模型可沿河道演进水流、确定水位与流量间的关系，并基于河段内的水流条件绘制洪水淹没潜力分布图。传统的河流几何特征量化方法依赖河道断面数据——例如HEC-RAS模型（HEC-RAS）的输入就要求此类数据。然而将这类基于断面的模型拓展至大尺度应用时，往往会面临复杂的实现难题。本次报告将介绍一种替代方案：最近邻排水高程（Height Above Nearest Drainage, HAND）。

我们所实现的HAND方法，依托TauDEM软件（TauDEM）中的Dinifinity算法从数字高程模型（Digital Elevation Model, DEM）中提取多流向水流路径，以此计算每个网格单元至其沿水流路径最近河道的相对高程。结合河道内的水流水深，即可确定洪水淹没的潜力与淹没水深。此外，通过将河道划分为河段或分段，可以分离出汇入各河段的汇水区，并针对该汇水区的HAND值网格应用一系列阈值水深，进而计算淹没体积、淹没表面积与河床湿润面积。将这些参数除以河段长度，即可得到河段平均断面面积、宽度与湿周。结合从DEM中提取的河道坡度与曼宁糙率系数（Manning's n），即可满足曼宁方程均匀流假设下水位-流量关系曲线构建，以及基于流量的洪水淹没范围绘制所需的全部输入参数。

本次报告将介绍该方法在美国本土的规模化应用案例：该案例结合美国国家海洋和大气管理局（National Oceanic and Atmospheric Administration, NOAA）国家水模型，针对国家水文数据集升级版（National Hydrography Plus, NHDPlus）中总计270万条河段的水位与洪水淹没潜力进行预测，其中绝大多数河段均无实测水文数据。本次美国本土尺度的应用，依托伊利诺伊大学国家超级计算应用中心（National Center for Supercomputing Applications, NCSA）与赛博地理信息科学中心（CyberGIS Center）的高性能并行计算得以实现。

本报告发表于2018年美国水资源协会（American Water Resources Association, AWRA）春季专业会议：地理信息系统（GIS）与水资源X，佛罗里达州奥兰多，2018年4月23-25日，http://awra.org/meetings/Orlando2018/。