Docstoc

Floodplain Modeling and Delineation Using HEC-GeoRAS for Support of HEC-RAS

Document Sample
Floodplain Modeling and Delineation Using HEC-GeoRAS for Support of HEC-RAS Powered By Docstoc
					Floodplain Modeling and Delineation Using HECGeoRAS for Support of HEC-RAS
A Proof of Concept Analysis

FOR 595N Dr. T.P. Colson Surface Water Investigations with Geographic Information Systems

Prepared by David R. Markwood December 14, 2008

PDF created with pdfFactory Pro trial version www.pdffactory.com

TABLE OF CONTENTS

INTRODUCTION .......................................................................................................................... 3 DATA SOURCES .......................................................................................................................... 3 METHODS & ANALYSIS............................................................................................................. 3 HEC-GEORAS GEOMETRY ............................................................................................. 5 HEC-RAS ANALYSIS ......................................................................................................10 RESULTS & DISCUSSION..........................................................................................................10

2

PDF created with pdfFactory Pro trial version www.pdffactory.com

INTRODUCTION Though the fundamentals of hydraulic analysis have persisted in the U.S. Army Corps of Engineers (USACE) Hydrologic Engineering Center’s one-dimensional flood modeling software, users have enjoyed an evolution in the capabilities of these engineering tools. Specifically, modeling input can be developed, and simulation results can be analyzed and further processed, in a GIS environment. HEC-GeoRAS works as an interface between HECRAS and an ArcGIS environment, both pre- and post-model simulation, supporting the iterative nature of hydraulic modeling. HEC-GeoRAS is an ArcGIS extension for use in supporting hydraulic model development and analysis using HEC-RAS. Application of many of the tools included in the GIS extension are demonstrated in the development of a HEC-RAS steady-flow flood model of a 1.7-mile stream reach. The floodplain delineation and analysis functionalities of the extension are also demonstrated, comparing the generated approximate model floodplain, mapped on a LiDARbased terrain model, with the current FEMA effective flood zone delineation for the reach. The project site is located in Catawba County, North Carolina, in the headwaters of the Catawba River Basin, in the Upper Catawba Sub-basin. A 1.7 mile reach of Lyle Creek, approximately 3.4 miles upstream of its confluence with the Catawba River, was studied in this proof of concept demonstration. The engineering analysis for the effective FEMA flood study for Lyle Creek was performed in 1974, by the NRCS. The water-surface elevations generated from this study were most recently delineated on a terrain model based on the North Carolina Floodplain Mapping Program’s LiDAR data collection.

DATA SOURCES Bare earth LiDAR data was used to develop the Digital Terrain Model (DTM), downloaded from the North Carolina Floodplain Mapping Information System (http://www.ncsparta.net/fmis/Download_LIDAR.aspx). Base data used in generating the hydraulic model included 2005 aerial photography and effective flood hazard data, obtained from NCOneMap (ftp://204.211.239.203/outgoing/raster/local_imagery/catawba2005/sid/ , http://www.nconemap.com/Default.aspx?tabid=286).

METHODS & ANALYSIS Building a HEC-RAS model using HEC-GeoRAS begins with a digital terrain model (DTM), from which elevation data is extracted for features such as cross-sections and streamlines. Optional data layers can be generated and included in the model build. The spatial data is then extracted to a format recognized by HEC-RAS, where a geometry model of the stream reach can be created, and remaining input parameters such as discharge and boundary conditions can be specified for a flood profile simulation. Upon successful run of the steady-flow analysis, the water-surface elevations are exported into HEC-GeoRAS for floodplain mapping and other

3

PDF created with pdfFactory Pro trial version www.pdffactory.com

1369000.000000

1372000.000000

Legend
Bridges/Culverts Stream Centerline Flowpaths XS Cut Lines Interstate-40 IneffAreas

XS 4

XS 3 XS 2 XS 1

NAD 1983 StatePlane North Carolina FIPS 3900 (Feet)
Base Map Stream Location River Basin Data Date LiDAR TIN Lyle Creek Catawba, NC Catawba
0 0 250 0.05 500 0.1 1,000 Feet 0.2 Miles

HEC-GeoRAS Layers

NCFMP: bare earth LiDAR, baseroad layer Scale: 1:9,665 Dec 14, 2008 .000000

1 inch equals 805.45 feet

1369000

1372000.000000

PDF created with pdfFactory Pro trial version www.pdffactory.com

representations of model results. In this section, the steps used in building the flood model for Lyle Creek using HEC-GeoRAS are explained, along with discussion of engineering concerns. HEC-GEORAS GEOMETRY A digital terrain model (DTM), either grid (raster) or TIN (triangular irregular network) format, is the only input required for HEC-GeoRAS. The DTM for this analysis was built from bare earth LiDAR data. Contours were generated for the study area using the ArcGIS 3D Analyst extension prior to loading HEC-GeoRAS. These were used later to assist in cross-section placement and model development. The HEC-GeoRAS toolbar consists of four main parts: RAS Geometry for generating flood model input, RAS Mapping for processing model output, an .xml converter utility for converting HEC-RAS output, and ApUtilties for managing data exchange and working locations. First, a new data frame must be created to establish a workspace. In a blank ArcMap document with the HEC-GeoRAS extension loaded, a new map is added using the Add New Map function under the ApUtilities pull-down menu. This tool also creates a geodatabase that houses layers generated in the specified Map data frame. A new Map named “LyleCreek” was created for the analysis. The bare earth TIN was then specified on the Required Surface tab, accessed from RAS Geometry | Layer Setup. Once working parameters are established, empty spatial layers can be created. Using the RAS Geometry | Create RAS Layers toolset, the following layers were created: Stream Centerline, Bank Lines, Flow Path Centerlines, XS Cut Lines, Bridges/Culverts, Ineffective Flow Areas, and Landuse Areas. It should be noted that the streamline and cross-section data are the only layers that must be generated in order to export geometry data to HEC-RAS. All data layers can also be created in batch. Some of the optional layers that were not generated as part of this analysis include Blocked Obstructions, Inline and Lateral Structures, Storage Areas, and Levee Alignments. Layers are edited using the Editor Toolbar and Sketch Tool, and it should be noted the streamline should be drawn from upstream to downstream. The model streamline and cross-sections were digitized using contours and aerial photography. Bank Lines were developed similarly to the streamline, such that left and right channel bank lines were created. The actual channel bank locations used for each cross-section in the HEC-RAS model were revised from the stations generated from the Bank Line HEC-GeoRAS takeoffs, based on a review of the LiDAR-based geometry at each section with the HEC-RAS geometry editor. Flow path centerlines were digitized similarly to the stream centerline for the left and right overbanks, representing the center of mass of conveyance on the overbanks, as well as the centerline. Cross-section geometry can be viewed in HEC-GeoRAS using the Plot Cross Section tool on the HEC-GeoRAS toolbar, and selecting a XS Cut Line. Cross-sections can also be generated using the Construct XS Cut Lines tool. However, it is the opinion of the author that crosssection placement is the most important and overlooked user-specified model input in developing a flood model. Hydraulic parameters set for each cross-section in the model, such as channel bank locations, surface roughness coefficients, contraction/expansion coefficients, and geometry, 5

PDF created with pdfFactory Pro trial version www.pdffactory.com

1369000.000000

1372000.000000

Legend
Bridges/Culverts XS Cut Lines Stream Centerline IneffAreas TIN Contours

XS 4

XS 3 XS 2 XS 1

TIN Contours

NAD 1983 StatePlane North Carolina FIPS 3900 (Feet)
Base Map Stream Location River Basin Data Date LiDAR TIN Contours Lyle Creek Catawba, NC Catawba NCFMP: bare earth LiDAR Dec 14, 2008
0 0 125 0.025 250 0.05 500 Feet 0.1 Miles

XS Cut Line Placement
1369000
.000000

Scale: 1:6,000

1 inch equals 500 feet

1372000.000000

PDF created with pdfFactory Pro trial version www.pdffactory.com

are significant in the computations, though cross-sections that capture constrictions in the floodplain which are appropriately spaced for RAS simulations, particularly at structures, are not likely achieved by automated procedures. A total of 18 cross-sections were placed such that constrictions in the floodplain were captured in the model geometry, with average spacing of approximately 500 feet. The cross-sections were drawn from left overbank to right overbank, facing downstream, and perpendicular to conveyance. Cross-sections were digitized with greater density in areas of rapid change in stream invert elevation and narrow floodplain width. Cross-section stationing along the stream reach was defined by the intersection of the XS Cut Lines with the Stream Centerline. This takeoff was performed by selecting the RAS Geometry | XS Cut Lines Attributes | Stationing takeoff tool. Remaining attributes are defined similarly, using this toolbar menu. The downstream study limit cross-section was located about 750’ upstream of Bunker Hill Bridge, while the upstream extent about 350’ downstream of Bunker Hill School Road. Interstate-40 is the only structure in the model, and the bridge was approximated based on aerial photography, the effective FEMA flood profile, and engineering judgment. The structure actually consists of two bridges, though based on review of aerial imagery, these bridges are approximately 30’ apart, and were modeled as a single bridge in HEC-RAS. The downstream and upstream bounding cross-sections for this structure, XS 2 and XS 3 for modeling bridgeflow in HEC-RAS, respectively, were placed as close to the structure as possible while not capturing the structure fill in the cross-section geometries. The most upstream bridge-flow crosssection, XS 4 for modeling flow through a bridge in HEC-RAS, was stationed at the point where fully-expanded flow begins to contract due to the obstruction. XS 1 for the bridge-flow simulation was placed where flow is fully expanded after experiencing the obstruction. Generally, XS 1 and XS 4 should not have ineffective conveyance areas defined, as these sections should represent fully-expanded flow. However, due to the meandering of the streamline downstream of Interstate-40, an ineffective conveyance area was specified on the left overbank of XS 1. Although HEC-GeoRAS does not support bridge deck data, the structure TopWidth, parallel to flow, and USDistance, distance from upstream face of bridge/culvert to cross-section representing upstream bounding face of structure (XS 3), must be specified in the Bridge/Culvert layer. These were set to 100’ and 45’, respectively. This layer was named “I-40”, and digitized similarly to the XS Cut Lines, such that the obstruction of flow due to fill is captured in the geometry. Ineffective conveyance polygons were digitized, containing portions of cross-sections within the floodplain that were not considered as contributing to effective conveyance. Using the RAS Geometry | Ineffective Flow Areas | Positions HEC-GeoRAS takeoff function, ineffective conveyance locations at each cross-section are defined based on the intersection of the XS Cut Lines with the IneffAreas polygons. The elevations of these ineffective stations can be revised within the HEC-RAS geometry editor, if necessary. For this analysis, all ineffective stations were set above the simulated water-surface elevation for the cross-sections with ineffective conveyance areas defined.

7

PDF created with pdfFactory Pro trial version www.pdffactory.com

1369000.000000

1372000.000000

Legend
Bridges/Culverts XS Cut Lines Stream Centerline IneffAreas

XS 4

XS 3 XS 2 XS 1

Aerial Photography

NAD 1983 StatePlane North Carolina FIPS 3900 (Feet)
Base Map Stream Location River Basin Data Date Aerial Photography Lyle Creek Catawba, NC Catawba NCOneMap Aerial Photography Dec 14, 2008
0 0 125 0.025 250 0.05 500 Feet 0.1 Miles

XS Cut Line Placement
1369000
.000000

Scale: 1:6,000

1 inch equals 500 feet

1372000.000000

PDF created with pdfFactory Pro trial version www.pdffactory.com

When dealing with floodplains having multiple or divided flow paths with variable depths, defining several areas of ineffective conveyance along many cross-sections using HEC-GeoRAS can be significantly more efficient than entering this data in the HEC-RAS Geometry Editor. This is also true for defining surface roughness for land areas. In order to extract n-value data to include with the GIS export, a table named “LUManning” was created and imported into the map geodatabase. This table contains “LUCode” and “N_Value” fields, with n values specified for all land use codes. Manning’s n-values were determined using aerial photography by creating a polygon containing the cross-section extents, and cutting the polygon feature into areas of common land cover, assigned with suitable, if not conservative, roughness coefficients. The n-values used ranged from 0.15 for high density forest, to 0.08 for open space, to 0.05 for the channel. The figure below displays polygons of common land cover intersecting model cross-sections, near the upstream limit of study.

9

PDF created with pdfFactory Pro trial version www.pdffactory.com

LandUse Polygons

Upstream Reach of Study

NAD 1983 StatePlane North Carolina FIPS 3900 (Feet)
Base Map Stream Location River Basin Data Date Aerial Photography Lyle Creek Catawba, NC Catawba NCOneMap Aerial Photography Dec 14, 2008
0 0 62.5 0.0125 125 0.025 250 Feet 0.05 Miles

Roughness Coefficient Polygons

Scale: 1:3,000

1 inch equals 250 feet

PDF created with pdfFactory Pro trial version www.pdffactory.com

Once all of the HEC-GeoRAS layers were digitized, the RAS Geometry menu functions were used to perform takeoffs for each layer, which compiled the model input data to be exported. Finally, the geometry data was exported using the RAS Geometry | Extract GIS Data function, which generated an .sdf file recognize by HEC-RAS. HEC-RAS ANALYSIS The .sdf HEC-GeoRAS export file was imported into RAS via the Geometry Data Editor, under File | Import GIS Data. A new project was created, and the geometry data imported. The main modifications to the model included revising channel banks as represented by cross-section geometry and aerial photography, and inputting a bridge deck at Interstate-40. In order to conservatively approximate the structure, the bridge opening was reduced from 120’ to 80’, and the contraction/expansion coefficients for the bounding cross-sections set to 0.4/0.6, respectively. The downstream boundary condition was set using the known effective water-surface elevation from the FEMA profile at the downstream study limit. Two flow profiles were specified in the Flow Data Editor, one using the effective 1% Annual-Chance event discharge value of 15,534 cfs, and another using the North Carolina Blue-Ridge Piedmont Urban Regression estimate of 10,028cfs for the same event, applying a conservative impervious percent value of 20. The simulated water-surface elevations and spatial data were extracted using the File | Export GIS Data menu function in the main HEC-RAS project window. The file generated when employing this function is an .sdf file, which must be converted to an .xml recognized by HECGeoRAS. Located next to the ApUtilities menu on the HEC-GeoRAS toolbar, the .xml converter tool was employed to convert the HEC-RAS .sdf export file, essentially an ASCII text file, to an .xml. It should be noted that the HEC-RAS geometry file can be manipulated within a text editor, as this can also be a useful way to make modifications for unique circumstances. The RAS Mapping | Layer Setup function was used to import the flood model results into HEC-GeoRAS. A new analysis window was specified, along with the RAS export .xml, the DTM type and source, and the output directory. Selecting the RAS Mapping | Read RAS GIS Export File tool will prompt HEC-GeoRAS to import HEC-RAS model data. Using the RAS Mapping | Inundation Mapping | Water Surface Generation and RAS Mapping | Inundation Mapping | Floodplain Delineation | Grid Intersection tools, sequentially, a watersurface grid and floodplain boundary was delineated for the flood model profiles.

RESULTS & DISCUSSION The effective FEMA flood profile of the study reach of Lyle Creek was generated by the NRCS in 1974, using their WSP-2 flood modeling software. Because the engineering analysis was performed over three-decades ago, it seems this stream may be a good candidate for a new study. Though the approximate flood model and floodplain delineation generated in this report are quite similar to that of the effective model. In fact, using the same discharge as the effective 1% Annual-Chance profile, the computed water surface elevation at the upstream face of Interstate40 is within a foot of the effective profile. 11

PDF created with pdfFactory Pro trial version www.pdffactory.com

1369000.000000

1372000.000000

Legend
Approximted Floodplains Effective Discharge USGS Regression Discharge Effective Floodplain XS Cut Lines Stream Centerline

1% Annual-Chance Floodplains

NAD 1983 StatePlane North Carolina FIPS 3900 (Feet)
Aerial Photography Lyle Creek Catawba, NC Catawba NCOneMap Aerial Photography, NCFMP LiDAR Scale: 1:6,000 Dec 14, 2008
0 0 125 0.025 250 0.05 500 Feet 0.1 Miles

Results
1369000
.000000

Base Map Stream Location River Basin Data Date

1 inch equals 500 feet

1372000.000000

PDF created with pdfFactory Pro trial version www.pdffactory.com

The RAS Mapping tool generates both a raster and polygon of the flooding extents, by intersecting the water-surface elevations at each cross-sections with the DTM. This capability to map water-surface profile results in a GIS environment, in concert with project base data used to develop the input, improves the engineering judgment involved in determining input parameters. Geographical representations of floodplain depths, velocities, and extents provide great insight into the model response, and ideally the behavior of the natural system under analysis. As previously stated, working with results, as well as spatial input data, in an interfaced GIS environment, is a tremendous benefit in support of the iterative nature of hydraulic modeling. Although there are several factors contributing to and diminishing the accuracy of the floodplain delineations generated from the approximate hydraulic model in this demonstration, or any floodplain model and delineation, the terrain source has the most impact. The relative accuracy achieved in the approximate flood model and floodplain delineations compared to the FEMA effective can be largely attributed to the quality of the DTM used in the model development and floodplain delineation. The reader is encouraged to visit the HEC website (http://www.hec.usace.army.mil/software/hecras/hec-georas.html) as well as view the HEC-GeoRAS User’s Manual for more information.

13

PDF created with pdfFactory Pro trial version www.pdffactory.com


				
DOCUMENT INFO
Shared By:
Stats:
views:3733
posted:12/15/2008
language:English
pages:13
Description: GIS-based Flood Modeling