Bozeman Creek LOMR Allied Engineering Services, Inc. (AESI) - PowerPoint PPT Presentation

Bozeman Creek LOMR Allied Engineering Services, Inc. (AESI) Jennifer Johnson, PE, CFM Paul Sanford, PE, CFM

Project Background & Objectives Modeling Approach 2D Modeling Presentation Outline 1D Modeling Results Questions and Answers

Project  Large-scale floodplain mapping was last completed for Background the Bozeman area in 1988 with a digital conversion of the data occurring in 2011.  Bozeman Creek & Tributaries Floodplain Study, Phase 2  Completed by FEMA, DNRC, Contractors  Draft hydraulic model and floodplain mapping – 2015  Completed for Incorporation into the Gallatin County and Incorporated Areas DFIRM and FIS  Supersedes 2011 FIS

Project Background Image Credit: DNRC

Project Background • Draft floodplain mapping for Bozeman Creek was completed around the time HEC-RAS 2D was emerging as a modeling tool • The City of Bozeman was interested in exploring 2D modeling in the downtown area Image Credit: Google Earth

 1D Modeling Challenges  Urban environment with complex flow splits  Pervasive shallow flooding  Long hydraulic structures with limited conveyance capacity  Olive-Rouse: ~230 feet; 12 ft x 5 ft opening  Babcock Street: ~185 feet; 16 ft x 5.3 ft opening  Main Street: ~320 feet; 12.6 ft x 6 ft opening  Overtopping of streets at Q < 10% Annual Chance Even t Project Background

 Update the floodplain mapping in the downtown area using HEC-RAS 2D Project  It is important to be able to use the best modeling tools available! Objectives  HEC-RAS 2D was a relatively new tool  Limited guidance available for 2D studies

Challenges associated with using a 2D model  Hydraulic structures  Floodways in 2D environments  Considered 2D floodway configurations including looking at (1) individual cell surcharges and (2) average WSEs at preselected cross sections Modeling Approach

 After discussions with the City of Bozeman, FEMA, and DNRC, we settled on a 1D model informed by a 2D model Modeling  Benefits of this approach: Approach  Better conceptualized 1D model  More confidence in split flows  We can still model a floodway per FEMA regulations

2D Modeling  General Setup  The 2D area encompassed downtown Bozeman and the area to the northeast  Rating curves and normal flow boundary conditions were used for outflows  1% Q = 1,007 cfs at upstream end of 2D model

General Setup  HEC-RAS 5.0.3  Full Momentum  Cell size 20 x 20 ft generally, 5 x 5 or 10 x 10 across main conveyance paths  Eddy viscosity coefficient of 1.5  Modeled buildings with a high Manning’s n 2D Modeling  Added buildings as breaklines to accurately model Manning’s n values  Simulations required a small time step

2D Modeling  How to model the hydraulic structures? • Modeled as bridges in 1D draft hydraulic model to account for geometry changes along each culvert and an open bottom • HEC-RAS 5.0.3 – no bridge modeling in 2D • Wormhole Culverts Introduced by a user • named Con Katsoulas on the RAS Solution forum Wormhole culverts • became obsolete with HEC-RAS Version 5.0.4

Calibration  We did not have gage data or other measurements to calibrate to  Structure modeling is not as advanced in the HEC-RAS 2D environment as in 1D  To justify the use of wormhole culverts, we compared their performance to the bridges in the draft hydraulic model 2D Modeling

 Complex split flows complicated the comparison between 1D and 2D structures  Used Hydraulic Property Table (HPT) values to compare 1D structures to 2D structures 2D Modeling

 Used HW, TW, and flow from 2D model culverts to pull bridge HW from 1D model HPT  Very similar results for low flow (impending road overtopping) comparison since no complex flow splits  Complex flow splits affect HW/TW relationship for the 1% event. More variance in comparison but still reasonable agreement. 2D Modeling

2D Modeling

2D Modeling  Results • The model showed Rouse as a significant conveyance path (~190 cfs to 350 cfs) • ~70 cfs of flow was leaving the Bozeman Creek/Rouse system via Main Street • More widespread shallow flooding

1D Modeling  General Model Setup • Modeled from upstream of Olive to the East Gallatin Confluence • Used the 2D model to establish reaches and flow exit/entrance points

1D Modeling  General Model Setup • Added a reach down Rouse and a reach at Main Street • Tied flow into the I90 split reach

1D Modeling  Split Flow Determinations • Used profile lines in 2D • Challenging accounting given the complexity of flow paths Rouse Split Church Split

 Flow Split comparison between Draft and LOMR 2015 Draft LOMR Model Model Location Flooding Source Q (cfs) Q (cfs) Bogert Park Bozeman Creek 1007 1007 US of East Olive Street Bozeman Creek 839 768 US of Main Street Bozeman Creek 839 540 US of E Mendenhall Street Bozeman Creek 839 540 US of E Lamme Street Bozeman Creek 853 777 Results US end of Church Ave Split Bozeman Creek 169 239 US end of Wallace Ave Split Bozeman Creek 35 113 Rouse Ave Downtown Split Bozeman Creek No Split 355 to 189 Main Street Split Bozeman Creek No Split 71  WSE comparison between Draft and LOMR  WSEs generally dropped ~0.5 feet along Bozeman Creek in downtown area  Floodway comparison between Draft and LOMR  Reduced FW along Bozeman Creek  Added ~440 feet of FW along Rouse Avenue  Added administrative FW along ~780 feet of Main Street

Results

Results

Questions and Answers