National Academies Press: OpenBook

Resilient Design with Distributed Rainfall-Runoff Modeling (2023)

Chapter: Chapter 4 - Case Examples

« Previous: Chapter 3 - Survey of State Practices for Distributed Rainfall-Runoff Modeling
Page 48
Suggested Citation:"Chapter 4 - Case Examples." National Academies of Sciences, Engineering, and Medicine. 2023. Resilient Design with Distributed Rainfall-Runoff Modeling. Washington, DC: The National Academies Press. doi: 10.17226/27051.
×
Page 48
Page 49
Suggested Citation:"Chapter 4 - Case Examples." National Academies of Sciences, Engineering, and Medicine. 2023. Resilient Design with Distributed Rainfall-Runoff Modeling. Washington, DC: The National Academies Press. doi: 10.17226/27051.
×
Page 49
Page 50
Suggested Citation:"Chapter 4 - Case Examples." National Academies of Sciences, Engineering, and Medicine. 2023. Resilient Design with Distributed Rainfall-Runoff Modeling. Washington, DC: The National Academies Press. doi: 10.17226/27051.
×
Page 50
Page 51
Suggested Citation:"Chapter 4 - Case Examples." National Academies of Sciences, Engineering, and Medicine. 2023. Resilient Design with Distributed Rainfall-Runoff Modeling. Washington, DC: The National Academies Press. doi: 10.17226/27051.
×
Page 51
Page 52
Suggested Citation:"Chapter 4 - Case Examples." National Academies of Sciences, Engineering, and Medicine. 2023. Resilient Design with Distributed Rainfall-Runoff Modeling. Washington, DC: The National Academies Press. doi: 10.17226/27051.
×
Page 52
Page 53
Suggested Citation:"Chapter 4 - Case Examples." National Academies of Sciences, Engineering, and Medicine. 2023. Resilient Design with Distributed Rainfall-Runoff Modeling. Washington, DC: The National Academies Press. doi: 10.17226/27051.
×
Page 53
Page 54
Suggested Citation:"Chapter 4 - Case Examples." National Academies of Sciences, Engineering, and Medicine. 2023. Resilient Design with Distributed Rainfall-Runoff Modeling. Washington, DC: The National Academies Press. doi: 10.17226/27051.
×
Page 54
Page 55
Suggested Citation:"Chapter 4 - Case Examples." National Academies of Sciences, Engineering, and Medicine. 2023. Resilient Design with Distributed Rainfall-Runoff Modeling. Washington, DC: The National Academies Press. doi: 10.17226/27051.
×
Page 55
Page 56
Suggested Citation:"Chapter 4 - Case Examples." National Academies of Sciences, Engineering, and Medicine. 2023. Resilient Design with Distributed Rainfall-Runoff Modeling. Washington, DC: The National Academies Press. doi: 10.17226/27051.
×
Page 56
Page 57
Suggested Citation:"Chapter 4 - Case Examples." National Academies of Sciences, Engineering, and Medicine. 2023. Resilient Design with Distributed Rainfall-Runoff Modeling. Washington, DC: The National Academies Press. doi: 10.17226/27051.
×
Page 57
Page 58
Suggested Citation:"Chapter 4 - Case Examples." National Academies of Sciences, Engineering, and Medicine. 2023. Resilient Design with Distributed Rainfall-Runoff Modeling. Washington, DC: The National Academies Press. doi: 10.17226/27051.
×
Page 58
Page 59
Suggested Citation:"Chapter 4 - Case Examples." National Academies of Sciences, Engineering, and Medicine. 2023. Resilient Design with Distributed Rainfall-Runoff Modeling. Washington, DC: The National Academies Press. doi: 10.17226/27051.
×
Page 59
Page 60
Suggested Citation:"Chapter 4 - Case Examples." National Academies of Sciences, Engineering, and Medicine. 2023. Resilient Design with Distributed Rainfall-Runoff Modeling. Washington, DC: The National Academies Press. doi: 10.17226/27051.
×
Page 60
Page 61
Suggested Citation:"Chapter 4 - Case Examples." National Academies of Sciences, Engineering, and Medicine. 2023. Resilient Design with Distributed Rainfall-Runoff Modeling. Washington, DC: The National Academies Press. doi: 10.17226/27051.
×
Page 61
Page 62
Suggested Citation:"Chapter 4 - Case Examples." National Academies of Sciences, Engineering, and Medicine. 2023. Resilient Design with Distributed Rainfall-Runoff Modeling. Washington, DC: The National Academies Press. doi: 10.17226/27051.
×
Page 62
Page 63
Suggested Citation:"Chapter 4 - Case Examples." National Academies of Sciences, Engineering, and Medicine. 2023. Resilient Design with Distributed Rainfall-Runoff Modeling. Washington, DC: The National Academies Press. doi: 10.17226/27051.
×
Page 63
Page 64
Suggested Citation:"Chapter 4 - Case Examples." National Academies of Sciences, Engineering, and Medicine. 2023. Resilient Design with Distributed Rainfall-Runoff Modeling. Washington, DC: The National Academies Press. doi: 10.17226/27051.
×
Page 64
Page 65
Suggested Citation:"Chapter 4 - Case Examples." National Academies of Sciences, Engineering, and Medicine. 2023. Resilient Design with Distributed Rainfall-Runoff Modeling. Washington, DC: The National Academies Press. doi: 10.17226/27051.
×
Page 65

Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

48 This chapter compiles the information that was obtained through interviews with personnel from three state DOTs that are involved in designs or studies that used hydrologic modeling, supporting the synthesis objective of documenting state DOT use of DRRMs. In particular, these agencies were selected from the list of 16 state DOTs that have reported using DRRMs in their hydrological practices. The criteria used to select states for this interview included the hydrologic characteristics observed in the selected state and the scope and history of DRRM application in transportation projects. North Carolina Department of Transportation North Carolina is a state located in the U.S. South Atlantic coast and has natural waters in very diverse landscapes with distinct hydrological characteristics. These range from tidal-influenced rivers discharging into the Atlantic Ocean to smaller streams located near the Appalachian Mountains. Due to its location, the state is affected by hurricanes and other coastal storms that result in extreme rainfall events, which in turn create severe stressors to roadways. Examples of these extreme events include Hurricane Matthew (2016), which created high- intensity rainfall, and Hurricane Florence (2018), which was slower moving but resulted in heavy rainfall (AECOM 2020), as shown in Figure 32. Data from the National Weather Service (NWS 2021) indicate that Matthew rainfall depths reached 18.95 inches and created flooding damages amounting to $1.5 billion in North Carolina alone. Hurricane Florence resulted in a larger rainfall depth up to 35.93 inches, resulting in 42 fatalities in the state and an estimated total of $16.7 billion in damage (NWS 2019). These events helped to catalyze changes to improve the state response to flooding (Park et al. 2021), which includes improved hydrological modeling practices. The North Carolina Department of Transportation (NCDOT), partnering with the North Carolina Department of Public Safety—Emergency Management (NCEM), has been employing DRRMs in the context of highway design. This section details one of these efforts, which is linked to the study of extreme flooding conditions in the River Lumber at the I-95 cross. This inves- tigation aimed to inform design and make recommendations with regard to flood impact and roadway resiliency to water stressors based on analyses of these events and flooding behavior of the River Lumber along the I-95 corridor, as shown in Figure 33. These recommendations would be reflected in terms of the recommended I-95 elevations to ensure the connectivity of the corridor during extreme hydrological events. This was the context in which two DRRM tools, specifically HEC-HMS and HEC-RAS 2D, were applied. One key characteristic of the hydrological study, in which the HEC-HMS model was applied, is the extreme rainfall events that were modeled. These include synthetic storms with a recurrence Case Examples C H A P T E R   4

Case Examples 49 Figure 32. Paths of Hurricanes Matthew (left) and Florence (right) (AECOM 2020). interval of 100 years (1% annual exceedance) event at 5th Street, Lumber, NC. An FFA using the data from the USGS station 02134170 River Lumber indicated that the peak flow discharge for the 100-year event was close to the observed peak flows from hurricanes Matthew and Florence (AECOM 2020). Modeling was also performed using the observed rainfall for both hurricanes, along with streamflows and flood elevation observations. The hydrologic–hydraulic study used an HEC-RAS rain-on-grid strategy to achieve this study’s goals. In terms of hydrological modeling, the study consisted of separate event-based simulations of the synthetic rainfall with TR = 100 years and the two hurricanes. The data sources for the study included the following: • Watershed geometry data: DA delineations used for computing USGS regression peak stream- flow estimates were initially developed using a hydro-corrected 50 ft × 50 ft grid size DEM generated from lidar data collected and processed for NCEM. For the modeling tasks, a more detailed 5-foot terrain DEM based on quality level 2 (QL2) lidar was acquired from NCEM, and the DEM was subdivided into four computational meshes shown in Figure 34: Drowning Creek, Upper Lumber, Maxton Lumber, and Lumberton. Pertinent structures in the watershed were also surveyed. • Rainfall data: The NWS and MesoWest observation stations provided rainfall data for the studies. The Design Event that was used in the hydrologic modeling was based on the Atlas 14 2nd Quartile 50% temporal distribution, reduced so that the water surface elevations match the observations from both hurricane events. Regarding the observed rainfalls linked with the hurricane events, Hurricane Matthew had a smaller average rainfall depth across River Lumber watershed (estimated between 5 and 8 inches) compared to Hurricane Florence (15 inches). Thiessen polygons were derived with five selected weather stations to assign rainfall data into the hydrologic catchments. Further details on the analysis regarding the rainfall data that were used in the study are skipped here for brevity. • Rainfall losses: The hydraulic model used in the analysis, HEC-RAS 2D, has the ability to consider rain on grid as a means to consider the effects of rainfall. However, the version applied in the River Lumber modeling (5.0.7) was not able to consider abstractions created by infiltration. An HEC-HMS hydrological model was used to compute the rainfall losses and

50 Resilient Design with Distributed Rainfall-Runoff Modeling derive effective rainfall hyetographs that could be used in the HEC-RAS 2D. Each of the four sub-watersheds presented in Figure 33 was supplied with a single hyetograph in HMS, and these were readjusted for input in HEC-RAS 2D. The computation of rainfall losses applied the CN approach (Cronshey et al. 1986) using land use data from the National Land Cover Database (Dewitz 2019). Initial CN values were obtained by combining land use data and soil characteristics following the pending update to “Part 630 Hydrology” of the NRCS National Engineering Handbook (NRCS 2017). These values of CN, however, were adjusted significantly during the hydrologic calibration process. Due to previous wet conditions associated with Hurricane Matthew, the modeled rainfall abstraction was smaller than the corresponding abstraction calibrated for Hurricane Florence. Greater, Figure 33. River Lumber watershed and the location of the I-95 corridor. The river watershed is divided into four major sub-watersheds (Drowning Creek, Upper Lumber, Maxton, and Lumberton), draining from the northwest to southeast toward Lumberton, NC (AECOM 2020).

Case Examples 51 more conservative CN values from Hurricane Matthew calibration were used for the Design Event using the 100-year hyetograph. • Hydraulic modeling: HEC-RAS 2D models were supplied with adjusted hyetographs that resulted from HEC-HMS. The inflows from upstream sub-watersheds were connected to the downstream sub-watersheds, and to these inflows the effective hyetographs were added to the grid via the rainfall boundary condition. A parameter that was used in the calibration of the hydraulic modeling was the values of Manning roughness, which were selected to improve the timing of the observed and modeled peak flows. Whereas larger computational grid cells (500 ft × 500 ft) were used in most of the model domain, these were much smaller (e.g., 30 ft × 30 ft) along the I-95 corridor, as shown in Figure 35. In addition to the 2-D model, a 1-D HEC-RAS model considering the bridges was constructed to inform and check the 2-D modeling process. Following the hydrologic–hydraulic modeling effort, a map presenting the extent of flooding was generated and is presented in Figure 36. The numbers presented in the figure correspond to the modeled water surface elevations minus the high water marks measured by the USGS. Figure 34. Elevation map of River Lumber watershed (AECOM 2020).

52 Resilient Design with Distributed Rainfall-Runoff Modeling Figure 35. HEC-RAS 2D grid at Lumberton, NC (AECOM 2020), with smaller grid elements near the I-95 corridor. Figure 36. Map of flooding extents created by River Lumber at the I-95 corridor. The numbers in the figure correspond to the differences between modeled water surface elevation and high water marks measured by USGS (AECOM 2020).

Case Examples 53 In general, the agreement between the model and observed water levels is very good, with dis- crepancies less than 0.4 feet. A similar map was derived for Hurricane Florence, also with good results and discrepancies between modeled and observed high water marks less than 0.7 feet (AECOM 2020). These results enabled recommendations regarding the new elevation for the I-95 corridor at Lumberton, NC, to ensure that the requirement of a minimum 1.5-foot free- board requirement is met and the road connectivity is preserved for extreme rainfall events of the magnitude of Hurricanes Matthew and Florence. As indicated, the application of DRRM tools, HEC-HMS and HEC-RAS 2D, enabled an accurate representation of the impacts of extreme hydrological events to the strategic I-95 cor- ridor in Lumberton, NC. The development of this study was performed by consultants, and NCDOT personnel involved in these studies had a close collaboration with NCEM. Due to the complex nature of the studies involving DRRMs and other time constraints on design staff working in NCDOT, future related studies are likely to be developed by consultants as well. However, projects such as the improvements in the I-95 corridor are costly and carry a poten- tial liability associated with changes in the road geometry. The added costs and efforts for the implementation of DRRMs are well justified according to the interviewees. In their evaluation, DRRM tools have improved significantly in the past 20 years. DRRMs can be used for creating critical hydrological conditions at key transportation corridors to determine the most cost- effective solutions to mitigate the impacts of water-related stressors. These can be through elevating roads or, in some cases, through hardening them against overtopping and verifying that roadway changes are not creating harmful changes elsewhere in the watershed. A final point raised in the interview is that the collaboration between NCDOT and NCEM has been very positive in the wider adoption of DRRMs for transportation projects. Also, the perception is that a more efficient dissemination of good examples of DRRM application can be achieved through better communication between state DOTs. Kentucky Transportation Cabinet Kentucky is located in the East South Central region of the United States, and it is a state with many important hydrological features. The state’s geography is characterized by the Appalachian Mountains on its eastern boundary and plains on its western boundary. The Ohio River borders the north, and the state has a portion of its western border limited by the Mississippi River. Within the commonwealth, the Kentucky Transportation Cabinet (KYTC) operates and main- tains more than 27,500 miles of roadways. With an increase in severe rainfall intensity and depths, there have been episodes of roadway flooding reported in the state. Although the vast majority of the hydrological studies performed by KYTC apply lumped approaches (such as the Rational Method or USGS regression equations), DRRMs have been used to evaluate roadway overtopping, as discussed in this section. The area of interest is in Stanton, KY, by Kentucky Route KY 11/15 (East College Avenue). According to Palmer Engineering (2022), flooding occurs on the south side of KY 11/15 generally between Ewen Street to the west and Derickson Lane to the east. Stormwater flooding of yards, parking lots, and overtopping of KY 11/15 have been reported during rain events. Overtopping of the roadway has been reported multiple times during 2021, and one instance is shown in Figure 37. The application of DRRMs was intended to help identify potential solutions to miti - gate stormwater flooding in these areas. In Stanton, KY, 11/15 is located parallel to and north of the Bert T. Combs Mountain Parkway (KY 9000) and drains much of the hilly forested area to the south of the parkway. Excess runoff drains northward down the hills and then under the Mountain Parkway through a series of culverts. Water then flows to KY 11/15 through a series of ditches that end at a roadside drainage system on the south side of KY 11/15, as indicated in Figure 38. These grass-lined ditches are complemented

54 Resilient Design with Distributed Rainfall-Runoff Modeling Figure 37. Overtopping and flooding of KY 11/15 (East College Avenue) on June 12, 2022 (Palmer Engineering 2022). Figure 38. Delineation of the drainage area and sub-catchments by Personal Computer Storm Water Management Model (PCSWMM) south of KY 11/15 in Stanton, KY (Palmer Engineering 2022).

Case Examples 55 by entrance pipes. For the development of the hydrological study, two watersheds, Study Area 1 and Study Area 2, were created to represent the west and east sections of the flooded roadway within the project corridor. The details of the stormwater conveyance are not presented here for brevity but can be found in Palmer Engineering (2022). In order to achieve the goals of the study, the selected DRRM was the EPA SWMM 5 and, more precisely, a commercial implementation known as Personal Computer Storm Water Management Model (PCSWMM) (https://www.pcswmm.com/). Building from the open-source SWMM 5.1.015, such commercial implementations add capabilities that include integration with GIS databases, model calibration tools, and the capability to simulate overland flows. The PCSWMM has the ability to map the location and extent of surface flooding, which was an important goal of this study. The PCSWMM integrated 1-D/2-D characteristics. In the model, various design storms were considered in the study: 25-year, 24-hour; 100-year, 24-hour; and the July 30, 2021 event. To compute rainfall losses, the Green-Ampt infiltration formula was used, with soil parameters sourced from the Soil Survey Geographic Database (SSURGO) (NRCS 2022). The 2-D mesh was generated in this study using a 5-foot DEM obtained from the Kentucky GIS geodatabase (KyFromAbove.ky.gov) and site surveys. For the selected rain events, the sub-catchments of the study area generated runoff, which was routed to junctions. As the water surface elevation exceeded a junction’s rim elevation, it flooded the adjacent 2-D meshes. The excess runoff was routed on the surface from 2-D cells to an adjacent cell with the steepest hydraulic gradient until it reached a 1-D junction. At that point, the flows were routed from junction to junction through the conveyance system comprising ditches, channels, pipes, and so forth. The schematic of this process is shown in Figure 39. Other relevant modeling parameters in the study included determining the impervious areas, which was also done using lidar, site surveys, and aerial imagery. The Manning roughness was assumed as 0.013 for concrete conveyance and 0.032 for grass-lined swales. The outfall in the model was fixed and far enough so that outfall conditions would not influence the hydrological calculations in the areas experiencing flooding. Figure 39. Schematic of 1-D/2-D flow pathways (arrows) used in the PCSWMM to represent the flooding of KY 11/15 (East College Avenue) (Palmer Engineering 2022).

56 Resilient Design with Distributed Rainfall-Runoff Modeling In order to assess the effectiveness of the PCSWMM, a total of 10 points of interest were identified—five in each of the two study areas—which included lots that experienced flooding. Modeling results for Study Areas 1 and 2 are shown in Figures 40 and 41, respectively. For Study Area 1, where the main direction of the flow is from east to west, the water crosses the roadway at a culvert (i.e., Culvert 1 in the study). Flooding is represented through a color gradient, with darker blue cells representing locations with larger depths. As shown in these two figures, the PCSWMM results indicate multiple locations of roadway overtopping, consistent with field observations. Figure 40. Model results of overtopping and flooding areas of KY 11/15 for Study Area 1 (Palmer Engineering 2022). Yellow lines show sub-catchment conveyances, and arrows show the runoff flow directions. Figure 41. Model results of overtopping of KY 11/15 for Study Area 2 (Palmer Engineering 2022). Yellow lines show sub-catchment conveyances, and arrows show the runoff flow directions.

Case Examples 57 The same modeling framework used to diagnose the issues of flooding was used to assess the impacts of various strategies to avoid these issues. The alternatives that were modeled in PCSWMM included increasing culvert conveyance, improving ditch grading, reducing Manning roughness of ditches, and using detention basins to retain flows and control discharges. Some of the proposed changes were very effective in controlling the roadway overtopping compared to the results in Figure 40, as illustrated in Figure 42 for Study Area 1. KYTC staff noted that DRRMs have been applied mostly when an unusual problem or condi- tion exists on sites. DRRMs were primarily developed and implemented by consulting com- panies. A wider application of these tools in the context of roadways would need to involve workforce training. Beyond the actual use of models, it is important that the engineers have a clear understanding of the fundamentals of hydrological processes so that a critical evaluation of the modeling results can be possible. These training tasks can highlight the benefits of applying DRRMs as a means of achieving more precise predictions and lowering costs in more complex hydrological design and analyses. California Department of Transportation The state of California is on the western coast of North America, with distinct hydrological characteristics in terms of its climate and dominant storm patterns. The historical average rainfall depth in California is 22.9 inches (OEHHA 2018), which is much less than the other states that are covered in this chapter. Moreover, a large diversity of hydrological conditions exists within the state, which brings additional challenges in terms of water resources modeling. The California Department of Transportation (Caltrans) owns and maintains the state’s highway system. The department is large and diverse, and the application of hydrological tools across Caltrans is also anticipated to be diverse. In the context of this synthesis, the Structure Hydraulics unit from the Office of Specialty Investigations at Structure Maintenance and Investigations was interviewed regarding their use of DRRMs in the context of bridge infrastructures. This unit Figure 42. Effects of countermeasures to prevent overtopping of KY 11/15 for Study Area 1 predicted by PCSWMM simulation (Palmer Engineering 2022).

58 Resilient Design with Distributed Rainfall-Runoff Modeling has been using hydrological models such as HEC-HMS and GSSHA for bridge scour estimates, providing peak flow estimates that can be used in hydraulic estimates of bridge scour. This section presents one case example in which DRRM results were used and compared with predictions from lumped and semi-distributed approaches. Specifically, the GSSHA model was used to study the discharges from the Big Sulphur Creek watershed, presented in Figure 43. The aim of performing hydrologic studies by this unit is to compute the 100-year return period discharge in support of scour evaluation. This study, performed by Caltrans staff, was initiated by providing GSSHA with the following data to perform computations: • Topography from the USGS National Elevation Data with the 30 m × 30 m spatial resolution. • Soil characteristics from the SSURGO database (NRCS 2022). • Land use data from the National Land Cover Database (Dewitz 2019). The modeling grid with soil type and land use data is presented in Figure 44. • Streamflows from the USGS “Big Sulphur C A G Resort Nr Cloverdale CA” gaging station site (USGS ID: 11463170) at a 15-minute frequency. • Recorded rainfall from the Department of Water Resources California Data Exchange Center (DWR CDEC) hourly data from the neighboring “Whispering Pines” recording station (CDEC ID: WSP). • Spatially distributed rainfall: The spatially distributed 3-hour precipitation data set was obtained from the Tropical Rainfall Measuring Mission (TRMM) Multi-Satellite Precipitation Analysis and processed in ArcGIS. Upon performing this step, the watershed was divided into four 0.25-degree × 0.25-degree grid cells, as presented in Figure 45. The combination of soil characteristics and land use data enabled the input of various hydro- logical parameters used in the GSSHA model setup. These included the HSGs, soil texture, hydraulic conductivity, initial moisture, field capacity, and wilting point. The infiltration calculation Figure 43. Location of the study area in California and Big Sulphur Creek watershed delineation (Vigneswaran 2022).

Case Examples 59 Figure 44. GSSHA grid of Big Sulphur Creek watershed overlaid on soil type data and land use (Vigneswaran 2022). Figure 45. Grid used for the spatially distributed Network Common Data Form (NetCDF) rainfall data for the GSSHA model (Vigneswaran 2022).

60 Resilient Design with Distributed Rainfall-Runoff Modeling in the model was computed using the Green-Ampt approach, and the channel routing used the diffusion wave (i.e., zero-inertia wave equation) approach. Given that the study was focused on the simulation of events rather than an extended period analysis, no evapotranspiration was considered. Models such as GSSHA typically undergo the calibration of the variables that are more impactful in the modeling results. In this study, the calibration process adjusted the values of hydraulic conductivity, initial moisture, and overland and channel roughness. These parameters were adjusted by a simple manual trial-and-error method so that the modeled hydrograph agrees with the single peak direct runoff rate. The flow hydrograph from the USGS stream gaging station 11463170 provided a peak flow for the rain event that occurred on January 3, 2017, close to 1,300 cubic feet per second (CFS) after deducting the stream’s base flow. The same event simu- lated with GSSHA corresponded to a peak flow of 1,310 CFS, indicating that the model calibration for the watershed was successful. In the work by Vigneswaran (2022), a comparison was made between the results of an FFA per- formed with the data from the USGS stream gaging station 11463170 and regional regression equation results. FFA peak flow results, based on the 41-year-long continuous record for the station and using a Log Pearson Type III distribution, yielded an estimate of 10,800 CFS peak flow rate for a 100-year return period rain event. The result derived from the regional regres- sion equation for the same return period was 4,300 CFS and far below the historically observed annual maximum flows (four records >7,000 CFS from 1981 to 2021). When considering the standard errors from the regression equation, the upper and lower peak flow prediction inter- vals were 2,160 and 8,580 CFS, respectively. One point that is relevant to this analysis is that the watershed is close to the boundary where the SCS design storms change from Type IA (Northern California) and Type I (Southern California). As shown in Figure 46, these two distributions are significantly different, with the latter yielding more intense rain events. When the calibrated GSSHA model was run using an SCS Type IA, 100-year return period synthetic rainfall, the resulting peak flow was 3,465 CFS, which is closer to the estimate from the regression equations. If instead, the same model was run using the SCS Type I, 100-year return period rainfall, the peak flow was 9,400 CFS, which is closer to FFA peak flow prediction. The results of the peak flow yielded by GSSHA for both of the rainfall distributions are presented in Figures 47 and 48. For the context of bridge scour computations, in which the flow velocity estimates are of paramount importance, reducing the uncertainty in the estimates of flows is very important. Figure 46. SCS rainfall distributions (24 hours), showing the differences between Type I and Type IA, and their geographical distribution (Cronshey et al. 1986).

Case Examples 61 Figure 47. Results of GSSHA modeling for the two 100-year peak flows based on the SCS Type IA rainfall distribution (Vigneswaran 2022). Figure 48. Results of GSSHA modeling for the two 100-year peak flows based on the SCS Type I rainfall distribution (Vigneswaran 2022). According to Vigneswaran (2022), this study presented one instance in which the use of regional regression equations and the FFA yielded discrepant estimates of peak flows. This difference is expected to be more pronounced for the border regions between two different hydrologic regions. Running a DRRM helped to verify and validate the discrepancy and identify the root cause for different outcomes. Texas Department of Transportation The state of Texas owns and maintains more than 14,000 miles of roadways through the Texas Department of Transportation (TxDOT). These roadways cross a variety of watersheds and geographic regions, ranging from a semi-arid west to a humid east. In particular, coastal storms originating from the Gulf of Mexico expose many roadways in east Texas to various hazards, such as flooding and the potential for damage.

62 Resilient Design with Distributed Rainfall-Runoff Modeling TxDOT has adopted DRRMs in the past for different roadway projects that evaluate the poten- tial impacts of large inflows to roadways. This synthesis presents the status of an ongoing project associated with the resiliency of the I-10 corridor (92 miles) that starts from approximately 30 miles east of downtown Houston to the Texas-Louisiana border. As presented in Figure 49, the study area crosses three counties and various large and important watersheds in the state, including the Sabine River, the Neches River, and the Trinity River. These river systems extend over hundreds of miles, but as they approach the Gulf of Mexico their conveyance is affected by coastal storms due to storm surges and large precipitation. As a result, these rivers may experience significant and widespread flooding, which in turn impacts roadways. The modeling approach used by TxDOT through consulting engineers has combined hydro- logical modeling using HEC-HMS and 1-D/2-D hydraulic modeling of overland flows and TUFLOW for representing flows through bridges and culverts. The focus of this discussion is placed on hydrologic modeling. The middle portions of the Sabine, Neches, and Trinity River basins were modeled only with the HEC-HMS tool, whereas the lower portion of each watershed was modeled using TUFLOW, as indicated in Figure 50(A). The delineated sub-watersheds in Figure  50(B) show rainfall zones determined from the National Oceanic and Atmospheric Administration (NOAA) Atlas 14. Different approaches were used within the DRRMs to enable a spatial distribution of rainfall in the studied watersheds. One of the methods was based on the data from the NOAA Atlas 14 that were adjusted using an area reduction factor approach. This approach was based on an earlier study from the USACE performed for the Neches River watershed. Rainfall depths were adjusted according to the distance of a watershed to the center of an ellipsoid shape where the maximum rainfall intensity was considered, as indicated in Figure 51. Figure 49. Extent of I-10 that is undergoing evaluation through DRRMs regarding the effects of coastal storms (Huitt-Zollars 2022).

Case Examples 63 (A) (B) Figure 50. (A) Texas watersheds modeled with DRRM, with the downstream portion modeled with the TUFLOW hydraulic model; (B) Delineation of sub-watersheds with the rainfall zones determined from the NOAA Atlas 14 (Huitt-Zollars 2022). Figure 51. Examples of distributed rainfall data using the NOAA Atlas 14 study with two derived elliptical areas (pink and green) reducing factors. Rainfall intensities decrease from the center of the ellipse and are applied in each sub-watershed (Huitt-Zollars 2022).

64 Resilient Design with Distributed Rainfall-Runoff Modeling The other approach to obtain distributed rainfall was based on gridded radar rainfall data complemented with ground rain gages. Two sources were used: the 4 km × 4 km gridded radar rainfall data from National Centers for Environmental Prediction/Environmental Modeling Center (NCEP/EMC) 4-km GRIdded Binary (GRIB) Stage IV and the 1 km × 1 km Multi-Radar Multi- Sensor (MRMS) data. The gridded rainfall data, as indicated in Figure 52, were from Tropical Storm Imelda (September 2019) for calibration of the hydraulic model (TUFLOW). The gridded radar rainfall was not directly applied on 2-D grids of TUFLOW, but average radar rainfall for a sub- basin was applied to all grids within the sub-basin. There are 15 sub-basins [Figure 50(B)] in the 2-D modeling area [Figure 50(A) and Figure 52]. Measured streamflow hydrographs from three rivers during Imelda were used as boundary conditions from upstream watersheds [Figure 50(A)]. The hydrologic modeling results derived with HEC-HMS provided inflow boundary condi- tions, and they were combined with coastal gage stage data and Federal Emergency Management Agency (FEMA) still-water elevations for design storms. These boundary conditions were added to the 2-D hydraulic modeling of river systems to compute the extents of flooding associated with extreme events and historical storms. The study also considered the possibility of sea-level rise, with the coastal stages raised by 3.54 feet, which is to assess the resiliency of an existing highway system. The modeling grids of the 2-D simulations varied from 50 ft × 50 ft for areas that were not near roads to a minimum of 6 ft × 6 ft to represent medians and conveyance structures. As a result, a very detailed representation of areal flooding was yielded by the model, as shown in Figure 53 for the case of a rain event with a 1% Annual Exceedance Probability (AEP). Summary of the Case Examples • The case example of NCDOT demonstrated the value of the combination of HEC-HMS and HEC-RAS tools to provide detailed spatial information on extent and severity of flooding scenarios. These detailed results from the rain-on-grid study informed the decision regarding the elevation of the I-95 corridor that is subject to coastal storm and, with that, improve the transportation resiliency. • The case example of KYTC presented flooding issues along KY 10/15 and was very useful in showing the benefits of using a SWMM 5 tool to confirm potential conveyance issues in the road drainage. These 2-D modeling results not only confirmed the sites in which flooding was STAGE IV (4x4km) MRMS (1x1km) Figure 52. Different gridded rainfall data representing Tropical Storm Imelda (September 2019) in the model area. The polygon shows the studied area along I-10 in the 2-D hydraulic modeling (Fang 2022).

Case Examples 65 observed but also simulated scenarios in which the conveyance was improved and attenuated the reported flooding. • The case example of Caltrans presented a case in which GSSHA was used to provide precise hydrological estimates of peak flows that were used in turn for scour calculations in bridges. For the studied watershed, the regional regression equations and the FFA yielded discrepant estimates of peak flows. This DRRM application helped to verify and validate the discrepancy and identified the root cause for different peak estimates. • The case example of TxDOT presented an ongoing hydrologic–hydraulic study that involves DRRMs applied along I-10 to evaluate its resiliency against coastal storms. The approach used different distributed rainfall approaches along with HEC-HMS based on an area reduc- tion methodology and radar data. The hydrological model results and coastal gage data were combined with a river hydraulic model to compute precise flooding extent maps. This is an example of an integrated and large-scale DRRM application in which various watersheds and roadway crossings are considered. Figure 53. Example of flooding extent map generated by the 2-D model with inflows supplied by the DRRM and the use of coastal boundary conditions for a rain event with a 1% Annual Exceedance Probability and no sea-level rise considered (Huitt-Zollars 2022).

Next: Chapter 5 - Summary of Findings »
Resilient Design with Distributed Rainfall-Runoff Modeling Get This Book
×
 Resilient Design with Distributed Rainfall-Runoff Modeling
MyNAP members save 10% online.
Login or Register to save!
Download Free PDF

The increased frequency of extreme rainfall events, inland and coastal flooding, and other water-related stressors poses challenges to roadway infrastructure.

The TRB National Cooperative Highway Research Program's NCHRP Synthesis 602: Resilient Design with Distributed Rainfall-Runoff Modeling documents the practices of state departments of transportation on the use of DRRMs and identifies state DOTs that have adopted DRRMs and the context in which these models are applied.

READ FREE ONLINE

  1. ×

    Welcome to OpenBook!

    You're looking at OpenBook, NAP.edu's online reading room since 1999. Based on feedback from you, our users, we've made some improvements that make it easier than ever to read thousands of publications on our website.

    Do you want to take a quick tour of the OpenBook's features?

    No Thanks Take a Tour »
  2. ×

    Show this book's table of contents, where you can jump to any chapter by name.

    « Back Next »
  3. ×

    ...or use these buttons to go back to the previous chapter or skip to the next one.

    « Back Next »
  4. ×

    Jump up to the previous page or down to the next one. Also, you can type in a page number and press Enter to go directly to that page in the book.

    « Back Next »
  5. ×

    To search the entire text of this book, type in your search term here and press Enter.

    « Back Next »
  6. ×

    Share a link to this book page on your preferred social network or via email.

    « Back Next »
  7. ×

    View our suggested citation for this chapter.

    « Back Next »
  8. ×

    Ready to take your reading offline? Click here to buy this book in print or download it as a free PDF, if available.

    « Back Next »
Stay Connected!