Sample__4___Flood_Report.pdf

Draft – 30 June 2008

Approved for public release; distribution is unlimited.

Flood Inundation Mapping – Camp Casey Dam Camp Casey, Korea

DRAFT – 30 June 2008

DRAFT – 30 June 2008 ii

DRAFT – 30 June 2008 iii

Contents

Preface...................................................................................................................................... vii

1 Introduction Purpose Flood History at Camp Casey Scope Study Overview Data and Document Acquisition

Geographic Information System (GIS) Data Meteorological Data Hydrologic Design Documents Historical Reports

2 Watershed Hydrology Watershed Overview Watershed Model Setup Model Validation Design Flow Calculations Comparison of Computed Design Flows with Previous Studies

3 Hydraulic Analysis Introduction Data Collection HEC-RAS Model Setup Geographic Extent of HEC-RAS Modeling Validating the HEC-RAS Model HEC-RAS Modeling of Return Period Events Summary of Estimated Overbank Flow Conditions Flow Conditions Downstream of Dam Limitations of HEC-RAS Modeling Hydraulic Modeling Summary

4 Inundation Modeling and Mapping Introduction Overland Flow Model Inundation Modeling Inundation Flood Maps

DRAFT – 30 June 2008 iv

5 Dam Breach Modeling Introduction Dam Failure Assumptions Dam Breach Modeling Scenarios Dam Breach Modeling Results

Sunny-Day Dam Failure Conditions Dam Failure at Peak of 100-Year Flood Event

Dam Breach Conclusions

6 Summary, Conclusions and Limitations Summary Conclusions Limitations

References

Appendix A – Photographs

Figures and Photographs

Figure 2.1. Watershed Boundary Upstream of Camp Casey Dam Figure 2.2. Hydrologic Model Sub-basins and Stream Network Figure 2.3 Computed Flow Hydrograph for August 5-6, 1998 Event Figure 2.4 Precipitation Temporal Distribution Figure 2.5 Hydrologic Model Results for Design Runoff Events Figure 2.6 Peak Runoff Values vs. Return Period for Various Methods Figure 3.1 Extent of Topographic Survey Data Conducted for Study Figure 3.2 Extent of Entire Topographic Data Coverage Figure 3.3 Overview of HEC-RAS Model Figure 3.4 Computed Water Surface Elevations with and without Dam Figure 3.5 Estimated Upstream Extent of Dam Influence at 100-year Return Period Flow Figure 3.6 Computed Flow Hydrograph at Dam for July 2006 Event Figure 3.7 Computed Water Surface for July 2006 Event Just Upstream of Dam Figure 3.8 Estimated Extent of Overbank Flow Conditions for July 2006 Event

DRAFT – 30 June 2008 v

Figure 3.9 Excerpt from IMCOM – Korea Presentation on July 2006 Flood Event Figure 3.10 Maximum Computed Water Elevation for 2-year Event Just Upstream of Dam Figure 3.11 Maximum Computed Water Elevation for 5-year Event Just Upstream of Dam Figure 3.12 Maximum Computed Water Elevation for 10-year Event Just Upstream of Dam Figure 3.13 Maximum Computed Water Elevation for 20-year Event Just Upstream of Dam Figure 3.14 Maximum Computed Water Elevation for 50-year Event Just Upstream of Dam Figure 3.15 Maximum Computed Water Elevation for 100-year Event Just Upstream of Dam Figure 3.16 Computed Water Elevations vs. Time for Design Flow Events Just Upstream of Dam Figure 3.17 Estimated Locations of Overbank Flow for 2-year Return Period Event Figure 3.18 Estimated Locations of Overbank Flow for 5-year Return Period Event Figure 3.19 Estimated Locations of Overbank Flow for 10-year Return Period Event Figure 3.20 Estimated Locations of Overbank Flow for 20-year Return Period Event Figure 3.21 Estimated Locations of Overbank Flow for 50-year Return Period Event Figure 3.22 Estimated Locations of Overbank Flow for 100-year Return Period Event Figure 4.1 Estimated Flood Inundation Area Due to Overbank Flow Upstream of Dam for 2-Year Return Period Event Figure 4.2 Estimated Flood Inundation Area Due to Overbank Flow Upstream of Dam for 5-Year Return Period Event Figure 4.3 Estimated Flood Inundation Area Due to Overbank Flow Upstream of Dam for 10-Year Return Period Event Figure 4.4 Estimated Flood Inundation Area Due to Overbank Flow Upstream of Dam for 20-Year Return Period Event Figure 4.5 Estimated Flood Inundation Areas Due to Overbank Flow Upstream of Dam for 50-Year Return Period Event Figure 4.6 Estimated Flood Inundation Areas Due to Overbank Flow Upstream of Dam for 100-Year Return Period Event Figure 5.1 Computed Flow Through 6-meter Wide Dam Breach under Sunny Day Conditions Photo A1. Casey Dam Looking Downstream from South Bank of Channel– October 2007

DRAFT – 30 June 2008 vi

Photo A2. Casey Dam Looking Upstream along South Bank of Channel – August 2007 Photo A3. View of North Channel Bank Looking Upstream – July 2006 Photo A4. View of North Channel Bank Looking Upstream – August Photo A5. View Looking along South Berm Towards Dam and Footbridge

– October 2007 Photo A6. Channel Looking Downstream from Bridge #1090 – October Photo A7. Channel Looking Upstream from Bridge #1090 – October Photo A8. Channel Looking Downstream from Footbridge Below Dam – October 2007 Photh A9. Channel Looking Downstream from Footbridge Below Dam – August 2007 Photo A10. Channel Looking Upstream from Western Edge of Camp Casey – October 2007 Photo A11. Security Grate across Channel at Western Edge of Camp Casey – October 2007

Tables

Table 2.1 Hourly Precipitation Data (mm) for Dongduchon Weather Station Table 2.2 Precipitation Totals for Various Return Period 24-Hour Duration Rainfall Events Table 2.3 Return Period Peak Flow Values from Hydrologic Model Table 5.1 Estimated Flow Rate Increase Above the 100-Year Flood Flow Due to Dam Breach Table 5.2 Estimated Water Level Increase the Above 100-Year Flood Level Due to Dam Breach

DRAFT – 30 June 2008 8

1 Introduction

Purpose

In 2005, a dam was constructed at Camp Casey, Korea for the purpose of providing a water intake source for the Camp Casey water plant. The new dam replaced a smaller dam that had been in place at the same location. In July 2006, a rainfall event occurred that resulted in the retaining walls upstream of the Camp Casey dam being overtopped along both the north and south banks of Casey Creek. The resultant floodwaters caused property damage within Camp Casey and in the local community of Dongduchon outside of Camp Casey. As a result of this flood and the associated property damage, the Camp Casey dam was reclassified by the Installation Management Command (IMCOM) as a significant hazard potential dam.

Army Regulation (AR) 420-72 dictates that an Emergency Action Plan (EAP) be developed for all dams classified as significant hazard potential dams, and flood inundation maps associated with the dam are an integral part of the

EAP.

The purpose of this study is to assess the potential flood risks associated with the Camp Casey dam and to provide flood inundation maps to support the eventual development of an EAP for the dam. The situation at Camp Casey associated with the dam is somewhat unique in that the primary risk of flooding conditions is not associated with potential failure of the dam and the resultant downstream flood effects. Rather, the potential flood conditions are associated with the upstream effects of the dam and flooding that may result from overtopping of the banks along either side of the stream channel. The results of this study are intended to provide an overview of areas within Camp Casey that are likely to be at risk of some level of flood inundation associated with the dam during various rainfall events.

Flood History at Camp Casey

Camp Casey has been subjected to a number of serious flood events through the years, and previous reports have been written that have detailed the flood conditions and provided recommendations for flood mitigation at Camp Casey. Among those reports are the 1984 report entitled Flood Hazard Assessment for Camp Casey prepared by the U.S. Army Engineer District, Far East and the 1998 report entitled Camp Red Cloud and Camps Hovey and Casey, Flood Mitigation Analysis prepared by the U.S. Army Engineer District, Alaska. These reports provide a thorough overview of the conditions that existed at Camp Casey during 1984 and 1998, summaries of the flood events that occurred during the 1984 and 1998 floods, and each contains some guidance on flood management and flood mitigation measures. The 2006 flooding was the most recent such event at Camp Casey, and in historical

DRAFT – 30 June 2008 9 terms the flow in Casey Creek during the July 2006 event was probably not a very large flood event. However, the presence of the dam in the stream channel impacted the flow conditions and contributed to the extent of flooding that occurred during that event.

Scope

The scope of this study is limited to the specific impact of the dam at Camp Casey on flood conditions under various rainfall events. The study is not intended to provide a complete flood risk analysis for the entire installation at Camp Casey. This study is specifically focused on the dam and the impact on flood conditions that may result due to the presence of the dam. As such, the hydrology of the drainage basin upstream of the dam was modeled to generate a series of design flow hydrographs for the main channel of Casey Creek and the primary tributary that joins the creek near the dam. These design flow hydrographs represent the 2-year through 100-year return period flood events at the Camp Casey dam. Hydraulic modeling was conducted of these flow hydrographs in the channel, and overland flood modeling was conducted for flows that were predicted to exceed the banks of the stream. The final output of these analyses is a series of flood inundation maps that depict the estimated areas within Camp Casey that are expected to be at risk of flooding during each of these events. These maps are presented in Chapter 5 of this report.

Study Overview

This study consists of several components, each of which is ultimately related to development of flood inundations maps associated with the Camp Casey dam. The tasks performed in this study included data acquisition, hydrologic modeling of the Camp Casey watershed, 1-dimensional hydraulic modeling of Casey Creek, 2-dimensional overland flow modeling and development of flood inundation maps. A chapter in this document is devoted to brief summary of each of those tasks is provided below, and results and final products are provided later in this report.

Data and Document Acquisition

A variety of data sources and documents were collected to support the study.

A brief summary of those data and documents is provided below.

Geographic Information System (GIS) Data

The U.S. Army Engineer District, Far East provided GIS data including geology, satellite imagery, datasets for various infrastructure and other relevant features, and digital terrain data.

Land use, land cover and soil data were obtained from the National Geospatial Intelligence Agency.

DRAFT – 30 June 2008 11

2 Watershed Hydrology

Watershed Overview

The watershed above the Camp Casey dam is bounded by mountains to the north, east and south, including Soyo Mt., Kuksabong Mt., Haepyong Mt. and Chilbong Mt. which range in elevation from just over 500 meters to just over 750 meters above sea level. The watershed below these mountains in which Camp Casey is located is drained by the Binyang Chon (hereafter referred to as Casey Creek in this report) which flows through the main installation and is a tributary to the Sin Chon just to the west of Camp Casey in the city of Dongduchon. Runoff through Camp Casey along Casey Creek is comprised of the combined flow from a series of small drainage areas that flow into Casey Creek from both sides of the creek. Within the main installation of Camp Casey and Camp Hovey upstream, a significant level of development and associated storm drainage infrastructure contributes to the total flow. The 1984 report stated that the upstream drainage area above the location of the Camp Casey dam was 41 square kilometers (km2).

Watershed Model Setup

The watershed drainage area upstream of the Camp Casey dam was modeled using the Watershed Modeling System (WMS) and the HEC-1 hydrologic model. Digital elevation data were used to represent the watershed topography. The overall watershed was delineated and then subdivided into 15 sub-basins based primarily on the drainage patterns of the tributaries to Casey Creek. The total drainage area of the watershed upstream of the Camp Casey dam was computed to be 41.7 km2 in this study, which compares quite well with the value reported in previous studies. The overall watershed boundary and the 15 sub-basins that were derived for the hydrologic model are shown in Figures 2.1 and 2.2, respectively.

The WMS provides tools for automatic calculation of the geometric properties of each sub-basin based on the digital terrain data, such as area and slope, and these computed values were used as model input parameters where needed. Infiltration parameters for each sub-basin were assigned based on the vegetative cover or land use and soil characteristics of each basin using standard hydrologic design data. The infiltration parameters were represented in the model using the Soil Conservation Service (SCS) curve number method. Unit hydrograph parameters for each sub-basin were established using the SCS dimensionless unit hydrograph method.

Dam Drainage Area = 41.7 km2

Figure 2.1. Watershed Boundary Upstream of Camp Casey Dam

Dam

Figure 2.2. Hydrologic Model Sub-basins and Stream Network

Model Validation

There are no stream gages within the Casey Creek watershed, and there are therefore no observed stream flow data with which to perform calibration of the hydrologic model. The 1998 report does provide one runoff event for which there are adequate rainfall data and anecdotal flow observations that can serve as the basis for some level of validation that the runoff computed by the hydrologic model provides a reasonable estimate of actual runoff conditions in the watershed. The peak runoff from the August 1998 flood event is reported to be an estimated 454 cubic meters per second (cms), or 16,050 cubic feet per second (cfs) as the flood passed through the lower end of Camp Casey in the vicinity of the existing Camp Casey dam. This estimate of peak flow for the 1998 event is not based on actual stream gage measurements, but was derived from observations of water surface elevations are several control sections along the channel. This represents the best and only runoff event for which any level reasonable information is available to validate the hydrologic model output at least in general terms.

Hourly rainfall data from the Dongduchon weather station were obtained for the August 1998 rainfall event. The Dongduchon weather station is located just over 2 kilometers (km) south of the Camp Casey dam. The rainfall during the August 1998 event started on August 5 and continued through August 8.

The largest 24-hour amount of approximately 353 mm (or nearly 14 inches) at the Dongduchon weather station was recorded between August 5 and 6. For validation of the hydrologic model, the rainfall that occurred during an approximately 19 hour period starting late on August 5 and continuing through part of August 6 was used as the forcing rainfall function in the model to compute the corresponding runoff. Table 2.1 shows the hourly rainfall data from the Dongduchon weather station for the period from July 30 through August 8, 1998. The rainfall data values that are highlighted in red were used as input to the hydrologic model to validate the runoff event.

Table 2.1 Hourly Precipitation Data (mm) for Dongduchon Weather Station

Adjustments were made to the hydrologic model infiltration and unit hydrograph parameters to account for the fact that rainfall had occurred over the watershed for 5 of the 6 days prior to the August 5 event. As such, the soil conditions in the watershed would have been generally wet. In terms of hydrologic model parameters, wet antecedent moisture conditions may be assumed if more than 53 mm of rainfall fell during the 5 days prior to the rainfall event being modeled. For the 5 days leading up to August 5, 1998, there were over 155 mm of rainfall recorded at the Dongduchon weather station. Thus, the model parameters were adjusted to reflect the wet moisture conditions in the watershed at the start of the August 5 rainfall event.

The model results showed a peak computed runoff of 453.79 cms, which compared extremely well with the estimated peak runoff of 454 cms as reported in the 1998 report. The runoff hydrograph at the location of the Camp Casey dam computed from the hydrologic model for the August 1998 event is shown in Figure 2.3. The starting time (Time = 0) for the hydrograph below corresponds with the start of the rainfall at 1600 on August 5. The peak flow occurred 630 minutes (10.5 hours) later, which corresponds to 0230 on August 6. Although no records are available to provide the actual timing of the peak flow during that event, eyewitness reports recounted in the 1998 report indicated that the peak runoff from the flood occurred in the early morning hours of August 6, 1998. Based on the estimated peak flow and timing of the peak as presented in the 1998 report, the computed runoff hydrograph as shown in Figure 2.3 is considered to be a very reasonable representation of the runoff that occurred in Casey Creek during the August 5- 6, 1998 flood event. The hydrologic model developed for this study was therefore also considered to be capable of providing a reasonable representation of runoff for a given rainfall event.

Figure 2.3 Computed Flow Hydrograph for August 5-6, 1998 Event

DRAFT – 30 June 2008 15

Design Flow Calculations

The end product of this study is a set of flood inundation maps showing the areas within Camp Casey expected to be at some risk of flooding during various levels of flood events. To define those different events, a series of design runoff hydrographs (stream flow vs. time) were generated to reflect the estimated 100-year, 50-year, 20-year, 10-year, 5-year and 2-year return period runoff events for Camp Casey. Ideally the statistical determination of these events is based on a long historic record of observed flow data for the area of interest. However, for Camp Casey there are no observed flow records.

In the absence of observed flow data, there are several other methods that may be used to estimate these design flows including regional regression methods and hydrologic modeling analysis. In the 1984 report for Camp Casey, the Rational Method (a simplified hydrologic modeling approach) and Regional Frequency Method were both presented. The flows resulting from the Rational Method were slightly more conservative and were used for the hydraulic computations in that report.

For this study, the hydrologic model was used to generate flow hydrographs for the various return period events. The model parameters determined from validation for the 1998 event were used, and rainfall input representing the 100-year, 50-year, 20-year, 10-year, 5-year and 2-year return period rainfall events was used to compute the corresponding return period flow hydrographs. Two methods were used to estimate the precipitation totals for the various return period rainfall events. First, rainfall data was taken from the Republic of Korea, Ministry of Construction and Transportation (MOCT) document entitled Year 1999 Water Resources Management Techniques Research Report: Appendix 1. Korea Rainfall Probability Map. The precipitation values contained in that document are based on data through 1999, and the values that are valid for Camp Casey were estimated from the isohyetal lines of rainfall depth for the various return period events provided on the maps. The second method used to estimate the precipitation totals for the various return period events was an annual exceedance series analysis of the hourly precipitation data from the Dongduchon weather station for the 1998-2007 period. These data offered the advantage of being more current since the period of record included data up through 2007, although the period of record is not particularly long. The results of these analyses are shown below in Table 2.2 for 24-hour duration storms. In the values below, adjustments have been made for the annual exceedance analysis and for the areal reduction factor for the Camp Casey watershed.

Since the data from the Dongduchon weather station represents more recent precipitation events in the vicinity of Camp Casey, and since the precipitation totals from the Dongduchon weather station represent more conservative (higher) estimates of these rainfall events, the Dongduchon weather station values were used as input to the hydrologic model for computing the runoff for each event.

Return Period MOCT Data Dongduchon Weather Station

100-year 460 mm 479 mm

50-year 407 mm 443 mm

20-year 340 mm 395 mm

10-year 287 mm 354 mm

5-year 238 mm 308 mm

2-year 157 mm 231 mm

Table 2.2 Precipitation Totals for Various Return Period 24-Hour Duration Rainfall Events

The data presented in Table 2.2 provide only the total rainfall during each event, and the temporal distribution of that rainfall must also be determined before the hydrologic model can be used. The Republic of Korea, Ministry of Construction and Transportation (MOCT) document entitled Year 1999 Water Resources Management Techniques Research Report: Appendix 2.

Temporal Distribution of Regional Design Rainfall provides detailed analysis of rainfall temporal distributions over Korea. Data from that document for the gage station closest to Camp Casey was used for this study. Figure 2.4 shows the cumulative rainfall distribution that was used in the hydrologic model for the rainfall amounts shown above in Table 2.2. The rainfall total values shown in Table 2.2 and the temporal distribution shown in Figure 2.4 were used as the precipitation input to the hydrologic model. The resultant flow hydrographs are shown in Figure 2.5 for the events from 2-year through 100-year return periods, and Table 2.3 shows the peak runoff values for each of these events.

Figure 2.4 Precipitation Temporal Distribution

Figure 2.5 Hydrologic Model Results for Design Runoff Events

Return Period Peak Flow (cms)

100-year 444

50-year 404

20-year 350

10-year 305

5-year 254

2-year 170

Table 2.3 Return Period Peak Flow Values from Hydrologic Model

Comparison of Computed Design Flows with Previous Studies

As stated above, the lack of observed flow data for Casey Creek creates cause for uncertainty in the analysis of return period flows, and it is prudent to compare the results from several methods before making a determination of the appropriate values to use for analysis. The 1984 report presented two completely separate methods for determining the design flows, and Figure 2.6 provides a plot of the results from those 2 other methods along with the results from the hydrologic modeling approach used in this study. From Figure 2.6, the results obtained in this study show peak flows that are slightly higher than those obtained with either of the methods presented in the 1984 report. This difference may be attributed to changes in the watershed that have occurred since 1984 such as increased urbanization, changes in the rainfall intensity and frequency patterns over the area, uncertainties in the assumptions and data associated with each method, or some combination of all of these factors. For the purposes of this analysis, the results obtained from the hydrologic modeling conducted for this study are considered to be very reasonable estimates of the runoff values for the various return period events and were used in the hydraulic analysis and flood inundation mapping presented in the following Chapters of this report.

Figure 2.6 Peak Runoff Values vs. Return Period for Various Methods

DRAFT – 30 June 2008 19

3 Hydraulic Analysis

Introduction

The design flow hydrographs computed with the hydrologic model as detailed in the previous chapter served as inflow hydrographs for a 1-dimensional hydraulic model of Casey Creek. The U.S. Army Corps of Engineers, Hydrologic Engineering Center – River Analysis System (HEC-RAS) model was chosen as the hydraulic model for this study because it allows for both steady and unsteady flow analysis and for a variety of hydraulic structures such as dams, weirs, bridges and levees to be accounted for in the model. The primary purpose of the hydraulic model in this study was to determine whether or not the flow in Casey Creek is expected to exceed the capacity of the stream channel for the various return period flow events and thus result in overbank flooding. The model was also used to determine the locations along the channel upstream of the dam where such overbank flow may occur for each event. Based on the results of this hydraulic analysis, 2-dimensional overland flow modeling was then conducted to determine the extent of flooding beyond the main channel and to ultimately generate flood inundation maps. Details of this aspect of the study and the final inundation maps are presented in Chapter 4.

Data Collection

The primary information required for the HEC-RAS model includes detailed topographic data for the stream channel and immediate floodplain or overbank areas, locations and elevations of structures along the channel such as retaining walls, details on bridges and other structures that impact the channel, details on dams or weirs across or within the channel, and information on the vegetation and other factors within the channel that effect the roughness parameters in the model. For this study, a detailed topographic survey of the study area along the stream channel was conducted by the U.S.

Army Engineer District, Far East. The topographic survey data were collected during late 2007 and early 2008, and the area for which the survey was conducted extended from the vicinity of the dam to a distance of approximately 1,000 meters upstream of the dam. This was considered to more than adequately cover the area along Casey Creek that is influenced by the presence of the dam. The hydraulic conditions farther upstream along Casey Creek are not impacted by the dam. Figure 3.1 shows an overview of the area for which topographic survey data was specifically collected for this study.

Dam

Figure 3.1 Extent of Topographic Survey Data Conducted for Study

Survey data collected for this study focused on the channel upstream of the dam. However, topographic data in the overbank areas, downstream of the dam, and across Camp Casey where overland flooding may occur were also required. Topographic data collected during previous surveys by the U.S.

Army Engineer District, Far East was provided to supplement the data collected specifically for this study. The resultant topographic dataset encompassed all of the areas along Casey Creek and within Camp Casey that may be impacted by flooding associated with the dam. Figure 3.2 provides an overview of the extent of this topographic dataset.

Dam

Figur Gate DRAFT – 30 June 2008 e 3.2 Extent of Entire Topographic Data Blue Dots = Elevation

Data Points Coverage

HEC-RAS Model Setup

The topographic data was used to create a model of Casey Creek using the HEC-RAS hydraulic model. The model for this study extended from the location where Casey Creek exits Camp Casey at the western edge of the golf course and continued upstream just beyond the footbridge near the weir for the vehicle wash rack. The total distance along the main channel covered by the model was approximately 1,380 meters, with 470 meters being downstream of the dam and 910 meters being upstream of the dam. A total of 70 cross-sections were included in the model along the main channel, and the cross-sections were spaced at an average interval of 20 meters along the channel. The drainage ditch that joins Casey Creek from the south just upstream of the dam was also included in the HEC-RAS model and was represented with 11 cross-sections along the ditch. The existing dam was included in the model, and bridge #1090 was also included in the model. The footbridge at the upper end of the HEC-RAS model and the footbridge just below the dam were not included in the model as it was judged that neither of these structures has a significant impact on the water surface profile upstream of the dam or the potential flood conditions that may result from the presence of the dam. Figure 3.3 provides an overview of the extent of the HEC-RAS model that was generated for this study.

Figure 3.3 Overview of HEC-RAS Model

Geographic Extent of HEC-RAS Modeling

The scope of this study was limited to assessing the potential flood conditions associated with the Camp Casey dam. As such, the distance upstream of the dam that the dam exerts influence on the hydraulics of the stream flow is a critical component of the study. To verify that the HEC-RAS model created

Dam

Bridge #1090

Wash Rack Weir

Ditch for this study extended upstream an adequate distance to model the full influence of the dam, the HEC-RAS model was run under two initial conditions after all of the cross-sections and input data were entered into the model. The first run was to compute the water surface elevation along the stream with the existing dam in place, and the second run was to compute the water surface elevations along the stream with the dam removed entirely from the model. Both model runs were made with the estimated 100-year return period flow rate of 444 cms. The results indicate that for the estimated 100-year return period flow, the influence of the dam extends upstream to model cross-section 0.86, or a distance of approximately 400 meters upstream of the dam. Figure 3.4 provides a graph of the computed water surface elevations from the model runs with the dam and without the dam. Figure 3.5 shows the approximate upstream extent of the dam’s influence within the stream based on this analysis.

The HEC-RAS model created for this study extends over 500 meters farther upstream from the location that was estimated to be the upstream limit of the dam’s hydraulic influence. Therefore, even allowing for some level of uncertainty associated with the modeling and assumptions in the hydrologic and hydraulic analyses, the HEC-RAS model easily extends far enough upstream of the dam to capture the hydraulic conditions associated with the dam.

Dam Location

XS 0.86

Figure 3.4 Computed Water Surface Elevations with and without Dam

Estimated Upstream Extent of Dam Influence at XS 0.86

Figure 3.5 Estimated Upstream Extent of Dam Influence at 100-year Return Period Flow

Validating the HEC-RAS Model

There are no actual measurements of observed flow or specific water surface elevations along Casey Creek with which to verify or validate output from the HEC-RAS model. Water surface profiles were computed and presented in the 1984 report. However, conditions along the channel have undoubtedly changed due to vegetation growth, erosion, sedimentation and channel improvements since that study was conducted, and comparing specific water surface elevations from 1984 with those computed in 2008 would not seem to prove particularly useful. The most recent information and observations regarding flow and flood conditions along Casey Creek are from the July 2006 flood event. This event occurred recently enough that the conditions of the channel have most likely not changed significantly since the event occurred.

In a presentation entitled Army Transformation Infrastructure Program – In-Progress Review by Mr. Son T. Ha of the Installation Management Command, Korea (IMCOM-Korea) dated 13 December 2006, a map is included that shows the general locations along the north and south sides of Casey Creek where over bank flooding occurred during the July 2006 flood.

Based on that information, the HEC-RAS model was run for the July 2006 event to determine if the results from the HEC-RAS model indicated over bank flow occurring at the same locations as shown in the IMCOM presentation.

Dam

Bridge #1090

The first step was to estimate the peak flow that occurred in Casey Creek at the dam during the July 2006 event. Hourly rainfall data from the Dongduchon weather station shows a total of 208 mm of precipitation on July 12, 2006. Comparing this to the data presented in Table 2.2 indicates that this rainfall amount was slightly less than a 2-year return period event for the 24-hour duration. The hourly rainfall data was then used as the precipitation input to the hydrologic model and a runoff hydrograph for the July 2006 event was computed. The computed hydrograph is shown in Figure 3.6 and shows an estimated peak flow of 186 cms. Comparing this peak flow to the data presented in Figure 2.6 indicates that the peak flow for this event was just slightly higher than a 2-year return period event. This means that in any given year, there is nearly a ½ or 50% chance of flow at this magnitude occurring. Note that there is not expected to be an exact correlation of the return periods between the rainfall and runoff for an actual rainfall-runoff event, i.e., an actual rainfall event with a total 24-hour precipitation of a 2-year rainfall event will not necessarily result in peak flow that equals the 2-year flow event. This may be due to differences in the temporal distribution of the actual rainfall compared to the hypothetical rainfall distribution used for the hydrologic return period analysis, and may be due to differences in the assumed initial soil moisture conditions when the rainfall event occurred. For the July 2006 event, it is likely that there were short durations of highly intense rainfall during the event which caused the peak flow to be higher than that computed using the hypothetical rainfall patterns as outlined in Chapter

2. Thus, the rainfall event with just less than a 2-year return period resulted in a runoff event with just more than a 2-year return period.

Figure 3.6 Computed Flow Hydrograph at Dam for July 2006 Event

The computed peak flow of 186 cms for the July 2006 event was run in the HEC-RAS model to determine the associated water surface elevations along the channel and to locate areas along the channel where the flow may have exceeded the banks of the channel and caused overland flooding. The results show that the flow does exceed the capacity of the channel as it existed in July 2006 for some distance upstream of the dam along both the north and south banks of the channel. Note that since 2006, a berm or short levee has been constructed to the south of the channel and a longer berm or small levee has been constructed along a section of the north bank of the channel.

Figure 3.7 shows a cross-section from the HEC-RAS model just upstream of the dam with the water surface shown for 186 cms. Labeled on that figure are the locations of the side walls along the channel as they existed in July 2006, and the new berms or levees that have been subsequently added. Clearly, the water level at this flow rate is shown to exceed the top of the channel walls on both sides of the stream at this location. This would have resulted in overland flooding on both the north and south sides of the channel during this event, as was actually experienced.

New South Berm

New North Berm

North Channel

Wall

South Channel

Wall

Figure 3.7 Computed Water Surface for July 2006 Event Just Upstream of Dam

Water surface elevations this flow rate at the other cross-sections upstream of the dam were reviewed to estimate the extent upstream of the dam where overbank flow would have occurred during the July 2006 event, i.e., where the water level would have exceeded the top elevation of the concrete channel walls. Figure 3.8 shows the general extent of where overbank flow is estimated to have occurred based on the results of this analysis. Figure 3.9 is taken from the December 2006 IMCOM presentation referenced above.

Estimated Extent of Overbank Flow Conditions along North Bank

Estimated Extent of Overbank Flow Conditions along South Bank

Water Level in Creek Would Prevent Ditch from

Flowing into Creek

Figure 3.8 Estimated Extent of Overbank Flow Conditions for July 2006 Event

Figure 3.9 Excerpt from IMCOM – Korea Presentation on July 2006 Flood Event

DRAFT – 30 June 2008 27

An important note is that the estimated water level in Casey Creek during the peak of this event would have been nearly 2 meters above the invert elevation of the drainage ditch that joins Casey Creek from the south just upstream of the dam. Thus, any water flowing down that ditch was prevented from flowing into Casey Creek and would have added to the flooding along the south side of the channel that eventually passed across the golf course and through the base.

Note that there is obviously some level of uncertainty based on the lack of specific data and on the assumptions that had to be made during this analysis.

However, the general extent of overbank flow as estimated from the hydrologic and hydraulic modeling conducted in this study does coincide well with the observations and anecdotal evidence from the July 2006 event.

HEC-RAS Modeling of Return Period Events

The HEC-RAS model of Casey Creek was run with the design flow hydrographs that were computed with the hydrologic model for the 2-year through 100-year through return period events. The purposes of this modeling were:

• to determine what return period flood events, if any, would result in overtopping of the channel under the current channel conditions

• to determine the locations or extent along the channel upstream of the dam where overtopping of the channel banks may occur

• to estimate the potential duration of overbank flooding along the channel based on the design flow events

Inflow points for the HEC-RAS model were established at the upstream end near the location of the wash rack weir and at the upper end of the drainage ditch that flows into Casey Creek from the south just upstream of the dam.

The design flow hydrographs from the hydrologic model at those two locations were entered as the upstream boundary conditions, and an unsteady flow analysis of each of the events was conducted with HEC-RAS to route each entire hydrograph through the system. The results of this analysis provided information on the location and severity of overbank flow during the peak conditions, and the unsteady analysis also provided information on the duration that overbank flow may be expected to persist for each event.

Figures 3.10 through 3.15 show the computed maximum water surface elevation just upstream of the dam for the 2-year through 100-year return period events.

New South Berm

New North Berm

North Channel

Wall

South Channel

Wall

Figure 3.10 Maximum Computed Water Elevation for 2-year Event Just Upstream of Dam

New South Berm

New North Berm

North Channel

Wall

South Channel

Wall

Figure 3.11 Maximum Computed Water Elevation for 5-year Event Just Upstream of Dam

New South Berm

New North Berm

North Channel

Wall

South Channel

Wall

Figure 3.12 Maximum Computed Water Elevation for 10-year Event Just Upstream of Dam

New North Berm

New South Berm

North Channel

Wall

South Channel

Wall

Figure 3.13 Maximum Computed Water Elevation for 20-year Event Just Upstream of Dam

New North Berm

New South Berm

North Channel

Wall

South Channel

Wall

Figure 3.14 Maximum Computed Water Elevation for 50-year Event Just Upstream of Dam

New North Berm

New South Berm

North Channel

Wall

South Channel

Wall

Figure 3.15 Maximum Computed Water Elevation for 100-year Event Just Upstream of Dam

DRAFT – 30 June 2008 31

As shown in Figures 3.10 through 3.15, the maximum computed water elevation does not exceed the top elevation of the south berm for any of the design flow events. For the 100-year and for the 50-year return period events, the maximum computed water elevation does overtop the north berm, and for the 20-year return period event the maximum computed water elevation comes extremely close to the elevation of the top of the north berm. For the 10-year, 5-year and 2-year return period events, the maximum computed water elevation remains well within the channel between the north and south berms upstream of the dam.

The above figures represent the maximum computed water elevation at the HEC-RAS model cross-section located approximately 15 meters upstream of the dam and show what return period events are expected to result in overtopping of the channel banks at peak flow conditions. During an actual flood event, the water level will obviously vary over time. Before the event, the water level will be low, it will rise and peak during the event, and then it will recede over time. The HEC-RAS unsteady flow analysis results provide the computed water elevations over time for each event, and these results can provide some general guidance as to the duration of time that the overbank flow conditions may last. Figure 3.16 provides a graph of the water elevation

vs. time for the 2-year through 100-year return period events, and includes a horizontal line that represents the top elevation of the north berm at the location where these flow data were computed. Notice from Figure 3.16 that the computed 2-year, 5-year and 10-year return period events do not overtop that north berm. The computed 20-year return period event is also shown as not overtopping the berm, but the computed maximum water elevation for this event is only 0.02 meters below the top of the berm. This is clearly within the level of uncertainty associated with this analysis, and the 20-year return period event may be considered to pose some risk of overtopping the north berm. The 50-year and 100-year return period events do overtop the north berm for some period of time during flood events. The 50-year return period event is above the top of the north berm for approximately 3.5 hours, and the 100-year return period event is over the top of the north berm for the duration of approximately 5 hours. Note that these durations are based solely on the hypothetical design flow events generated for this study, and actual conditions during a flood event may result in longer or shorter durations of overtopping at these flow rates.

Figure 3.16 Computed Water Elevations vs. Time for Design Flow Events Just Upstream of Dam

Summary of Estimated Overbank Flow Conditions

The model output presented in Figures 3.10 through 3.16 provides a sample of the results for one cross-section location approximately 15 meters upstream of the dam. The corresponding results for the remainder of the model cross-sections upstream of the dam were reviewed in the same manner to determine at what locations overbank flow was predicted by the model. The 2-year through 20-year return period flows were not shown to overtop either the south or north berms at any location upstream of the dam. Flows for those return period events were shown to overtop the vertical concrete channel wall along the south bank of the stream just upstream of the dam. The 50-year and 100-year return period event flows were shown by the model to overtop the north berm just upstream of the dam and overtop the vertical concrete channel wall along the south bank upstream of the dam. The south berm is not show to be overtopped under any flow conditions up through the 100-year return period event. Figures 3.17 through 3.22 provide overviews of the estimated extent of these overbank flow conditions for the 2-year through 100-year return period events.

North Berm

South Berm

Blue arrows indicate the approximate extent along the south channel bank where flow will overtop the concrete channel wall. Modeling indicates that neither the north berm nor the south berm would be overtopped.

Figure 3.17 Estimated Locations of Overbank Flow for 2-year Return Period Event

North Berm

South Berm

Blue arrows indicate the approximate extent along the south channel bank where flow will overtop the concrete channel wall. Modeling indicates that neither the north berm nor the south berm would be overtopped.

Figure 3.18 Estimated Locations of Overbank Flow for 5-year Return Period Event

North Berm

South Berm

Blue arrows indicate the approximate extent along the south channel bank where flow will overtop the concrete channel wall. Modeling indicates that neither the north berm nor the south berm would be overtopped.

Figure 3.19 Estimated Locations of Overbank Flow for 10-year Return Period Event

North Berm

South Berm

Blue arrows indicate the approximate extent along the south channel bank where flow will overtop the concrete channel wall. Modeling indicates that neither the north berm nor the south berm would be overtopped.

Figure 3.20 Estimated Locations of Overbank Flow for 20-year Return Period Event

North Berm

South Berm

Blue arrows indicate the approximate extent along the south channel bank where flow will overtop the concrete channel wall, and the approximate extent along the north berm where flow will overtop that berm. The south berm is not overtopped.

Figure 3.21 Estimated Locations of Overbank Flow for 50-year Return Period Event

North Berm

South Berm

Blue arrows indicate the approximate extent along the south channel bank where flow will overtop the concrete channel wall, and the approximate extent along the north berm where flow will overtop that berm. The south berm is not overtopped.

Figure 3.22 Estimated Locations of Overbank Flow for 100-year Return Period Event

DRAFT – 30 June 2008 36

Flow Conditions Downstream of Dam

The scope of this study was limited to the potential flood conditions that may occur as a result of the presence of the Camp Casey Dam. The dam itself will not cause water levels downstream of the dam to be any higher or more severe during flood conditions than the water levels would be without the dam in place. To some degree, the dam may actually act to reduce water levels in the stream channel downstream of the dam. This would be due in part to the fact that the dam may serve to attenuate the flood peak as it passes, thus reducing the peak flow downstream of the dam. Another factor contributing to this under higher flow conditions is that some water upstream of the dam may exceed the banks of the stream channel and continue flowing outside of the stream, thus the total amount of water flowing in the stream channel downstream of the dam is slightly reduced by this loss of the overbank flow upstream. Thus, under normal flow conditions with the dam intact, the dam should not adversely affect flood conditions downstream of the dam. In the event of a dam failure during a high flow event, there may be some adverse impact on downstream flood conditions. This scenario is discussed in greater detail in Chapter 5.

Note that flooding along Casey Creek downstream of the dam may occur under higher flow conditions, but this flooding would not be associated with the Camp Casey dam. Thus, detailed analysis of this section of Casey Creek was beyond the scope of this study. Flow in this lower part of Casey Creek as it leaves Camp Casey to the west is potentially influenced to a large degree by downstream conditions, notably the level of the Sinchon River. Modeling the conditions of the Sinchon River was beyond the scope of this study. Also of note is the grate that serves as a security barrier at the western edge of Camp Casey in the stream. If debris is caught in that structure such that downstream flow is impeded to any degree as it leaves Camp Casey, the water levels and potential for flooding along the golf course are certainly increased.

Additional data collection and further study of the lower reach of Casey Creek, its interaction with the Sinchon River, and the impacts of the various structures in the stream channel would be needed to provide a full analysis of potential flood conditions in this area. The 1984 report does provide delineations of the 100-year and 25-year flood extents downstream of the current dam location and those results may or may not still be relevant if conditions have changed significantly since the report was produced. The flood extents downstream of the dam in that report generally start at the current dam site and extend across roughly the western and southern 2/3 of the golf course area that lies north of Casey Creek.

Limitations of HEC-RAS Modeling

The hydraulic analysis that was conducted for this study using the HEC-RAS model was intentionally limited to the stream channel and some portions of the immediate overbank along the stream, such as the area between the

DRAFT – 30 June 2008 37 channel and the south berm just upstream of the dam. The HEC-RAS model is a one-dimensional hydraulic model, i.e., the computations are conducted for flow in the direction parallel to the main flow along the centerline of the stream channel. Once overbank flow occurs and flow overtops a berm or levee, HEC-RAS could still be an appropriate model for those flow conditions provided that the overbank flow remained generally within an area along the channel such that the overall direction of flow was more or less parallel to the centerline of the channel. However, if overbank flow occurs and the resultant flow outside of the channel continues on a path that is essentially disconnected from the channel hydraulically and if that flow continues in directions that are much different than the general direction of flow along the main channel, then HEC-RAS is not an appropriate model to depict that flow.

To properly model these flow conditions, a two-dimensional model is required that is capable of computing the overland flow as the flow follows the terrain beyond the stream channel in whatever direction that flow path leads. For Camp Casey, once flow escapes beyond the confines of the channel, it may continue through areas of Camp Casey away from the stream in a direction generally downstream, it may flow back in an upstream direction in some areas, it may flow perpendicular to the stream, or any combination thereof depending on where the water leaves the stream and what the local terrain dictates. Thus, modeling the direction of flow and overall extent of flooding that may result from the overbank flow conditions was not conducted with HEC-RAS.