ABSTRACT Witt, Anne Carter.

Using a GIS (Geographic Information System) to model slope instability and debris flow hazards in the French Broad River watershed, North Carolina. (Under the direction of Dr. Michael M. Kimberley.) Catastrophic, storm-generated mass wasting is a destructive erosional process in the portion of the southern Appalachians that extends through western North Carolina. Steep slopes, a thin soil mantle, and extreme precipitation events all increase the risk of slope instability, slope movement and failure. Since the late 1800s, several intense storms and hurricanes have tracked through the French Broad watershed initiating thousands of debris flows and causing severe flooding. Studying the history of debris flows has identified triggering mechanisms that are particular to North Carolina and the recurrence interval of these events. This study was initiated to investigate and predict the spatial distribution of regional slope instability within the French Broad watershed by comparing the results of two GIS-based modeling applications: SINMAP (Stability Index Mapping) and SHALSTAB (Shallow Landsliding Stability Model). As extensions to ArcView 3.x, SINMAP and SHALSTAB use a modified form of the infinite slope equation to compute and map slope-instability by calculating either a factor of safety (SINMAP) or the critical steady-state rainfall intensity necessary to trigger slope instability (SHALSTAB). In both models, topographic slope is derived from digital-elevation data while parameters for soil and climate are considered more variable and can be adjusted to better match existing conditions. An inventory of actual debris flow locations, collected from aerial photography, field reconnaissance, and a literature review of historic mass wasting events, are used to verify model results. SINMAP model runs have been completed for the watershed using a 30-meter and 10-meter digital elevation model (DEM) and 142 landslide point locations. Results

using the programs default parameters were compared with those for four recharge events (50, 125, 250, and 375 mm/d). In the latter, parameters for soil density, cohesion, internal soil friction angle, and transmissivity were adjusted to better match existing watershed conditions. As with the SINMAP model, SHALSTAB was used to model instability using both a 10-meter and 30-meter DEM. Limitations in the SHALSTAB program only allow smaller (county-size) DEMs to be processed. Because of these limitations Haywood County was chosen for several model runs in SHALSTAB for comparison to the SINMAP results. Parameters for soil density, soil depth, cohesion, and soil friction angle were adjusted and results were compared to 23 mapped debris flow locations. The modeled results for the default SINMAP and SHALSTAB parameter values underestimate the extent of instability in the study area. By adjusting soil parameters, SINMAP calculated 88% -to- 94% of the inventoried landslides would fall into the lower threshold, upper threshold, and defended stability classes. Generally, predicted areas of unstable land did not change, even as recharge increased. SHALSTAB calculated 91% to 100% of the mapped debris flows would occur in the three most unstable stability classes for low values of soil friction angle (26) and cohesion (0). Overall, SHALSTAB seems to over-predict areas of instability for these values. An increase in cohesion or soil friction angle decreases the amount of land predicted as unstable (88%) but increases the landslide density. When the results of the two programs are compared directly, the results are very similar. Visually the SHALSTAB results seem to cluster while the SINMAP results are more dispersed. In the field, it was noted that bedrock foliation and fracturing concentrate

groundwater flow in the watershed and failure tends to occur along these planes of weakness. Neither model takes into account either antecedent moisture or the effect that geologic structure can have on concentrating groundwater flow.

USING A GIS (GEOGRAPHIC INFORMATION SYSTEM) TO MODEL SLOPE INSTABILITY AND DEBRIS FLOW HAZARDS IN THE FRENCH BROAD RIVER WATERSHED, NORTH CAROLINA

by ANNE CARTER WITT

A thesis submitted to the Graduate Faculty of North Carolina State University in partial fulfillment of the requirements for the Degree of Master of Science MARINE, EARTH AND ATMOSPHERIC SCIENCES Raleigh 2005

APPROVED BY: Chair of Advisory Committee Committee Member Committee Member Committee Member ______________________________________ Dr. Michael M. Kimberley ______________________________________ Dr. Jeffery C. Reid ______________________________________ Dr. Elana L. Leithold ______________________________________ Dr. Helena Mitasova

BIOGRAPHY Anne Carter Witt was born in Lynchburg, VA on August 10, 1977. She grew up in Forest, VA and graduated from Brookville High School in 1995. In 1999, she graduated from Mary Washington College in Fredericksburg, VA with an undergraduate degree in Geology. Before being accepted into the Masters degree program at North Carolina State University, she worked for three years as a GIS analyst with Dewberry and Davis, LLC in Fairfax, VA.

ii

ACKNOWLEDGEMENTS I would like to express sincere thanks and appreciation to the following individuals: Rick Wooten with the North Carolina Geological Survey who provided me with landslide data and an open discussion on the complexities of SINMAP. His guidance, patience and knowledge of slope movements and the geomorphology of western North Carolina are greatly appreciated. Thanks are also extended to Jody Kuhne from the North Carolina Department of Transportation whose tattered, long-forgotten landslide map was really the foundation for a digital landslide inventory in North Carolina. In addition, I would like to thank Drs. Helena Mitasova and Lonnie Liethold for their advice and counsel is serving as members on my committee. I would also like to thank Dr. Mary Schweitzer for being a last minute substitution during my thesis defense on one of the busiest days of her life. Many thanks are also extended to Dr. Jeff Reid for his time, guidance, critical comments, and dogged determination to see me finish. Thanks are also extended to Dr. Michael Kimberley for his editorial expertise, assistance, and support. He generously donated both a computer and his own office space to help with my research. I would also like to express my sincere gratitude to all of my family and friends. Thanks go out to my fellow graduate students for their sometimes backwards but wellmeaning encouragement and enlightened conversations. I also would like to acknowledge the 2003 UNC-System Wide field camp students and staff who ultimately changed the direction of my life, whether they knew it or not. I also deeply appreciate the kindness, well wishes, and the open ear of my fellow Mary Washington College graduates and my

iii

best friend, Jamie Gibson. Finally, I would like to thank my parents and brother who provided the foundation for me to be the best that I could be and instilled a lifelong love of learning. Thank you for your love and support.

iv

TABLE OF CONTENTS Page LIST OF TABLES .......................................................................................................... vii LIST OF FIGURES ....................................................................................................... viii CHAPTER 1: INTRODUCTION.................................................................................... 1 1.1 RESEARCH AIM AND OBJECTIVES ........................................................................ 3 1.2 SIGNIFICANCE AND SCIENTIFIC APPLICATIONS .................................................... 4 1.3 LANDSLIDE CLASSIFICATION ............................................................................... 6 1.3.1 Classification of Sharpe (1938) ...................................................................... 7 1.3.2 Classification of Varnes (1978) and of Cruden and Varnes (1996) ............... 8 1.4 COMMON CAUSES OF SLOPE MOVEMENTS .......................................................... 9 1.4.1 Precipitation ................................................................................................. 10 1.4.2 Human Interference ...................................................................................... 10 1.4.3 Tectonic Activity............................................................................................ 11 1.4.4 Geologic Material and Structures ................................................................ 11 1.5 DEBRIS FLOWS ................................................................................................... 12 1.6 DEBRIS FLOWS WITHIN THE BLUE RIDGE AND WESTERN NORTH CAROLINA ..... 14 1.7 SINMAP AND SHALSTAB .............................................................................. 16 1.8 LIMITATIONS OF RESEARCH ............................................................................... 18 CHAPTER 2: PROJECT SETTING ............................................................................ 21 2.1 2.2 2.3 2.4 2.5 INTRODUCTION .................................................................................................. 21 GEOLOGY ........................................................................................................... 21 SOILS ................................................................................................................. 23 CLIMATE ............................................................................................................ 25 VEGETATION ...................................................................................................... 26

CHAPTER 3: PRE-HISTORIC AND HISTORIC DEBRIS FLOWS IN WESTERN NORTH CAROLINA ..................................................................................................... 33 3.1 QUATERNARY DEBRIS FLOWS ........................................................................... 33 3.2 MODERN FLOODING AND DEBRIS FLOWS .......................................................... 37 3.2.1 June, 1876 ..................................................................................................... 38 3.2.2 May, 1901 ..................................................................................................... 39 3.2.3 July, 1916...................................................................................................... 40 3.2.4 August, 1940 ................................................................................................. 42 3.2.5 November, 1977 ............................................................................................ 44 3.2.6 September, 2004............................................................................................ 46 3.2.7 Precipitation Thresholds and Recurrence Intervals ..................................... 51 CHAPTER 4: METHODOLOGY................................................................................. 64 4.1 INFINITE-SLOPE EQUATION ................................................................................ 64 4.2 PARAMETERIZATION OF THE MODELS ................................................................ 65 4.2.1 Digital Elevation Models .............................................................................. 66 v

4.2.2 Debris Flow Inventory .................................................................................. 68 4.2.3 Soil Data ....................................................................................................... 68 4.2.4 Soil Density ................................................................................................... 69 4.3 SINMAP PARAMETERS ..................................................................................... 69 4.3.1 T/R (Ratio of Transmissivity to Effective Recharge)..................................... 71 4.3.2 Dimensionless Cohesion ............................................................................... 72 4.3.3 Internal Soil Friction Angle .......................................................................... 72 4.4 SHALSTAB PARAMETERS ............................................................................... 73 CHAPTER 5: RESULTS AND DISCUSSION ............................................................ 80 5.1 INTRODUCTION .................................................................................................. 80 5.2 DEBRIS FLOW INVENTORY ................................................................................. 80 5.3 SINMAP RESULTS ............................................................................................ 81 5.3.1 Results 30-meter DEM............................................................................... 83 5.3.2 Results 10-meter DEM............................................................................... 84 5.3.3 SINMAP Interpretation................................................................................. 85 5.4 SHALSTAB RESULTS....................................................................................... 87 5.4.1 Results 30-meter DEM............................................................................... 89 5.4.2 Results 10-meter DEM............................................................................... 90 5.4.3 SHALSTAB Interpretation ............................................................................ 92 5.5 SINMAP VS. SHALSTAB................................................................................ 93 5.6 GEOLOGY AND SOILS ......................................................................................... 94 5.7 JOINTING, FRACTURING AND FOLIATION ........................................................... 97 CHAPTER 6: CONCLUSIONS .................................................................................. 116 REFERENCES.............................................................................................................. 122 APPENDICES ............................................................................................................... 133 APPENDIX A: GEOLOGIC UNITS ................................................................................... 134 APPENDIX B: GENERAL SOIL DATA ............................................................................. 139 APPENDIX C: SLOPE MOVEMENT DATA FOR THE 2004 HURRICANES FRANCES AND IVAN ..................................................................................................................................... 148 APPENDIX D: DEBRIS FLOW INVENTORY ..................................................................... 152 APPENDIX E: SINMAP RESULTS................................................................................. 157 APPENDIX F: SHALSTAB RESULTS ........................................................................... 161

vi

LIST OF TABLES Page Table 2.1: Table of the average, median, minimum, and maximum precipitation totals (mm/month) from 1895 to 2001 for the mountains of North Carolina (NCDC Climate Data Online, 2003)................................................................31 Table 3.1: Prehistoric debris flow studies in the southern Blue Ridge and the age-dating techniques utilized. ..........................................................................................53 Table 3.2: Major storms within the French Broad Watershed and their minimum, average, and maximum precipitation amounts. ...............................................63 Table 4.1: Table of the hydraulic conductivity (K, m/hr), transmissivity (T, m/hr), and T/R (m) values used for each precipitation threshold (50 mm/d, 125 mm/d, 250 mm/d, and 375 mm/d) in the SINMAP analysis. The numbers in blue are the lower bound values while the numbers in red are the upper bound values. ...................................................................................................78 Table 5.1: The parameters used in all of the SINMAP model runs...................................99 Table 5.2: SINMAP stability index definitions (Pack et al., 1998b). ..............................100 Table 5.3: Mapped instability classes used in the SHALSTAB model analysis. ............104 Table 5.4: Table comparing q/T and log (q/T) values and the precipitation rate required to initiate instability for soils with a transmissivity of 65 md and 17 m/d (after Dietrich and Asua, 1998). .......................................................105 Table 5.5: Parameters used in the SHALSTAB model runs............................................105 Table 6.1: The advantages and disadvantages of SINMAP and SHALSTAB. ...............121

vii

LIST OF FIGURES Page Figure 1.1: The morphology of a typical debris flow found in the Southern Appalachians (courtesy of the North Carolina Geological Survey, 2003). ............................19 Figure 1.2: Threshold precipitation values necessary for producing debris flows in the southern Appalachian Mountains. Storms likely to start debris flows occur above the 125 mm/d threshold. Storms with precipitation values higher than 250 mm/d are deemed rare but do occur in North Carolina (after Eschner and Patric, 1982). .............................................................................................20 Figure 2.1: Location Map of the French Broad Watershed in western North Carolina. ...27 Figure 2.2: General geologic map for the French Broad Watershed. Individual geologic unit descriptions can be found in Appendix A (adapted from North Carolina Geological Survey, 1985). ...............................................................................28 Figure 2.3: General soil map for the French Broad Watershed. Individual soil descriptions can be found in Appendix B (adapted from U.S. Department of Agriculture, 1998). ...............................................................................................................29 Figure 2.4: The U.S. Department of Agriculture guide for the textural classification of soils. This guide is only for soils with a particle size of less than 2 mm in diameter. A rock fragment modifier (gravelly, cobbly, stony, bouldery) prefaces the textural name if particles larger than 2mm compose more than 15% of the soil (Buol et al., 2003). ..................................................................30 Figure 2.5: Graph based on the data from Table 2.1 (NCDC Climate Data Online, 2003). ..............................................................................................................31 Figure 2.6: Average annual precipitation in inches within the French Broad Watershed. (Adapted from data provided by North Carolina Center for Geographic Information and Analysis map server (http://204.211.135.111)). ...................32 Figure 3.1: Old slope movement deposit along a private drive in Maggie Valley, North Carolina (age not determined). There is significant soil development in this poorly sorted colluvial deposit. It is located along a chute where a modern debris flow occurred. .......................................................................................52 Figure 3.2: Areas of major debris flows and landslides in western North Carolina (after Scott, 1972). .....................................................................................................54 Figure 3.3: A sketch map of debris flows that occurred along Gouges Creek in Mitchell County, North Carolina in May, 1901 (Myers, 1902). ....................................55 Figure 3.4: Map showing some of the hurricane and storm paths that have affected western North Carolina as reported by the U.S. National Hurricane Center and the U.S. Geological Survey Water Resources Branch (1949). .....................56 Figure 3.5: Total storm precipitation for July 14-16, 1916................................................57 Figure 3.6: U. S. Geological Survey hydrograph from the river gauge located on the French Broad River in Asheville for the month of July, 1916. In early July, there is a large hydrographic spike generated by rainfall from the remnants of a hurricane that passed through the area. There is also a second large peak between July 16 and 17 generated by the flooding associated with a hurricane

viii

moving northwestward over the watershed. This is the greatest recorded streamflow for the gauge at Asheville. ............................................................58 Figure 3.7: Total storm precipitation for August 14-15, 1940 adapted from U. S. Geological Survey, 1949). ...............................................................................59 Figure 3.8: Total storm precipitation for August 28-31, 1940 (adapted from U. S. Geological Survey, 1949). ...............................................................................59 Figure 3.9: Total storm precipitation for November 2-5, 1977 (adapted from Neary and Swift, 1987)......................................................................................................60 Figure 3.10: Total storm precipitation (inches) for the remnants of Hurricane Frances (National Weather Service, 2004b)..................................................................61 Figure 3.11: Total storm precipitation (inches) for the remnants of Hurricane Ivan (National Weather Service, 2004c)..................................................................61 Figure 3.12: A debris flow that blocked the westbound lanes of Interstate-40 near Old Fort Mountain in McDowell County (North Carolina Geological Survey, 2004a). .............................................................................................................62 Figure 3.13: The initiation zone of the debris flow that occurred on Fishhawk Mountain and devastated the Peeks Creek area of Macon County on September 17, 2004 (Wilett, 2004)...................................................................................................62 Figure 4.1: The infinite slope equation as defined by Hammond et al., (1992) and Pack et al., (1998b) where Cr is root cohesion, Cs is soil cohesion, is slope angle, s is soil density, w is the density of water, g is acceleration due to gravity, D is the vertical soil depth, Dw is the vertical height of the water table, and is the internal soil friction angle. In the SINMAP model, the ratio of the vertical soil depth to the vertical soil height is simplified so that depth is measured perpendicular to the slope (h). (Diagram after Hammond et al., 1992 and Otteman, 2001) ................................................................................................76 Figure 4.2: Default parameters used in the SINMAP model analysis. The values for the gravitational constant and the density of water were not adjusted in this study.................................................................................................................77 Figure 4.3: Default values used for SHALSTAB. .............................................................79 Figure 5.1: Landslide inventory map for the French Broad Watershed. The cluster of location points in southern Buncombe County is due to the extensive mapping of 1977 debris flows done by Pomeroy (1991) and Otteman (2001). .............98 Figure 5.2: A comparison of 30-meter and 10-meter DEMs in Haywood County, North Carolina. In the coarser 30-meter DEM, there is a great deal of pixelation. Lower resolution can lead to an underestimation of both the slope and instability in the study area. .............................................................................99 Figure 5.3: SINMAP results for a 30-meter DEM and using default parameters............101 Figure 5.4: SINMAP results for 125 mm/d recharge and using a 30-meter DEM. .........102 Figure 5.5: SINMAP results for a 10-meter DEM and using default parameters............103 Figure 5.6: SINMAP results for 125 mm/d recharge using a 10-meter DEM.................104 Figure 5.7: The map of log (q/T) for Haywood County produced using default parameters in SHALSTAB...............................................................................................106 Figure 5.8 Cumulative percent of Haywood County found in each log (q/T) category for a variety of soil parameters for the 30-meter DEM (after Dietrich et al., 2001).

ix

In the legend, the first number is the degree if soil friction angle and the second number is the amount of cohesion (N/m). ........................................107 Figure 5.9: Cumulative percent of debris flows for each log (q/T) instability category for a variety of soil parameters for a 30-meter DEM (after Dietrich et al., 2001). .............................................................................................................107 Figure 5.10: Cumulative percent of the area if Haywood County for each log (q/T) instability category for a variety of soil parameters for the 10-meter DEM (after Dietrich et al., 2001).............................................................................108 Figure 5.11: Cumulative percent of debris flows for each log (q/T) instability category for a variety of soil parameters for a 10-meter DEM (after Dietrich et al., 2001). .............................................................................................................108 Figure 5.12: Mapped log (q/T) results for a 10-meter DEM of Haywood County. The parameters used for this model run are essentially that same as those used in the SINMAP lower bounds: 26 soil friction angle, and zero cohesion. .......109 Figure 5.13: SHALSTAB results for a soil friction angle of 35 and cohesion of 2000 N/m. ..............................................................................................................110 Figure 5.14: Comparison of the output of SINMAP and SHALSTAB for 125 mm of recharge, 35 soil friction angle, 1922 kg/m soil density, and zero soil cohesion for a location in Haywood County. The red areas are calculated as unstable by both programs whereas the grey areas are calculated to be stable. Visually the SHALSTAB results seem to cluster better while the SINMAP results are more scattered. Overall, the results are very similar. ..................111 Figure 5.15: Results for SINMAP and SHALSTAB. Red areas are the areas calculated by both models as being unstable. The purple areas were only calculated by SHALSTAB and the green areas were only calculated by SINMAP............112 Figure 5.16: The mean stability index for each geologic unit in the French Broad Watershed. The most unstable unit, Zchs, is located in the northwestern portion of Haywood County. The most stable units, bz and Ctzp, are located in the southeastern portion of the study area......................................................113 Figure 5.17: The mean stability index for each soil unit in the French Broad Watershed. The most unstable soil unit in the watershed is NC104, which is located in western Haywood County. The most stable soils are located around the floodplains of the Pigeon and French Broad Rivers. .....................................114 Figure 5.18: Picture taken in October, 2003 along the Blue Ridge Parkway, near Mt. Mitchell. Even during this light rain event, water is pouring out from fractures in the rock. .....................................................................................................115

CHAPTER 1: INTRODUCTION Catastrophic, storm-generated mass wasting is a destructive erosional process in the portion of the Southern Appalachians that extends through western North Carolina. This study was initiated to investigate and predict the spatial distribution of regional slope instability within the French Broad watershed by comparing the results of two GISbased modeling applications: SINMAP (Stability Index Mapping) (Pack et al., 1998b) and SHALSTAB (Shallow Landsliding Stability Model); (Dietrich and Montgomery, 1998). Mass wasting is a general term used to describe the dislodgement and downslope transport of soil and rock material under the direct application of gravitational body stresses (Jackson, 1997). Mass wasting can range from very slow creep to a rock avalanche. The term is often synonymous with the term mass movement (Crozier, 1986). Both Crozier (1986) and Varnes (1978) advocate the use of the term slope movements for mass movement restricted to slopes, as slope movement appears to be a suitably neutral, all-encompassing term. Throughout this paper the term slope movement will be used interchangeable with the terms mass wasting and landslide to describe the movement of a mass of rock, soil and debris downslope. Of the several types of slope movements that occur in the Appalachians, rapid mass movement, particularly debris flows, are considered the most dangerous and will be the focus of this study. In the Appalachian Mountains, steep slopes, a thin soil mantle, and extreme precipitation events all increase the risk of slope instability, slope movement and failure (Gryta and Bartholomew, 1983; Neary and Swift, 1987; Wieczorek, 1996). Since the

early 1900s, several intense storms and hurricanes have tracked through western North Carolina, initiating hundreds of debris flows and causing severe flooding. In the Appalachian Mountain chain, it has been estimated that as many as 1,700 debris flows occurred in the 20th century, killing at least 200 people and destroying thousands of acres of farm and forested land (Scott, 1972). As extensions to ArcView 3.x GIS software, both SINMAP and SHALSTAB compute and map areas of potential slope instability based upon digital-elevation data and observed landslide locations. These models combine steady-state hydrologic concepts and an infinite-slope-stability analysis with a digital elevation model (DEM) to calculate either a factor of safety (SINMAP) or the critical steady-state rainfall intensity necessary to trigger slope instability (SHALSTAB). As in any landslide investigation using modeling software, model results should be compared with mapped landslide locations whenever possible. In this study, preliminary field investigations were completed in three key locations in Haywood, Yancey, and Transylvania Counties, initially modeled as having a high probability for slope instability and failure. Aerial photography was used to search for debris-flow scars, in the development of a landslide database for the region, and to complete more-detailed SINMAP and SHALSTAB model runs for those locations. During field work in these areas, soil and geologic data was collected for each location; debris-flow locations were more precisely located with a hand-held global positioning system, and the overall accuracy of the modeled stability index was assessed by identifying actual debris-flow locations.

1.1

Research Aim and Objectives Few comprehensive research studies comparing slope-stability-modeling software

have been completed within a specific drainage basin in North Carolina. Most of these studies have tended to focus on identifying landslide hazard areas on the U.S. West Coast or in other countries (Appt et al., 2002; Dietrich et al., 2001; Montgomery et al., 2001; Zaitchick et al., 2003). Other studies have focused on identifying relict landforms and debris aprons in the Appalachians formed during periglacial conditions (Gryta and Bartholomew, 1987; Clark, 1987; Kochel, 1987; Jacobson et al., 1989; Liebens and Schaetzel, 1997). This study focuses on identifying debris-flow hazard areas and the factors that affect this instability. The objectives of this research are as follows: 1. To identify areas within the French Broad watershed prone to debris flows. 2. To identify triggering mechanisms particular to the watershed that promote instability and failure. 3. To determine which soil and geologic units are the most prone to instability and those that are the most stable, in general. 4. To compare the results, reliability, and effectiveness of two widely used, and theoretically similar, GIS-based programs that model shallow landslide potential in a watershed in North Carolina. 5. To study the effect of parameter variation on each of the modeling programs. As part of this study, preliminary debris-flow hazard maps have been produced, for comparison, using SINMAP and SHALSTAB. Results from both computer programs were compared to an inventory of recent debris-flow and landslide locations determined through a review of previous research and literature, aerial photography, and field verification.

1.2

Significance and Scientific Applications The French Broad river basin is located in western North Carolina, in the central

portion of the Appalachian Mountain chain. The French Broad River itself flows through the City of Asheville, a major commercial and manufacturing center, and a popular mountain resort area. According to data collected by the U.S. Census Bureau (2000), approximately 426,000 people live within the French Broad watershed and this population is predicted to increase, particularly in and around the City of Asheville. Debris flows hazards are a major concern in mountainous areas as debris fans are favored areas of development due to their flat building surface and location above the floodplain (Ritter et al., 1995). With continued development and tourism in the forested areas of the Blue Ridge, the risk to people and property will increase because of debris flows, especially during periods of high precipitation. The major hazard to human life and property from debris flows is from burial or impact by boulders and other debris. Usually starting on steep hillsides, debris flows can accelerate to speeds between 15-55 kph (10-35 mph) (Highland et al., 2004) and often strike without warning. Because of their relatively high density and viscosity, debris flows can move and even carry away vehicles, bridges and other large objects (Cruden and Varnes, 1996). The economic costs of slope movements are can be both event-related and preventative (Crozier, 1986). Event-related costs may involve the removal of debris from rivers, streams and roadways as well as damage to vehicles, buildings, and utility systems. In agricultural areas, there may be replacement costs necessary for livestock, crop and timberland, and fence repair (Crozier, 1986). Preventative costs involved in risk assessment are more difficult to assess but may involve research in the form of surveying,

mapping, and laboratory testing. After an initial analysis has been completed, policies for the establishment, management, and inspection of preventative measures should be completed (Crozier, 1986). This may also include the restriction of development in areas considered landslide-prone or the removal or conversion of existing developments. Control measures in prone areas may also be designed and implemented including controlled drainage, planting, slope-geometry modification, and structures such as rock fences or other barriers (Schuster and Kockelman, 1996). Continued poor land-management practices and deforestation add to the risk of soil mass movement by increasing runoff, erosion, and flooding. Human activity also disturbs large volumes of geologic materials with the construction of housing developments, commercial buildings, mines and quarries, dams and reservoirs, and particularly the emplacement of transportation systems along steep slopes (Schuster, 1996). Roadcuts and other altered or excavated areas along slopes are particularly susceptible to debris flows. Repeated landslides and rock falls have plagued the Interstate 40 corridor as it winds through the mountains of Western North Carolina and the Pigeon River Gorge (Leith et al., 1965; Glass, 1981; Kaya, 1991; Wieczorek, 1996). By combining a theoretical model of slope instability with actual debris-flow locations, landslide hazard areas may be efficiently identified and mapped for regulatory purposes. Identifying areas already at risk for slope failure could prevent loss of property and lives. This would be a valuable resource to government agencies, as well as city and county planners, in their attempts to develop land-use and building ordinances that regulate construction. Both SINMAP and SHALSTAB are freeware, readily available for

download from the World Wide Web at no cost, making them economical for both government and businesses. Debris flows are not just an agent of destruction, but also play a critical role in the processes of erosion and sediment transport in the Blue Ridge Mountains. They provide a major means of removing weathered material from steep areas that normally experience little concentrated surface drainage (Scott, 1972). By studying the causes, characteristics, and effects of debris flows in the Southern Appalachians, we can better understand their erosional importance, as well as their destructive influences. Improved knowledge of the Blue Ridge may also be extended to comparable sub-tropical mountainous areas in other parts of the world. 1.3 Landslide Classification There have been several attempts to create a suitable landslide classification

system that is useful for both scientific and engineering purposes. Early attempts occurred in Europe during the late 1800s and early 1900s (Sharpe, 1938). These schemes tended to be incomplete; in that slow movements such as creep were ignored, they were developed for a single specific purpose or were developed based on only regional observations. The more successful classification systems are those that provide clear and unambiguous terminology that can be applied in a number of different situations (Crozier, 1986). Generally, these systems classify landslides on the basis of the kind and quality of material moved, its water content, and the type, size, and rate of movement. The classification system most likely to be encountered, in the Englishspeaking world, is that authored by D.J. Varnes (1978) from the U. S. Geological Survey

in Denver, CO. But this classification system, like many before it, was influenced by the classification system proposed by C.F.S. Sharpe (1938). 1.3.1 Classification of Sharpe (1938) In the late 1930s, C. F. S. Sharpe became one of the first American scientists to

create a widely accepted landslide classification system. Sharpe created his classification based primarily on the kind and rate of movement (making a distinction between slides and flows) and the forms of the resulting deposits. Other important factors included the moisture content in the moving mass and the type of material involved. Based on these factors, mass movement was separated into four groups, i.e. slow flowage, rapid flowage, sliding, and subsidence. Sharpe (1938) defined the mass-wasting process that typically occurs in the Southern Appalachians as debris-avalanche, a type of rapid flowage. A debris avalanche is a rapidly-moving, sliding flow with a long, narrow track that occurs on steep terrain in humid mountainous areas with significant vegetative cover. This type of slope movement tends to have higher water content than does a slow flow. They are usually preceded by a period of heavy precipitation which increases the weight of the unadjusted material and aids in its lubrication (Sharpe, 1938, p. 61). The material of a debris avalanche is variable and can consist of soil, rock, vegetation and ice. Initial failure typically occurs between the soil-rock interface or within loose debris on slopes between 20 and 40 degrees. When there is less water present in the same type of material, the movement would be termed a debris-slide.

1.3.2

Classification of Varnes (1978) and of Cruden and Varnes (1996) The landslide classification system of D. J. Varnes (1978) was well-received and

has been repeatedly updated, the latest update having been published in Cruden and Varnes (1996). The goal of the current version of this classification was to provide definitions and vocabulary that allow an investigator to observe and describe a landslide in the field succinctly and unambiguously. The Cruden and Varnes (1996) classification scheme emphasizes the type of material and the type of movement in a slide. Two terms are needed to describe any landslide, i.e. one that describes the material (rock, debris, earth) and one that describes the movement (fall, topple, slide, spread, flow). Other descriptors can then be added in front of the two-term classification as more information about the movement becomes available. These include the water content, rate of movement, the current activity of the slide (reactivated, inactive, etc.), distribution, and overall style (complex, composite, etc.). Like the classification scheme of Sharpe (1938), Cruden and Varnes (1996) uses the terms debris avalanche, debris slide, and debris flow to describe rapidly-flowing mass-wasting events. The terms debris avalanche and debris slide in this classification are varieties of the more general debris flow. Debris avalanche is an older term used to in the literature to describe extremely large, rapidly-moving mass movements that involve large amounts of ice, snow, and soil. A debris slide generally has less water content, is more coherent, and can occur on more gentle slopes than a flow. 1.4 Common Causes of Slope Movements The mechanics of slope stability can be divided into two basic forces i.e., the

gravitational driving force and frictional resistance of slope material (Easterbrook, 1999).

Driving forces cause material to move downslope and can increase depending on the mass of the material involved in the movement and the slope angle (Easterbrook, 1999). Frictional resistance opposes deformation or motion and is caused by the friction between grains within the material or by the material at the base of the slope. The resistance of a material to shear along a slip surface due to an applied external force can be referred to as the materials shear strength. Shear strength is determined by analyzing the amount of cohesion, effective normal stress, and internal friction between material particles (Ritter et al., 1995). When the gravitational driving force is greater than the frictional resistance the slope will fail (Easterbrook, 1999). While a single trigger may actually cause a slope to fail, a number of other destabilizing factors are often needed to contribute to a reduction in a materials shear strength in order to ultimately initiate movement (Crozier, 1986). First, though, the topography must have a sufficient slope. Usually slopes of 35 or greater are prone to instability simply because of the effects of gravity (Sidle et al., 1985). But often other factors, such as soil and root cohesion, increase the shear strength of materials on a slope. 1.4.1 Precipitation A number of slope movements are generated by an excessive amount of water, usually due to heavy precipitation or a moderate rainfall lasting several days. Slope movements in soil and weathered rock are usually produced on steep slopes during the most intense part of a storm (Wieczorek, 1996). Intense rainfall results in a higher groundwater table, saturation of the soil, a temporary increase in soil pore-water pressure, and a decrease in internal cohesion between soil particles (Easterbrook, 1999). These factors decrease the frictional resistance along a slope and may induce failure within the

material or along an existing plane of weakness. Thin, loose soils on hillslopes with sparse vegetation are particularly prone to failure during an intense rainfall. Soils with a high rate of hydraulic conductivity (the amount of water that will move through a porous medium in unit time) have the ability to transmit water more quickly downslope. This decreases soil pore pressure and may help to increase shear strength. Typically, an unconsolidated, coarse grained, well- rounded, and well-sorted soil will have a higher value of hydraulic conductivity (Fetter, 1994). 1.4.2 Human Interference Road construction is a major contributor to slope failure and their mitigation can often incur enormous public cost. Excavation of the toe of a hillslope by emplacing a road, quarry, canal, or other type of cut, removes support and may induce anthropogenic slope moment (Cruden and Varnes, 1996). Road fill and traffic also increases weight on a hillslope, increasing shear stress on materials (Sidle et al., 1985). In developed areas, slope saturation may occur, even during moderate recharge events, because of concentrated run-off from rerouting of drainage systems during road construction and from man-made structures such as drainpipes, buildings, and paved impervious surfaces. Vegetation often acts to stabilize slopes and increases shear strength and root cohesion. Foliage intercepts rainfall whereas roots and stems extract moisture from the soil (lowering pore-water pressure), increase surface roughness (Greenway, 1987), and provide an interlocking network that strengthens unconsolidated sediment (Easterbrook, 1999). The removal of vegetation by such activities as harvesting timber, decreases root cohesion and may decrease shear strength and initiate failure.

10

1.4.3

Tectonic Activity In tectonically-active areas, volcanic eruptions and earthquakes have often caused

slope instability and failure by causing uplift or tilting (Cruden and Varnes, 1996). Volcanic ash deposited on steep volcanic peaks, combined with the rapid melting of snow or heavy rainfall, can initiate deadly lahars, debris flows, and mudflows. Slope movements involving loose, saturated, low-cohesion soils commonly occur as a result of earthquake-induced liquefaction, a process in which shaking temporarily raises porewater pressure and reduces shear strength. Fault zones may also contain fractured, crushed, or low-metamorphic-grade rock that contain inherent weakness and may be susceptible to failure (Sidle et al., 1985). 1.4.4 Geologic Material and Structures Some particular rock and soil types may be inherently weak and can influence

slope instability. Organic soils and clays naturally have low shear strength and are particularly prone to weathering processes (Cruden and Varnes, 1996). A predominantly clay soil layer can prevent vertical infiltration of water and cause a buildup of pore-water pressure, while also providing a smooth, low-friction surface for failure. In contrast, sandy soils commonly erode quickly despite rapid infiltration and sandy slopes may be undermined by rivers and run-off. A permeable shallow soil overlying a hard, impermeable rock layer may also be susceptible to mass movement. As water builds up and travels along the rock-regolith interface, cohesion between the layers diminishes and provides a smooth, low-friction slip surface (Sidle et al., 1985). Weak rock types may occur where unfavorable geologic structures and neotectonics combine to induce mass movement (Sidle et al., 1985). Fault planes, joints, fractures, and other geologic structures also can act as a potential rupture surface by providing a conduit

11

for groundwater during rain events and reducing cohesion between layers or bedding (Sidle et al., 1985). Structures parallel to the ground surface and downslope-dipping beds, particularly those that separate two distinct lithologic units, may also act as a viable failure plane. 1.5 Debris Flows Although several types of slope movements have been described in the high-relief

portions of the French Broad watershed, this study focuses on debris flows, i.e., rapid downslope movement of regolith. It was noted during field work that a number of the regularly-occurring mass-wasting events were actually simple rock falls or slides along over-steepened roadcuts. Debris flows seem to be a less common mass-wasting event than those mentioned above, but may be devastating when they occur in populated areas. Unfortunately, there is a considerable amount of inconsistency in terminology for modern rapid channelized downslope movement of poorly sorted sediment. The terms, debris torrent, debris avalanche, debris flow, debris slide, mudflow, and mud flood are occasionally used interchangeably (Pierson and Costa, 1987). These terms all have similar descriptive characteristics for grain-size and the rate of downslope movement, but individually are often difficult to distinguish from one another in the field (Ritter et al., 1995). Other authors also use the terms, alluvial fan, debris fan, hillslope, toe-slope or foot-slope deposit to refer to relict fan-shaped features found throughout the Appalachians produced by repeated depositional episodes by debris flow processes (Kochel, 1990). For the purpose of this study, the general term debris-flow will be used to describe the slope movements in this study.

12

The term debris flow is used herein to describe swift-moving mass-wasting events that occur predominantly in shallow, silty-to-gravelly soil on steep slopes during periods of exceptionally heavy precipitation (Cruden and Varnes, 1996) (Figure 1.1). Debris defines a material that contains 20-80 percent coarse-grained particles larger than 2mm. These materials may include boulders to clay with varying amounts of water (Ritter et al., 1995). The flows begin in depressions or hollows on steep slopes and tend to move downslope following preexisting drainage channels. The most common movement interface is between the bedrock-soil contact, but slippage may also occur within deep soils (Clark, 1987). Debris flows from several different sources often converge into one main drainage channel, increasing the flows overall volume of water and material (Highland et al., 2004). Debris flows can travel for several kilometers before releasing their suspended load and coming to rest upon reaching an area of low gradient (Ritter et al., 1995). Debris flows may be triggered in a variety of ways. The most common trigger is an abundant amount of moisture, either from intense rainfall or rapid snowmelt, or a combination of heavy precipitation and antecedent soil moisture. Topography also influences debris-flow initiation by concentrating subsurface flow and determining slope. Soil thickness, conductivity, soil strength, bedrock-fracture flow, and root strength also influence the spatial distribution of debris flows and other types of shallow landslides (Montgomery and Dietrich, 1994). During a heavy rain event, piezometric head in the stratum increases while shear strength and cohesion decrease to the point where failure may occur (Neary et al., 1986). A debris flow typically begin as a debris and sediment-

13

laden slurry that gains material as it travels downslope or as a shallow-slope movement that is mobilized into a flow (Ritter et al., 1995). 1.6 Debris Flows within the Blue Ridge and western North Carolina Debris flows have been reported to be the most common form of rapid slope

movement in the Blue Ridge (Mills et al., 1987). Debris flows have occurred throughout the Blue Ridge province and have been specifically documented in Virginia (Woodruff, 1971; Williams and Guy, 1973; Gryta and Bartholomew, 1987; Gryta and Bartholomew, 1989; Mazza and Wieczorek, 1997), North Carolina (Holmes, 1917; Gryta and Bartholomew, 1983; Pomeroy, 1991; Wooten et al., 2001), West Virginia (Jacobson et al., 1989) and Tennessee (Clark et al., 1987). Mills et al. (1987) suggest that the abundance of this type of slope movement may be due to the amount of thick, unconsolidated colluvium derived from crystalline rock. This colluvium is highly permeable and susceptible to weathering, both of which contribute to the generation of slope movements. In western North Carolina, debris flows are activated primarily by either localized severe storms that produce intense rainfall for several hours or by more regional moderate storms that may last for several days (Wieczorek, 1996). Most debris-flowproducing storms can be linked to the incursion of warm, tropical air masses over the mountains between May and November, or the remnants of hurricanes and tropical storms (Kochel, 1990). The heavily forested slopes of the Appalachians are generally quite stable under normal rainfall and snowmelt conditions (Kochel, 1990). Certain thresholds of rainfall intensity and duration therefore must be reached before failures will occur (Figure 1.2).

14

Precipitation rates that readily induce debris flows in western North Carolina range from 125 mm/day (Neary and Swift, 1987) to the upper end of observed precipitation (560 mm/day). These thresholds may vary due to lithology, vegetation, and topography but generally catastrophic rainfall is required to initiate debris flows in heavily vegetated areas (Kochel, 1990). Under these conditions, rapid infiltration and a corresponding increase in soil saturation brings the soil mantle to field capacity. This tends to occur in shallow (< 1 m thick) mountain soils on slopes averaging 25-40 degrees, overlying an impermeable horizon of metamorphic rock or saprolite (Eschner and Patric, 1982). A temporary rise in piezometric pressure within slope sediment causes an increase in shear stress while decreasing shear strength. This, combined with a decrease in soil cohesion, reduces the shear resistance force enough to lessen the stability of the soil and eventually induce failure (Neary and Swift, 1987). Typically, debris flows have a characteristic long narrow shape. In the Southern Appalachians, the width of a debris flow, or the chute, may range from only a few meters to 60 m wide, although most tend to be narrow (Mills et al., 1987). Runout lengths rarely exceed 600 m, but a few have been known to extend as far downslope as 1600 m (Scott, 1972). Although single linear chutes are common, larger tracks are often produced by coalescing debris flows that form a dendritic pattern upslope (Scott, 1981). This shape is due to the formation of a majority of debris flows in topographic hollows (Scott, 1972). At their bases, debris accumulates as a washout fan of colluvium or may merge with alluvium and become washed away by floodwaters (Graham et al., 1990).

15

1.7

SINMAP and SHALSTAB There are many approaches to the problem of predicting shallow slope movements,

and almost as many predictive models. Simple models, based on identifying and classifying high-hazard areas on the basis of critical slope angle, do not take into account the effects of topographic form or position and lithology. More complex approaches to prediction should consider a wide range of variables such as drainage area, bedrock geology, soil thickness and cohesion, precipitation, vegetation and land use. Two such models, which take many of these variables into account, are SINMAP and SHALSTAB. SINMAP was designed as an extension to ArcView GIS, a product of Environmental Systems Research Institute, Inc. SINMAP is applied to shallow transitional landsliding phenomena controlled by shallow groundwater convergence (Pack et al., 2001), as is generally found in western North Carolina. The SINMAP methodology is based on the infinite-slope equation, an equation that has been found to be adequately accurate in the analysis of debris flows for planning purposes in the U.S. Western Cordillera (Hammond et al., 1992) as well as in western North Carolina (Otteman, 2001). A digital elevation model (DEM) is used to determine inputs of topographic slope and catchment area. Landslide point data is added to compare locations of predicted instability with areas of actual instability and to evaluate the accuracy of the model results. Other parameters for both natural material and climatological properties are assumed to be more uncertain and are specified in terms of upper and lower boundary values (Pack et al., 1998b). SINMAP uses these parameters to establish a stability index based on the probability that a location is stable. These values range from 0 (most unstable) to 10 (most stable). If the most destabilizing limits of the specified parameter

16

ranges are reached and yet stability is still retained, the stability index (defined as the factor of safety) is calculated as greater than one (Pack et al., 2001). The default output of a SINMAP session is a series of map grids that define areas of potential terrain instability; shaded green areas are considered stable whereas dark red areas have a high probability of failure, based on parameter inputs (Pack et al., 1998b). Like SINMAP, SHALSTAB is an extension of ArcView and is theoretically similar to SINMAP. The SHALSTAB model is also based on the infinite-slope equation and uses a DEM for the input of topographic grid information. SHALSTAB assumes steady-state, saturated flow parallel to the slide surface and uses Darcy's law to estimate the spatial distribution of pore pressures (Dietrich and Montgomery, 1998). The authors of SHALSTAB intended the model to be as simple as possible and nearly parameter free (Dietrich et al., 2001). In other words, most of the input variables are derived directly from the slope and area grids created from the input DEM and do not require the user to calculate specific parameters based on known soil properties which, without direct sampling and laboratory investigations, are often difficult to derive (Dietrich et al., 2001). Although SHALSTAB does allow for adjustments to certain soil parameters to better match existing conditions, the model does not provide upper and lower boundaries for parameters like SINMAP. Moreover, the model does not allow for an adjustment to recharge, but rather calculates the effective recharge (precipitation minus evapotranspiration) needed to induce failure for each cell within the DEM grid. Given that parameters for soil properties and precipitation are important to modeling shallow debris flows, the following authors reiterate that the SHALSTAB predictions

17

should be compared with mapped landslide features whenever possible (Dietrich and Montgomery, 1998). 1.8 Limitations of Research The prime limitation to both SINMAP and SHALSTAB is the availability of

high-quality digital-elevation data i.e. a 10-meter scale or smaller. Surface topography, determined from digital-elevation data, has great bearing on the location and frequency of shallow landsliding predicted by both models. Currently, only 30-meter and 10-meter DEMs are available for the western portion of North Carolina. Both SINMAP and SHALSTAB require the input of high-quality digital-elevation data to identify areas of steep slope. As more accurate DEMs and LIDAR data are released for this portion of the state, more accurate slope-instability and landslide-hazard maps can be developed. Limitations to SINMAP and SHALSTAB include the lack of high-quality soil data necessary to parameterize the models, the size of the study area, and the need to generalize soil and climate parameters over such a large area. Better quality results can be derived from larger-scale studies covering towns or individual neighborhoods. Such studies should collect and analyze soil samples as well as create a detailed inventory of all present and past debris flows for the study area. The authors of SINMAP and SHALSTAB did not intend for these models to provide a complete forecast of all types of landslide potential. These models were designed to predict only the potential for shallow landsliding, not deep-seated slides, rock fall, or any number of other slope movements found in the Southern Appalachians. Both models should be used in conjunction with geologic fieldwork as part of the number of tools available for a complete and effective landslide investigation.

18

Figure 1.1 - The morphology of a typical debris flow found in the Southern Appalachians (courtesy of the North Carolina Geological Survey, 2003).

19

Figure 1.2: Threshold precipitation values necessary for producing debris flows in the southern Appalachian Mountains. Storms likely to start debris flows occur above the 125 mm/d threshold. Storms with precipitation values higher than 250mm/d are deemed rare but do occur in North Carolina (after Eschner and Patric, 1982).

20

CHAPTER 2: PROJECT SETTING 2.1 Introduction The French Broad watershed is approximately 7330 km and includes the counties

of Buncombe, Haywood, Henderson, Madison, and Transylvania, and portions of Avery, Mitchell, and Yancey counties (Figure 2.1). Major tributaries of the French Broad River include the Nolichucky, Toe, and Pigeon Rivers. The watershed includes large portions of Great Smoky Mountains and Pisgah National Parks. Two major interstates, Interstate40 and Interstate-26, cross the basin, as does the Blue Ridge Parkway. The French Broad River actually begins in Rosman, North Carolina (35 miles southwest of Asheville, NC) where four tributaries converge. Topography within the French Broad basin ranges greatly, from the relatively gently sloping floodplains along the banks of the French Broad River to steep slopes in the mountains and along roadcuts. The highest point in the watershed, and the entire Appalachian mountain chain, is Mount Mitchell at an elevation of 2073 m. 2.2 Geology Given the large size of the French Broad River basin, 40 geologic units and 17

general soil types have been mapped within the watershed by North Carolina agencies. The watershed lies within both the Blue Ridge Belt and, to the east, a small portion of the Inner Piedmont Belt. Bedrock generally consists of sedimentary, metasedimentary, and intrusive igneous rock of Proterozoic and Paleozoic age (North Carolina Geological Survey, 1985) (Figure 2.2). More specifically, sedimentary rocks include sandstone and siltstone; metamorphic rocks range from biotite gneiss to mica schist and amphibolite, and igneous rocks include granite and dunite (Kaya, 1991; see Appendix A for details). Strike is generally towards the northeast with a dip to the southeast.

21

The Blue Ridge geologic province reaches its greatest east-west extent in the Carolinas, Tennessee and Georgia. The province is bounded on the northwest by the Blue Ridge fault systems (Holston-Iron Mountain, Great Smoky, and Cartersville Faults) and on the southeast by the Brevard fault zone, while the Hayesville and Greenbrier fault zones intersect in the middle of the French Broad basin (Hatcher and Goldberg, 1994). These faults transported a series of large crystalline thrust sheets over the Paleozoic rocks of the Valley and Ridge province, each with different tectonic histories and degrees of metamorphism (Hatcher, 1987). The western Blue Ridge is composed of a rift-facies sequence of clastic sedimentary rocks deposited on basement rock. The eastern Blue Ridge records a series of slope-and-rise sequences associated with rifting and continental collision (Hatcher and Goldberg, 1994). The Brevard Fault zone separates the Blue Ridge province from the Inner Piedmont block to the east. The block is a composite stack of thrust sheets containing a variety of metamorphic rocks, intrusive granitoids, and sparse ultramafic bodies (Horton and McConnell, 1994). Within the French Broad Watershed, only a few units from the Inner Piedmont block are represented i.e., the Henderson gneiss and middle-to-late Paleozoic granite gneiss. The Brevard fault zone itself represents a linear, southeast-dipping belt of cataclastic and mylonitic rock, 1-to-2 km wide (Horton and McConnell, 1994). According to Scott (1972), geologic structure and bedrock orientation play a more important role in slope stability than rock type in the Southern Appalachians. When soils are formed on weathered bedrock surfaces that are nearly coincident with the dip surface, sliding is more likely to occur between the soil-rock interface. Control on groundwater flow by joints and other fractures also can create areas of slope instability. This is

22

particularly true when fracture surfaces are parallel to the dip surface. It was observed in the study area that even during a light precipitation event, groundwater flow through fracture zones was swift. This concentration of groundwater could quickly cause an increase in pore-water pressure in soils on a slope or create ephemeral channels for debris flows to follow. A similar correlation between joint orientation, direction of groundwater flow, and debris-flow initiation was noted in the Coweeta Basin, an experimental forest and research station just south of the watershed (Grant, 1987). In the SINMAP and SHALSTAB models, groundwater flow is incorporated by using a value for soil transmissivity (m/s). 2.3 Soils The types of soil in the French Broad watershed reflect the regional geology

because variation in bedrock mineralogy partly controls soil mineralogy (Figure 2.3). Herein soil will be defined as unconsolidated mineral or organic material on the immediate surface of the earth that has been subjected to and shows effect of genetic and environmental factors (Jackson, 1997). Steep relief, broad ridges, and humid temperatures allow for a wide range of soil-forming conditions. On steep side-slopes, Inceptisols are common whereas Ultisols are found on gently sloping areas (Graham and Buol, 1990). Inceptisols, particularly Dystrocrepts, are soils that have not developed many diagnostic features due to rapid erosion rates and downslope movement. Ultisols develop from acidic parent materials, such as granite, and have an increase in clay content with depth (Buol et al., 2003). Soil textures range from fine clay and silt to sandy- and gravelly-loam (U.S. Department of Agriculture, 1998) (Figure 2.4). The dominant soil temperature regime is mesic, but frigid soils are found on top of Mt.

23

Mitchell and at very high elevations. Mesic soils have mean soil temperatures of 8 to 15 C (46 to 59 F) during June, July, and August, whereas frigid soils have mean temperatures less than 8 C (46 F) during the summer months (Buol et al., 2003). Generally, soils with a high susceptibility of failure tend to have a large mica content and develop over micaceous schist, slate, and phyllite (Scott, 1972). Soil cover varies in thickness and development depending upon slope and weathering and can range from less to one meter to several meters in depth (Clark, 1987). Steep slopes create a shallow soil veneer held together by plant roots that may overlie a rock or saprolite layer. More moderate slopes, i.e., those between 30 and 35 degrees, develop thicker soil profiles and are more prone than shallow soils to debris flows. This may seem contradictory, but steep slopes are not able to develop and retain soil cover and can easily be washed away, even during moderate rainfalls. Thus, an increase in soil depth and a decrease in slope increase the risk of slope instability (Scott, 1972). In the unglaciated portions of the Blue Ridge, chemical weathering plays an important role in the breakdown of rock into soil regolith (Grant, 1987). According to Graham et al. (1990), three types of regolith are found in the Southern Appalachians i.e., saprolite, colluvium, and soil residuum. Thick sequences of saprolite, or rotten rock, develop on gentle slopes composed of metasedimentary and metaigneous rock, promoting stability and chemical weathering (Jacobson, et. al., 1989). Typically, saprolite is covered by soil residuum. Residuum is separated from saprolite where the visual evidence of the rock structure is lost and soil begins. On steep slopes, either soil residuum or saprolite is overlain by colluvium (Grant, 1987). Thick, unsorted, coarse-grained

24

colluvium and alluvium are deposited in valley bottoms and along rivers (Hatcher, 1987; Otteman, 2001). 2.4 Climate Due to the variation of altitude (460-to-2073 m) within the French Broad

watershed, temperature and moisture regimes vary greatly from one place to another within this area. In fact, the mountains have some of the wettest and driest weather in North Carolina (Daniels et al., 1999). The greatest 24-hour rainfall total in the State (565 mm) was measured in the watershed at Altapass in Mitchell County on July 15-16, 1916 when a hurricane passed through the area. In contrast, the station with the driest weather on average is located in downtown Asheville in Buncombe County (State Climate Office of North Carolina, 2003). Mean annual rainfall in the southern Appalachians ranges from 1000 to 2700 mm (1-to-2.7 m) with snowfall only contributing 5 percent of the total precipitation (Neary and Swift, 1987). Rainfall occurs frequently as small, low-intensity rains in all seasons but localized heavy precipitation is uncommon. Precipitation is usually greater during the winter and spring, with March being the wettest average month (Table 2.1 and Figure 2.5). The highest maximum precipitation amounts have been recorded in the summer months when localized, high-intensity thunderstorms and hurricanes are more common. As a result, a majority of debris-forming rainfall events in the Blue Ridge occur in June, July, and August (Clark, 1987). Recently, in September 2004, several debris flows were also generated in the study area. No debris flows have been reported in the months of December, January or February.

25

Orographic influences generate extremely heavy rainfall in localized mountainous areas, even in storms with weak pressure gradients and gentle air circulation (Scott, 1972). Generally, rainfall increases with elevation at a rate of 5 percent per 100 m (Swift et al., 1988), but altitude is not as important as orographic boundaries. The Blue Ridge produces an elongate area of high values of mean precipitation (Jacobson et. al., 1989). As can be seen in Fig. 2.6, there is a marked increase in the amount of annual precipitation in northern Transylvania County and southern Haywood County. 2.5 Vegetation Like rainfall, vegetation within the watershed varies with the topography. Slope

aspect and shading by adjacent higher mountains also influences the distribution of major tree species (Daniels et al., 1999). At lower elevations (below 1400 m) hardwoods, oak, hemlock and pine forests dominate. Hardwoods such as yellow poplar, ash, and black cherry are found in coves and along steep slopes whereas several varieties of pine and oak thrive in open areas (Scott, 1972). Except for the most rugged terrain, the regions forestland has been cut or burned at least once since European settlement (Clark, 1987). In the very high mountainous areas of the watershed (above 1400 m) distinctive ecological systems have been established as a result of the cool year-round temperatures. Areas are often wind-swept and trees are damaged by ice and winter wind. Red spruce, mountain ash and Fraser fir are common with the latter dominating above 1890 m (Daniels et al., 1999). Grass balds and areas dominated by low shrub like rhododendron and laurel are common on southern-facing exposures (Daniels et al., 1999). These plants create extensive root systems or mats that increase soil and root cohesion, imparting stabilizing influences to the underlying soil.

26

Figure 2.1: Location Map of the French Broad Watershed in western North Carolina.

27

Figure 2.2: General geologic map for the French Broad Watershed. Individual geologic unit descriptions can be found in Appendix A (adapted from North Carolina Geological Survey, 1985).

28

Figure 2.3: General soil map for the French Broad Watershed. Individual soil descriptions can be found in Appendix B (adapted from U.S. Department of Agriculture, 1998).

29

Figure 2.4: The U.S. Department of Agriculture guide for the textural classification of soils. This guide is only for soils with a particle size of less than 2 mm in diameter. A rock fragment modifier (gravelly, cobbly, stony, bouldery) prefaces the textural name if particles larger than 2 mm compose more than 15% of the soil (Buol et al., 2003).

30

Table 2.1: Table of the average, median, minimum, and maximum precipitation totals (mm/month) from 1895 to 2001 for the mountains of North Carolina (NCDC Climate Data Online, 2003).

North Carolina Climate Division 01 - Southern Mountains Average Total Precipitation (mm/month) 1895-2001
JAN Average Median Minimum Maximum FEB MAR APR MAY JUN JUL AUG SEP OCT NOV DEC

TOTAL
1393.18 1312.93 244.35 3337.81

117.86 115.82 142.49 114.81 110.74 121.67 136.14 133.60 99.06 92.20 92.71 116.08 109.98 112.52 131.83 111.76 100.58 118.36 131.06 120.65 91.95 82.55 90.93 110.74 8.38 13.21 20.83 23.11 29.21 37.34 49.02 19.81 8.89 1.27 24.13 9.14 279.15 254.00 282.96 208.79 271.02 245.87 315.21 398.27 249.43 257.56 303.02 272.54

Monthly Precipitation for the NC Southern Mountains 1895-2001

400 350 Millimeters (mm) 300 250 200 150 100 50 0
JA N FE B M AR AP R M AY JU N SE P O C T N O V D EC L AU G JU

Average Median Minimum Maximum

Month
Figure 2.5: Graph based on the data from Table 2.1 (NCDC Climate Data Online, 2003)

31

Figure 2.6: Average annual precipitation in inches within the French Broad Watershed. (Adapted from data provided by North Carolina Center for Geographic Information and Analysis map server (http://204.211.135.111)).

32

CHAPTER 3: PRE-HISTORIC AND HISTORIC DEBRIS FLOWS IN WESTERN NORTH CAROLINA The Appalachian Mountains have a long history of producing destructive debris flows. Through the Pleistocene, temperature and moisture fluctuations associated with the transition from glacial to interglacial ages, destabilized exposed soil and rock. These pre-historic debris flows have formed prominent modern landforms and a rolling topography (Jacobson et. al., 1989). Records of flooding in western North Carolina associated with hurricanes and other strong storms exist back into the late 1700s. Recorded instances of debris flows and other slope movements during major rain events began in the late 1800s, but their generation and mechanics have been poorly understood until recently. This study is only a part of the ongoing natural hazards research being conducted by the North Carolina Geological Survey (Wooten et al., 2004) and North Carolina Department of Transportation. Continued study of the history of debris flows can help identify triggering mechanisms that are particular to North Carolina and the recurrence interval of these events. 3.1 Quaternary Debris Flows The Quaternary history of the Appalachian Mountains consists of slow uplift and

subsequent denudation. Although extensive terrain is mantled with colluvial deposits, individual ancient mass movements have only recently been identified in the Blue Ridge Province. Studies of these deposits have elucidated the rate of soil development and erosion, catastrophic debrisflow frequency and triggering events, and the possible role of periglacial processes in the Appalachians during the late Pleistocene (Table 3.1). Pre-historic debris-flow deposits form undulating, hummocky topography and elongated lobes or fans that are expressed as step-like landforms. Debris fans tend to be coarse-grained and

33

poorly sorted, but may be either matrix-or-clast-supported (Figure 3.1). Typically, fans are composites of several mass-wasting events with a weathered surface on each colluvial unit in the sequence. This indicates that there may be great differences in age between the units and upwards several thousand years have elapsed between debris-flow-forming events (Kochel, 1990). In the Great Smoky Mountains, characteristic recurrence intervals for major debris flows are on the order of 400 to 1600 years (Kochel, 1990) whereas catastrophic debris flows have been estimated to occur every 3000-6000 years in Nelson County, Virginia (Kochel, 1984; Kochel and Johnson, 1984). Catastrophic geomorphic events, such as debris flows, have been the principal means of erosion of the central Appalachian mountain chain during the Quaternary. The long-term denudation rates in the Appalachians average 40mm/kyr (Hack, 1980). In the short term, a single debris flow can remove enough material to account for 1 kyr of erosion during a single storm (Mills, et. al., 1987). Numerous studies (Table 3.1) hypothesize that extreme precipitation events, associated with major Quaternary climate change and periglacial environmental conditions have encouraged the formation of a number of debris flows in the Southern Blue Ridge during the Pleistocene (Liebans and Schaetzel, 1997). During the last 850 kyr there have been at least ten major ice advances that have brought glaciation to much of the northern Appalachians and periglacial conditions to the southern Appalachians (Braun, 1989). The term periglacial refers to the area near a glacier that is affected significantly by cold weather and has a landscape appreciably modified by the action of freeze-thaw cycles; it has been suggested that periglacial conditions extended as far south as the mountains of Georgia during glacial maxima (Jackson, 1997).

34

During glacial periods, North Carolina experienced a greater frequency of freeze-thaw cycles and physical weathering. Rock exposed at high elevations decomposed to a thin loose soil mantle (Mills, 2000). A dry polar climate dominated the region. In modern polar climates, monthly temperatures average below 10C (50F) year-round, resulting in little to no tree growth (Lydolph, 1985). Although a polar climate can create a ready supply of sediment through erosion and physical weathering, the lack of precipitation inhibits the formation of debris flows. In contrast, slow mass movements, such as solifluction and creep, are common (Ritter et al., 1995). In Virginia, slope wash of material may have proliferated more than debris flows during the Pleistocene (Eaton et al., 1997; Eaton, 1999). In the Great Smoky Mountains National Park, block field and slope deposits produced by Pleistocene frost wedging have been identified Clark and Torbett (1987). After the late Wisconsin glacial maximum, near the end of the Pleistocene, the northward migration of the polar front would have allowed tropical moisture to reenter the area during the summer months (Kochel, 1990). Previously undisturbed soil and rock then became exposed to heavy precipitation. Slopes that were still sparsely vegetated (due to cold winter temperatures) apparently became saturated and unstable, creating numerous large debris flows. Repeated intervals of glacial and interglacial climate on a periglacial landscape probably created episodic sequences of catastrophic mass-wasting during the Pleistocene and early Holocene (Kochel, 1990). Hillslopes remained generally unstable during the late Wisconsin glaciation, but transitioned to a period of less-frequent landsliding during the Holocene (Jacobson et al., 1989). Several large prehistoric debris flows have also been identified in the western Blue Ridge and the region of the Great Smoky Mountains by Hatcher et al. (1996). These debris flows originated at

35

high elevations (>1100 m), produced large volumes of colluvium (>106 m3) and may have transported material as far as 8 km in a single event (Hatcher et al., 1996). Giant Pleistocene rock and block slides also have been identified in both southwestern Virginia (Shultz, 1986) and the Great Smoky Mountains (Hadley and Goldsmith, 1963). These slides indicate that large sections (100-400 million m) of sandstone detached along bedding planes and joints and slid downslope (Schultz, 1986). If so, these debris flows and slides were significantly larger than those produced during modern times. Studies of pre-historic debris flows and fans tend to use three major dating techniques, i.e., relative-age dating of soil-horizon development and clast weathering, carbon-14 dating of organic material, and thermoluminescence dating of colluvium and related sediments (Table 3.1). These studies quantitatively provide evidence of repeated debris-flow activity during the Quaternary (Clark, 1987). Relative dating, when used as the only means for quantifying the age of a colluvial or fan deposit, may be less reliable than the other two techniques. Nonetheless, relative-age relationships have been used to correlate the development of debris flows from several different sites in western North Carolina (Mills, 1982; Mills and Allison, 1995b; Liebens and Schaetzl, 1997). Generally, these studies use weathering-rind development, Munsell soil color, clay content, and superposition to differentiate between older and younger debris-flow deposits. Although these studies reach reasonable conclusions, none of them provide a probable date for the surface deposits being studied, only a generalized time period. Results could be improved if the deposits could be correlated to similar deposits that have absolute age data. Unfortunately, this can be problematic as the affects of weathering may vary greatly, depending on lithology, climate, and other variables. This makes it difficult to correlate results from one area to another

36

unless the weathering conditions in both areas are very similar. Nonetheless, relative dating is a useful descriptive tool and a good first approximation for modern debris flow hazard areas. These studies also provide evidence of periods of episodic mass-wasting that was more extensive than the processes occurring today. Few radiocarbon-dating studies have been completed in North Carolina due to a lack of datable organic material in the stratigraphic record and the presumption that some deposits are too old (Table 3.1; Jacobson et al., 1989; Mills and Allison, 1995b). Existing dates in nearby Virginia range from as recent as 2,200 kyr to as old as 50,800 kyr (Eaton et al., 1997). Dates for North Carolina range from 16,000 25,000 kyr. Kochel (1990) believes that these older dates indicate that while debris-flow activity may have been impeded during glacial maxima, tropical moisture occasionally has invaded the area. Further radiocarbon dating, where possible, would greatly improve our knowledge of the Pleistocene environment and pre-historic Holocene mass-wasting in western North Carolina and the Appalachians. Few historical accounts of debris flows have actually been recorded and this is a field that could fruitfully be explored in more detail, particularly with new dating techniques using longer-half-life techniques such as K-Ar dating.

3.2

Modern Flooding and Debris Flows The first recorded instance of a major flood in the French Broad watershed occurred in

April 1791, six years before Asheville, NC was incorporated with its present name (Tennessee Valley Authority, 1960). Anecdotal accounts of this flood describe the water level as having been as high or a few meters higher than the well-documented flood of 1916 (Tennessee Valley Authority, 1960). The occurrence of this 1791 flood in April is consistent with the modern rainfall data that show April to be a particularly wet month on average (Table 2.1).

37

3.2.1

June, 1876 The first recorded instance of debris flows affecting the N.C. Blue Ridge occurred on

June 15-17, 1876. The debris flows accompanied flooding that is often called the June Freshet, one of the greatest floods in the upper reaches of the French Broad watershed (Tennessee Valley Authority, 1960). At the time, a debris flow was generally attributed to a waterspout, i.e., a sudden funnel-shaped cascade of water falling from the sky during a torrential rain event (Clingman, 1877). It was believed that the force of the falling water ripped away the soil from the side of the mountain, leaving only solid bedrock. The term waterspout was used not only to describe a meteorological event but also a geomorphic feature. The closest modern term to this "waterspout" is microburst. Clingman (1877) reports that at least 40-60 waterspouts were reported in portions of Macon and Jackson counties during the June, 1876 storm. Although Clingman (1877) did not provide a mechanism for the waterspouts, his detailed descriptions of the erosion and deposition are excellent, e.g., that of two debris flows that occurred in Macon County near the crest of Fishhawk Mountain and the Tessantee River on the afternoon of June 15, 1876 (Clingman, 1877). There were no known fatalities but the Conley family witnessed the debris flow across the river from their home: They saw a large mass of water and timber, heavy trees floating on the top, which appeared ten or fifteen feet high, moving rapidly towards them, as if it might sweep directly across the Tessantee and overwhelm them. Fortunately, however, sixty or seventy yards beyond the creek the ground became comparatively level, and the water expanded itself, became thus shallower, and leaving many of the trees strewn for a hundred yards along the ground, entered the creek with a moderate current. (Clingman, 1877, p. 69) Another flow also occurred on the opposite side of Fishhawk Mountain. The lengths of both of these debris flows were estimated to be two miles. The location of these slides is worth

38

noting as Fishhawk Mountain is the same area where four people where killed and 15 houses destroyed in a debris flow that occurred on Sept. 16, 2004 (see below). 3.2.2 May, 1901 From May 18-to-23, 1901 a series of low-pressure systems passed through western North

Carolina and brought heavy rain, with the heaviest precipitation occurring on May 21-22. The storm was centered near the Black Mountains of North Carolina. Total precipitation amounts ranged from 8.99 (22.8 cm) in Marion, North Carolina to 5.04 (12.8 cm) in Asheville, North Carolina (Myers, 1902). Extreme flooding affected portions of the Nolichucky, Watauga, Little Tennessee, and Catawba Rivers systems (Myers, 1902; Scott, 1972). Later flooding in the spring and summer only added to the destruction. Total damage to farms, bridges, highways, and buildings in the French Broad watershed was estimated to be four million dollars (U. S. Department of Agriculture, 1902). Most of the debris flows associated with the 1901 storm occurred in Buncombe, Henderson, Mitchell and McDowell Counties (Scott, 1972) (Figure 3.2). The Southern Railroad was particularly affected as a number of slides buried tracks for hundreds of meters or had portions of track washed away in the associated flooding. A resident of Marion, George Bird, reported that a number of slides affected the hills near his home and generated large piles of timber (Holmes, 1917). Landslides and waterspouts seemed to have been particularly prevalent in Mitchell County where as many as 17 slides were observed on one hill by Myers (1902) (Figure 3.3). Myers (1902, p. 104) describes in detail one of the largest slides he encountered: the excavated area was roughly heartshaped, having an extreme breath of about 100 ft., the distance from head to point being about 300 ft., and it was located on a hillside, sloping from 80 to 45 and having its head about 200 ft. below the crest of the hill, which was as high as any nearby. From the lower

39

end of the cavity a sharp and well-defined channel led down the hill to the stream at the base, this channel being from 5 to 6 ft. wide and from 4 to 5 ft. deep with side walls practically vertical cut down though a gravelly clay. It is estimated that the excavation has a total content of about 2,500 cu. yds. of earth which seems to have disappeared utterly. The particular slide described by Myers (1902) destroyed a log house that was in the flow path. Other accounts by area residents describe cloudbursts of extreme intensity accompanying the waterspouts and that water bubbled and then burst from the ground at the head of many smaller slides (Myers, 1902). It can be assumed from these descriptions that the mass movements in Mitchell County were debris flows, given their high water-and-debris content, their characteristic flow path, and their rupture surface. 3.2.3 July, 1916 During July of 1916 two tropical cyclones moved through the French Broad watershed

causing extensive flooding and numerous debris flows. On July 5-6, 1916 a weak hurricane passed over the Mississippi and Alabama coast and moved northeast, eventually deteriorating into a tropical depression by the time that it reached western North Carolina (Southern Railway, 1917). This storm produced 4 to 10 inches (10 to 25 cm) of rain but did not create any known debris flows. On July 14, a hurricane made landfall in Charleston, South Carolina and traveled rapidly northwest into the mountains of North Carolina (Figure 3.4). By the morning of July 15, the center of the powerful storm had already reached western North Carolina. The flood of July 14-16, 1916 was the largest recorded flood ever to have affected western North Carolina. It also produced numerous debris flows in the mountains. Rainfall totals for the 1916 storm were exceedingly heavy with nearly all of the eastern slopes of the North Carolina Blue Ridge received 10 inches (25 cm) of rain or more (Scott, 1972). The greatest amount of rain was recorded in the French Broad watershed at Altapass,

40

where 22.22 inches (56.4 cm) fell in a 24-hour period (Hudgins, 2000). This is also the greatest 24-hour rainfall total ever recorded in North Carolina. Generally, the storms of 1916 produced two distinct regions of exceptionally heavy precipitation, i.e., one in Mitchell, Avery, and Caldwell counties, and the other in Transylvania and Henderson counties (Figure 3.5). Runoff from the second storm was estimated to be as high as 80-90 percent (Southern Railway, 1917). The first storm had already thoroughly soaked the soil, increasing antecedent moisture conditions, and filled most streams nearly to flood stage (Scott, 1972). Rainfall from the second storm exacerbated flood conditions (Figure 3.6). The July, 1916 storms killed about 80 people and caused $22M in damages (Southern Railway, 1917). In Asheville, several homes and buildings were destroyed and four of the main river bridges were washed away (Tennessee Valley Authority, 1960). The Southern Railway Company suffered extreme financial losses and transportation within western North Carolina was disrupted for several days. Many railway lines were covered by debris flows, trapping freight and passenger trains between terminals. The Southern Railway (1917) reported that almost every mile of track between Asheville and Statesville was covered by debris or washed out. At some places, track was suspended in mid-air after the fill below was washed away (Southern Railway Company, 1917). Generally, debris flows were reported along the Blue Ridge Mountains to the east, southeast, and south of Asheville (Holmes, 1917; Scott, 1972) (Figure 3.2). Most slides occurred between 5 pm., July 15 and 7 a.m., July 16. The flows began before dark and could be heard throughout the night during the period of heaviest rainfall. They typically initiated in topographic hollows where the soil was thick, near the head of surface streams. Flow thicknesses ranged from

41

0.6 to 6.1 m (2 to 20 ft) and averaged 1.5 to 1.8 m (5 to 6 ft). Bedrock was seldom exposed anywhere along any slide (Holmes, 1917). 3.2.4 August, 1940 In August of 1940, a pair of storms caused significant flooding and numerous debris

flows in the western mountains of North Carolina; the first occurred from August 11-17 and the other from August 28-31. These storms also brought record flooding to portions of Virginia, Tennessee, and South Carolina. Approximately 30-40 lives were lost and there were at least $30M in damages (U.S. Geological Survey, 1949). The situation was similar to that of 1916, with two large storms occurring in the same month. The 1940 mid-August storm was strikingly similar to the to the second 1916 storm in terms of rainfall intensity and storm path (Figure 3.4). However, unlike the 1916 storm, the antecedent moisture conditions in 1940 were relatively dry, allowing for increased infiltration, hence lower overall flood discharge levels (U.S. Geological Survey, 1949). The first storm in 1940, an unnamed hurricane, made landfall between Beaufort, South Carolina and Savannah, Georgia on August 11, 1940. Although no wind speeds were recorded, damage reports indicate that trees were uprooted and broken, many buildings were damaged or destroyed, and 20 coastal residents were killed. An unusually high tide was reported, reflecting the storm surge. The storm then moved inland and curved northward following the Savannah River Valley, weakening significantly. It followed a semi-circular path through Georgia, Tennessee and Virginia, and then back into North Carolina before it moved offshore on August 16, just south of Norfolk, VA (Figure 3.4) (U.S. Geological Survey, 1949). This mid-August hurricane of 1940 did not affect the French Broad Watershed until August 13-14 (Tennessee Valley Authority, 1960). While rainfall intensities were moderate, the

42

slow rate of movement allowed for heavy precipitation for several days over the North Carolina Blue Ridge, resulting in high rainfall totals. Maximum precipitation totals ranged from 13 to 16 inches (33 to 41 cm) at to as little as 5 inches (13 cm) in Asheville (Tennessee Valley Authority, 1960). A series of well-defined storms centers over the Appalachians Mountains extended toward the northeast from Blue Ridge, Georgia to Luray, Virginia (Figure 3.7), apparently due to an orographic influence on the storm precipitation (U.S. Geological Survey, 1949). The second storm in 1940 occurred during the period of August 28-31 but intense rainfall did not begin until the morning of August 29. Rainfall continued to fall until August 30 when it abruptly ended around noon. Only passing showers remained by August 31 (U.S. Geological Survey, 1949). This storm was a relatively local meteorological disturbance that only affected the French Broad and Little Tennessee watersheds (Figure 3.8). Precipitation was of both shorter time duration and aerial extent than the mid-August storm, but of higher intensity. Rainfall amounts ranged from 8 to 13 inches (20 to 33 cm) on the western slopes of the Blue Ridge in 20 to 30 hours (U.S. Geological Survey, 1949). Given the antecedent moisture conditions due to the earlier storm, flooding was more severe near the storm center but overall was not as widespread. The 200-300 debris flows associated with both storms of 1940 attributed greatly to the devastation wrought by the floods (Scott, 1981). These slides occurred near the centers of both storms in shallow saturated soils on steep slopes (Figure 3.2). They originated on slopes 300 to 400 feet (91.5 to 122 m) from the tops of mountains, near the slope break. The debris flows ranged in width from 6 to 8 feet (1.8 to 2.4 m) and 40 to 50 feet (12 to 15 m) long to 200 to 300 feet (61 to 91 m) wide and 1320 to 2640 feet (402 to 805 m) in length (U.S. Geological Survey, 1949). Many of the larger debris flows then continued downslope following stream valleys, uprooting trees and destroying structures (Wieczorek et al., 2004).

43

During the mid-August storm of 1940, debris flows mainly occurred in the Blue Ridge Mountains, from the North Fork of the Catawba River northward into Watauga County near the North Carolina Virginia border (Figure 3.2). During the late August storm, debris flows occurred mainly in the Upper Pigeon and Tuskasgee River basins (Figure 3.2). Because of the concentration of high-intensity rainfall within a small area, more than 200 debris flows occurred in an area of only 150 mi2 (388.5 km2) (U.S. Geological Survey, 1949). 3.2.5 November, 1977 In early November 1977, a storm system that had formed as a low-pressure system in the

Gulf of Mexico moved northwestward into the Appalachian Mountains (Neary and Swift, 1987). Rainfall began in western North Carolina in the early morning of November 2 and continued at a steady rate (20-50 mm/day) until November 5. This steady rain was followed by intense downpours (102 mm/hr) on the night of November 5-6 during which most of the debris flows were initiated (Nearly and Swift, 1987). This heavy precipitation, as in 1916 and 1940, was produced by convection associated with orographic lifting over the southern Appalachians. Four areas of exceptionally heavy precipitation (200-320 mm) were produced along the southeast ridges of the North Carolina Blue Ridge. Two of these areas were within the French Broad Watershed (Neary and Swift, 1987) (Figure, 3.9). Although the heaviest rainfall in 1977 occurred in the vicinity of Mt. Mitchell, the best information about debris flows and flooding came from the Bent Creek watershed, located about 15 km southwest of Asheville. A survey was conducted here immediately following the storm (Neary and Swift, 1987; Otteman, 2001). At least seven major flows and other small failures were identified in this area (Neary and Swift, 1987). Most of these debris flows occurred on steep slopes (26-46) at high elevations (945-1100 m) and flowed downhill following

44

ephemeral creekbeds or along hillslope depressions (Pomeroy, 1991). Scarps occurred in shallow residual soils less than 1 m deep over gneissic bedrock (Neary et al., 1986). All of the flows occurred in undisturbed, forested areas (Neary et al., 1986). Topography in the Bent Creek watershed is at least partially controlled by the underlying concentration of tension joints in the bedrock. Where there is a greater amount of jointing, topographic hollows tend to develop. These joints allow for the infiltration of groundwater, enhancing breakdown of the rock. This accelerates weathering, providing loose material for mass wasting (Pomeroy, 1991). Debris flows seem to originate on the bedrock-soil or bedrockcolluvium interface within these hollows. The November 1977 flood killed at least thirteen people; sixteen counties in western North Carolina were declared disaster areas. The most serious flooding occurred along the French Broad River downstream from Asheville and in Yancey County where nearly every bridge was washed out (Eshner and Patric, 1982; Stewart et al., 1978). Flooding destroyed 384 homes, 389 miles (622 km) of highway, and 12 dams. In total there was over $50M in damages associated with this storm (Stewart, 1978). In 1977, precipitation, slope, and topography all contributed to the initiation of debris flows southeast of Asheville (Pomeroy, 1981). In comparison to the historical range of rainfall intensities for the entire Tennessee Valley area, the maximum intensities associated with the 1977 storm were in the middle-to-low range but antecedent moisture was exceptionally high (177% above normal) for the two months preceding the storm (Neary and Swift, 1987). The combination of very wet antecedent conditions and high-intensity, short-duration rainfall created excellent conditions for debris flows to form. Slope aspect also showed a definite affinity

45

towards the northeast, east, and southeast (Pomeroy, 1991). These slopes would only receive sunlight in the morning and thus would have higher soil moisture. 3.2.6 September, 2004 The 2004 Atlantic hurricane season was exceptionally brutal. Fifteen tropical or

subtropical storms formed in the North Atlantic. Nine of these storms became named hurricanes and six of these struck the United States (National Weather Service, 2004a). In North Carolina, the remnants of three tropical systems, i.e., hurricanes Frances, Ivan and Jeanne, impacted the western part of the state in September. Frances and Ivan caused extreme flooding in Asheville and several debris flows and rockslides in the mountains, affecting Interstate-40. Rainfall totals for the month over much of western North Carolina ranged from 10 to 25 inches (25 to 64 cm). This was 2-to-5 times greater than normal (Badgett et al., 2004). Hurricane Frances struck the east coast of Florida early on Sept. 5, 2004 and quickly weakened into a tropical storm (National Weather Service, 2004a). The storm then rapidly moved across the state, through the panhandle of Florida, and northeastward across the eastern United States (Figure, 3.4). The effects of hurricane Frances could first be felt in North Carolina on September 6 around 6:00 p.m. (Boyle, 2004) but most of the flooding and mass wasting occurred on September 8 (Appendix C). In North Carolina, the heaviest precipitation in 2004 occurred slightly east of the French Broad Watershed but heavy rain also fell in the eastern portions of Tennessee and Kentucky, and through most of West Virginia, Virginia, Maryland, Ohio and southwestern Pennsylvania (Associated Press, 2004). Two areas experienced exceptional rainfall over Transylvania, Yancey and McDowell Counties (Figure, 3.10). The highest precipitation total was recorded in Edgemont, NC, 80 miles northeast of Asheville, which received 16.6 inches (42.2 cm) of rain.

46

Sixty miles southwest of Asheville, Lake Toxaway received 14 inches (35.6 cm) of rain (Nowell, 2004). In total, 17 western counties were affected by flooding (Anonymous, 2004). Hundreds of people were evacuated from their homes and several had to be rescued from the rising water (Nowell, 2004). Areas of Asheville located near the Swannanoa River were flooded, particularly the shopping center near the entrance to the Biltmore Estate, were water stood as much as 5 feet (1.5 m) deep (Nowell, 2004). In Haywood County, flooding along the Pigeon River also inundated downtown Canton and Clyde. The remnants of hurricane Frances caused at least 21 reported incidents of mass wasting (Appendix C) along several major roadways in seven western North Carolina counties. However, only three counties within the boundaries of the French Broad Watershed experienced debris flows, i.e., Avery, Henderson, and Transylvania. The largest reported debris flow occurred east of Asheville on Interstate-40, near Old Fort Mountain in McDowell County (Figure 3.12). This slide crossed the westbound lane and the median to block four of the six lanes of a five-mile stretch of Interstate-40 (Nowell, 2004). In Watauga County, one house was destroyed and eight others condemned when a debris flow moved through a subdivision (North Carolina Geological Survey, 2004a). Portions of the Blue Ridge Parkway were closed when at least six debris flows destroyed the roadway in four areas between Linville Falls and Waynesville (Ball, 2004). About 250 roads became impassable or were closed due to flooding and mass wasting (Barrett, 2004). Most of that road damage was in Buncombe County (Ball, 2004). Ivan was an unusually long-lived hurricane that made landfall along the United States coast twice. Ivan struck the Alabama coast early on September 16 as a Category 3 hurricane and gradually weakened as it moved northeastward into the southeastern United States (Figure 3.4). After emerging off of the Delmarva Peninsula on September 19, a portion of the storm moved

47

southwestward, crossed over Florida and then into the Gulf of Mexico. By September 23, the remnants of Ivan had strengthened into a tropical storm and the re-born storm made landfall for the second time on September 24 over southwestern Louisiana (National Weather Service, 2004a). The remnants of Ivan moved into western North Carolina early on September 16. Although Ivan had weakened to a tropical storm by the time it reached North Carolina, it still packed powerful winds and heavy rain. Rainfall was not as heavy as that which fell during Frances, mainly because the storm moved rapidly northeastward, but the western portion of the state still received between 4 to 8 inches (10 to 20 cm) of rain. The heaviest precipitation fell in Transylvania, Jackson and McDowell Counties at high elevations. Black Mountain (near Asheville) received 11.5 inches (29.2 cm) of precipitation and Sapphire (in Transylvania County) reported 15 inches (38 cm) (Figure 3.11) (Badgett et al., 2004). Although Ivan produced less rain than Frances, high antecedent-moisture conditions and saturated soils allowed for more slope movements to be produced. A total of 53 slope movements have been attributed to hurricane Ivan (Cabe, 2004). Those recorded in Appendix C occurred along major roadways (20 slope movements), but several other slope movements may also have occurred in undisturbed or rural areas and were not reported by either the North Carolina Department of Transportation (NCDOT) or major news agencies. Further work will have to be conducted to obtain a complete record of these slope movements. Slope movements, downed trees, and flooding obstructed several roads throughout western North Carolina, stranding residents in several communities in Avery, Jackson and Haywood Counties (Appendix C). A major slope movement occurred in the westbound lane of Interstate-40 at milepost 35 in Haywood County at the edge of the study area (Figure 3.11).

48

Further west, near the North Carolina-Tennessee border, a large portion of the eastbound lane of Interstate-40 collapsed due to undercutting by the swollen Pigeon River. A major debris flow also destroyed a home in Candler in Buncombe County (Cantley-Falk, 2004). The worst damage occurred in the community of Peeks Creek in Macon County. At around 10:10 p.m. on September 16, a debris flow originated near the peak of Fishhawk Mountain (Figure 3.13) destroying at least fifteen houses, injuring several people, and resulting in the deaths of four people (and an unborn baby). The debris flow traveled approximately 2.25 miles (3.6 km), possibly in pulses, dropping nearly 2200 feet (670 m) in elevation as it progressed down a mountain cove and into the north fork of Peeks Creek (Cabe, 2004). The velocity of the flow was estimated to be 20.3 mi/hr (32.7 kph) near the scarp and 33.2 mi/hr (5305 kph) just upstream of the area of major damage (Cabe, 2004). The force of the flow scoured the streambed, ripped trees down and left other striped of bark; houses were removed from their foundations (North Carolina Geological Survey, 2004b; Ostendorff, 2004). The flow probably originated as a debris slide; a slab of cohesive rock, debris and earth the size of a football field detached from the side of the mountain and quickly disintegrated into a debris flow as more water mixed in with the slide material (Cabe, 2004). What is remarkable about the Peeks Creek disaster is that this location is the same area where a two large debris flows occurred in 1876 (Clingman, 1877). Observations of residents living in the area were strikingly similar in both incidents. Clingman (1877) describes trees stripped of bark and limbs, a clean, broad furrow more than two miles long carved into the side of the mountain, and boulders weighing several tons moved by the flow. Residents during both the 1876 and 2004 incidents reported seeing or hearing a tornado or waterspout just before or during the debris flow. In the light of an exploding electrical

49

transformer, one resident described seeing debris spinning and flying around in the air in a circular motion above their house (Biesecker and Shaffer, 2004). In 1876, residents described seeing funnel-shaped spinning masses of water near the crest of the mountain (Clingman, 1877). Tornadoes are fairly rare in mountainous areas, but do occasionally develop. While there was wind damage throughout the Peeks Creek area after the passage of Ivan, this damage was more consistent with wind shear. So far, the National Weather Service has not been able to conclude if a tornado actually did touch down on Fishhawk Mountain, but they do not discount the eyewitness accounts of local residents (Cabe, 2004). The question remains, "Why did mass movements occur in these Macon County areas as opposed to elsewhere?" In the Peeks Creek flow, fracture planes in the rock, sloping 35-55 degrees, provided a smooth slip surface. Soil layers over this bedrock were thin, generally less than three feet (1 m) deep (Cabe, 2004). These physical properties of the terrain probably facilitated the debris flows on Fishhawk Mountain. Meteorologically, the rainfall rates from the remnants of hurricanes Frances and Ivan were not unusually intense for either event. However, the combined rainfall totals were exceptionally heavy. The rainfall produced by Frances initially saturated mountain soils and slopes. Before the soil had a chance to drain sufficiently, Ivan moved through the area, bringing even more rain to already soaked areas. Rainfall from Ivan may have caused even higher soil-water pressure on slopes than during Ivan, explaining why there was more mass wasting during the second storm. Nonetheless why, exactly, debris flows occurred in one area and not another, even under similar meteorological and physical conditions will take more research.

50

3.2.7 Precipitation Thresholds and Recurrence Intervals Generally, historical rainfall totals from 1901 to 2004 are well within the 125 mm/d to 250 mm/d precipitation thresholds suggested by Eschner and Patric (1982) as necessary for debris-flow generation (see Figure 1.2). Average rainfall amounts were greater than 125 mm/d in all cases, i.e., greater than the minimum amount of precipitation necessary to saturate soil and set the stage for debris flows (Table 3.2). While storms with rainfall totals of 250 mm/d are described as extremely rare, all but one storm produced maximum precipitation amounts that exceeded 250 mm/d (Eschner and Patric, 1982). Extreme precipitation does not necessarily guarantee that debris flows will occur, as during the July 5-6, 1916 storm, but extreme precipitation certainly increases the risk of slope instability. Escher and Patric (1982) suggested a return interval of 100 years or less for storms producing debris flows in western North Carolina. Nonetheless, historical rainfall data from 1901 to 2004 indicates that recurrent intervals may be smaller. According to Cabes (2004), there have been 14 storms or hurricanes that have triggered slope movements in western North Carolina since 1901. Based on the occurrence of major debris-flow triggering storms from the historical record (i.e. those that have extensive amounts of precipitation and debris-flow location data), a preliminary recurrence interval for slope instability can be estimated. The return interval for mass wasting from 1876 to 2004 (for six events) averages 25.6 years. This return interval has varied from as few as 15 years to as many as 37 years. Based on Cabes data (2004), the recurrence interval for slope movements from 1901 to 2004 may be as few as 7.4 years. On human timescales, this is still enough time for people to forget that mass wasting can occur in their area. Nearly all of the major events that caused debris flows in western North Carolina occurred when two storm systems, producing heavy precipitation greater than 12.5 cm/d,

51

traveled over the area within 10 to 20 days of each other. Antecedent moisture seems to play a crucial role in predisposing slope to debris-flow generation. Geoscientists, emergency management, and citizens living in these mountainous areas, must be vigilant in monitoring weather conditions, particularly with repeated sequences of heavy rain events.

Figure 3.1: Old slope movement deposit along a private drive in Maggie Valley, North Carolina (age not determined). There is significant soil development in this poorly sorted colluvial deposit. It is located along a chute where a modern debris flow occurred.

52

Table 3.1: Prehistoric debris flow studies in the southern Blue Ridge and the age-dating techniques utilized.

Reference
Kochel Jacobson et al. Behling et al. Kochel Eaton et al. Shafer Mills Mills and Allison Mills and Allison Liebens and Schaetzl Mills

Year Dating Technique

1987 Radiocarbon 1989 Radiocarbon 1993 Radiocarbon 1990 Radiocarbon 1997 Radiocarbon 1984 Thermoluminescence 1982 Relative-age Relative-age 1995a /paleomagnetism 1995b Relative-age 1997 Relative-age 2000 Relative-age

Location
Davis Creek, VA West Virginia

Age of Features
> 11,000 BP 10,000 - 12,000 BP; 315 BP 17,000 - 22,000 BP

West Virginia Appalachian Mountains, 16,000 - 25,000 BP NC Upper Rapidan River Basin, 2,200-50,800 BP VA Flat Laurel Gap, NC Late Quaternary North Carolina ? Watauga County, NC Haywood County, NC 780 ka - 1Ma ?

Macon and Swain Co., NC ? Appalachians ?

53

Figure 3.2: Areas of major debris flows and landslides in western North Carolina (after Scott, 1972).

54

Figure 3.3: A sketch map of debris flows that occurred along Gouges Creek in Mitchell County, North Carolina in May, 1901 (Myers, 1902).

55

Figure 3.4: Map showing some of the hurricane and storm paths that have affected western North Carolina as reported by the U.S. National Hurricane Center and the U.S. Geological Survey Water Resources Branch (1949).

56

Figure 3.5: Total storm precipitation for July 14-16, 1916.

57

Figure 3.6: U. S. Geological Survey hydrograph from the river gauge located on the French Broad River in Asheville for the month of July, 1916. In early July, there is a large hydrographic spike generated by rainfall from the remnants of a hurricane that passed through the area. There is also a second large peak between July 16 and 17 generated by the flooding associated with a hurricane moving northwestward over the watershed. This is the greatest recorded streamflow for the gauge at Asheville.

58

Figure 3.7: Total storm precipitation for August 14-15, 1940 adapted from U. S. Geological Survey, 1949).

Figure 3.8: Total storm precipitation for August 28-31, 1940 (adapted from U. S. Geological Survey, 1949).

59

Figure 3.9: Total storm precipitation for November 2-5, 1977 (adapted from Neary and Swift, 1987).

60

Figure 3.10: Total storm precipitation (inches) for the remnants of Hurricane Frances (National Weather Service, 2004b).

Figure 3.11: Total storm precipitation (inches) for the remnants of Hurricane Ivan (National Weather Service, 2004c).

61

Figure 3.12: A debris flow that blocked the westbound lanes of Interstate-40 near Old Fort Mountain in McDowell County (North Carolina Geological Survey, 2004a).

Figure 3.13: The initiation zone of the debris flow that occurred on Fishhawk Mountain and devastated the Peeks Creek area of Macon County on September 17, 2004 (Wilett, 2004).

62

Table 3.2: Major storms within the French Broad Watershed and their minimum, average, and maximum precipitation amounts. STORM DATE MIN (in) AVG (in) MAX (in) MIN (mm) AVG (mm) MAX (mm) 1901 5 7 8.9 128 177.8 228.4 1916 (1)* 4 7 10 101.6 177.8 254 1916 (2) 1 10 22.2 25.4 254 564.4 1940 (1) 4 13 16 101.6 330.2 406.4 1940 (2) 3 8 13 76.2 203.2 330.2 1977 2 8 14 50.8 203.2 355.6 2004 - FRANCES 4 10.3 16.6 101.6 261.6 421.6 2004 IVAN 4 9.5 15 101.6 241.3 381 *no debris flows produced

63

CHAPTER 4: METHODOLOGY In order to quantify debris-flow potential in the French Broad Watershed, two deterministic process-based computer-modeling programs have been chosen, i.e., SINMAP (Stability INdex MAPping) (Pack et al., 1998) and SHALSTAB (Shallow Landsliding Stability Model) (Dietrich and Montgomery, 1998). Both use similar numerical models and steady-state hydrologic assumptions to quantify the influence of topography on pore pressure. These computer programs, instructions, and examples are freely available for download from their respective websites. 4.1 Infinite-slope Equation SINMAP and SHALSTAB are based on the infinite-slope form of the Mohr-

Coulomb failure law, an equation commonly applied as part of a slope-stability model within the GIS environment (Hammond et al., 1992; Montgomery and Dietrich, 1994; Wu and Sidle, 1995; Pack et al., 1998). The infinite slope equation is:

where Cr is root cohesion, Cs is soil cohesion, is slope angle, s is soil density, w is the density of water, g is acceleration due to gravity, D is the vertical soil depth, Dw is the vertical height of the water table, and is the internal soil friction angle. Ultimately, the infinite-slope equation calculates a dimensionless stability index, or factor of safety (FS), by comparing the destabilizing components of gravity with the stabilizing components of friction and cohesion, as illustrated in Fig. 4.1 (Pack et al., 1998a). The infinite-slope equation makes several simplifying assumptions. First, the failure plane and groundwater surface are assumed to be parallel to the ground surface.

64

This assumption is reasonable because the drainage barrier, e.g., top of bedrock, and the ground surface are often parallel on colluvial slopes. The contrast in hydraulic conductivity between the overlying soil and the material forming the drainage barrier causes groundwater to flow nearly parallel to the ground surface, providing a surface for slope failure (Hammond et al., 1992). Another assumption is that the failure plane is assumed to be infinite in extent and the resistance to movement along the sides and ends of the slope movement are so insignificant that they may be ignored (Hammond et al., 1992). Finally, the soil is assumed to be of uniform thickness. Although SINMAP and SHALSTAB share similar physical assumptions and equations, they use different indices to quantify instability. SINMAP uses the infiniteslope equation to calculate a factor of safety, i.e., an estimated potential for instability whereas SHALSTAB quantifies terrain instability in terms of the effective recharge required to trigger instability. A comparison of the results and relative performance of these programs is difficult due to their dissimilar output. To simplify comparison, model parameters have been kept as similar as possible. 4.2 Parameterization of the Models SINMAP and SHALSTAB share specific parameters that must be input into the

program to optimize the models for local soil and meteorological conditions. The models have three necessary components in common, i.e., digital terrain data in the form of a digital elevation model (DEM) or some other type of grid-based digital topographic data, e.g., LIDAR, a local landslide database or inventory, and site-specific soil data.

65

4.2.1

Digital Elevation Models Both SINMAP and SHALSTAB assume that the dominant control on debris-flow

occurrence is surface topography, specifically the interplay between slope and shallow subsurface flow convergence (Montgomery and Dietrich, 1994). To calculate topography, both computer programs require digital elevation data. In this study, DEMs at 10-meter and 30-meter scales were obtained from the United States Geologic Survey National Elevation Dataset (http://seamless.usgs.gov/). Both SINMAP and SHALSTAB are highly sensitive to the accuracy of the DEM. Verification of data quality is advisable prior to running these models (Zhang et al., 1999; Dietrich et al., 2001; Guimaraes et al., 2003). As explained by Michael Oimoen, a scientist from the USGS National Center for Earth and Resources Observations and Science (pers. comm., 2005), USGS DEMs used for this study were created from the digitized contours of scanned 1:24,000-scale topographic quadrangle sheets. The contours from these maps are digitized and either a 10 x 10 meter or 30 x 30 meter mesh grid is overlaid on top of the digitized contours. For each grid cell, inverse distance weighting interpolation is used to calculate the elevation value of each cell (Maune et al., 2001). Since the 10-meter and 30-meter DEM are created from the same scale map and source data, neither is more accurate than the other. But as the cell size is smaller for the 10-meter DEM, the 10-meter spacing better captures the detail in the contours. Given both 10- and 30-meter DEM data are available for the watershed, SINMAP and SHALSTAB runs were completed in both resolutions and the modeled results have been compared herein. Once downloaded, the DEMs were clipped using ArcInfo to the shape of the French Broad Watershed before being input into SINMAP and SHALSTAB for processing. Both programs use the DEM to extract elevation data, calculate the specific

66

catchment or drainage area, and compute the slope for each grid cell (Dietrich and Montgomery, 1998; Pack et al., 2001). Both programs have procedures for increasing accuracy and eliminating potential errors inherent in the DEM grid. Pit-filling corrections are automatically executed within both programs to eliminate artificial sinks or depressions in the DEM. Given that such grid elements are rare in natural topography, they are assumed to be errors created during the initial preparation of the DEM (Pack et al., 2001). Both programs use a flooding approach where pits are filled, raising the elevation to that of the lowest neighboring grid cell. The processing time for DEMs with large spatial areas, such as a DEM for the entire watershed, can be extremely long, depending on the speed and memory availability of the users personal computer. Both SINMAP and SHALSTAB seem to have a DEM size limitation. Although SINMAP could process the entire watershed at the 30-meter scale, the 10-meter DEM had to be broken down into county-size pieces. SHALSTAB cannot process DEMs larger than a county at a 30-meter scale without creating run-time errors and premature termination. At the 10-meter scale, SHALSTAB can run a maximum of four 1:24,000-scale quad-sized DEM blocks before shutting down. This significantly increases process time. For this reason, the original 30-meter DEM had to be clipped to county-size when run in SHALSTAB. In the interest of time, only Haywood County was chosen for several model runs in SHALSTAB for comparison to the SINMAP results. The 10-meter DEM for Haywood County then had to be broken down into five smaller components.

67

4.2.2

Debris Flow Inventory In order to complete the slope-stability model runs for SINMAP and

SHALSTAB, basic digital-line and polygon coverages were collected and formatted for use in ArcView 3.x and ArcGIS 8.3. Both programs require the input of landslide location data to verify the model results; SINMAP requires point data whereas SHALSTAB requires polygonal shapes of each landslide feature. Landslide location data for this study was obtained through a combination of aerial-photography interpretation, previous studies, field investigation, and data provided by both the North Carolina Geological Survey and the North Carolina Department of Transportation. Most of this data was not in a digital (e.g., GIS) format and was digitized into a single landslide-location database for the French Broad Watershed (Appendix D). Other GIS data, e.g., roads, hydrology, and geology, and digital orthophoto quadrangles (DOQs) were obtained through the North Carolina State University Geodigital Library website (http://www.lib.ncsu.edu/stacks/gis/) and the North Carolina Center for Geographic Information and Analysis map server (http://204.211.135.111). 4.2.3 Soil Data Various soil parameters, such as cohesion, density, and hydraulic conductivity,

have been estimated from the general soil maps of North Carolina and from the advice of experienced geologists who routinely work with soils in the study area. Basic soil maps were downloaded from the State Soil Geographic database (STATSGO) available through the U.S. Department of Agriculture (1998). These STATSGO maps were created by generalizing more detailed soil survey maps for each county. Given the scale associated with the study area, it is considered appropriate to use the general soil information provided by this database. Each soil unit is linked to descriptive attributes

68

that give the proportional extent of the component soils and their properties. There are eighteen general soil types found within this watershed (see Figure 2.2). The STATSGO soils coverage for the State of North Carolina was downloaded as a polygon shapefile, imported into ArcGIS, and clipped to the extent of the French Broad Watershed. 4.2.4 Soil Density Both programs require the input of a value for soil density. This value is used to

represent the total bulk density of the soil over the entire study area. In SINMAP the default value for soil density is 2000 kg/m3 while in SHALSTAB the value is 1700 kg/m3. Otteman (2001) used a value of 1450 kg/m3 for her study area in the Bent Creek Experimental Forest in Buncombe County, North Carolina whereas in Madison County Virginia, a value of 1200 kg/m3 was used by Morrissey et al. (2001). A geologist with the North Carolina Department of Transportation estimates typical density values of 1441.6 kg/m3 to 2082 kg/m3 in soils of western North Carolina (Jody Kuhne, pers. comm., 2004). Although no independent soil-density analysis was completed, a value of 1922 kg/m was chosen as representative value for the entire watershed and was used in all model runs. This number was suggested by Richard Wooten, a geologist with the North Carolina Geological Survey who has done extensive slope-stability work in western North Carolina. 4.3 SINMAP Parameters SINMAP has simplified the infinite-slope equation (1) originally proposed by

Hammond et al. (1992). In SINMAP, the soil thickness (h) is specified perpendicular to the slope, rather than vertically downward (D) (Figure 4.1). Soil and root cohesion are

69

combined into one dimensionless cohesion factor. The SINMAP stability-index equation is given by:

where the variables a and are the specific catchment area and slope, respectively, and are derived from the topography determined using a DEM. The other parameters, C (cohesion), (soil friction angle), R/T (recharge divided by transmissivity), and r (the ratio of water and soil density) are manually entered into the model. These parameters are considered more uncertain and are specified in terms of upper and lower boundary values (Pack et al., 1998b). The default values for the foregoing parameters are provided in Figure 4.2. Values may be adjusted to represent local conditions, to identify landslide-prone areas more precisely, and to assure that a significant amount of the landslides are captured in areas considered unstable. The following text rationalizes the specific input values and distributions used in this study (T/R, C, and ). For each cell of the DEM, SINMAP calculates a factor of safety (FS), a ratio of stabilizing-to-destabilizing forces found in equation (2). This is a dimensionless index number with a value between zero and 10. If the index falls below one, there is a high probability that the area is unstable whereas high index values (greater than one) indicate better stability (Pack et al., 2001).

70

4.3.1

T/R (Ratio of Transmissivity to Effective Recharge) The ratio of transmissivity of the soil (m2/hr) to the effective recharge (m/hr) has a

default range of 2000 to 3000 (m). When multiplied by the sine of the slope, the T/R value can be interpreted as the length of the hillslope (in meters) required to develop saturation (Pack et al., 1998b). The recharge rates used in this study have been derived from four precipitation thresholds, i.e., 50 mm/d, 125 mm/d, 250 mm/d, and 375 mm/d. Two of these precipitation thresholds, 125 mm/d and 250 mm/d, are described by Eschner and Patric (1982) as necessary for debris-flow initiation in the Appalachians (see Figure 1.2). The 50 mm/d rate was chosen as a minimum rate, i.e., one where SINMAP should not indicate a large mapped area of instability. The last threshold amount, 375 mm/d, is used as a maximum, i.e., an extreme example of the precipitation that can produce debris flows in the French Broad Watershed. Only a few times have recorded rainfall totals actually exceeded this rate (see Table 3.2). The transmissivity rate (m/hr) was calculated using the basic equation: T = Kb (3)

where K (m/hr) is the permeability or hydraulic conductivity of the soil and b (m) is the soil depth. Permeability rates within each soil unit are available in the STATSGO soil attribute database. The permeability of each of the eighteen soil units within the watershed was calculated by determining the average permeability of each of the component soils within each major unit (Appendix B). These permeability values were then multiplied by the soil depth (b), which was held at a constant two meters, to calculate the transmissivity for each soil unit. Finally, the transmissivity was divided by

71

the recharge rate to find the upper and lower T/R (m) values. The final T/R values used were an average of each of the eighteen soil units (Table 4.1). 4.3.2 Dimensionless Cohesion In SINMAP, root and soil cohesion is combined with soil density and thickness to

calculate a dimensionless cohesion factor, C (Pack et al., 1998b). Conceptually, this is the ratio of the cohesive strength of the soil and roots relative to the weight of a saturated thickness of soil, or the contribution of cohesion to the stability of a slope (Pack et al., 1998b). The equation used to determine dimensionless cohesion is: C = (Cr + Cs)/(hsg) (4)

where Cr is root cohesion (N/m2), Cs is soil cohesion (N/m2), h is the soil thickness (m), s is the wet soil density (kg/m3), and g is the acceleration due to gravity (9.81 m/s2). Various values for cohesion have been suggested in diverse SINMAP studies. The default values used for cohesion in SINMAP are 0 and 0.25 but Morrissey et al. (2001) used values between 0 and 1.28. Wooten uses values of 0.6 to .96, suggesting that zero cohesion is unlikely in the Southern Appalachians due to the considerable amount of low vegetation growing on mountain slopes (e.g., rhododendron), which would add to the overall root cohesion found in the soil. Because of the lack of more precise soil cohesion data, the SINMAP default values were deemed reasonable for the watershed and used in all of the SINMAP model runs. 4.3.3 Internal Soil Friction Angle Like cohesion, the internal angle of friction is a component of a materials

shearing resistance. Internal friction is the friction between individual grains within a mass of material. Failure occurs when internal friction is overcome along a given

72

shearing angle. The angle is constant for a specified material and depends on the size, shape, and surface roughness of the grains, the density of the soil, moisture content, and material saturation (Easterbrook, 1999). Values for the soil friction angle for certain soil types can be estimated from tables provided by Hammond et al., (1992). These tables require that the soil be classified according to the Unified Soil Classification (USC) system (ASTM D-2487-85 and D2488-84) which are part of the STATSGO attribute data. Although no independent soil analysis was completed, the 26 and 45 soil-friction angles used to calibrate the model were considered realistic for the study area (Hammond et al., 1992). Given that this study requires that parameters be generalized over large areas, a wide variation encompassing the properties of several different soil types, from clayey to sandy and gravelly soils, is more realistic than a small range of values. 4.4 SHALSTAB Parameters Like SINMAP, SHALSTAB is an extension of ArcView and is theoretically

based on the infinite-slope equation. SHALSTAB assumes steady-state, saturated flow parallel to the slide surface and uses Darcys Law to estimate the spatial distribution of pore water pressure. However, the equation that SHALSTAB uses to formulate a slopestability map is quite different from that used by SINMAP. The SINMAP equation solves for a factor of safety whereas SHALSTAB calculates the ratio of log q/T (effective precipitation over transmissivity) (Dietrich et al., 2001). Essentially, this is the amount of precipitation necessary to cause instability in a slope with a known soil transmissivity rate.

73

The two hydrologic-slope stability equations solved by SHALSTAB are shown above. Equation (5) solves for the hydraulic ratio while equation (6) solves for the area per outflow boundary length. The equations have three topographic terms defined by the DEM: drainage area (a), outflow boundary length (b), and hillslope angle (). Of the other four parameters, soil bulk density (s) and internal friction angle () can be assigned by the user. The ratio of transmissivity (T) and effective precipitation (q) are solved by SHALSTAB and are given as the final output of the model (Dietrich et al., 2001). The authors of SHALSTAB made even more simplifications to the infinite-slope equation, and their slope-stability model, than did the authors of SINMAP. First, they decided to set the cohesion to zero and eliminate the variable of root strength altogether. Although this approximation is clearly incorrect in some situations, the elimination of root cohesion makes parameterization easier for the user. Root strength varies widely, both spatially and over time. It is especially hard to estimate over large areas, such as at the watershed-scale. They also consider setting the cohesion to zero to be a very conservative thing to do as this maximizes the extent of possible instability within the users study area (Dietrich and Montgomery, 1998). The value of the SHASTAB model, particularly in this study, is that it can be used with fixed parameters over large areas (Dietrich et al., 2001). SHALSTAB is intended to

74

be parameter free when run with the default parameter values. In other words, they wanted a slope stability model where a user did not have to have any available soil data parameters to calibrate the model. When the default version of SHALSTAB is run, only the soil density and the internal soil friction angle can be adjusted. Later versions of SHALSTAB have included an option where the cohesion value and soil depth can be adjusted if necessary (Figure 4.3). The SHALSTAB model needed to be calibrated in a way in which it could be directly comparable with SINMAP. Some of the variables, such as soil density (1922 kg/m) and soil thickness (2 m), could remain the same between the two models. But other soil property values, such as cohesion and soil friction angle, vary widely over the study area. This problem is overcome by SINMAP, which uses a range of values for cohesion, soil friction angle, and transmissivity over recharge. SHALSTAB only uses a single input value for cohesion and soil friction angle. As specific soil data is not available at the watershed-scale, a variety of parameters were tested during several different model runs to find which values best-approximated landslide susceptibility.

75

Figure 4.1: The infinite slope equation as defined by Hammond et al., (1992) and Pack et al., (1998b) where Cr is root cohesion, Cs is soil cohesion, is slope angle, s is soil density, w is the density of water , g is acceleration due to gravity, D is the vertical soil depth, Dw is the vertical height of the water table, and is the internal soil friction angle. In the SINMAP model, the ratio of the vertical soil depth to the vertical soil height is simplifed so that depth is measured perpendicular to the slope (h). (Diagram after Hammond et al., 1992 and Otteman, 2001)

76

Figure 4.2: Default parameters used in the SINMAP model analysis. The values for the gravitational constant and the density of water were not adjusted in this study.

77

Table 4.1: Table of the hydraulic conductivity (K, m/hr), transmissivity (T, m/hr), and T/R (m) values used for each precipitation threshold (50 mm/d, 125 mm/d, 250 mm/d, and 375 mm/d) in the SINMAP analysis. The numbers in blue are the lower bound values while the numbers in red are the upper bound values.
SOIL HYDRAULIC HYDRAULIC T (m2/hr) MAPPING COND (m/hr) COND (m/hr) LOWER UNIT LOWER UPPER T/R T/R T/R T/R T/R T/R T/R T/R T (m2/hr) 50 (mm/d) 50 (mm/d) 125(mm/d) 125(mm/d) 250(mm/d) 250(mm/d) 375(mm/d) 375(mm/d) UPPER UPPER LOWER UPPER LOWER UPPER LOWER UPPER LOWER

NC005 NC006 NC088 NC089 NC090 NC091 NC092 NC093 NC094 NC095 NC096 NC097 NC098 NC099 NC100 NC102 NC103 NC104

0.024 0.014 0.005 0.003 0.005 0.009 0.009 0.01 0.005 0.003 0.008 0.025 0.004 0.01 0.01 0.007 0.005 0.012

0.331 0.152 0.038 0.027 0.115 0.053 0.091 0.032 0.075 0.021 0.027 0.156 0.015 0.081 0.099 0.035 0.025 0.046

0.048 0.028 0.010 0.006 0.010 0.018 0.018 0.020 0.010 0.006 0.016 0.050 0.008 0.020 0.020 0.014 0.010 0.024

0.662 0.304 0.076 0.054 0.230 0.106 0.182 0.064 0.150 0.042 0.054 0.312 0.030 0.162 0.198 0.070 0.050 0.092
AVERAGE

24.0 14.0 5.0 3.0 5.0 9.0 9.0 10.0 5.0 3.0 8.0 25.0 4.0 10.0 10.0 7.0 5.0 12.0 3.0

331.0 152.0 38.0 27.0 115.0 53.0 91.0 32.0 75.0 21.0 27.0 156.0 15.0 81.0 99.0 35.0 25.0 46.0 331.0

9.6 5.6 2.0 1.2 2.0 3.6 3.6 4.0 2.0 1.2 3.2 10.0 1.6 4.0 4.0 2.8 2.0 4.8 1.2

132.4 60.8 15.2 10.8 46.0 21.2 36.4 12.8 30.0 8.4 10.8 62.4 6.0 32.4 39.6 14.0 10.0 18.4 132.4

4.6 2.7 1.0 0.6 1.0 1.7 1.7 1.9 1.0 0.6 1.5 4.8 0.8 1.9 1.9 1.3 1.0 2.3 0.6

63.7 29.2 7.3 5.2 22.1 10.2 17.5 6.2 14.4 4.0 5.2 30.0 2.9 15.6 19.0 6.7 4.8 8.8 63.7

3.1 1.8 0.6 0.4 0.6 1.2 1.2 1.3 0.6 0.4 1.0 3.2 0.5 1.3 1.3 0.9 0.6 1.5 0.4

42.4 19.5 4.9 3.5 14.7 6.8 11.6 4.1 9.6 2.7 3.5 20.0 1.9 10.4 12.7 4.5 3.2 5.9 42.4

78

Figure 4.3: Default values used for SHALSTAB.

79

CHAPTER 5: RESULTS AND DISCUSSION 5.1 Introduction The purpose of this chapter is to provide model results for SINMAP and

SHALSTAB for a variety of parameter ranges, to describe the effectiveness of each model, and to identify the most sensitive model parameters. SINMAP and SHALSTAB results are compared herein to identify weaknesses and strengths inherent to each model. Finally, results will be compared with respect to slope, aspect, geology, and soil type. 5.2 Debris Flow Inventory A total of 142 debris flows have been mapped in the 7330 km study area (Figure

5.1 and Appendix D). All of these slope movements have been located using historical documents and maps, aerial photography, field identification, and data provided by both the North Carolina Geological Survey and the North Carolina Department of Transportation. Landslides were identified in all counties except for Avery County. Of the 42 geologic units that occur within the watershed, debris flows have developed on fourteen. Most of these units are gneiss with a large component of biotite or muscovite (coded Ybgg, Ymg, Zabg, Zatm, Zatb, Zybn on the geologic map). Debris flows also have occurred in quartz diorite (Dqd), metagraywacke (Zatw, Zgs), quartzite (Zsl), siltstone (Zsp), sandstone (Zsr), and slate (Zwc). The greatest number of debris flows (52) have occurred in soil map unit NC095, generally a stony fine-to-sandy loam, and NC093 (46), a stony silt-to-sand loam. The other soils with debris flows are stony-to-gravelly loam or sandy-to-silty loam. All of the soils have a permeability rate that varies from 0.02 to 20 inches per hour (0.05 to 51 cm/h) (U. S. Department of Agriculture, 1998). The affinity of debris flows to these

80

coarse-grained soils may be attributed to rapid infiltration through these soils. In the thin soils of the Southern Appalachians, soil pore pressure can quickly increase as groundwater builds above the impermeable bedrock, increasing the risk of slope failure. The gradient and aspect of the debris-flow locations have been calculated using the 10-meter DEM (with ArcGIS Spatial Analyst extension), given that the 30-meter DEM seems to underestimate slope (Zhang et al, 1999). For the entire watershed, slope varies from 0 to 74 degrees. Debris flows have occurred on slopes ranging from 10 to 50 degrees, although 88% of the slopes are at least 20 degrees. The average slope on which debris flows have initiated is 28 degrees. The aspect slope angle, as calculated in ArcGIS, is the compass direction towards which a slope faces. The "aspects" of the debris-flow headscarps occur in diverse directions, but show an affinity toward the east (32), southeast (23), southwest (21), and south (21). East-facing slopes receive only morning sunlight during the winter and thus have higher soil moisture than south- and west-facing slopes. During high rainfall events, this leads to higher levels of antecedent moisture, faster saturation, and greater soil porewater pressure, leading to an increased incidence of slope failure. The incidence of debris flows on south-facing slopes may be due to the direction of storm tracks over the watershed. Several of the hurricanes that have tracked over western North Carolina have come from south, out of the Gulf of Mexico (see Figure 3.4). 5.3 SINMAP Results Using both a 30-meter DEM and 10-meter DEM, the SINMAP program has been

used to derive stability-index maps for four precipitation thresholds, i.e., 50 mm/d, 125 mm/d, 250 mm/d, and 375 mm/d. The parameters used in each SINMAP model run are

81

summarized in Table 5.1. Although there is an option in SINMAP to create multiple calibration regions, usually based on the properties of individual soil types, only one general calibration region has been used in this study. This was deemed appropriate due to the general nature of the soil data and the regional scale of the study. SINMAP defines six different stability-class definitions based upon the stability index (SI) (Table 5.2). The terms stable, moderately stable, and quasi-stable, are used to represent areas that should not fail under the most conservative input parameters. The areas modeled as having lower threshold and upper threshold, are those that respectively have a <50% or >50% probability of instability. Areas defined as defended slopes are unstable throughout the range of the specified parameter. Where these slopes occur in the field, they are only stable due to factors not modeled by SINMAP. For example, they may be bedrock outcrop (Pack et al., 1998b). Parameters are adjusted so that the resulting map captures the maximum amount of observed landslides in regions with a low SI (stability index), while minimizing the spatial extent of low-stability regions (Pack et al., 2002). Initially, SINMAP model runs for this study involved using 30-meter DEMs, default parameters, and the four aforementioned precipitation thresholds. At the outset of this study, only 30-meter DEMs were publicly available for western North Carolina. In the summer of 2004, 10-meter DEMs also became available. As a result, the SINMAP model runs were completed at both scales to see if the coarseness of the 30-meter DEM seriously affected the accuracy of the stability maps produced (Figure 5.2).

82

5.3.1

Results 30-meter DEM The default SINMAP values seem to underestimate instability in the watershed.

For the 30-meter DEM, the model predicts that 52% (73) of the observed debris flows would have occurred in the lower threshold for instability whereas none were found in the upper and defended stability classes (Figure 5.3). All of other 67 debris flows (48%) occurred in the more stable stability classes. In the next SINMAP calculation, a precipitation threshold of 50 mm/d was used with the 30-meter DEM. Soil friction angle, soil density, cohesion and T/R were adjusted (Table 5.1). Under these conditions, 93.6% (131) of the inventoried debris flows are predicted to occur in the unstable classes, but only one of these was predicted to fail unconditionally. In this model run, 62.7% (4431.8 km) of the study area is predicted to be unstable. Thresholds of 125 mm/d, 250 mm/d, and 375 mm/d were subsequently used with SINMAP. For all three simulated rain events, 131 observed debris flows occurred in unstable zones whereas 9 slides (6.4%) were predicted to form in stable zones. Both the total area of instability and the area predicted to be defended increased slightly as recharge increased. For a precipitation event of 125mm/d, the unstable zones accounted for 4418.1 km (62.6%) of the study area while 28.2 km (0.4) of the watershed was considered unconditionally unstable (Figure 5.4). The defended class contains four (2.9%) of the inventoried debris flows. A 250 mm/d rainfall event is predicted to cause 4434.4 km (62.8%) of the watershed to be unstable. The area considered defended increases in extent slightly to 53.6 km (0.8%) and includes seven (5%) of the slides. For the heaviest modeled recharge event, i.e., 375 mm/d, 4442.2 km (62.8%) of the area is predicted to be unstable; the

83

defended zone increases to 61.4 km (0.9%) and includes eight (5.7%) of the debris flows. In the final calculation, the low value for dimensionless cohesion was increased slightly from zero to 0.1 for a recharge event of 125 mm/d, with all other properties remaining the same. This should better model the effects of root cohesion in the watershed, a factor which often has a strong effect on slope stability, even during heavy precipitation. With cohesion increased, the area of predicted unstable land decreased to 3170.6 km or 44.6% of the watershed, but still contained 124 (88.6%) of the inventoried debris flows. Increased cohesion also increased the stability zones so that 35.7% of the study area (2534.7 km) became classified as unconditionally stable. 5.3.2 Results 10-meter DEM When comparing the area of the 10-meter grid calculated by SINMAP to the 30-

meter grid, the total area of the 10-meter grid is less than the 30-meter grid (Appendix E). The 10-meter SINMAP model run seems to have assigned NO DATA values instead of a stability index to many cells in the raster. Moreover, the stability index classes captured only 122 of the 142 debris flows used in the parameter run. No explanation or repair has yet been discovered for this problem. The default results for the 10-meter DEM differ slightly from 30-meter results but still underestimate instability (Figure 5.5). The same number of debris flows (73) occurred in unstable stability classes, accounting for 18.4% of the study area (1274.3 km). Four of these mass-wasting events were captured in the upper threshold for instability while all of the others occur in the lower threshold class. Unconditionally stable areas were predicted in 53% of the area (3673.9 km).

84

As with the 30-meter DEM model, the results for the four precipitation thresholds are very similar. The extent of the stable, moderately stable, and quasi-stable zones (2409 km) and the number of debris flows (7) occurring within these classes are the same for all four recharge events (Figure 5.6). Only the lower, upper, and defended stability classes change slightly for each precipitation threshold, shifting to a more unstable class as the recharge increases. The number of debris flows that fall in the unstable classes (115) also remains the same for all four recharge events. Compared to the 30-meter DEM, a larger portion of the watershed is predicted to be unconditionally unstable in the 10-meter DEM. For the 50 mm/d threshold, 24.8 km of the area is predicted to be defended although no debris flows fall into this zone. For 125 mm/d of recharge, the defended zone increases to 70.8 km (1.0%) and includes eight (6.6%) slides. In the calculation for 250 mm/d and 375 mm/d recharge, the unconditionally unstable area increases to 126.4 km and 138.8 km, respectively, and contains 12 (9.8%) of the debris flows. As with the 30-meter DEM, when the value for dimensionless cohesion is increased (0.1), the areas of instability decrease while stability increases. Of the 122 inventoried debris flows, 108 (88.5%) fall within the three unstable zones. This is equivalent to 3278.7 km (47.7%) of the watershed. More of the watershed falls within the defended stability zones (50.2 km) than in the 30-meter DEM, whereas 33.6% of the area is predicted to be unconditionally stable (2308.1 km). 5.3.3 SINMAP Interpretation In both the 30-meter and 10-meter DEM, SINMAP predicts that the most stable

areas can be found in the flattest regions, e.g., along floodplains, independent of the

85

parameters used. The areas of greatest instability also match well with the steepest terrain and generally follow steep ridgelines. Overall, the model does well, accurately modeling approximately 94% of the inventoried debris flows in unstable zones with an SI less than 1.0 for both DEM scales and for all four precipitation thresholds. With a slightly increased cohesion value, accuracy decreases to about 88%, but landslide density increases as the model minimizes the extent of the areas mapped as unstable. The 10-meter DEM models a slightly greater percentage of instability in the watershed than the 30-meter DEM. Slopes that are derived from DEMs vary with the spatial resolution, becoming lower at larger pixel sizes (Zhang et al., 1999). Given the coarser resolution of the 30-meter DEM, the elevation data is more generalized and slopes are typically underestimated. Using the 30-meter DEM, SINMAP could not predict small areas of increased instability along narrow ridges and valleys. The greater resolution of the 10-meter DEM allows for better prediction of unstable areas (2.7% improvement) and a slightly greater percentage of inventoried debris flows within the three unstable classification zones (an 0.7% improvement). In this study, SINMAP is not particularly sensitive to changes in the value for recharge. Even with significant changes in the precipitation threshold, there was little change in the predicated areas of instability. This is unexpected because one of the most important factors in triggering debris flows surely is the rate of recharge. SINMAP was more sensitive to changes in the values for dimensionless cohesion, a value increased minimally to model the possible effects of root cohesion. A shallow value for soil thickness (2 m) has been used in the calculation of transmissivity, reducing the resulting value for T/R. SINMAP was developed initially for

86

British Columbia so the default parameters reflect representative soil data for that coolclimate area (Pack et al., 1998b). Model calibration was based on the thick packages of coarse subangular till and colluvium found in British Columbia. Even with low values of hydraulic conductivity (K), a large value for soil thickness (b) would result in a greater transmissivity value. Deeper soils occur in the study areas of Pack et al. (1998b) than in the present study area, hence their use of T/R default values of 2000 m and 3000 m. Given these default values with a 2-meter soil thickness and a moderate (125 mm/d) recharge rate, the hydraulic conductivity lies between 5.21 m/hr and 7.81 m/hr. These rates are representative of well-sorted sand and gravel but are unrealistic for the French Broad Watershed (Fetter, 1994). Clearly, there are other factors at work in the Southern Appalachians that trigger debris flows, factors that are not yet taken into account by SINMAP. 5.4 SHALSTAB Results As with the SINMAP model, SHALSTAB was used to model instability using

both a 10-meter and 30-meter DEM. Due to limitations in the SHALSTAB program, only Haywood County was run in the model (see discussion in Chapter 3). Haywood County was chosen for its rugged topography and the dispersed nature of the inventoried debris flows in the county. Slopes range from 0 to 67 with a median of 20. Some of the steepest terrain in the French Broad Watershed is located in the northwestern corner of the county, within the Great Smoky Mountains National Park. The output of SHALSTAB is the log of the ratio of effective precipitation (q) to soil transmissivity (T), a value that can range from 10 to 10. The low values of log (q/T) are the upper bounds of the model, the unconditionally unstable or chronic instability

87

condition. Consequently, if the tangent of the slope () is greater than or equal to the soil friction angle (), the slope will be unstable, even under dry conditions (Dietrich and Real de Asua, 1998). High values of log (q/T) are the SHALSTAB lower bounds, the unconditionally stable condition. Shallow slopes do not allow for high enough pore pressure in the soil, even during full saturation, and are rarely unstable (Table 5.3). For every grid cell, SHALSTAB calculates the amount of critical effective precipitation necessary to trigger pore-pressure-induced instability, at a constant rate of transmissivity. Areas with lower values are considered more unstable because they require less precipitation to cause them to fail than do areas with higher values (Table 5.4). SHALSTAB requires that landslide locations be supplied to verify the model results. Instead of simple point locations, polygonal shapes of each landslide scar are necessary for this model. In the case of Haywood County, 23 landslide polygons were interpreted and digitized from digital orthophotos and DEMs using ArcGIS 3-D Analyst. The landslide location has been placed within a stability class based upon the lowest q/T value that it intersects. The authors of SHALSTAB consider the model to be successful if the majority of landslides occur in grid cells with low values of log (q/T) (Dietrich and Montgomery, 1998). The four parameters that can be adjusted in SHALSTAB are soil friction angle, soil density, soil depth, and cohesion. Whereas SINMAP uses a range of input parameters, SHALSTAB allows for only a single variable to be adjusted and run at a time. A number of different tests for parameter sensitivity have been completed using a range of variables like those used in the SINMAP calculations, including the SINMAP

88

upper and lower bounds. The parameters used in each SHALSTAB run are summarized in Table 5.5. 5.4.1 Results 30-meter DEM The results for all parameters run in SHALSTAB with a 30-meter DEM are

summarized in Appendix F. Even though 23 landslide polygons were digitized for the county, only 22 of the landslides were actually calculated by SHALSTAB and used in the final results. As with SINMAP, the SHALSTAB default parameters seem to underestimate instability (Figure 5.7). Seven of the mapped debris flows occur in the unconditionally stable zone, with land characterized by gradients too low to fail even when saturated. Collectively, this comprises 974.9 km (31.82% of the county). All of the other debris flows occur in stability classes between 3.1 and 2.2, approximately 38% (421 km) of the county. No landslides fall into the chronic instability zone, the category where areas are defined as potentially unstable even without the addition of significant rainfall (Dietrich et al., 2001). To estimate which parameter values best predict instability in the study area, the number of cells in each log (q/T) category have been determined for each parameter run. The resulting cumulative frequency (percent area) of the county falling into each instability class is plotted on Figure 5.8 (Dietrich et al., 2001). The cumulative percent of debris-flows-per-instability category also is calculated (Figure 5.9). The most unstable combination of parameters is for the SINMAP lower bound, i.e., a soil friction angle of 26 and zero cohesion. The difference between this set of parameters and the others is that there is a greater incidence of observed mass-wasting

89

(20) assigned to the chronic, the <3.1, and the -3.1 - -2.8 categories. This is equivalent to 979.3 km or 59.7% of the county. The next test involved increasing cohesion slightly to 2000 (N/m); this value is equivalent to a dimensionless cohesion value of 0.1 as used in SINMAP. For soil friction angles of 26 and 35, this increase in cohesion decreased the cumulative percentage of the area and number of landslides predicted to occur in unstable areas. For log (q/T) values less than 2.8, the area decreased by 13% given a soil friction of 26 and 12% with a soil friction angle of 35. Strangely, for a soil friction angle of 45, there was no change in the calculated results for values of cohesion between zero and 2000 (N/m). This seems to indicate that SHALSTAB is more sensitive to changes in soil-friction angle than soil cohesion. The most stable parameters were equivalent to the SINMAP upper bound, i.e., a soil friction angle of 45 and cohesion equal to 9427.41 N/m. This high cohesion number is equal to a dimensionless cohesion value of 0.25 when the soil depth is 2 m and soil density is 1922 kg/m. All of the landslides and nearly the entire area (99.92%) fell into the stable category. 5.4.2 Results 10-meter DEM A benefit of the SHALSTAB program is that the model does not lose grid cells,

like SINMAP, during individual calculations. There was a slight loss of grid-cell area from the 30-meter to the 10-meter DEMs, but not nearly as much as with the SINMAP model. The total area used in the 30-meter model was 1639.9 km whereas in the 10meter model, the area was 1642.2 km, a difference of only 1%. This is probably due to

90

the clipping procedure in ArcInfo where some pixels were excluded from the final clipped DEM in the 30-meter scale due to the coarse nature of the data. Appendix F summarizes results for a 10-meter DEM and each of the tested parameters. The default parameters seem to underestimate instability in the county, even though more of the mapped landslides are captured in higher stability classes. Only two landslides (8.8%) fall into the stable class but the area is slightly smaller than in the 30meter model run, i.e., 870.1 km or 53% of the county. The rest of the debris flows (21) are contained in the stability classes between 3.1 to 2.2 that encompass 37% (615.7 km) of Haywood County. As with the 30-meter DEM, a variety of parameters were tested with the 10-meter DEM and both the cumulative percent of the county area and the area of landslides were calculated (Figure 5.10 and Figure 5.11). The smaller grid size of the 10-meter DEM increases the predicted extent of unstable ground in these latter calculations. The lower bound values (26, 0) were found to contain the most debris flows in the category of chronic instability: 20 slides were assigned to an area of 534.8 km (32.6% of the county). Slightly less of the area, 23%, is predicted to be unconditionally stable. This is a difference of only 3% from that defined by the 30-meter DEM. As with the 30-meter DEM, the influence of cohesion on the model was tested by increasing the value slightly from zero to 2000 N/m for each soil friction angle (26, 35, 45). For log (q/T) values less than 2.8, the area decreased by 12.7% with a soil-friction of 26, 10.8% with a soil-friction angle of 35, and 3.2% for a soil-friction angle of 45. Again with the 10-meter DEM, SHALSTAB proved to be more sensitive to changes in cohesion than in soil-friction angle.

91

The upper bound parameters, 45 and 9427.41 N/m calculate the most stable

land. A majority of the debris flows (21) fell into the stable category that comprises 1633.7 km of the county. But the areas of instability were also increased by 7.2% from those found by the 30-meter DEM. Two slides occurred in the -3.1 -2.8 and -2.5 - 2.2instability classes and unstable zones with a log (q/T) of greater than 2.2 make up 5.4 km of the county (8.5%). 5.4.3 SHALSTAB Interpretation Like SINMAP, SHALSTAB is more sensitive to changes to some parameters than

others. A significant increase in cohesion causes more of the study area to be designated as unconditionally stable, but overall the model is less sensitive to this parameter. SHALSTAB seems to be most sensitive to changes in soil-friction angle. With only a 9 degree increase in friction angle from 26 to 35, the total extent of predicted instability, for a log (q/T) less than 2.8, decreases by approximately 29%. Unconditionally stable areas increase by around 20%. SHALSTAB calculates the unstable terrain depending on the tangent value of slope. Given that a greater percentage of a county has a lower gradient when lower soil-friction angle values are used in the model runs, larger regions will be calculated as having a potential for instability. Accurate topography, as determined from the DEM, plays an important part in the SHALSTAB calculation. The authors of SHALSTAB based their interpretation of the success or failure of the model on the mapped log (q/T) results. The model is considered successful if the majority of landslides are captured in grid cells with low values of log (q/T) (Dietrich and Montgomery, 1998). However, if a model considers a large portion of an area to be unstable even though historic mass-wasting is restricted to a small portion of the terrain,

92

the interpretation of model effectiveness is based upon the potential for future masswasting in areas that have not yet failed (Guimaraes et al., 2003). This interpretation is necessary with the lowest tested values, a soil-friction angle of 26o and cohesion of zero. Using these parameters, SHALSTAB accurately models 90.9% to 100% of the mapped debris flows in the upper three instability categories (log (q/T) less than 2.8). Despite the fact that these low values could be interpreted as a worst-case scenario for debris-flow initiation, they seem to overestimate the extent of instability in Haywood County (Figure 5.12). If one compares the debris-flow density for each instability category, mid-range parameters capture the most landslides in the smallest area. Using the 10-meter DEM with a 35 soil-friction angle and soil cohesion value of 2000 N/m, six debris flows are captured in the chronic instability class within an area of 51.9 km (3.2% of total area). This is a landslide density of 0.116 slides per square kilometer, but an accuracy of only 26% for the chronic instability region, and 60.9% for the three highest instability classes (Figure 5.13). All of these debris-flow scars have been interpreted from DOQs, but field verification of the landslides may improve the modeled results. SHALSTAB is very sensitive to the accuracy of mapped debris-flow locations, requiring that the standard for mapping be higher than normal, particularly when 10-meter or higher-resolution DEMs are used (Dietrich and Montgomery, 1998b). 5.5 SINMAP vs. SHALSTAB Only two previous studies have compared the results from SINMAP directly with

those from SHALSTAB (Dietrich et al., 2001; Chinnayakanahalli et al., 2003). A direct

93

comparison of the relative performance of the programs is difficult because they calculate completely different indices to quantify instability. In order to compare them, the SINMAP stability index and the SHALSTAB log (q/T) must be transformed into an analogous format. This can be accomplished using the Spatial Analyst extension of ArcGIS. For direct comparison, the parameters used in SINMAP and SHALSTAB must be identical. A moderate soil-friction angle of 35, a zero value for cohesion, and a soil density 1922 kg/m were chosen. Transmissivity and recharge were also kept at a constant rate of 1 m/d and 125 mm/d, respectively, which set the SINMAP T/R value equal to 200 m. The SHALSTAB log (q/T) value for these parameters is equal to 2.28. This procedure is similar to the one used by Dietrich et al. (2001). Figure 5.14 shows a comparison of the two programs for the same area in Haywood County using identical parameters. Both programs calculated 705.4 km (49%) of Haywood County to be unstable. SHALSTAB calculated slightly more of the county as unstable (870.3 km or 61%) than SINMAP (740.5 km or 52%). Visually, the SHALSTAB results appear more clustered, while the SINMAP results seem more scattered. These differences are due to the different ways that the models calculate area and slope (Dietrich et al., 2001). Overall, the calculated results for both programs are strikingly similar i.e., both programs predicted 81-95% of the same area to be unstable (Figure 5.15). 5.6 Geology and Soils Comparing SINMAP and SHALSTAB results to the geology and soil units in the

entire watershed and just Haywood County reveals some similarities. Independent of scale, both SINMAP and SHALSTAB predict the most unstable geologic unit to be

94

Zchs, i.e., slate of the Copperhill Formation (Figure 5.16). This unit is a graphitic to sulfidic slate-to-phyllite found in the Great Smoky Mountains National Park in Haywood County (North Carolina Geological Survey, 1985). Rock descriptions of the Copperhill Formation are scarce but in the geologic map of the Great Smoky Mountains of Hadley and Goldsmith (1963), "Zchs" corresponds to a map unit called pa, i.e., the Anakeesta Formation. According to Merschat (pers. comm., 2005), a geologist with the North Carolina Geologic Survey, the slate of the Copperhill and the Anakeesta are separate formations but are nearly identical. The Anakeesta is part of the Smoky Mountain Group and the Ocoee Supergroup. It is a pyritic dark, fine-grained argillaceous rock with interbedded metasiltstone and coarse sandstone (Hadley and Goldsmith, 1963). Both the Anakeesta and the Copperhill formations represent repetitive turbidite sequences at different stratigraphic levels (Merschat, pers. comm., 2005). Topography in this portion of the Great Smoky Mountains National Park is exceptionally narrow and steep with serrate crests and craggy pinnacles, and thin residual soils (Hadley and Goldsmith, 1963). Near the bottom of the Anakeesta, sulfidic argillaceous rock commonly intertongues with sandstone like that within the underlying Thunderhead sandstone. The Thunderhead, a member of the Smoky Mountain Group (Zgs), also can be extremely unstable based on the SINMAP and SHALSTAB results (Figure 5.16). It is also known to produce boulder fields and slope in western North Carolina and Tennessee. In the southeast, and western North Carolina, chemical weathering may play a significant role in slope instability and the development of mass wasting (Grant, 1988). The slate of the Copperhill Formation, being sulfidic, would be extremely prone to

95

chemical weathering and could account, in part, for the steep weathered topography and instability associated with this unit. Transverse and strike jointing also cause angular cliff exposures in the Anakeesta and Thunderhead sandstone in the Great Smoky Mountains National Park (Hadley and Goldsmith, 1963). The most stable geologic units in the French Broad Watershed were found to be that coded bz on the Geologic Map of North Carolina (1985), i.e., the metamorphosed rocks of the Brevard Fault Zone, and that coded CZtp, i.e., porphyroblastic gneiss of the Sauratown Formation. Both of these units are located in the southeastern portion of the watershed and lie within the geomorphologic region known as the Western Piedmont. The Brevard Fault zone seems to form a dividing line between moderately unstable rock and stable rock (Figure 5.16). Both SINMAP and SHALSTAB predict the most unstable soil unit in the watershed to be that coded NC104, i.e., a general soil group that occurs in northern Haywood County, generally overlying the Copperhill Formation. This soil unit is principally composed of the Tanasee (21%), Burton (19%), Oconaluftee (16%), and Porters (14%) soil series (U.S. Department of Agriculture, 1998) (see Appendix B). The Tanasee soil series is sandy loam derived from colluvium and often forms toe slopes, fans and benches on coves in the high elevations of the Appalachians. The Burton and Porters series are a sandy-to-fine loam derived from residuum on ridge and side slopes. The Ocanaluftee is a channery loam, also derived from residuum, or from slate, phyllite, or metasandstone. All of these soils are well drained and have moderate-to-high permeability and are derived from weathering and erosion of the underlying soil units

96

(U.S. Department of Agriculture, 2005). All of the stable soil units are located on the flat topography along the floodplain of the French Broad and Pigeon Rivers (Figure 5.17). 5.7 Jointing, Fracturing and Foliation One factor that neither SINMAP nor SHALSTAB takes into consideration is the

influence of geologic structure on groundwater flow. Groundwater flow through a fracture is notoriously difficult to model because the rate of transmissivity is strongly controlled by the size of the fracture aperture and the connectivity of the fracture network (Renshaw, 2000). Nonetheless, flow through joints and fractures, and over bedrock foliation planes parallel to the dip slope, may play a significant role in triggering debris flows in western North Carolina (Figure 5.18). In the Great Smoky Mountains National Park, it was noted that the heads of debris flows originate at the intersection of cleavage and joints or beds, or on the cleavage or bedding plane (Southworth et al, 2003). Intersecting joints and fractures can cause groundwater flow to become concentrated in topographic hollows. During a heavy rainfall event, these hollows could periodically be flushed out of weathered colluvial material, triggering a debris flow. Foliation planes provide a smooth surface on which sliding may initiate, particularly if the dip angle for bedding is close to the dip of the foliation.

97

Figure 5.1: Landslide inventory map for the French Broad Watershed. The cluster of location points in southern Buncombe County is due to the extensive mapping of 1977 debris flows done by Pomeroy (1991) and Otteman (2001).

98

Figure 5.2: A comparison of 30-meter and 10-meter DEMs in Haywood County, North Carolina. In the coarser 30-meter DEM, there is a great deal of pixelation. Lower resolution can lead to an underestimation of both the slope and instability in the study area.

Table 5.1: The parameters used in all of the SINMAP model runs.

Precipitation Threshold Default 50mm 125mm 125mm 250mm 375mm

Soil Density 2000 1922 1922 1922 1922 1922

T/R Max. 3000 331 132.4 132.4 63.7 42.4

T/R Min. 2000 3 1.2 1.2 0.6 0.4

Cohesion Max. 0 0 0 .10 0 0

Cohesion Min. .25 .25 .25 .25 .25 .25

Max. 30 26 26 26 26 26

Min. 45 45 45 45 45 45

99

Table 5.2: SINMAP stability index definitions (Pack et al., 1998b).

Condition SI > 1.5

Class 1

Predicted State Stable Slope Zone Moderately stable slope zone Quasi-stable slope zone Lower threshold slope zone Upper threshold slope zone Defended slope zone

1.5 > SI > 1.25 1.25 > SI > 1.0 1.0 > SI > 0.5 0.5 > SI > 0.0 0.0 > SI

3 4 5 6

Possible Influence of Factor Not Modeled Range cannot Significant model instability destabilizing factors required for instability Range cannot Moderate model instability destabilizing factors required for instability Range cannot Minor destabilizing model instability factors could lead to instability Pessimistic half Destabilizing factors of range required are not required for for instability instability Optimistic half of Stabilizing factors range required for may be responsible stability for stability Range cannot Stabilizing factors model stability are required for stability

Parameter Range

Map Color green

blue

yellow pink red tan

100

Figure 5.3: SINMAP results for a 30-meter DEM and using default parameters.

101

Figure 5.4: SINMAP results for 125 mm/d recharge and using a 30-meter DEM.

102

Figure 5.5: SINMAP results for a 10-meter DEM and using default parameters.

103

Figure 5.6: SINMAP results for 125 mm/d recharge using a 10-meter DEM.

Table 5.3: Mapped instability classes used in the SHALSTAB model analysis.

Log (q/T) Interval Chronic Instability < - 3.1 -3.1 - -2.8 -2.8 - -2.5 -2.5 - -2.2 >-2.2 Stable

Color Dark red Red Pink Orange Yellow Green Blue

104

Table 5.4: Table comparing q/T and log (q/T) values and the precipitation rate required to initiate instability for soils with a transmissivity of 65 md and 17 m/d (after Dietrich and Asua, 1998).

q/T (1/m) .00079 .00158 .00316 .00833 .01266

log (q/T) (1/m) -3.1 -2.8 -2.5 -2.2 -1.9

Precip for T = 65 m/d (mm/d) 52 103 206 410 818

Precip for T = 17 m/d (mm/d) 14 27 54 103 214

Table 5.5: Parameters used in the SHALSTAB model runs.

Soil Friction Density Angle (kg/m) *45 1700 **45 1922 45 1922 45 1922 35 1922 35 1922 26 1922 ***26 1922 * SHALSTAB default values ** SINMAP upper bounded values *** SINMAP lower bounded values

Soil Depth (m) 1 2 2 2 2 2 2 2

Cohesion (N/m) 0 9427.41 2000 0 2000 0 2000 0

105

Figure 5.7: The map of log (q/T) for Haywood County produced using default parameters in SHALSTAB.

106

Figure 5.8 Cumulative percent of Haywood County found in each log (q/T) category for a variety of soil parameters for the 30-meter DEM (after Dietrich et al., 2001). In the legend, the first number is the degree if soil friction angle and the second number is the amount of cohesion (N/m).

Figure 5.9: Cumulative percent of debris flows for each log (q/T) instability category for a variety of soil parameters for a 30-meter DEM (after Dietrich et al., 2001).

107

Figure 5.10: Cumulative percent of the area if Haywood County for each log (q/T) instability category for a variety of soil parameters for the 10-meter DEM (after Dietrich et al., 2001).

Figure 5.11: Cumulative percent of debris flows for each log (q/T) instability category for a variety of soil parameters for a 10-meter DEM (after Dietrich et al., 2001).

108

Figure 5.12: Mapped log (q/T) results for a 10-meter DEM of Haywood County. The parameters used for this model run are essentially that same as those used in the SINMAP lower bounds: 26 soil friction angle, and zero cohesion.

109

Figure 5.13: SHALSTAB results for a soil friction angle of 35 and cohesion of 2000 N/m.

110

Figure 5.14: Comparison of the output of SINMAP and SHALSTAB for 125 mm of recharge, 35 soil friction angle, 1922 kg/m soil density, and zero soil cohesion for a location in Haywood County. The red areas are calculated as unstable by both programs whereas the gray areas are calculated to be stable. Visually the SHALSTAB results seem to cluster better while the SINMAP results are more scattered. Overall, the results are very similar.

111

Figure 5.15: Results for SINMAP and SHALSTAB. Red areas are the areas calculated by both models as being unstable. The purple areas were only calculated by SHALSTAB and the green areas were only calculated by SINMAP.

112

Figure 5.16: The mean stability index for each geologic unit in the French Broad Watershed. The most unstable unit, Zchs, is located in the northwestern portion of Haywood County. The most stable units, bz and Ctzp, are located in the southeastern portion of the study area.

113

Figure 5.17: The mean stability index for each soil unit in the French Broad Watershed. The most unstable soil unit in the watershed is NC104, which is located in western Haywood County. The most stable soils are located around the floodplains of the Pigeon and French Broad Rivers.

114

Figure 5.18: Picture taken in October 2003 along the Blue Ridge Parkway, near Mt. Mitchell. Even during this light rain event, water is pouring out from fractures in the rock.

115

CHAPTER 6: CONCLUSIONS Debris flows and other forms of mass-wasting in western North Carolina are a destructive force that deserves more attention from both local and federal-level planners and citizens living in these mountainous areas. Debris flows are difficult to predict and have such a low frequency that it is rare for them to recur in the same area within the average life span of area residents. Unlike a flood-hazard map, a landslide-hazard map is inherently imprecise with regard to the extent, location, and timing of future masswasting. After the tragic results of the September 2004 hurricane season, North Carolina Senate Bill 7, the Hurricane Recovery Act of 2005 was written and passed (available online at http://www.ncga.state.nc.us/). It recognizes the need for better understanding of debris flows and provides funds for disaster relief and to identify areas of potential slope instability. This study has incorporated historical rainfall data, geology, soil types, and geomorphology to help predict debris flows in the French Broad Watershed. A combination of factors have been found to trigger a debris flow: 1. Topography is the most important factor in generating a debris flow. Slopes with historic incidences of mass-wasting have averaged 28 degrees in this study area but range up to 50 degrees. 2. Intense precipitation, generally greater than 125 mm/d, tends to increase the probability of debris flows. This is especially true when a second large storm system travels over the area soon after the first. The first storm may thoroughly saturate the soil, producing high soil-antecedent-moisture conditions for the second storm. The second storm then increases soil pore

116

pressure, decreases cohesion, and may induce slope failure. Steeply sloping areas may be of potential concern under these certain weather conditions and should be closely monitored. 3. Both SINMAP and SHALSTAB predict that certain geologic units are more prone to failure than others. Sulfidic shale beds of the Copperhill Formation in the Great Smoky Mountains National Park are particularly susceptible, largely due to their steep topography, thin soil cover, and vulnerability to chemical weathering. 4. Both SINMAP and SHALSTAB predict that certain soil units are more prone to failure than others. Particularly vulnerable to mass-wasting are those that form at high elevations on steep slopes, are well drained, have a moderate to rapid rate of permeability, and develop over less permeable bedrock. 5. The movement of groundwater may be concentrated through joints, fractures, and foliation, quickly increasing pore pressure and decreasing cohesion and friction between the soil and underlying bedrock. Areas with thin soils overlying fractured bedrock may be more prone to slope instability. SINMAP and SHALSTAB have both proven adequately accurate in the prediction of slope instability on the regional scale of the French Broad Watershed. Both ArcView and ArcGIS provide useful visual representations of landslide-hazard maps that may readily be updated as new information becomes available. The performance of both programs may be summarized as follows: 1. Both SINMAP and SHALSTAB are highly dependent on the quality of the DEM. The coarser, 30-meter USGS DEM data generalizes and underestimates

117

both slope and relief. A finer 10-meter grid maximizes the overall extent of predicted instability, but significantly increases computer-processing time. Extremely large 10-meter DEMs, covering more than one county, may cause either computer program to abort. 2. The modeled results for the default SINMAP and SHALSTAB parameter values underestimate the extent of instability in the study area. 3. A SINMAP stability index of less than 1.0, designating instability, has been calculated for 88% -to- 94% of the inventoried landslides whereas the corresponding stability threshold in SHALSTAB has designated 91% to 100% of the mapped debris flows. The SHALSTAB threshold is a log (q/T) value less than 2.8 for a soil-friction angle of 26 and cohesion of zero. Overall, these results seem to over-predict the areas of instability. 4. The results from both models are similar, calculating 81-95% of the same area of Haywood County as unstable. Visually, the SINMAP results seem dispersed whereas the SHALSTAB results are more clustered. The advantages and disadvantages of SINMAP and SHALSTAB are discussed below and summarized in Table 6.1. Neither model was conclusively found to predict instability better than the other and each should be used with caution and with the most accurate input data available. 1. Some of the advantages of the SINMAP model are as follows: SINMAP has a range of values that may be entered into the model, allowing for uncertainty in soil parameters that are often difficult to measure.

118

It can run large DEMs, e.g., watershed size, without aborting the program or causing errors.

Recharge values are used in the calculation of the T/R values so these precipitation thresholds can be tested.

2. Some of the disadvantages of the SINMAP model are as follows: During some model runs, some grid cells calculated as NO DATA, effectively changing the area calculated by the model. The model was created for study areas that had thick soil packages and high transmissivity rates, neither of which characterize the French Broad Watershed. 3. Some of the advantages of the SHALSTAB model are as follows: When run with only the default values, the user needs little knowledge of the actual soil parameters, or even mapped landslides, to produce a map of relative instability. No grid cells are lost in the SHATSTAB model calculation, making a direct comparison of total watershed area between the 10-meter and 30meter DEM easier and more accurate. 4. Some of the disadvantages of the SHALSTAB model are as follows: To obtain more accurate results, specific soil-parameter data must be introduced into the model, however the model does not allow for a range of values. The stability index is a parameter that is not very useful, i.e., log (q/T).

119

5. Neither model takes into account either antecedent moisture or the effect that geologic structure can have on concentrating groundwater flow. Nonetheless, both of these factors probably have a significant effect on instability. Deterministic slope-stability analytical methods, e.g., SINMAP and SHALSTAB, are generally more useful in areas where ground conditions are fairly uniform throughout the study area (Dai et al., 2002). Due to the scale of this study, assumptions and generalizations about soil parameters had to be made, thus forcing these parameters to be uniform over a large area. From this broad reconnaissance study, the French Broad Watershed offers several avenues for further study: 1. Subsequent studies should use a smaller study area, so that a site-specific soil and geologic analysis may be completed. Precise measurements of hydraulic conductivity (permeability), soil density, soil-friction angle, and soil cohesion increase the accuracy of the predicted SINMAP and SHALSTAB results. 2. An accurate and detailed debris flow inventory should be constructed when using these models. Verification of the modeled results using actual debrisflow locations is essential. 3. Mapping the extent of ancient debris-flow deposits in the study area could be used to better understand the return interval for such catastrophic events. 4. High-quality (finer-resolution) precipitation data must be acquired. Although generalized rainfall data is useful, debris flows are often generated by intense, localized downpours. 5. It is recommended that the user obtain the highest-quality digital-elevation data available, given that both SINMAP and SHALSTAB rely upon accurate

120

topographic information. This kind of data would include 10-meter DEMs or smaller and LIDAR data.

Table 6.1: The advantages and disadvantages of SINMAP and SHALSTAB.

SINMAP
Advantages Allows for a range of input values/ recharge thresholds can be tested Factor of safety easy to interpret Disadvantages Requires knowledge of specific soil/recharge parameters Must adjust parameters repeatedly to capture landslides Can run large DEMs NO DATA grid cells/shifting Does not take into account geologic -structure or antecedent moisture

SHALSTAB
Advantages Can be run without knowing soil parameters Does not require recharge values, can back-calculate Retains calculated grid cell area -Disadvantages Does not allow for a range of input values Log (q/T) value difficult to interpret Aborts when using large DEMs Does not take into account geologic structure or antecedent moisture

121

REFERENCES Appt, J., Skaugset, A., Pyles, M., 2002, Discriminating between landslide sites and adjacent terrain using topographic variables: Geological Society of America, Abstracts with Programs, v. 34, no. 5, p. 93. Associated Press, 2004, Frances Remnants Cause Flooding in N.C., viewed September 9, 2004 at http://www.abcnews.com. Badgett, P., Locklear, B., Blaes, J., 2004, September 2004 NC Weather Review, 8 p., viewed January 27, 2005 at http://www.erh.noaa.gov/rah/climate/data/MonthlySummary.Sep.2004.pdf. Ball, J., 2004, Collapsed sections of parkway force travelers to detour: Asheville CitizenTimes, viewed September 14, 2004 at http://www.citizentimes.com/cache/article/print/61363.shtml. Barrett, M., 2004, Old Fort mudslide could be clear by noon; many roads still closed: Asheville Citizen-Times, viewed September 9, 2004 at http://www.citizentimes.com/cache/ article/print/61068.shtml. Behling, R. E., Kite, J, S., Cenderelli, D.A., Stuckenrath, R., 1993, Buried organic-rich sediments in the unglaciated Appalachian Highlands: A Stratigraphic model for finding pre-late Wisconsin paleoenvironmental data: Geological Society of America, Southeastern Section, Abstracts with Programs, v. 25, no. 6, p. 60. Biesecker, M. and Shaffer, J., 2004, Ivan starts fatal slide: The News & Observer, viewed September 20, 2004 at http://www.newsobserver.com/print/sunday/front/vprinter/story/1652067p-7880420c.html. Boyle, J., 2004, WNC on alert for floods: Asheville Citizen-Times, viewed September 8, 2004 at http://www.citizen-times.com/cache/article/print/60934.html. Braun, D. D., 1989, Glacial and periglacial erosion of the Appalachians: Geomorphology, v. 2, p. 233-256. Buol, S.W., Graham, R.C., McDaniel, P.A., Southard, R.J., 2003, Soil Genesis and Classification: Iowa, Iowa State Press, 494 p. Cabe, W. J., 2004, Summary of Peeks Creek Incident Task Force findings, viewed January 27, 2005 at http://www.maconnc.org/ems/peeks%20creek%20summary.pdf. Cantley-Falk, R., 2004, 82-year old woman survives after mountain crashes through home: Asheville Citizen-Times, viewed September 18, 2004 at http://www.citizentimes.com/cache/article/print/61724.shml.

122

Chinnayakanahalli, K., Tarboton, D.G., Pack, R.T., 2003, An objective method for the intercomparison of terrain stability models: San Francisco, Eos Transactions, AGU Fall Meeting, 8-13 December. Clark, G.M., 1987, Debris Slide and debris flow historical events in the Appalachians south of the glacial border in Costa J.E. and Wieczorek G.F., eds, Debris flows/avalanches: process, recognition, and mitigation: Geological Society of America, Reviews in Engineering Geology, v. 7, p. 125-138. Clark, M.G., Ryan, P.T., Drumm, E.C., 1987, Debris slides and debris flows on Anakeesta Ridge, Great Smoky Mountains National Park, Tennessee in Schultz, A. P and Southworth, S.C. eds., Landslides of Eastern North America: Geological Survey Circular 1008, U.S. Dept. of the Interior, U.S. Geological Survey, Denver, CO, p. 18-19. Clark, M.G., and Torbett, C.A., 1987, Block fields, block slopes, and block streams in the Great Smoky Mountains National Park, North Carolina and Tennessee in Schultz, A. P and Southworth, S.C. eds., Landslides of Eastern North America: U.S. Dept. of the Interior, Geological Survey Circular 1008, U.S. Geological Survey, Denver, CO, p. 2021. Clingman, T.L., 1877, Selections from the speeches and writings of the Honorable Thomas L. Clingman of North Carolina with Additions and Explanatory Notes: John Nichols, Raleigh, N.C. Crozier, M.J., 1986, Landslides: causes, consequences, and environment: Croon Helm, Dover, N.H., 252 p. Cruden, D.M. and Varnes, D.J., 1996, Landslide Types and Processes, in Turner A.K., and Schuster, R. J., eds., Landslides: Investigation and Mitigation - Special Report 247: Washington, D.C., Transportation Research Board, National Research Council, p. 36-75. Dai, F.C., Lee, C.F., Ngai, Y.Y., 2002, Landslide risk assessment and management: an overview: Engineering Geology, v. 64, p. 65-87. Daniels, R.B., Buol, S.W., Kleiss, H.J., Ditzler, C.A., 1999, Soil Systems in North Carolina: North Carolina State University, Soil Science Department, Raleigh, NC, 118 p. Dietrich, W. E. and Real de Asua, R., 1998, A validation of the shallow slope stability model, SHALSTAB, in forested lands of Northern California, Dietrich, W.E., Bellugi, D., Real de Asua, R., 2001, Validation of the Shallow Landslide Model, SHALSTAB, for Forest Management, in Wigmosta, M.S. and Burges, S.J. eds., Landuse and Watersheds: Human Influence on Hydrology and Geomorphology in Urban and Forest Areas: Washington, D.C., American Geophysical Union, v. 2, p. 195-227.

123

Dietrich, W.E. and Montgomery, D.R., 1998, SHALSTAB: A digital terrain model for mapping shallow landslide potential, viewed October 23, 2003 at http://socrates.berkeley.edu/ ~geomorph/shalstab/. Eaton, L.S., Kochel, R.C., Howard, A.D., Sherwood, W.C., 1997, Debris flow and stratified slope wash deposits in the central Blue Ridge of Virginia, Geological Society of America, Abstracts with Programs, v. 29, no. 6, p. 410. Easterbrook, D.J., 1999, Surface Processes and Landforms: Macmillan Publishing Company, New York, NY, 546 p. Eschner, A.R. and Patric, J.H., 1982, Debris Avalanches in Eastern Upland Forests: Journal of Forestry, v. 80, p. 343-347. Fetter, C.W., 1994, Applied Hydrogeology: Prentice Hall, Englewood Cliffs, NJ. Glass, F.R., 1981, Unstable Rock Slopes Along Interstate 40 Through Pigeon River Gorge, Haywood County, North Carolina in Proceedings, 32nd Annual Highway Geology Symposium, v. 32: Atlanta, GA, Highway Geology Symposium, 15 p. Graham, R.C., Daniels, R.B., Buol, S.W., 1990, Soil-Geomorphic Relations on the Blue Ridge Front: I. Regolith Types and Slope Processes: Soil Science Society of America Journal, v. 54, p. 1362-1367. Graham, R.C. and Buol, S.W., 1990, Soil-Geomorphic Relations on the Blue Ridge Front: II. Soil Characteristics and Pedogenesis: Soil Science Society of America Journal, v. 54, p. 1367-1377. Grant, W.H., 1988, Debris Avalanches and the Origin of First-Order Streams in Swank, W.T. and Crossley, D.A., eds., Forest hydrology and ecology at Coweeta: New York, N.Y., Springer-Verlag, p. 103-110. Greenway, D.R., 1987, Vegetation and Slope Stability, in Anderson, M.G., and Richards, K.S., eds., Slope Stability: geotechnical engineering and geomorphology: Chichester, New York, John Wiley & Sons Ltd., p. 187-517. Gryta, J.J. and Bartholomew, M.J., 1983, Debris-Avalanche type features in Watauga County North Carolina, in Lewis, S.E., ed., Geological Investigations in the Blue Ridge of Northwestern North Carolina: Carolina Geological Society Field Trip Guidebook: North Carolina Division of Land Resources, p. 1-22. Gryta, J.J. and Bartholomew, M.J., 1987, Frequency and susceptibility of debris avalanching induced by Hurricane Camille in Central Virginia in Schultz, A. P and Southworth, S.C. eds., Landslides of Eastern North America: Geological Survey Circular 1008, U.S. Dept. of the Interior, U.S. Geological Survey, Denver, CO, p. 16-18.

124

Gryta, J.J. and Bartholomew, M.J., 1989, Factors influencing the distribution of debris avalanches associated with the 1969 Hurricane Camille in Nelson County, Virginia in Landslide Processes of the Eastern United States and Puerto Rico, Special Paper Geological Society of America, no. 236, p. 15-28. Guimaraes, R.F., Montgomery, D.R., Greenberg, H.M., Fernandes, N.F., Gomes, R.A.T., Junior, O.A., 2003, Parameterization of soil properties for a model of topographic controls on shallow landsliding: application to Rio de Janeiro: Engineering Geology, v. 69, p. 99-108. Hack, J. T., 1980, Rock control and tectonism; Their importance in shaping the Appalachian Highlands: US Geologic Survey Professional Paper 1126-B, 17 p. Hadley, J.B., and Goldsmith, R., 1963, Geology of the Eastern Great Smoky Mountains, North Carolina and Tennessee: U.S.Geological Survey Professional Paper 349-B, p. Hammond, C., Hall, D., Miller, S., Swetik, P., 1992, Level 1 Stability Analysis (LISA) documentation for version 2.0 Gen. Tech. Rep. INT-285: Ogden, UT, Department of Agriculture, Forest Service, Intermountain Research Station, 190 p. Hatcher, R.D., Jr., 1987, Bedrock geology and regional geologic setting of Coweeta Hydrologic Laboratory in the eastern Blue Ridge in Swank, W.T. and Crossley, D.A., eds., Forest Hydrology and Ecology at Coweeta, p. 81-92. Hatcher, R.D., Jr., Carter M. W., Clark, G.M., Mill, H.H., 1996, Large landslides in western Blue Ridge of TN & NC; normal mass-wasting phenomena, products of late Pleistocene climates or smoking gun for earthquake(s) in East TN?, Geological Society of America, Abstracts with Programs, v. 28, no. 7, pg. 299. Hatcher, R.D., Jr. and Goldberg, S.A., 1994, The Blue Ridge Geologic Province in Horton, J.W. Jr. and Zullo, V.A., eds., The Geology of the Carolinas: The University of Tennessee Press, Knoxville, TN, p. 11-35. Highland, L.M., Ellen, S.D., Christian, S.B., Brown, W.M., 2004, Debris-Flow Hazards in the United States, U.S. Geological Survey Fact Sheet 176-97, viewed November 10, 2003 at http://geohazards.cr.usgs.gov/factsheets/html_files/debrisflow/fs176-97.html. Holmes, J.S., 1917, Some notes on the occurrence of landslides: Journal of Elisha Mitchell Society, v. 33, p. 100-105. Horton, J.W., Jr. and McConnell, K.I., 1994, The Western Piedmont in Horton, J.W., Jr. and Zullo, V.A., eds., The Geology of the Carolinas: The University of Tennessee Press, Knoxville, TN, p. 36-58. Hudgins, J.E., 2000, Tropical cyclones affecting North Carolina since 1586 - an historical perspective: U.S. Dept. of Commerce, National Oceanic and Atmospheric

125

Administration, National Weather Service, Eastern Region Headquarters, Scientific Services Division, Bohemia, N.Y., 83 p. Jackson, J.A., 1997, Glossary of Geology: Alexandria, VA, American Geological Institute, 769 p. Jacobson, R.B., Cron, E.D., McGeehin, J.P., 1989, Slope movements triggered by heavy rainfall, November 3-5, 1985, in Virginia and West Virginia, U.S.A.: Geological Society of America, Special Paper 236, 13 p. Jacobson, R.B., Miller, A.J., Smith, J.A., 1989, The Role of catastrophic events in Central Appalachian landscape evolution, in Gardner, T.W., and Sevon, W.D. eds. Appalachian Geomorphology, v. 2, p. 257-284. Kaya, A., 1991, Rock slope stability analysis along the Pigeon River, Interstate Highway 40 in Western North Carolina: M.S. Thesis, Purdue University, West Lafayette, IN, 127 p. Kochel, R.C., 1984, Quaternary debris avalanche frequency, sedimentology, and stratigraphy in Virginia, Geological Society of America, Abstracts with Programs, v. 16, no. 6, p. 562. Kochel, R.C., 1987, Holocene debris flows in Central Virginia: Geological Society of America, Reviews in Engineering Geology, v. 7, p. 139-155. Kochel, R.C., 1990, Humid fans of the Appalachian Mountains in Rachocki, A.H., and Church, M. eds., Alluvial Fans: A field approach: New York, NY, John Wiley & Sons Ltd., p. 109-129. Kochel, R.C., and Johnson, R.A., 1984, Geomorphology and sedimentology of humidtemperate alluvial fans, central Virginia in Koster, E.H. and Steel, R.J., eds., Sedimentology of gravels and conglomerates: Canadian Society of Petroleum Geologists, v. 10, p. 109-122. Leith, C.J., Fisher, C.P., Deal, C.S., Meyer, H.D., 1965, An investigation of the stability of highway cut slopes and embankments in North Carolina: Raleigh, NC, Engineering Research Department of North Carolina State University, 30 p. Liebens, J. and Schaetzl, R.J., 1997, Relative-age relationships of debris flow deposits in the Southern Blue Ridge, North Carolina: Geomorphology, v. 21, p. 53-67. Lydolph, Paul E., 1985, Weather and Climate: Rowman & Allanheld, Totowa, NJ, 216 p. Maune, D.F, Kopp, S.M., Crawford, C.A., Zervas, C.E., 2001, Introduction in Maune, D.F, ed., Digital elevation model technologies and applications: The DEM users manual: Bethesda, MD, American Society for Photogrammetry and Remote Sensing, p. 1-31.

126

Mazza, N.A. and Wieczorek, G.F., 1997, A volumetric sediment analysis for Kinsey Run debris flow, triggered during an extreme storm event, Madison County, Virginia: Geological Society of America, North Eastern Section, Abstracts with Programs, v. 29, no. 1, p. 64. Mills, H.H., 1982, Piedmont-cove deposits of the Dellwood quadrangle, Great Smoky Mountains, North Carolina, USA: Morphometry: Zeitschrift fuer Geomorphologie, v. 26, no. 2, p. 163-178. Mills, H.H., 2000, On the distribution of old colluvial deposits in the Appalachians: Abstracts with Programs Geological Society of America, v. 32, no. 7, p. 182. Mills, H.H. and Allison, J.B., 1995a, Controls on the variation of fan-surface age in the Blue Ridge Mountains of Haywood County, North Carolina: Physical Geography, v. 15, no. 5, p. 465-480. Mills, H.H. and Allison, J.B., 1995b, Weathering rinds and the evolution of Piedmont slopes in the Southern Blue Ridge Mountains: The Journal of Geology, v. 103, p. 379394. Mills, H.H., Brakenridge, G.R., Jacobson, R.B., Newell, W.L., Pavich, M.J., Pomeroy, J.S., 1987, Appalachian mountains and plateaus, in Graf, W.L, ed., Geomorphic systems of North America: Boulder, CO, Geological Society of America, Centennial Special Volume 2, p. 5-50. Montgomery, D.R. and Dietrich, W.E., 1994, A physically based model for the topographic control on shallow landsliding: Water Resources Research, v. 30, no. 4, p. 1153-1171. Montgomery, D.R., Greenberg, H.M., Laprade, W.T., Nashem, W.D., 2001, Sliding in Seattle: Test of a Model of Shallow Landsliding Potential in an Urban Environment in Wigmosta, M.S. and Burges, S. J. eds., Landuse and Watersheds: Human Influence on Hydrology and Geomorphology in Urban and Forest Areas: Washington, D.C., American Geophysical Union, v. 2, p. 59-73. Morrissey, M.M., Wieczorek, G.F. and Morgan, B.A., 2001, A comparative analysis of hazard models for predicting debris flows in Madison County, Virginia: U.S. Geological Survey Open-File Report 01-0067, 15 p. Myers, E. W., 1902, A study of the southern river floods of May and June, 1901: Engineering News and American Railway Journal, v. 48, p. 102-104. National Weather Service, 2004a, Tropical Weather Summary: Miami, FL, Tropical Prediction Center/National Hurricane Center, viewed January 27, 2005 at http://www.nhc.noaa.gov/archive/2004/tws/MIATWSAT_nov.shtml.

127

National Weather Service, 2004b, Preliminary storm total rainfall amounts for the western Carolinas and northwest Georgiaassociated with Frances, viewed September 23, 2004 at http://www.erh.noaa.gov/gsp/localdat/cases/6-8Sep2004/frances.htm. National Weather Service, 2004c, Rainfall totals from around the area associated with the remnants of Ivan, viewed September 23, 2004 at http://www.erh.noaa.gov/gsp/localdat/cases/16-17Sep2004/ivan.htm. Neary, D.G. and Swift, L.W., 1987, Rainfall thresholds for triggering a debris avalanching event in the Southern Appalachian Mountains: Geological Society of America, Reviews in Engineering Geology, v. 7, p. 81-92. Neary, D.G., Swift, L.W., Jr., Manning, D.M., Burns, R.G., 1986, Debris avalanching in the Southern Appalachians: An Influence on Forest Soil Formation: Soil Science Society of America Journal, v. 50, p. 465-471. NCDC Climate Data Online, 2003, Monthly surface data, selected climate divisions in North Carolina, viewed October 23, 2003 at http://cdo.ncdc.noaa.gov/pls/plclimprod. North Carolina Geological Survey, 1985, State Geologic Map of North Carolina: Raleigh, North Carolina Geological Survey, scale 1:500,000. North Carolina Geological Survey, 2004a, Landslides from Hurricane Frances, viewed January 23, 2005 at http://www.geology.enr.state.nc.us/Geologic_hazards_Landslides_Hurricane_Frances_1/ Landslides_Hurricane_Frances.htm. North Carolina Geological Survey, 2004b, Landslides from Hurricane Ivan, viewed January 23, 2005 at http://www.geology.enr.state.nc.us/Geologic_Hazards_Landslides_Hurricane_Ivan/Land slides_Hurricane_Ivan.htm. Nowell, P., 2004, Far-stretching Frances disrupts N.C. from mountains to coast, viewed September 6, 2004 at http://abclocal.go.com/wtvd/news/print_090904_Apstate_francesroundup.html. Ostendorff, J., 2004, Flood work in Peeks Creek moves from rescue to recovery; one man remains missing: Asheville Citizen-Times. Otteman, R., 2001, Using GIS to model debris flow susceptibility for the Bent Creek Experimental Forest near Asheville, North Carolina: M.S. Thesis, East Carolina University, Greenville, North Carolina, 181 p. Pack, R.T., Maynard, D., Geertsema, M., 2002, The use of SINMAP in terrain stability mapping in the Prince George region two case studies in Jordan, P., and Orban, J., ed.,

128

Terrain stability and forest management in the interior of British Columbia: Workshop Proceedings, May 23-25, 2001, Nelson, British Columbia, Canada, no. 3, p. 209-210. Pack, R.T., Tarboton, D.G., Goodwin, C.N., 1998a, The SINMAP approach to terrain stability mapping: 8th Congress of the International Association of Engineering Geology, Vancouver, British Columbia, Canada. Pack, R.T., Tarboton, D.G., Goodwin, C.N., 1998b, Terrain stability mapping with SINMAP, technical description and users guide for version 1.00: Terratech Consulting Ltd., Salmon Arm, B.C., Canada, Report Number 4114-0, 68 p. (Report and software available online: http://moose.cee.usu.edu/sinmap/sinmap.htm) Pack, R.T., Tarboton, D.G., Goodwin, C.N., 2001, Assessing terrain stability in a GIS using SINMAP: Presented at the 15th Annual GIS Conference, GIS 2001, February 1922, Vancouver, B.C., Canada, 9 p. Pierson, T.C. and Costal, J.E., 1987, A rheologic classification of subaerial sedimentwater flows: Geological Society of America, Reviews in Engineering Geology, v. 7, p.112. Pomeroy, 1991, Map showing late 1977 debris avalanches southwest of Asheville, western North Carolina: U.S. Geological Survey Open File Report 91-334, scale 1:24,000. Renshaw, C.E., 2000, Fracture spatial density and the anisotropic connectivity of fracture networks in Faybishenko, B., Witherspoon, P.A., Benson, S.M., eds., Dynamics of fluids in fractured rocks: Washington, D.C., American Geophysical Union, p. 203-211. Ritter, D.F., Kochel, R.C., and Miller, J.R., 1995, Process geomorphology: Third Edition, Wm. C. Brown Publishers, Dubuque, Iowa, 546 p. Schuster, R.L., 1996, Socioeconomic significance of landslides in Turner A.K., and Schuster, R.J., eds., Landslides: Investigation and Mitigation - Special Report 247: Washington, D.C., Transportation Research Board, National Research Council, p. 12-35. Schuster, R.L. and Kockelman, W.J., 1996, Principles of Landslide Hazard Reduction in Turner A.K., and Schuster, R.J., eds., Landslides: Investigation and Mitigation - Special Report 247: Washington, D.C., Transportation Research Board, National Research Council, p. 91-105. Scott, R.C., 1972, The geomorphic significance of debris avalanching in the Appalachian Blue Ridge Mountains: Unpublished Ph.D. dissertation, University of Georgia, Athens, GA, 124 p. Scott, R.C., 1981, Meteorological Aspects of Major Debris Avalanche Occurrences in the Southern Appalachians: Virginia Geographer, v. 14, p. 27-37.

129

Sharpe, C.F.S., 1938, Landslides and related phenomena: a study of mass-movements of soil and rock: New York, Columbia University Press, 137 p. Sidle, R.C., Pearce, A.J., OLoughlin, C.L., 1985, Hillslope stability and landuse: Water Resources Monograph 11 Edition, American Geophysical Union, p. 1940. Southern Railway, 1917, The floods of July 1916: how the Southern Railway organization met an emergency: Washington, D.C., Southern Railway Company, 131 p. Southworth, S., Schultz, A., Denenny, D., Triplett, J., 2003, Surficial geologic map of the Great Smoky Mountains National Park region, Tennessee and North Carolina, U.S. Geological Survey Open File Report 03-381: Reston, VA, U.S. Geological Survey, 44 p. State Climate Office of North Carolina, 2003, Aspects of NC Climate Extreme weather records, viewed October 17, 2003 at http://www.ncclimate.ncsu.edu/climate/extremes.html. Stewart, J.M., Heath, R.C., Morris, J.N., 1978, Floods in Western North Carolina, November 1977: Water Resources Research Institute of the University of North Carolina, 25 p. Swift, L.W., Cunningham, G.B., Douglass, J.E., 1988, Climatology and hydrology at Coweeta in Swank, W.T. and Crossley, D.A., eds., Forest hydrology and ecology at Coweeta: New York, N.Y., Springer-Verlag, p. Tennessee Valley Authority, Division of Water Control Planning, 1960, Floods on French Broad and Swannanoa Rivers in vicinity of Asheville, North Carolina: Knoxville, TN, Tennessee Valley Authority, 88p. U.S. Census Bureau, 2000, County and city data books (electronic resource): Washington, D.C., viewed November 16, 2002 at http://www.census.gov/statab/www/ccdb.html. U.S. Department of Agriculture, 1902, Message from the President of the United States transmitting a report of the Secretary of Agriculture in relation to the forests, rivers, and mountains of the southern Appalachian region: Washington, Government Printing Office, 210 p. U.S. Department of Agriculture, 1998, Soils - NC General, State Soil Geographic (STATSGO) database: Raleigh, U.S. Department of Agriculture, Natural Resource Conservation Service, scale 1:250,000. (Available Online: http://www.ftw.ncrs.usda.gov/statsgo2_ftp.html)

130

U.S. Department of Agriculture NRCS Soil Survey Division, 2005, Official Soil Series Descriptions, viewed March 1, 2005 at http://ortho.ftw.nrcs.usda.gov/cgibin/osd/osdname.cgi. U. S. Geological Survey - Water Resources Branch, 1949, Floods of August 1940 in the Southeastern States: Geological Survey Water Supply Paper 1006: U.S. Government Printing Office, Washington D.C., 554 p. U.S. Geological Survey, 1999, National Elevation Dataset: United States Geological Survey (USGS), EROS Data Center: United States Geological Survey, Sioux Falls, SD, Available Online: http://gisdata.usgs.net/ned/. Varnes, D.J., 1978, Slope movement types and processes in Schuster, R., and Krizek, R., eds., Landslides: Analysis and Control: Washington, D.C., National Academy of Science, p. 12-33. Wieczorek, G.F., 1996, Landslide triggering mechanisms, in Turner, A.K. and Schuster, R.J., eds., Landslides: Investigation and Mitigation: Washington, D.C., Transportation Research Board Special Report 247, National Research Council, p. 76-90. Wieczorek, G.F., Mossa, G.S., Morgan, B.A., 2004, Regional debris-flow distribution and preliminary risk assessment from severe storm events in the Appalachian Blue Ridge Province, USA: Landslides, v. 1, pp. 53-59. Williams, G.P. and Guy, H.P., 1973, Erosional and depositional aspects of Hurricane Camille in Virginia, U.S.A: U.S. Geological Survey Professional Paper 804, 80 p. Woodruff, J.F., 1971, Debris avalanches as an erosional agent in the Appalachian Mountains: The Journal of Geography, v. 70, p. 399-406. Wu, W. and Sidle R.C., 1995, A distributed slope stability model for steep forested basins: Water Resources Research, v. 31, n. 8, p. 2097-2110. Wooten, R.M., Clark, T.W., Bateson, J.T., 2001, Geologic features related to recent debris flows and weathered-rock slides in the Blue Ridge and Piedmont of North Carolina: Abstracts with Programs - Geologic Society of America, v. 33, no. 2, p. 14-15. Wooten, R. M., Reid, J.C., Clark, T.W., Medina, M.A., Bateson, J.T., Davidson, V.A., 2004, Selected geologic hazards in North Carolina An overview (poster), Northeastern Section and Southeastern Section Joint Meeting, March 2527, 2004. Zaitchick, B.F., Van Es, H.M., Sullivan, P.J., 2003, Modeling Slope Stability in Honduras: Parameter Sensitivity and Scale of Aggregation: Soil Science Society of America Journal, v. 67, p. 268-278.

131

Zhang, X., Drake, N.A., Wainwright, J., Mulligan, M., 1999, Comparison of slope estimates from low resolution DEMs: Scaling issues and a fractal method for their solution: Earth Surface Processes and Landforms, v. 24, 763-799.

132

Appendices

133

Appendix A: Geologic Units Geologic Rock Type Unit

CZab Metamorphic

Formation Name (if given)

Geologic Description

Composition

Interlayered; minor layers and lenses of Amphibolite and Biotite Gneiss hornblende gneiss, metagabbro, mica schist, and granitic rock Inequigranular, locally abundant potassic feldspar and garnet; interlayered and gradational with calc-silicate rock, sillimanite-mica schist, mica schist, and amphibolite. Contains small masses of granitic rock Interlayered with amphibole Feldspathic arenite, white to yellowish gray. Minor silty shale, feldspathic siltstone, and conglomerate in lower part

CZbg

Metamorphic

Biotite Gneiss and Schist

CZgms CZtp Ccl

Metamorphic SAURATOWN MOUNTAIN Metamorphic SAURATOWN MOUNTAIN Sedimentary LOWER CHILHOWEE

Garnet-Mica Schist Porphyroblastic gneiss Arenite

Ccu Chg Cr DSc Dqd

Sedimentary

UPPER CHILHOWEE

Metamorphic HENDERSON GNEISS Sedimentary Metamorphic Igneous ROME FORMATION CAESARS HEAD GRANITE GNEISS INTRUSIVE

Vitreous quartz arenite, white to light gray; interbedded sandy siltstone and shale Monzonitic to granodioritic, Gneiss inequigranular Shale and siltstone, variegated red to Shale and Siltstone brown; interbedded fine-grained sandstone Equigranular to porphyritic, massive to well foliated; contains biotite and Granite Gneiss muscovite Contains biotite, muscovite, and Quartz Diorite and Granodiorite xenocrysts Arenite

134

Dsc PzZu SOgg

? Meta-Ultramafic Rock Granite Gneiss

Metamorphic INTRUSIVE Metamorphic INTRUSIVE

? Metamorphosed dunite and peridotite; serpentinite, soapstone, and other altered ultramafic rock Poorly foliated; interlayered with biotite augen gneiss Equigranular, massive to well foliated, interlayered, rarely discordant, metamorphosed intrusive and extrusive mafic rock; may include metasedimentary rock Pinkish gray to light gray, massive to well foliated, granitic to quartz monzonitic; includes variably mylonitized orthogneiss and para-gneiss, interlayered amphibolite, calc-silicate rock, and marble Greenish gray to pinkish gray, porphyroclastic to mylonitic; epidote, sericite, and chlorite common Equigranular, massive to well foliated, interlayered, rarely discordant, metamorphosed intrusive and extrusive mafic rock; may include metasedimentary rock Layered biotite-granite gneiss, biotitehornblende gneiss, amphibolite, calcsilicate rock; locally contains relict granulite facies rock Poorly foliated to well foliated, equigranular to inequigranular, granitic

Ybam

Metamorphic UNCONFORMITY

Amphibolite

Ybgg

Metamorphic UNCONFORMITY

Biotite Granite Gneiss

Ygg

Metamorphic UNCONFORMITY

Granodioritic Gneiss

Ymam

Metamorphic UNCONFORMITY

Amphibolite

Ymg

Metamorphic UNCONFORMITY

Migmatitic Biotite-Hornblende Gneiss Gneiss

Ytg

Metamorphic TOXAWAY GNEISS

135

Zaba

Metamorphic

ALLIGATOR BACK FORMATION

Amphibolite

Zabg

Metamorphic

ALLIGATOR BACK FORMATION

Gneiss

Equigranular, massive to well foliated, interlayered, rarely discordant, metamorphosed intrusive and extrusive mafic rock; may include metasedimentary rock Finely laminated to thin layered; locally contains massive gneiss and micaceous granule conglomerate; includes schist, phyllite and amphibolite Equigranular, massive to well foliated, interlayered, rarely discordant, metamorphosed intrusive and extrusive mafic rock; may include metasedimentary rock Interlayered with biotite-garnet gneiss, biotite-muscovite schist, garnet-mica schist, and amphibolite Locally sulfidic, interlayered and gradational with mica schist, minor amphibolite, and hornblende gneiss Foliated to massive. Locally conglomeratic; interlayered and gradational with mica schist, muscovitebiotite gneiss, and rare graphitic schist Metagraywacke, massive, graded bedding common; includes dark-gray slate, mica schist, and nodular calcsilicate rock Slate to phyllite, dark gray, graphitic, sulfidic; includes metagraywacke with local graded bedding Sericitic, conglomeritic, locally crossbedded interlayered metasiltstone and slate

Zata

ASHE METAMORPHIC Metamorphic SUITE AND TALLULAH FALLS FORMATION ASHE METAMORPHIC Metamorphic SUITE AND TALLULAH FALLS FORMATION ASHE METAMORPHIC Metamorphic SUITE AND TALLULAH FALLS FORMATION ASHE METAMORPHIC Metamorphic SUITE AND TALLULAH FALLS FORMATION

Amphibolite

Zatb

Biotite gneiss

Zatm

Muscovite-biotite gneiss

Zatw

Metagraywacke

Zch

Metamorphic COPPERHILL FORMATION

Metagraywacke

Zchs

Metamorphic

SLATE OF COPPERHILL FORMATION

Slate

Zgma

Metamorphic

GRANDFATHER MOUNTAIN Meta-Arkose FORMATION

136

Zgmg

Metamorphic

GRANDFATHER MOUNTAIN Greenstone FORMATION

Schistose to massive, amygdaloidal; interlayered with metasedimentary rock Thick metasedimentary sequence of massive to graded beds of metagraywacke and metasiltstone with interbedded graphitic and sulfidic slate and schist Mottled pink and light green, coarse grained to porphyritic, massive; contains biotite Feldspathic; interbedded with dark argillaceous layers and laminae Feldspathic metasiltstone, metasandstone, and phyllite. Basal schist contains lenses of quartz-pebble conglomerate Cross-bedded. feldspathic, locally conglomeratic; includes dark slate and metasiltstone Thin bedded to laminated, commonly cross-bedded, metamorphosed; locally includes argillite and calcareous and ankeritic metasiltstone grading to silty metalimestone Greenish gray, fine to medium grained, locally cross-bedded, metamorphosed; interbedded metasiltstone and phyllite Slate and metasiltstone, dark green to black. Metaconglomerate lentils in upper part; calcareous metasandstone, sandy metalimestone, and quartzite in lower part

Zgs Zlm Zm Zrb

Metamorphic

GREAT SMOKY GROUP, UNDIVIDED

Metagraywacke and metasiltstone Metadacite Granite Sandstone

Metamorphic Lincolnton Metadacite Igneous Sedimentary MAX PATCH GRANITE RICH BUTT SANDSTONE SNOWBIRD GROUP, UNDIVIDED

Zs

Metamorphic

Metasiltstone and sandstone

Zsl

Metamorphic LONGARM QUARTZITE

Quartzite

Zsp

Sedimentary

PIGEON SILTSTONE

Siltstone

Zsr

Sedimentary

ROARING FORK SANDSTONE

Sandstone

Zss

Metamorphic SANDSUCK FORMATION

Slate

137

Zsw

WADING BRANCH Metamorphic FORMATION WALDEN CREEK GROUP, UNDIVIDED

Slate

Zwc

Metamorphic

Slate

ZYbn

Metamorphic

Biotite gneiss

bz

Metamorphic BREVARD FAULT ZONE

Schist and phyllonite

Sandy slate to coarse-grained pebbly metagraywacke with local graded bedding. Basal quartz-sericite schist or phyllite Slate to metasiltstone, local limy beds and pods; interbedded with quartz-pebble metaconglomerate and metasandstone Migmatitic; interlayered and gradational with biotite-garnet gneiss and amphibolite; locally abundant quartz and alumino-silicates. Stratigraphic position uncertain Fish scale schist and phyllonite, graphitic; interlayered with feldspathic metasandstone, marble lenses

138

Appendix B: General Soil Data

MUSEQ NAME S5ID % OF MAPUNIT SURFACE TEXTURE UNIFIED AASHTO DEPTH LOW DEPTH PERM HIGH LOW PERM HY COND HY COND HIGH LOW HIGH

NC005 NC005 NC005 NC005 NC005 NC005 NC005

1 2 3 5 6 7 8

TOXAWAY ROSMAN DELANCO COMUS HATBORO BRADSON SUNCOOK

NC0021 NC0024 MD0155 MD0050 PA0016 NC0028 CT0001

35 17 26 9 7 4 2

SIL L SIL FSL L GR-L LS

CL ML ML ML ML SM SM

A-4 A-4 A-4 A-2 A-4 A-2 A-2

0.0 2.5 1.0 6.0 0.0 6.0 3.0

1.0 5.0 2.5 6.0 0.5 6.0 6.0

0.60 2.00 0.20 0.60 2.00 0.60 6.00

20.00 20.00 2.00 6.00 6.00 20.00 20.00

0.21 0.34 0.05 0.05 0.14 0.02 0.12 0.94 0.024 0.09 0.14 0.11 0.10 0.05 0.05 0.55 0.014 0.14 0.30 0.09 0.04 0.04 0.10 0.12

7.00 3.40 0.52 0.54 0.42 0.80 0.40 13.08 (IN/HR) 0.331 (M/HR) 0.90 1.38 0.38 1.02 1.80 0.54 6.02 (IN/HR) 0.152 (M/HR) 0.46 0.90 0.30 0.12 1.20 0.30 0.36

NC006 NC006 NC006 NC006 NC006 NC006

CLIFTON BRADDOCK EVARD BRADDOCK TOXAWAY URBAN LAND NC006 7 CLIFTON

1 2 3 4 5 6

NC0015 VA0054 SC0083 VA0231 NC0021 DC0035 NC0015

15 23 19 17 9 8 9

L L L GR-L L VAR L

ML CL ML SM CL ML

A-4 A-2 A-4 A-2 A-4 A-4

6.0 6.0 6.0 6.0 0.0 2.0 6.0

6.0 6.0 6.0 6.0 1.0 2.0 6.0

0.60 0.60 0.60 0.60 0.60 0.60

6.00 6.00 2.00 6.00 20.00 6.00

NC088 NC088 NC088 NC088 NC088 NC088 NC088

1 2 3 4 5 6 7

CHESTER ASHE CHESTER CHESTER CODORUS ASHE CHANDLER

MD0001 NC0186 MD0001 MD0001 PA0015 NC0019 NC0263

23 15 15 6 6 5 6

L ST-FSL L L L FSL ST-FSL

CL SM CL CL ML SM SM

A-4 A-2 A-4 A-4 A-4 A-4 A-4

6.0 6.0 6.0 6.0 1.0 6.0 6.0

6.0 6.0 6.0 6.0 2.0 6.0 6.0

0.60 2.00 0.60 0.60 0.60 2.00 2.00

2.00 6.00 2.00 2.00 20.00 6.00 6.00

139

NC088 NC088 NC088 NC088 NC088 NC088 NC088 NC088

8 9 10 11 12 13 14 15

ASHE CHANDLER CHESTER SUNCOOK TATE WATAUGA WATAUGA FANNIN

NC0019 NC0263 MD0001 CT0001 NC0025 NC0091 NC0091 NC0020

3 3 4 3 3 3 3 2

FSL ST-FSL L LS L L L SIL

SM SM CL SM ML SM SM ML

A-4 A-4 A-4 A-2 A-4 A-4 A-4 A-4

6.0 6.0 6.0 3.0 6.0 6.0 6.0 6.0

6.0 6.0 6.0 6.0 6.0 6.0 6.0 6.0

2.00 2.00 0.60 0.60 2.00 0.60 0.60 0.60

6.00 6.00 2.00 20.00 6.00 6.00 6.00 6.00

0.06 0.06 0.02 0.02 0.06 0.02 0.02 0.01 0.21 0.005 0.19 0.08 0.07 0.05 0.03 0.05 0.02 0.02 0.02 0.02 0.02 0.02 0.01 0.01 0.11 0.003 0.15 0.09 0.08 0.07 0.05 0.04

0.18 0.18 0.08 0.60 0.18 0.18 0.18 0.12 1.52 (IN/HR) 0.038 (M/HR) 0.64 0.78 0.22 0.54 0.10 1.60 0.24 0.24 0.18 0.18 0.18 0.18 0.06 0.06 1.08 (IN/HR) 0.027 (M/HR) 1.50 0.90 2.60 2.20 1.80 0.42

NC089 NC089 NC089 NC089 NC089 NC089 NC089 NC089 NC089 NC089 NC089 NC089 NC089 NC089

1 2 3 4 5 6 7 8 9 10 11 12 13 14

CHESTER CLIFTON CHESTER CLIFTON CHESTER CODORUS WATAUGA WATAUGA CLIFTON FANNIN FANNIN PORTERS CLIFTON TATE

MD0001 NC0264 MD0001 NC0015 MD0001 PA0015 NC0091 NC0091 NC0015 NC0020 NC0020 NC0152 NC0015 NC0025

32 13 11 9 5 8 4 4 3 3 3 3 1 1

L ST-L L L L L L L L SIL SIL ST-L L L

CL SM CL ML CL ML SM SM ML ML ML ML ML ML

A-4 A-4 A-4 A-4 A-4 A-4 A-4 A-4 A-4 A-4 A-4 A-2 A-4 A-4

6.0 6.0 6.0 6.0 6.0 1.0 6.0 6.0 6.0 6.0 6.0 6.0 6.0 6.0

6.0 6.0 6.0 6.0 6.0 2.0 6.0 6.0 6.0 6.0 6.0 6.0 6.0 6.0

0.60 0.60 0.60 0.60 0.60 0.60 0.60 0.60 0.60 0.60 0.60 0.60 0.60 0.60

2.00 6.00 2.00 6.00 2.00 20.00 6.00 6.00 6.00 6.00 6.00 6.00 6.00 6.00

NC090 NC090 NC090 NC090 NC090 NC090

1 2 3 4 5 6

HAYESVILLE HAYESVILLE BRADSON CODORUS BRADSON EVARD

NC0013 NC0013 NC0028 PA0015 NC0028 SC0083

25 15 13 11 9 7

L L GR-L L GR-L SL

SM SM SM ML SM SM

A-4 A-4 A-2 A-4 A-2 A-2

6.0 6.0 6.0 1.0 6.0 6.0

6.0 6.0 6.0 2.0 6.0 6.0

0.60 0.60 0.60 0.60 0.60 0.60

6.00 6.00 20.00 20.00 20.00 6.00

140

NC090 NC090 NC090 NC090 NC090

7 8 9 10 11

ROSMAN DELANCO EDNEYVILLE HATBORO TOXAWAY

NC0024 MD0155 NC0023 PA0016 NC0021

6 4 4 4 2

L FSL FSL L SIL

ML ML SM ML CL

A-4 A-4 A-2 A-4 A-4

2.5 1.0 6.0 0.0 0.0

5.0 2.5 6.0 0.5 1.0

0.60 0.60 0.20 0.60 0.60

20.00 6.00 6.00 6.00 20.00

0.04 0.02 0.01 0.02 0.01 0.20 0.005 0.16 0.36 0.07 0.05 0.06 0.04 0.20 0.02 0.01 0.01 0.01 0.35 0.009 0.19 0.25 0.06 0.02 0.04 0.28 0.00 0.00 0.01 0.01 0.02 0.35

1.20 0.24 0.24 0.24 0.40 4.54 (IN/HR) 0.115 (M/HR) 1.62 1.08 0.72 0.48 0.60 0.42 0.60 0.24 0.12 0.06 0.06 2.10 (IN/HR) 0.053 (M/HR) 0.64 2.46 0.18 0.06 0.12 2.80 0.12 0.40 0.02 0.06 0.06 3.58 (IN/HR)

NC091 NC091 NC091 NC091 NC091 NC091 NC091 NC091 NC091 NC091 NC091

1 2 3 4 5 6 7 8 9 10 11

DITNEY UNICOI JUNALUSKA BRASSTOWN LONON NORTHCOVE SOCO JUNALUSKA BRASSTOWN LONON NORTHCOVE

TN0075 TN0054 NC0181 NC0206 NC0203 NC0204 NC0180 NC0181 NC0206 NC0203 NC0204

27 18 12 8 10 7 10 4 2 1 1

CB-SL STV-L CN-FSL CN-FSL CB-SL STV-FSL CN-FSL CN-FSL CN-FSL CB-SL STV-FSL

ML GM SM SM SM GM SM SM SM SM GM

A-4 A-2 A-4 A-4 A-2-4 A-2-4 A-4 A-4 A-4 A-2-4 A-2-4

6.0 6.0 6.0 6.0 6.0 6.0 6.0 6.0 6.0 6.0 6.0

6.0 6.0 6.0 6.0 6.0 6.0 6.0 6.0 6.0 6.0 6.0

0.60 2.00 0.60 0.60 0.60 0.60 2.00 0.60 0.60 0.60 0.60

6.00 6.00 6.00 6.00 6.00 6.00 6.00 6.00 6.00 6.00 6.00

NC092 NC092 NC092 NC092 NC092 NC092 NC092 NC092 NC092 NC092 NC092

1 2 3 4 5 6 7 8 9 10 11

SYLCO DITNEY TUSQUITEE DITNEY DITNEY CATASKA SPIVEY HAYWOOD SYLCO UNICOI UNICOI

TN0014 TN0075 NC0158 TN0075 TN0075 TN0133 TN0109 NC0095 TN0014 TN0054 TN0054

32 41 3 1 2 14 2 2 1 1 1

SIL CB-L ST-L CB-L CB-L STV-L STV-L ST-L L STV-L STV-L

GC ML SM ML ML CL-ML GM SM GC GM GM

A-4 A-4 A-2-4 A-4 A-4 A-4 A-2 A-2-4 A-4 A-2 A-2

6.0 6.0 6.0 6.0 6.0 6.0 6.0 6.0 6.0 6.0 6.0

6.0 6.0 6.0 6.0 6.0 6.0 6.0 6.0 6.0 6.0 6.0

0.60 0.60 2.00 2.00 2.00 2.00 0.06 0.06 0.60 0.60 2.00

2.00 6.00 6.00 6.00 6.00 20.00 6.00 20.00 2.00 6.00 6.00

141

0.009 ASHE CHESTER CHESTER ASHE ROCK OUTCROP NC093 6 ASHE NC093 7 CLIFTON NC093 8 TUSQUITEE NC093 NC093 NC093 NC093 NC093 1 2 3 4 5 NC0186 MD0001 MD0001 NC0019 DC0015 NC0019 NC0264 NC0158 53 15 12 9 8 1 1 1 ST-FSL L L FSL UWB FSL ST-L ST-L SM CL CL SM SM SM SM A-2 A-4 A-4 A-4 A-4 A-4 A-2-4 6.0 6.0 6.0 6.0 6.0 6.0 6.0 6.0 6.0 6.0 6.0 6.0 6.0 6.0 6.0 6.0 2.00 0.60 0.60 0.60 2.00 2.00 0.60 0.60 6.00 2.00 2.00 6.00 0.00 6.00 6.00 6.00 1.06 0.09 0.07 0.05 0.16 0.02 0.01 0.01 0.41 0.010 0.46 0.24 0.07 0.04 0.04 0.03 0.10 0.04 0.02 0.02 0.02 0.01 0.01 0.05 0.18 0.005 0.10 0.06 0.06 0.18

0.091 (M/HR) 3.18 0.30 0.24 0.54 0.00 0.06 0.06 0.06 1.26 (IN/HR) 0.032 (M/HR) 1.38 0.72 0.66 0.42 0.42 0.30 0.30 0.36 0.24 0.18 0.24 0.04 0.12 1.80 2.98 (IN/HR) 0.075 (M/HR) 1.02 0.60 0.60 1.80

NC094 NC094 NC094 NC094 NC094 NC094 NC094 NC094 NC094 NC094 NC094 NC094 NC094 NC094

1 2 3 4 5 6 7 8 9 10 11 12 13 14

CHANDLER WATAUGA CHANDLER CLIFTON WATAUGA ASHE FANNIN CLIFTON CLIFTON TATE CHANDLER CHESTER CLIFTON CODORUS

NC0263 NC0091 NC0263 NC0264 NC0091 NC0186 NC0020 NC0015 NC0015 NC0025 NC0017 MD0001 NC0015 PA0015

23 12 11 7 7 5 5 6 4 3 4 2 2 9

ST-FSL L ST-FSL ST-L L ST-FSL SIL L L L L L L SIL

SM SM SM SM SM SM ML ML ML ML ML CL ML ML

A-4 A-4 A-4 A-4 A-4 A-2 A-4 A-4 A-4 A-4 A-2 A-4 A-4 A-4

6.0 6.0 6.0 6.0 6.0 6.0 6.0 6.0 6.0 6.0 6.0 6.0 6.0 1.0

6.0 6.0 6.0 6.0 6.0 6.0 6.0 6.0 6.0 6.0 6.0 6.0 6.0 2.0

2.00 2.00 0.60 0.60 0.60 0.60 2.00 0.60 0.60 0.60 0.60 0.60 0.60 0.60

6.00 6.00 6.00 6.00 6.00 6.00 6.00 6.00 6.00 6.00 6.00 2.00 6.00 20.00

NC095 NC095 NC095 NC095

1 2 3 4

EVARD FANNIN PORTERS BREVARD

SC0083 NC0020 NC0152 NC0012

17 10 10 9

SL SIL ST-L L

SM ML ML ML

A-2 A-4 A-2 A-4

6.0 6.0 6.0 6.0

6.0 6.0 6.0 6.0

0.60 0.60 0.60 2.00

6.00 6.00 6.00 20.00

142

NC095 NC095 NC095 NC095 NC095 NC095 NC095 NC095 NC095 NC095 NC095 NC095 NC095 NC095

5 6 7 8 9 10 11 12 13 14 15 16 17 18

FANNIN TUSQUITEE ASHE PORTERS ASHE TUSQUITEE EDNEYVILLE ROCK OUTCROP BREVARD CHANDLER TUSQUITEE EVARD EVARD TUSQUITEE

NC0020 NC0158 NC0186 NC0152 NC0186 NC0026 NC0023 DC0015 NC0012 NC0263 NC0158 SC0083 SC0083 NC0026

8 8 8 7 5 5 3 1 2 2 2 1 1 1

SIL ST-L ST-SL ST-L ST-SL L FSL UWB L ST-L ST-L SL SL L

ML SM SM ML SM ML SM ML SM SM SM SM ML

A-4 A-2-4 A-2 A-2 A-2 A-4 A-2 A-4 A-4 A-2-4 A-2 A-2 A-4

6.0 6.0 6.0 6.0 6.0 6.0 6.0 6.0 6.0 6.0 6.0 6.0 6.0 6.0

6.0 6.0 6.0 6.0 6.0 6.0 6.0 6.0 6.0 6.0 6.0 6.0 6.0 6.0

0.60 0.60 2.00 2.00 2.00 2.00 2.00 2.00 2.00 0.60 2.00 0.60 0.60 0.60

6.00 6.00 6.00 6.00 6.00 6.00 6.00 0.00 20.00 6.00 6.00 6.00 6.00 6.00

0.05 0.05 0.16 0.14 0.10 0.10 0.06 0.02 0.04 0.01 0.04 0.01 0.01 0.01 0.13 0.003 0.62 0.32 0.06 0.05 0.04 0.04 0.02 0.08 0.08 0.06 0.06 0.04 0.01 0.01 0.33 0.008

0.48 0.48 0.48 0.42 0.30 0.30 0.18 0.00 0.40 0.12 0.12 0.06 0.06 0.06 0.82 (IN/HR) 0.021 (M/HR) 1.86 0.96 2.00 0.16 0.42 0.12 0.24 0.24 0.24 0.18 0.18 0.12 0.06 0.06 1.08 (IN/HR) 0.027 (M/HR)

NC096 NC096 NC096 NC096 NC096 NC096 NC096 NC096 NC096 NC096 NC096 NC096 NC096 NC096

1 2 3 4 5 6 7 8 9 10 11 12 13 14

PORTERS PORTERS CODORUS CHESTER TUSQUITEE CHESTER CHANDLER PORTERS PORTERS TUSQUITEE TUSQUITEE FANNIN FANNIN WATAUGA

NC0152 NC0152 PA0015 MD0001 NC0158 MD0001 NC0263 NC0022 NC0022 NC0026 NC0026 NC0020 NC0020 NC0091

31 16 10 8 7 6 4 4 4 3 3 2 1 1

ST-L ST-L SIL L ST-L L ST-FSL L L L L SIL SIL L

ML ML ML CL SM CL SM ML ML ML ML ML ML SM

A-2 A-2 A-4 A-4 A-2-4 A-4 A-4 A-4 A-4 A-4 A-4 A-4 A-4 A-4

6.0 6.0 1.0 6.0 6.0 6.0 6.0 6.0 6.0 6.0 6.0 6.0 6.0 6.0

6.0 6.0 2.0 6.0 6.0 6.0 6.0 6.0 6.0 6.0 6.0 6.0 6.0 6.0

2.00 2.00 0.60 0.60 0.60 0.60 0.60 2.00 2.00 2.00 2.00 2.00 0.60 0.60

6.00 6.00 20.00 2.00 6.00 2.00 6.00 6.00 6.00 6.00 6.00 6.00 6.00 6.00

143

NC097 NC097 NC097 NC097 NC097 NC097

1 2 3 4 5 6

EDNEYVILLE EDNEYVILLE TUSQUITEE ASHE TUSQUITEE TOXAWAY

NC0023 NC0023 NC0026 NC0019 NC0026 NC0021

73 11 9 3 3 1

L L L GR-FSL L L

SM SM ML SM ML CL

A-2 A-2 A-4 A-2 A-4 A-4

6.0 6.0 6.0 6.0 6.0 0.0

6.0 6.0 6.0 6.0 6.0 1.0

0.60 2.00 2.00 2.00 2.00 2.00

6.00 6.00 6.00 6.00 6.00 20.00

0.44 0.22 0.18 0.06 0.06 0.02 0.99 0.025 0.26 0.18 0.16 0.16 0.16 0.14 0.12 0.02 0.06 0.04 0.02 0.00 0.01 0.02 0.02 0.17 0.004

4.38 0.66 0.54 0.18 0.18 0.20 6.17 (IN/HR) 0.156 (M/HR) 2.58 0.54 0.48 0.48 0.48 0.42 0.36 0.06 0.18 0.12 0.06 0.06 0.06 0.06 0.06 0.60 (IN/HR) 0.015 (M/HR) 2.16 1.20 0.34 1.80 0.36 0.36 0.18

NC098 NC098 NC098 NC098 NC098 NC098 NC098 NC098 NC098 NC098 NC098 NC098 NC098 NC098 NC098

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15

WAYAH TANASEE PORTERS WAYAH WAYAH TANASEE PORTERS CULLASAJA EDNEYVILLE EDNEYVILLE CHESTNUT SPIVEY TUSQUITEE CHESTNUT TUCKASEGE E

NC0188 NC0197 NC0152 NC0188 NC0188 NC0197 NC0152 NC0237 NC0115 NC0115 NC0242 TN0109 NC0158 NC0242 NC0226

43 9 8 8 8 7 6 1 3 2 1 1 1 1 1

GR-L ST-L ST-FSL GR-L GR-L ST-L ST-FSL STV-L ST-FSL ST-FSL ST-FSL STV-L ST-FSL ST-FSL ST-L

SM SM ML SM SM SM ML SM SM SM SM GM SM SM ML

A-2-4 A-2-4 A-2 A-2-4 A-2-4 A-2-4 A-2 A-1-B A-4 A-4 A-2-4 A-2 A-2-4 A-2-4 A-2

6.0 6.0 6.0 6.0 6.0 6.0 6.0 6.0 6.0 6.0 6.0 6.0 6.0 6.0 6.0

6.0 6.0 6.0 6.0 6.0 6.0 6.0 6.0 6.0 6.0 6.0 6.0 6.0 6.0 6.0

0.60 2.00 2.00 2.00 2.00 2.00 2.00 2.00 2.00 2.00 2.00 0.06 0.60 2.00 2.00

6.00 6.00 6.00 6.00 6.00 6.00 6.00 6.00 6.00 6.00 6.00 6.00 6.00 6.00 6.00

NC099 NC099 NC099 NC099 NC099 NC099 NC099

1 2 3 4 5 6 7

FANNIN FANNIN TALLADEGA IOTLA EVARD EVARD TATE

NC0020 NC0020 GA0037 NC0140 SC0083 SC0083 NC0025

36 20 17 9 6 6 3

SIL SIL SIL L SL SL FSL

ML ML SM SM SM SM ML

A-4 A-4 A-4 A-2 A-2 A-2 A-4

6.0 6.0 6.0 1.5 6.0 6.0 6.0

6.0 6.0 6.0 3.5 6.0 6.0 6.0

2.00 0.60 0.60 2.00 0.60 0.60 0.60

6.00 6.00 2.00 20.00 6.00 6.00 6.00

0.72 0.12 0.10 0.18 0.04 0.04 0.02

144

NC099 8 HATBORO NC099 9 TUSQUITEE

PA0016 NC0026

2 1

L L

ML ML

A-4 A-4

0.0 6.0

0.5 6.0

0.60 0.60

6.00 6.00

0.01 0.01 0.39 0.010 0.22 0.11 0.02 0.15 0.01 0.02 0.01 0.07 0.39 0.010 0.13 0.22 0.10 0.07 0.12 0.10 0.04 0.08 0.05 0.10 0.02 0.01 0.02 0.01 0.28 0.007

0.12 0.06 3.22 (IN/HR) 0.081 (M/HR) 0.74 1.08 0.24 1.50 0.12 0.06 0.20 0.72 3.92 (IN/HR) 0.099 (M/HR) 1.26 0.66 1.02 0.66 0.36 0.30 0.36 0.24 0.48 0.30 0.18 0.06 0.06 0.06 1.38 (IN/HR) 0.035 (M/HR)

NC100 NC100 NC100 NC100 NC100 NC100 NC100 NC100

1 2 3 4 5 6 7 8

EVARD OTEEN HAYESVILLE SALUDA TUSQUITEE TATE BREVARD HAYESVILLE

SC0083 NC0107 NC0013 SC0082 NC0158 NC0025 NC0012 NC0151

37 18 4 25 2 1 1 12

L L L L ST-L ST-L L ST-L

ML ML SM SM SM ML ML SM

A-4 A-4 A-4 A-2 A-2-4 A-4 A-4 A-4

6.0 6.0 6.0 6.0 6.0 6.0 6.0 6.0

6.0 6.0 6.0 6.0 6.0 6.0 6.0 6.0

0.60 0.60 0.60 0.60 0.60 2.00 0.60 0.60

2.00 6.00 6.00 6.00 6.00 6.00 20.00 6.00

NC102 NC102 NC102 NC102 NC102 NC102 NC102 NC102 NC102 NC102 NC102

1 2 3 4 5 6 7 8 9 10 11

NC102 12 NC102 13 NC102 14

JUNALUSKA TSALI SPIVEY SANTEETLA H STECOAH SOCO JUNALUSKA TSALI CHEOAH SPIVEY SANTEETLA H JUNALUSKA TSALI BRASSTOWN

NC0181 NC0179 TN0109 NC0208 NC0184 NC0180 NC0181 NC0179 NC0190 TN0109 NC0208 NC0181 NC0179 NC0206

21 11 17 11 6 5 6 4 8 5 3 1 1 1

CN-L CN-L STV-L CN-FSL CN-L CN-L CN-L CN-L CN-L STV-L CN-FSL CN-L CN-L CN-L

SM SM GM SM SM SM SM SM SM GM SM SM SM SM

A-4 A-4 A-2 A-4 A-4 A-4 A-4 A-4 A-4 A-2 A-4 A-4 A-4 A-4

6.0 6.0 6.0 6.0 6.0 6.0 6.0 6.0 6.0 6.0 6.0 6.0 6.0 6.0

6.0 6.0 6.0 6.0 6.0 6.0 6.0 6.0 6.0 6.0 6.0 6.0 6.0 6.0

0.60 2.00 0.60 0.60 2.00 2.00 0.60 2.00 0.60 2.00 0.60 0.60 2.00 0.60

6.00 6.00 6.00 6.00 6.00 6.00 6.00 6.00 6.00 6.00 6.00 6.00 6.00 6.00

145

NC103 NC103 NC103 NC103 NC103 NC103 NC103 NC103 NC103 NC103 NC103 NC103 NC103 NC103 NC103 NC103 NC103 NC103 NC103 NC103 NC103

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21

EDNEYVILLE CHESTNUT STECOAH SOCO STECOAH SOCO SPIVEY TUSQUITEE PORTERS COWEE EVARD EDNEYVILLE CHESTNUT PORTERS SPIVEY TUSQUITEE SANTEETLA H SAUNOOK STECOAH DELLWOOD DILLARD

NC0115 NC0242 NC0184 NC0180 NC0184 NC0180 TN0109 NC0158 NC0152 NC0171 SC0135 NC0115 NC0242 NC0152 TN0109 NC0158 NC0208 NC0195 NC0184 NC0183 GA0061

13 9 8 5 6 4 5 3 6 6 8 4 2 6 6 1 2 3 1 1 1

ST-FSL ST-FSL CN-L CN-L CN-L CN-L STV-L ST-L ST-FSL GR-L GR-L ST-FSL ST-FSL ST-FSL STV-L ST-L CN-L GR-L CN-L GR-FSL L

SM SM SM SM SM SM GM SM ML SM SM SM SM ML GM SM SM SM SM SM ML

A-4 A-2-4 A-4 A-4 A-4 A-4 A-2 A-2-4 A-2 A-2-4 A-2 A-4 A-2-4 A-2 A-2 A-2-4 A-4 A-2 A-4 A-2-4 A-4

6.0 6.0 6.0 6.0 6.0 6.0 6.0 6.0 6.0 6.0 6.0 6.0 6.0 6.0 6.0 6.0 6.0 6.0 6.0 2.0 2.0

6.0 6.0 6.0 6.0 6.0 6.0 6.0 6.0 6.0 6.0 6.0 6.0 6.0 6.0 6.0 6.0 6.0 6.0 6.0 4.0 3.0

0.60 2.00 2.00 2.00 2.00 2.00 2.00 0.60 2.00 2.00 0.60 0.60 2.00 2.00 2.00 0.60 2.00 0.60 0.60 2.00 0.60

6.00 6.00 6.00 6.00 6.00 6.00 6.00 6.00 6.00 6.00 6.00 6.00 6.00 6.00 6.00 6.00 6.00 6.00 6.00 20.00 2.00

0.08 0.18 0.16 0.10 0.12 0.08 0.10 0.02 0.12 0.12 0.05 0.02 0.04 0.12 0.12 0.01 0.04 0.02 0.01 0.02 0.01 0.22 0.005 0.04 0.38 0.32 0.28 0.14 0.06 0.04 0.06 0.04

0.78 0.54 0.48 0.30 0.36 0.24 0.30 0.18 0.36 0.36 0.48 0.24 0.12 0.36 0.36 0.06 0.12 0.18 0.06 0.20 0.02 1.00 (IN/HR) 0.025 (M/HR) 1.26 1.14 0.96 0.84 0.42 0.60 0.12 0.18 0.12

NC104 1 TANASEE NC104 2 BURTON NC104 3 OCONALUFT EE NC104 4 PORTERS NC104 5 SPIVEY NC104 6 CHEOAH NC104 7 STECOAH NC104 8 TANASEE NC104 9 BURTON

NC0197 NC0114 NC0192 NC0152 TN0109 NC0190 NC0184 NC0197 NC0114

21 19 16 14 7 10 2 3 2

ST-L ST-L FL-L ST-FSL CB-L CN-L CN-L ST-L ST-L

SM SM SM ML GM SM SM SM SM

A-2-4 A-2 A-4 A-2 A-2 A-4 A-4 A-2-4 A-2

6.0 6.0 6.0 6.0 6.0 6.0 6.0 6.0 6.0

6.0 6.0 6.0 6.0 6.0 6.0 6.0 6.0 6.0

0.20 2.00 2.00 2.00 2.00 0.60 2.00 2.00 2.00

6.00 6.00 6.00 6.00 6.00 6.00 6.00 6.00 6.00

146

NC104 10 SANTEETLA NC0208 H NC104 11 SOCO NC0180

5 1

L CN-L

SM SM

A-4 A-4

6.0 6.0

6.0 6.0

2.00 2.00

6.00 6.00

0.10 0.02 0.46 0.012

0.30 0.06 1.80 (IN/HR) 0.046 (M/HR)

147

Appendix C: Slope Movement Data for the 2004 Hurricanes Frances and Ivan HURRICANE FRANCES - SEPTEMBER 6-8, 2004
County Location Date Reported Avery County, Near Crossnore US-221 Both Directions Avery County, Near Crossnore SR-1504 Both Directions, Pineola Rd. Blue Ridge Parkway Mileposts: 322, 345, 348, 9/9/2004 349, 413, and 429 Henderson County Henderson County Henderson County Henderson County Henderson County, Near Hendersonville Henderson County Jackson County Source NCDOT NCDOT Description Road is one lane in certain locations HIGH WATER, MUDSLIDE

McDowell County, Near Old Fort

Asheville Citizen- Most of those between mile 322 and 349 took out major Times portions of the motor road, from just south of Linville Falls to south of Buck Creek Gap at NC 80 near Marion. SR-1710 Both Directions, 9/10/2004 17:06 NCDOT Trees down and Mudslide, Road has collapsed along with a Bald Rock Rd. mudslide near waterfall. SR-1799 Both Directions, 9/8/2004 NCDOT Deep Gap has many landslides blocking the road Deep Gap Rd. SR-1194 Both Directions, 9/8/2004 NCDOT Two mudslides into road Patterson Rd. SR-1613 Both Directions, 9/8/2004 NCDOT Road is washed out. Mudslide covering entire road Slick Rock Rd. SR-1706 Both Directions, NCDOT Sinkhole - road washed away about 1 mile off Sugarloaf Little Creek Rd. Mountain Road (SR 1868). Shepard Street portion of 9/8/04 11:00AM NCDOT One lane closed due to mudslide N.C. 191 NC-281 (Mile Marker 13 9/8/2004 NCDOT Road is closed due to slide from SR 1140 (Fanny Mae Brown to 17) Both Directions, Road) to Rock Bridge near Tannassee Gap. Detour is signed. Little Canada Detour distance is 41 miles. I-40 (Mile Marker 72 to early morning, NCDOT A slide on I-40 at Old Fort Mountain has blocked 2 of 3 lanes 67or 69-67) Both 9/8/04 on westbound I-40 as well as 2 of the 3 eastbound lanes. Directions Emergency crews are working to reopen the rest of the lanes as quickly as possible. Expect delays.

148

McDowell County, Near Woodlawn

US-221 Both Directions

NCDOT

There are several mudslides and dangerous areas at the top of US 221 North (Linville Mtn.). The road is closed and will be reopened as soon as the crews can complete the work and make the road passable again. US 221 North (Linville Mtn.) is closed due to Hurricane Frances. Approximately 70 feet of the road is washed away at the Mountain Paradise Campground. It will take a couple of months to rebuild this portion. A mudslide closed Pearson Falls Road near Saluda. between Saluda and Tryon, one lane closed from mudslides Temporarily closed Tuesday near the Henderson-Polk County line when mud and trees slide down a bank Portion of road closed due to mudslides near the Jackson County line This appears to have been an embankment failure. One home was destroyed, and eight condemned for occupancy

Polk County Polk County Polk County Transylvania County Watauga County

Pearson Falls Rd., near Saluda U.S. 176 U.S. 176

9/9/2004

Old Highway 64/ N.C. 281N White Laurel subdivision, 9/8/2004 near Boone

Asheville CitizenTimes 9/9/2004 Asheville CitizenTimes 9/9/2004 Asheville CitizenTimes 9/8/04 11:00AM NCDOT Rick Wooten, personal communication

HURRICANE IVAN - SEPTEMBER 16-18, 2004

County Avery County Location N.C. 184 and 194 Date Reported Source Asheville Citizen Times 9/18/2004 9/18/2004 9/18/2004 9/18/2004 9/18/2004 Asheville Citizen Times Asheville Citizen Times Asheville Citizen Times Asheville Citizen Times Asheville Citizen Description There was so much rain that high water and mudslides covered N.C. 184 and 194, shutting Banner Elk off from the rest of the region. mudslide mudslide mudslide mudslide mudslide

Buncombe County Buncombe County Buncombe County Buncombe County Buncombe County

Mink Farm Road Hookers Gap Gibbs Road Freedom Farm Road Newfound Rd

149

Buncombe County Buncombe County Buncombe County Buncombe County Haywood County Haywood County Haywood County Haywood County

Times North Turkey Creek Rd 9/18/2004 Asheville Citizen Times Sluder Branch Road - #99 9/18/2004 Asheville Citizen Times N.C. 151 Asheville Citizen Times Arrowood Rd. near 9/17/2004 2:00 Asheville Citizen Starnes Cove Rd. Times I-40 mudslide in the 9/17/2004 13:44 NCDOT westbound lane at MM 35 U.S. 19-23 9/17/2004 0:00 Asheville Citizen Times Dutch Cove at Turnpike 9/17/2004 Asheville Citizen Times I-40, between the TN state 9/17/2004 NCDOT line and MM 20 U.S.276 9/17/2004 Asheville Citizen Times 9/17/2004 Asheville Citizen Times 9/16/2004 23:02 NCDOT

Mudslide at Early's Mtn. Road Mudslide & Pvt. Bridge damaged Closed due to major slide near Blue Ridge Parkway access debris flow occurred on side of mountain near the town of Enka-Candler, destroyed at least one home Interstate 40 is closed in Haywood County from Exit 451 in Tennessee to Exit 20 in North Carolina due to a slope failure. mudslides mudslide I-40 is closed in Haywood County between the Tennessee State Line and mile marker 20 due to the road being washed away. rockslide south of Waynesville mudslide mudslide - N.C. 107 is closed form the Thorpe Power Plant south to Cashiers. Downed trees, boulders and mudslides are blocking the road, the main access the area. US 64 is closed between Cashiers and the Transylvania Co. Line due to slides and debris. US 64 is closed between Cashiers and the Macon Co. Line due to slides and debris. Hwy 64 at Spring Forest Road, three-fourths of highway has slid off Down Trees and Slides Rockslide on NC226-A will probably take all day to move. Mud Slide and shoulder broke off

Haywood County Henderson County

Middle Fork (one lane closed) Jackson County, Near Cashiers NC-107 (Mile Marker 18 to 18) Both Directions Jackson County, Near Cashiers US-64 Both Directions

9/16/2004 23:19 NCDOT

Macon, near Franklin McDowell County, Near Woodlawn Swain, Near Bryson City

SR-1310, Wayah Rd. NC-226 ALT Both Directions SR-1195, Hwy 19a

9/17/2004 7:46 9/17/2004 6:21 9/17/2004 8:02

NCDOT NCDOT NCDOT

150

Transylvania County, near Brevard Watauga, near Boone

SR-1540, Wilson Rd. SR-1130, Lee Gualtney Rd.

9/18/2004 3:35

NCDOT

9/17/2004 17:16 NCDOT

Due to high water, trees, power lines, and slides road is closed. Slide blocking road

151

Appendix D: Debris Flow Inventory

MATERIAL DATA SOURCE REMARKS LAT LONG COUNTY YEAR TYPE GEO UNIT SOIL UNIT SLOPE ASPECT

NCDOT Cv Scott, R NCGS Pomeroy NCDOT NCDOT Pomeroy NCDOT NCDOT NCDOT NCDOT NCDOT Lambe NCDOT NCDOT Pomeroy NCDOT Dockal FS Cv? NCGS Dockal Pomeroy NCDOT

SR 1607, .5mi S of Madison County line Brevard slide BRP Mile 357.7 MP 17.6, E of SR 1336 WBL Intersection of BRP, 694, SR 2053 .6 mi E of SR 1130 WBL .2mi N of Buncombe County line, US 25-70 Balsam Gap Landslide Rt. 206 NBL NC 226, .6mi S of 1253 Balsam Gap II Landslide 600 ft. S of SR 1319 Fines Creek Slide NC 262, 1.8mi S of TN line Off FR475, W of Davidson River, elev. 2500 feet Possible landslide Richland Ridge Rd. Maggie Valley debris flow Cove Creek Slide Number 50 N-side of NC19

35.7300 82.7000 35.1900 35.7305 35.4800 35.6300

Buncombe

1990

Slump

Ymg Dqd Zatw Zatm Zch Zatm Zatm ZYbn Ybgg ZYbn ZYbn Zatm ZYbn Zsr Ybgg Zatm Ymg Dqd Zata Zgs Zatb Zatm Zabg

NC090 NC093 NC098 NC093 NC006 NC095 NC093 NC090 NC100 NC095 NC006 NC093 NC103 NC092 NC006 NC093 NC098 NC094 NC006 NC103 NC094 NC095 NC093

10 11 11 13 13 13 15 15 16 17 17 17 18 18 19 19 19 20 20 21 21 21 21

N SE SW E SE N E SW NW E W SW NE W N S S S S NE SE E N

Rd Rd Cv Rd Rd Rd

82.8700 Transylvania 82.3082 Buncombe 82.6500 Buncombe 82.9900 Haywood Buncombe Buncombe Haywood Madison Haywood Haywood Mitchell Haywood Madison Haywood Buncombe Mitchell

1970 Debris Flow 1995 Debris Flow 1977 Slump 1968 Debris Flow 1982 Slump

35.6500 82.4900 35.4800 82.6700 35.4700 82.9500 35.7600 82.6100 35.4400 83.0800 35.6000 82.9400 35.9900 82.1600 35.4500 35.8900 35.6700 35.4800 36.0900 83.0700 82.7500 82.9900 82.6500 82.0900

1977 Slump 1990 Slump 1989 Debris Flow 1989 1996 1990 Slump Slump Slump

Cv Rd

Debris Flow 1986 Slump 1978 Debris Flow 1977 Slump Slump 1993 Slump

35.2800 82.8100 Transylvania 35.1800 82.8900 Transylvania 35.5014 35.2800 35.4500 35.9200 83.0944 Haywood 82.8100 Transylvania 82.6600 Henderson 82.0600 Mitchell

2003 Debris Flow Debris Flow Slump 1977 Debris Flow 1976 Slump

Rd

152

Cv? Rd

USFS Pomeroy Pomeroy Lambe NCDOT NCGS Pomeroy Pomeroy Pomeroy Pomeroy Pomeroy Pomeroy Dockal

Rd Rd

Rd rk Cv?

Dockal NCDOT NCDOT NRCS Pomeroy Pomeroy NCDOT NCDOT

Rd

Pomeroy Pomeroy Pomeroy Pomeroy Dockal NCDOT

Microburst; dual slides near Dry Branch Number 24 Number 49 Waterville Landslide SBL NC 25-70, S of NC 208 Deadly debris flow (12/11) - Maggie Valley Debris Number 15 Number 6 Number 41 Number 43 Number 35 Debris Torrent Near Maxwell Cove, east of drainage Cove Creek Slide SBL NC 215, 200' N of SR 1216, approximate I-40, World's Fair Slide Several debris flows triggered by trop. depression Number 36 Number 37 SBL SR 1137 SR 1318, .55mi W of SR 1334 Number 16

35.7700 83.0400 35.4700 35.4400 35.6900 35.9100 82.6600 82.6600 83.0100 82.7500

Haywood Buncombe Henderson Haywood Madison Haywood

1994 Debris Flow 1977 Debris Flow 1977 Debris Flow Debris Flow 1986 Slump 2003 Debris Flow 1977 1977 1977 1977 1977 1977 1996 Debris Flow Debris Flow Debris Flow Debris Flow Debris Flow Debris Flow Slump

Zsl Zatm Zatm Ybgg Zsr Ybgg Zatm Zatm Zatm Zatm Zatm Zatm Zatb Zatb Zatm Zsp Zatw Zatm Zatm Zabg Ybgg Zatm Zatm Zatm Zatm Zatb Ybgg

NC103 NC093 NC093 NC102 NC092 NC103 NC093 NC093 NC093 NC093 NC093 NC093 NC095 NC094 NC095 NC102 NC098 NC093 NC093 NC093 NC093 NC093 NC093 NC095 NC093 NC094 NC093

22 22 22 23 23 23 23 23 23 23 23 23 24 24 24 24 25 25 25 25 25 26 26 27 27 27 27

SE SE E SW SW NE NE NE SE S SE W SW SE E SW W SE S SW W E SW NE S E N

35.5048 83.8500 35.5000 35.4800 35.4300 35.4300 35.4600 35.4800 35.3200

82.6500 Buncombe 82.6900 Buncombe 82.6100 Henderson 82.6400 Henderson 82.6400 Buncombe 82.6800 Buncombe 82.7400 Transylvania

35.2800 82.8100 Transylvania 35.4000 82.9400 Haywood 35.7700 83.0900 35.7647 82.2656 35.4700 35.4700 35.9300 35.9200 82.6200 82.6100 82.5000 82.6700 Haywood Yancey Buncombe Buncombe Mitchell Madison

Slump 1989 Debris Flow 1982 Slump 1977 Debris Flow 1977 Debris Flow 1977 Debris Flow 1990 Slump 1990 Debris Flow 1977 Debris Flow 1977 Slump 1977 Debris Flow 1977 Slump Slump 1990 Debris Flow

35.4966 35.4800 Number 34 35.4500 35.4800 Cove Creek Slide 35.2800 WBL SR 1334, 0.2 mi NW 35.9100 of SR 1425

82.6504 Buncombe 82.6800 Buncombe 82.6500 Buncombe 82.6500 Buncombe 82.8100 Transylvania 82.6900 Madison

153

Rd wr Cv Cv Cv Rd

Pomeroy Pomeroy Pomeroy Pomeroy Pomeroy Pomeroy NCDOT Pomeroy NCDOT NCDOT NCDOT NCGS NCGS Pomeroy Pomeroy Pomeroy Pomeroy NCGS Pomeroy Pomeroy

Number 17 Number 39 Number 51 Number 34

W of SR 1395 Number 16 MP 8.8, E of Harmon Den NC 1318, .25 W of 1334 Hickey Fork Debris Flow Allen Stand Debris Flow No. 23, Runnout length of 2820 ft. Number 14 Number 7 Blackstone Knob West Slide Number 4 No. 31, Chute 1800 ft. long, into Laurel Branch Number 30 Number 25 MP 1.0, I-40 Number 10 Number 48 Number 12 MP 7.95 EBL, N of single tunnel Rt. 215 Number 2 Number 40

35.5200 35.4400 35.4500 35.4500 35.4800 35.4900 35.8900 35.5200 35.5400 35.7200 35.9200 36.0068 35.9865 35.4700 35.4800 35.4900 35.4800 35.7353

82.6400 82.6400 82.6600 82.6500 82.6700 82.6600 82.6600 82.6400 82.8100 83.0400 82.6700 82.6977 82.7611 82.6600 82.6500 82.6600 82.6900 82.3174

Buncombe Buncombe Henderson Buncombe Buncombe Buncombe Madison Buncombe Haywood Haywood Madison Madison Madison Buncombe Buncombe Buncombe Buncombe Buncombe Buncombe Buncombe Buncombe Buncombe Haywood Buncombe Henderson Buncombe Haywood

1977 1977 1977 1977 1977 1977 1980 1977 1984 1979 1990 1999 1999 1977 1977 1977 1977 2004

Debris Flow Debris Flow Debris Flow Debris Flow Debris Flow Debris Flow Slump Debris Flow Slump Slump Debris Flow Debris Flow Debris Flow Debris Flow Debris Flow Debris Flow Debris Flow Debris Flow

Zatm Zatm Zatm Zatm Zatm Zatm Ybgg Zatm ZYbn Zsl Ybgg Zwc Zwc Zatm Zatm Zatm Zatm Zatm Zatm Zatm Zatm Zatm Zsr Zatm Zatm Zatm Zsl Zatm Zatm Zatm Zatm

NC093 NC093 NC095 NC095 NC095 NC095 NC095 NC093 NC006 NC102 NC093 NC092 NC092 NC090 NC093 NC095 NC093 NC095 NC093 NC095 NC095 NC095 NC102 NC093 NC093 NC095 NC102 NC093 NC095 NC095 NC093

27 27 27 28 28 28 28 28 28 28 28 28 28 29 29 29 29 29 29 30 30 30 30 30 30 30 30 31 31 31 31

S N E NE SE S E S S SW N SE E S SE NE NE S NE E NE E SW SW E S N SW E E NE

35.4788 82.6863 35.4600 82.6500 35.4600 35.4800 35.7700 35.4900 35.4300 35.4900 35.7200 35.4900 35.2900 35.4700 35.4400 82.6700 82.6700 83.0800 82.6800 82.6600 82.6600 83.0300

1977 Debris Flow 1977 Debris Flow 1977 Debris Flow 1977 Debris Flow Debris Flow 1977 Debris Flow 1977 Debris Flow 1977 Debris Flow 1973 Debris Flow 1977 Debris Flow Slump 1977 Debris Flow 1977 Debris Flow

Rd Cv

Rd/Cv

Pomeroy Pomeroy NCDOT Pomeroy Pomeroy Pomeroy NCDOT Pomeroy Witt Pomeroy Pomeroy

82.6500 Buncombe 82.9100 Transylvania 82.7000 Buncombe 82.6200 Buncombe

154

Rd Rd Cv? Rd

Pomeroy Pomeroy NCDOT Otteman USFS Pomeroy Otteman Pomeroy Witt Pomeroy DOQQ Pomeroy Pomeroy Pomeroy Pomeroy NCGS Pomeroy Pomeroy Pomeroy NCDOT NCGS Pomeroy Pomeroy Pomeroy Wooten NCDOT Pomeroy Pomeroy NCDOT Pomeroy Witt

Number 44 No. 8, Debris Torrent NC 63, 5.1mi N of Buncombe County line

35.4300 82.6400 35.4800 82.6800 35.7000 82.8200

Henderson Buncombe Madison

1977 Debris Flow 1977 Debris Flow 1980 Slump 1977 Debris Flow 1994 Debris Flow 1977 Debris Flow 1977 Debris Flow 1977 Debris Flow Slump 1977 Debris Flow Slump 1977 Debris Flow 1977 Debris Flow 1977 Debris Flow 1977 Debris Flow 2004 Debris Flow 1977 Debris Flow 1977 Debris Flow 1977 Debris Flow 1985 Slump 2001 Debris Flow 1977 Debris Flow 1977 Debris Flow 1977 Debris Flow 2003 Debris Flow 1986 Slump 1977 Debris Flow 1977 Debris Flow 1950 Slump 1977 Debris Flow Slump

Zatm Zatm Ybgg Zatm Zsl Zatm Zatm Zatm Zsl Zatm Zsp Zatm Zatm Zatm Zatm Zatm Zatm Zatm Zatm Zsl Zwc Zatm Zatm Zatm Zatw Zsr Zatm Zatm Zbgg Zatm Zatm

NC093 NC093 NC095 NC093 NC102 NC095 NC095 NC095 NC102 NC095 NC102 NC095 NC093 NC093 NC093 NC095 NC093 NC095 NC093 NC102 NC092 NC095 NC095 NC095 NC098 NC092 NC095 NC095 NC006 NC095 NC095

31 31 31 31 32 32 32 32 32 33 33 33 33 33 33 33 34 34 34 34 34 35 35 35 35 35 35 35 35 36 36

S SW NE S SE E NE S W E SW E E SE E SW SE NE E SE S E SE E S W E W S E E

Rd rk

Cv

rk

Cv MGW Rd

wr

35.4967 82.6431 Buncombe Microburst; dual slides near 35.7700 83.0400 Haywood Dry Branch Number 32 35.4500 82.6500 Buncombe Number 29 35.4600 82.6800 Buncombe 35.4900 82.6500 Buncombe identified in 3D Analyst 35.7200 86.0400 Haywood Number 28/29 35.4700 82.6600 Buncombe identified from DOQQ 35.7700 83.0900 Haywood Number 11 35.4900 82.6700 Buncombe Number 42 35.4200 82.6200 Henderson Number 42 35.4200 82.6200 Henderson Number 38 35.4400 82.6300 Buncombe Blackstone Knob East Slide 35.7338 82.3154 Buncombe 35.4700 82.6600 Buncombe Number 14 (second head) 35.4900 82.6600 Buncombe Number 5 35.4800 82.6900 Buncombe I-40, East of Twin Tunnels 35.7600 83.0400 Haywood Pounding Mill Branch Slide 35.9951 82.7325 Madison Number 33 35.4600 82.6500 Buncombe Chute 920 ft. long, debris 35.4600 82.6500 Buncombe torrent Number 28/29 35.4700 82.6600 Buncombe Mt. Mitchell Slide 35.7300 82.3000 Buncombe SBL NC 25-70, 250 ft. N of 35.8900 82.7500 Madison SR 1319 Number 3 35.4800 82.6900 Buncombe Debris Torrent 35.4900 82.6800 Buncombe Lesser Fines Creek Slide 35.6700 82.9900 Haywood Number 33 35.4600 82.6500 Buncombe Rt. 215 35.2900 82.9100 Transylvania

155

Rk

Cv

Cv Rd

Pomeroy Pomeroy NCDOT Pomeroy Otteman Pomeroy NCDOT Pomeroy Pomeroy Otteman NCDOT Pomeroy Pomeroy Pomeroy Witt NCDOT NCDOT Pomeroy Pomeroy Pomeroy Pomeroy Pomeroy Pomeroy Pomeroy Pomeroy Pomeroy Pomeroy Pomeroy Pomeroy Pomeroy

Number 8 Debris Torrent MP 0.4, I-40 number 27 Number 27 SBL NC 261 Debris Torrent Debris Torrent number 28 NBL NC 25-70, S of NC 208 Number 46 Number 45 Number 13 Rt. 215 I-40, may have been rockfall, not sure NC 215 debris torrent Ext. rock rubble at toe, multi-headed flow, No. 28 Number 29 Number 27 Number 9 Number 9 Debris Torrent Number 26 Number 1 Number 47 Number 22

35.4800 35.4900 35.7700 35.4468 35.4600 35.4800 36.1000 35.4900 35.4900 35.4600 35.9100 35.4400 35.4400 35.4900 35.2900 35.7800

82.6800 82.6600 83.0900 82.6474 82.6700 82.6700 82.1000 82.6600 82.6600 82.6700 82.7500

Buncombe Buncombe Haywood Henderson Buncombe Buncombe Mitchell Buncombe Buncombe Buncombe Madison

1977 1977 1997 1977 1977 1977 1977 1977 1977 1977 1989

Debris Flow Debris Flow Slump Debris Flow Debris Flow Debris Flow Debris Flow Debris Flow Debris Flow Debris Flow Slump

Zatm Zatm Zsp Zatm Zatm Zatm Ymg Zatm Zatm Zatm Zsr Zatm Zatm Zatm Zatm Zsp Zatm Zatm Zatm Zatm Zatm Zatm Zatm Zatm Zatm Zatm Zatm Zatm Zatm Zatm

NC093 NC095 NC102 NC095 NC095 NC095 NC098 NC095 NC095 NC095 NC092 NC093 NC093 NC095 NC095 NC102 NC095 NC095 NC095 NC095 NC095 NC093 NC093 NC093 NC095 NC095 NC095 NC093 NC095 NC095

36 36 36 36 37 37 37 37 37 38 38 38 38 38 39 39 39 40 40 41 41 41 41 41 41 42 42 46 46 1977

NW SW S S E NE SW SW SW NE NW SE SE SW E S NE SE SE E E W SW W E SE NW SE E SE

Cv wr Rd Rd

82.6500 Henderson 82.6500 Henderson 82.6700 Buncombe 82.9100 Transylvania 83.1000 Haywood Haywood Buncombe Buncombe Buncombe Buncombe Buncombe Buncombe Buncombe Buncombe Buncombe Buncombe Henderson Buncombe Buncombe

1977 Debris Flow 1977 Debris Flow 1977 Debris Flow Slump 1989 Slump 1988 Slump 1977 Debris Flow 1977 Debris Flow 1977 1977 1977 1977 1977 1977 1977 1977 1977 1977 1977 Debris Flow Debris Flow Debris Flow Debris Flow Debris Flow Debris Flow Debris Flow Debris Flow Debris Flow Debris Flow Debris Flow

35.3500 82.9100 35.4600 82.6700 35.4600 82.6700 35.4600 35.4800 35.4900 35.4900 35.4900 35.5000 35.4800 35.4700 35.4400 35.4900 35.4800 82.6700 82.6700 82.6800 82.6800 82.6700 82.6700 82.6700 82.7000 82.6500 82.6600 82.6600

156

Appendix E: SINMAP Results

Default Parameters - 30m DEM Moderately Quasi-Stable Lower Stable Threshold 949.2 1216.3 1114.8 13.4 17.1 15.7 14 41 73 10.0 29.3 52.1 0.015 0.034 0.065 Recharge 50mm 30m DEM Moderately Quasi-Stable Lower Stable Threshold 358.0 544.1 3112.1 5.1 7.7 44.0 2 5 46 1.4 3.6 32.9 0.006 0.009 0.015 Recharge 125mm 30mDEM Moderately Quasi-Stable Lower Stable Threshold 358.0 544.1 3103.3 5.1 7.7 44.0 2 5 46 1.4 3.6 32.9 0.006 0.009 0.015 Recharge 250mm 30m DEM Moderately Quasi-Stable Lower Stable Threshold 358.0 544.1 3102.1

Stable Area (km) % of Region # of Landslides % of Slides LS Density 3796.6 53.5 12 8.6 0.003

Upper Threshold 15.3 0.2 0 0.0 0.000

Defended 0.1 0.0 0 0.0 0.000

Total 7092.3 100.0 140 100.0

Stable Area (km) % of Region # of Landslides % of Slides LS Density 1735.2 24.5 2 1.4 0.001

Upper Threshold 1315.9 18.6 84 60.0 0.064

Defended 3.8 0.1 1 0.7 0.263

Total 7069.1 100.0 140 100.0

Stable Area (km) % of Region # of Landslides % of Slides LS Density 1735.2 24.5 2 1.4 0.001

Upper Threshold 1286.7 18.2 81 57.9 0.063

Defended 28.1 0.4 4 2.9 0.142

Total 7055.4 99.9 140 100.1

Stable Area (km) 1735.2

Upper Threshold 1278.7

Defended 53.6

Total 7071.7

157

% of Region # of Landslides % of Slides LS Density

24.5 2 1.4 0.001

5.1 2 1.4 0.006

7.7 5 3.6 0.009

43.9 46 32.9 0.015

18.1 78 55.7 0.061

0.8 7 5.0 0.131

100.1 140 100.0

Stable Area (km) % of Region # of Landslides % of Slides LS Density 1735.2 24.5 2 1.4 0.001

Recharge 375mm 30m DEM Moderately Quasi-Stable Lower Stable Threshold 358.0 544.1 3102.1 5.1 7.7 43.8 2 5 46 1.4 3.6 32.9 0.006 0.009 0.015

Upper Threshold 1278.7 18.1 77 55.0 0.060

Defended 61.4 0.9 8 5.7 0.130

Total 7079.5 100.1 140 100.0

Stable Area (km) % of Region # of Landslides % of Slides LS Density 2524.7 35.7 8 5.7 0.003

Recharge 125mm, C = .1-.25 30m DEM Moderately Quasi-Stable Lower Upper Stable Threshold Threshold 535.7 844.5 2355.6 811.2 7.6 11.9 33.3 11.5 4 4 63 60 2.9 2.9 45.0 42.9 0.007 0.005 0.027 0.074

Defended 3.8 0.1 1 0.7 0.263

Total 7075.5 100.1 140 100.1

Stable Area (km) % of Region # of Landslides % of Slides LS Density 3673.9 53.0 14 11.5 0.004

Default Parameters - 10m DEM Moderately Quasi-Stable Lower Stable Threshold 868.0 1120.2 1229.4 12.5 16.1 17.7 11 24 69 9.0 19.7 56.6 0.013 0.021 0.056 Recharge 50mm- 10m DEM Quasi-Stable Lower Threshold

Upper Threshold 44.3 0.6 4 3.3 0.090

Defended 0.6 0.0 0 0.0 0.000

Total 6936.4 100.0 122.0 100.0

Stable

Moderately Stable

Upper Threshold

Defended

Total

158

Area (km) % of Region # of Landslides % of Slides LS Density

1591.9 23.1 0.0 0.0 0.000

322.7 4.7 2.0 1.6 0.006

494.4 7.2 5.0 4.1 0.010

2944.9 42.7 33.0 27.0 0.011

1514.0 22.0 82.0 67.2 0.054

24.8 0.4 0.0 0.0 0.000

6892.7 100.0 122.0 100.0

Stable Area (km) % of Region # of Landslides % of Slides LS Density 1591.9 23.2 0.0 0.0 0.000

Recharge 125mm - 10m DEM Moderately Quasi-Stable Lower Stable Threshold 322.7 494.4 2932.5 4.7 7.2 42.7 2.0 5.0 33.0 1.6 4.1 27.0 0.006 0.010 0.011 Recharge 250mm - 10m DEM Moderately Quasi-Stable Lower Stable Threshold 322.7 494.4 2930.9 4.7 7.2 42.5 2.0 5.0 33.0 1.6 4.1 27.0 0.006 0.010 0.011 Recharge 375mm -10m DEM Moderately Quasi-Stable Lower Stable Threshold 322.7 494.4 2930.8 4.7 7.2 42.4 2.0 5.0 33.0 1.6 4.1 27.0 0.006 0.010 0.011

Upper Threshold 1455.2 21.2 74.0 60.7 0.051

Defended 70.8 1.0 8.0 6.6 0.113

Total 6867.5 100.0 122.0 100.0

Stable Area (km) % of Region # of Landslides % of Slides LS Density 1591.9 23.1 0.0 0.0 0.000

Upper Threshold 1434.2 20.8 70.0 57.4 0.049

Defended 126.4 1.8 12.0 9.8 0.095

Total 6900.5 100.0 122.0 100.0

Stable Area (km) % of Region # of Landslides % of Slides LS Density 1591.9 23.0 0.0 0.0 0.000

Upper Threshold 1432.4 20.7 70.0 57.4 0.049

Defended 138.8 2.0 12.0 9.8 0.086

Total 6911.0 100.0 122.0 100.0

Stable Area (km) 2306.1

Recharge 125mm, C = .1-.25 10m DEM Moderately Quasi-Stable Lower Upper Stable Threshold Threshold 495.1 789.9 2269.2 959.3

Defended 50.2

Total 6869.8

159

% of Region # of Landslides % of Slides LS Density

33.6 5.0 4.1 0.002

7.2 4.0 3.3 0.008

11.5 5.0 4.1 0.006

33.0 45.0 36.9 0.020

14.0 56.0 45.9 0.058

0.7 7.0 5.7 0.139

100.0 122.0 100.0

160

Appendix F: SHALSTAB Results

Cohesion 2000, SFA 26 30m DEM NUMSLIDES PERSLIDES CUMSLIDES Area km CUMAREA LS DEN 10 45.45 45.45 243.8 243.8 0.041 5 22.73 68.18 336.2 580.0 0.015 2 9.09 77.27 186.9 766.9 0.011 2 9.09 86.36 173.1 940.0 0.012 0 0.00 86.36 105.0 1044.9 0.000 1 4.55 90.91 13.3 1058.2 0.075 2 9.09 100.00 581.8 1639.9 0.003

PERCENT CUMPERCENT INSTABILITY 14.87 14.87 Chronic Instability 20.50 35.37 < -3.1 11.40 46.76 -3.1 - -2.8 10.55 57.32 -2.8 - -2.5 6.40 63.72 -2.5 - -2.2 0.81 64.53 > -2.2 35.47 100.00 Stable

PERCENT CUMPERCENT INSTABILITY 0.04 0.04 Chronic Instability 2.00 2.04 < -3.1 4.19 6.23 -3.1 - -2.8 9.63 15.85 -2.8 - -2.5 9.20 25.05 -2.5 - -2.2 1.99 27.04 > -2.2 72.96 100.00 Stable

Cohesion 2000, SFA 45 30m DEM NUMSLIDES PERSLIDES CUMSLIDES Area km CUMAREA LS DEN 0 0.00 0.00 0.7 0.7 0.000 3 13.64 13.64 32.8 33.5 0.092 2 9.09 22.73 68.7 102.2 0.029 5 22.73 45.45 157.8 260.0 0.032 1 4.55 50.00 150.8 410.8 0.007 0 0.00 50.00 32.6 443.4 0.000 11 50.00 100.00 1196.6 1639.9 0.009

PERCENT CUMPERCENT 0.00 0.00 0.00 0.00 0.00 0.00 0.01 0.01 0.04 0.05 0.03 0.08 99.92 100.00

Cohesion 9427.41, SFA 45 (SINMAP Upper Bound) 30m DEM INSTABILITY NUMSLIDES PERSLIDES CUMSLIDES Area km CUMAREA LS DEN Chronic Instability 0 0.00 0.00 0.0 0.0 0.000 < -3.1 0 0.00 0.00 0.0 0.0 0.000 -3.1 - -2.8 0 0.00 0.00 0.0 0.0 0.000 -2.8 - -2.5 0 0.00 0.00 0.2 0.2 0.000 -2.5 - -2.2 0 0.00 0.00 0.6 0.8 0.000 > -2.2 0 0.00 0.00 0.5 1.3 0.000 Stable 22 100.00 100.00 1638.7 1639.9 0.013

161

PERCENT CUMPERCENT INSTABILITY 0.04 0.04 Chronic Instability 4.17 4.21 < -3.1 7.18 11.39 -3.1 - -2.8 14.33 25.72 -2.8 - -2.5 12.31 38.03 -2.5 - -2.2 2.53 40.55 > -2.2 59.45 100.00 Stable

Default Results 30m DEM NUMSLIDES PERSLIDES CUMSLIDES Area km CUMAREA LS DEN 0 0.00 0.00 0.7 0.7 0.000 3 13.64 13.64 68.3 69.0 0.044 4 18.18 31.82 117.7 186.7 0.034 6 27.27 59.09 235.1 421.8 0.026 2 9.09 68.18 201.8 623.6 0.010 0 0.00 68.18 41.4 665.0 0.000 7 31.82 100.00 974.9 1639.9 0.007

Cohesion 0, SFA 26 (SINMAP Lower Bound) 30m DEM PERCENT CUMPERCENT INSTABILITY NUMSLIDES PERSLIDES CUMSLIDES Area km CUMAREA LS DEN 25.59 25.59 Chronic Instability 11 50.00 50.00 419.7 419.7 0.026 24.52 50.11 < -3.1 6 27.27 77.27 402.1 821.8 0.015 9.60 59.71 -3.1 - -2.8 3 13.64 90.91 157.5 979.3 0.019 8.72 68.43 -2.8 - -2.5 0 0.00 90.91 143.0 1122.2 0.000 5.20 73.63 -2.5 - -2.2 2 9.09 100.00 85.2 1207.5 0.023 0.38 74.01 > -2.2 0 0.00 100.00 6.3 1213.8 0.000 25.99 100.00 Stable 0 0.00 100.00 426.2 1639.9 0.000

PERCENT CUMPERCENT INSTABILIT 0.04 0.04 Chronic Instability 2.00 2.04 < -3.1 4.19 6.23 -3.1 - -2.8 9.63 15.85 -2.8 - -2.5 9.20 25.05 -2.5 - -2.2 1.99 27.04 > -2.2 72.96 100.00 Stable

Cohesion 0, SFA 45 30m DEM NUMSLIDES PERSLIDES CUMSLIDES Area km CUMAREA LS DEN 0 0.00 0.00 0.7 0.7 0.000 3 13.64 13.64 32.8 33.5 0.092 2 9.09 22.73 68.7 102.2 0.029 5 22.73 45.45 157.8 260.0 0.032 1 4.55 50.00 150.8 410.8 0.007 0 0.00 50.00 32.6 443.4 0.000 11 50.00 100.00 1196.6 1639.9 0.009

Cohesion 2000, STA 35 30m DEM PERCENT CUMPERCENT INSTABILIT NUMSLIDES PERSLIDES CUMSLIDES Area km CUMAREA LS DEN 1.08 1.08 Chronic Instability 3 13.64 13.64 17.6 17.6 0.170 7.55 8.62 < -3.1 4 18.18 31.82 123.8 141.4 0.032 9.66 18.28 -3.1 - -2.8 5 22.73 54.55 158.4 299.8 0.032

162

13.97 9.61 1.73 56.40

32.25 41.86 43.60 100.00

-2.8 - -2.5 -2.5 - -2.2 > -2.2 Stable

2 2 0 6

9.09 9.09 0.00 27.27

63.64 72.73 72.73 100.00

229.1 157.6 28.4 925.0

528.9 686.6 715.0 1639.9

0.009 0.013 0.000 0.006

PERCENT CUMPERCENT 3.34 3.34 14.04 17.38 13.00 30.38 14.94 45.31 9.42 54.73 1.49 56.23 43.77 100.00

INSTABILIT Chronic Instability < -3.1 -3.1 - -2.8 -2.8 - -2.5 -2.5 - -2.2 > -2.2 Stable

Cohesion 0, STA 35 30m DEM NUMSLIDES PERSLIDES CUMSLIDES Area km CUMAREA LS DEN 3 13.64 13.64 54.8 54.8 0.055 9 40.91 54.55 230.2 285.1 0.039 2 9.09 63.64 213.1 498.2 0.009 3 13.64 77.27 244.9 743.1 0.012 1 4.55 81.82 154.5 897.6 0.006 0 0.00 81.82 24.5 922.1 0.000 4 18.18 100.00 717.8 1639.9 0.006

Total 10m DEM Haywood - 26 1922 0 2 PERCENT CUMPERCENT INSTABILIT NUMSLIDES PERSLIDES CUMSLIDES Area km CUMAREA LS DEN 32.57 32.57 Chronic Instability 20.0 86.96 86.96 534.8 534.8 0.037 17.09 49.66 < -3.1 0.0 0.00 86.96 280.7 815.5 0.000 8.52 58.18 -3.1 - -2.8 3.0 13.04 100.00 139.9 955.4 0.021 9.14 67.32 -2.8 - -2.5 0.0 0.00 100.00 150.1 1105.5 0.000 6.60 73.92 -2.5 - -2.2 0.0 0.00 100.00 108.3 1213.8 0.000 3.05 76.97 > -2.2 0.0 0.00 100.00 50.2 1264.0 0.000 23.04 100.01 Stable 0.0 0.00 100.00 378.3 1642.3 0.000 Total 10m DEM Haywood - 26 1922 2000 2 PERCENT CUMPERCENT INSTABILIT NUMSLIDES PERSLIDES CUMSLIDES Area km CUMAREA LS DEN 21.35 21.35 Chronic Instability 13.0 56.52 56.52 350.6 350.6 0.037 14.85 36.20 < -3.1 5.0 21.74 78.26 243.7 594.3 0.021 9.28 45.48 -3.1 - -2.8 2.0 8.70 86.96 152.4 746.7 0.013 10.92 56.40 -2.8 - -2.5 2.0 8.70 95.65 179.3 926.0 0.011 8.33 64.73 -2.5 - -2.2 0.0 0.00 95.65 136.8 1062.8 0.000 4.02 68.75 > -2.2 0.0 0.00 95.65 66.0 1128.8 0.000 31.26 100.01 Stable 1.0 4.35 100.00 513.3 1642.1 0.002

163

PERCENT CUMPERCENT 3.16 3.16 6.85 10.01 7.26 17.27 12.95 30.22 12.78 43.00 6.83 49.83 50.17 100.00

Total 10m DEM Haywood - 35 1922 2000 2 INSTABILIT NUMSLIDES PERSLIDES CUMSLIDES Area km CUMAREA LS DEN Chronic Instability 6.0 26.09 26.09 51.9 51.9 0.116 < -3.1 7.0 30.43 56.52 112.6 164.5 0.062 -3.1 - -2.8 1.0 4.35 60.87 119.2 283.7 0.008 -2.8 - -2.5 4.0 17.39 78.26 212.7 496.4 0.019 -2.5 - -2.2 3.0 13.04 91.30 209.9 706.3 0.014 > -2.2 0.0 0.00 91.30 112.1 818.4 0.000 Stable 2.0 8.70 100.00 823.9 1642.3 0.002 Total 10m DEM Haywood - 35 1922 0 2 INSTABILIT NUMSLIDES PERSLIDES CUMSLIDES Area km CUMAREA LS DEN Chronic Instability 10.0 43.48 43.48 115.2 115.2 0.087 < -3.1 5.0 21.74 65.22 183.9 299.1 0.027 -3.1 - -2.8 1.0 4.35 69.57 161.4 460.5 0.006 -2.8 - -2.5 5.0 21.74 91.30 239.0 699.5 0.021 -2.5 - -2.2 1.0 4.35 95.65 204.7 904.2 0.005 > -2.2 0.0 0.00 95.65 103.0 1007.2 0.000 Stable 1.0 4.35 100.00 635.0 1642.2 0.002

PERCENT CUMPERCENT 7.02 7.02 11.20 18.22 9.83 28.05 14.55 42.60 12.46 55.06 6.27 61.33 38.67 100.00

Total 10m DEM Haywood - 35 1922 9427.41 2 PERCENT CUMPERCENT INSTABILIT NUMSLIDES PERSLIDES CUMSLIDES Area km CUMAREA LS DEN 0.00 0.00 Chronic Instability 0.0 0.00 0.00 0.0 0.0 0.000 0.11 0.11 < -3.1 2.0 8.70 8.70 1.8 1.9 1.111 0.24 0.35 -3.1 - -2.8 1.0 4.35 13.04 3.9 5.8 0.256 0.92 1.27 -2.8 - -2.5 2.0 8.70 21.74 15.1 20.9 0.132 2.40 3.67 -2.5 - -2.2 3.0 13.04 34.78 39.4 60.3 0.076 2.59 6.26 > -2.2 2.0 8.70 43.48 42.6 102.9 0.047 93.74 100.00 Stable 13.0 56.52 100.00 1539.3 1642.2 0.008 Total 10m DEM Haywood - 45 1922 0 2 PERCENT CUMPERCENT INSTABILIT NUMSLIDES PERSLIDES CUMSLIDES Area km CUMAREA LS DEN 0.34 0.34 Chronic Instability 0.0 0.00 0.00 5.6 5.6 0.000 2.29 2.63 < -3.1 8.0 34.78 34.78 37.6 43.2 0.213 3.37 6.00 -3.1 - -2.8 2.0 8.70 43.48 55.4 98.6 0.036

164

8.46 11.85 7.70 65.99

14.46 26.31 34.02 100.00

-2.8 - -2.5 -2.5 - -2.2 > -2.2 Stable

3.0 7.0 0.0 3.0

13.04 30.43 0.00 13.04

56.52 86.96 86.96 100.00

139.0 194.6 126.5 1083.6

237.5 432.1 558.6 1642.2

0.022 0.036 0.000 0.003

PERCENT 0.09 0.99 1.69 4.92 8.50 6.39 77.41

Total 10m DEM Haywood - 45 1922 2000 2 CUMPERCENT INSTABILIT NUMSLIDES PERSLIDES CUMSLIDES Area km CUMAREA LS DEN 0.09 Chronic Instability 0.0 0.00 0.00 1.6 1.6 0.000 1.09 < -3.1 5.0 21.74 21.74 16.3 17.9 0.307 2.78 -3.1 - -2.8 2.0 8.70 30.43 27.8 45.7 0.072 7.70 -2.8 - -2.5 3.0 13.04 43.48 80.8 126.5 0.037 16.20 -2.5 - -2.2 4.0 17.39 60.87 139.5 266.0 0.029 22.59 > -2.2 2.0 8.70 69.57 104.9 370.9 0.019 100.00 Stable 7.0 30.43 100.00 1271.3 1642.2 0.006 Total 10m DEM Haywood - 45 1922 9427.41 2 INSTABILIT NUMSLIDES PERSLIDES CUMSLIDES Area km CUMAREA LS DEN Chronic Instability 0.0 0.00 0.00 0.0 0.0 0.000 < -3.1 0.0 0.00 0.00 0.1 0.1 0.000 -3.1 - -2.8 1.0 4.35 4.35 0.1 0.2 8.598 -2.8 - -2.5 0.0 0.00 4.35 0.5 0.7 0.000 -2.5 - -2.2 1.0 4.35 8.70 2.4 3.1 0.417 > -2.2 0.0 0.00 8.70 5.4 8.5 0.000 Stable 21.0 91.30 100.00 1633.7 1642.2 0.013

PERCENT CUMPERCENT 0.00 0.00 0.00 0.00 0.01 0.01 0.03 0.04 0.15 0.19 0.33 0.52 99.48 100.00

Total Default Values - 10m DEM Haywood PERCENT CUMPERCENT INSTABILIT NUMSLIDES PERSLIDES CUMSLIDES Area km CUMAREA LS DEN 0.34 0.34 Chronic Instability 0 0.00 0.00 5.6 5.6 0.000 3.96 4.30 < -3.1 8 34.78 34.78 65.0 70.6 0.123 5.21 9.51 -3.1 - -2.8 2 8.70 43.48 85.5 156.1 0.023 12.38 21.89 -2.8 - -2.5 5 21.74 65.22 203.2 359.3 0.025 15.61 37.50 -2.5 - -2.2 6 26.09 91.30 256.3 615.6 0.023 9.52 47.02 > -2.2 0 0.00 91.30 156.4 772.0 0.000 52.98 100.00 Stable 2 8.70 100.00 870.1 1642.1 0.002

165