Copyright (c) 2017 Association of Environmental & Engineering Geologists (AEG). Creative Commons license Attribution-Non-Commercial 4.0 International (CC BY-NC 4.0)

The North Carolina Geological Survey’s Response to Landslide Events: Methods, Findings, Lessons Learned, and Challenges Wooten, R.M., Rick.Wooten@ncdenr.gov; Cattanach, B.L., Bart.Cattanach@ncdenr.gov; Bozdog, G.N., Nick.Bozdog@ncdenr.gov; Isard, S.J., Sierra.Isard@ncdenr.gov; North Carolina Geological Survey, 2090 US 70, Swannanoa, NC 28778 Fuemmeler, S.J., Stephen@appalachianlandslideconsultants.com , Bauer, J.B., Jennifer@appalachianlandslide.com ; Appalachian Landslide Consultants, P.O. Box 5516, Asheville, NC 28813 Witt, A.C. Anne.Witt@dmme.virginia.gov ; Virginia Div. of Minerals and Geology, 900 Natural Resources Dr., Charlottesville, VA 22903 Douglas, T.J., Tommy.Douglas@ncdot.gov; North Carolina Dept. of Transportation, 4142 Haywood Road, Mills River, NC 28759 Gillon, K.A., gillonk@bellsouth.net; 18 Amber Drive, Horse Shoe, NC 28742 Latham, R.S., Rebeccaslatham@yahoo.com 13709 Portpatrick Lane, Matthews NC 28105 ABSTRACT: From 1990 to 2016 the North Carolina Geological Survey (NCGS) responded to 130 requests for assistance on landslide events from government agencies, the public, and consultants. As a result, NCGS geologists investigated 202 landslides in the Blue Ridge Mountains of western North Carolina which have destroyed 27 homes and damaged 42 others, damaged 79 roads, and killed 5 people and an unborn child. Field investigations of 183 landslides determined that 89% occurred where slope modifications by humans were contributing factors, with the majority being fill failures that mobilized into damaging debris flows and debris slides. Landslides that occurred where there were existing landslide hazard maps correlate strongly with areas where the maps show an increased potential for debris flows and debris slides. Most landslides corresponded with periods of heavy rainfall from tropical cyclones Frances and Ivan (2004) and Cindy and Ernesto (2005), or with periods of above normal rainfall in 2009-2010 and 2013. Correlations between rainfall and debris flow occurrences indicate that debris flows on slopes modified by human activity can be triggered by rainfall events with lower rates and durations than events that trigger debris flows on unmodified slopes. Development of a landslide geodatabase used with field computers, LiDAR digital elevation models and orthophotography has

359

3rd North American Landslide Symposium, Roanoke, VA, June 4-8, 2017

improved pre-response preparation, data collection and analysis, and communication of results and delivery of geospatial data to stakeholders. This paper highlights case examples; discusses methods and objectives, lessons learned, and challenges in responding to landslide events as a state agency.

INTRODUCTION From 1990 to 2016 the North Carolina Geological Survey (NCGS) responded to 130 requests for assistance on landslide events from government agencies, the public, and consultants. As a result, NCGS geologists investigated 202 landslides in the Blue Ridge Mountains of western North Carolina (Fig. 1) which have destroyed 27 homes and damaged 42 others, damaged 79 roads, and killed 5 people and an unborn child. The main objective of the NCGS in providing assistance on landslide events is to help protect life, health, safety and property by providing stakeholders with unbiased scientific information. This information is aimed at helping stakeholders understand the scope and magnitude of a landslide, the associated immediate and future potential hazards and risks, and to provide general guidance that they may pursue in the short and long terms. Importantly, the responses and subsequent investigations are neither intended to be, nor carried out as, forensic investigations for the purposes of establishing responsibility or liability of the parties involved.

In the aftermath of a landslide, information is gathered to increase situational awareness and improve the safety of emergency responders during rescue and recovery operations (Taylor, 2017). Response efforts can include assisting emergency managers with post-landslide contingency planning, such as preparing for rainfall scenarios when evacuations may be warranted where unstable slopes remain above locations vulnerable to damage. Post-event assistance can include providing documentation to support efforts by state and local agencies in obtaining funds for recovery and mitigation efforts such as the stabilization of hazardous slopes. Emergency situations involve communicating with the media in order to convey objective, scientific information about the nature of landslides and recovery efforts. Important aspects of the responses are the investigations into the geologic characteristics of these landslides, factors contributing to their causes and triggers, and the resulting damage. Landslide data are documented in a geographic information system (GIS) geodatabase, and in formal reports

and publications when appropriate. Existing information in the landslide geodatabase serves an important function in pre-response planning, and data collected during a response goes into the geodatabase as a feedback loop for future responses. The compilation and analysis of this data is to make scientific information available on landslide events so that individuals or agencies can take actions towards preparing for future landslide events and reducing losses from them.

RESPONSE SUMMARY Table 1 shows landslide types and corresponding slope configurations for responses by emergency, urgent, and routine categories as defined in Table 2. Fast moving debris flows and debris slides dominated among landslide types and, the majority of the investigated landslides originated slopes where modifications by human activity (e.g., fill slopes, cut slopes, retaining walls) contributed to the slope instability. Landslides originating on modified slopes occurred in 92.3% of the emergency cases; 96.8% of the urgent cases, and 63.9% of the routine cases. Fill slope failures dominated within each of the response categories comprising 76.9% of the emergency cases, 57.6% of the urgent cases, and 46.4% of the routine cases. Field reviews of 183 of the 202 landslides investigated in total determined that 89% occurred on modified slopes.

METHODS Data collection methods have evolved considerably from the paper maps and field notebooks in the 1990’s, superseded by data acquisition with hand-held global positioning satellite (GPS) units and then importing the data into a landslide database and a GIS. An improvement to landslide documentation and response was the development of a GIS landslide geodatabase compatible with GPS-equipped field computers (Witt et al., 2008, Bauer et al., 2012). We also use personal GPS-equipped mobile phones with internet access to augment our landslide response efforts.

360

Wooten, R.M., Cattanach, B.L., Bozdog, G.N., Isard, S.J., Fuemmeler, S.J., Bauer, J.B., Douglas, T.J.

Figure 1. Map showing landslide responses by category and other landslide locations documented in the NCGS landslide geodatabase. Labeled names correspond to case examples of landslide responses described in the paper.

The geodatabase is structured to allow the input of mapped landslide features such as scarps, tension cracks, outlines of the areas affected by landslides, and landslide deposits. Point data can also be collected to record information on soil, sediment, and bedrock lithology and structures, slope configurations, and other landslide attributes. Bauer et al. (2012), and Witt and Wooten (2015) give further descriptions of the geodatabase. Field laptops utilize the geodatabase with a variety of base maps derived from LiDAR digital elevation models (e.g., shaded relief, slope, contour elevation maps), and ortho-imagery, and digital geologic maps.

Using the geodatabase augments the data collection and mapping aspects of field investigations, and improves our ability to provide stakeholders with map products and geospatial data during various stages of an investigation. NCGS geologists use the geodatabase to gain advance knowledge of landslide features that have previously been recorded in an area prior to the field phase of a landslide response. In addition to its application to landslide responses, the geodatabase is a tool used by emergency managers to update state and local hazard mitigation plans in order to quantify the severity and frequency of events with single or multiple landslides. Figure 2 outlines the conceptual framework for the iterative

phases of landslide response activities and documentation in the landslide geodatabase. Organizational procedures followed in landslide responses have evolved from the early, less structured responses that occurred from 1990-2001. Through the emergency response experiences in 2003, 2004, 2009, 2010, and 2015, the NCGS procedures became more integrated with state and local emergency management protocols (Fig. 2) primarily though improvements in the chain-of-communications, and response documentation. In 2015 the NCGS developed guidelines for landslide emergency responses that includes a chain-of-communications for notifications through the State Emergency Operations Center. The guidelines outline preparations for field mobilization, safety considerations, and investigation objectives. The guidelines are available online on the North Carolina Geological Survey website.

361

3rd North American Symposium on Landslides, Roanoke, VA, June 4-8, 2017

Table 1. Chart showing landslide types and corresponding slope configurations by emergency, urgent and routine response categories. Total landslides = 202.

Table 2. Definitions of emergency, urgent and routine landslide response categories.

Figure 2. Conceptual framework for the iterative phases of landslide response activities and documentation in the landslide geodatabase. Asterisked items typically recur in the various phases of a response. EOC = Emergency Operations Center; EM = Emergency Management.

EmergencyLandslide

TypeUnmodi-

fiedFill Cut Retaining

WallTotal

debris flow 1 4 − 1 6debis slide − 5 − − 5

debris slide-flow

−1

− −1

weathered-rock slide

− −1

−1

rock slide − − − − −rock fall − − − − −

Total 1 10 1 1 13% Total (n=13) 7.7% 76.9% 7.7% 7.7% 100.0%

Urgentdebris flow 1 19 5 1 26debis slide 1 33 10 2 46weathered-rock slide

−1 11

−12

rock slide − − 6 − 6rock fall 1 − 1 − 2

Total 3 53 33 3 92% Total (n=92) 3.3% 57.6% 35.9% 3.3% 100.0%

Routinedebris flow 28 24 52debis slide 4 21 5 1 31weathered-rock slide

− −8

−8

rock slide 1 − 1 − 2rock fall 2 − 2 − 4

Total 35 45 16 1 97% Total (n=97) 36.1% 46.4% 16.5% 1.0% 100.0%

Emergency – response time is critical; involves loss of life, serious injury, severe damage or blocked access to homes; poses an immediate threat to safety or property damage; immediate actions may be required to reduce the level of threat, (e.g., evacuations, repairs); involves emergency management agencies. Urgent - response time is important, but not critical; poses a potential, but not immediate threat to safety or property damage; actions may be required to mitigate a future serious threat. Routine - response time is not critical; poses a low potential threat to safety or property damage; actions may be required to prevent an increased level of threat. Response may be for informational purposes only.

362

Wooten, R.M., Cattanach, B.L., Bozdog, G.N., Isard, S.J., Fuemmeler, S.J., Bauer, J.B., Douglas, T.J.,

CASE EXAMPLES The seven examples of landslide responses summarized in this section highlight a variety of damaging landslide types, corresponding slope configurations, and response efforts. (Table 3.) Table 3. Case example landslide events by date, type, slope configuration, and affected area.

Lands Creek Debris Flows The NCGS responded to two debris flows in the Lands Creek watershed in Swain County (Fig. 1). An emergency investigation of the December 23, 1990 Lands Creek I debris flow determined that the 670 m-long debris flow that destroyed a mobile home and the chlorinator for the Bryson City water system had originated on a logging road fill slope (Wooten et al., 2007). During a May 5-7, 2003 rainfall event the Lands Creek II debris flow originated on a fill slope on the same logging road, and deposited sediment and large woody debris into the reservoir that supplies water to the chlorinator rebuilt after the 1990 event. In both cases, poorly constructed road fills placed in colluvial hollows contributed to the slope failures. The Lands Creek debris flows, among others, are also examples of the potential for fill and cut slope instability related to acid-producing, sulfidic rock horizons (Wooten and Latham, 2004; Latham et al., 2009). Peeks Creek Debris Flow The September 16, 2004 Peeks Creek debris flow (Fig. 1) triggered by rainfall from the remnants of Hurricane Ivan led to a major emergency response effort (Wooten et al., 2008a, Latham et al., 2007). The debris flow initiated on an unmodified, forested

slope, and killed four people and an unborn child, and destroyed 16 homes along its 3.6 km-long path. The NCGS provided consultation to emergency managers on stability of the slopes during recovery operations, mapped the extent of the debris flow as part of the landslide hazard mapping of Macon County (Wooten et al., 2006), and served on the interdisciplinary Peeks Creek task force that prepared a report documenting the debris flow. Bear Trail and Ghost Town Debris Flows Two emergency responses included the Bear Trail and Ghost Town debris flows in Maggie Valley, Haywood County (Fig. 1). The January 7, 2009 Bear Trail debris flow initiated in a fill placed in a colluvial hollow, and destroyed an occupied home (Fig. 3), and severely damaged a N.C. Department of Transportation (NCDOT) road (Wooten et al., 2009b). In 2005, County inspectors had identified tension cracks in the fill indicating the potential instability of the slope.

Figure 3. Photograph of the house destroyed by the Bear Trail debris flow (fill failure). The 200 m-long debris flow originated on the slope in the upper right of the photograph. Surprisingly, occupants of the house received relatively minor physical injuries. The Ghost Town debris flow of February 5, 2010 initiated as a retaining wall failure and damaged three homes, a NCDOT road, and caused the temporary evacuation of about 40 people. The nearly 38,000 m3 of unstable fill and retaining wall material that remained on the slope of the amusement park after the debris flow threatened

Landslide Event Date

Landslide Type

Slope Configuration

Area (m2)

Lands Creek I Dec. 23, 1990 debris flow fill 16,550

Lands Creek II May 5-7, 2003 debris flow fill 26,060

Peeks Creek Sept. 16, 2004 debris flow unmodified 138,570

Bear Trail Jan. 7, 2009 debris flow fill 4075

Ghost Town Feb. 5, 2010 debris flow retaining wall 26,010Occoneechee

Mtn. Feb. 18, 2001 rock slide cut 1010

Chimney Rock Nov. 14, 2012 rock fall debris slide

unmodified 7235

363

3rd North American Symposium on Landslides, Roanoke, VA, June 4-8, 2017

public safety, and required stabilization. Photographs by a local newspaper reporter revealed that scarps had developed in the fill behind the wall by May 2009. Contingency planning with emergency managers included modeling the potential inundation area should the slope fail prior stabilization (Witt, et al., 2012). The NCGS assisted local governments by mapping and estimating the volume of landslide material to be removed from a stream. We also provided documentation to support a successful application for federal emergency watershed protection funds to remove the damaged retaining wall and unstable fill, and flatten the slope. The regraded fill slope failed during the January 14, 2013 rain event; however, a sediment basin constructed during the 2011 repair prevented major damage below Ghost Town property. Figure 4 shows the extents of a 1997(?) failure, and the 2010 and 2013 failures.

Figure 4. Paths of (1997(?), 2010 and 2013 debris flows associated with the Ghost Town Amusement Park in Maggie Valley. Map base is 2015 orthophotography.

Occoneechee Mountain - Chimney Rock Rock slope failures constitute only 7% of the total landslide response efforts; however, two events in state parks are examples of urgent responses to

assist the N.C. Division of Parks and Recreation. A 2001 rockslide on a cut slope in an abandoned quarry threatened the safety of an overlook platform on Occoneechee Mountain (Fig. 1) in Eno River State Park (Wooten and Clark, 2001; Wooten et al., 2006). The NCGS and the NCDOT Geotechnical Unit jointly assessed the stability of the slope by analyzing bedrock lithology and structure, and past rock slides in the quarry. We developed conceptual stabilization and mitigation alternatives including removal of the observation deck, the action taken by the Division of Parks and Recreation.

During the night of November 14, 2012 a major rock fall occurred in Chimney Rock State Park in Hickory Nut Falls Gorge (Figs. 1, 5). A 1270-1800 metric ton rock slab dislodged from an overhanging outcrop ledge about 100m above the trail to Hickory Nut Falls. The block fell and shattered into boulders when it struck the rock slope immediately below the overhang. A portion of the boulder severely

Figure 5. Locations of the Nov. 14, 2012 rock fall-debris slide in Chimney Rock State Park, and debris flows near the Town of Chimney Rock. Point locations for landslide initiation sites are color-coded by event date: green dots - July 15-16, 1916; purple dots - Sept. 4, 1996; red dots - Nov. 14, 2012 rock fall-debris slide; blue dots - other dates; Qd = Quaternary debris deposits. Sources: NCGS landslide geodatabase; Soplata, 2016.

damaged a 65-meter-long section of trail and destroyed a section of a steel girder footbridge. The impact of some of the rock fall boulders triggered the movement of an existing colluvial deposit that resulted in approximately 2,300m3 of unstable

364

Wooten, R.M., Cattanach, B.L., Bozdog, G.N., Isard, S.J., Fuemmeler, S.J., Bauer, J.B., Douglas, T.J.,

material remaining on the slope above the trail. Fortunately, the rock fall occurred after park hours, as the Hickory Nut Falls trail is a popular trail. Our report on the rock fall event (Wooten et al., 2012) supported the Division of Parks and Recreation in obtaining funds to reconstruct the trail and remove the hazardous material above the trail before it reopened to the public.

The Hickory Nut Gorge area (Fig. 5) has a long history of historical landslide events including those in 1916, 1994, 1996, 2008 and 2014 (Soplata, 2016). Extensive foot slope deposits of rock blocks and boulders reveal that the gorge walls have been prone to numerous Quaternary landslides (Soplata, 2016). The mapping and documentation of these landslides and deposits proved to be particularly relevant following a November 2016 wildfire (Party Rock fire) which burned areas of the steep, south-facing slopes above the Town of Chimney Rock (Fig. 5). The NCGS is currently coordinating with the U.S. Geological Survey, and Chimney Rock State Park to monitor post-fire landslide occurrence in the Party Rock fire area.

CORRELATIONS OF LANDSLIDES WITH LANDSLIDE HAZARD MAPS

Stability index maps are components of the NCGS landslide hazard maps for Macon, Watauga, Buncombe and Henderson Counties (Wooten et al., 2006, 2008b, 2009a, 2011). We generated the stability index maps using the SINMAP (Stability Index MAPping) software developed for use in a GIS by Pack et al., (1998). Wooten et al., (2007), and Bauer et al., (2012) give further information on the methods used to create the stability index maps. The stability index maps show by color code the relative susceptibility for shallow translational slope failures (e.g., debris flows and debris slides) to initiate on unmodified slopes in response to a rainfall event of 125mm (5in) or more within 24 hours (Fig. 6). We used the mapped initiation sites of debris flows and debris slides on unmodified slopes to calibrate the stability index maps as recommended by Pack et al., (1998). These initiation sites, therefore, are strongly correlated with the high hazard zones (i.e., factors of safety <1) of the stability index maps. We did not use the initiation sites of debris flows and debris slides originating on modified slopes to calibrate the

stability index maps because the infinite slope model, the governing equation for SINMAP, is not ideally suited for analyzing the stability of non-uniform slopes with cuts and fills (Pack, et al., 1998). During the course of work on the landslide hazard maps and since their completion, the NCGS has documented 55 landslides that initiated on modified slopes in those counties, and have responded to 51 of these landslide events. Although the stability index maps are not intended to apply directly to the stability of cut and fill slopes, there is a strong correlation between landslides on these slopes with areas mapped in the high and moderate hazard zones. Of the 55 debris flows and debris slides on modified slopes, 85% originated in the high hazard zone, or within areas with both high and moderate hazard zones on the stability index maps. Within the two map categories above, 45% originated in the high hazard zone; whereas, 40% originated within areas with both high and moderate hazard zones on the stability index maps. Figure 6 shows a map of the January 16, 2013 Watauga Vista debris flow (fill failure) in Macon County that originated in the high hazard zone of the stability index map.

One example where the spatial correlation is not strong with a landslide and the high hazard zone of the stability index map is an 8600m2 reactivated debris slide in Buncombe County. In this case, only 7% of the debris slide (located in the head scarp area) occurred in contiguous high and moderate hazard zones of the stability index map. The entire slide, however, is within mapped debris deposits that are identified on the county landslide hazard map (Wooten et al., 2009a) as potentially unstable, and where ground disturbance should be done with caution.

The correlations of landslides on modified slopes with the high and moderate hazard zones of the stability index maps occurs partly because the steep slopes of cuts and fills are captured by the LiDAR digital elevation model used to create the maps. The infinite slope model, the governing equation for the stability index mapping, is sensitive to ground slope (Pack, et al., 1998); therefore, steep cut or fill slopes are typically included in the high and moderate hazard zones. Where these slopes coincide with steep, high cuts in residual or

365

3rd North American Symposium on Landslides, Roanoke, VA, June 4-8, 2017

Figure 6. Map showing the initiation site (yellow dot) of the Watauga Vista debris flow (fill failure) in the high hazard zone of the stability index map. The chart shows the relative stability index map rankings based on factors of safety in response to a rainfall event of 125mm (5in) or more within 24 hours. Map base is 2015 orthophotography. colluvial soil, or poorly constructed fills, they can be prone to failure. For this reason, the stability index map can be used as a screening tool to identify cut and fill slopes that could be prone to failure.

RAINFALL AND LANDSLIDES

Most landslide events involving a response by the NCGS corresponded with periods of widespread

heavy rainfall from tropical cyclones Frances and Ivan in 2004, and Cindy and Ernesto in 2005; with storms that occurred during periods of regional above normal (30-yr average) rainfall in 2009-2010 and 2013; or, with more localized heavy rainfall events in 1990 (Wooten et al., 2007), 1996 (Fig. 7), 2003 (Wooten and Latham, 2004), and 2011 (Fig. 7).

The fatal Dogwood Drive debris flow (Fig. 1) of December 11, 2003 in Maggie Valley was a notable exception to a landslide triggered by a major rainfall event (Wooten and Carter, 2004). Nearby rain gauges recorded only 18-21mm of rainfall in the 24 hours preceding the debris flow. Litigation over liability for the slope failure revealed that a leaking water supply line was the likely source of water which infiltrated a road fill and triggered the debris flow that destroyed a house killing the occupant.

Two rainfall scenarios suggest that rainfall thresholds necessary to induce slope failures on modified slopes are less than those necessary to trigger slope failures on forested slopes not modified by human activity. Western North Carolina received approximately 1066mm of rainfall from August 2009 through February 2010, nearly 660mm above normal (Bauer et al., 2010). The NCGS investigated 15 landslides, mainly debris flows, during this period. All of these slope failures, (including the Bear Trail and Ghost Town debris flows (Figs. 1, 3, 4), originated on modified slopes; and, the NCGS has not documented any landslides that initiated on unmodified slopes during the 2009-2010 period. From January through August, 2013 record rainfall fell throughout most of western North Carolina. The National Weather Service (NWS) recorded 1730 mm of cumulative rainfall for the year at the Asheville airport by the end of August, 585 mm above normal. From July through August 2013, 335 reported landslides occurred throughout western North Carolina in response to four storm events (Wooten et al. 2016). The NCGS investigated 33 of these landslides that resulted in 5 destroyed or condemned homes and damage to 4 other homes and 24 roads. Information to date indicates that only two of the 335 landslides originated on unmodified slopes, with the remainder involving slopes modified by human activity; again, mainly fill slope failures that mobilized into damaging debris flows and slides.

366

Wooten, R.M., Cattanach, B.L., Bozdog, G.N., Isard, S.J., Fuemmeler, S.J., Bauer, J.B., Douglas, T.J.,

Figure 7 shows values of peak rainfall rates and durations that triggered selected debris flow events on unmodified and modified slopes. Examples are given where reliable estimates of rainfall rates and durations correspond with reasonably well- constrained times of the debris flow events. Figure 8 shows the locations of the investigated debris flows and the respective rain gauges used in this analysis. Although limited, these data indicate that, in general, debris flows originating on modified slopes with decreased stability can be triggered by rainfall events with lower levels than those needed to trigger debris flows on forested, unmodified slopes.

Figure 7. Graph showing peak rainfall rate and duration for selected debris flow events in western North Carolina. Debris flow slope configurations: UM = modified and unmodified slopes; U = unmodified slopes. M = modified slopes. *Times of debris flows known. **Times of debris flows reasonably inferred. See Figure 8 for landslide and rain gauge locations, and Table 4 for sources of landslide and rainfall data. Table 4. Sources of debris flow and rainfall data for Figure 7.

Figure 8. Map showing locations of debris flows and rain gauge locations used for Figure 7. *Times of debris flows known; **Times of debris flows reasonably inferred. Debris flow locations from the NCGS landslide geodatabase. High antecedent moisture conditions prior to storm events are important triggering scenarios for debris flows in western North Carolina (Furhmann et al, 2008; Wooten et al., 2016). Above normal rainfall preceded the debris flows at Bent Creek (Neary and Swift, 1987), Peeks Creek and Wayah (Wooten et al., 2008a), and Bear Trail, and Ghost Town (Bauer et al., 2010). The elevated moisture levels prior to these debris flows may be why the triggering rainfall rates and durations fall into the lowest group of values shown in Figure 7.

Because the majority of North Carolina landslides are triggered by rainfall, coordination between the NCGS and meteorologists at the NWS Greenville-Spartanburg Weather Forecast Office has been ongoing since the tropical cyclones Frances and Ivan in September 2004. This coordination involves interdisciplinary workshops (Bauer, et al., 2011), data sharing, and collaboration on publications (e.g., Wooten et al., 2008a). Through these interactions the NWS developed operational guidelines for including statements on landslide hazards in their public advisories for floods, and flash floods.

LESSONS LEARNED AND CHALLENGES

Responding to requests for assistance on landslides remains a priority for the NCGS. A challenge is to balance the resources committed to such responses

Rainfall-Debris Flow Event Data Source

Watauga 1940 Wieczorek et al., 2004; Wooten et al., 2008Bent Creek 1977 Neary and Swift, 1987; Wooten et al., 2009aHickory Nut Gorge 1996 Johnstone and Burrus, 1998; Soplata, 2016Peeks Creek, Wayah 2004 Wooten et al., 2008Bear Trail 2009 Wooten et al, 2009bGhost Town 2010 Wooten et al., 2010; Witt et al., 2012Balsam Mountain 2011 Miller et al., 2012; Tao and Barros, 2014

367

3rd North American Symposium on Landslides, Roanoke, VA, June 4-8, 2017

with other program requirements within the agency’s mission and authority. Fifteen of the landslide events covered here involved litigation. Although the NCGS was not a party in any litigation, producing documents and digital files for discovery, depositions, and court testimony required significant staff time. The NCGS policy is to remain a neutral party in litigation with the objective to provide unbiased and defensible information to all parties about the geologic characteristics of a landslide. As geologists we do not make engineering recommendations for stabilizing slopes, and refer stakeholders to geotechnical engineers in these cases. We have identified cut and fill slopes as significant factors that contributed to instability in the majority of our landslide responses. In some cases, we identified pre-existing indicators of instability (e.g., subsidence, tension cracks, and scarps) in our investigations of these slopes. The extents of our investigations, however, do not determine the parties responsible for the design, construction or monitoring of the slopes. The term “landslide” in particular, proves to be problematic when legal cases involve denied insurance claims for damages from landslides. Insurance policies may use terms like earth movements, landslides, or mudslides; but not necessarily standard landslide classification terms such as debris flow, debris slide, and rock slide, as used by the scientific community (e.g., Cruden and Varnes, 1996). In one case, a significant part of the inquiry during deposition and testimony involved the question if a debris flow is a type of landslide. The distinction between landslide causes and triggers are an important scientific concept (e.g., Wieczorek, 1987), and in context of litigation requires sometimes lengthy explanation with regard to the various factors contributing to landslides.

CONCLUSIONS

The NCGS landslide geodatabase is an effective way to record geospatial and temporal attributes of landslides. The geodatabase facilitates data acquisition and analysis during landslide responses, aids in the delivery of geospatial data to stakeholders, and builds upon an inventory of landslide information applicable to future landslide events and responses. The strong correlation of

debris flows and debris slides originating on unmodified and modified slopes with areas where landslide hazard maps show an increased potential for such failures indicates that the maps are useful screening tools for identifying potentially hazardous slopes. A challenge is to gain broad-based stakeholder support for landslide hazard maps so that they are not perceived as threats to private property rights and property values.

Our analysis of limited data indicates that debris flows originating where slope modifications have had a destabilizing effect can be triggered by rainfall events with significantly lower levels than those needed to trigger debris flows on forested, unmodified slopes. The frequency of debris flows on inadequately constructed and maintained slopes, therefore, is likely to be greater than the frequency of debris flows on forested, unmodified slopes.

Our experiences with landslide losses in western North Carolina are not unlike those documented for many other locations. Poorly constructed and maintained slopes, and fill slopes in particular, are a well-documented source of debris flows in mountainous terrain and can increase the rates of slope instability (Collins, 2008; Sidle and Ochiai, 2006). Social and economic losses from landslides, however, can be reduced by the careful development of hill slopes (Turner and Schuster, 1996; Laprade, 2007). The challenge in western North Carolina, as in many other places, is to encourage the careful development of mountain slopes, and the long term maintenance of slopes constructed in the mountains.

AKNOWLEDGEMENTS The cooperation and support of the Emergency Management and Erosion Control Departments of Macon and Haywood Counties during our landslide responses is gratefully acknowledged. LuVerne Haydock graciously provided family history and photographs used to help locate the 1916 and 1996 debris flows in Hickory Nut Gorge. The many thought-provoking discussions on landslides with Tom Collins are appreciated. Reviews by Chris Wills and an anonymous reviewer resulted in substantial improvements to the manuscript.

368

Wooten, R.M., Cattanach, B.L., Bozdog, G.N., Isard, S.J., Fuemmeler, S.J., Bauer, J.B., Douglas, T.J.,

REFERENCES Bauer, J.B., Wooten, R.M., Gillon, K.A., Douglas, T.J.,

Fuemmeler, S.J., Witt, A.C., and Latham, R.S., 2010, “Your country is falling apart:” Response to recent landslides by the North Carolina Geological Survey [abstract]: Association Environmental & Engineering Geologists, 2010 Annual Meeting, Charleston, S.C., Program with Abstracts, p.62.

Bauer, J.B, Wooten, R.M., and Moore, P., 2011. Landslides and Weather: An Interdisciplinary Approach to Anticipating the Hazard and Communicating the Threat [abstract]: Association Environmental & Engineering Geologists 2011 Annual Meeting, Anchorage, AK Program with Abstracts, p.65.

Bauer, J.B., Fuemmeler, S.F., Wooten, R.M., Witt, A.C., Gillon, K.A., and Douglas, T.J, 2012, Landslide hazard mapping in North Carolina – overview and improvements to the program, In Eberhardt, E., Froese, C., Turner, S., Leroueil, S., (Editors), Landslides and Engineered Slopes: Protecting Society through Improved Understanding, 11th International, and 2nd North American Symposium on Landslides, Banff, AB, pp. 257-263.

Collins, T.K., 2008, Debris flows caused by failure of fill slopes: early detection, warning, and loss prevention. Landslides 5 pp. 107-119.

Cruden, D.M. and Varnes, D.J., 1996, Landslide types and processes, In Turner, A.K. and Schuster, R.L., (Editors)., Landslides: Investigation and Mitigation: Transportation Research Board Special Report No. 247, National Research Council, National Academy Press, Washington, D.C., pp. 36-75.

Fuhrmann C.M., Konrad, C.E. III, Band, L.E., 2008, Climatological perspectives on the rainfall characteristics associated with landslides in western North Carolina, Physical Geography, 29, pp.289-305.

Johnstone, T.P., Burrus, S.A., 1998, An analysis of the 4 September 1996 Hickory Nut Gorge flash flood in western North Carolina, 16th Conf. on Weather Analysis and Forecasting, American Meteorological Society, Phoenix, AZ, pp. 275 -277.

Laprade, W.T., 2007, Effects of urbanization and development on mass wasting, In Turner, A.K., Schaefer, V.R., (Editors), Landslides and Society, Association of Environmental & Engineering Geologists, Special Publication, No. 22, pp. 376-383

Latham, R.S., Wooten, R.M., Witt, A.C., Gillon, K.A., Douglas, T.J., Fuemmeler, S.F, Bauer, J.B., and Brame, S., 2007, Investigation of the Peeks Creek debris flow of September 2004 and its relationship to landslide hazard mapping in Macon County, North Carolina: 2007 Southeastern Friends of the Pleistocene Field Trip Guidebook, 35p.

Latham, R.S., Wooten, R.M., Cattanach, B.L., Merschat, C.E., and Bozdog, G.N., 2009, Rock slope stability analysis along the North Carolina section of the Blue Ridge Parkway: using a geographic information system (GIS) to integrate site data and digital geologic maps, 43rd U.S., and 4th U.S.-Canada Rock Mechanics Symposium,

Asheville, NC, American Rock Mechanics Association, 09-171, 12p.

Neary D.G. , Swift L. W., Jr., 1987, Rainfall thresholds for triggering a debris avalanching event in the southern Appalachian Mountains. In Costa J.E., Wieczorek GF (Editors) Debris flows, avalanches: process, recognition, and mitigation. Geological Society America, Reviews in Engineering Geology 7: pp. 81-92

Miller D. K., Barros A. P., Lee L, Wilson A. M., and Wooten, R., 2012 Weather conditions leading to the Gunter Fork Trail debris flows of 14-15 July, 2011 [abstract]: 23rd Annual Great Smoky Mountains National Park Science Colloqium, Gatlinburg, TN.

Pack, R.T., Tarboton, D.G., and Goodwin, C.N., 1998, Terrain stability mapping with SINMAP, technical description and users guide for version 1.00: Terratech Consulting Ltd., Salmon Arm, BC, Report Number 4114-0, 68 p.

Sidle RC, Ochiai H (2006) Landslides: Processes, Prediction and Land Use, American Geophysical Union Water Resources Monograph 18 312p.

Soplata, C.A., 2016, Case study of historically destructive landslides in Hickory Nut Gorge near Chimney Rock North Carolina, Unpublished M.S. Thesis, Univ. of Memphis, 77p.

Tao J., Barros A.P., 2014, Coupled prediction of flood response and debris flow initiation during warm- and cold-season events in the southern Appalachians, USA. Hydrologic Earth Systems Science, 18367-388.

Taylor, K.B., 2017, Landslides and Other Geologic Hazards: A Perspective from an Emergency Management Professional who is also a Geoscientist, in Shakoor, A., DeGraff, J.V. (Editors), 3rd North American Symposium on Landslides, Roanoke, VA (this volume).

Turner, R.L., Schuster, 1996, Principles of landslide hazard reduction, In Turner, A.K. and Schuster, R.L., (Editors)., Landslides: Investigation and Mitigation: Transportation Research Board Special Report No. 247, National Research Council, National Academy Press, Washington, D.C., pp. 91-105.

Wieczorek, G.F, 1987, Landslide triggering mechanisms, In Turner, A.K. and Schuster, R.L., (Editors)., Landslides: Investigation and Mitigation: Transportation Research Board Special Report No. 247, National Research Council, National Academy Press, Washington, D.C., pp. 76-90.

Wieczorek GF, Mossa GS, Morgan BA (2004) Regional debris flow distribution and preliminary risk assessment from severe storm events in the Appalachian Blue Ridge Province, USA, Landslides 1:53-5.

Witt, A.C., Wooten, R.M., Latham, R.S., Gillon, K.A., Douglas, T.J., Fuemmeler, S.J., and Bauer, J.B., 2008, Migration to a Geospatial Database to Improve Data Management in the N.C. Geological Survey's Landslide Hazard Mapping Program, Geological Society America. Abstracts with programs, Vol. 40, No. 6.

Witt, A.C., Gillon, K.A., Wooten, R.M., Douglas, T.J., Bauer, J.B., and Fuemmeler, S.F., 2012, Determining potential debris flow inundation zones for an emergency response using LAHARZ: The Ghost Town Debris Flow, Maggie

369

3rd North American Symposium on Landslides, Roanoke, VA, June 4-8, 2017

Valley, N.C., USA: In Eberhardt, E., Froese, C., Turner, S., Leroueil, S., (Editors), Landslides and Engineered Slopes: Protecting Society through Improved Understanding, 11th International and 2nd North American Symposium on Landslides, Banff, AB, 5p.

Witt, A.C., Wooten, R.M., 2015, The laws of physics don’t stop at state boundaries: challenges and accomplishments in documenting landslide activity in the Central and Southern Appalachian, (abstract, poster) AEG Professional Forum, Time to Face the Landslide Dilemma: Bridging Science, Policy, Public Safety, and Potential Loss, Seattle, WA., p.104.

Wooten, R.M., and Clark, T.W., 2001, Occoneechee Mountain rockslide, Eno River State Park, Orange County, North Carolina: North Carolina Geological Survey Report of Investigation to N.C. Div. of Parks and Recreation, 20p.

Wooten, R., Carter, M., 2004, Maggie Valley Debris Flow of Dec. 11, 2003, AEG-AIPG Geonews, Winter 2003, 2p.

Wooten, R.M., and Latham, R.S., 2004, Report on the May 5-7, 2003 debris flows on slopes underlain by sulfidic bedrock of the Wehutty, Nantahala, and Copper Hill Formations, Swain County, North Carolina: unpublished N. C. Geological Survey report of investigation to N.C. Division of Emergency Management, 20p.

Wooten, R.M., and Clark, T.W., and Latham, R.S., 2006. Occoneechee Mountain rockslide of February 17-18, 2001, Eno River State Park, Orange County, North Carolina: In Bradley, P.J., and Clark, T.W. (Editors)., The Geology of the Chapel Hill, Hillsborough and Efland Quadrangles, Orange and Durham Counties, Carolina Terrane, North Carolina, Carolina Geological Society field trip guidebook, pp.23-27.

Wooten R.M., Latham R.S., Witt A.C., Fuemmeler S.J., Gillon K.A., Douglas T.J., and Bauer J.B., 2006, Slope movement hazard maps of Macon County, North Carolina: N.C. Geological Survey Geologic Hazards Map Series 1, , map scale 1:48,000, Digital Data Series DDS-1.

Wooten, R.M., Latham, R.S, Witt, A.C., Gillon, K.A, Douglas, T.D., Fuemmeler, S.J., Bauer, J.B., and Reid, J.C., 2007, Landslide hazards and landslide hazard mapping in North Carolina: In Schaefer, V.R., Schuster, R.L., and Turner, A.K., (Editors.), 1st North American Landslide Conference, Vail CO, Association Environmental & Engineering Geologists Special Publication 23, pp. 458-471.

Wooten, R.M., Gillon, K.A., Witt, A.C., Latham, R.S., Douglas, T.J., Bauer, J.B., Fuemmeler, S.J., and Lee, L.G., 2008a, Geologic, geomorphic, and meteorological aspects of debris flows triggered by Hurricanes Frances and Ivan during September 2004 in the Southern Appalachian Mountains of Macon County, North Carolina (southeastern USA): Landslides Vol. 5, p.31-44.

Wooten R.M., Witt A.C., Gillon K.A., Douglas T.J., Latham R.S., Fuemmeler S.J., and Bauer J.B., 2008b, Slope movement hazard maps of Watauga County, North Carolina: N.C. Geological Survey Geologic Hazards Map Series 3, map scale 1:36,000, Digital Data Series DDS-3.

Wooten R.M., Witt A.C., Gillon K.A., Douglas T.J., Fuemmeler S.J., Bauer J.B., and Latham R.S., 2009a,

Slope movement hazard maps of Buncombe County, North Carolina: N.C. Geological Survey Geologic Hazards Map Series 4, map scale 1: 52,000, Digital Data Series DDS-4.

Wooten, R.M., Gillon, K. A., Douglas, T.J., Witt, A.C., Bauer, J.B., and Fuemmeler, S.J., 2009b, Report on the Bear Trail debris flow of January 7, 2009: N. C. Geological Survey report of investigation, to Haywood County, 33p.

Wooten R.M., Witt A.C., Douglas, T.J., Gillon K.A., Fuemmeler, S.J., Bauer, J.B, Latham, R.S. Nickerson, J.G., 2010, The North Carolina Landslide Program: 3,864 km2 mapped and 6,042 landslide features later – findings and observations [abstract]: Association Environmental & Engineering Geologists, 2010 Annual Meeting, Charleston, S.C., Program with Abstracts, p. 98.

Wooten R.M., Witt A.C., Douglas T.J., Fuemmeler S.J., Bauer J.B., Gillon K.A., Latham R.S., 2011, Slope movement hazard maps of Henderson County, North Carolina: N.C. Geological Survey Geologic Digital Data Series DDS-5.

Wooten, R.M., Cattanach, B.L., Bozdog, G.N., 2012, Initial and supplemental reports on the November 14, 2012 rock fall at Chimney Rock State park; unpublished N.C. Geological Survey reports to the N.C. Division of Parks and Recreation, 6p, 13p.

Wooten, R.M., Witt, A.C., Miniat, C.F., Hales, T.C., Aldred, J.A., 2016, Frequency and Magnitude of Selected Historical Landslide Events in the Southern Appalachian Highlands of North Carolina and Virginia: Relationships to Rainfall, Geological and Ecohydrological Controls, and Effects, In Greenberg, C.H. and Collins, B.S., (Editors) Natural Disturbances and Historic Range of Variation: Type, Frequency, Severity, and Post-disturbance Structure in Central Hardwood Forests USA, pp. 203-262.

370