Geologic Controls of Basin Denudation from Debris Flows in Rockbridge County, Virginia

Sas, Robert J. JR and Eaton, L. Scott

  1. Introduction

Landslide activity in the Eastern United States is normally associated with heavy rainfall due to large NW and NE tracking storms[g1]. Rainfall intensity is considered the main mechanism of mountainous slope failure. However, debris-flow activity on the western flank of the Blue Ridge Mountains in Virginia suggests that some slope failure processes are heavily dependent upon structural deformation of the ridge forming rocks. Orthogonal joint sets in the Antietam quartzite act to reduce slope-stability by providing failure surfaces and discrete zones of hydraulic pressure head increases. The denudation for one of these mountainous basins is nearly twice the magnitude of any other debris flow impacted basin in Virginia.

  1. Storm
  2. June 27, 2005[g2]?
  3. Parameters in MadisonCounty, VA (Smith, et al., 1996)
  4. RapidanRiver Basin, 14-hr period, more than 775mm[g3] of precipitation
  5. 6-km column of precipitable air mass
  6. “Strong boundary layer winds directed upslope, weak upper level winds”
  7. Strong topographic controls on storm motion
  8. Rockbridge County, VA probably similar results of rainfall and orographics given location and aspects of flows
  9. 213.28 mm precipitation, June 28, 2005[g4], accumulated, estimated (NOAA)
  10. 122.17 mm highest antecedent rainfall, June 23, 2005[g5] (NOAA)
  11. Water table was already high, soils saturated by antecedent rainfall
  1. Methods
  2. Method of reconstructing the pre-erosional slope for volumetric analysis as described by Shroyer (1997)
  3. Cross-sectional area x Length of thalweg = Volume of Material
  4. Volume of material (m3) / Area of zero order basin (m2) = Basin Denudation (m)
  5. Joint and bedding plane orientations measured using Brunton geological compass
  6. Joints and bedding planes selected randomly on outcroppings within the denuded basin
  7. Statistical analysis and stereoplotting performed using Rick Allmendinger’s (Cornell University) computer program “StereoWin v1.2”
  1. Field observations
  2. Bedrock controls evident in debris? flows
  3. Regolith deposits are wedges of material that thin towards the apex, with the thickest deposits in the thalweg.
  4. Bedrock exposures common near lateral escarpments and often form lateral escarpments
  5. In steepest upper portions of at least 2 flows, bedrock is main exposure without any regolith
  6. Quartzite jointing main bedrock control on extents of failure laterally and vertically
  7. Vertical incision, especially in thalweg, controlled by resistant bedding planes with abundant Skolithus
  8. Bedrock ranges in competence from intact to sapprolitic
  9. Well-saprolitized bed that has maintained joint structure is the layer stratigraphically above resistant Skolithus layer
  10. Major strike of main joint on failure plane is 217°-252° which is strike of mountains, strong correlation to topography
  11. Other joints penetrating exposed faces were 160°-164° which is approximately 60°-70° difference between strikes
  1. One joint system dips at 90° or nearly 90°, which likely allows expedited infiltration
  2. Description of regolith
  3. Most evident surficial deposits are orthogonal cobbles of quartzite, with angular boulders less common
  4. Soil is sandy in upper third of flows and becomes more clay rich towards the flow base
  5. Lower thirds of sediments often included weathered, rounded cobbles, sapprolitic rock layers stratified sub-parallel to the dip of the slope
  6. Knick points are common in the larger flows but are poorly understood; some contain evidence of previous debris flow deposits
  7. Previous deposits have a thin soil profile in the O-horizon; deposits directly below are clayey, sandy soils containing matrix supported cobbles with some boulders
  8. Initiation sites
  9. Most are small and open into a wider escarpment
  10. Apex failure commonly has rotational and/or slide component
  11. Failures are spoon shaped in map view, shorter width of spoon facing upslope
  12. Soils are sandy with a smaller amount of clays
  13. Depth of failure, order of magnitude 3 m.
  14. Bedrock exposure with jointingis consistently resistant layer with abundant Skolithus
  15. Minor failures
  16. Primarily cuspate shaped features
  17. Rotational slump, differential stabilities between regolith deposits, related to colluvial/alluvial fans
  18. Occur near to lower reaches of flow
  19. Tree blocks common in larger slumps
  20. Best re-vegetation, Eastern White Pine dominant specie

III.Methods

A.Method of reconstructing the pre-erosional slope for volumetric analysis as described by Shroyer (1997)

B.Cross-sectional area x Length of thalweg = Volume of Material

C.Volume of material (m3) / Area of zero order basin (m2) = Basin Denudation (m)

  1. Denudation Volume Results
  2. Actual sediment yield (2447.1 m3 and 25185.4 m3) one to two orders of magnitude smaller than largest yields in Madison and NelsonCounties (SEE Figure 1)
  3. Due to small zero order basins, basin denudation is one to two orders of magnitude greater in RockbridgeCounty (.086 m and .130 m)
  4. Highest basin denudations measured in VA
  5. Micro scale to mesoscopic structures of foliation and compositional layering found in Nelson County (Gryta and Bartholmew 1989) (possible foliations in Madison County)
  6. Macroscopic (regional) joint systems in Rockbridge County
  1. Mechanisms of Slope FailureStatistical Analysis of Joint and Bedding Planes
  1. Fisher test for uniformity of spherical distribution used to test for the presence of more than one statistically significant set of joints (Davis 2002)
  2. Test utilizes von Mises distribution, which is the circular equivalent to the normal distribution with a mean direction and a concentration parameter
  3. The mean resultant length “R-bar”, concentration parameter “k”, and the population size “n” are all analyzed
  4. Joint plane data results
  5. R-bar = .54
  6. Minimum value of R-bar for 95% confidence = .24
  7. k = 2.1
  8. n = 45
  9. Bedding plane data results
  10. R-bar = .98
  11. Minimum value of R-bar for 95% confidence = .48
  12. k = 50.9
  13. n = 11
  1. Mechanisms of Slope Failure[g6]
  2. Rainfall intensity[g7]
  3. Joint systems
  4. Orthogonal sets and conjugate sets
  5. Fissile plane of weakness
  6. Fails easily when pried with shovel
  7. Allow water to quickly infiltrate
  8. Steep slope 30-35
  1. Discussion (some questions ideas to ponder for inclusion or exclusion of in this paperthe manuscript)
  2. Neither the quartzite lithology nor clastic lithologies have been reported to have any sites of debris flow study in Virginia. The greatest basin denudation reported in a clastic basin in West Virginia is 2.150 * 10-3 m (Cenderelli and Kite, 1998). In RockbridgeCounty, the greatest basin denudation measured is 1.30. * 10-13 m, nearly twice the magnitude of the basin in West Virginia. Additionally, 5.071.0 * 10-23 m, the previously greatest basin denudation reported in Virginia was in a Precambrian granite, gneiss, and schist lithology (Williams and Guy, 1973) (SEE Figure 2a).

B. Eaton et al. (2003) reported a plot of rainfall total vs. denudation for mountainous drainage basins in Virginia and calculated a least-squares regression line for the data. They found that for two basins in the Valley and Ridge province that experienced similar total rainfall values as RockbridgeCounty yielded less than 3.0 * 10-23 m of denudation. A pPossible explanation for the great difference in denudation, given total rainfall, is the role of hydraulic pressure head increase in discrete joint systems.

C.C. While the main mechanism in landslide initiation in the mountainous basins of Virginia is intensity of rainfall (what papers? Wieczorek, Williams/Guy, Hack/Goodlett, Eaton?), the structural character of these basins seems to have a major influence on the denudation. Orthogonal joint systems provide a flow path for rainwater to infiltrate and pressurize the bedrock. Due to the unstable nature of the orthogonal joints in quartzite (Lee, 1988; Panda and Kulatilake, 1995;), these basins experience preferential failure along joint pplanes. Statistical analysis shows that the joint data for two joint sets passes the Fisher Test for a spherical distribution within a 95% confidence interval. The bedding planes have the highest correlation within the subset shown but the high value of k (SEE Figure 3). However, the value of k = 2.1 for the joint dataset is also statistically significant because the null hypothesis is H0: k = 0 and the alternative is H1: k > 1. A contour analysis of correlation of individual poles to adjacent poles within a 1% area of the stereoplot suggests a minimum of two subsets of joints within a population size of n = 45 (SEE Figure 4).

Figure 1. Graph of sediment yield for basins in Virginia and West Virginia (Eaton et al 2000) as compared to quartzite basins in Rockbridge County, VA[g8].

Things to do:

-Run soil gradations

-Stereo-plot of structure

-Calculate theoretical pressure heads in joint systems

-Plot in total rainfall vs. denudation

Figure 2a. Graph of Denudation vs. Rock type for the same basins listed in Figure 1.

Sediment yield (m^3) / Drainage area (m^2) / Denudation (m) / Lithology / Source
20900 / 9710000 / 0.00215 / quartz sandstone / Cenderelli and Kite, 1998
8500 / 5010000 / 0.00170 / siltstone/shale/sandstone / Cenderelli and Kite, 1998
13900 / 17480000 / 0.00080 / quartz sandstone / Cenderelli and Kite, 1998
3300 / 1780000 / 0.00185 / siltstone/shale/sandstone / Cenderelli and Kite, 1998
13364 / 400000 / 0.03341 / gneissic granodiorite/mm greenschist dike / Springer et al, 2001
2492 / 120000 / 0.02077 / granite/foliated granitoid gneisses / Springer et al, 2001
544000 / 29500000 / 0.01844 / metabasalt-greenstone / Eaton, 1999
88727 / 1750000 / 0.05070 / gneiss/granite/schists / Williams and Guy, 1973
173488 / 4080000 / 0.04252 / gneiss/granite/schists / Williams and Guy, 1973
87707 / 2470000 / 0.03551 / gneiss/granite/schists / Williams and Guy, 1973
25185.4 / 193858.2 / 0.12992 / quartzite / Sas, In Progress
2447.061 / 28355.3 / 0.08630 / quartzite / Sas, In Progress

Figure 2b. Table of values used in Figure 1 and Figure 2a.

Figure 3. Stereoplot of bedding planes measured on Lowry Run debris flow.

Figure 4. Stereoplot of joint planes measured on Lowry Run debris flow. Blue contours show poles within a 1% area of stereoplot hemisphere with a 2% contour interval per 1% area.

[g1]Hurricanes have occasionally been the storms from the SE that have triggered landslides, e.g. Hurricane Camille in Nelson County, VA. For some detailed information on previous historic events, see “Debris slide and debris flow historical events in the Appalachians south of the glacial border” by G. Michael Clark (1987, GSA Reviews in Engineering Geology Vol. VII, p. 125-138)

[g2]Are you referring to the June 27, 1995 storm?

[g3]Technically, the maximum recorded storm total of 770 mm was reported by two different observers near Graves Mill (Wieczorek et al., 2000, p. 5) E&EG, v. VI, n. 1, pp. 3-23

[g4]1995?

[g5]1995?

[g6]How would you characterize the initial type of slope failure before it transpired into a debris flow?

[g7]Is there any hourly rainfall data available for the RockbridgeCounty region? Was NEXRAD data available? If so, it would need to be carely calibrated as for MadisonCounty. The detailed rainfall examination of landslides by Meghan Morrissey et al (2004) “Transient Hazard Model Using Radar Data for Predicting Debris Flows in MadisonCounty, Virginia” would be useful to examine in E&EG, v. X, n. 4, pp. 285-296

[g8]The graph is a little unclear having two quartzite sites and three gneiss/granite/schists. How does this figure compare VA and WVA with Rockbridge County?