2008 Griswold

Mobility Statistics and Automated Hazard Mapping for
Debris Flows and Rock Avalanches
Scientific Investigations Report 2007–5276

Version 1.1
U.S. Department of the Interior

U.S. Geological Survey
Mobility Statistics and Automated Hazard
Mapping for Debris Flows and Rock
Avalanches
By Julia P. Griswold and Richard M. Iverson
Scientific Investigations Report 2007–5276

Version 1.1

DIRK KEMPTHORNE, Secretary

Mark D. Myers, Director
U.S. Geological Survey, Reston, Virginia: 2008

Revised 2014
This report and any updates to it are available at:

http://pubs.usgs.gov/sir/2007/5276/
For product and ordering information:

World Wide Web: http://www.usgs.gov/pubprod
Telephone: 1-888-ASK-USGS
For more information on the USGS--the Federal source for science about the Earth, its natural and living resources,
natural hazards, and the environment:

World Wide Web: http://www.usgs.gov
Telephone: 1-888-ASK-USGS
Any use of trade, product, or firm names is for descriptive purposes only and does not imply endorsement by the
U.S. Government.
Although this report is in the public domain, permission must be secured from the individual copyright owners to
reproduce any copyrighted materials contained within this report.
Suggested citation:
Griswold, J.P., and Iverson, R.M., 2008, Mobility statistics and automated hazard mapping for debris flows and rock
avalanches (ver. 1.1, April 2014): U.S. Geological Survey Scientific Investigations Report 2007-5276, 59 p.
Manuscript approved for publication, December 12, 2007

Text edited by Tracey L. Suzuki
Layout by Luis E. Menoyo
iii
Contents
Abstract............................................................................................................................................................1
Introduction.....................................................................................................................................................1
Description of Mass-Flow Phenomena..............................................................................................1
Hazard-Zone Prediction Methods................................................................................................................2
Historic and Geologic Evidence..........................................................................................................2
Physically Based Models.....................................................................................................................3
Empirical Models ..................................................................................................................................3
Statistical Models Motivated by Physical Scaling...........................................................................3
Objectives........................................................................................................................................................5
The Database..................................................................................................................................................6
Statistical Analysis of the Data....................................................................................................................7
Test for Differences in Slope ...............................................................................................................7
Test for Difference in Intercepts.........................................................................................................9
Error and Uncertainty in Calibrated Prediction Equations............................................................18
Discussion and Interpretation of Predictive Equations.................................................................18
Debris-Flow Hazard Delineation by using DEMs and GIS.....................................................................21
Data Input .............................................................................................................................................21
Application Area for Debris-Flow Hazard Mapping—Scottsburg, Oregon...............................22
History of the Scottsburg Area.................................................................................................22
Topography and Geology...........................................................................................................22
Topographic Dataset—Acquisition and Description............................................................23
Background for Flow Volume Assessment for Scottsburg, Oregon...................................24
LAHARZ Application...................................................................................................................25
Hazard Map Summary........................................................................................................................25
Conclusions...................................................................................................................................................27
Acknowledgments........................................................................................................................................29
References Cited..........................................................................................................................................29
Appendix A—Data Table.............................................................................................................................33
Appendix B—Data Bibliography................................................................................................................49
Figures
1. Schematic inundation limits of a debris flow moving down-valley from a
source area.....................................................................................................................................4
2. Diagram showing the maximum inundated cross-sectional area, and total inundated
planimetric area, of a lahar runout path downstream from a source area
on a volcano...................................................................................................................................5
3. Scatter plot of all data compiled for debris flows, rock avalanches, and lahars...............8
4. Debris-flow data and three regression models for cross-sectional area.........................10
5. Debris-flow data and three regression models for planimetric area.................................10
iv
Figures—Continued
6. Rock-avalanche data and three regression models for cross-sectional area.................10
7. Rock-avalanche data and three regression models for planimetric area.........................10
8. Rock-avalanche and lahar data used to compute statistics as a combined dataset......17
9. Debris-flow and lahar data used to compute statistics as a combined dataset..............17
10. Debris-flow and rock-avalanche data used to compute statistics as a
combined dataset........................................................................................................................18
11. Lahar data and best-fit regression line with 99-percent confidence interval for
regression and 99-percent confidence interval for prediction after Iverson and
others.............................................................................................................................................19
12. Debris-flow data and best-fit regression line with 99-percent confidence interval
for regression and 99-percent confidence interval for prediction......................................19
13. Rock-avalanche data and best-fit regression line with 99-percent confidence
interval for regression and 99-percent confidence interval for prediction.......................20
14. Aerial photograph of the Scottsburg, Oregon, area..............................................................21
15. Shaded-relief map with highways and logging roads for the study area west of the
town of Scottsburg in the central Coast Range of Oregon, at about 40 river
kilometers from the coast...........................................................................................................23
16. Shaded relief map constructed from a DEM, overlaid with orange where slopes
exceed 30 degrees......................................................................................................................24
17. Shaded relief map of upper half of the test basin..................................................................25
18. Examples of nested, inundation-hazard zones computed for four flow volumes.............26
19. Debris-flow inundation-hazard map computed for four flow volumes ranging
from 103 to 104.5 m3........................................................................................................................27
20. Oblique perspectives of the test basin with debris flow-hazard zone overlays,
view to the southeast..................................................................................................................28
Tables
1. Parameters and analysis-of-variance statistics for alternative linear models of
log-transformed debris-flow data for inundated cross-sectional area, as a
function of flow volume..............................................................................................................11
log-transformed debris-flow data for inundated-planimetric area, as a
log-transformed rock-avalanche data for inundated cross-sectional area, as
a function of flow volume...........................................................................................................11
log-transformed rock-avalanche data for inundated-planimetric area, as a
5. Summary table of the best-fit regression equations with calibrated slopes and
intercepts......................................................................................................................................12
v
Tables—Continued
6. Summary table of the best-fit regression equations with specified 2/3 slopes and
calibrated intercepts...................................................................................................................12
7. Summary of results of F-tests comparing the specified-slope models to the
best-fit regression model with adjustable slope....................................................................13
8. Dummy variable terms and equations for null model and alternate models.....................14
9. Parameters for computing the F-statistic to compare the specified 2/3-slope
models with combined and individual datasets.....................................................................14
10. Parameters for computing the F-statistic to compare the best-fit regression
models with combined and individual datasets.....................................................................14
11. F-test comparison between the combined dataset and individual datasets for
rock avalanches and lahars.......................................................................................................15
12. F-test comparison between the combined dataset and the individual datasets for
debris flows and lahars..............................................................................................................15
13. F-test comparison between the combined dataset and the individual datasets for
rock avalanches and debris flows............................................................................................16
14. F-test comparison between combined rock avalanche and lahar datasets and the
listed individual datasets for a best-fit regression slope......................................................16
15. F-test comparison between combined debris flow and lahar datasets and the
listed individual datasets for a best-fit regression slope......................................................16
16. F-test comparison between combined rock avalanche and debris flow datasets
and the listed individual datasets for a best-fit regression slope.......................................16
Mobility Statistics and Automated Hazard Mapping for
Debris Flows and Rock Avalanches
By Julia P. Griswold and Richard M. Iverson
Abstract Introduction
Debris flows and rock avalanches are two types of
Power-law equations that are physically motivated and
rapid mass movement that can pose great hazards to areas
statistically tested and calibrated provide a basis for fore-
downslope and downstream. This report presents a methodol-
casting areas likely to be inundated by debris flows, rock
ogy to assess areas likely to be affected by downslope and
avalanches, and lahars with diverse volumes. The equations
downstream runout of debris flows and rock avalanches by
A=α1V2/3 and B=α2V2/3 are based on the postulate that the
deriving and employing inundation-area statistics, analogous
maximum valley cross-sectional area (A) and total valley pla-
to an approach used by Iverson and others (1998) for lahars.
nimetric area (B) likely to be inundated by a flow depend only
Establishing inundation-area relationships for differing types
on its volume (V) and the topography of the flow path. Testing
of flows aids in understanding flow mobility and enables
of these equations involves determining whether or not they
delineation of hazard zones that may be inundated by future
fit data for documented flows satisfactorily, and calibration
lahars, nonvolcanic debris flows, and rock avalanches.
entails determining best-fit values of the coefficients α1 and
The lahar-inundation equations developed by Iverson and
α2 for debris flows, rock avalanches, and lahars. This report
others (1998) are currently used to compute hazard zones in
describes statistical testing and calibration of the equations by
areas adjacent to volcanoes around the Pacific Rim. Schilling
using field data compiled from many sources, and it describes
(1998) implemented the set of lahar-inundation equations in
application of the equations to delineation of debris-flow
LAHARZ, a Geographic Information System (GIS)-based
hazard zones.
computer program that facilitates rapid calculation and
Statistical results show that for each type of flow (debris
delineation of the hazard zones for a range of probable flow
flows, rock avalanches, and lahars), the dependence of A and
volumes and any flow-path topography represented by a
B on V is described well by power laws with exponents equal
digital elevation model (DEM). This study expands the use of
to 2/3. This value of the exponent produces fits that are effec-
LAHARZ with modifications for hazard mapping of nonvol-
tively indistinguishable from the best fits obtained by using
canic debris flows and rock avalanches.
adjustable power-law exponents. Statistically calibrated values
of the coefficients α1 and α2 provide scale-invariant indices of
the relative mobilities of rock avalanches (α1 = 0.2, α2 = 20), Description of Mass-Flow Phenomena
nonvolcanic debris flows (α1 = 0.1, α2 = 20), and lahars
(α1 = 0.05, α2 = 200). These values show, for example, that Rapid mass flows on Earth’s surface include the fol-
a lahar of specified volume can be expected to inundate a lowing related phenomena: dry, granular-rock avalanches,
planimetric area ten times larger than that inundated by a rock water-saturated debris flows, and hyperconcentrated stream
avalanche or nonvolcanic debris flow of the same volume. flows (for example, Pierson and Costa, 1987; Iverson and Val-
The utility of the calibrated debris-flow inundation equa- lance, 2001). Flows are described by many terms in common
tions A=0.1V2/3 and B=20V2/3 is demonstrated by using them practice and in the scientific literature, but our nomenclature
within the GIS program LAHARZ to delineate nested hazard is restricted to rock avalanche, debris flow (nonvolcanic), and
zones for future debris flows in an area bordering the Umpqua lahar (volcanic debris flow). In this section, we discuss the
River in the south-central Oregon Coast Range. This applica- general nature of these flows.
tion requires use of high-resolution topographic data derived Rock avalanches commonly initiate as rock-slab failures
form LIDAR surveys, knowledge of local geology to specify a or rockfalls and transform into unsaturated, granular flows.
suitable range of prospective debris-flow volumes, and devel- Rockfalls result from failure along a bedrock discontinuity
opment and use of a new algorithm for identification of pro- (fracture, bedding, foliation), or from failure of a pocket of
spective debris-flow source areas in finely dissected terrain. rubble perched in a bedrock face. Their motion is influenced
2 Mobility Statistics and Automated Hazard Mapping for Debris Flows and Rock Avalanches
by gravity, intergranular vibrational energy, and Coulomb fric- degree of fragmentation of volcanic rubble makes for a readily
tion (Iverson, 2003). Rock avalanches tumble from mountain- erodible and commonly clay-rich source material (Vallance
sides and commonly travel to distal areas where they terminate and Scott, 1997). The failed volcanic material can incorporate
on flatter slopes. Deposits may include massive hummocks large volumes of water, ice, and snow to reach full saturation.
composed of coarse debris and partially saturated sediment Lahars generally inundate the lower flanks of volcanoes, as
(Crandell, 1989). Although small rock avalanches occur well as downstream reaches of streams that originate on those
commonly, they are seldom well documented, and volumes flanks. Volumes of well-documented lahars generally exceed
of well-documented rock avalanches generally exceed 106 m3 those of nonvolcanic debris flows and are typically >105 m3.
(cubic meters). At Mount Shasta in California, the deposit of Observed runout distances of lahars indicate they are typically
one prehistoric rock avalanche is estimated to have a volume more mobile than rock avalanches. For example, large lahars
~45 km3 (>1010 m3; Crandell, 1989). The large volume of some originating on Mount Rainier have traveled more than
rock avalanches can result in burial and filling of entire val- 120 km to Puget Sound, filling the White River Valley to
leys, potentially damming rivers and creating breeching haz- depths greater than 100 m (Vallance and Scott, 1997). Addi-
ards for some time after the initial event. Additional descrip- tional descriptions and reports about lahars are provided by
tions and reports on rock avalanches are provided by Voight Crandell and Mullineaux (1967, 1975), Janda and others
(1979), Eisbacher and Clague (1984), Ui and others (1986), (1981), Major (1984), and Pierson (1985).
Siebert (1984), Ui (1983), and Siebert and others (1987).
Debris flows differ from rock avalanches primarily
because of their high water content and resulting fluidity; they
typically initiate in steep drainages (generally sloping Hazard-Zone Prediction Methods
> 30º) where sufficient unconsolidated sediment and water are
Numerous methods have been proposed to predict areas
available. The water-saturated sediment liquefies during initial
likely to be inundated by future granular mass flows. These
slope failure or subsequent entrainment and stays fluid due to
methods include (1) use of historic and geologic evidence of
the persistence of high pore-fluid pressure, which is facilitated
past flows to estimate inundation limits of future flows, (2) use
by the presence of fine matrix sediment (Iverson and others,
of physically based models that invoke conservation of mass,
1997). Steep slopes, the presence of shallow soils, colluvium,
momentum, and energy and rheological properties to calculate
alluvium, or poorly consolidated bedrock, and the presence
prospective inundation limits, (3) use of statistically calibrated
of abundant surface water (for example, ponded water, snow,
empirical equations derived from analysis of inundation data,
ice, runoff, outbreak flood) or shallow ground water (perched
and (4) use of statistically calibrated inundation equations that
or return flow) provide conditions that favor development of
are motivated by physical-scaling arguments. Method 4 is
debris flows. Local topographic features commonly play a role
employed in this report, as described below. The context for
in focusing water flow on slopes where failure occurs. These
use of this method, however, is clarified by providing a brief
topographic features include gullies, swales, hollows, or con-
synopsis of the first three methods.
vergent slopes and also can include constructed features such
as road-fill prisms. Focusing of shallow ground water also is
influenced by stratigraphy that may aid saturation and devel- Historic and Geologic Evidence
opment of high pore-water pressures (for example, Reid and
Iverson, 1992). Termination of debris-flow motion downval- Traditionally, mass-flow hazard maps have been derived
ley on lesser slopes of stream channels, alluvial fans, or other from inspection of historic and geologic evidence and use of
gently sloping areas commonly results in a coarse depositional this evidence to posit future inundation patterns (for example,
snout and bounding levees that enclose a liquefied interior of Scott and others, 1998). Although this method is well estab-
finer material (Iverson, 1997). Volumes of nonvolcanic debris lished, it has two inherent limitations. First, documented past
flows rarely exceed 106 m3, and although debris flows are events do not necessarily provide an adequate sample of the
commonly smaller than rock avalanches, they occur frequently population of all events (both past and future) that might occur
and are responsible for similar numbers of deaths as reported in a particular area. Therefore, the extent of inundation during
in Schuster’s (1996) list of “The 25 most catastrophic land- future events can exceed limits forecast on the basis of past
slides of the 20th century”. Additional descriptions and reports events. This problem is most serious where a dearth of docu-
about debris flows are provided by Jakob and Hungr (2005), mentation of past events results from limited historical records
Iverson (1997), Iverson and others (1997), Major (1996), and geological fieldwork. A second problem with this forecast-
Fannin and Rollerson (1993), Takahashi (1991), and Johnson ing method is reliance on geological inference and consequent
(1984). lack of reproducibility. Different geologists may examine the
Lahars are debris flows that originate on the flanks of same historical records and field evidence, but draw different
volcanoes where abundant, loose sediment is available, and conclusions regarding the potential for future inundation. Lack
they may be triggered by volcanic eruptions, earthquakes, of reproducibility is best overcome by formalizing predictions
glacier or lake break-out floods, or torrential rains (Myers through use of mathematical models, which may have a physi-
and Brantley, 1995). Hydrothermal alteration and (or) a high cal basis, a statistical basis, or both.
Hazard-Zone Prediction Methods 3
Physically Based Models by rock avalanches has led several authors to propose that
H/L, as used in the Heim (1932) equation, depends systemati-
Physically based mathematical models for predicting cally on avalanche mass or volume (for example, Heim, 1932;
inundation by mass flows have varying degrees of sophistica- Scheidegger, 1973; Hsu, 1975). Calibration of the relationship
tion, but all such models are built on a foundation of physical- between H/L and V then provides a basis for prediction. How-
conservation laws. The most elementary models invoke only ever, this method takes no account of the effect of runout-path
one-dimensional momentum conservation for a translating topography on the distal or lateral limits of inundation, an
point mass (that is, Newton’s second law of motion). The first effect that can be apparent to casual observers.
model of this type was presented by Heim (1932), and it led Some authors used empirically calibrated limits for
to the famous equation H/L = tan φ, where H is the vertical debris-flow stoppage to predict debris-flow progress through
distance descended by the mass, L is the predicted horizontal successive channel cross-sections. Benda and Cundy (1990)
distance traversed by the mass, and φ is the Coulomb angle of used an empirical model based on channel-junction angles
sliding friction, which typically ranges from about 30 to (≥70°) and channel gradients (<20°) to predict runout ter-
40 degrees in experimental tests. This model famously under- mination for small debris flows in finely dissected, forested
predicts the extent of runout (L), particularly if mass flows landscapes. Fannin and Rollerson (1993) used channel con-
are saturated with water (for example, debris flows, Iverson, finement (width to depth ratios) to track debris-flow progress
1997), or if their volumes exceed about 106 m3 (Heim, 1932; down valley through analysis of the channel at successive
Hsu, 1975; Scheidegger, 1973). sections. Fannin and Wise (2001) used slope geometry and net
One approach to remedying the failing of the Heim changes in volume for successive sections of a channel to infer
(1932) model involves use of resistance formulae other than whether or not a debris flow was likely to entrain or deposit
that for Coulomb friction. For example, viscosity coefficients sediment. Changes in flow volume were then assessed for each
or fixed yield-strength coefficients have been suggested as section until the cumulative volume was zero.
alternatives to tan φ (for example, Voight and others, 1983; Other authors have focused on debris-flow flow volume
Johnson, 1984; Dade and Huppert, 1998). A significant prob- as a basis for predicting runout. Adopting the methodology
lem with this approach is that the relevance of such resistance developed by Iverson and others for lahars (1998), Crosta
coefficients is not supported by experimental data (Iverson, and others (2003) found that planimetric areas inundated by
2003). diverse debris flows were nearly proportional to flow volume
Basal fluid pressure can be invoked as a means of raised to the 2/3 power. Rickenmann (1999) used the product
modifying Coulomb friction (for example, Shreve, 1968; of debris-flow volume and height of source area to predict
Sassa, 1988). Pore-fluid pressure effects are clearly evident
runout distance. Cannon (1989) reported that the lengths of
in experimental data, but it is difficult to estimate the degree
debris-flow runout paths are proportional to the rate at which
to which high basal fluid pressures will develop and persist in
flow volume is lost due to progressive deposition.
any particular mass flow. Some authors have assigned a basal
For rock avalanches, Li Tianchi (1983) related volume
pore-pressure distribution that fits experimental observations
to inundated planimetric area, (log(Area) = 1.8807 + 0.5667
(for example, Iverson, 1997), whereas others have calcu-
logV), and with statistical methods, produced a pair of pre-
lated pore-pressure distributions based on the porous-media
diction curves for runout-path length and width for a given
consolidation theory (for example, Savage and Iverson, 2003),
vertical relief and flow volume. Kilburn and Sorenson (1998)
but good constraints for applying these approaches to field
concluded that the runout distance L is described by L=αV1/2
phenomena are still lacking.
The most elaborate physically based models take into where α = 3-40 is a calibrated coefficient. For both volca-
account multidimensional mass and momentum conservation, nic rock avalanches and lahars, Vallance and Scott (1997)
as well as pore-pressure evolution, and they thereby reduce the observed a dimensionally homogenous power-law relation-
need for calibration of flow resistance (for example, Iverson ship between flow volume and planimetric area of inundation,
and Denlinger, 2001; Denlinger and Iverson, 2001). However, analogous to the result reported by Crosta and others (2003)
such physically-based models demand considerable input data, for debris flows.
as well as computationally intensive solution techniques, and
they remain an active area of research (Denlinger and Iverson, Statistical Models Motivated by Physical
2004; Iverson and others, 2004; Iverson, 2005). Application
of such models to practical hazard assessment is in its earliest Scaling
stages.
Several authors have used scaling arguments to bolster
the proposition that planimetric areas inundated by rock
Empirical Models avalanches should be proportional to V2/3 (for example, Davies,
1982; Hungr, 1990; Dade and Huppert, 1998; Kilburn and
Empirical equations, that are statistically calibrated, pro- Sorensen, 1998; Legros, 2002). All of these authors presented
vide an alternative to physically based mathematical models. data that broadly supported this contention, resulting in a com-
For example, analysis of data on the distal limits of inundation pelling body of evidence.
Iverson and others (1998) extended these ideas by postu- These equations have quantitative confidence limits,
lating that both the total, planimetric area (B) and the maxi- which were determined by statistics associated with the regres-
mum valley cross-sectional area (A) inundated by a passing sion analysis. The coefficients 0.05 and 200 in equations 1
lahar should be proportional to flow volume raised to the and 2 were specified by using only one significant digit, as the
2/3 power. Statistical analysis of a dataset for 36 lahars at 9 confidence limits imply that greater precision is inappropri-
volcanoes was used to test and confirm the validity of these ate. Another important aspect of equations 1 and 2 is that they
2/3 power laws. Optimal values of the proportionality coef- describe the maximum extent of inundation by an average
ficients were determined by using regressions of log-trans- flow. Such inundation limits do not necessarily correspond
formed data, yielding the following predictive equations: with the limits of deposits left by the flow because some
reaches of the flow path may not be subject to deposition.
A = 0.05 V 2/3 for cross-sectional area, and (1) Together, equations 1 and 2 suffice for delineating
B = 200 V 2/3 for planimetric area. (2) inundation limits on maps, provided that V is known, and the
Figure 1. Schematic inundation limits of a debris flow moving down-

valley from a source area. The maximum inundated valley cross-
sectional area, A, is shown for four transects (yellow). These highlighted
cross sections represent inundation limits during peak flow passage,
not necessarily the limits of post-flow deposition. The total inundated
planimetric area, B, is shown by the dashed red line. Photograph is from
the July 20, 2003, debris flow at Minamata in Kyushu, Japan, (Sidle and
Chigira, 2004).
Objectives 5
topography downslope or downstream from the lahar source

area is known (fig. 1). Relative to other empirical methods,
Objectives
an advantage of the Iverson and others (1998) method is that The overall objective of this study is to extend the
it makes full use of three-dimensional topographic constraints hazard-zone delineation methodology of Iverson and others
for forecasting inundation. (1998) to nonvolcanic debris flows and rock avalanches. To
Because flow-volume V is the independent variable in the attain this objective, several steps are taken and are enumer-
method of Iverson and others (1998) and because the volumes ated here.
of future lahars are indeterminate, forecasts of inundation (1) Assemble a database consisting of flow volumes (V)
limits generally postulate a range of prospective V-values. paired with maximum inundated valley cross-sectional areas
Inundation limits A and B are then calculated for the range (A) and (or) total inundated planimetric areas (B) for a large
of postulated V’s. This procedure results in a nested set of number of nonvolcanic debris flows and rock avalanches
inundation hazard zones, which depict the combined effect of (fig. 2). This database parallels that assembled for lahars by
uncertainties about the volumes and behaviors of future flows. Iverson and others (1998) and partly reproduces the rock-
Selection of appropriate V-values depends on geological avalanche databases assembled by Li Tianchi (1983), Legros
knowledge, and ideally, on recurrence probabilities determined (2002), and others.
for flows with various V’s (Iverson and others, 1998). How- (2) Use the database to test whether power-law equations
ever, data suitable for calculating such probabilities commonly with specified 2/3-exponents satisfactorily predict inundated-
are unavailable, and geological inferences about hydrologic planimetric and cross-sectional areas as functions of flow
contributing areas, thicknesses of soil mantles, and the mag- volume for nonvolcanic debris flows and rock avalanches.
nitudes of events in similar settings provide the main basis for This test involves determining the goodness of fit of the 2/3
selecting prospective flow volumes. power-law equations, as well as statistical comparison of these
Trimline
Figure 2. Diagram showing the maximum inundated cross-sectional area, A, and total inundated
planimetric area, B, of a lahar runout path downstream from a source area on a volcano. The downstream
edge of the source area is delineated by using an H/L cone in this instance. Figure modified from Iverson
and others (1998).
equations to alternative statistical models. If the 2/3 power-law gravels, or the height of stripped bark on the trunks of adjacent
relationships are satisfactory, then for each type of flow it is trees, and the height of log jams. Therefore, area A depends
appropriate to pursue a subsequent step— testing the hypoth- on the maximum stage (for example, elevation) of inundation
esis that the inundation-area equations, during passage of flow through a cross section—and not on
the area blanketed by deposits. Evidence used to infer total,
A=α1V2/3, and (3) inundated planimetric area, B, includes the extent of coarse
deposits that form the levees and snout of a granular flow, or
B=α2V2/3, (4) the lateral limits of any evidence of high-flow lines. This total,
planimetric area includes all parts of the flow path, whether or
with calibrated α-coefficients, provide distinguishable models not they are blanketed by deposits, but does not include areas
for different types of flows. In this context, differing values of affected by subsequent flooding or hyperconcentrated stream
the calibrated α-coefficients imply differing degrees of flow flow. Pairing of data on inundated cross-sectional area A and
mobility. (or) planimetric area B with flow-volume data is necessary
(3) Implement the calibrated predictive equations by because, from a statistical standpoint, flow volume, V, is the
embedding them in the LAHARZ computer program and independent variable, and A and B are the dependent variables.
modifying the criteria for identification of source areas to Ideally, values of all three variables are known for any particu-
extend the use of the software to nonvolcanic flows. Iverson lar event.
and others (1998) identified prospective lahar source areas Determination of the inundation variables A, B, and V
within any valley that drains the upper slopes of a volcano. commonly is complicated by lack of accurate knowledge of
Upper slopes were defined by using an “H/L cone” in which topography before and immediately after an event. The most
L is the horizontal distance from the volcano summit to any useful data come from surveys made shortly after an event that
point downslope, and H is the elevation difference between record high-flow marks and from surveys that include detailed
those two points (fig. 2). Generally, H/L-values between topographic maps of the land surface before and after an
0.2 and 0.3 are well suited for identifying lahar source areas event. However, because the motivations behind the various
on large composite volcanoes (for example, Mount Rainier in papers and reports on debris flows and avalanches differ, the
the Cascade Range of the United States), and values up to 0.7 type and quality of data vary significantly. In many cases areas
are used for smaller or multipeaked volcanoes (for example, A and B implicitly include the effects of any channel scour that
Santa Ana volcano in El Salvador). However, it is more dif- occurred during or following flow events because areas A and
ficult to identify prospective source areas in dissected terrain B are based entirely on post-event evidence.
that is not characterized by a dominant topographic feature, Many reports did not document the values of A, B, and
such as a volcano. Therefore, a new method for identification V explicitly, thus the quality and quantity of diverse infor-
of source areas of nonvolcanic debris flows is described in the mation dictated whether or not the report could be used to
LAHARZ Application section. reconstruct the values of A, B, and V. Reports fell into three
(4) Use GIS and a modified version of LAHARZ to categories. Reports that included detailed maps of deposit
create a debris-flow hazard map for a range of hypothetical extent and channel cross-section surveys that identified the
flow volumes in a drainage basin in a geographic area where pre-flow surface were most useful. The second type of report
precise digital-topographic data are available. offered sketches with descriptive details on total runout
distance; maximum- and average- deposit dimensions at road,
railway, or trail crossings; and maximum and average flow
widths in well-constrained channels. These dimensions were
The Database used to calculate one or more of the inundation variables in a
piece-wise fashion. In these cases, calculations made by using
Diverse data sources and events in diverse geographic average, maximum, or minimum dimensions were used to
locations (with diverse climate, bedrock, topography, and veg- estimate the difference in calculated outcomes. As long as the
etation) were used to assemble a dataset suitable for extending results agreed to one significant digit, calculations from these
the methodology of Iverson and others (1998) to nonvolcanic descriptive reports were included in the database. The third
debris flows and rock avalanches. All data and data sources report category either provided data that could not be used to
are tabulated in appendixes A and B. reconstruct the inundation variables objectively and reproduc-
In assembling the dataset, descriptions of documented ibly, or it provided data suitable for reconstructing only one
runout paths of debris flows and rock avalanches of prehistoric of the three inundation variables (V, A, B). Reports from this
and historic events were found in scientific literature, unpub- third category were excluded from the database.
lished reports and maps, and personal communications or field If reports included maps of the deposit without a quanti-
notes. As described in the original sources, evidence used to tative assessment of inundation variables, a simple method for
infer maximum inundated cross-sectional area, A, includes determining area B was used, which entailed overlaying a fine
high-flow marks indicated by strandlines, levees, embedded grid on the map, counting the boxes within the mapped extent,
Statistical Analysis of the Data 7
and using the map scale to calculate the total area. The same tions. We focused on power-law equations because prelimi-
procedure was used to determine cross-sectional area, A, for nary analyses by Griswold (2004) showed that power laws fit
any surveyed channel profiles that included topography from the data better than did alternative simple regression models
before and after an event. (that is, linear, exponential, or quadratic).
Despite the availability of some high-precision field mea- First, we used results from an analysis of variance
surements obtained through detailed surveying or calculations (ANOVA) to determine whether equations 3 and 4, with speci-
done in GIS, the database in appendix A lists only one signifi- fied 2/3 exponents and statistically calibrated α-coefficients,
cant digit for each volume and area entry. Values of volume provide appropriate models for both debris flows and rock
and area generally are accurate to only one or two significant avalanches; and we compared these models to best-fit power-
digits due to the following factors (1) the small scale at which law regression models with adjustable exponents. Second,
surveys were done and the maps were produced, (2) the degree we examined the statistical difference between the calibrated
of erosion of deposits and loss of reconstructable evidence α-coefficients by using a technique involving a dummy
between the time of the event and the time of the study; and variable and multiple linear-regression analysis to determine
(3) the uncertainty of inferences from descriptive reports used whether one set of predictive equations would be adequate
to reconstruct the inundation variables where they were not for multiple types of flows. Third, we described the statistical
specifically reported. It is important to point out that precision uncertainty inherent in the predictive equations. Finally, we
greater than one significant digit would have little effect on considered geological interpretation and physical implications
results reported in this study as the data are analyzed in the of the calibrated inundation equations for the various types of
context logarithmic plots and power-law equations. Therefore, flows.
the order of magnitude of the data is paramount. Prior to statistical testing, the data and power-law equa-
Some atypical and ambiguous events were intentionally tions are transformed logarithmically. Use of log-transformed
excluded from the dataset. For example, events in which mul- data acknowledges that data scatter increases roughly in
tiple debris flows coalesced, such that inundation variables for proportion to the data magnitude, as shown in figure 3. Log
a single event were obscured, were excluded. Also excluded transformation also enables use of standard least-squares
were events that followed in rapid succession such that flow- regression methods when fitting power-law equations to the
path features were overprinted or undifferentiable, or where data. Following log-transformation, the power-law equations
pre- and post- event topography was not discernible. Cases in to be tested and calibrated become
which the flow would be characterized as a hyperconcentrated
log A = log α1 + β1 log V , and (5)
water flood during a portion of the runout also were excluded
from the dataset.
Data for 64 debris flows (44 V-B pairs and 50 V-A area log B = log α 2 + β 2 log V ,
pairs) and 143 rock avalanches (142 V-B pairs and 12 V-A (6)
pairs) are included in the database (appendix A). Rock-ava-
lanche volumes range between 105 and 1011 m3 and correspond where the logα terms represent y-intercepts, and the
respectively to the Felsberg event (Heim, 1921) and the Flims β-coefficients (slopes) are hypothesized to be 2/3.
event (Jackli, 1957), both in The Alps. Debris-flow volumes
range between 101 and 107 m3, which correspond to USGS
debris-flow flume experiments (Iverson and others, 1992) and
Test for Differences in Slope
the 900 years B.P. (before present) Upper Lillooet River debris In this section we examine whether the proposed equa-
flow in British Columbia (Jordan, 1994). For lahars, we used tions with a specified slopes (β=2/3) and calibrated inter-
data compiled by Iverson and others (1998), which include cepts (log(α)) provide viable alternatives to linear regression
events ranging from about 105 to 1010 m3 in volume. equations with best-fit slopes and intercepts. A preliminary
All data assembled for lahars, debris flows, and rock question is whether or not there is a significant relation-
avalanches are plotted in figure 3. A log-log graph is used ship between the dependent and independent variables, and
to depict the data because on such a graph, power laws of answers to both questions rely on the outcome of an analysis
the form A = αV2/3 plot as straight lines with slopes of 2/3. of variance (ANOVA) and the associated F-statistic.
Although considerable data scatter is evident in figure 3, the For each paired data set (V,A or V,B for each type of
linear-data trends evident in the figure help motivate investiga- flow) the best-fit linear regression models (referred to hereaf-
tion of such 2/3-power laws as suitable models. ter as “Model 1”) have two adjustable parameters (slope and
y-intercept) to be calibrated by using a standard procedure
of minimizing the residual sum of squares. The specified
Statistical Analysis of the Data 2/3-slope models (“Model 2”), as well as specified zero-slope
models (“Model 3”), each have one adjustable parameter (the
The second objective of this report is to use the assem- y-intercept) to be calibrated by minimizing the sum of squares.
bled datasets to develop and test predictive inundation equa- The two questions to be addressed are (1) “Do the models with
107
106
rock avalanches
Cross-sectional area, A (m2)
105
104
lahars
103
1010
debris flows
102
109
101
108
lahars
Planimetric area, B (m2)

100
107
2/3
e=
106
p rock avalanches
slo
105
104
103
debris flows
102
101 102 103 104 105 106 107 108 109 1010 1011
Flow volume, V (m3)
Figure 3. Scatter plot of all data compiled for debris flows, rock avalanches, and lahars. Flow volume, V, is the independent variable
and maximum inundated cross-sectional area, A, and total, planimetric area, B, are the dependent variables. All data and data sources
are tabulated in appendixes A and B.
nonzero slopes fit the data significantly better than a speci- where SS is the residual sum of squares and DF is the residual
fied zero-slope model (a horizontal line expressing no linear degrees of freedom. For each model, DF = n-N, where n is
relationship between the variables)?” and (2)” If the models the number of data pairs and N is the number of calibrated
with nonzero slopes fit the data best, is there a significant parameters in the model. (N=2 for the best-fit regression mod-
difference between the fit attained by using the best-fit linear els and N=1 for the specified-slope models.) The computed
regression model and the specified 2/3-slope model?” If there F statistic-value is compared against tabulated values of the
is no statistically significant difference, then the 2/3-slope F-distribution (for example, Haan, 1977), and a probability of
model provides a suitable representation of the data. the F-statistic falling within a specified range of the F-dis-
The datasets and regression lines representing the three tribution is determined. For example, a probability (p-value)
alternative models for each dataset are plotted in figures 4–7. of 0.01 indicates that, if the null hypothesis were true, there
The corresponding ANOVA results are listed in tables 1–4 would be at least a 1 percent chance of obtaining a calculated
under column headings DF-A1 for debris flow V,A dataset F-value smaller than the tabulated F-value. Our key questions
Model 1; DF-B2 for debris flow V,B dataset Model 2; and so are whether or not calculated F-statistics support rejection of
forth. For debris flows and rock avalanches with 2 datasets Model 3 (specified zero slope), but do not support rejection of
each (V,A pairs and V,B pairs) and 3 statistical models of each Model 2 (specified 2/3 slope).
dataset, there are 12 models to consider. Tables 5 and 6 sum- For both debris flows and rock avalanches, the calculated
marize the candidate predictive equations for Models 1 and F-statistics for the specified 2/3-slope models (Model 2) are
2 (the best-fit regression and specified 2/3 slope models) for smaller than tabulated F-values for p=0.01, and confidence
rock avalanches and debris flows and also list the analogous that the associated null hypothesis can be rejected conse-
predictive equations for lahars (reproduced from Iverson and quently is small (table 5). The computed F-statistics for the
others, 1998). specified zero-slope models (Model 3) are much greater than
Values of the coefficient of determination (r2) listed in tabulated F-values for p=0.01, and therefore, the confidence
tables 1–4 indicate how well the sloped models surpass the with which this null hypothesis can be rejected is great
zero-slope model in describing the data. Following Weisberg (table 7). Thus, there is little statistical evidence that the best-
(1985), the coefficient of determination, r2, is computed as fit regression models fit the data significantly better than the
specified 2/3-slope models, but there is strong evidence that
the best-fit regressions provide a better fit to the data than does
, (7) a horizontal line. Therefore, for purposes of forecasting hazard
zones, we adopted the 2/3-slope models (table 6) as acceptable
where SS is the residual sum of squares that summarizes fits to the data.
deviations of data values from a model trend line. The residual
sum of squares for the sloped models is necessarily smaller
than those of the zero-slope models, provided that some linear Test for Difference in Intercepts
relationship exists between V and A and between V and B. Our
Having shown that power-law equations with specified
datasets, all have r2-values greater than 0.77 (tables 1–4), and
2/3 slopes and calibrated intercepts provide suitable models
we infered that sloped models are better than the zero-slope
of the data, we examined the assumption that the inunda-
models that use only the mean value to represent the data.
tion areas for the three types of landslides (debris flows, rock
The F-test is used to compare the specified-slope models
avalanches, and lahars) yield power-law equations that are
(Models 2 and 3 in tables 1–4) against the best-fit regression
statistically distinct. This can be done by showing that the
model (Model 1) for each dataset. More specifically, the F-test
y-intercepts of the log-transformed power-law equations are
evaluates the credibility of a null hypothesis (H0) and alternate
statistically different.
hypothesis (H1) that state
Testing the intercepts for significant differences can be
H0, The specified-slope model (Model 2, with β1=2/3; or
accomplished by combining any two datasets (for example,
Model 3, with β1=0) fits the data as well as the best-fit regres-
rock avalanche and debris flow V, A data), calibrating a regres-
sion model (Model 1, with β1 adjustable);
sion equation for the combined group, and testing whether the
H1, The specified-slope model (Model 2 or Model 3) does
resulting regression equation fits the data as well as either of
not fit the data as well as the best-fit regression model
the two regression equations for the individual datasets. If the
(Model 1).
individual and combined regressions are shown to have statis-
The F statistic is computed as (Weisberg, 1985)
tically similar intercepts, then one equation adequately repre-
sents the data for both flow types. Alternatively, the intercept
for the single dataset may be significantly different from that
for the combined datasets, in which case separate equations for
each flow type are warranted. Comparing pairs of datasets (for
(8)
10 5 10 8
CROSS-SECTIONAL AREA, A, IN SQUARE METERS

10 4 10 7
10 3 10 6
10 2 10 5
10 1 10 4
10 0 10 3
10 -1 10 2
debris-flow data debris-flow data
specified 2/3 slope specified 2/3 slope
10 -2 best-fit regression 10 1 best-fit regression
specified zero slope specified zero slope
10 -3 10 0
10 0 10 1 10 2 10 3 10 4 10 5 10 6 10 7 10 8
10 4 10 5 10 6 10 7 10 8 10 9 10 10 10 11 10 12
FLOW VOLUME, V, IN CUBIC METERS
FLOW VOLUME, V, IN CUBIC METERS
Figure 4. Debris-flow data and three regression models for Figure 6. Rock-avalanche data and three regression models
cross-sectional area (denoted by DF-A1, A2, A3 in table 1). Model for cross-sectional area (denoted by RA-A1, A2, A3). Model 1 is
1 is the best-fit regression. Model 2 is the specified 2/3-slope the best-fit regression. Model 2 is the specified 2/3-slope model.
model. Model 3 is the specified zero-slope model expressing that Model 3 is the specified zero-slope model expressing that cross-
cross-sectional area has no dependence on flow volume. sectional area has no dependence on flow volume.
10 8 10 10
10 7 10 9
PLANIMETRIC AREA, B, IN SQUARE METERS
10 6 10 8
10 5 10 7
10 4 10 6
10 3 10 5
10 2 10 4
debris-flow data rock-avalanche data
specified 2/3 slope specified 2/3 slope
10 1 best-fit regression 10 3
best-fit regression
specified zero slope specified zero slope
10 0 10 2
10 0
10 1
10 2
10 3
10 4
10 5
10 6
10 7
10 8 10 4 10 5 10 6 10 7 10 8 10 9 10 10 10 11 10 12
FLOW VOLUME, V, IN CUBIC METERS FLOW VOLUME, V, IN CUBIC METERS
Figure 5. Debris-flow data and three regression models for Figure 7. Rock-avalanche data and three regression models for
planimetric area (denoted by DF-B1, B2, B3 in table 2). Model planimetric area (denoted by RA: B1, B2, B3). Model 1 is the
1 is the best-fit regression. Model 2 is the specified 2/3-slope best-fit regression. Model 2 is the specified 2/3-slope model.
model. Model 3 is the specified zero-slope model expressing that Model 3 is the specified zero-slope model expressing that
planimetric area has no dependence on flow volume. planimetric area has no dependence on flow volume.
Table 1. Parameters and analysis-of-variance statistics for alternative linear models of log-transformed debris-flow data for
inundated cross-sectional area, A, as a function of flow volume, V.
Models for prediction of cross-sectional area of inundation, A
Parameter Best-fit regression, Specified 2/3 slope, Specified zero slope,
(Model DF-A1) (Model DF-A2) (Model DF-A3)
Calibrated slope of the line 0.59 0.67 0
Calibrated intercept of line at log V =0 -0.66 -0.97 1.6
α-coefficient log -1(-0.66) = 0.22 log -1(-0.97) = 0.11 log -1(1.6) = 44
Number of data pairs, n 50 50 50
Residual degrees of freedom, DF 48 49 49
Residual sum of squares, SS 9.0 9.6 41
Residual mean square, MS 0.19 0.20 0.84
Standard error of model, σ 0.43 0.44 0.91
Coefficient of determination, r2 0.78 0.77 0.0
F-statistic, (comparison to model A1) NA 3.2 170
Table 2. Parameters and analysis-of-variance statistics for alternative linear models of log-transformed debris-flow data for
inundated-planimetric area, B, as a function of flow volume, V.
Models for prediction of planimetric area of inundation, B
(Model DF-B1) (Model DF-B2) (Model DF-B3)
Calibrated intercept of line at log V =0 1.0 1.3 4.0
α-coefficient log-1(1.0) = 10 log-1(1.3) = 19 log-1(4.0) = 9,000
F-statistic, (comparison to model B1) NA 3.7 470
Table 3. Parameters and analysis-of-variance statistics for alternative linear models of log-transformed rock-avalanche data for
inundated cross-sectional area, A, as a function of flow volume, V.
Models for prediction of cross-sectional area of inundation, A
(Model RA-A1) (Model RA-A2) (Model RA-A3)
Calibrated intercept of line at log V =0 -1.0 -0.64 4.7
α-coefficient log -1(-1.0) = 0.10 log -1(-0.64) = 0.23 log -1(4.7) = 52,000
F-statistic, (comparison to model A1) NA 2.7 58
Table 4. Parameters and analysis-of-variance statistics for alternative linear models of log-transformed rock-avalanche data for
inundated-planimetric area, B, as a function of flow volume, V.
Models for prediction of planimetric area of inundation, B
(Model RA-B1) (Model RA-B2) (Model RA-B3)
Calibrated intercept of line at log V =0 0.73 1.4 6.8
α-coefficient log-1(0.73) = 5.3 log-1(1.4) = 24 log-1(6.8) = 6,000,000
Residual sum of squares, SS 27 28 130
F-statistic, (comparison to model B1) NA 6.2 540
Table 5. Summary table of the best-fit regression equations with calibrated slopes and intercepts (Model 1).
Debris flows, models DF-A1 and DF-B1 A=0.22V 0.59 B=10V 0.73
Rock avalanches, models RA-A1 and RA-B1 A=0.10V 0.71 B=5.3V 0.75
Lahars, (Iverson and others, 1998) models LA-A1 and LA-B1 A=0.062V 0.65 B=110.V 0.69
Table 6. Summary table of the best-fit regression equations with specified 2/3 slopes and calibrated intercepts (Model 2). Only one
significant digit is reported for the α-coefficients.
Debris flows, models DF-A2 and DF-B2 A=0.1 V 2/3 B=20 V 2/3
Rock avalanches, models RA-A2 and RA-B2 A=0.2 V 2/3 B=20 V 2/3
Lahars, (Iverson and others, 1998) models LA-A2 and LA-B2 A=0.05 V 2/3 B=200 V 2/3
lahars, debris flows, and rock avalanches), for both V, A and V, each case, then the residual degrees of freedom are reduced by
B data, to each of the individual datasets results in 12 tests. 1. Thus, the best-fit regression model for an individual dataset
The test procedure is formalized by adding a term to has n-3 residual degrees of freedom in this test. Tables 9 and
the simple linear-regression model (equation 9) to produce a 10 list the parameters used to compute the F statistics.
multiple-regression model (equation 10): The results of the F-test comparing combined and
individual models for cases with a specified slope of 2/3
log A = log α1 + β1 log V , and (9) are summarized in tables 11–13, and results for equivalent
tests for cases with best-fit slopes calibrated by regression
are summarized in tables 14–16. The tabulated values show
log A = log α1 + β1 log V + β Z Z that at the p=0.01 level, the implications of the F-test results
. (10) are the same, regardless of whether or not the models with
specified 2/3 slopes or freely calibrated slopes are considered.
The additional term in equation 10 includes a dummy Therefore, because we are interested primarily in the 2/3-slope
regression variable, Z, commonly used with categorical data— models, our discussion below focuses exclusively on the F-test
also referred to as an indicator variable or binary variable (for results for those models.
example, Wesolowsky, 1976). Following Wesolowsky (1976), The F-tests comparing the intercepts for the combined
values for Z are assigned as 0 or 1 to differentiate between two datasets (for example, rock avalanches and lahars) to those
datasets used in any test. The Z-value assignment is 0 for the for the individual datasets (either rock avalanches or lahars)
combined datasets (for example, rock avalanches and lahars yield mixed results (fig. 8–10 and tables 11–13). In half of the
planimetric areas) and 1 for the individual datasets. The form cases, the null hypothesis can be rejected at the p=0.01 level,
of equations 9 and 10 then implies that the individual data- and thus, separate y-intercepts are warranted for individual
set has y-intercept, logα1+βZ; and the combined dataset has datasets. In the other cases, the F-test results imply that little
intercept logα1. The null hypothesis states that the intercepts statistical difference exists between dataset y-intercepts.
for both the combined and individual datasets are the same, for Detailed examination of the statistical results for specific
example, that βZ is zero. Rejecting the null hypothesis means F-tests, however, reveals some subtleties that complicate this
that the difference in intercepts for the combined and indi- inference.
vidual datasets is significantly different from zero. The F-tests of greatest interest compare inundation equa-
As in the preceding section, we used the F-test to evaluate tions for lahars and rock avalanches, which have flow-volume
the confidence with which a null hypothesis can be accepted magnitudes that largely overlap (fig. 8, table 11). The F-tests
or rejected. In this case, however, we considered whether or generally confirm the visual impression that the data sets are
not null models (in which Z=0 in equation 10) for combined distinct, and there is little doubt that planimetric-area inunda-
datasets are statistically distinct from models for individual tion equations for lahars and rock avalanches have distinct
data sets (denoted by Z=1 in equation 10). For the models y-intercepts and warrant different inundation equations. On
summarized by equations 9 and 10, the model equations and the other hand, there is some ambiguity in the F-test results for
residual degrees of freedom for this series of F-tests are sum- cross-sectional area inundation. The tests show that the lahar
marized in table 8. For the case in which β1=2/3, the residual data are distinguishable from the combined data, but the rock
degrees of freedom for the null models (combined dataset) are avalanche data are not. The paucity of cross-section inundation
n-1, whereas the residual degrees of freedom for the alternate data for rock avalanches is at least partly responsible for this
models (individual datasets) are n-2. (The null model with 2/3 outcome, as it results in a small value for the residual degrees
slope has one free coefficient to calibrate, logα1, whereas the of freedom, which weakens the statistical test. The second
alternate model representing an individual dataset with 2/3 comparison is for lahars and debris flows, which have flow-
slope has two coefficients to calibrate, β0 and βZ). If the slope volume magnitudes that overlap between 105 and 107 m3
β1 is treated as a free parameter (calibrated by regression) in (fig. 9, table 12). On this basis alone, the datasets might be
Table 7. Summary of results of F-tests comparing the specified-slope models to the best-fit regression model with adjustable slope.
Specified 2/3-slope regression model Specified zero-slope regression model
Degrees of freedom F-value,
compared to null model compared to null model
(numerator, denominator) p= 0.01
F-statistic p-value F-statistic p-value
Debris flows
V,A 1, 48 7.2 3.2 0.080 170 2.2 x 10 -17
V, B 1, 42 7.3 3.7 0.061 470 2.0 x 10 -24
Rock avalanches
V, A 1, 11 9.6 2.7 0.13 58 1.0 x 10 -5
V, B 1, 140 6.8 6.2 0.014 540 6.8 x 10 -50
Table 8. Dummy variable terms and equations for null model (combined datasets) and alternate models (individual datasets).
Model Equation Residual Degrees of Freedom
Null with 2/3 slope log A = log α 1 + 2 log V n-1

3
Alternate with 2/3 slope log A = log α1 + 2 logV + β Z Z n-2

3
Null with best-fit slope log A = log α1 + β 1 logV n-2
Alternate with best-fit slope log A = log α 1 + β 1 log V + β Z Z n-3
Table 9. Parameters for computing the F-statistic to compare the specified 2/3-slope models with combined and individual datasets.
Combined datasets (null model)
Rock avalanche Lahar Debris flow Residual sum of squares, SSnull Number of samples, n1+n2
V, A 13.01 63
X X
V, B 32.83 186
V, A 8.008 31
X X
V, B 46.13 169
V, A 13.34 68
X X
V, B 21.55 71
Individual Datasets (multiple-regression model)
Rock avalanche Lahar Debris flow Residual sum of squares, SSregr Number of samples, n
V, A 2.314 13
X
V, B 27.95 142
V, A 2.008 18
X
V, B 2.064 27
V, A 9.557 50
X
V, B 4.505 44
V, A denotes the paired dataset for volume-cross sectional area.
V, B denotes the paired dataset for volume-planimetric area.
Table 10. Parameters for computing the F-statistic to compare the best-fit regression models with combined and individual
datasets.
Combined datasets (null model)
Rock avalanche Lahar Debris flow Residual sum of squares, SSnull Number of samples, n1+n2
V, A 12.81 63
X X
V, B 31.38 186
V, A 7.410 31
X X
V, B 46.09 169
V, A 11.18 68
X X
V, B 10.60 71
Individual datasets (multiple-regression model)
Rock avalanche Lahar Debris flow Residual sum of squares, SSregr Number of samples, n
V, A 2.274 13
X
V, B 26.77 142
V, A 1.967 18
X
V, B 1.867 27
V, A 8.968 50
X
V, B 4.145 44
V, A denotes the paired dataset for volume-cross sectional area.
V, B denotes the paired dataset for volume-planimetric area.
considered separate because the physical processes they rep- from 10 to 1010 m3), some aspects of their behavior probably
resent typically occur at different scales. (Nonvolcanic debris are similar, as discussed below.
flows generally do not achieve volumes greater than The third comparison is between debris flows and rock
106 m3 because the terrain in which they initiate, transport, and avalanches (fig. 10, table 13). For this case, the y-intercepts
deposit does not provide adequate relief and erodible material for the planimetric inundation equations are known to be simi-
to produce large flows. Lahars entrain and deposit material lar (see the discussion of α-values in the previous section).
across large distances and through great relief where ample Thus, the F-tests indicate that there is ambiguity in differen-
unconsolidated material is available, and they are, therefore, tiating the planimetric-area inundation equations. Moreover,
able to achieve greater volumes.) The equations for inundated the F-tests show that, for prediction of cross-sectional area,
planimetric area have distinct y-intercepts based on F-test the difference between the debris-flow equations and rock-
results, and therefore, separate equations for predicting inun- avalanche equations is not statistically significant. Therefore,
dated planimetric areas are warranted. On the other hand, the areas inundated by modest-sized debris flows and by great
intercepts for the cross-sectional area-inundation equations are rock avalanches exhibit statistically indistinguishable depen-
statistically indistinguishable. Therefore, although lahars and dencies on flow volume.
debris flows can have widely differing flow volumes (ranging
Table 11. F-test comparison between the combined dataset and individual datasets for rock avalanches and lahars.
Combined datasets (rock avalanches and lahars)
V, A
Note: log(y-intercept) for combined datasets is 0.09
Proposed
Individual dataset for comparison log (y-intercept), F p Outcome
α
Rock avalanches 0.2 1.42 0.28 accept null
Lahars 0.05 3.41 0.01 reject null
V, B
Note: log(y-intercept) for combined datasets is 30
Proposed
Individual dataset for comparison log (y-intercept), F p Outcome
α
Rock avalanches 20 3.25 0.000002 reject null
Lahars 200 3.73 0.0002 reject null
Table 12. F-test comparison between the combined dataset and the individual datasets for debris flows and lahars.
Combined datasets (debris flows and lahars)
V, A
Individual dataset for Proposed
F p Outcome
comparison log (y-intercept), α
Debris Flows 0.1 1.00 0.48 accept null
Lahars 0.05 1.77 0.10 accept null
V, B
F p Outcome
comparison log (y-intercept) , α
Debris Flows 20 5.68 0.0000003 reject null
Lahars 200 5.25 0.00002 reject null
Table 13. F-test comparison between the combined dataset and the individual datasets for rock avalanches and debris flows.
Combined Datasets (debris flows and rock avalanches)
V, A
F p Outcome
comparison log (y-intercept), α
Rock avalanches 0.2 1.00 0.54 accept null
Debris flows 0.1 1.24 0.28 accept null
V, B
Individual Dataset for Proposed
F p Outcome
Comparison log (y-intercept) , α
Rock avalanches 20 0.54 0.99 accept null
Debris flows 20 1.85 0.01 reject null
Table 14. F-test comparison between combined rock avalanche and lahar datasets and the listed individual datasets for a best-fit
regression slope (non-2/3). Statistical results are the same as the case where the regression slope is forced to 2/3.
V, A F p Outcome
Rock avalanches 1.19 0.40 accept null
Lahars 2.96 0.02 reject null
V, B F p Outcome
Rock avalanches 3.58 0.0000003 reject null
Table 15. F-test comparison between combined debris flow and lahar datasets and the listed individual datasets for a best-fit
regression slope (non-2/3). Statistical results are the same as the case where the regression slope is forced to 2/3.
V, A F p Outcome
Debris flows 0.61 0.88 accept null
Lahars 1.38 0.25 accept null
V, B F p Outcome
Debris flows 2.28 0.008 reject null
Table 16. F-test comparison between combined rock avalanche and debris flow datasets and the listed individual datasets for a
best-fit regression slope (non-2/3). Statistical results are the same as the case where the regression slope is forced to 2/3.
V, A F p Outcome
Rock avalanche 0.91 0.62 accept null
Debris flows 1.44 0.17 accept null
V, B F p Outcome
Rock avalanche 0.53 0.99 accept null
Debris flows 1.88 0.01 reject null
108 1011
CROSS-SECTIONAL AREA, A (m 2)
PLANIMETRIC AREA, B (m2)

107 1010
rock avalanches
106 109
lahars
105 108
104 107
103 106
lahars
102 105
101 104
rock avalanches
100 103
10-1 102
10-2 101
101 102 103 104 105 106 107 108 109 1010 1011 10 10 10 10 10 10 10 10 10 10 10
1 2 3 4 5 6 7 8 9 10 11
FLOW VOLUME, V (m3) FLOW VOLUME, V (m3)
Figure 8. Rock-avalanche and lahar data used to compute statistics as a combined dataset (see table 11). On the cross-sectional
area plot, the data for rock avalanches and lahars are not obviously separated. On the planimetric-area plot, the data trends are visibly
separate.
108 1011

107 1010
106 109
lahars lahars
105 108
104 107
103 106
102 105
101 104
debris flows
100 103
10-1
debris flows 102
10-2 101
101 102 103 104 105 106 107 108 109 1010 1011 101 102 103 104 105 106 107 108 109 1010 1011
Figure 9. Debris-flow and lahar data used to compute statistics as a combined dataset (see table 12). On the cross-sectional area plot,
the data for rock avalanches and lahars generally occupy different potions of the flow volume range, but the data trends are not visibly
separate. On the planimetric-area plot, the data trends appear somewhat separate.
108 1011

107 1010
106 109
rock avalanches rock avalanches
105 108
104 107
103 106
102 105
101 104
debris flows
100 103
debris flows
10-1 102
10-2 101
101 102 103 104 105 106 107 108 109 1010 1011 10 10 10 10 10 10 10 10 10 10 10
1 2 3 4 5 6 7 8 9 10 11
Figure 10. Debris-flow and rock-avalanche data used to compute statistics as a combined dataset (see table 13). On the cross-
sectional area plot, the data for rock avalanches and debris flows occupy different potions of the flow-volume range, but the data trends
are not visibly separate. On the planimetric-area plot, the data trends appear somewhat separate.
Error and Uncertainty in Calibrated Prediction natural events, but if a lower degree of confidence can be tol-
erated, then the range of prediction uncertainty is reduced. We
Equations
chose to stipulate a high degree of confidence (99-percent),
The standard errors and predictive uncertainties of the resulting in portrayal of wide confidence limits.
2/3-slope models with best-fit calibrated α-coefficients
(table 6) have important ramifications for how the equations Discussion and Interpretation of Predictive
are interpreted and used. The standard errors, σ, of these
models (tables 1–4) characterize the variability of behavior Equations
exhibited by past flows, and the 99-percent confidence interval
Statistical results show that the power-law equations sum-
curves for prediction (figs. 11–13) characterize the uncertain-
marized in table 6 adequately relate inundation areas to flow
ties of predicting areas inundated by future flows if volume is
volumes, and that power-law equations with a specified slope
known with certainty (cf. Helsel and Hirsh, 1992). The factors
of 2/3 are indistinguishable from best-fit power-law regression
of error (10σ ) for predicting inundation areas (A, B) for rock
equations. Therefore, equations of the form Area = α Vol-
avalanches (10σ = 2.7, 2.8) are somewhat larger than the fac-
ume 2/3 are adopted for purposes of hazard-zone delineation.
tors of error for debris flows (10σ = 2.7, 2.1) and lahars
Only one significant digit is used to specify the α-coefficients
(10σ = 2.2, 1.9).
in the predictive equations listed in table 6; greater precision is
As an example of the effect of uncertainty, we considered
unwarranted given the uncertainties inherent in the predictive
prediction of planimetric areas inundated by debris flows.
models.
Taken alone, the inundation-area equation predicts that a
Constant power-law exponents (constant slopes on log-
debris flow of 105 m3 would inundate a planimetric area of
log plots) in our predictive equations imply a fractal scaling,
43,000 m2. The 99-percent confidence interval for prediction
and such geometric fractals are said to be scale invariant (Peit-
of the inundation area (outer set of curves, fig. 12) for the
gen and others, 1992). Thus, for a wide range of flow volumes,
same flow volume, however, indicates a wide range of pos-
inundated areas will appear similar when portrayed on maps
sible inundation areas, between 6,000 and 300,000 m2. At the
of any scale. Clearly, however, there is a practical limit below
95-percent confidence level (not shown in figures) this range
which the power-law equations do not apply. For example, as
shrinks to between 10,000 and 200,000 m2, but the uncertainty
volume falls below the smallest observed flow volume
remains considerable. The confidence intervals for predic-
(101 m3) and approaches zero, we should not expect the equa-
tion are great owing to the data scatter associated with diverse
tions to apply. This is an important consideration because
10 7 10 11

10 6 10 10
10 5 10 9
10 4 10 8
10 3 10 7
10 2 10 6
lahar data lahar data

10 1 10 5
best fit regression best fit regression
99-percent confidence 99-percent confidence
interval for regression interval for regression
10 0 10 4
interval for prediction interval for prediction
10 -1 10 3
10 3
10 4
10 5
10 6
10 7
10 8
10 9
10 10
10 11 10 3 10 4 10 5 10 6 10 7 10 8 10 9 10 10 10 11
Figure 11. Lahar data and best-fit regression line (solid) with 99-percent confidence interval for regression (inner pair of dashed
curves) and 99-percent confidence interval for prediction (outer pair of dashed curves; computed following Helsel and Hirsch, 1992)
after Iverson and others (1998). Note that Iverson and others used 95-percent confidence intervals.
10 5 10 8
10 4 10 7
10 3 10 6
10 2 10 5
10 1 10 4
10 0 10 3
debris-flow data debris-flow data

10 -1 10 2 best fit regression
best fit regression
10 -2 10 1 99-percent confidence
99-percent confidence
10 -3 10 0
10 0 10 1 10 2 10 3 10 4 10 5 10 6 10 7 10 8 10 0 10 1 10 2 10 3 10 4 10 5 10 6 10 7 10 8
Figure 12. Debris-flow data and best-fit regression line (solid) with 99-percent confidence interval for regression (inner pair of dashed
curves) and 99-percent confidence interval for prediction (outer pair of dashed curves; computed following Helsel and Hirsch, 1992).
10 8 10 11

10 7 10 10
10 6 10 9
10 5 10 8
10 4 10 7
10 3 10 6
rock-avalanche data rock-avalanche data

10 2 10 5
best fit regression best fit regression
10 1 10 4
10 0 10 3
10 4 10 5 10 6 10 7 10 8 10 9 10 10 10 11 10 12 10 4 10 5 10 6 10 7 10 8 10 9 10 10 10 11 10 12
Figure 13. Rock-avalanche data and best-fit regression line (solid) with 99-percent confidence interval for regression (inner pair of
dashed curves) and 99-percent confidence interval for prediction (outer pair of dashed curves; computed following Helsel and Hirsch,
1992).
log-transformed versions of the equations would imply a 99-percent confidence level. This lack of distinction implies
non-zero y-intercept where volume is equal to zero, and this is that there is considerable overlap in the physical behavior of
physically impossible. This observation doesn’t imply that the these three types of flows, at least insofar as behavior deter-
equations do not hold true, but that they operate across a wide mines cross-sectional inundation. For inundation-prediction
but finite range of scales. purposes, we gained some resolving power by retaining
According to the coefficients of the calibrated 2/3-power- distinct equations for the three types of flows, and this is the
law models (table 6), lahars typically inundate a planimetric course we adopted in practice.
area roughly ten times greater than do debris flows or rock With respect to equations for prediction of planimetric
avalanches of the same volume. Rock avalanches typically inundation areas, however, there is a clear statistical distinc-
inundate valley cross-sectional areas four times greater than tion between the lahar equation B = 200 V2/3 and the equation
does a lahar of equal volume, and two times greater than a that applies to both debris flows and rock avalanches, B =
debris flow of equal volume. The relatively large cross-sec- 20 V2/3. Separate inundation equations should be applied for
tions inundated by rock avalanches imply the existence of a separate types of flows.
relatively resistive, bulky cross-sectional surge front, whereas The development and comparison of the inundation-area
debris flows flow more fluidly, and lahars flow more fluidly equations listed in table 6 is the lack of clear statistical distinc-
still. The explanation for these differences probably lies in tion between the equations that apply for debris flows and the
the typical compositions and typical topographies traversed equations that apply for rock avalanches. When considering
by each type of flow (for example, Vallance and Scott, 1997). this statistical similarity, however, it must be borne in mind
However, physical or geological interpretations of differing that the rock-avalanche data included in the dataset apply to
α coefficients in our power-law equations must be tempered in events with volumes that are, on average, several orders of
view of statistical results that show that differing coefficients magnitude larger than the volumes of debris flows included
may not be truly distinct. Statistical comparison of the cali- in the data set. In terms of inundation patterns, large rock
brated equations for inundated cross-sectional areas avalanches tend to appear like small, nonvolcanic debris flows
(A = 0.05V2/3, A = 0.1V2/3 and A = 0.2V2/3 for lahars, debris viewed at a proportionally smaller scale. The reasons for this
flows, and rock avalanches, respectively) shows that the geometric similarity may be severalfold. Nonvolcanic debris
α-coefficient values (0.05, 0.1, and 0.2) are not distinct at the flows (with volumes generally less than 106 m3) commonly
Debris-Flow Hazard Delineation by using DEMs and GIS 21
entrain large boulders, logs, and other coarse debris that Data Input
concentrate at the flow front and produce a high-friction zone
there (Iverson, 1997). In many cases, individual clasts at the Data input for LAHARZ includes a range of postulated
flow front have dimensions comparable to the flow thickness flow volumes, digital topographic data, a suite of flow-path
or the flow width. Nonvolcanic debris flows are commonly data derived from the topography, and identification of source
channeled into narrow stream courses in finely dissected areas where debris flows may originate. Specification of
terrain, and a large degree of form drag thereby retards flow appropriate flow volumes generally requires an estimate of
runout. In this sense, the fronts of nonvolcanic debris flows available surficial material on slopes or within channels, as
may behave much like the fronts of large rock avalanches well as a historical perspective of the range of volumes that a
that surmount small-terrain obstacles and respond only to the given geographic area has produced. For example, the Coast
largest features of topography. Much evidence exists to sug- Range in Oregon would not likely produce a flow with a
gest that large rock avalanches are, to some degree, fluidized volume of 1 km3 because the topographic relief in prospective
at their bases, facilitating distal runout (for example, Legros, source areas is typically of the order of 100 m, and therefore,
2002). Although there is still controversy concerning the exact a maximum credible flow volume might be 0.001 km3. The
mechanism by which this fluidization can occur in the absence range of postulated flow volumes should span 2 or 3 orders of
of saturation with water, there is no doubt that large rock ava- magnitude to encompass a range of probability of occurrence.
lanches can to some extent behave like much smaller, lique- Volumes ranging from smaller (more frequent) to larger (less
fied debris flows. Thus, the similarity in power-law inundation frequent) depict areas of decreasing inundation hazard because
equations for nonvolcanic debris flows and much larger rock areas that are likely to be inundated more frequently have
avalanches is perhaps unsurprising. greater hazard associated with them.
Selecting the increment between successive flow vol-
umes used to forecast hazards can be done so as to mirror the
intrinsic statistical error of the models. For debris flows the
Debris-Flow Hazard Delineation by standard error, σ, of the calibrated power-law models (tables
using DEMs and GIS 6–9) is < 0.5 (ranging from 0.32 to 0.45), such that a conserva-
As an example of application of our inundation-area

equations to hazard-zone delineation, we considered an area
in the south-central Oregon Coast Range where significant
hazard from debris flows exists (fig. 14). Creating a hazard
zone map requires (1) obtaining topographic data of adequate
grid cell size, accuracy, and precision to represent the size
of the anticipated events and of sufficient extent to cover the
area of interest, (2) identifying potential debris-flow source
areas, (3) selecting appropriate potential flow volumes, (4)
using inundation-area equations to calculate the predicted A-
and B-values corresponding to the selected V-values, and (5)
computing and displaying the delineated planimetric areas as
nested hazard zones that depict uncertainty and relative levels
of hazard posed by the suite of potential flows. Automation
of the computations and display process is important, espe-
cially for high-resolution data covering large areas, and it also
ensures reproducibility.
As a platform for automation, we use GIS-based soft-
ware, LAHARZ, written in the ArcInfo Macro Language
(AML) (Schilling, 1998). This software has been used to
construct dozens of volcano hazard assessment maps around
the world (for example, Schilling and others, 2001; Vallance
and others, 2001; Scott and others, 2001; Gardner and others,
2004) and also used and compared to other methods of runout Figure 14. Aerial photograph (Google Earth) of the Scottsburg,
prediction (for example, Sheridan and others, 1999; Haapala Oregon, area. The dark ribbon is the Umpqua River. Mapped
and others, 2005; Sorenson, 2003). Below we summarize the recent debris-flow paths are in red, perennial streams are in blue,
data requirements and programming modifications necessary and the yellow line denotes the extent of the Oregon Department
to adapt LAHARZ for application to nonvolcanic debris flows. of Forestry’s research area (Robinson and others, 1999).
tive factor of error (10σ) is half an order of magnitude (100.5). Used in combination, the initiation criteria 1–3 are
Deliberate use of such a flow-volume increment is a technique somewhat similar to the criteria used to forecast sites of
that effectively enables the display of error bars in the form of shallow-landslide initiation in spatially distributed models,
nested hazard zones (Iverson and others, 1998). Depicting suc- such as SHALSTAB (Montgomery and Dietrich, 1994). Like
cessive hazard zones in a nested fashion implies that any given the criteria in SHALSTAB, our criteria are intended to identify
flow volume produces a unique, predicted inundation area ± all prospective source areas in a drainage basin simultane-
the area of the closest larger and smaller nested-hazard zones. ously. This simultaneous delineation does not imply that all
The demand for resolution and precision of topographic source areas will generate debris flows. Rather, it implies
data depends on the volumes of the flow events that are postu- that many source areas have equal probabilities of spawning
lated. For example, the digital elevation model (DEM) used in debris flows, thereby leading to simultaneous delineation of
the lahar hazard assessment at Mount Rainier, Washington, has many possible inundation paths downstream. In contrast to the
62.5-m grid cells for a map spanning much of western Wash- criteria used in SHALSTAB, our simple flow-initiation criteria
ington (Iverson and others, 1998). This resolution is adequate are not intended to emulate physical processes, but instead are
for large-volume lahars (> 107 m3) that inundate broad areas, entirely empirical — although they do exploit the common
as potential inundation areas are much greater than the size of observation that slope angle and upslope contributing area are
the individual grid cells. Small-volume debris flows (<105 m3) important morphometric parameters in determining flow-
may have flow widths less than 10 m and, therefore, require initiation sites.
cell-size resolution and accuracy of roughly a few meters.
The main LAHARZ algorithm computes hazard zones
downslope or downstream from source areas, but these areas Application Area for Debris-Flow Hazard
must be identified by some independent means. Source-area Mapping—Scottsburg, Oregon
identification in highly dissected topography is particularly
challenging because many prospective source areas may exist.
Our modified version of LAHARZ addresses the source-area History of the Scottsburg Area
problem by using a combination of three criteria appropriate
for identifying debris-flow initiation sites in the Oregon Coast In the south-central Oregon Coast Range (fig. 14), the
Range, our target for model application: town of Scottsburg and the Rock Creek-Hubbard Creek
(1) A specified minimum contributing-upslope source area watershed (also known as Stump Acres) in southern Douglas
must exist, analogous to contributing areas commonly County, Oregon, became infamous when two debris flows
used to evaluate surface-water runoff in hydrologic caused five fatalities between November 17 and 19, 1996.
models. In our implementation, the default minimum The deaths were the result of debris flows inundating a home,
contributing source area for debris flows is 103 m2. This overtaking a pedestrian, and pushing a car off the road (The
value derives from evidence of initial failure scars and the Oregonian, 1996; Mapes and Tims, 1996; Harvey and Squier,
onset of debris-flow inundation in site-specific reports for unpub. data, 1998). Highway 38 was blocked at multiple loca-
the Coast Range of southwestern Oregon (for example, tions by debris flows that inundated the road, and many more
Harvey and Squier, unpub. data, 1998). debris flows occurred in the vicinity that winter season (fig.
(2) A specified minimum slope angle must exist within the 14).
source area or along the computer-generated thalweg fed
by the source area. This criterion is based on the observa-
tion that debris flows generally initiate on slopes Topography and Geology
θ>30 degrees (Iverson and others, 1997), and we used The map area shown in fig. 15 (~5 km2) is characterized
30 degrees as the default minimum slope. by narrow, dissected valleys and sharp ridges, and the eleva-
(3) Slopes must be uniformly steep in the vicinity of a pro- tions within this area range from ~3 to 390 meters msl (above
spective initiation site. Here, uniformly steep is defined mean sea level). Several unnamed perennial streams flow
as 95-percent of grid cells, exceeding 30 degrees, in a along lower portions of tributary valleys feeding the Umpqua
100-m2 area centered on a prospective initiation cell. This River. The average and maximum slopes for the map area are
criterion excludes small patches of steep ground that are 35 degrees and 77 degrees, respectively, if flat areas (river and
unlikely to produce slope failures of significant size. roads) are excluded from consideration. These hillslopes are
Criteria 1 and 2 serve to identify many potential initiation exceedingly steep, even by Coast Range standards.
cells in a DEM, and criterion 3 limits the number of cells that The entire map area is underlain by the Tyee Formation,
define the starting points for debris-flow inundation. Crite- a late Eocene, eastward dipping, rhythmically bedded sand-
rion 3 also serves to limit the potential for DEM errors to stone and siltstone (Baldwin, 1961). The weakly consolidated
lead to false identifications of sufficiently steep ground. The bedrock is easily weathered and eroded, and it yields a suc-
specifications listed here are default settings for our modified cession of highly dissected ridges and small valleys. Where
LAHARZ program, but they may be adjusted after a sensitiv- the sandstone dominates, ridgelines are sharp; and where the
ity analysis is performed for any particular study area. siltstone dominates, ridges are moderately rounded. Based
123°52'0"W 123°51'0"W
Highway
38
r
R ive
ua
pq
Um
43°39'0"N
43°39'0"N
Mi l
l
St
r ea
m
Meters
0 500
123°52'0"W 123°51'0"W
Figure 15. Shaded-relief map with highways and logging roads for the study area west of the town of Scottsburg in the central Coast
Range of Oregon, at about 40 river kilometers from the coast.
on the Oregon Department of Forestry (ODF) Storm Impacts Topographic Dataset—Acquisition and
Study (Robinson and others, 1999) and the State of Oregon’s
Emergency Management Plan (OEMP, 2000), the highest Description
debris flow/torrent hazard is within steeply sloped areas in LIDAR (LIght Detection And Ranging) data acquired
the Tyee Formation (or similar sedimentary rocks) in Western by ODF on November 9, 1997 for the area around Scottsburg
Douglas, Coos, and Western Lane Counties. Most hillslopes were collected as part of a pilot study of damage assessment
that are steeper than 35 degrees can produce rapidly mov- of the impacts of the torrential storms and debris flows of
ing landslides, regardless of the underlying geologic unit. In 1996. LIDAR technology uses an airborne laser transmitter
designating high-risk areas, the ODF uses lower slope criteria and receiver coupled with GPS to map elevation profiles as
in the Tyee Formation than for other geologic types (Robinson the beam reflects off of surfaces such as tree canopy or bare
and others, 1999). earth. The elevation data are filtered to recover only bare-earth
surface return times for the production of a 1-meter DEM
with high resolution and accurate X, Y, and Z locations (for
example, Schickler and Thorpe, 2001).
The topographic data were processed by ODF to mini-
mize the effects of the dense tree canopy and derive elevations
of the bare earth surface. The study area excludes regions of Background for Flow Volume Assessment for
less dense data coverage around the periphery and therefore Scottsburg, Oregon
avoids coarse interpolation and false slopes. Within the DEM,
however, some areas have relatively few elevation data due to Some constraints on applicable flow volumes are pro-
dense tree coverage and fewer laser beams reaching the bare vided by the work of Ketcheson and Froelich (1978), who
earth surface. The accuracy of the data in these areas is still completed a landslide inventory in the Mapleton Ranger Dis-
high, but the elevation postings or resolution is reduced. The trict of the Oregon Coast Range and found that initial volumes
data points were used to create a TIN (triangulated irregular of slope failures ranged from 1.5 to 150 m3 for 104 events
network), interpolated to a grid, and drainage enforced. Drain- during a ten-year period (1966-1976). Although these initial
age enforcement is the filling of small topographic “sinks” volumes cannot be used to estimate maximum flow volume,
or single cell depressions that inhibit continuous flow across they do provide a minimum flow-volume estimate.
a DEM. This enforces continuous drainage by raising the Harvey and Squier (unpub. data, 1998) describe two
elevation of a single cell depression to the “pour” level of its debris flows that occurred within the same sub-basin as the
neighbors using a standard GIS hydrologic function. Hwy 38 MP 13 event and describe evidence observed in aerial
We used the TIN to generate a one-meter resolution DEM photographs for the Scottsburg area. In January 1990, a flow
that has not been checked for 1-m accuracy, and thus, GIS blocked the highway but did not enter the Umpqua River. In
products based on this topographic data should be interpreted December 1992, a flow deposited 500-600 m3 of material on
as preliminary. Such resolution and accuracy, however, are the highway. Aerial photographs taken in 1986 indicate recent
needed for forecasting debris flow inundation where flow
widths can be expected to be less than 10 meters.
""
""
"" ""
" "
"
" ""
"
" "
"
"" "
"
" "
"
" "
"
"
" ""
"" ""
" "
"""" """"
"
"
"" ""
""
"" "
"
"" "
"" "
" ""
"
" " " ""
""" " " """"""
" "
"
" ""
"
" "
"
" "
" "
" "
"
" """"
" ""
"" ""
"
" ""
" "
"
" """"
"
" """ ""
"
" "" """
"
" "
" ""
" " ""
"
" "
" ""
"
" "
" "" """""""""""
"
" "
" """""
" " "
"
" "
" """
"
"
" "
" """
"" "
" "" """"""""""
""
" " "" "
"
" "
" "" """"""
"
" "
" "
"" """""""
"
" "
" ""
"" ""
" " """""
""""""""""
"
" "
" "
" """""""""""""
"
" "
" "" """"
"" "
" "" """""""""""""""""""""""""""""""
" " " "
"
" "" "" ""
"
" "
" "" ""
"
" "" "
" "
" " "
"
" "
" "
"
"
" "
" "
"
"
" "" ""
" " "
" " ""
"
er
" " ""
"
" "
" ""
" "
" " " "
R iv
" " " "
" " " "
"
" "
" "
" """"
""
"
" "
" "
" ""
" " " "
a "
"
"
"
"
"
"
"
qu
"
" "" "
" ""
"
" "
" "" ""
"
" "
" "
" "
"
p
"
" "" "
" "" "
"" " " """ ""
"
""" "
" "
" "
Um
""
" " "" "
" ""
" "
" " "
" "
" ""
" " " " "
"
" "
" "" " "
" "" "" "
" """
" " """
""
"
" "" " "
" """""
" " "
" " "
"
" "" "" "
" ""
" "
" "
" "
" ""
"
" "
" "
" ""
" " " "
"" "
" "
" ""
"
"
" "
" "
" """
"""
"
" "
" "
" ""
" " " "
"
" "
" "
" ""
"" "
" "
" "
"
"
" "
" "
" ""
" " " "
"
" "
" "
" "
"
"
" "
" ""
"
" "
" ""
" " "
"
" "
" ""
"" "
" ""
"" "" ""
" " "
"
" "
" "
"
"
" "
" "
"
"
" "
" ""
" "
"
" ""
"
"" ""
""
"" ""
""
"" "
"
"" "
"" "" "
"
""
"
"
"
0 50
Meters
Meters
0 500
Figure 16. A, Shaded relief map constructed from a DEM, overlaid with orange where slopes exceed 30 degrees. The red box identifies
the test basin where the modified LAHARZ program is applied. B, Example of locations (red cells) meeting initiation criteria along
potential flow paths.
""" "
" ""
"" "" """
""""" "
" """"
""""" "" """
""
"" """"""""""
"" ""
"
"""
""
"" "
" """
"""" "
" """"
"" "
" """
""""" "
" """""
""""" "
" ""
""""""""""""" """""" "
" """"
"" "" "
" ""
"""""""""""""
""
""
"" "
" ""
"" "
" ""
"" ""
er
""
""""""""""""" "
" ""
""""" ""
R iv
"" "" "
""
"
"" """"" "
" ""
"
""
" """
"
" " "
" ""
""
""
ua
" " " "
"" """""""
" ""
""""
"
"" ""
" ""
""""
" " ""
" "" """
pq
""" " " " """
"" "" "" """
"
""""""""""
"" ""
"
" " "
" "" """" ""
""""" " " "" """""" """
Um
"""" ""
"
" """""""" " "" ""
"""""" """""""""""
"" "
"" " " """" "
" ""
"" "" " """"" " ""
""""""""""""""""" "
" "" "
" ""
"" "
" """"""""""""" "
"
"" "" """
"" "
" ""
"""""" "
" """"""
""" "
" ""
"
" "
""
"
""
"
""
"
"
"
"
""""
"" """
"
""
""
" "
"" "
"
"" "
"
"" """
"
""
""
" ""
"""" """"
"
"" ""
""
"""
" "
" " "
"
" " ""
" ""
""" "
""""""
"
"""""
""
" "
" ""
" "
"""
" "
""""""" " ""
""
"
""
" " "" ""
""" " " """"""
""""""
"
" ""
"""
"
" "
" ""
"" " "
"""""""""
"
" """""
""
"
" "
" """ """
" " "" ""
""
" "
" " """
"""""
" "
" "
" " ""
"
" "
" ""
" "" """
" " " """""""""""""
""
" "
" "
" "
" ""
"" " " """"
""
" "
" " ""
""
"
""
" "
"
" """ """""""""
"""
"""" "" "
" "
" "" """"""
""""""""""""""" "
""""
" "
" "
" "
""
"" """""""
"""""""" "" "
"
"
"
"
" """"""""""""""""
""
"""" " " " " """""""""""""
""""""""""" ""
" " "
" " " """"
"""""" ""
"
""
""
" " "" " "" """""""""""""""""""""""""""""""
""""" "
" "
" "
" "" ""
""" "
" "
" "
" "" ""
"
" "
" "
" " ""
"" "
" " "
""
"
"
" "
" " "
"
"
" "
" "
" "
""
"" "
" "" "
"" "
" "
" "
"""
"" "
" "
" "
"
"
" " " """
""
" "
" " " """
""
" " "
" "
" "
" "
" """""
" " " " ""
"" ""
" "
" " "
""
" " " "
" "
" " " "
""
" "
"" "
" " "
" ""
" " "" "" "
""
" "
" "
" "" "
" "
"
" " " " """"""
"" """
" "" "
" " ""
""
" "
" "
" " " "" ""
" ""
" "
"" " " ""
" "" "" "
" "
"" "
"" "
" "
" "" " "
"
"
" " " " " ""
"""
""
"" " "
" " " """""
""
" "" " "
" "" "" """ "
" ""
" """ "
" " " "
" ""
""" "" "
" "" "
" "
" ""
"" "" "
" "
" "
" "
" ""
"
" "" "" "
" "
" "
""
""
"
" "" "
" "
" "
" """"
"" "
" "
" "
" "
" """
" " " " "
""
" " " " ""
"" " " "
""
"" " " "
"" " "
" "" "
""
" " " "
""
" " " "
" " " "
""
" " " "
" "
" "
" ""
Meters
"" "
" "
" ""
"
" "" "
" ""
"
" """ "" ""
"
" "" "
"" ""
" " ""
0 500
"
" "
" "
"
"
" "
" "
"
"
" "
" ""
" "
"
" ""
"
""
" ""
"
"" ""
"
""
" "
"
"" "
"" ""
"
" "
"
"
"
"
"
"
"""
"
"
"
"
"
""
"
Meters ""
"
"
""
"
0 50
""
"
""
""
"
"
""
"
"
""
"
"" """""
" ""
"
"" "
" "
"
Figure 17. Shaded relief map of upper half of the test basin. Green-shaded polygons illustrate minimum contributing source areas
that define the start of potential flow paths (blue). Flow paths are not perennial stream channels, but are the paths of steepest descent
below an area of potential flow initiation. The image on the right shows the location of the upper half of the test basin (also fig. 16A) and
the prevalence of steep slopes in the test basin as indicated by the presence of the orange overlay where slopes exceed 30 degrees.
debris flows in adjacent sub-basins also reached the highway. initiation sites within the study basin (fig. 16B). Each poten-
Aerial photographs from 1997 indicate that the fatal 1996 tial initiation cell, shown in red, represents one square meter.
debris flow event was accompanied by multiple events within A total of 42 contributing source areas (≥103 m2 each) were
the upper reaches of the same subbasin and on adjacent slopes identified within the basin (shades of green, fig. 17).
outside of the sub-basin. Many of those flows reached the The selected potential flow volumes are 103, 103.5, 104,
highway. Total flow volumes, total runout lengths, and initia- and 104.5 m3 based on historic volumes for this region of the
tion locations were not recorded except for the single fatal Coast Range and a rough estimate of available erodible mate-
event. The initiation site for the 1996 event was ~45 m from rial. Reports about the Scottsburg area suggest that several
the ridgeline (roughly sketched on a 1:2,000 scale map) and debris flows in the last two decades had volumes of ~103 m3.
the runout length was about 400 m. This minimum flow volume may represent events of roughly
Harvey and Squire (unpub. data, 1998) note that the decadal frequency, and the larger flow volumes would repre-
November 1996 storm that produced the debris flows was sent less frequent events. In the study basin the longest poten-
preceded by 5 inches of precipitation in a 24-hour period and tial flow path is ~800 meters and the entrainment swath might
a total of 6 inches in 2 days. Harvey and Squier (unpub. data, be 30-50 meters in width with one or two meters of erosion,
1998) indicate that this 2-day intensity of rain was the great- yielding a conservative upper limit for flow volume ~104.5 m3.
est in the past 48 years, and their recurrence analysis indicates For the selected series of flow volumes (103, 103.5, 104,
that the storm event was a 50-year storm. and 104.5 m3), the inundation-area equations from table 6,
A=0.1V2/3 and B=20V2/3, predict cross-sectional areas (A) of 10,
22, 47, and 100 m2, and planimetric areas (B) of 2,000; 4,300;
LAHARZ Application 9,300; and 20,000 m2. With the aid of LAHARZ software, the
One basin (red box, fig. 16A) was selected from the four pairs of successively larger predicted A- and B- values
5 km2-area of LIDAR coverage to run trial applications. One produce a set of four nested inundation zones. One set of inun-
of the November 1996 debris flows emerged from this basin dation zones is produced for every initiation point (fig. 18A).
along the Umpqua River and pushed a traveling car off the Several amalgamated inundation-hazard zones are shown in
road and into the Umpqua River channel. figure 18B. Merging all the hazard zones from each of the ini-
Combining the three criteria for debris-flow initiation tiation points creates coalesced hazard zones (fig. 19) for each
(see section on “Data Input”) identifies hundreds of potential postulated flow volume. The resulting debris-flow hazard map
for the whole basin is shown in figures 19 and 20.
Figure 18. Examples of nested, inundation-hazard zones computed for four flow volumes (orange is 103 m3, yellow is
103.5 m3, green is 104 m3, and brown is 104.5 m3). A, Inundation-hazard zones computed downstream from a single initiation point within a
flow path. B, Amalgamated inundation-hazard zones computed from seven initiation points (red cells along blue flow paths).
Figure 19. Debris-flow inundation-hazard map computed for four flow volumes ranging from 103 to 104.5 m3. Hazard zones show the
predicted inundation limits for debris flows that start anywhere within the basin where the initiation criteria are met. The spikiness of
the terminus of the LAHARZ hazard zones as the hypothetical flows spill out of the constricted channel are a result of the limitation of
flow directions across gridded elevation data.
Hazard Map Summary nels) near channel heads are less likely to be inundated than
are narrow paths.
The modeled hazard zones do not predict the runout of a The ragged edges of the computed inundation zones
particular debris flow but instead delineate the paths likely to result from using a computational grid having only four poten-
be inundated by a suite of debris flows with various volumes. tial transect directions (NW-SE, N-S, NE-SW, W-E) in which
The four nested hazard zones are colored from “cool” to “hot” to calculate any cross-section. Furthermore, the computational
to reflect which areas are most hazardous because they are methodology and, therefore, the hazard map do not account
subject to inundation by even the smallest debris flows. In explicitly for changes in flow volume (due to entrainment or
general, hazard is greatest along channel thalwegs (fig. 20), deposition) or run-up potential because the same predicted
and it decreases with distance and elevation above the valley A-value is used to compute the inundation cross-section along
floor. Moreover, bracketing flow volumes and using graded the entire length of a potential debris-flow path.
shades of color generates a hazard map that conveys visual
uncertainty. The nested hazard zones are equivalent to error
bars where the increments between flow volumes (100.5) con-
servatively exceed the statistical errors inherent in the inunda-
tion equations (0.32–0.45). The nested hazard zones imply that
wide paths (sometimes overlapping between adjacent chan-
Figure 20. Oblique perspectives of the test basin with debris flow-hazard zone overlays, view to the southeast. A, Oblique aerial
view. B, Up-valley view from the north bank of the Umpqua River, Oregon.
References Cited 29
Conclusions those in the historical or geologic record. A limitation of this

methodology is that flow dynamics and run-up potential where
The main objective of this work is to provide a statisti- granular mass flows encounter channel bends or obstacles are
cally tested and calibrated model that can be used to fore- not considered. Another limitation is that events representing
cast inundation by future debris flows and rock avalanches. statistical outliers (defined by the curves of 99-percent level of
To accomplish this objective, a database consisting of flow confidence for prediction) will not be predicted accurately.
volumes, maximum inundated cross-sectional areas, and total Quantifying recurrence intervals for specific flow vol-
inundated planimetric areas was compiled for 64 nonvolcanic umes in a frequency-magnitude analysis for a particular region
debris flows ranging in volume from 101 to 107 m3 and for could assign probabilities to each flow volume and, therefore,
143 rock avalanches ranging from 105 to 1011 m3 in volume. assign annual likelihood that a specific postulated flow would
The database complements the lahar inundation-area database reach a certain point down valley. Until such work is under-
of Iverson and others (1998) for events that range between 105 taken, postulating a range of future flow volumes provides an
and 1010 m3 in volume. indication of the range of inundation possibilities.
Power-law equations of the form Area = α·Volume 2/3
with calibrated α-coefficients provide a good fit
(r2 = 0.76 – 0.91) to the debris-flow and rock-avalanche data Acknowledgments
with standard errors ranging from 0.45 to 0.32. The result-
ing inundation-area equations for debris flows are A = 0.1 The authors thank USGS colleagues Dennis Helsel
V2/3 and B = 20 V2/3, and the inundation-area equations for and Susan Cannon for insightful reviews of this manuscript.
rock avalanches are A = 0.2 V2/3 and B = 20 V2/3, where V is Professor Albyn Jones of Reed College provided additional
flow volume, A is inundated cross-sectional area, and B is assistance in the statistical review of this work. Steve Schilling
inundated planimetric area. These equations are similar to a (USGS) wrote the original version of the computer program,
set of inundation-area equations (A = 0.05 V2/3 and B = 200 LAHARZ, which was applied in this project. Michelle Howell
V2/3) developed for lahars by Iverson and others (1998). The collected a significant portion of the rock avalanche data.
equations imply no scale dependence of the runout process We thank Andrew Fountain and Scott Burns (Portland State
for rock avalanches, debris flows and lahars, but the differing University) for serving on the first author’s Master’s thesis
α-coefficients of the equations indicate that the bulk mobili- committee and supervising phases of this project. Jason Hinkle
ties of different types of flow can differ. Specifically, the (ODF) and Jon Hofmeister (DOGAMI) provided field guid-
α-coefficients imply that rock avalanches and nonvolcanic ance in the area around Scottsburg, Oregon.
debris flows are less mobile than lahars because they inundate
planimetric areas roughly ten times smaller than do lahars of
similar volume. The maximum cross-sectional area inundated
by rock avalanches is, on average, four times greater than that References Cited
of lahars and twice as large as that of debris flows of the same
volume, although the statistical support for this distinction Baldwin, E. M., 1961, Geologic map of the lower Umpqua
is weak. The similarity of the inundation-area equations for River area, Oregon: U.S. Geological Survey Oil and Gas
relatively small, nonvolcanic debris flows and much larger Inventory Map OM–204, scale 1:62,500.
rock avalanches indicates that these two types of flows may
have strong mechanical similarities, even though the processes Benda, L.E., and Cundy, T.W., 1990, Predicting deposition of
operate on different scales. debris flows in mountain channels: Canadian Geotechnical
Implementation of the inundation-area equations within Journal, v. 27, p. 409–417.
the GIS-based program, LAHARZ, provides a repeatable,
Cannon, S.H., 1989, An evaluation of the travel-distance
objective way to produce hazard maps. Use of a range of
potential of debris flows: Utah Geological and Mineral
hypothetical flow volumes that span orders of magnitude pro-
Survey Miscellaneous Publications, 89–2, 35 p.
vides graphical output that depicts uncertainty in the predicted
inundation areas. Not knowing the volume of the next event Crandell, D.R., 1989, Gigantic debris avalanche of Pleistocene
and, therefore, the inundation limits for that volume, is toler- Age from Ancestral Mount Shasta Volcano, California, and
able where hazard is assessed in terms of relative likelihood of debris-avalanche hazard zonation: U.S. Geological Survey
inundation for a range of flow volumes instead of the inunda- Bulletin 1861.
tion limits of the next event.
A strength of this methodology compared to traditional Crandell, D.R., and Mullineaux, D.R., 1967, Volcanic hazards
hazard-mapping techniques is that a broad range of possible at Mount Rainier, Washington: U.S. Geological Survey Bul-
volumes is considered in delineating maximum inundation letin 1238, 26 p.
areas. Rather than making predictions of future runout based
on the maximum extent of deposits of past events, this method
allows for the possibility that future events will be larger than
Crandell, D.R., and Mullineaux, D.R., 1975, Technique and Helsel, D.R., and Hirsch, R.M., 1992, Statistical methods in
rationale of volcanic-hazards appraisals in the Cascade water resources: Amsterdam, Elsevier, 529 p.
Range, northwestern United States: Environmental Geol-
ogy, vol.1, no.1, p. 23–32. Heim, A., 1921, Geologie der Schweiz. Band 2: die Schweizer
Alpen 1. Halfte. Tauchniz Leipzig.
Crosta, G.B., Cucchiaro, S., and Frattini, P., 2003, Validation
of semi-empirical relationships for the definition of debris- Heim, A., 1932, Der bergsturz und Menschenleben: Zurich,
flow behavior in granular materials in Rickenmann, D., and Fretz und Wasmuth, 218 p.
Chen, C.-L., eds., Debris-flow hazards mitigation; mechan- Hsu, K.J., 1975, Catastrophic debris streams (sturzstroms)
ics, prediction, and assessment: Rotterdam, Millpress, p. generated by rockfalls:
821–831. Geological Society of America Bulletin, v. 86, p. 129–140.
Dade, W.B., and Huppert, H.E., 1998, Long-runout rockfalls: Hungr, O., 1990, Mobility of rock avalanches: Report of the
Geology, v. 26, no. 9, p. 803–806. National Research Institute for Earth Science and Disaster
Davies, T.R.H., 1982, Spreading of rock avalanche debris by Prevention, Japan, v. 46, p. 11–20.
mechanical fluidization: Rock Mechanics, v. 15, p. 9–24. Iverson, R.M., 1997, The physics of debris flows: Reviews of
Denlinger, R.P., and Iverson, R.M., 2001, Flow of variably Geophysics, v. 35, no. 3, p. 245–296.
fluidized granular masses across three-dimensional terrain; Iverson, R.M., 2003, The debris-flow rheology myth in Rick-
2— Numerical predictions and experimental tests: Journal enmann, D., and Chen, C.-L., eds., Debris-flow hazards
of Geophysical Research, v. 106, no. B1, p. 553–566. mitigation; mechanics, prediction, and assessment: Rotter-
Denlinger, R.P., and Iverson, R.M., 2004, Granular avalanches dam, Millpress, p. 303–314.
across irregular three-dimensional terrain; 1—Theory and Iverson, R.M., 2005, Regulation of landslide motion by dilat-
computation: Journal of Geophysical Research, v. 109, ancy and pore pressure feedback: Journal of Geophysical
F01014. Research, v. 110, F02015.
Eisbacher, G.H., and Clague, J.J., 1984, Destructive mass Iverson, R.M., Costa, J.E., and LaHusen, R.G., 1992, Debris-
movements in high mountains; Hazard and management: flow flume at H.J. Andrews Experimental Forest, Oregon:
Geological Survey of Canada, Paper 84–16, 230 p. U.S. Geological Survey Open-File Report 92–483, 2 p.
Fannin, R.J., and Rollerson, T., 1993, Debris flows; some Iverson, R.M., and Denlinger, R.P., 2001, Flow of variably
physical characteristics and behaviour: Canadian Geotech- fluidized granular masses across three-dimensional ter-
nical Journal, v. 30, no. 1, p. 71–82. rain; 1— Coulomb mixture theory: Journal of Geophysical
Fannin, R.J., and Wise, M.P.W., 2001, An empirical-statistical Research, v. 106, no. B1, p. 537–552.
model for debris Iverson, R.M., Logan, M., and Denlinger, R.P., 2004, Granular
flow travel distance: Canadian Geotechnical Journal, v. 38, avalanches across irregular three-dimensional terrain; 2—
no. 5, p. 982–994. Experimental tests: Journal of Geophysical Research,
Gardner, C.A., Scott, W.E., Ewert, J.W., Griswold, J.P., John- v. 109, F010115.
son, M.C., and Devoli, G., 2004, Volcano hazard zones for Iverson, R.M., Reid, M.E., and LaHusan, R.G., 1997, Debris-
Volcàn Momotombo, Nicaragua: U.S. Geological Survey flow mobilization from landslides: Annual Review of Earth
Open-file Report 2004–1012. and Planetary Sciences, v. 25, p. 85–138.
Griswold, J.P., 2004, Mobility statistics and hazard mapping Iverson, R.M., Schilling, S.P., and Vallance, J.W., 1998,
for non-volcanic debris flows and rock avalanches: Port- Objective delineation of lahar-hazard inundation zones:
land, Oregon, Portland State University, master’s thesis, Geological Survey of America Bulletin, v. 110, no. 8,
102 p. p. 972–984.
Haan, C.T., 1977, Statistical methods in hydrology: Ames, Iverson, R.M., and Vallance, J.W., 2001, New views of granu-
Iowa State University Press, 378 p. lar mass flows: Geology, v. 29, no. 2, p. 115–118.
Haapala, J.M., Escobar Wolf, R., Vallance, J.W., Rose, W.I., [Jackli, 1957] compiled in Li Tianchi, 1983, A mathemati-
Griswold, J.P., Schilling, S.P., Ewert, J.W., and Mota, M., cal model for predicting the extent of a major rockfall:
2005, Volcanic hazards at Atitlan Volcano, Guatemala: U.S. Zeitschrift fur Geomorphologi N.F., Bd. 27, heft. 24,
Geological Survey Open-File Report 2005–1403, 19 p. p. 472–482.
References Cited 31
Jakob, M., and Hungr, O., eds., 2005, Debris-flow hazards and Myers, B., and Brantley, S. R., 1995, Volcano hazards fact
related phenomena, New York, Springer, 739 p. sheet; hazardous phenomena at volcanoes: U.S. Geological
Survey Open-File Report 95–231, 2 p.
Janda, R.J., Scott, K.M., Nolan, K.M., and Martinson, H.A.,
1981, Lahar movement, effects and deposits, in Lipman, Oregon Emergency Management Plan (OEMP), 2000, Land-
P.W., and Mullineaux, D.R., eds., The 1980 eruptions of slides and debris flows: Natural hazards mitigation plan,
Mount St. Helens, Washington, U.S. Geological Survey v. 1, June 2000.
Professional Paper 1250, p. 461–478.
The Oregonian, 1996, Landslides near Reedsport a ‘night-
Jefferys, W.H., and Berger, J.O., 1992, Ockham’s Razor and mare’: Portland, Oregonian Publishing Company, Sunrise
Bayesian Analysis: American Scientist, v. 80, p. 64–72. edition, November 22, 1996, from wire reports, p. B04.
Johnson, A.M., 1984, Debris flow, in Brunsden, D., and Prior, Peitgen, H.O., Hartmut J., and Dietmar, S., 1992, Chaos and
D.B., eds., Slope instability: New York, John Wiley and fractals; new frontiers of science: New York, Springer-
Sons, p. 257–361. Verlag, 984 p.
Jordan, P.R., 1994, Debris flows in the Southern Coast Moun- Pierson, T.C., 1985, Initiation and flow behavior of the 1980
tains, British Columbia: Dynamic Behaviour and Physical Pine Creek and Muddy River lahars, Mount St. Helens,
Properties: Vancouver, University of British Colombia, Washington: Geological Survey of America Bulletin, v. 96,
Ph.D. dissertation, 260 p. p. 1056–1069.
Ketcheson, G., and Froelich, H.A., 1978, Hydrologic factors Pierson, T.C., and Costa, J.E., 1987, A rheologic classifica-
and environmental impacts of mass soil movements in the tion of subaerial sediment-water flows, in Costa, J.E., and
Oregon Coast Range, Corvallis, Oregon: Water Resources Wieczorek, G.E., eds., Debris flows/avalanches—Process,
Research Institute Bulletin 56, Oregon State University. recognition, and mitigation, Geological Society of America
Reviews in Engineering Geolgoy, v. 7, p.1–12.
Keselman, H.J., Othamn, A.R., Wilcox, R.R., and Fradette, K.,
2004, The new and improved two-sample t test: Psychologi- Rickenmann, D., 1999, Empirical relationships for debris
cal Science, v. 15, no. 1, p. 47–51. flows: Natural Hazards, v. 19, p. 47–77.
Kilburn, C.R.J. and Sorenson, S.A., 1998, Runout lengths of Reid, M.E., and Iverson, R.M., 1992, Gravity-driven ground-
sturztroms; the control of initial conditions and of fragment water flow and slope failure potential, 2— Effects of slope
dynamics: Journal of Geophysical Research, v. 102, no. B8, morphology, material properties, and hydraulic heterogene-
p. 17877–17884. ity: Water Resources Research, v. 28, no. 3, p. 939–950.
Legros, F., 2002, The mobility of long-runout landslides, Engi- Robinson, E.G., Mills, K., Paul, J.T., Dent L., and Skaugset,
neering Geology, v. 63, 301–331. A., 1999, 1996 storm impacts monitoring project; Final
Report—Forest Practices Technical Report #4: Salem,
Li Tianchi, 1983, A mathematical model for predicting the Oregon Dept. Forestry, 157 p.
extent of a major rockfall: Zeitschrift für Geomorphologie,
v. 27, p. 473–482. Sassa, K., 1988, Geotechnical model for the motion of land-
slides, in Bonnard, C., ed., Proceedings, 5th international
Major, J.J., 1984, Geological and rheologic characteristics symposium on landslides, v. 1: Rotterdam, Netherlands,
of the May 18, 1980 southwest flank lahars at Mount St. A.A. Balkema, p. 37–55.
Helens, Washington: University Park, Pennsylvannia State
University, Master’s thesis, 225 p. Savage, S.B., and Iverson, R.M., 2003, Surge dynamics
coupled to pore-pressure evolution in debris flows in
Major, J.J., 1996, Experimental studies of deposition by debris Rickenmann, D., and Chen, C.-L., eds., Debris-flow hazards
flows; Process, characteristics of deposits, and effects of mitigation; mechanics, prediction, and assessment: Rotter-
pore-fluid pressure: Seattle, University of Washington, dam, Millpress, p. 503–514.
Ph.D. dissertation, 341 p.
Scheidegger, A., 1973, On the prediction of the reach and
Mapes, J., and Tims, D., 1996, Storm claims five lives: velocity of catastrophic landslides: Rock Mechanics, v. 5,
Portland, Oregonian Publishing Company, Sunrise edition, p. 231–236.
November 20, 1996, p. A01.
Schickler, W. and Thorpe, A., 2001, Surface estimation based
Montogmery, D.R., and Dietrich, W.E., 1994, A physically on LIDAR: Proceedings of ASPRS Annual Conference, St.
based model for the topographic control on shallow land- Louis, Missouri, April 2001.
sliding: Water Resources Research, v. 30, p. 1153–1172.
Scott, K.M., Wolfe, E.W. and Driedger, C.L., 1998, Mount Siebert, L., 1984, Large volcanic debris avalanches; Charac-
Rainier— Living with perilous beauty: U.S. Geological teristics of source areas, deposits, and associated eruptions:
Survey Fact Sheet 065–97, 4 p. Journal of Volcanology and Geothermal Research, v. 22,
p. 163–197.
Scott, W.E., Iverson, R.M., Schilling, S.P., and Fisher, B.J.,
2001, Volcano hazards in the Three Sisters Region, Oregon: Siebert, L., Glicken, H., and Ui, T., 1987, Volcanic hazards
Vancouver, Washington, U.S. Geological Survey Open-File from Bezymianny— and Bandai-type eruptions: Bulletin of
Report 99–437, 18 p. Volcanology, v. 49, p. 435–459.
Schilling, S.P., 1998, LAHARZ; GIS programs for automated Sorenson, O.E., 2003, Hazard mapping at Tungurahua
mapping of lahar-inundation hazard zones: U.S. Geological volcano, Ecuador: Houghton, Michigan Tech University,
Survey Open-File Report 98–638, 80 p. Master’s thesis.
Schilling, S.P., Vallance, J.W., Matías, O., and Howell, M.M., Takahashi, T., 1991, Debris flow: Rotterdam, Balkema, 165 p.
2001, Lahar hazards at Agua Volcano, Guatemala: U.S.
Geological Survey Open-File Report 01–432. Ui, T., 1983, Volcanic dry avalanche deposits— Identification
and comparison with nonvolcanic debris stream deposits:
Schuster, R.L., 1996, The 25 most catastrophic landslides of Journal of Volcanology and Geothermal Research, v. 18,
the 20th century in Chacón, J., Irigaray, C., and Fernádez, p. 135–150.
T., eds., Proceedings of the Eighth international conference
and field trip on landslides, Granada City, Spain, Ui, T., H. Yamamoto, and Suzuki-Kamata, K., 1986, Charac-
Sept. 27–28, 1996: Rotterdam, A.A. Balkema, p. 53–62. terization of debris avalanche deposits in Japan: Journal of
Volcanology Geothermal Research, v. 29, p. 231–243.
Sheridan, M.F., Hubbard, B., Hooper, D., and Abrams, M.
1999, Hazard zones demarking inundation limits for debris Vallance, J.W., Schilling, S.P., Devoli, G., Reid, M.E., and
flows and debris avalanches at Volcán Colima, México: Howell, M.M., 2001, Lahar hazards at Casita volcano,
EOS, Transactions American Geophysical Union Fall Meet- Nicaragua: Vancouver, Washington, U.S. Geological Survey
ing, v. 80, no. 46, p. 1141. Open-File Report 01–468, 24 p.
Shreve, R.L., 1968, Leakage and fluidization in air-layer lubri- Vallance, J.W., and Scott, K.M., 1997, The Osceola mudflow
cated avalanches: Geological Survey of America Bulletin, from Mount Rainer— Sedimentology and hazards implica-
v. 79, p. 653–658. tions of a huge clay-rich debris flow: Geological Survey of
America Bulletin, v. 109, p. 143–163.
Sidle, R.C., and Chigira, M., 2004, Landslides and debris
flows strike Kyushu, Japan: EOS Transactions, v. 85, no. Voight, B., ed., 1979, Rockslides and avalanches; natural phe-
15, p. 145, 151. nomena in Developments in geotechnical engineering 14A:
Elsevier, Amsterdam, 833 p.
Voight, B.R., Janda, J., Glicken, H., and Douglas, P.M., 1983,
Nature and mechanics of the Mount St. Helens rockslide-
avalanche on 18 May 1980: Geotechnique, v. 33,
p. 243–273.
Weisberg, S., 1985, Applied linear regression 2d ed.: New
York, Wiley, 324 p.
Wesolowsky, G.O., 1976, Multiple linear regression and

analysis of variance— An introduction for computer users
in management and economics: New York, Wiley, 292 p.
Maximum Inundated Total Inundated
Flow Volume Cross-sectional Area Planimetric Area
Ref # (cubic meters) (square meters) (square meters) Name of Event/ Location Region/ State/ Country Year *
V A B
DF 1 10,000,000 -- 1,900,000 Upper Lillooet River British Columbia, Canada 900 B.P.
DF 2 3,000,000 7,000 -- Klattasine Creek southern Rockies, British 1971-1973

Columbia, Canada
DF 3 3,000,000 3,000 200,000 Devastation Cr. British Columbia, Canada 1931
DF 4 2,000,000 -- 1,300,000 Caraballeda Fan/ Sierra de Vargas State, Venezuela December 1, 1999
Appendix A—Data Table
Avila
DF 5 1,600,000 1,000 -- Sesa Landslide/ Grigna Bienno, northern Italy 1993
valley
DF 6 1,200,000 300 -- Capricorn Creek Mount Meager, British 1998
Columbia, Canada
DF 7 1,000,000 200 570,000 Ophir Creek Washoe County, Nevada, May, 1983
USA
DF 8 1,000,000 400 -- Turbid Cr. British Columbia, Canada 1984
DF 9 200,300 -- 72800 Shadow Canyon vicinity of Boulder,

Colorado, USA
DF 10 200,000 -- 70,000 Capricorn Cr. British Columbia, Canada 1972
DF 11 195,000 3,700 160,000 Bullock Creek Mt Thomas, New Zealand April, 1978
DF 12 180,000 300 90,000 St. Peter's Dome Columbia Gorge, Oregon, November, 2001
USA
DF 13 162,700 -- 59200 Fern Canyon vicinity of Boulder,
Colorado
DF 14 150,000 40 -- Sourgrass Debris Flow Sierra Nevadas, California January 1, 1997
Appendix A—Data Table 33
V A B
DF 15 136,000 -- 77,000 Cathedral Mountain southern Rockies, British 1978
Columbia, Canada
DF 16 92,000 230 140,000 Hummingbird Creek Mara Lake, British July 11, 1997
Columbia, Canada
DF 17 90,000 4,200 50,000 Cathedral Mountain southern Rockies, British 1946
Columbia, Canada
Columbia, Canada
Columbia, Canada
DF 20 80,000 90.0 100,000 West Dodson Columbia Gorge, Oregon,
USA
DF 21 76,000 300 -- Polallie Creek Mount Hood, Oregon December 25, 1980
DF 22 63,000 77 -- Pierce Creek southern British Columbia November 28, 1995
DF 23 60,000 -- 35,000 Hot Springs Cr. British Columbia, Canada 1984
DF 24 55,000 48 -- Wahleach A British Columbia, Canada
DF 25 50,000 80 -- Hope Creek southern Rockies, British November 8, 1995

Columbia, Canada
DF 26 50,000 32 21,000 Boundary Cr. British Columbia, Canada 1987

DF 28 24,000 -- 12,000 Cathedral Mountain southern Rockies, B.C. 1962

V A B
DF 29 20,000 78 10,000 Canyon Cr. British Columbia, Canada 1990
DF 30 20,000 70 7,000 Lower Ryan R. tributary British Columbia, Canada 1984
DF 31 20,000 400 -- Mt. Currie gully British Columbia, Canada 1989
DF 32 20,000 48 -- M-Creek British Columbia, Canada
DF 33 20,000 47 -- Charles Creek British Columbia, Canada
DF 34 17,000 30.0 16,000 Mayflower Tenmile Range, Colorado August 18, 1961
DF 35 10,000 -- 5,000 No Good Cr. British Columbia, Canada 1990
DF 36 10,000 55 8,000 Canyon Cr. British Columbia, Canada 1987
DF 37 9,990 50.4 2,180 Black Cr. Olympics, Washington

Columbia, Canada
DF 39 6,000 16 -- McGillivray Cr. gully British Columbia, Canada 1989
DF 40 5,790 19.4 1,540 Marlow#1 Coast Range, Oregon, USA

Columbia, Canada
DF 42 5,000 12 8,600 Newton Canyon Santa Monica Mountains, December 29, 1965
California
V A B
DF 44 4,800 22.8 2,400 Slide Cr. San Bernardinos, California
DF 45 3,500 20 -- Fountain Ridge gully British Columbia, Canada 1989
DF 46 3,390 23.0 2,110 SlumpAcres Coast Range, Oregon, USA
DF 47 3,000 44 1,570 Tom McDonald Creek California (Redwood Creek December 16, 1982
basin), USA
DF 48 2,300 15.0 -- Oddstad California, USA
DF 49 2,000 11 -- Lillooet R. east fan British Columbia, Canada 1989
DF 50 1,050 30 1,100 New Zealand
DF 51 690 20.0 492 SlumpAcres 2 Coast Range, Oregon, USA
DF 52 660 38.8 621 Marlow#3 Coast Range, Oregon, USA
DF 53 660 5.0 -- Big Bend California, USA
DF 54 610 11.0 -- Yosemite site California, USA
DF 55 300 3.3 2,000 B1 California, USA

DF 56 200 5.6 -- Angel B gully British Columbia, Canada 1990

V A B
DF 57 200 3 -- United Kingdom
DF 58 150 -- 260 Old Topanga Canyon Santa Monica Mountains, January 26, 1969
California
DF 59 135 20.6 240 Marlow#4 Coast Range, Oregon, USA
DF 60 100 15 100 No Good Cr. British Columbia, Canada 1988
DF 61 100 2.7 600 N32 Canada
DF 62 27.87 0.9 200 Levan debris flow Utah, USA
DF 63 10 0.7 200 N2 Canada
DF 64 10 0.5 250 USGS flume experiments USA
* Dates are year A.D. unless specified as year B.P. (years before present).
Ref# Reference code and number indicates landslide type and key # for Appendix B/ data bibliography. DF = nonvolcanic debris flow; RA = rock avalanche;
LA = lahar.
-- Information is not available in source report.
V A B
RA 1 12,000,000,000 -- 51,000,000 Flims Alps
RA 2 45,000,000,000 2,000,000 675,000,000 Shasta California, USA 300,000-380,000 B.P.
RA 3 28,000,000,000 -- 300,000,000 Popocatepetl Mexico Pleistocene
RA 4 25,000,000,000 -- 1,085,000,000 Raung Java Quaternary
RA 5 18,000,000,000 -- 400,000,000 Avachinsky Kamchatka Pleistocene
RA 6 16,000,000,000 -- 400,000,000 Peteroa South America
RA 7 15,000,000,000 -- 480,000,000 Socompa South America Holocene
RA 8 15,000,000,000 -- 200,000,000 Antuco South America
RA 9 15,000,000,000 -- 1,400,000,000 Meru Africa
RA 10 12,600,000,000 -- 840,000,000 Wrangell Alaska, USA 200,000 B.P.
RA 11 12,000,000,000 -- 1,550,000,000 Volcan de Colima Mexico 4,028 B.P.
RA 12 8,100,000,000 -- 150,000,000 Chimborazo South America Pleistocene
RA 13 8,000,000,000 -- 80,000,000 Tungurahua South America 3,000 B.P.

RA 14 7,500,000,000 -- 250,000,000 Egmont, (1) New Zealand 23,000 B.P.

V A B
RA 15 7,100,000,000 -- 1,150,000,000 Mawenzi Africa
RA 16 7,000,000,000 -- 200,000,000 Drum USA Pleistocene
RA 17 3,500,000,000 800,000 170,000,000 Roque Nublo Canary Islands Pliocene
RA 18 3,000,000,000 -- 80,000,000 Calbuco South America Holocene
RA 19 2,900,000,000 -- 175,000,000 Galunggung Indonesia 23,000 B.P.
RA 20 2,800,000,000 -- 80,000,000 Jocotitlan Mexico 10,000 B.P.
RA 21 2,500,000,000 -- 9,000,000 Engelberg Alps
RA 22 2,500,000,000 -- 64,000,000 Mount St. Helens Washington, USA 1980
RA 23 2,100,000,000 -- 12,000,000 Kofels Alps
RA 24 2,000,000,000 -- 28,000,000 Siders (Sierre) Alps
RA 25 2,000,000,000 -- 90,000,000 Asama Japan
RA 26 2,000,000,000 1,600,000 -- Usoi Landslide Dam Pamir Mountains, Tajikistan 1911
RA 27 1,800,000,000 -- 70,000,000 Iriga Philippines 1628
RA 28 1,500,000,000 -- 98,000,000 Shiveluch Kamchatka 1964

V A B
RA 29 1,300,000,000 -- 65,000,000 Iwaki Japan
RA 30 1,500,000,000 -- 34,000,000 Bandai Japan 1888
RA 31 1,000,000,000 64,000 2,600,000 Mayunmarca Rockslide & Andes Mountains, Peru April 25, 1974
Debris Flow
RA 32 1,000,000,000 -- 14,500,000 Fernpass Alps
RA 33 1,000,000,000 -- 45,000,000 Mombacho Central America
RA 34 1,000,000,000 -- 20,000,000 San Pedro-Pellado South America Holocene
RA 35 900,000,000 -- 6,800,000 Kandertal Alps
RA 36 800,000,000 -- 8,800,000 Glarnisch-Guppen Alps
RA 37 800,000,000 -- 27,000,000 Popa Burma Pleistocene
RA 38 800,000,000 -- 30,000,000 Bezymianny Kamchatka 1956
RA 39 600,000,000 -- 4,300,000 Totalp Alps
RA 40 600,000,000 -- 6,400,000 Dejenstock Alps
RA 41 550,000,000 -- 11,600,000 Monte Spinale Alps

RA 42 500,000,000 -- 30,000,000 Augustine, (2) West Island Alaska, USA 500 B.P.
V A B
RA 43 500,000,000 -- 40,000,000 Sierra Velluda South America Holocene
RA 44 500,000,000 -- 25,000,000 Soufriere, W.I. Guadeloupe 3,000 B.P.
RA 45 498,000,000 -- 4,660,000 Maligne Lake Canadian Rocky Mountains
RA 46 469,000,000 -- 12,000,000 Marocche im Sarcatal Alps
RA 47 400,000,000 -- 8,000,000 Parpan-Lenzerheide Alps
RA 48 400,000,000 -- 3,600,000 Lago di Molveno Alps
RA 49 400,000,000 -- 11,000,000 Eibsee Alps
RA 50 400,000,000 -- 96,000,000 Cotopaxi, South America < 20,000 B.P.
RA 51 360,000,000 -- 16,000,000 Dobratsch (2) Alps
RA 52 350,000,000 -- 120,000,000 Egmont, (2) New Zealand 6,570 B.P.
RA 53 340,000,000 -- 15,000,000 Unzen (Mayu-yama) Japan 1792
RA 54 300,000,000 -- 21,000,000 Augustine, (1) Burr Point Alaska, USA 1883
RA 55 285,000,000 -- 1,900,000 Vaiont Alps 1963
RA 56 250,000,000 -- 5,200,000 Lago di Tovel Alps

V A B
RA 57 250,000,000 -- 800,000 Cayley, (2) Canada 4,800 B.P.
RA 58 230,000,000 -- 10,000,000 Myoko Japan 7,780 B.P.
RA 59 224,000,000 -- 4,560,000 Chaski Bay Crater Lake, Oregon, USA
RA 60 210,000,000 -- 13,200,000 Tschirgant Alps
RA 61 180,000,000 -- 4,000,000 Bormio Alps
RA 62 170,000,000 -- 7,000,000 Masiere di vedane Alps
RA 63 170,000,000 -- 8,000,000 Dobratsch (1) Alps
RA 64 165,000,000 -- 1,000,000 Lago di Poschiavo Alps
RA 65 150,000,000 -- 3,500,000 Lavini de Marco Alps
RA 66 150,000,000 -- 15,000,000 Abimes de Myans Alps
RA 67 150,000,000 -- 8,000,000 Chaos Crags USA 1650
RA 68 150,000,000 -- 15,000,000 Callaqui South America

RA 69 150,000,000 -- 10,000,000 Augustine (3) Alaska, USA ~1700
RA 70 140,000,000 -- 2,500,000 Obersee Alps

V A B
RA 71 140,000,000 -- 18,000,000 Papandayan Indonesia 1772
RA 72 120,000,000 -- 1,700,000 Oeschinensee Alps
RA 73 115,000,000 -- 1,840,000 Cal de la Madeleine Alps
RA 74 100,000,000 120,000 22,500,000 Huascaran Peru 1970
RA 75 100,000,000 -- 1,000,000 Am Saum Alps
RA 76 100,000,000 -- 5,000,000 Oberterzen Alps
RA 77 100,000,000 -- 2,400,000 Mallnitz Alps
RA 78 91,000,000 -- 9,300,000 Lastarria South America prehistoric
RA 79 86,000,000 -- 1,550,000 Medecine Lake Canadian Rocky Mountains
RA 80 80,000,000 -- 4,000,000 Pletzachkogel Alps
RA 81 80,000,000 -- 3,580,000 Lofer Alps
RA 82 68,000,000 -- 1,600,000 Lac Lauvitel Alps
RA 83 56,500,000 -- 2,700,000 Obernbergtal Alps
RA 84 56,000,000 12,000 6,750,000 Ontake Japan

V A B
RA 85 55,000,000 -- 8,520,000 Mt. Cook New Zealand 1991
RA 86 54,000,000 15,000 6,000,000 Mageik USA historic
RA 87 50,000,000 -- 2,000,000 Pontives Alps
RA 88 50,000,000 -- 2,300,000 Marquartstein Alps
RA 89 50,000,000 -- 2,200,000 Diablerets Alps 1714, 1749
RA 90 49,400,000 190,000 700,000 Madison Canyon Wyoming, USA 1959
RA 91 40,000,000 -- 1,000,000 Monte Avi Alps
RA 92 40,000,000 -- 1,000,000 Kleines Rinderhorn Alps
RA 93 39,100,000 -- 3,630,000 Mt. Kitchener Canadian Rocky Mountains
RA 94 38,000,000 -- 1,260,000 Danger Bay, west Crater Lake, Oregon, USA
RA 95 35,000,000 -- 4,000,000 Goldau Alps 1806
RA 96 34,000,000 -- 1,150,000 Danger Bay, east Crater Lake, Oregon, USA
RA 97 30,000,000 -- 700,000 Haslensee Alps

RA 98 30,000,000 -- 5,000,000 Dobratsch Alps

V A B
RA 99 30,000,000 -- 850,000 Voralpsee Alps
RA 100 30,000,000 -- 1,100,000 Torbole Alps
RA 101 30,000,000 13,000 3,000,000 Frank Slide Turtle Mountain, Alberta, April, 1903
Canada
RA 102 29,500,000 -- 1,700,000 Haiming Alps
RA 103 25,000,000 28,000 1,300,000 North Long John rock Inyo Mountains, Owens prehistoric
avalanche Valley, California
RA 104 25,000,000 -- 1,200,000 San Giovanni Alps
RA 105 21,000,000 -- 790,000 St Andre Alps
RA 106 20,000,000 -- 450,000 Mordbichl Alps
RA 107 20,000,000 -- 500,000 Lago di Alleghe Alps
RA 108 20,000,000 -- 1,000,000 Kals Alps
RA 109 17,500,000 -- 700,000 Monte Corno Alps
RA 110 16,000,000 -- 700,000 Brione Alps
RA 111 16,000,000 -- 410,000 Grand Clapier Alps
RA 112 15,000,000 -- 6,000,000 Huascaran Peru 1962

V A B
RA 113 15,000,000 -- 800,000 Disentis Alps
RA 114 15,000,000 -- 1,000,000 Biasca Alps
RA 115 14,000,000 -- 600,000 Ludiano Alps
RA 116 14,000,000 -- 720,000 Llao Bay Crater Lake, Oregon, USA
RA 117 13,000,000 -- 950,000 Hintersee Alps
RA 118 13,000,000 -- 580,000 Eagle Point, west Crater Lake, Oregon, USA
RA 119 12,100,000 -- 8,250,000 Sherman Glacier USA 1964
RA 120 12,000,000 -- 1,000,000 Lago de Antrona Alps
RA 121 11,000,000 8,000 5,000,000 Rainier Washington, USA 1963
RA 122 10,000,000 -- 2,200,000 Mount Munday Canada 1997
RA 123 10,000,000 -- 580,000 Elm Alps 1881
RA 124 8,500,000 -- 880,000 Oberes Vallesinella Alps
RA 125 8,000,000 -- 600,000 Val Brenta Alta Alps

RA 126 7,000,000 -- 720,000 Melkode Alps

Maximum Innundated Total Innundated
Flow Volume
Cross-sectional Area Planimetric Area
Ref # (cubic meters) Name of Event/Location Region/State/Country Year*
(square meters) (square meters)
V
A B
RA 127 6,000,000 -- 470,000 North Nahanni District of Mackenzie, NWT, Canada 1985
RA 128 6,000,000 -- 430,000 Tucketthutte Alps
RA 129 5,000,000 -- 200,000 Prayon Alps
RA 130 5,000,000 12,000 600,000 Cayley, (1) Canada 1963
RA 131 4,820,000 -- 300,000 Beaver Flats south Canadian Rocky Mountains
RA 132 4,130,000 -- 300,000 Beaver Flats north Canadian Rocky Mountains
RA 133 4,000,000 -- 400,000 Fionnay Alps
RA 134 4,000,000 -- 400,000 Adams, USA (1) Washington, USA 1921
RA 135 3,000,000 -- 250,000 Winkelmatten Alps
RA 136 3,000,000 -- 700,000 Adams, USA (2) 1997
RA 137 2,400,000 -- 580,000 Cloudcap Bay Crater Lake, Oregon
RA 138 1,820,000 1,000 550,000 Puget Peek USA 1964
RA 139 2,100,000 -- 137,000 Jonas Creek south Canadian Rocky Mountains
RA 140 4,500,000 -- 226,000 Jonas Creek north Canadian Rocky Mountains

V A B
RA 141 400,000 -- 200,000 Fidaz Alps
RA 142 150,000 -- 75,000 Haltenguet Alps
RA 143 100,000 -- 100,000 Felsberg Alps
* Dates are year A.D. unless specified as year B.P. (years before present).
Ref # indicates landslide type and key # for bibliography. DF = non-volcanic debris flow; RA = rock avalanche; LA = lahar.
Appendix B—Data Bibliography 49
Appendix B—Data Bibliography

Data References for Appendix A, Debris Flows
DF 1 Jordan, P., 1994, Debris flows in the southern Coast Mountains, British Columbia: dynamic behaviour and physical proper-
ties: University of B.C. PhD thesis, 260 p.
DF 2 Clague, J. J., and Evans, S. G., 1994, Formation and Failure of Natural Dams in the Canadian Cordillera: Geological Survey
of Canada, Geological Survey of Canada Bulletin 464.
Clague, J. J., Evans, S. G., and Blown, I. G., 1985, A debris flow triggered by the breaching of a moraine-dammed lake, Klat-
tasine Creek, British Columbia: Canadian Journal of Earth Sciences, v. 22, no. 10, p. 1492-1502.
DF 4 Wieczorek, G. F., Larsen, M. C., Eaton, L. S., Morgan, B. A., and Blair, J. L., 2001, Debris-flow and flooding hazards associ-
ated with the December 1999 storms in coastal Venezuela and strategies for mitigation: U.S. Geological Survey Open File
Report 01-0144.
DF 5 Crosta, G. B., 2001, Failure and flow development of a complex slide: the 1993 Sesa landslide: Engineering Geology, v. 59,
p. 173-199.
DF 6 Bovis, M. J., and Jakob, M., 2000, The July 29, 1998, debris flow and landslide dam at Capricorn Creek, Mount Meager
Volcanic Complex, southern Coast Mountains, British Columbia: Canadian Journal of Earth Sciences, v. 37, p. 1321-1334.
DF 7 Glancy, P. A., and Bell, J. W., 2000, Landslide-Induced Flooding at Ophir Creek, Washoe County, Western Nevada, May 30,
1983: U.S. Geological Survey Professional Paper 1617.
DF 9 Miller, H.F., 1979, Debris flows in the vicinity of Boulder, Colorado: Boulder, Colorado, University of Colorado M.S. The-
sis, 93 p.
DF 11 Pierson, T. C., 1980, Erosion and Deposition by Debris Flows at Mt Thomas, North Canterbury, New Zealand: Earth Surface
Processes, v. 5, p. 227-247.
DF 12 Janssen, K., and Marshall, C., 2003, unpublished report prepared for Fall 2002 Environmental Geology taught by Scott
Burns, Debris flow along
I-84.
DF 13 Miller, H.F., 1979, Debris flows in the vicinity of Boulder, Colorado: Boulder, Colorado, University of Colorado M.S. The-
sis, 93 p.
DF 14 DeGraff, J. V., 1997, Geologic Investigation of the Sourgrass Debris Flow, Calaveras Ranger District, Stanislaus National
Forest: U.S. Department of Agriculture; Forest Service, FS 6200-7.
Clague, J. J., and Evans, S. G., 1994, Formation and Failure of Natural Dams in the Canadian Cordillera: Geological Survey
DF 15 of Canada, Geological Survey of Canada Bulletin 464.
Jackson, E., Jr., Hungr, O., Gardner, J. S., and Mackay, C., 1989, Cathedral Mountain Debris Flows, Canada: Bulletin of the
International Association of Engineering Geology, v. 40, p. 36-54.
DF 16 Jakob,M., Anderson, D., Fuller,T., Hungr,O. and Ayotte, D., 2000. An Unusually Large Debris Flow at Hummingbird Creek,
Mara Lake, British Columbia. Canadian Geotechnical Journal Vol. 37, no. 5, p.1109-1125.
DF 20 Iverson, R.M., Schilling, S.P., Vallance, J.W., 1998, Objective delineation of lahar-inundation hazard zones: GSA Bulletin,
vol. 100, no. 8, p. 972-984.
DF 21 Gallino, G. L., and Pierson, T. C., 1985, Polallie Creek debris flow and subsequent dam-break flood of 1980, East Fork Hood
River basin, Oregon, U.S. Geological Survey Water-Supply Paper 2273.
DF 22 Jakob, M., Hungr, O., and Thomson, B., 1997. Two debris flows with anomalously high magnitude. Debris Flow Hazards
Mitigation, Mechanics, Prediction and Assessment. Procs., The First International Conference on Debris Flow Hazards
ASCE, C.L.Chen, Ed.,
DF 24 Thurber Consultants Ltd, 1985, Debris torrent assessment, Wahleach and Floods, Highway 1, Hope to Boston Bar Creek
summit, Coquihalla Highway. Report to B.C. Min. Transportation and Highways, Victoria, B.C.
DF 25 Jakob, M., Hungr, O., and Thomson, B., 1997. Two debris flows with anomalously high magnitude. Debris Flow Hazards
Mitigation, Mechanics, Prediction and Assessment. Procs., The First International Conference on Debris Flow Hazards
ASCE, C.L.Chen, Ed., San Francisco, pp. 382-394.
DF 32 Thurber Consultants Ltd and Ker Priestman & Associates Ltd, [no date], Site visit to debris torrent facilities at Charles
Creek, Harvey Creek, Magnesia Creek, Alberta Creek, Lions Bay, B.C. Field Guide.
DF 33 Thurber Consultants Ltd and Ker Priestman & Associates Ltd, [no date], Site visit to debris torrent facilities at Charles
Creek, Harvey Creek, Magnesia Creek, Alberta Creek, Lions Bay, B.C. Field Guide.
DF 34 Curry, R.R., 1966, Observations of alpine mudflows in the Tenmile Range, Colorado: Geological Society of America Bul-
letin, v. 77, p. 771-776.
DF 37 Stock, J.D., 2001, unpublished data in personal communication to R.M.Iverson, 1/16/2001, Berkeley, California.
Clague, J. J., and Evans, S. G., 1994, Formation and Failure of Natural Dams in the Canadian Cordillera: Geological Survey
DF 41 of Canada, Geological Survey of Canada Bulletin 464.
DF 42 Campbell, R. H., 1975, Soil Slips, Debris Flows, and Rainstorms in the Santa Monica Mountains and Vicinity, Southern
California, U.S. Geological Survey Professional Paper 851.
DF 47 Walter, T., unpublished data, 1983, Memorandum: Subj: Debris slide of December 16, 1982.
DF 48 Howard, T.R., Baldwin, J.E., and Donley, H.F., 1988, Landslides in Pacifica, California, caused by the storm, in Landslides,
floods, and marine effects of the storm January 3-5, 1982, in the San Francisco Bay Region, California: U.S. Geological
Survey Professional Paper 1434, p. 163-183.
DF 50 McDonnell, J.J., 1990, The effect of macropores on debris flow initiation: Quarterly Journal of Engineering Geology, vol. 23,
p. 325- 332.
DF 55 Owens, I.F., 1972, Morphological characteristics of alpine mudflows in the Nigel Pass area, in Slaymaker, O., and McPher-
son, H.J., eds., Mountain Geomorphology: Vancouver, Tantalus Research, p. 93-100.
DF 57 Bevin, K. and E.F. Woods, 1983, Catchment geomorphology and the dynamics of runoff contributing areas: Journal of Hy-
drology, vol. 65, p. 139-158.
DF 58 Campbell, R. H., 1975, Soil Slips, Debris Flows, and Rainstorms in the Santa Monica Mountains and Vicinity, Southern
California, U.S. Geological Survey Professional Paper 851.
DF 62 Cannon, S.H., 1989, An approach for estimating debris flow runout distance, in Proceedings Conference XX, International
Erosion Control Association, Vancouver, B.C.
DF 64 Major, J.J., 1996, Experimental studies of depositionby debris flows: Process, characteristics of deposits, and effects of pore-
fluid pressure [Ph.D. dissert.]: Seattle, University of Washington, 341 p.
Iverson, R.M., 1997, The physics of debris flows: Reviews in Geophysics, v. 35, p. 245-296.
Data References for Appendix A, Rock Avalanches

RA 1 Heim, A., 1921, Geologie der Schweiz. Band 2: die Schweizer Alpen 1. Halfte. Tauchniz Leipzig.
RA 2 Crandell, D.R., Miller, C.D., Glicken, H.X., Christiansen, R.L., and Newhall, C.G., 1984, Catastrophic debris avalanche
from ancestral Mount Shasta volcano, California: Geology, vol. 12, p. 143-146.
Crandell, D.R., 1989, Gigantic debris avalanche of Pleistocene age from ancestral Mount Shasta volcano, California, and
debris-avalanche hazard zonation: USGS Bulletin 1861.
RA 3 Siebert, L., Glicken, H., Tadahide, U., 1987, Volcanic hazards from Bezymianny- and Bandai- type eruptions: Bulletin of
Volcanology, vol. 49, p. 435-459.
RA 4 Siebert, L., Bronto, S., Supriatman, I., Mulyana, R., 1996, Massive debris avalanche from Raung Volcano, Eastern Java:
EOS, vol. 77, p. S291.
RA 6 Haller, M J, Mendia, J E, and Ostera, H A, 1991, Mapa preliminar de riesgo en la vertiente Argentina del Volcan Peteroa;
Preliminary map of risk on the Argentinian slopes of Peteroa Volcano in Proceedings for Congreso Geologico Chileno,
Vina del Mar, Chile, Aug. 5-9, 1991, vol.6, pp.355-358.
RA 8 Thiele, R., Moreno, H., Elgueta, S., Lahsen Azar, A., Rebolledo, S., Petit-Breuilh, M.E., 1998, Evolucion geologico-geomor-
fologica cuaternaria del tramo superior del valle del rio Laja; Quaternary geological-geomorphological evolution of the
uppermost course of the Rio Laja Valley: Revista Geologica de Chile, vol.25, no.2, pp.229-253.
RA 9 Guest, N.J., Leedal, G.P., 1956, The volcanic activity of Mount Meru: Records of the Geological Survey of Tanganyika
1953, Rec., v. 3, Geological Survey of Tanganyika, Tanzania, p. 40-7.
Stoopes, G.R., Sheridan, M.F., 1988, Giant debris avalanches from the Colima Volcanic Complex, Mexico- Implications for
long-runout landslides (>100 km) and hazard assessment: Geology, vol. 20, p. 299-302.
RA 13 Hall, M.L., Robin, C.R., Beate, B., Mothes, P., Monzier, M., 1999, Tungurahua Volcano, Ecuador- Structure, eruptive his-
tory, and hazards: Journal of Volcanology and Geothermal Research, vol. 91, p. 1-21.
RA 15 Omenge, J.M., Mosley, P.N., 1995, Evidence for catastrophic flank collapse of Mt. Kilimanjaro and implication for geo-
logical hazard assessment, in Proceedings of 10th Conference of the Geological Society of Africa; GSA 95 international
conference, Nairobi, Kenya, Oct. 9-13, 1995, Conference Programme, vol.10, pp.99-100.
RA 17 Cacho, L.G., Iez-Gil, J.L., Arana, V., 1994, A large volcanic debris avalanche in the Pliocene Roque Nublo Stratovolcano,
Gran Canaria, Canary Islands: Journal of Volcanology and Geothermal Research, vol. 63, p. 217-229.
RA 20 Siebe, C., Komorowski, J.C., Sheridan, M.F., 1992, Morphology and emplacement of an unusual debris avalanche deposit at
Jacotitlan volcano, Central Mexico: Bulletin of Volcanology, vol. 54, p.573-589.
RA 21 Arbenz, P., 1934, Helvetische Region, in ed. Schweiz Geol. Kommission, Geologischer Fuhrer der Schweiz, Wepf & Cie,
Basel.
Rowley, P.D., Kuntz, M.A., MacLeod, N.S., 1981, Pyroclastic-flow deposits: USGS Professional Paper 1250, p. 489-512.
Voight, B., Janda, R.J., Glicken, H., and Douglass, P.M., 1983, Nature and mechanics of Mount St. Helens rockslide-ava-
lanche of 18 May 1980: Geotechnique, vol. 33, p. 243-273.
RA 23 [Ampferer, 1939] compiled in Li Tianchi, 1983, A mathematical model for predicting the extent of a major rockfall:
Zeitschrift fur Geomorphologi N.F., Bd. 27, heft. 24, p. 472-482.
RA 24 [Lacger, 1903] compiled in Li Tianchi, 1983, A mathematical model for predicting the extent of a major rockfall: Zeitschrift
fur Geomorphologi N.F., Bd. 27, heft. 24, p. 472-482.
RA 26 Schuster, R.L., 2002. Usoi Landslide Dam and Lake Sarez, Pamir Mountains, Tajikistan in Proceedings for Annual Confer-
ence of Geological Society of America, Denver, Colorado, October 27-30, 2002.
Aguila, L., Newhall, C.G., Miller, C.D., and Listanco, E., unpublished document, Reconnaissance geology of a large debris
avalanche from Iriga Volcano, Philippines.
Schuster, R.L., and Crandell, D.R., 1984, Catastrophic debris avalanches from volcanoes in Proceedings of the Fourth An-
nual Conference on Landslides, Toronto.
RA 31 Kojan, E., and Hutchinson, J. N., 1978, Mayunmarca rockslide and debris flow, Peru, in Voight, B., ed., Rockslides and
Avalanches: Amsterdam, Netherlands (NLD), Elsevier Sci. Publ. Co., p. 315-361.
RA 32 Abele, G., 1974, Bergstürze in den Alpen: Wissenschaftliche Alpenvereinshefte 25: 230 S.
RA 35 [Turnau, 1906] compiled in Li Tianchi, 1983, A mathematical model for predicting the extent of a major rockfall: Zeitschrift
RA 36 Heim, A., 1932, Bergsturz und Menschenleben, Fretz & Wasmuth, Zurich.
Siebert, L., 1996, Hazards of Large Volcanic Debris Avalanches and Associated Eruptive Phenomena in Monitoring and
Mitigation of Volcano Hazards: Washington D.C., Springer-Verlag Berlin Heidelberg, p. 541-572.
RA 39 [Bornhauser, 1950] compiled in Li Tianchi, 1983, A mathematical model for predicting the extent of a major rockfall:
RA 41 [Schwinner, 1912] compiled in Li Tianchi, 1983, A mathematical model for predicting the extent of a major rockfall:
Siebert, L., Beget, J.E., Glicken, H., 1995, The 1883 and late-prehistoric eruptions of Augustine volcano, Alaska: Journal of
Volcanology and Geothermal Research, vol. 66, p. 367-395.
RA 45 Voight, B., and Pariseau, W.G., 1978, Rockslides and avalanches, an introduction in Rockslides and Avalanches, 1, B.
Voight, ed., Elsevier Publishing Company, p. 2-63.
RA 46 [Trevisan, 1941] compiled in Li Tianchi, 1983, A mathematical model for predicting the extent of a major rockfall:
RA 47 [Arbenz, 1934] compiled in Li Tianchi, 1983, A mathematical model for predicting the extent of a major rockfall: Zeitschrift
RA 51 Till, A., 1907, Das Naturereignis von 1348 und die Bergsturze des Dobratsh, Mitteilungen du K.K. Geographischen Gesell-
schaft in Wein 50 Wein.
RA 55 Muller, L., 1964, The rock slide in the Vaiont Valley: Felsmechanik und Ingenieur-geologie, vol. 2, p. 148-212.
RA 57 Evans, S.G., and Brooks, G.R., 1991, Prehistoric debris avalanches from Mount Cayley volcano, British Columbia: Cana-
dian Journal of Earth Science, vol. 28, no. 9, p.xxx.
RA 59 Ramsey, D.W., Robinson, J.E., Dartnell, P., Bacon, C.R., Gardner, J.V., Mayer, L.A., Buktenica, M.W., 2002, Crater Lake
revealed; Using GIS to visualize and analyze postcaldera volcanoes beneath Crater Lake, Oregon: U.S. Geological Survey
Geologic Investigations Series I-2790.
RA 60 Heuberger, H., 1968, Die Alpengletscher im Spat und Postglazial: Eine chronologische Ubersicht, Eiszeitalter und Gegen-
wart, vol. 19, p. 270-275 Ohringen (Hohenlohe’sche Buchhandl).
RA 61 [Turrer, 1962] compiled in Li Tianchi, 1983, A mathematical model for predicting the extent of a major rockfall: Zeitschrift
RA 62 [Bruckner, 1901] compiled in Li Tianchi, 1983, A mathematical model for predicting the extent of a major rockfall:
RA 63 [Aichinger, 1951] compiled in Li Tianchi, 1983, A mathematical model for predicting the extent of a major rockfall:
RA 66 Gignoux, M., and Barbier, R., 1955, Geologie des barrages et des amenagements hydrauliques: Masson et Cie, Paris, 343 pp.
RA 68 Naranjo, J.A., Young, S.R., Moreno, H., 1998, Mitigation of volcanic risk in the Biobio River basin, Chile in Proceedings for
Cities on volcanoes, International meeting on Cities on volcanoes, Rome and Naples, Italy, June 28-July 4, 1998.
RA 74 Plafker, G. and Erickssen, G.E., 1978, Nevados Huascaran Avalanches, Peru in Rockslides and Avalanches, 1, B. Voight, ed.,
Elsevier Publishing Company, p.277-314.
RA 77 [Hammer, 1927] compiled in Li Tianchi, 1983, A mathematical model for predicting the extent of a major rockfall:
RA 78 Naranjo, J.A., and Francis, P., High velocity debris avalanche at Lastarria volcano in the north Chilean Andes: Bulletin of
RA 80 Klebelsberg, R., 1935, Geologie von Tirol.- 872 S., 1 Kt., 11 Beil.; Berlin (Borntraeger)[Leidlmair, 1956] compiled in Li
Tianchi, 1983, A mathematical model for predicting the extent of a major rockfall: Zeitschrift fur Geomorphologi N.F.,
Bd. 27, heft. 24, p. 472-482.
RA 81 [Leidlmair, 1956] compiled in Li Tianchi, 1983, A mathematical model for predicting the extent of a major rockfall:
RA 83 [Pashinger, 1953] compiled in Li Tianchi, 1983, A mathematical model for predicting the extent of a major rockfall:
RA 84 Inokuchi, T., 1985, The Ontake Rock Slide and Debris Avalanche Caused by the Naganoken-Seibu Earthquake in Proceed-
ings IVth International Conference and Field Workshop on Landslides, Tokyo, 10 p.
Nagaoka, Masatoshi, 1987, Geomorphological Characteristics and Causal Factors of the 1984 Ontake Landslide Caused by
the Naganoken-Seibu Earthquake: Bulletin of the Geographical Survey Institute, vol. 31, p. 72-89.
RA 85 Hancox, G.T., Chin, T.J., and McSaveney, M.J., 1991, Immediate report, Mt. Cook Rock Avalanche, 14 December 1991,
Report by the New Zealand Department of Scientific and Industrial Research, 23 December 1991.
RA 87 Fuganti, A., and V.G. Antonio, 1969, Previsioni sul progettato traforo ferroviario del Brennero in base studio geologico del
tracciato: Convegno Int. sui problemi tecnini nella costruzione di gallerie, Torino, 1969.
RA 88 [Broili, 1914] compiled in Li Tianchi, 1983, A mathematical model for predicting the extent of a major rockfall: Zeitschrift
RA 90 Hadley, J.B., 1978, Madison Canyon Rockslide, Montana, USA in Rockslides and Avalanches, 1, B. Voight, ed., Elsevier
Publishing Company, p. 167-180.
RA 91 [Grasso, 1968] compiled in Li Tianchi, 1983, A mathematical model for predicting the extent of a major rockfall: Zeitschrift
RA 92 Furrer, E., 1962, Der Bergsturz von Bormio: Vierteljahrsschrift der Naturforschenden Gesellschaft in Zürich, vol. 107, p.
233-242.
RA 98 Till, A., 1907, Das Naturereignis von 1348 und die Bergsturze des Dobratsh, Mitteilungen du K.K. Geographischen Gesell-
schaft in Wein 50 Wein.
RA 100 [Penck, 1901] compiled in Li Tianchi, 1983, A mathematical model for predicting the extent of a major rockfall: Zeitschrift
RA 101 Cruden, D.M. and O. Hungr, 1986, The debris of Frank Slide and theories of rockslide-avalanche mobility: Can. J. Earth
Science, vol. 23, 425-432.
RA 102 Heuberger, H., 1966, Gletschergeschichtliche Untersuchungen in den Zentralalpen zwischen Sellrain und Ötztal: Wissen-
schaftliche Alpenvereinshefte, heft 20, 126 pp.
RA 103 Blair, T.C., 1999, Form facies and depositional history of the North Long John rock avalanche, Owens Valley, California:
Canadian Journal of Earth Science, vol. 36, p. 855-870.
RA 104 Montandon, F., 1933, Chronologie des grands eboulements alpins, du debut de l’ere chretienne a nos jours, in Materiaux
pour l’Etude des Calamites, Societe de Geographie Geneve, vol. 32, p. 271-340.
RA 105 [Onde, 1938] compiled in Li Tianchi, 1983, A mathematical model for predicting the extent of a major rockfall: Zeitschrift
RA 106 [Klebelsberg, 1935] compiled in Li Tianchi, 1983, A mathematical model for predicting the extent of a major rockfall:
RA 108 [Cornelius, 1936] compiled in Li Tianchi, 1983, A mathematical model for predicting the extent of a major rockfall:
RA 110 [Cadisch, 1953] compiled in Li Tianchi, 1983, A mathematical model for predicting the extent of a major rockfall: Zeitschrift
RA 111 [Bourdier, 1961] compiled in Li Tianchi, 1983, A mathematical model for predicting the extent of a major rockfall:
RA 112 Plafker, G. and Erickssen, G.E., 1978, Nevados Huascaran Avalanches, Peru in Rockslides and Avalanches, 1, B. Voight, ed.,
Elsevier Publishing Company, p.277-314.
RA 113 [Jackli, 1957] compiled in Li Tianchi, 1983, A mathematical model for predicting the extent of a major rockfall: Zeitschrift
RA 117 [Gillitzer, 1912] compiled in Li Tianchi, 1983, A mathematical model for predicting the extent of a major rockfall: Zeitschrift
RA 119 McSaveney, M.J., Sherman Glacier Rock Avalanche, Alaska, USA in Rockslides and Avalanches, 1, B. Voight, ed., Elsevier
Publishing Company, p. 197-258.
RA 121 Crandell, D.W, and Fahnestock, R.K, 1965, Rockfalls and avalanches from Little Tahoma Peak on Mount Rainier Washing-
ton: Geological Survey Bulletin 1221-A, Washington, 30 p.
RA 122 Evans, S.G., and Clague, J.J., 1998, Rock avalanche from Mount Munday, Waddington Range, British Columbia, Canada:
Landslide News, no. 11, p. 23-25.
RA 126 [Schmit-Thome, 1960] compiled in Li Tianchi, 1983, A mathematical model for predicting the extent of a major rockfall:
RA 127 Evans, S.G., Aitken, J.D., Wetmiller, R.J., and Horner, R.B., 1987, A rock avalanche triggered by the October 1985 North
Nahanni earthquake, District of Mackenzie, N.W.T.: Canadian Journal of Earth Science, vol. 24, p. 179-184.
RA 130 Clague, J.J., and Souther, J.G., 1983, The Dusty Creek landslide on Mount Cayley, British Columbia: Canadian Journal of
Earth Science, vol. 19, p. 524-539.
RA 134 Vallance, J.W., 1999, Postglacial lahars and potential hazards in the White Salmon River system on the southwest flank of
Mount Adams, Washington: USGS Bulletin No. 2161.
RA 136 Iverson, R.M., written communication, August 1997: fieldnotes.
RA 138 Hoyer, M., 1971, Puget Peak Avalanche, Alaska: Geological Society of America Bulletin, vol. 82, p. 1267-1284.
RA 139 Jakob, M., 2008, unpublished data in personal communication to J.W. Vallance, 20 June 2008, Vancouver, British
Columbia, Canada.
RA 140 Jakob, M., 2008, unpublished data in personal communication to J.W. Vallance, 20 June 2008, Vancouver, British
Columbia, Canada.
RA 141 [Niederer, 1941] compiled in Li Tianchi, 1983, A mathematical model for predicting the extent of a major rockfall:
Griswold—Mobility Statistics and Automated Hazard Mapping for Debris Flows and Rock Avalanches—Scientific Investigations Report 2007–5276

2008 Griswold

Uploaded by

Document Informationclick to expand document informationMobility Statistics and Automated Hazard Mapping for Debris Flows and Rock Avalanches

Document Informationclick to expand document information

Copyright:

Available Formats

2008 Griswold

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

2008 Griswold

Uploaded by

Copyright:

Available Formats

Mobility Statistics and Automated Hazard Mapping for

Debris Flows and Rock Avalanches

Scientific Investigations Report 2007–5276

U.S. Department of the Interior

By Julia P. Griswold and Richard M. Iverson

Scientific Investigations Report 2007–5276

U.S. Department of the Interior

U.S. Geological Survey

U.S. Geological Survey, Reston, Virginia: 2008

This report and any updates to it are available at:

For product and ordering information:

natural hazards, and the environment:

Manuscript approved for publication, December 12, 2007

By Julia P. Griswold and Richard M. Iverson

Figure 1. Schematic inundation limits of a debris flow moving down-

topography downslope or downstream from the lahar source

Planimetric area, B (m2)

CROSS-SECTIONAL AREA, A, IN SQUARE METERS

FLOW VOLUME, V, IN CUBIC METERS FLOW VOLUME, V, IN CUBIC METERS

Null with 2/3 slope log A = log α 1 + 2 log V n-1

Alternate with 2/3 slope log A = log α1 + 2 logV + β Z Z n-2

Alternate with best-fit slope log A = log α 1 + β 1 log V + β Z Z n-3

PLANIMETRIC AREA, B (m2)

FLOW VOLUME, V (m3) FLOW VOLUME, V (m3)

PLANIMETRIC AREA, B (m2)

PLANIMETRIC AREA, B (m2)

FLOW VOLUME, V (m3) FLOW VOLUME, V (m3)

PLANIMETRIC AREA, B, IN SQUARE METERS

lahar data lahar data

FLOW VOLUME, V, IN CUBIC METERS FLOW VOLUME, V, IN CUBIC METERS

PLANIMETRIC AREA, B, IN SQUARE METERS

debris-flow data debris-flow data

FLOW VOLUME, V, IN CUBIC METERS FLOW VOLUME, V, IN CUBIC METERS

PLANIMETRIC AREA, B, IN SQUARE METERS

rock-avalanche data rock-avalanche data

FLOW VOLUME, V, IN CUBIC METERS FLOW VOLUME, V, IN CUBIC METERS

As an example of application of our inundation-area

Conclusions those in the historical or geologic record. A limitation of this

Wesolowsky, G.O., 1976, Multiple linear regression and

DF 2 3,000,000 7,000 -- Klattasine Creek southern Rockies, British 1971-1973

DF 9 200,300 -- 72800 Shadow Canyon vicinity of Boulder,

DF 22 63,000 77 -- Pierce Creek southern British Columbia November 28, 1995

DF 23 60,000 -- 35,000 Hot Springs Cr. British Columbia, Canada 1984

DF 24 55,000 48 -- Wahleach A British Columbia, Canada

DF 25 50,000 80 -- Hope Creek southern Rockies, British November 8, 1995

DF 27 25,000 50 18,000 Boundary Cr. British Columbia, Canada 1989

DF 28 24,000 -- 12,000 Cathedral Mountain southern Rockies, B.C. 1962

DF 30 20,000 70 7,000 Lower Ryan R. tributary British Columbia, Canada 1984

DF 31 20,000 400 -- Mt. Currie gully British Columbia, Canada 1989

DF 32 20,000 48 -- M-Creek British Columbia, Canada

DF 33 20,000 47 -- Charles Creek British Columbia, Canada

DF 35 10,000 -- 5,000 No Good Cr. British Columbia, Canada 1990

DF 36 10,000 55 8,000 Canyon Cr. British Columbia, Canada 1987

DF 37 9,990 50.4 2,180 Black Cr. Olympics, Washington

DF 38 7,000 -- 5,000 Cathedral Mountain southern Rockies, British 1982

DF 40 5,790 19.4 1,540 Marlow#1 Coast Range, Oregon, USA

DF 41 5,000 -- 5,000 Cathedral Mountain southern Rockies, British 1984