Copyright q 1997,SEPM (Society for Sedimentary Geology) 1073-130X/97/067-776/$03.00
FIRE-RELATED SEDIMENTATION EVENTS ON ALLUVIAL FANS,
YELLOWSTONE NATIONAL PARK,U.S.A.
Department of Geology,Middlebury College,Middlebury,Vermont 05753,U.S.A.
Quaternary Studies Center,Desert Research Institute,PO Box 60220,Reno,Nevada 89506,U.S.A.
:We document initiation and ¯ow processes,deposit facies,
and geomorphic effects of forest-®re-related sedimentation on small
alluvial fans in Yellowstone National Park.Brief,intense convective-
storm precipitation on steep basins burned in the 1988 ®res produced
sedimentation events involving a variety of depositional processes on
fans.Over the course of all documented events,¯ows on fans pro-
gressed from higher to lower sediment concentration.Events were of-
ten dominated by either debris ¯ows or relatively sediment-poor
stream¯ow processes;in some events,however,¯ows ranging a over
wide spectrum of sediment concentration produced signi®cant fan de-
posits.Debris ¯ows were generated by progressive sediment bulking
involving pervasive surface runoff and rill erosion on steep upper basin
slopes,followed by deep incision as ¯ows progressed down channels.
Debris-¯ow deposits show a marked decline in thickness and coarse
gravel content downfan,often with extensive distal gravel-poor facies.
We recognized a relatively minor percentage of noncohesive debris-
¯ow and hyperconcentrated-¯ow facies,with sorting and strati®cation
intermediate between muddy debris-¯ow and stream¯ow facies;these
were deposited where dilute ¯ows bulked with coarse sediment by
eroding channel alluviumor earlier deposits of the event.Belowincised
fan channels,stream¯ows expanded as sheet¯oods,which prograded
lower fans with distally ®ning deposits.Basins
duced stream¯ow events on fans,but sediment texture and availability
on slopes and in channels are primary factors determining ¯ow pro-
cesses on fans of smaller basins.Burned soil surfaces provided abun-
dant silt and clay for debris-¯ow generation,but because soil surface
sediment was stripped and/or compacted over time,the lack of avail-
able ®nes resulted in dominance of stream¯ow processes in later events.
Deposition on many alluvial fans is dominated by infrequent,catastroph-
ic events,making direct observations dif®cult (e.g.,Beaty 1963;Blair and
McPherson 1994).Relatively few investigations have been made of the
processes and deposits of modern fan sedimentation events (e.g.,Blair
1987;Wells and Harvey 1987).In Yellowstone National Park,devegetation
by the major forest ®res of 1988 promoted a number of debris-¯ow and
¯ood events on alluvial fans,providing an excellent opportunity to study
sedimentation processes.Previous studies have highlighted the importance
of ®res in generating runoff and sediment transport in mountain drainage
basins (e.g.,White and Wells 1979;Swanson 1981;Wells 1987;Florsheim
et al.1991;Meyer et al.1992;Meyer et al.1995).In this paper,we doc-
ument and interpret the generative factors,¯ow processes,and fan deposits
of three ®re-related sedimentation events in northeastern Yellowstone:the
``Twelve Kilometer'',``Slough Creek Ranger Station''(Slough R.S.),and
``Frenchy's Meadow''events (Fig.1,Table 1).We also examined several
other ®re-related events in basins of varying lithologic and geomorphic
character in order to interpret intrinsic controls on fan sedimentation.
The three main ®re-related sedimentation events studied occurred in the
Yellowstone backcountry of the Slough Creek drainage,a broad glacial
trough valley (Fig.1).The valley walls have up to 1200 m of relief and
are composed mainly of friable``alluvial facies''volcaniclastic rocks of
the Eocene Absaroka Volcanic Supergroup (Smedes and Prostka 1972;
Prostka et al.1975).These laharic deposits (cf.Smith and Lowe 1991)
contain abundant ®ne matrix and yield silt-rich weathering products (Miller
and Drever 1977).Colluvium is relatively thin over the glacially scoured
slopes (typically,0.3 m).Soil A horizons are usually silt loam,with ®nes
of probable eolian and slopewash origin.Lateral moraines and kame de-
posits lie along lower slopes (Pierce 1974) and are minor sediment sources
in the study basins.Tributaries have been building small fans with radial
lengths of 0.4±1.3 km along the valley sides since ice retreat;14±12 ka
(Pierce 1979;Meyer et al.1995).Intense ®res during the extreme 1988
summer drought burned a large percentage of the mixed-conifer forest can-
opy in the Slough Creek basin above Elk Tongue Creek (Balling et al.
1992;Romme and Despain 1989) (Fig.1).Tributary basins that produced
fan sedimentation had
,10% vegetative cover over their 20±408 head-
We mapped deposits on the Twelve Kilometer fan using a 1:6000-scale
air photo combined with compass and tape methods.Total station surveying
accurate to;1 cm in x±y±z coordinates was used to map the Slough R.S.
and Frenchy's Meadow deposits,construct topographic pro®les,and derive
fan slopes.Topographic maps (1:24,000) were used to estimate slopes on
the Twelve Kilometer fan and measure morphometric parameters of basins.
Deposit texture was described in the ®eld using sieve and visual estimation
of the coarse-fraction percentage,and by soil texture classes for the
mm matrix,later converted to terminology of Folk (1974).Particle size
analyses were conducted on the pebble and ®ner (
,16 mm;,24 f)
fraction using standard wet-sieving and hydrometer methods (Day 1965).
To investigate ¯ow initiation processes,erosional and depositional fea-
tures were mapped throughout the drainage basin of the Twelve Kilometer
event a few days after its occurrence,and a volumetric sediment budget
was constructed using tape and topographic map measurements.Basin
headslopes were subdivided into areas of similar rill depth,and cross-sec-
tional areas of rills measured along representative transects were multiplied
by slope length (Meyer 1993).For the volume of channel erosion,the area
of each channel was averaged from several cross sections and multiplied
by channel length.To calculate fan deposit volume,an isopach map was
constructed using;50 thickness measurements.
Facies Classi®cation.ÐFacies were classi®ed according to ¯ow process
as debris-¯ow (DF),hyperconcentrated-¯ow (HCF),or stream¯ow (S) fa-
cies (Pierson and Costa 1987).Flow process interpretations were based on
the sedimentology and morphology of related deposits as de®ned by pre-
vious studies (summarized brie¯y below).Debris ¯ows are non-Newtonian
slurries with substantial yield strength,and consist of high concentrations
of sediment in water ¯owing as a single phase.Deposit characteristics
include sharp margins,lack of internal strati®cation,very poor sorting,and
except where it has drained or been washed away,a ®ner-grained matrix
®lling interstices (e.g.,Bull 1964a;Costa 1984,1988;Blair and McPherson
1994).Hyperconcentrated ¯ows have lower sediment concentrations and
are non-Newtonian liquids with a small yield strength (Beverage and Cul-
bertson 1964;Pierson and Costa 1987).Only sand-dominated deposits re-
sulting from actual measured hyperconcentrated ¯ow have been described
(Pierson and Scott 1985);coarse-grained deposits are generally inferred to
be intermediate in sorting and strati®cation between debris-¯ow and
777FIRE-RELATED SEDIMENTATION ON ALLUVIAL FANS
.1.ÐTopographic maps of the three main study areas for ®re-related sedimentation events in the Slough Creek drainage of northeastern Yellowstone (c ontour interval
40 feet),showing the A) Twelve Kilometer,B) Slough R.S.,and C) Frenchy's Meadow alluvial fans and drainage basins.Dark pattern shows primary areas of fan deposits
in the events;facies abbreviations as in Table 1.Pro®le lines indicate positions of radial fan pro®les in Figures 6B and 9B.D) Index map showing location of the main
study areas,and sites of the Cutoff Creek,West Fork Lodgepole Creek,and Gibbon Canyon fan sedimentation events.
778 G.A.MEYER AND S.G.WELLS
1.Ð Morphometry of basins of ®re-related fan sedimentation events.
BASIN (Dominant Facies)
Twelve Kilometer (DF)
Frenchy's Meadow (S)
Cutoff Creek (S)
West Fork Lodgepole Cr.(S)
La Familia (DF,HCF)
West Slickrock (DF)
Dominant deposit facies types of initial events in basins:DF 5 debris ¯ow,HCF 5 hyperconcentrated
¯ow,S 5 stream¯ow.
Relief from highest point on basin divide to lowest point on fan.
Maximum relief from basin mouth to divide,divided by length of longest stream channel extended to
2.Ð Facies descriptions for ®re-related alluvial fan deposits in Yellowstone.
Facies Morphology Thickness (m) Texture and Composition Fabric and Structure
Levees Curvilinear ridges 0.5±1.5 Poorly sorted
Boulders and woody debris,variable matrix
Usually matrix-poor;muddy coating on clasts
Random clast orientation to crude,steep,up-
Boulder lobes Lobate to short digitate
Narrow to moderate width
Moderate to steep margins
Sharp margins (except where transi-
tional to gravel-poor facies)
Low- to high-relief surface (low where
0.3±1.51 Very poorly sorted (s
boulders;matrix-rich except where drained
from margins or upper deposit
Matrix is coarser,less muddy in noncohesive
Charcoal- and organic-rich
Matrix- to clast-supported
Vertical clasts common
May show crude coarse-tail grading,otherwise
Gravel-poor lobes Broad to digitate lobes
Moderate-angle,sharp discrete margins
Very poorly sorted (s
5 3±4f) mud to
Boulders and cobbles absent to minor (some
transitional to above facies)
Very charcoal- and organic-rich
Coarse-tail normal grading sometimes present
Local organic debris concentration zones near
margins and surface
Boulder and cobble splays Splays;lobate planform is common
Moderate to broad width
Margins clear;low-angle except where
0.3±1.0 Poorly sorted (s
Matrix-poor in upper zones
Coarse surface lags common
Crude surface-parallel strati®cation
May have surface-parallel clast orientation
May be normally graded
Sandy deposits Splays,sheets
0.03±0.3 Poorly sorted (s
Structureless to crude surface-parallel strati®-
Boulder bars and splays Narrow curvilinear bars with steep mar-
Small splays with high-relief surfaces
0.3±1.2 Moderate to well-sorted
Cobble bars and splays Splays with low-moderate relief sur-
0.3±0.8 Moderate to well-sorted
Minor sandy matrix
Upper matrix-free zone
Sheet¯ood deposits Broad fan-like splays and sheets;elon-
gate where partly con®ned
Clear to diffuse margins
0.1±0.3 Mod.to well-sorted (s
Cobble to sand-dominated beds with moderate
sorting within beds
Minor sandy matrix
0.5±20 cm beds,may be normally graded or in
Surface-parallel planar strati®cation to low-an-
gle scour-and-®ll structure
Channel lags Irregular,high-relief surface with local
bars and scours
0.3±1.0 Minor poorly sorted matrix
Poorly sorted cobbles to large boulders
Local weak imbrication
Graphical standard deviation;Folk 1974) for the,24ffraction (Meyer 1993).
f 2 f
stream¯ow deposits (Smith 1986;Wells and Harvey 1987;Scott 1988;
Costa 1988).Bull (1964a) recognized gravelly,clast-supported``transition-
al''fan deposits with textures intermediate between debris-¯owand stream-
¯ow facies and inde®nite,low-relief deposit margins.
Stream¯ows are fully turbulent Newtonian ¯ows with relatively low sed-
iment concentrations (Pierson and Costa 1987).Stream¯ows that are sig-
ni®cant in fan sedimentation in Yellowstone are typically ¯ash-¯ood dis-
charges,either con®ned within incised channels or expanded over fan sur-
faces as sheet¯oods (Hogg 1982).Sheet¯oods on fans are relatively shallow
and broad,typically supercritical ¯ows (Blair 1987;Blair and McPherson
1994).Stream¯ow (or``water-laid'') deposits on fans share a relatively
well-sorted and clast-supported character (e.g.,Bull 1964a;Wells and Har-
vey 1987;Costa 1988).Facies characteristics for ®re-related events in
Slough Creek are summarized in Table 2.
RELATED SEDIMENTATION EVENTS
Twelve Kilometer Event:Debris-Flow-Dominated Deposition
Basin Characteristics and Flow Generation.ÐThe Twelve Kilometer
fan is fed by a 1.6 km
basin that was almost entirely burned in 1988 (Fig.
1A,Table 1).A major sedimentation event occurred on 9 July 1989 in
response to a brief,intense convective storm (Meyer 1993).Stripping of
wood ash and ®ne sediment from the soil surface provided evidence for
overland ¯ow on most of the 25±35 8 basin headslopes.Broad swaths where
;1 cm of ash and ®ne sediment were removed on low-angle crestslopes
narrowed into 1±12 cm deep rills on the main slopes.Colluvium in slope
hollows and alluvium in low-order channels is thin (,1 m),and was not
evacuated by en masse failure (cf.Reneau and Dietrich 1987).Instead,
where rills converged in these areas,deep,narrow gullies were incised,and
in lower reaches the stored sediment was commonly stripped to underlying
bedrock.We interpret these erosional features as the result of pervasive
779FIRE-RELATED SEDIMENTATION ON ALLUVIAL FANS
.2.ÐFacies map of the Twelve Kilometer event deposits.Boundaries between debris-¯ow facies are locally gradational.
generation of surface runoff,which caused shallow erosion of burned soil
surfaces by sheetwash and rill¯ow,and deeper incision in hollows and
channels by concentrated ¯ow.Drop tests,however,showed little water
repellency in soils,a condition that limits in®ltration and generates surface
runoff in some burned areas (cf.Wells 1987).
Channels in the upper basin did not show evidence of debris-¯ow pas-
sage (e.g.,levees or mud coatings;Costa 1988),and deposits along the
main channel were mostly moderately sorted gravels with little matrix (Fig.
1A).Incision of the main channel increased below,however,forming a
steep-walled slot up to 2 m deep and 4 m wide that was ¯oored by a
bouldery,matrix-poor lag along most of its length.Debris-¯ow levees and
mud coatings on trees and channel margins were ®rst noted in the middle
channel reach and were common in the 500 m reach above the fan.The
Twelve Kilometer fan has an unincised fanhead channel.The proximal fan
is forested,but only patches were burned in 1988.
Debris-Flow Deposits.ÐA large bouldery lobe up to 165 cm in thick-
ness (average;50 cm) was deposited on the 7±98 upper fan.This deposit
;60±65% cobbles and boulders enclosed in a muddy sand matrix,
and has moderate-relief margins except where buttressed by log jams (Figs.
2,3A).The downfan part of the lobe is mostly matrix-supported.Flow
lineations showed that matrix drained from the surface and margins (Fig.
3A),and thick matrix accumulations in low areas on top of the lobe suggest
ponding.Flow thinning was indicated by mud coatings up to 44 cm above
the deposit surface on the down¯ow side of trees.The upfan part of the
lobe is more matrix-poor and clast-supported.
We interpret this lobe as the depositional expression of a bouldery de-
bris-¯ow head,where coarse clasts are typically concentrated due to friction
and particle interlocking (Pierson 1986).Where the debris-¯ow head be-
came uncon®ned at the fan apex,¯ow expansion and thinning,settling of
coarse clasts,and out¯ow of matrix led to greater frictional interactions
and deposition of the boulder lobe (Table 2).Settling of coarse clasts in a
relatively low-strength slurry is suggested by local concentration of boul-
ders in the lower part of the deposit.Matrix was washed from the upfan
part of the lobe by later,more dilute ¯ow in the event.The ten largest
clasts in the main debris-¯ow lobe average 93
361 cm (a 3b axes),and
the ten largest clasts in debris-¯ow levees just above the fan apex average
3 61 cm.Because larger boulders remained in the feeder channel,
these clasts indicate ¯ow competence at the fanhead (Hampton 1975;Pier-
780 G.A.MEYER AND S.G.WELLS
.3.ÐDebris-¯ow facies of the Twelve Kilometer event;locations are shown in Figure 2.A) Lower part of main bouldery debris-¯ow lobe.Note gravel-poor muddy
matrix below log jam that drained from the main boulder lobe.B) View across gravel-poor facies (foreground and right margin of deposit),transitional boulder- and
cobble-bearing facies (center),and main bouldery lobe (background).C) Medial gravel-poor debris-¯ow lobe showing sharp,steep margin over dark burned soil surface
(foreground),marginal woody debris concentrations,and desiccation cracking.
son 1981).Meter-size boulders are much more common in levees than in
the main lobe,thus the largest clasts may have been moved toward ¯ow
margins,as observed by Pierson (1980).
Deposits of very poorly sorted pebbly muddy sand,most with,25%
gravel,were emplaced around and below the boulder lobe margin (Figs.
3B±C).Transitional facies with scattered cobbles and small boulders extend
from the lower boulder lobe (Fig.3B).These unstrati®ed deposits thin
distally,from 20 cm to as little as 1±2 cm in the fronts of digitate lobes.
Sharp margins showed no evidence of dewatering.Intact charred litter lies
beneath the deposits,indicating nonerosive ¯ow.They contain abundant
charcoal,wood,and other organic matter,especially near the surface and
lobe margins (Fig.3C).The,24 ffractions of the debris-¯ow boulder
lobe and these deposits are very similar (Fig.4).These characteristics in-
dicate a debris-¯ow origin for the gravel-poor lobes (Table 2).
Because the volume of gravel-poor debris-¯owfacies is about 50%great-
er than the volume of bouldery facies (Fig.5),it is unlikely that gravel-
poor facies were derived entirely by draining of matrix from the bouldery
lobe.Direct observations of moving debris ¯ows indicate that matrix com-
monly ¯ows out of the bouldery front,and that the gravel-poor tail of the
¯ow often overtakes the more slowly moving front,forming highly mobile
precursor surges (Costa and Williams 1984;Johnson 1984;Pierson 1986).
The wide deposits of gravel-poor facies alongside the boulder lobe also
suggest that the tail of the ¯ow was diverted around the lobe.
The height of debris-¯ow runup on trees as shown by mud coatings was
used to estimate ¯ow velocity (Chow 1959):
v 5 (2gh)
where v is average velocity,g is gravitational acceleration,and h is the
vertical distance of runup.This equation is strictly valid only for ¯ows with
no strength,and thus gives minimum estimates of debris-¯ow velocity
(Gallino and Pierson 1985).Velocities of 2.1±3.7 m/s were estimated in
the area of the lower boulder lobe and transitional gravel-poor facies,where
781FIRE-RELATED SEDIMENTATION ON ALLUVIAL FANS
.4.ÐCumulative curves of particle size
distribution (,24 ffraction) for the Twelve
Kilometer,Slough R.S.,and Frenchy's Meadow
debris-¯ow deposits.Analyses encompass the
entire size range of gravel-poor lobes.
.5.ÐSediment budget for the Twelve
Kilometer event (HCF 5 hyperconcentrated-¯ow
facies;S 5 stream¯ow facies).Sediment is
accounted by volume for erosional source areas
and depositional facies.
the fan gradient is;78.Maximum velocity and runup probably occurred
as the more mobile gravel-poor ¯ows passed the trees.
Gravel-poor debris ¯ow samples were reconstituted by adding water until
the mixture ¯owed as a slurry and retained margins with relief similar to
those observed in the ®eld.The resulting mixtures contained 64±69% sed-
iment by weight,slightly lower than the 70±90% typically observed in
debris ¯ows (Costa 1984,1988),and had a density of;1.54±1.57 g/cm
Properties of the gravel-poor deposits,including abundant low-density or-
ganic matter,imply that debris-¯ow conditions would have been reached
at a relatively low weight concentration.A very ®nes-rich texture (24%silt
and 9% clay) relative to many debris ¯ows (Costa 1984) and smectite clays
from weathered Absaroka volcanic rocks (Miller and Drever 1977) en-
hanced yield strength in the gravel-poor ¯ows (Hampton 1972;Pierson and
Costa 1987).For comparison,Weirich (1989) measured peak sediment con-
centrations of 56±65% by weight in southern California ®re-related debris
¯ows,where the matrix contained 10±25% very ®ne sand to clay.
Hyperconcentrated-Flow and Stream¯ow Deposits.ÐSmall deposits
of boulder gravel with a coarse sandy matrix were found along the fanhead
channel and are inset within the head of the main debris-¯ow lobe,thus
represent later deposition (Fig.2).Internal strati®cation and grading is
crude or absent,but weak imbrication is present and sorting is better than
in debris-¯ow deposits,thus we tentatively interpret these sediments as
hyperconcentrated-¯ow deposits (Table 2).Small splays and planar-bedded
sheets of relatively well-sorted sandy cobble and pebble gravel along shal-
low channels were interpreted as stream¯ow facies.Inset and erosional
relations show that these gravels represent the last stage of deposition.
Stream¯ow gravels were probably derived in part by reworking of debris-
¯ow deposits upfan.Adjacent to stream¯ow channels,pebbly lags with
patchy,muddy matrix preserved only in lower interstices represent eroded
remnants of the gravel-poor debris-¯ow facies.By 1991,minor additional
stream¯ow gravels had been deposited along the main fan channel.
Sediment Budget and Summary.ÐA sediment budget for the Twelve
Kilometer event shows that the total volume of sediment eroded from the
basin (11,500 6 3300 m
) and the total deposit volume (10,200 6 1700
) balance within the estimated range of error (Fig.5;Meyer 1993).
Sediment mass balance is uncertain,however,because of changes in den-
sity between eroded and deposited material.The large volume of matrix-
size debris-¯ow material implies a proli®c source of ®ne sediment.A min-
imum of 30% of the total volume eroded consisted of silty soils and ash
removed from slopes.By comparison,much of the sediment entrained by
782 G.A.MEYER AND S.G.WELLS
.6.Ð A) Facies map of the Slough R.S.
event deposits.Cross-pro®le locations A±A9 and
B±B9 are shown for B),pro®les of the Slough
R.S.fan.Radial pro®le location is shown in
Figure 1B;locations of important facies
transitions are indicated.
783FIRE-RELATED SEDIMENTATION ON ALLUVIAL FANS
.7.ÐDeposits of the Slough R.S.event;location of photo is shown in Figure
6A.Boulder debris-¯ow lobe in foreground is overlapped by a large splay of cobbly
hyperconcentrated-¯ow to stream¯ow deposits (below person).The near margin of
the debris-¯ow lobe (lower right) was eroded by late-stage,channelized stream¯ow.
3.Ð Area of fan deposits by facies for the Twelve Kilometer,Slough R.S.,
and Frenchy's Meadow events.Percentage of the total deposit area for each
facies is shown in parentheses below.
Area of Facies Type
DF HCF S Total
(% of total)
(% of total)
(% of total)
DF debris ¯ow,HCF 5 hyperconcentrated ¯ow,S 5 stream¯ow.
main channel incision was sandy gravel alluvium.We infer that ®ne sed-
iment stripped fromburned soil surfaces by surface runoff was an important
factor in generation of debris-¯ow conditions.Deposition on the Twelve
Kilometer fan was strongly dominated by debris-¯ow processes,but pro-
gressed through minor,successively more dilute ¯ow stages;hyperconcen-
trated and stream¯ow deposits make up only 7%of the total deposit volume
(Fig.5).The most important geomorphic effect of the Twelve Kilometer
event was fan aggradation generally tapering from the proximal to distal
Slough R.S.Event:Multiple Flow Processes on a Channeled Fan
Basin Characteristics and Flow Generation.ÐThe``Slough Creek
Ranger Station''(Slough R.S.) event was produced by the same 1989 storm
that generated the Twelve Kilometer event.The steep 0.8 km
is underlain by Archean granitic gneiss and amphibolite and was intensely
burned in 1988 (Fig.1B,Table 1).Shallow rills were visible in 1991 aerial
reconnaissance,and the feeder channel was incised to bedrock above the
fanhead,but no en masse failures were seen.The 148±98 proximal fan is
broadly con®ned by glacial deposits and bedrock,and has the convex cross
pro®le characteristic of fans (Blair and McPherson 1994) (Fig.6A±B).The
fanhead channel is incised;1 m,and a number of smaller channels cut
the upper fan.The distal fan expands over nearly 90
8 and declines in slope
to;28 at the limit of deposition in the event.
Debris-Flow Deposits.ÐDeposits of the initial phase of this event have
a very poorly sorted muddy sand matrix,and display sharp margins locally
abutting log jams.Proximal boulder lobes contain clasts to
lobes bearing 50% or greater cobbles but no coarse boulders were found
in the midfan area.Thin (
,10 cm) pebbly muddy sand deposits extend
;100 m beyond the map area along the main northeast channel.These
deposits were identi®ed as debris-¯ow facies (Figs.6A±B,7;Table 2).
Clast-supported bouldery ridges;0.5±1.5 m high along proximal and
midfan channels have muddy sand matrix coating interior interstices and
steeper faces proximal to the channel,and are interpreted as debris-¯ow
levees (Fig.6A;Table 2).Levee-like deposits with lower relief and greater
matrix appear to be debris-¯ow lobes that were breached by later,more
dilute ¯ows in this event (cf.Wells and Harvey 1987).As in the Twelve
Kilometer event,maximum clast size in debris-¯ow facies decreases down-
fan.The Slough R.S.deposits were not as distinctly segregated into boulder
lobes and gravel-poor facies,however,and the latter constitute a much
smaller percentage of the debris-¯ow facies than in the Twelve Kilometer
event (Table 3).Debris-¯ow matrix is also poorer in ®nes and slightly
better sorted than in the Twelve Kilometer debris ¯ows,with an average
of 6% clay and 12% silt in the,24 fmm fraction (Fig.4).Overall,
debris-¯ow facies accounted for 55% of the mapped deposit area (Table
Hyperconcentrated-Flow and Stream¯ow Deposits.ÐPoorly sorted,
structureless to crudely strati®ed gravel splays were emplaced following
debris-¯ow deposition,as shown by stratigraphic relations and inset posi-
tions within breached and eroded debris-¯ow deposits (Fig.7).Matrix tex-
tures are poorly sorted coarse sands with little mud,and matrix is mostly
absent in the upper parts of deposits.Margins are readily de®nable but are
less sharp and have lower relief than debris-¯ow lobes.These deposits were
interpreted as hyperconcentrated-¯ow facies and constitute about one-
fourth of the deposit area (Fig.6A,Tables 2,3).They are largely restricted
to the upper fan,and generally ®ne downfan.
Bars and splays of moderately to well-sorted cobble gravels with little
matrix were emplaced on the upper fan (Fig.6A).Deposits with similar
sorting include planar-bedded pebble gravel sheets near channels,and thin,
discontinuous sand sheets on the distal fan north of the mapped area.These
deposits were interpreted as stream¯ow facies (Table 2).Inset and erosional
relationships indicate that stream¯ow deposits were emplaced following
hyperconcentrated-¯ow deposition,although some may have formed con-
currently with hyperconcentrated-¯ow deposition upfan.Stream¯ow boul-
der bars are rare,suggesting relatively low discharge and competence in
the stream¯ow phase;the upper main fan channel contains a matrix-poor
bouldery lag (Fig.6A,Table 2).
Summary.ÐThe Slough R.S.event was approximately evenly divided
between debris-¯ow deposition and the hyperconcentrated-¯owand stream-
¯ow deposition that followed,producing a wide range of facies from a
784 G.A.MEYER AND S.G.WELLS
.8.ÐDeposits of the Frenchy's Meadow event.A) Hyperconcentrated-¯ow gravels in the fan-like deposit above the main fan apex (see Figure 1C for location).Note
crude surface-parallel orientation of platy clasts (e.g.,near scale),upward-coarsening trend,and coarse,matrix-poor surface lag.B) High-relief stream¯ow boulder bar on
the fan-like deposit above the main fan apex (Fig.1C) showing sorting,lack of matrix,and strong imbrication.Boulder bars on the upper main fan (Fig.9A) are similar.
C) Noncohesive debris-¯ow lobe at the main fan apex (Fig.9A) with greater matrix than in hyperconcentrated-¯ow deposits in A (note lower left),vertic al variations in
matrix content,and variable clast orientations (note vertical clasts as well as local crude imbrication).Deposit is capped by;15 cm of stream¯ow sand and gravel.
Exposure was cut by later,more dilute ¯ow that breached the lobe and eroded below its base;knife point rests on pre-event soil surface.Note bouldery c hannel lag at
785FIRE-RELATED SEDIMENTATION ON ALLUVIAL FANS
bottom of photo.D) Intact sample on shovel blade of hyperconcentrated-¯ow sand facies (Fig.9A) and underlying;3 cm of darker soil;lines indicate surface and base
of deposit.Note very crude surface-parallel strati®cation and reverse grading.E) View downfan over cobble-dominated sheet¯ood deposits on the western fan (Fig.9A);
intensely burned Slough R.S.basin is in center background.F) Intact sample on shovel blade of thin sheet¯ood sand facies near the fan-to-¯oodplain transition (Fig.9A).
Deposit;6 cm thick has discontinuous lamina and shallow scour-and-®ll structures and overlies;2 cm of bioturbated soil.
single event.Incised fan channels allowed bouldery debris ¯ows to progress
farther downfan than if ¯ows had expanded (Whipple and Dunne 1992).
Frenchy's Meadow Event:Stream¯ow-Dominated Deposition
Basin Characteristics and Flow Generation.ÐThe Frenchy's Meadow
basin is slightly larger and less steep than the Twelve Kilometer basin (Fig.
1C,Table 1).In both basins,bedrock in headslope areas consists of an-
desitic volcaniclastic rocks,and the lower main channel is shallowly inset
in glacial deposits,with up to;2 m of gravelly alluvial ®ll.A major
sedimentation event occurred in late April or early May 1991,nearly three
years after the basin was burned in 1988.Snowmelt in the basin was largely
complete by late April (Lauryl Mack,NPS,personal communication 1991).
Aerial reconnaissance of the upper basin in July 1991 revealed a very
sparse cover of herbaceous plants,and little evidence of en masse failures
other than a few small,shallow debris slides.
Incision of the main channel was observed to begin in the central basin.
About 400 m above the main fan apex,a steep (;108) fan-like deposit
100 m long was constructed at a channel widening below a deeply incised
reach.This deposit is composed largely of coarse,poorly sorted and crudely
strati®ed gravels with surface-parallel clast orientations and a coarse sandy
matrix,capped by matrix-poor coarse gravel.High-relief bars of stacked
and imbricated boulders mantle the middle and upper fan (Fig.8A±B).
This sequence was interpreted to indicate rapid hyperconcentrated-¯owde-
position by expanding ¯ow,followed by high-energy stream¯ow bar de-
position and reworking of the deposit surface leaving a coarse lag.No
evidence of debris-¯ow activity was noted.Between this fan-like deposit
and the main fan apex,the feeder channel was incised as much as 2±3 m.
Gradients on the main Frenchy's Meadow fan range from about 8±6
the proximal area,to 48 on the middle fan,to a distal fan of about 1.5±
8 that merges with the Slough Creek ¯oodplain (Fig.9A±B).Prior to
the event,the fanhead channel was shallowly entrenched (,1 m) along
the west side of the fan.
Debris-Flow Deposits.ÐInitial deposition on the main fan involved em-
placement of a small,high-relief,clast-supported gravelly lobe abutting a
log-and-boulder jam at the fan apex (Fig.8C;site 1,Fig.9A).The lobe
thickens downfan to a maximum of about 1.5 m and contains boulders with
a axes up to 55 cm,with clasts.20 cm common.The matrix is composed
of slightly muddy coarse sand and granules (Fig.4) with little charcoal;
organic content is 3%.The,2 mm (21 f) fraction is similar to that of
the Slough R.S.debris-¯ow facies,but constitutes only 48% of the matrix
sample,whereas 62±75% of the Slough R.S.debris-¯ow matrix is,2
mm.Matrix ®lls only the bottom of some interstices,and is absent in others
except for discontinuous mud coatings.The basal contact shows little ero-
sion (Fig.8C).The vertical sequence includes a matrix-poor basal cobble
zone 5±10 cm thick with crude imbrication and moderate sorting,a matrix-
rich lower cobble zone,and a middle zone of locally imbricated and matrix-
poor ®ner cobbles.The upper zone is bouldery and matrix-rich,and the
deposit as a whole shows crude inverse grading.
Although this mud-poor deposit differs substantially from debris-¯ow
facies of the Twelve Kilometer and Slough R.S.events,the matrix content,
common vertical clast orientations,and steep margins of the lobe suggest
deposition by a ®nes-poor,noncohesive debris ¯ow (Blair 1987) (Figs.4,
9A;Table 2).Emplacement of the lobe where ¯ow became uncon®ned
immediately below the fan apex implies a high-strength ¯ow,and loss of
matrix may have increased its frictional strength.The vertical changes in
matrix content and imbricated zones suggest varying sediment concentra-
tion during deposition,perhaps near the hyperconcentrated-¯ow to debris-
¯ow transition.The lack of evidence for debris ¯ow in the fan-like deposit
above the main fan suggests that sediment bulking through deep incision
of the 400-m-long,8±98 intervening channel was responsible for producing
the noncohesive debris ¯ow.The amount of ®nes in available sediment is
important in determining ¯ow characteristics (e.g.,Blair 1987),and allu-
vium along the incised channel was mostly sand and gravel.
Hyperconcentrated-Flow and Stream¯ow Deposits.ÐThin sheets of
poorly sorted silty coarse sand with crude planar strati®cation were mapped
marginal to and downfan of the noncohesive debris-¯ow lobe (Fig.8D;
Fig.9A,site 2).These sandy deposits are better sorted and contain less
®nes and charcoal than gravel-poor debris-¯ow deposits of the Twelve
Kilometer or Slough R.S.events;because of their association with the lobe,
however,we tentatively interpret these deposits as hyperconcentrated-¯ow
facies formed by segregation of sandy matrix from the noncohesive debris
¯ow (Fig.10;Table 2).
The fan-apex debris-¯ow lobe was breached by subsequent ¯ow (Fig.
8C) that split into distributary branches northeast of the fan axis and along
the western fan margin (Fig.9A).Deposits along these channels include
high-relief curvilinear bars constructed of steeply imbricated openwork
boulders (cf.Fig.8B),especially in the west fan margin channel below the
incised proximal reach.Cobble bars and splays with moderate sorting and
variable imbrication merge downfan with broad gravelly sheets where ¯ow
became uncon®ned (Fig.8E).Small ®lls of moderately sorted to well-
sorted,planar-bedded sandy gravels overlie or are inset within coarser de-
posits in broad,shallow channel reaches.This set of deposits indicates
deposition by stream¯ow (Table 2) of decreasing energy,both over time
A remnant gravel deposit with abundant poorly sorted silty sand matrix
and a distinct depositional margin lies along the west channel (site 3,Fig.
9A;Fig.10).Inset stream¯ow gravels and an eroded inner margin show
that it was deposited before the stream¯ow phase,and we tentatively in-
terpret it as hyperconcentrated-¯ow facies.Hyperconcentrated ¯ow condi-
tions may have resulted from sediment bulking during breaching of the
debris-¯ow lobe and channel incision below the lobe.
Sheet¯ood Subfacies.ÐTwo large distally ®ning sheets of relatively
well-sorted deposits extend from the lower fan over the Slough Creek
¯oodplain (Fig.9A).Boundaries between mapped textural facies are gra-
dational and re¯ect the near-surface texture,although the deposits generally
coarsen upward.The western deposits thin from;30 cmdown to a single-
clast thickness in the central cobble facies (Fig.8E),and form a discontin-
uous veneer in the distal sand to mud facies.The eastern sheet thins from
50 cm of cobble gravel to;10 cm in the central sand facies,but distal
mud accumulations in abandoned ¯oodplain channels are highly variable
in thickness.The deposit surface slope declines from 1.2 8 over the cobble
and pebble facies to 0.38 over the mud facies on the Slough Creek ¯ood-
plain (Fig.9B).Low-relief bar-and-swale morphology is pervasive over the
surface of the gravel and sand facies.The central sand facies contain dis-
continuous millimeter-scale laminae,with signi®cant mud in ®ner laminae
(site 4,Fig.9A;Figs.8F,10).Sorting and surface-parallel strati®cation in
these deposits are indicative of stream¯ow processes,where deposition oc-
curred under uncon®ned and expanding sheet¯ood conditions (e.g.,Wells
and Harvey 1987;Blair and McPherson 1994) (Table 2);however,velocity
and depth conditions may have caused ¯ows to become subcritical over
786 G.A.MEYER AND S.G.WELLS
.9.Ð A) Facies map of the Frenchy's
Meadow event.Boundaries between sheet¯ood
textural facies are gradational,thus approximate.
B) Radial pro®les along the main western
channel area and the large eastern sheet¯ood
deposit of the Frenchy's Meadow fan showing
locations of selected deposits and facies
boundaries;pro®le locations are shown in Figure
the Slough Creek ¯oodplain.Gravel facies have a coarse-and-®ne couplet
stratigraphy similar to sheet¯ood deposits described by Wells and Harvey
(1987) and Blair (1987),albeit weakly developed.The sand facies are also
similar to sand-skirt sheet¯ood facies described by Blair and McPherson
(1994) on distal fans.
Summary.ÐDebris-¯ow and hyperconcentrated-¯ow deposits make up
only about 2% of the Frenchy's Meadow deposit area;the remainder con-
sists of stream¯ow facies,93% of which are sheet¯ood deposits (Table 3).
Processes and deposits are thus in strong contrast to the debris-¯ow-dom-
inated Twelve Kilometer event,despite similar basin characteristics.The
primary geomorphic effect of the event was fan progradation over the
Slough Creek ¯oodplain.In addition to export of sediment from the basin,
fan sediment was transported to more distal sites by fan channel incision
Other Fire-Related Sedimentation Events
The Gibbon Canyon area in west-central Yellowstone was intensely
burned in 1988 (Fig.1D).On 10 August 1989,debris ¯ows were generated
in several small,steep basins on the north canyon wall by high-intensity
convective-storm precipitation lasting 20±30 minutes (Meyer 1993) (Table
1).Bedrock in this area is welded,rhyolitic Lava Creek Tuff (Christiansen
and Blank 1974),which weathers to a ®nes-poor sandy pebble gravel.
Slopes have a thin (mostly,0.3 m) mantle of colluvium with this grus-
like texture.As in northeastern Yellowstone,however,the upper soil is
enriched in silt and clay,and erosion transects and particle-size data show
that ®nes were selectively entrained by sheetwash and rilling on steep
burned slopes (Meyer 1993).The uppermost 1 cm of soil in an uneroded
area was loose and powdery and contained 54% silt,13% clay,and 10%
organic material,whereas the top 1 cm of soil in a sheetwash-eroded area
was more compact and contained 26% silt,10% clay,and 6% organic
material.In the``La Familia''basin (Meyer 1993) (Table 1),mud coatings
on upper channel walls and trees indicated that debris-¯ow conditions were
achieved near the midpoint of the main channel.
Witnesses at the La Familia fan observed a small initial ¯ow of water,
followed by two major debris-¯ow surges carrying scattered``¯oating''
boulders;the surges were separated by a few minutes of shallow sheet-
787FIRE-RELATED SEDIMENTATION ON ALLUVIAL FANS
.10.ÐCumulative curves showing particle
size distribution in the,24 ffraction of
Frenchy's Meadow fan deposits;circled numbers
show location of sample in Figure 9A.Arrows
show inferred time sequence of deposition,
where each succeeding deposit was emplaced
farther downfan.Note the general decrease of
particle size and increase in sorting with distance
downfan.The hyperconcentrated-¯ow cobble
splay matrix (3) shows slightly poorer sorting,
probably because of incision and associated
sediment bulking in the channel above this
¯ooding (Rob Danno,NPS,personal communication 1990).Corresponding
deposits consist of two relatively ®nes-poor debris-¯ow units (4±8% silt
and clay in the,24 ffraction) (Fig.11),separated by several upward-
coarsening beds of sand to pebble gravel,each
;5 cm thick.Overlying
strati®ed,well-sorted coarse sand to cobble gravel 10±20 cmthick indicates
a ®nal interval of sheet¯ooding.In nearby basins,deposits with large ma-
trix-supported boulders,steep margins,prominent levees,and short runout
distances gave evidence for relatively high-strength,poorly cohesive debris
¯ows.Sorting is a useful discriminator of facies for Yellowstone fan de-
posits;however,textures of a few deposits interpreted as noncohesive de-
bris-¯ow and hyperconcentrated-¯ow facies in the Gibbon Canyon area are
similar (Fig.11).Stream¯ow deposits were typically minor,although a
debris-¯ow deposit and underlying fan sediments were deeply incised by
late-stage dilute ¯ow on one fan.Several additional sedimentation events
occurred in the Gibbon Canyon area through 1991,but debris ¯ows were
small and were overwhelmed by stream¯ow deposition and erosion in later
events (Meyer 1993).
Fire-related sedimentation events were also investigated in larger basins
in the northeastern Yellowstone area.Cutoff Creek drains a steep,8.3 km
basin on volcaniclastic rocks.This Slough Creek tributary ¯ooded in the
same storm that produced the Twelve Kilometer and Slough R.S.events
(Fig.1D,Table 1).Moderately sorted cobble to pebble gravel bars and
splays along the main fan channel gave evidence for stream¯ow processes.
A large ®re-related ¯ood also occurred in the West Fork of Lodgepole
Creek in the Clarks Fork of the Yellowstone River drainage in July 1989
(Fig.1D,Table 1).Bedrock in this 5.1 km
basin is a porphyritic intrusive
rock that weathers into a coarse,talus-like colluvium.This material ap-
peared to have protected the intensely burned slopes from erosion,and rills
were shallow except in deeper footslope colluvium.Elongate bars of ma-
trix-free,imbricated boulders and cobbles and inset planar-bedded sheet-
¯ood gravels indicate stream¯ow deposition (Table 2) on the tributary-
junction fan at the con¯uence with the main stem of Lodgepole Creek.
Initiation of Debris Flows by Progressive Sediment Bulking
A common mechanism for debris-¯ow initiation is the transformation of
en masse slope failures into debris ¯ow by addition of water (e.g.,Johnson
1970,1984;Costa 1984).This process is typical of unglaciated slopes with
thick colluvial mantles (e.g.,Campbell 1975;Reneau and Dietrich 1987),
and is promoted in burned areas when evapotranspiration is reduced and
roots anchoring colluvium decay (e.g.,Swanson 1981;Swanson and Dyr-
ness 1975).Despite thorough examination,however,no en masse failures
were found in the Twelve Kilometer and Gibbon Canyon basins,and aerial
reconnaissance of many debris-¯ow source basins over northeastern Yel-
lowstone revealed only a few isolated debris slides (Meyer 1993).A less
common debris-¯ow initiation process was described by Johnson (1970,
1984) as the``®rehose effect'',whereby high-energy dilute ¯ows in very
steep bedrock channels debouch onto talus slopes and entrain enough loose
sediment to achieve debris-¯ow concentration.Generation of noncohesive
debris ¯ows has similarly been linked to highly turbulent ¯ow in steep
channels containing abundant available sediment (Church and Desloges
In contrast,evidence exists in active volcanic settings for initiation of
laharic debris ¯ows by progressive sediment bulking of runoff,involving
erosion by widespread surface runoff as well as channelized ¯ow (Scott
1988).Slopes mantled by loose volcaniclastic materials are broadly anal-
ogous to burned slopes in terms of runoff potential and availability of ®ne
sediment.In Yellowstone debris-¯ow basins,densely rilled slopes indicat-
ing pervasive surface runoff were ubiquitous (Johansen 1991;Meyer 1993).
Johnson (1984) and Wells (1987) recognized the connection between rilling
and ®re-related debris-¯ow generation in California,but interpreted that
rills initiate through many small saturation-induced en masse failures,as
suggested by shallow spoon-shaped scarps at some rill heads.Little evi-
dence was seen to support this model in Yellowstone debris-¯ow basins.
Small scarps at the heads of rills were not observed,although small knick-
points were observed in some rills.Broad areas of sheetwash erosion on
gentle crestslopes were transitional to poorly integrated rills on steep mids-
lopes.Although small pebbly levees along a few rills suggested that surface
runoff reached debris-¯ow concentrations,no footslope debris-¯ow accu-
mulations were observed,even where rills ended on low-gradient areas.
Debris-¯ow evidence such as levees and mud coatings on trees typically
®rst appeared in the middle reaches of the main basin channels,indicating
that additional sediment bulking by channel erosion was required to pro-
duce debris-¯ow concentrations.Undercutting of banks caused very small
slumps and topples of sediment into larger channels in Yellowstone,but
few extended outside of the channel alluvium itself.Similarly,in a study
of ®re-related debris ¯ows near Helena,Montana,Parrett (1987) noted the
788 G.A.MEYER AND S.G.WELLS
.11.ÐGraphical standard deviation of
particle size (i.e.,sorting;Folk 1974) as a
function of mean particle size for ®re-related
alluvial fan deposits (,24 ffraction).
Localities for debris-¯ow samples are indicated.
HCF 5 hyperconcentrated-¯ow deposits,NCDF
5 noncohesive or poorly cohesive debris-¯ow
rarity of en masse slope failures and interpreted a downslope sequence of
sheetwash,rilling,and channel incision,with progressive sediment bulking
to debris-¯ow conditions.
Importance of Fine Sediment Sources in Determining Flow
Effects of the 1988 forest ®res common to all of the event basins include
widespread exposure of loose,silty A horizon surfaces on steep slopes,and
the addition of ®ne wood ash to those surfaces.Soil surfaces provide abun-
dant ®nes for entrainment by surface runoff,but sheetwash and rill¯ow
typically lack the competence to transport gravel.Observations and parti-
cle-size data from burned soil surfaces in Yellowstone (Meyer 1993) and
in New Mexico and Colorado (White and Wells 1979;Morris and Moses
1987) indicate that burned slopes contribute mostly sand and ®nes to sur-
face runoff.The importance of ®nes in producing debris-¯ow conditions
has been demonstrated in numerous laboratory experiments (e.g.,Hampton
1972;Rodine and Johnson 1976;Major and Pierson 1992) and ®eld studies
(e.g.,Harvey 1984;Wells and Harvey 1987).Yield strength in sediment±
water ¯ows is promoted by cohesion in clays and by intergranular friction
between abundant silt- and sand-size particles (Rodine 1974;Pierson and
Costa 1987).The contribution of ®ne sediment to runoff from burned
slopes is thus important in generating debris ¯ows,because sediment stored
in larger channels is predominantly sand and gravel.
In the 1989 Twelve Kilometer,Slough R.S.,and La Familia events,
abundant ®ne soil-surface sediment was available one year after burning,
and debris-¯ows were prevalent.The Twelve Kilometer event in particular
was dominated by ®nes-rich debris ¯ows (Figs.2±5;Table 3),probably
because of abundant ®ne sediment derived from the muddy volcaniclastic
bedrock in that basin.By comparison,Blair (1987) attributed the domi-
nance of sheet¯ood processes in the Roaring River fan sedimentation event
to a lack of available ®ne sediment.
Range of Flow Processes in Single Fire-Related Events
Flows evolved through a wide range of sediment concentration in indi-
vidual ®re-related sedimentation events in Yellowstone.In the three events
studied in detail,depositional evidence showed a progression from sedi-
ment-rich debris ¯ows toward ¯ows with lower sediment concentrations.
Hyperconcentrated ¯ows commonly breached and deposited inset gravels
within debris-¯ow deposits.Late-stage stream¯ows also breached and erod-
ed hyperconcentrated-¯ow and debris-¯ow deposits (Figs.2,6A,7,8C,
9A),and on the distal Twelve Kilometer and Slough R.S.fans,pebbly
stream¯ow facies overlie gravel-poor debris-¯ow facies.Evidence for a
similar progression toward more dilute ¯ow,from noncohesive debris ¯ow
through successively lower-discharge sheet¯ood deposition and ®nally ero-
sion,was noted in the Roaring River fan deposits (Blair 1987).Depositional
processes in this event resembled those in the Frenchy's Meadow event,
but the maximum size of transported clasts indicates that peak discharge
was much greater in the Roaring River dam-break ¯ood,and sediment was
derived almost entirely by channel incision in a gravelly boulder till with
only 0.5% silt and clay (Blair 1987).
Convective storms in Yellowstone typically reach maximum intensity
quickly.It is unlikely that precipitation intensity increased during an event
such that runoff eventually overwhelmed available sediment,producing the
observed progression toward lower sediment concentration in ¯ows on fans.
This progression is more likely caused by decreasing sediment availability.
Sediment supply may be limited by downcutting to less erodible substrates,
or armoring through development of coarse lags.Evidence for these pro-
cesses were observed over a wide range of scales,from shallow rills to
main channels.Rills developed pebbly armors in some sections,and lo-
cally,roots and shallow bedrock limited rill incision.Most commonly,rills
cut down to more compact,less erodible soil horizons as loose upper soil
was stripped away.Upper-basin channels were frequently incised to bed-
rock,and lags of coarse boulders in most other channels limited further
entrainment except by channel widening.The control of ¯ow process by
sediment availability is also supported by observations of repeated storm-
runoff events in Gibbon Canyon basins:later events were dominated by
dilute-¯ow processes,in part because of removal of loose soil-surface ®nes
and stored channel alluvium in earlier events.
In small fans generated by an extreme precipitation event in the Howgill
Fells,England,Wells and Harvey (1987) observed depositional evidence
for a rapid initial increase in sediment concentration in ¯ows,followed by
a slower,progressive decrease in concentration.Evidence for an initial
depositional phase of increasing sediment concentration was rarely seen in
Yellowstone fan deposits.A primary sediment source in the Howgill Fells,
however,was en masse debris slides,which likely took up to 1 hour to
develop through saturation.Earlier ¯ows were dilute because intact vege-
tated surfaces limited sediment supply to surface runoff.In contrast,initial
storm runoff reaching fans in Yellowstone was probably of high sediment
789FIRE-RELATED SEDIMENTATION ON ALLUVIAL FANS
.12.ÐMorphometry and dominant ¯ow processes in Yellowstone basins pro-
ducing ®re-related sedimentation on alluvial fans (Table 1).The pattern inside basin-
area bars indicates the dominant depositional process(es) in the event.Bars are
ordered with smallest area at left.Relief ratio,a measure of basin steepness,is the
maximum relief from basin mouth to divide divided by length of the longest stream
channel extended to the divide.
concentration because surface runoff entrained ®ne sediment from bare
soils,with additional sediment bulking by channel incision.
Differences in Flow Processes Between Events in Different Basins
Major differences in the relative importance of debris ¯ows and stream-
¯ows in the Yellowstone events suggest that basin characteristics may exert
a strong in¯uence on ¯ow processes.Several studies have suggested that
in general,smaller,steeper basins tend to produce debris ¯ows on alluvial
fans because of greater erosion and sediment supply per unit runoff-pro-
ducing area,whereas larger basins tend to produce stream¯ow deposition
(e.g.,Kostaschuk et al.1986;Wells and Harvey 1987).In a large basin,
debris ¯ows generated on steep upper slopes may terminate in gentler
reaches before reaching the fan.In a general sense,this relationship also
exists within a sample of Yellowstone ®re-related events (Fig.12).In initial
events,basins of,2 km
produced mostly debris-¯ow deposits on fans,
whereas larger basins produced stream¯ow-dominated deposits.
Factors other than basin area,however,must account for the observed
differences in processes between basins.The small,steep Slough R.S.basin
yielded 55% debris-¯ow deposits by area;during the same storm,the
Twelve Kilometer basin produced 96% debris-¯ow deposits,despite being
twice the size of the Slough R.S.basin and less steep (Tables 1,3).This
difference probably stems in large part from contrasting bedrock and sur-
®cial deposits within the basins.The Twelve Kilometer basin is underlain
by friable,muddy volcaniclastic rocks,and has more glacial sediment and
alluvium along its channel than the Slough R.S.basin,which is underlain
by granitic gneiss.Differences in basin lithology and morphometry,how-
ever,do little to explain the major differences in ¯ow process between the
Frenchy's Meadow and Twelve Kilometer events (Fig.12).The two basins
are similar in size,relief ratio,overall morphology,bedrock and sur®cial
geology,and burn intensity;yet stream¯ow-facies constituted 98% of the
deposit area in the Frenchy's Meadow event (Tables 1,3).Differences in
the timing of these events may be critical.The Twelve Kilometer event
occurred in July 1989,nine months after the 1988 ®res and following a
period of warm,dry weather.The Frenchy's Meadow event occurred in
late spring of 1991,
;30 months after the ®res,and shortly after snowmelt.
Soil moisture may have been high prior to the Frenchy's Meadow event,
leading to widespread,rapid surface runoff.
Nonetheless,increased surface runoff does not explain why more ®ne
soil-surface sediment was not entrained with little vegetative cover present.
By 1991,burned soil surfaces had been subject to compaction by raindrop
impact,in®ltration,and snow accumulation.Although compaction was not
directly measured in the Frenchy's Meadow basin,repeated ®eld obser-
vations in Gibbon Canyon indicated compaction of soil surfaces between
1988 and 1991.In®ltration was reduced such that one day after a large
¯ood event in 1991,soil underlying the wet,compacted surface was dry.
The soil surface hardened on drying,indicating crusting and surface sealing
(Swanson 1981;RoÈmkens et al.1990).Although runoff was increased,®ne
sediment availability decreased because of compaction and increased co-
hesion.Therefore,the difference in ¯ow processes between the Twelve
Kilometer and Frenchy's Meadow events may be explained largely by
changes in soil surface properties with time after burning.
Where basins experienced repeated events,stripping of the soil surface
by runoff would also reduce the availability of ®ne sediment in subsequent
events.In sediment-trap studies in the Colorado Front Range,Morris and
Moses (1987) observed rapidly declining sediment yield from slopes with
time after ®re,despite minimal recovery of vegetation or litter cover.Par-
ticle-size data showed that soil surfaces coarsened over time on burned
slopes because of removal of ®ner sediments.Field observations and lim-
ited particle-size data (Meyer 1993) indicate a similar trend in Yellowstone.
Generation of surface runoff in burned areas is frequently ascribed to
water-repellent compounds formed on or near soil surfaces by burning of
organic material,particularly in chaparral environments (e.g.,DeBano
1980;Wells 1987;Florsheim et al.1991).In Yellowstone burns,water
repellency was not commonly present in soils immediately after the ®res
in 1988 (Shovic 1988) and was sporadic or absent in event basins in 1989,
despite abundant evidence of surface runoff (Johansen 1991;Meyer 1993).
These observations suggest that physical properties of bare burned soils are
suf®cient to cause low in®ltration and surface runoff.Regrowth of herba-
ceous plant cover after ®re reduces both slope sediment yield and surface
runoff.In the Gibbon Canyon basins,transects indicated that burned slopes
.30±35% cover of post®re herbaceous plants experienced minimal
surface runoff and rill erosion in the same stormthat generated major rilling
of bare slopes and debris ¯ows (Meyer 1993).Revegetation caused ®re-
related sedimentation events to become rare in Yellowstone after 1991.
Depositional Patterns on Alluvial Fans
A general trend expressed within each major facies type as well as the
fan deposits as a whole was ®ning of the coarse fraction downfan.Several
other studies document proximal to distal coarse-tail ®ning of debris-¯ow
deposits (e.g.,Sharp and Nobles 1953;Bull 1964a;Suwa and Okuda 1983).
The progressive loss of coarser material may occur because it is rolling or
sliding instead of being fully supported (Pierson 1986),so that expanding,
thinning,and decelerating debris ¯ows allow grounding of larger clasts and
deposition due to decreased ¯ow energy and competence.Debris-¯ow fa-
cies with abundant coarse material imparting relatively high strength and
low mobility are emplaced in proximal fan areas and at higher depositional
gradients (Figs.2,6,9).Gravel-poor debris-¯ow facies were also deposited
on the upper Twelve Kilometer and Slough R.S.fans but are minor in
abundance (Figs.2,6A).Bouldery deposits also resist downfan reworking
by erosive dilute ¯ows.
Downfan coarse-tail ®ning trends are apparent in hyperconcentrated-¯ow
790 G.A.MEYER AND S.G.WELLS
and stream¯ow facies deposits as well.Boulder splays are replaced by
cobble splays in hyperconcentrated-¯ow facies down the Slough R.S.fan.
The maximum clast size in Frenchy's Meadow stream¯ow deposits de-
creases downfan,and sheet¯ood facies as a whole ®ne downfan in a regular
and gradational manner (Fig.9A).Sheet¯ood expansion and decreasing fan
slope caused reduction of stream power and deposition of progressively
®ner material.The concave-up eastern sheet¯ood deposit pro®le in Fig.9B
was built over a nearly ¯at ¯oodplain surface by deposition of coarser
material at a steeper slope.
During single events,the main area of deposition tended to migrate
downfan as ¯ows became more dilute.Downfan transfer of the depositional
locus occurred through breaching of,or avulsion around,earlier and coarser
deposits higher on the fan,often with channel incision and rebulking of
the ¯ow (Figs.6A,8C,9A).Deposition resumed below intersection points.
This downfan-shifting pattern of deposition was noted by Blair (1987) in
the Roaring River sheet¯ood-dominated fan event.
A concave pro®le is a feature of most fans (Blissenbach 1954) and is
expectable because a positive correlation exists between slope and mean
particle size on a fan (e.g.,Denny 1965;Blair and McPherson 1994),and
sediments usually become ®ner downfan (e.g.,Beaty 1963;Bull 1964a;
Heward 1978).Segmented radial fan pro®les (i.e.,those with abrupt
changes in slope between major segments) have been interpreted to re¯ect
tectonic tilting (Hooke 1972) or long-term changes in depositional pro-
cesses stemming from climatic variations (e.g.,Beaty 1961;Bull 1964b).
In the Yellowstone event deposits,marked changes in slope are often as-
sociated with intersection points and abrupt,facies-related downfan ®ning
(Figs.6B,9B).As noted by Blair (1987),persistence of such depositional
patterns could lead to segmented fan pro®les through intrinsic fan pro-
SUMMARY AND CONCLUSIONS
Forest ®res in Yellowstone dramatically increase runoff generation and
sediment availability,thus increasing the probability of sedimentation
events on alluvial fans.Flows in a single event varied widely in sediment
concentration,from debris ¯ow through hyperconcentrated ¯ow and
stream¯ow,producing a variety of facies on the fan.Debris ¯ows were
generated by progressive bulking of sediment in slope runoff and in chan-
nels.Sediment concentration in ¯ows on fans declined over the course of
an event,probably because of diminishing sediment availability as loose
®nes were stripped from slopes and as armors developed in channels.The
distribution of ¯ow processes in an event is likely determined by a variety
of intrinsic basin characteristics,including morphometry,lithology,and
regolith characteristics,as well as timing of the event with respect to ®re.
Rapid evolution of the upper soil toward a more compact,erosionally re-
sistant surface after ®re can greatly reduce the ®ne sediment input to runoff,
even before slopes are revegetated.Thus,seemingly subtle differences in
soil properties may control whether fan deposition is dominantly by debris-
¯ow or stream¯ow processes.
This project was undertaken as part of the Ph.D.research of Grant Meyer at the
University of New Mexico and was supported by grants from National Science
Foundation (EAR 9005058),Geological Society of America±Quaternary Geology
and Geomorphology Division,Sigma Xi,and the University of New Mexico Grad-
uate Student Association.We are especially grateful for assistance from Gary Smith
on interpretation of ¯ow processes and sedimentology.Christy Terhune skillfully
conducted laboratory analyses of sediments and soils.We thank Ralph Mason,Chris
Inoue,Danny Katzman,John Rogers,and many other enthusiastic individuals for
®eld assistance,and the National Park Service and U.S.Forest Service for cooper-
ation and logistical support.The paper was greatly improved by comments from
John Shaw,Terence C.Blair,and two anonymous reviewers.
,S.G.,1992,Relation of surface climate and burned
area in Yellowstone National Park:Agricultural and Forest Meteorology,v.60,p.285±293.
,C.B.,1961,Topographic effects of faulting,Death Valley,California:Association of
,C.B.,1963,Origin of alluvial fans,White Mountains,California and Nevada:Asso-
ciation of American of Geographers,Annals,v.53,p.516±535.
,J.K.,1964,Hyperconcentrations of suspended sediment:
American Society of Civil Engineers,Journal of the Hydraulics Division,v.90,no.HY6,
,T.C.,1987,Sedimentary processes,vertical strati®cation sequences,and geomorphology
of the Roaring River alluvial fan,Rocky Mountain National Park,Colorado:Journal of
,J.G.,1994,Alluvial fans and their natural distinction fromrivers
based on morphology,hydraulic processes,sedimentary processes,and facies assemblages:
Journal of Sedimentary Research,v.A64,p.450±489.
,E.,1954,Geology of alluvial fans in semiarid regions:Geological Society of
,W.B.,1964a,Alluvial fans and near-surface subsidence in western Fresno county,Cal-
ifornia:U.S.Geological Survey,Professional Paper 437-A,71 p.
,W.B.,1964b,Geomorphology of segmented alluvial fans in western Fresno county,
California:U.S.Geological Survey,Professional Paper 352-E,p.89±129.
,R.H.,1975,Soil slips,debris ¯ows,and rainstorms in the Santa Monica Mountains
and vicinity,southern California:U.S.Geological Survey,Professional Paper 851,51 p.
,V.T.,1959,Open-Channel Hydraulics:New York,McGraw-Hill,680 p.
.,1974,Geologic map of the Madison Junction quad-
rangle,Yellowstone National Park,Wyoming:U.S.Geological Survey,Map GQ-1193,1:
,J.R.,1984,Debris torrents and natural hazards of steep mountain
channelsÐeast shore of Howe Sound:Canadian Association of Geographers,University of
British Columbia,Dept.of Geography,Vancouver,Field Trip Guidebook 7,83 p.
,J.E.,1984,Physical geomorphology of debris ¯ows,Chapter 9,in Costa,J.E.,and
Fleisher,P.J.,Developments and Applications of Geomorphology:Berlin,Springer-Verlag,
,J.E.,1988,Rheologic,geomorphic,and sedimentologic differentiation of water ¯oods,
hyperconcentrated ¯ows,and debris ¯ows,in Baker,V.R.,Kochel,R.C.,and Patton,P.C.,
eds.,Flood Geomorphology:New York,Wiley,p.113±121.
,G.P.,1984,Debris-Flow Dynamics (videotape):U.S.Geological
Survey,Open-File Report 84-606,22.5 min.
,P.R.,1965,Particle fractionation and particle size analysis,in Black,C.A.,ed.,Methods
of Soil Analysis,Pt.1:American Society of Agronomy,Madison,Wisconsin,Monograph
,L.F,1980,Water-repellent soils:A state-of-the-art:U.S.Department of Agriculture,
Forest Service,Paci®c Southwest Forest and Range Experiment Station,General Technical
Report PSW-46,21 p.
,C.S.,1965,Alluvial fans in the Death Valley region,California and Nevada:U.S.
Geological Survey,Professional Paper 466,62 p.
,D.W.,1991,Fluvial sediment transport following
chaparral wild®re,Ventura County,southern California:Geological Society of America,
,R.L.,1974,Petrology of Sedimentary Rocks:Austin,Texas,Hemphill,182 p.
,T.C.,1985,Polallie Creek debris ¯ow and subsequent dam-break
¯ood of 1980,East Fork Hood River Basin,Oregon:U.S.Geological Survey,Water-Supply
Paper 2273,22 p.
,M.A.,1972,The role of subaqueous debris ¯ow in generating turbidity currents:
Journal of Sedimentary Petrology,v.42,p.775±793.
,M.A.,1975,The competence of ®ne-grained debris ¯ows:Journal of Sedimentary
,A.M.,1984,Debris ¯ows and ¯uvial deposits in Spanish Quaternary alluvial fans:
Implications for fan morphology,in Koster,E.H.,and Steel,R.J.,eds.,Sedimentology of
Gravels and Conglomerates:Canadian Society of Petroleum Geologists,Memoir 10,p.123±
,A.P.,1978,Alluvial fan sequence and megasequence models:with examples from
Westphalian D±Stephanian B coal®elds,northern Spain,in Miall,A.D.,ed.,Fluvial Sedi-
mentology:Canadian Society of Petroleum Geologists,Memoir 5,p.669±702.
,R.L.,1972,Geomorphic evidence for Late Wisconsin and Holocene tectonic defor-
mation,Death Valley,California:Geological Society of America,Bulletin,v.83,p.2073±
,E.A.,1991,Quantitative analysis of the variables causing debris ¯ows [unpublished
M.S.thesis]:Montana College of Mineral Science and Technology,Butte,89 p.
,A.M.,1970,Physical Processes in Geology:San Francisco,Freeman Cooper,577 p.
,J.R.),1984,Debris ¯ow,Chapter 8,in Brunsden,D.,and Prior,
D.B.,Slope Instability:New York,Wiley,p.257±362.
,P.E.,1986,Depositional processes and
alluvial fan±drainage basin morphometric relationships near Banff,Alberta,Canada:Earth
Surface Processes and Landforms,v.11,p.471±484.
,T.C.,1992,Debris ¯ow rheology:Experimental analysis of ®ne-
grained slurries:Water Resources Research,v.28,p.841±857.
,G.A.,1993,Holocene and modern geomorphic response to forest ®res and climate
791FIRE-RELATED SEDIMENTATION ON ALLUVIAL FANS
change in Yellowstone National Park [unpublished Ph.D.thesis]:University of NewMexico,
Albuquerque,New Mexico,402 p.
,A.J.T.,1995,Fire and alluvial chronology in Yellowstone
National Park:Climatic and intrinsic controls on Holocene geomorphic processes:Geolog-
ical Society of America,Bulletin,v.107,p.1211±1230.
,A.J.T.,1992,Response of alluvial
systems to ®re and climate change in Yellowstone National Park:Nature,v.357,p.147±
,J.I.,1977,Chemical weathering and related controls on surface
water chemistry in the Absaroka Mountains,Wyoming:Geochimica et Cosmochimica Acta,
,T.A.,1987,Forest ®re and the natural soil erosion regime in the
Colorado Front Range:Association of American Geographers,Annals,v.77,p.245±254.
,C.,1987,Fire-related debris ¯ows in the Beaver Creek drainage,Lewis and Clark
County,Montana:U.S.Geological Survey,Water-Supply Paper 2330,p.57±67.
,K.L.,1974,Sur®cial geologic map of the Tower Junction quadrangle and part of the
Mount Wallace quadrangles,Yellowstone National Park,Wyoming and Montana:U.S.Geo-
logical Survey,Miscellaneous Geological Investigations Map I-647,1:62,500.
,K.L.,1979,History and dynamics of glaciation in the northern Yellowstone National
Park area:U.S.Geological Survey,Professional Paper 729-F,90 p.
,T.C.,1980,Erosion and deposition by debris ¯ows at Mt.Thomas,North Canterbury,
New Zealand:Earth Surface Processes,v.5,p.227±247.
,T.C.,1981,Dominant particle support mechanisms in debris ¯ows at Mt.Thomas,
New Zealand,and implications for ¯ow mobility:Sedimentology,v.28,p.49±60.
,T.C.,1986,Flow behavior of channelized debris ¯ows,Mount St.Helens,Washing-
ton,in Abrahams,A.D.,ed.,Hillslope Processes:Boston,Allen & Unwin,p.269±296.
,J.E.,1987,A rheologic classi®cation of subaerial sediment±water
¯ows,in Costa,J.E.,and Wieczorek,G.F.,eds.,Debris Flows/Avalanches:Geological So-
ciety of America,Reviews in Engineering Geology 7,p.1±12.
,K.M.,1985,Downstream dilution of a laharÐtransition fromdebris
¯ow to hyperconcentrated stream¯ow:Water Resources Research,v.21,p.1511±1524.
,E.T.,1975,Geologic map of
the Tower Junction Quadrangle,Yellowstone National Park,Wyoming:U.S.Geological
Survey,Geological Quadrangle Map GQ-1247,1:62,500.
,W.E.,1987,The importance of hollows in debris-¯ow studies:
Examples from Marin County,California,in Costa,J.E.,and Wieczorek,G.F.,eds.,Debris
Flows/Avalanches:Geological Society of America,Reviews in Engineering Geology 7,p.
,J.R.,1974,Analysis of mobilization of debris ¯ows [unpublished Ph.D.thesis]:Stan-
ford University,Palo Alto,California,226 p.
,A.M.,1976,The ability of debris,heavily freighted with coarse
clastic materials,to ¯ow on gentle slopes:Sedimentology,v.23,p.213±234.
,F.D.,1990,Surface sealing and in®ltration,
Chapter 5,in Anderson,M.G.,and Burt,T.P.,eds.,Process Studies in Hillslope Hydrology:
,D.G.,1989,The Yellowstone ®res:Scienti®c American,v.261,
,K.M.,1988,Origins,behavior,and sedimentology of lahars and lahar-runout ¯ows in
the Toutle±Cowlitz River system:U.S.Geological Survey,Professional Paper 1447-A,74
,L.H.,1953,Mud¯ow of 1941 at Wrightwood,southern California:
Geological Society of America,Bulletin,v.64,p.547±560.
,H.F.,1988,Post®re soil research:Report to the Greater Yellowstone Post®re Ecological
Assessment Workshop,November 17,1988,Mammoth,Wyoming.
,H.J.,1972,Stratigraphic framework of the Absaroka Volcanic
Supergroup in the Yellowstone National Park region:U.S.Geological Survey,Professional
Paper 729-C,33 p.
,G.A.,1986,Coarse-grained nonmarine volcaniclastic sediment:terminology and de-
positional process:Geological Society of America,Bulletin,v.97,p.1±10.
,D.R.,1991,Lahars:Volcanohydrologic events in the debris ¯ow±
hyperconcentrated ¯ow continuum,in Fisher,R.V.,and Smith,G.A.,eds.,Sedimentation
in Volcanic Settings:SEPM,Special Publication 45,p.59±70.
,S.,1983,Deposition of debris ¯ows on a fan surface,Mt.Yakedake,
Japan:Zeitschrift fuÈr Geomorphologie,N.F.,Supplement Band 46,p.79±101.
,F.J.,1981,Fire and geomorphic processes,in Mooney,H.A.,Bonnicksen,T.M.,
Christensen,N.L.,Lotan,J.E.,and Reiners,W.A.,eds.,Fire Regimes and EcosystemProp-
erties:U.S.Department of Agriculture,Forest Service,General Technical Report WO-26,
,C.T.,1975,Impact of clear-cutting and road construction on soil
erosion by landslides in the western Cascade Range,Oregon:Geology,v.3,p.393±396.
,F.H.,1989,The generation of turbidity currents by subaerial debris ¯ows,California:
Geological Society of America,Bulletin,v.101,p.278±291.
,A.M.,1987,Sedimentologic and geomorphic variations in storm-
generated alluvial fans,Howgill Fells,northwest England:Geological Society of America,
,W.G.,II,1987,The effect of ®re on the generation of debris ¯ows,in Costa,J.E.,and
Wieczorek,G.F.,eds.,Debris Flows/Avalanches:Geological Society of America,Reviews
in Engineering Geology 7,p.105±114.
,T.,1992,The in¯uence of debris ¯ow rheology on fan morphology,
Owens Valley,California:Geological Society of America,Bulletin,v.104,p.887±900.
,S.G.,1979,Forest-®re devegetation and drainage basin adjustments
in mountainous terrain,in Rhodes,D.D.,and Williams,G.P.,eds.,Adjustments of the Flu-
vial System:State University of New York at Binghamton,Binghamton,New York,Tenth
Annual Geomorphology Symposium Series,Proceedings,p.199±223.
Received 17 October 1996;accepted 29 April 1997.