Kenneth A. Cameron and Patrick Pringle, USGSCascades Volcano Observatory VancouverWashinoton 98661

Post-GlacialLahars of the SandyRiver Basin, MountHood,

Abstract Vithin the last I0,000 years,three significant lahar'producingperiods have occurr€dar , Oregon.The Timberline eruptire period occurred belween 1400and 1800years BP. It wasby far the nosr voluminousof lhe three periods,producing enoughclastic debris to bury the glacialtopography ofthe southwestface of the mountain beneatha snootb debris fan. Snaller debris fans were forned in the upper reachesof the Sandy and Salmon Rivers. Timberline-agelahars trareled the length of the Zigzag and SandyRivers, a disrancein ercesso{90 krn. Thesejahars and associatedfluvial depositshelped forrn flat-floored valleys near the confluenceof these rivers and cr€ated a deha at the mouth of the Sandy at the Colunbia River. The flons probably atlained depthsofg to 12 m above modern river level for nost of their passage.Pyroclastic flows of this age traveled at least 13 km frorn the rent a.€a at Crat€r Rock along the Zigzag or Litle Zigzag Rirers. Between400 and 600 yearsBP, ihe Zigzag erupriveperiod produceddeposirs along the middle reachesof the Zigzag River and the upper . Bouldert labars and flurial depositscreated a terrace B to I0 n above present river lerel along the Zigzag and veneereda Timberlin€-aget€rrace on the Sandy River. The Old Maid eruptive period occured b€tw€€n lB0 and 2?0 years BP. A single lahar flowed down the Sandy River at least as {ar as Brightwood,30 }n from Crarer Rock. Sand d€posilsthat may b€ related to this erent are found down to the nouth of the SandyRiver. A single lahar also flowed at least l8 km down th€ Zigzag River. The lahars,which were probably about 9 n deep,buried maiure cedar forestsalong both rivers.

lntroduction to a discussionof the depositsin the Sandyand Zigzag drainagehasins (Figure l). Thesebasins Since the end of the Fraser alpine glaciatiod werechosen due to the concentrationof popula- abour 10,000years ago, three significant lahar- tion and recent intensedevelopment, which producing eruptive periods have occurred at would exacerbatethe effectsof future eruptions. Mount Hood. The lahars were confined to river basinsoriginating on the southwestflank of the CraterRock mountain.The older, Timberline eruptiveperiod occuried 1400ro lB00 yearsBP (Crandell,1980), A compositedome known as Crater Rock lies at the Zigzag eruptive period 400 ro 600 yearsBP, the apex of the smooth debris fan ruhich gives and the more recent Old Maid eruptive period the southwestside of Mount Hood its distinctive occurredlB0 ro 270 yearsBP (Crandell,1980). profile (Figure 2). This massof amphibole-rich Depositsfrom the Timberlineeruptive period (in- phyric dacite sits atop the modern vent, and its formally termed Timberline-agein this report) extrusion was the probable causative force can be tracedthe length of the Zigzagand Sandy behindlahar generation. Rivers to the Columbia River. Flows from the Field studiesand air-photointerpretation Zigzag eruptive period (informally termed showthat the Crater Rock dome is composedof Zigzag-agein this report)have been identified at leastthree lobes(Figure 3). This is basedupon only in the niddle Zigzag basin and in the up- degreeof alteralion.lexture. and fabric conlrasts per Sandybasin at Old Maids Flat. Depositsfrom betweendifferent areasof the dome.The dacite (informally the Old Maid eruptive period termed of LobeI hasan oxidized,brick red or pink col- 0ld Maid-agein this report)have been identified ored matrix, but phenocrystsof plagioclaseand from the upper Sandy basin downstreamto 2.5 amphiboleare unaltered.Localized areas of high km belov the confluenceof the Sandy and Sal- fumarolic activity havecreated patches of white mon Riversnear the hamletof Brightwoodand and yellowalteration and encrustation.Spires of at a single locality below the confluenceof the whitish,pumiceous rock rise 5 to l0 m abovethe Zigzag ard Little Zigzag Rivers. surfaceof the domeon the north-northwestside Although the lahars from the three eruptive of the lobe,The generalsurface is very irregular periods affected all drainageson the soulhwest and angularwith local relief of over I m. The flank of Mount Hood,this paperwill be limired lobe showsno obviousfabric.

NorthweslScience. Vol. 60, \o. 4, lq86 225 T

J

1:

Figure l. Generallocation nap for the Sandy and Zigzag River basinr.

.:{&

Figure 2. South*est flank o{ Mount Hood shoving the smooth surfac€ of the main debris fan forin€d during rhe Tinb€rlin€ eruptive p€riod, wifi Crater Rock at its apex.Mount Adans, Washington,is in the background.Photograph by Ellen Cameron.

226 Cameronand Pringle SLrm mrt

.RATERAREAMAp .€>S)t "I K\-7 ' -- - (./----,r ).. /? t:':t{'}-ta' €

Figure 3 Map vi€{ of C.ater Rock and su.rounding area shosing location of the rhree lobes.Dots representfunrarotic areas. "X" indicales spines on the dome. Dashed lin€s indicate inferred linirs ot rhe breadcrusr-bonb teohra Dlun€.

Lobe 2 has an unaltered grey matrix and a Two main regionsof fumarolic activity flank much smoothersurface texture with Iocal relief Crater Rock and bound CoalmanGlacier (Figure of underI m. It possessesa general fabric in the 3). Temperaturesof the fumaroles(B4oC to 90oC) form of platyjointing that dipsto the southwest are at or belowlocal boiling point. Hydrother- (downslope) and is the result of flor.vwithin the mal alteration has reducedthe surrounding lobe during extrusion.It is separatedlocally from fragmentaldebris to sticky yellow,red and blue- Lobe I by deepcrevasses eroded along the con- grey clay.Both fumarolicareas remain snow-free tact. No current fumarolic activity was seenon throughout the winter despite their elevation Lobe 2. (3200m) and preventCoalman Glacier frorn join- ig Zigzagand White River Glaciers. Lobe 3 appearsto be the youngestarea of the dome.Whether it is a true lobe with exten- TimberlineEruptive Period sivetalus coveror a depositof angularclastic debriscould not be deterrnineddue to snowcover Timberline-agelahars are the mostvoluminous and lack of penetratingexposures. Nevertheless, of the postglacial lahar sequences.They form Lobe3. as it l'ill be calledfor convenience,is the bulk of the valley-fill depositsin the Sandy an areaof grey,vesicular, unaltered, angular rub- and Zigzagbasins and the broad fan that extends ble.Individual clasts average 0.5 m in diameter, from CraterRock to the baseof the nourrarn. but a few iange up to 2 m and all sizescommonly Pyroclasticflow depositsof this age have been showthermally related prismatic fracture pat- found as far as Barlow Campground on the terns.No current fumarolicactivity wasobserved ZlgzagRiver (El2 NEI/4 Secl5 T3S RBE),l3 on this lobe. km from the vent area, and lahar and hyper-

Post-ClacialLahars of the SandyRiver Basin,Mount Hood, 0reqon 227 concentratedrunout depositsas far as the mouth abruptly to bouldery lahars and flood deposits of the SandyRiver at the townof Troutdale,S0 of the upper half of the sequence.Here clasts km from the source area. make up 40 to 70 percent of the flow units. Averageclast size is 20 to 50 cm in diameterwith Ageand Orlgin rare bouldersto 2 m. Individualflows are I to 1.5m thick. Lensesof beddedfluvial sandand entrained Radiocarbondating, of organic debris gravel20 to 50 cm rhick and I to 4 m long are in the lahars,dates the Timberlineflows at 1440 occasionallyfound in the upper half. Weather- (Table + 155 to l?80 t 200 years BP l). ing of the top unit extendsdownrvard for at least in Depositsof Timberline age are found all 40 cm at all sitesmeasured on the fan, wilh 40 drainageson the southwestside of the mountain to 70 cm of weatheringtypical for depositsof this (upper from the SalmonRiver reachesonly) west age in the Sandy River basin (Crandell, 1980). to the Sandy River. The exhusionand mass wasringof Crater Rock producedthe pyroclastic The timberline-agedeposits of the debrisfan flowsand lahars.Prismatically jointed boulders areveneered by sand-and gravel-sizedangular arecommon throughout the lower75 percentof particlesof grey,phyric dacite. This is inferred the sequencebut rare in the upper25 percent, to be air-fall materialfrom smallevents reported in mid l800's (Folsom, suggestingthat lahar generationaccompanied to have occurred the I dome building. 1970).This depositis rarelyover cm thick and lies directly on the ground surface. Areal D stributionand General Stratigraphy Belowthe main debrisfan and aboveOld Maids Flat the SandyRiver flows through a nar- The main debrisfan (Figure4) is boundedon the row gorge cut through the Pliocene volcanic Illumination Ridge, on the eastby the noth by rocks of the SandyGlacier volcano (Wise, 1969) lateral moraine of White River Glacier,on right and the Pleistoceneand Holocenedeposits of the Multorpor Mountain and on the the south by Mount Hood cone.The high gradientand nar- southwestby Laurel Hill. The actualthickness row nature of the gorgeprecludes the accumula- exposuresof over of the depositsis unknown,but tion of more than a traceof deposits. l5 m occur along Highway26 at Laurel Hill (SEl/4 SWI/4 Sec 14 T3S R8E). Exposures At the mouth of the canyon,a wideningvalley measuredfrom aerialphotographs at the head- and decreasinggradient have allowed the forma- cuts of the Sandy River and Rushing Water tion of a debris fan at Old Maids Flat which is (Figure Creekare in excessof 100rn in height.These composedof Timberline-agedeposits 4). thicknessesprobably representfilling of the pre- The fan, coveringan area0.5 km wide by 1.5km eruptionglacial valleys of the headwatersof these long, hasbeen dissected by the SandyRiver and rivers. yieldsexposures up to 60 m high.The deposits of the Old Maids Flat debris fan showa distinct Exposuresat LaurelHill (SEl/4 SWI/4 Sec textural variation similar to that of the main 14 T3S RgE)and in the villageof Government debrisfan. The lowerone-third is a sequenceof Campnext to the fire station(NW1/4 Sec23 T3S sand-rir"hlaharic and flurial bedsranging in R8l/2E), showtwo discretetextural zoneswithin thicknessfrom 1.5to 3 m, whichappear to con- the Timberline deposits.Matrix-rich, sandy tain lessthan l0 percentclasts. The uppertwo- lahars and fluvial deposits constitute approx- thirds of the exposureis composedof intercalated imatelythe lowerone half ofboth outcrops.Cob- cobbleand boulder laharsand fluvial beds,rrith ble or largerclasts make up lessthan l0 percent thicknessesranging from 1.5to 2 m. Clastsmake of theseflow units.(In all subsequentdescrip- up 20 ro 70 percentof the laharsby volume,the tionsof deposits,the term "clast" will be used remainderbeing a matrix of coarselithic sand. to nean lithic fragmentsobviously larger than The fluvial beds are generallythinner than the the matrix material.)Maximum clast size is l0 laharsand are composedof stratifiedsand and to 20 cm in diameter.Thinly laminatedand cross- gravel.Most of the beds are laterallydiscon- beddedfluvial layersup to l0 cm thick inter- tinuous and probably representfill deposits calatewith laharsthat average1.5 to 2 m thick. within shallow,meandering channels on the ag- The sandy depositsof the lower half change grading fan surface.Cut and fill {eaturesare

228 Cameronand Pringle TABLE l. Radioca.bon dates from drainases on rhe southwestflank of Mount Hood.

Locationand Setting

Sandy Rirer near Zigzag, charcoal in duff betweenlahars 200 Crandell, 1980 SEI/4 SEI/4 Sec33 T2S R7E Whire River, te.race near timberline, linb in lahar 185:t120 Caneron, Pringle Nl/2 Sec 8 T3S Rgf, SandyRiver, north bank.2 km NW ot CuardStation, 220r!l50 crandell, r980 Old Maids Flai, charcoal in lahar Sandy River, nargin of Old Maids Flat, exacr location unlnown, 250 crande , l9B0 standing stump buried bl labar White River, upper canyon, standing stump buried by lahar 250a150 Lawrence, I959 NI/2 S€C8 T3S R9E White River near Highway 35 crossins, wood in ptroclastic flow 260a 160 Crandell, 1980 swl/4 sec 16 T3s R9E Zigzag Rfter at Twin Bridges, standing cedar snag killed by lahar 270a 150 Cameron, Pringl€ NWI/4 Sec 15 T3S R8E

Vhite River at Tygh Valley, wood ftagm€nts in lahar 425:r 150 Cameron,Prinele Nr/2 Sec 12 T4S Rt3E "bayonet Sandy River at OId Maids Flat, tr€e" buried by lahar 455a135 Caneron, Prinsle NEI/4 Sec23 T2S R8E Zigzag Rftet near Twin Bridges, sranding stunp buri€d by lahar 550r 130 Cameron,Prinsle NWI/4 Sec 15 T3S R8E Muddl Fork Sandy River at Portase Trail crossing,old Maids 560,r150 Carneron,Pringle Flai, wood fron ash layer b€lo{ upper lahlr

White River at Vhite River Station, charcoal in soil below lahar 1340+160 caneron, Pringle NWI/4 Sec 30 T4S Rlof, Zigzag River at Barlow canpground, charcoal fron upper 1440a155 Caneron, Prinsle pyroclastic flow S[l/4 NEI/4 Sec 15 T3S R8E SandyRiver. bourh banl,.0.2 lm Str of GuardSrar;on, 1530,r200 Crandell, I980 old Maids Flat, los catied in lahar Near GovernmentCamp, along road, charcoal in ash above l6I0a 200 Crandell, 1980 glacial age depositsSli2 Sec l8 T3S R9E Base of Mulrorpor Mountain, log carried in lahar I670ir 200 Crandell, t980 Nl/2 S€c 23 T3S R81/2f, White River, top of lefr lareral noraine cha.coal under ash layer 1690+t40 Cam€ron, Pringle NWl/4 SEl/4 Sec 16 T3S R9E Sandy River near Zig Zag,log carried in lahar 1780r 200 Crandell, I980 SEl/4 SEl/4 SCC33 T2S R?E

While Riv€. at Tygh Valley, charcoal in sands below iahar 2080rlt5 Csmeron, Pringle Nr/2 Sec 12 T4S Rl3[ Sandy River 4 km from Brightwood, wood from lahar 40,000 crand€ll, r980 NEI/4 Sec 22 T2S R6E

Post-GlacialLahars of the Sandy River Basin, Mount Hood, Oreqon 229 TIMBERLIN€ EBL]PTIVEPEAIOD OEPOSITS

ili

'./)/" ) ";r,'r,/'

Iigure 4. Inundation boundaries of the Timberline eruptiye period. common, with some channelsbeing up to 4 m 1.2rn. The uppermostunit of eachexposure is wideand2mdeep. oxidizedto a depth of 40 to 60 cm. Timberline-agelahars fill the Sandy River The powerdiversion dam at Marmot(SEl/4 valley from the OId Maids Flat debris fan NEI/4 Sec l3 T2S RsE) marksrhe transition downstreamto the powerdiversion dam at Mar- from valley filling by Timberline-agelahars to mot, forming a flat valley floor 0.5 to 3 km wide. migration of the lahars along the river with lit- The true thicknessof the depositsis unknown, tle deposition.The river flows through narrow, but well-logs from two geothermal test wells bedrockgorges between river miles 3 and B, and locatedon Old MaidsFlat (NEl/4 SEI/4 Secl5 between15 and 30.(River miles are markedon T2S RBEand SWI/4 SEI/4Sec l5 T2S R8E)list most quadranglesfor this region and this use is thicknessesof the mudflowsequence at B0 and the reasonfor citing the Englishmeasurement 30 m, respectively(Priest and Vogt, 1982). rather than the metric equivalent.)Deposits ExposuresI to 2 km downstreamof rhe Old within thesereaches are limited to point barsand Maids Flat debris fan show only the lower, san- veneerson lower terraces. dy part of the Timberline-agesequence. In- Betweenthe dam and the mourh of the Sandy dividual flows are I to 3 m thick and conrainless River Timberline-age deposits were positively rhan l0 percentcobble or larger clasts.Maximum identified at only four locations:near the town clastsize is I m. Beyond2 km, only the upper, of Troutdale(NWl/4 Sec25 TIN R3E),Oxbow coarserportion of the sequenceis found. The Park (SWl/4 Sec l0 TIS R3E), Dodge Park coarser lahars are generally matrix-suppofied (NEl/4 Sec36 TIS R4E),and at Marmot dam cobbleand boulderlahars with clastscomposing (NEl/4 Sec13 T2S R5E)(Figure l). At eachsite 20 to 70 percentof the unit. Averageclast size a maximum of two lahars was found. Average is 5 to l0 cm with a maximumof B0cm. The flows thicknessofeach lahar deposit was just over0.5 are separaled by fluvial sand and gravel. In- m with a maximumof 1.5n at Troutdale.The dividualflow depositsaverage I m thick with an laharsare generallyseparated by fluvial or hyper- observedrninimum of 0.4 m and a maximumof concentratedrunout deposits. Average clast size

230 Cameronand Pringle is from 2 to 5 cm, with a maximumof l0 cm. exposureof I m or lessin road cuts and drainage AII lahar depositsare matrix-supportedwith ditches(SWl/4 NEl/4 Sec16 T3S RBE,road cut; clastscornposing 30 to 70 percentof the unit. SEI/4 NEI/4 Secl? T3S R8E,road cut; NEI/4 A well"developeddelta occurs at the mouth SEI/4 SeclB T3S RBE,road cut; NEli4 NEI/4 of the SandyRiver which, until the constructron Sec 19 T3S R8E, drainageditch). of a rockfill dam,split the river into two chan- Timberline-agedeposits have been identified nels.The surfaceof the delta is veneeredby only in the upperreaches of the SalmonRiver, modernflood deposits fron both the Sandyand including a well-definedalluvial fan (Figure 4) the ColurnbiaRivers but it is assumedto be cored at the mouthof the upperSalmon River canyon by Timberline-agedeposits. (Sec19, 20, 29, 30 T3S R9E).Numerous mead- As with the SandyRiver, the Zigzag and Lit- ows,ponds and bogsoccupy a large,low-gradient tle ZigzagRivers pass through narrow, deep can- reachof the river immediatelybelow the fan. No yons in their upper reachesthat precludethe ac- phyric rock typesindicative of a Crater Rock curnulationof large deposits.Below the rivers' origin werefound in streambank depositsin the confluence,Timberline-age lahars have filled the canyonbelow the low-gradientreach. Evidently valleyto an unknowndepth, producing a flar the Timberline-agelahars in the SalmonRiver floored valley extendingto the Sandy River wereneither voluminous nor energeticenough valley. to crossthe low-gradientreach below the alluvial fan. Exposuresare rare in the valleybelow the canlons.The riversare not artivelyincising or PhysicalCharacteristics side cutling, and vegetationextends to the Iyater'sedge. A l0 n thick sequenceof cobble Timberline-agelahars, exclusive of the debris and boulderlahars is exposedin a road cut at fans,are generallyinterbedded with fluvial sands the abandonedbridge over the Little Zigzag and gravels, this feature becoming more pro- River (SWl/4 NWI/4 Sec 14 T3S RBE).These nounced as distancefrom the mountain in- flows came down the Little Zigzag River or creases.The fluvial interbedsprobably represent spilledover the low saddledivide from the up- laharmaterial reworked by floodsthat followed per Camp Creekarea of the main debrisfan. shortly after the lahars.Lack of soil horizonsbe- Herethe lower,sandy portion ofthe Timberline- tween the lahar and fluvial material showsthat agesequence is not present.A quarrynear the the time spanbetween the eventswas minimal. HiddenLake trailhead (SEl/4 NEI/4 Secl5 T3S 0bservationsof the lahardeposits of the Toutle R8E)exposes at least8 m of pyroclasticflow and River systemat Mount St. Helensshow that rivers lahar deposits.Two pyroclasticflows are sep- affectedby laharscarry large sedimentloaris un- aratedby a discontinuousgravel layer and overlie til the thalweghas been flushed of sediment.Un- at leastone sandylahar. The upper pyroclastic til then, aggradationof the river bed in low- flow wasdated at 1440 a 155years BP (Table energy reachescan greatly decreasethe carry- 1)which places it nearthe endof the Timberline ing capacityof the river channel,making floods eruptiveperiod. This exposureis in a l0 ro 15 more commonand severe. m high terraceformed along a I km reachbelow The texturalvariation of the Timberline-age the confluence of the Little Zigzag and the deposits,with a lowersandy portion and an up- Zigzag River. It is not known down which river per coarserportion, suggest early, fine-grained' the flovs came.The height of the terraceand producingevents and later, coarseclastlc ma- the thicknessof the pyroclasticflow depositsis terial production.The extrusionof the Crater inconsistentwith the lack of downstreamde- Rock dome might havebeen preceded by ash posits.Either erosion has removed the deposits emissionsand ash-richpyroclastic flows which, or the flows simply did not extend any farther. becauseof the high gradientand availablewater A flattenixg of the valleygradient combined with sourr"ein thpform of snohand glar"ier ice. pro. a constrictionin valleywidth at this point might ducedthe sandylahars that form the lowerpor- havecaused ponding of the flow. tion of the Timberline-agedeposits. Once the The only other known Timberline-agede- domebecame established, rnass wasting or col- positsfound in the Zigzagbasin are in shallow lapse of the growing extrusionwould have

Post-ClacialLahars of the SandyRiver Basin,Mount Hood, Oregon 231 suppliedthe coarseclastics of the upper portions trusiverocks from smallstocks to the westand of the sequence. soulhtteslol lhemounlain. and very s.oria.cous "other" Clastsize generallydecreases downstream. and alteredrocks from ]avaflows. These Averageclast size in the boulderysection of the clastsare assumedto havebeen incorporated into main debrisfan B km from the sourceis 20 to the lahar as it moveddown streamchannels. 50 cm in diameterwith a maximumof 2 m. On Relativepercentages among the rock types 01d Maids FIat, 12 km from the sourceand vary somewhatfrom site to site,which may be downstreamfrom the Old MaidsFlat debrisfan, a functionof the smallsample size (usually 15 thearerage lize is l0 to 20 cm rnitha marimu m to 20 clasts,or all that wouldfit into an easily of I m. NearBrightwood, 40 km fromthe source, carriedsample bag). A few trends,however, are the averageclast size is B to 15cm and the max- distinguishable.Crater Rock-typelithologies imum is 50 cm. At Troutdale,B0 km from the generallymake up about60 percentof the clasts source,the averagesize is I to 3 cm and the max- by number.This indicates thar incorporation of imum is l0 cm. materialduring flow played an importantrole in determiningthe actualvolume of the lahar.The Cenerationof laharsby the extrusionand "other masswasting of CraterRock shouldhave pro- rock types representedin the rocks" duceda clastlithology predominantly the same categoryare found only high on the slopesof the as CraterRock. Samples of averagesized clasts mountainand not from the formationsexposed weretaken fiom threesites along the SandyRiver in the valley walls away from the cone, such as and the clastsdirided into lithologirgroups olivinebasalt of the SandyRiver GlacierVolcano, "hand basedupon lens" petrology.Sample sites alteredbasalt of the MioceneRhododendron For- \r'erenear Brightwood(SWl/4 NWI/4 Sec22 T2S mation,and flowsof the YakimaBasalt Subgroup R6E,two laharssampled), Marmot Dam (NWl/4 of the MioceneColumbia River BasaltCroup. In- NEI/4 Secl3 T2S R5E,one lahar sampled) and corporationof exoticclasts must have occurred OxbowPark (SWl/4 SWI/4 Sec l0 TIS R4E,one primarily high on rhe coneof the mountain.This lahar sampled).The sites are 25, 30 and 45 km premiseis supportedb1 personalobservalions by riverfrom the source area on lhe mounlain. at MountSt. Helensof recent,small scale lahars that showedthe laharsto be erosionalonly in the Clastswere divided into threedominant rock early stagesof flow and to rapidly become types: grey-matrixphyric dacite, red-matrix depositional. phyric dacite,and "other rocks." Both phyric daciterock typesare similarin grosspetrology EstirnatedDepth of Timberline-ageLahars to lithologic types found on Crater Rock. DuringFlow Phenocrystscornprise 20 to 40 percentof the Depthof the laharsduring peak flow cannotbe dacites.with an aphanilicor phancriticmalrix. determinedfrom the height of depositsabove Phenocrystsare predominantly(80 percent) river level,but personalobservations of lahars euhedralplagioclase I to 3 mm acrosswith ac- and their depositsat Mount St. Helensprovides cessorysubhedral amphibole 0.5 to 3 mm across. a basisfor someinterpretations. On the Toutle The matrix varies from being truly aphanitic to River and its North Fork, which resemblethe possessingdiscernible microlites of plagioclase SandyRiver below the 0ld MaidsFlat debris fan set in an aphaniticgroundmass. Oxidation, which in channeland valley morphology, the heightof givesa reddishcolor to someof the clasts,is con- the lahar depositsabove river level variesbe- fined to the aphaniticportions of the matrix with tween20 and?0 percentof the true depthof the neithermicrolites nor phenocrystsshowing any lahar at peakflow asdetermined by mudlinesand alteration.At the domeon MountSt. Helens this trimlines.In constrictedreaches, deposits ranged type of matrix oxidationhas at timesoccurred from 20 to 40 percent of the peak flow height. within rnonthsof the extrusionof a lobe.Angular In moreopen reaches, deposits ranged from 50 voidsless than I mm acrossform up to 5 per- to 70 percentof the peakflow height.In very cent of the volumeof someof the clasts. narow, steepreaches, such as in the headwaters The "other rock" categoryincludes all rock of the SouthFork of the Toutle and in Sheep typesother than the phyricdacite. Included are Canyon,the heightof the depositsis lessthan tabular to rounded clastsof aphyric andesite,in- I to 10 percentof the peakflow height.

232 Cameron and Pringle Depositheight on the SandyRiver on Old beiowits confluencewith the Little ZigzagRiver MaidsFlat is 7.3m abovewater level; at the con- and veneer a terrace in the Timberline-age fluencewith the ZigzagRiver, 8.5 m; at Sleepy depositsof the Old Maid debrisfan (Figure5). Hollowbridge, 8.5 m; at OxbowPark, 7.0 m; and The Zigzagbasin terrace is at leastI km long at Troutdale,5.8 m. All of thesesites are in open and averages100 m wide.It mergeswith the val- reaches;therefore, it can be conservativelyesti- ley wall at its upstreamend and terminates matedthat peakflow heightsof the laharsap- abruptlydownstream in a seriesoferosional ter- proachedone and a half timesthe heightof the races.The valleynarrows considerably at this deposits.This wouldproduce a peakstage 9 m point and may havecaused ponding of the flows, present abovethe river level at Troutdaleand creatingthe thick, localizeddeposits and may ex, l0 to 12 m at the upstreamsites. plain the lack of any exposuresdownstream. It is in this reachthat the thick but not laterally ZigzagEruptive Period extensiveTimberline-age pyroclastic flows are ponded Depositsfrom a previouslyundescribed sequence found.They, too, may have behindthis of laharsand associatedfluvial deposits forrn an vqllpv nnnstrietinn I rn high terrace along a single reach of the Theterrace is cornposedof a lower,bouldery ZigzagRiver just belowits confluencewith the deposit3 to 4 m thick that is massiveand made Little Zigzag River (Figure 5) and caps part of up of roundedclasts up to 0,5 m in diameterset a 30 m high erosionalterrace in the Old Maid in a sandymatrix. Portionsof the depositare debrisfan. This sequencehas the lowestvolume clast-supported.The flowsengulfed a forestof and extentof the threemajor post-glacial lahar matureDouglas fir treesup to 0.5m in diameter sequences.Within this paperit will be referred but left the bark intact on the upstreamsides and to as the Zigzageruptive period. undamagedbrushy branches wrapped around the trunks.The rootedbases of someof thesestumps Age and Orign are at the modernriver leveland showthat the ZigzagRiver hasreturned to the baselevel which Along the ZigzagRiver (Nl/2 Sec15 T3S RBE) existedbefore the formationof the terrace. the flowsburied a stream-bankforest of mature Douglasfir, and standingtrunks still enistthat Overlyingthe boulderydeposit are 2 to 3 m of areat least2 m tall. The bark from oneof these intercalatedlahar, hyperconcentrated runout and stumpshas beendated at 550t 130years BP fluvial sanddeposits. These units are generally (Table l). This date correspondswell with the 560 lessthan I m thick and are not laterallytraceable. + 150years BP ageof woodfrom a tephralayer They representthin flows which vashedover the in the upperSandy River basin and suggestsan flat surfaceof the fill material immediatelyafter eruptionwhich produced both lahars and air fall ils emplacement.The thin units are toppedby material.On the Old Maid debrisfan I km down- l0 cm of soil and forestduff rhich is in turn streamfrom the UpperSandy River GuardSta- overlainby a lahar of Old Maid-age. tion(NEl/4 Sec23 T2SRBE) a singlesandy, cob- The depositon the Old Maid debrisfan ter- ble lahar capsan erosionalterace 30 m above raceis 0.9 m thick and representsa single,sandy presentriver level.The flow overranand charred cobblelahar. The height of the terraceabove a standof youngDouglas fir trees5 to l0 crn river levelat the time of the laharis unknown, in diameter.These trees are alignedin down- streamorientations similar to the "bayonet" but it is 30 m abovernodern levels. The deposit treesflattened by the laharsat MountSt. Helens nowsupports a Douglasfir forestof greaterden- (Jandaand others, 1980). Wood from thesetrees sity and sizethan elsewhereon the Old Maids has beendated at 455 a 135 yearsBP. Flat surface.This vegetativedifference is read- ily apparenton air photosand showsthat the depositis limited to a terrace0.5 km long and Area Extent and Stratigraphy 100 m wide.No other evidenceof this flow is Laharicdeposits of this age are restrictedto a found in the Sandybasin. terraceon the right bank of the ZigzagRiver

PostGlacialLahars of the SandyRiver Basin,Mount Hood, Oregon 233 oLD r!1AlOANO ZTCZAC EI]UP] IVE IE R OD OEPOSTS

"'"" N,;":.: 7Zl'*'"*" E"',,",".,.

L] IiT

Figure 5. Inundation boundaries of the Zigzag and Old Maid eruptive periods and the location o{ two Old Maid age lakes.

Old Maid Eruptive Period in hisjournal (Thvaites, 1959), reproduced here with originalspellings: More than 20 exposuresof Old Maid agelahar ". depositshave been locatedin the Sandyand . . I arrived al the €nterenceof a river which appearedto Zigzag Ri,rer drainages.Although a successron scatterove. a Sa.d bar,the bottom otwhich I couldsee quil of Old Maid age laharsis found in the White acrossand did no1sppear to be 4Inches deeprn any pan; Rivercanyon on the southside of Mount Hood, I attempted to wade this Srream and to my astonrsnmenr only a singlelahar unit hasbeen identified in the found lhe boltom a quick Sand,and impassable.. . this rirel \{hich found Sandyand Zigzagbasins in any one exposure. 11e to be a verry considerableStrean Discharg- ing ils saters through 2 chanelswhich forms an Island of Old Maid agedeposits essentially form a veneer about3 niles in lengthon thc rivcr and I li2 miles{ide, on the surfaceof the voluminousTimberline-age composedof corse $nd which is lhrown out of ihis quick depositsand orr the Zigzag-agedeposits along sandriver comprcssingthe {aters ot the Columbiaand lhrow- the Zigzag River (Figure 5). No depositsof Old ing lhc wholecurrenl of its watersagainst its Northern Maid agewere identified in the Little Zigzagand banks.. . Thisstrean has much the appearance ofthe River SalmonRivers or in Still and Camp Creeks. Platt, roleing its quick sands into thc botloms {ilh greal lelocity after {hicb it is divided inlo 2 chanelsby a large Age and Or g n sandbar beforenentioned, the na.ro{estpart olthis River ". is 120yards." (3 Noy. 1805) .. aborerh€ poinrwhich il Radiocarbondating of wood samplessuggests dirided itselfinto two chanels,it is alrout 300yards wide rho' that the Old Maid eruptiveperiod occurred be- the chanel is not more than 50 yards, and only 6 {eet deep. tween200 and 300 years BP (Tablel). Historical the o!her part of lh€ rirer from 2 to I inches{ater, the b€ad accountsand dendrochronologicalwork extend of this river is forned entirely of quick sa.di i1s banks are the lower limit to about lB0 yearsBP. low and at present overflown.the sater is turbed and cur- rent rapid." (1 April In Novemberof lB05 and April of 1806,Lewis 1806) and Clark exploreda river they namedQuick- The name of this river was shortened in later sandRiver. CaptainClark describedit asfollows yearsto the SandyRiver. The SandyRiver is now

234 Cameronand Pringle a boulder-armoredriver and the channelim- weatheringprovide evidence for a recentorigin. mediatelyabove the delta is at least3 to 4 m The close proximity of the lobe to Coalman deep.Modern fall and springfloods do not gen- Clacier(Figure 3) and the normalsnowpack ex- erallyovertop its banks.The expeditionappears pectedat this elevationwould provide the water to havearrived while the river wasstill flushing neededfor lahargeneration through melting of the sedimentof the Old Maid eventthrough the the snowand glacialice. system.Taking into accountthe rapidity with A depositof greytephra mantles the normally which someof the streamssurrounding Mount light coloredmaterial of the summitridge. From St,Helens have recovered from damageinduced its shape,the deposithad its origin at Craler by the l9B0 eruption(staff members of the Cas- Rock.The tephrahas a rnaximumobserved thick- cadesVolcano Observatorv, personal communtca- nessof 20 cm at the summitand exhibitssome tions, 1984),it is possiblethat the Lewis and stratification that may be due to rer.'orkingby Clark Expeditionarrived within 5 to l0 years wind.Breadcrust bombs are common throughout after a laharicevent. the deposit,averaging l5 cm acrossbut with oc" Dendrochronologicsamples were taken in the casionalbombs up to 0.8 m. The actualage of vicinityof 0ld MaidsFlat in an attemptto refine this ashis not known,but the fact that the loose the agerange established by radiocarbonanalysis ashis still quitevisible in a verywindy location and suggestedby historicalaccounts. At Mount and that the highly fractured surfacesof the St. Helens,trees at the peripheryof the May 18, breadcrustbombs are sharpand unalteredby the 1980mudflows commonly had largepatches of freeze-thawcycle requires a relativelyyoung age. bark removedon their upstreamsides by abra- The Old Maid eruptiveperiod is the last of the sionbut managedto suruive.It is assumedthat major eruptionsat Mount Hood; the ash and damageof this magnitudewould affect the bombdeposit may be contemporaneouswith it. growth rate of the tree and causean abrupt nar- rowing of the annualrings. Areal Distribution About20 old growthfir andcedar trees were The highestirrefutable exposure of Old Maid age cored around the edgesof Old Maids Flat. depositsin the SandyRiver basin was found Severaltrees near the confluenceof the Muddy about two km downstreamfrom the Upper Sandy Forl andthe Sandr Rirer show a consistenlnar- GuardStation. At thispoint the laharceased be- rowingof ringsbetween 1760 A.D. and 1805A.D. ing channelizedand beganoverbank deposition with a pos"ibledrastic thin sequencestarting acrossmost of the surfaceof Old MaidsFlat. A about 1794A.D. and lastingfor an averageof possibleOld Maid age depositis found in the l0 years.The ensuingrings alsoshow the pres- veryhigh-gradient upper and middlereaches of enceof traumaticresin canals, which have been RushingWater Creek. This grey,sandy deposit associatedwith the recoveryresponse of a tree is found in patchesalong lov ridge crestsand to physicalinjury. Treeshigher on the hillslopes terraces. do not showthe narrowingor the concentratton In the 01dMaids Flat area,the Old Maidage of resincanals. These preliminary findings sug- Iaharforms a veneerfrom 0.2 to I m thick on gestthat a tree-damagingevent, possibly a lahar, pre-existingterraces and over mostof the sur- occurredaround the year1794 A.D. The narrow face of the Flat. Most of the exposedsections rings starting in 1760 A.D. may indicaterhe rangebetween 0.5 and 0.8 m thick.Deposits up beginningof ashfalls in the basinwhich slowed to 3 m thick arefound along the streamchannel the growthrate of the trees.Ring countsin l9B4 of the of the SandyRiver down- on stumpsrooted in the Old Maid-agesurface streamfrom the PortageTrail crossing(E1/2 Sec shorvmaximum ages of 173years in the Sandy 15 T2S RBE).The preiahar channelsof the Mud- basinand 190years in the Zigzagbasin, or l8l I dy Fork and RamonaCreek were incised into the A.D. and 1794A.D., respectively. Timberline agesurface of the Flat and lined with The formationof lobe3 of CraterRock is a old grolvthcedar trees up to 2 m in diameter. possiblemethod of generationfor the 0ld Maid The Old Maid agelahar filled thesechannels and lahars.The very angularnature of the lobema- adjacentlow areas,killing the trees.Their re- terial and its completelack of oxidationand mainsare still apparentin the form of treervells

Post-GlacialLahars of the SandyRiver Basin,Mount Hood, Oregon 235 and standing snags up to 30 m tall. Inferred dale are up to 6 m abovemodern water level, depthsof the pre-laharchannels of 1.6to 6.5 m much higher than normal flood deposits,and are wereobtained by soundingtree wellsand the in- atleast3mthick. teriorsof hollowsnags. Along the ZigzagRiver depositsof an Old None of the exposuresof the Old Maid age Maid agelahar mantlethe terraceof the Zigzag- lahar within the Sandy River drainageprovides age lahar sequenceto an averagedepth of about evidenceof multiple flows. Apparent stratifica- I m and a maximumdepth of l.B m. Thissandy- tion in a depositforming the north bank of Mud- rnatrixcobble lahar contains clasts up to 30 cm dy Fork 100m downstreamfrom the crossingof with an averageclast size of 5 to l0 cm. As in the PortageTrail may reflectindividual surges the 01d Maids Flat area,the lahar buried and of pulsesas a single flow spilled into and flowed killed a maturecedar forest, leaving standing down the old channel.A thin (2 to 6 cm) layer snagsup to 5 m tall. Thesesnags have been dated of moderatelywell sortedsands, possibly the sole at 270 t 150years BP. This date is from the or basallayer (Kevin Scott,personal communlca- outermostlayer of woodfrom one of thesesnags, tion, 1983),underlie. the aboveunil. but the bark and an unknownnumber of annual At leastone and possiblytwo temporarylakes ringshave been lost to decay.The actualage of wereformed along the marginsof Old Maids Flat the lahar is thereforeyounger by an undeter- through dammingof tributaries by the Old Maid mtneoamouu. agelahar (Figure5). The largestand bestdefined Downstreamfrom the terrace,where the of the temporary lakes formed at the mouth of Zigzag River valley broadensand flattens,there the Muddy Fork valley where it joins the main is an obviousvegetation zone very different from SandyRiver valley(Sec 14 and 15 T2S RBE). the surroundingforest. Unlike the surrounding Over I m of lacustrinedeposits in the form of flora of Douglas fir, ferns, rhododendronsand interbeddedsand, silt and organic debris in- huckleberry,this zone is sparselycovered by dicatesthat the lake existedfor at leasta few stuntedpines l0 to l5 crnin diameterwith little years. The fine-grainedsediment supports a or no undergrowlh.This typeof regelalionis densegrowth of deciduoustrees, mostly alders, very sinilar to that found growing on the Old very different frorn the dominant fir coverof the Maid-agelahar sulface on Old MaidsFlat. It is surroundingregion, making the extentofthe lake theresult oI la,"hoI soilformation owing to ils readilyapparent. This broad-leafforest covers youth and the relatively dry subsurfacecondi- about 0.3 kmZ. The valleyfloor of the Muddy tions causedby drainagethrough the sandylahar "dry" Fork was carpetedby a mature cedar forest deposit.This vegetationzone extends for before the formation of the lake and over a hun- about I km onto the flat areaof the ZigzagRiver dred 30 to 50 m high snagsdrowned by rhe im- valley(Sec l7 T3S RBE)and probablyrepresents poundedwater are found within the lake deposit rhe path of the distalportion of the Old Maid boundaries, age lahar. The secondlake, at the mouthof LostCreek Ph'/c..1 ah.r2.tarieri.c (Sec21 T2S RBE),is not nearlyas well defined asthe Muddy Fork lakebed. lls e\islenceis sug- Although only a single lahar is recordedin gestedby sandand silt depositsthat appearto depositsalong the Sandyand ZigzagRivers, it be of lacustrinerather than fluvial origin. Lost wasnot necessarilythe samelahar in both rivers. Creek is much steeperin gradient than the Radiometricdating ofwood from the laharsgives MuddyFork so the lakewould have covered less about a 50-yearrange for emplacement(250 and area. No snagsrvere found in the area but logs 220 + 150years BP for the SandyRiver deposits werefound buried in the lacustrinedeposits. and 270 t 150years BP for the ZigzagRiver No Old Maid-agelahar depositswere posi- deposits)and the dendrochronologicand histor- tively identifiedbelow Marmot Dam. However, ical materialextends that range to 90 to 100 thick sandbeds found near Dodge Park,0xbow years.It is possiblethat multipleOld Maid-age Park and nearTroutdale may be terracesformed laharswere produced over a spanof time andthat by a lahar runout or postJahar flow with high only two, alongseparate rivers, produced over- sedimentconcentration. The depositsat Trout- bank depositswhich have survived to the present.

236 Cameron and Pringle The smallersize and exrentof the Old Maid probably1.5 to 2 timesthe depositthickness. In age lahar (comparedto the Timberline age virtually all locationsthis wouldgive a flow depth Iahars)produces a correspondingsmaller clast of 0.8 to 1.5 m abovethe terracesurface. The sizeand range.Maximum clast size in the Old bulk of the flow remainedwithin the channel.The Maid agelahar is 0.5 m in the upperreaches of flat-toppednature of the terracesallowed the flow Old MaidsFlat and 0.3 m near Brightwood,at to travelup to 0.5 km beyondthe channelin some the distalportion of the lahar.Average clast size placeson Old Maids Flat. for the tuo locations is 0.2 and 0.1 m, respectively. Summary Clastlithology was subdivided into 3 groups During the last 10,000years, at least3 major "hand ba.edon lens" petrologl:grey matri\ eruptiveperiods have occurred at Mount Hood. "other dacite,black matrix daciteand rocks." The oldest, and volumetricallylargest, the Both typesof dacite are aphanitic-porphyritic Timberlineeruptive period, occurred between with phenocrystsof plagioclaseand amphibole. 1400and 1800years BP. The Zigzageruptive Maximumphenocryst size is 3 to 5 mm for the period,volumetrically the smallestof the three, plagioclaseand bimodal I to 2 and 3 to 5 mm occurredbetween 400 and 600 yearsBP. The Old for the amphiboleneedles. The grey matrix Maid eruptiveperiod occurred between 180 and daciteis generallymore vesicular vith up to l0 270years BP andits aftereffectsmay have been percentvoid space,some of the voids being witnessedby early explorers. plasticpull-aparts with numerousglass needles. Greymatrix dacitemakes up 30 percentof the The combinedclastic debris from these clasts by number. The black matrix dacite periodsburied the southwestflank of the moun- generallyhas 0 to 5 percentangular void space tain, producing a smooth debris fan. Lahars and makesup 65 percentof the clastsby nurnber. travelledat least90 km and pyroclasticflows at Ciastswith poorly developedbreadcrust or radi- least13 krn from their sourcearea near Crater ally fractured textures are common in both da- Rock.Flow-depths of the laharsprobably aver- cites.The remaining5 percentof the clastsare agedB to 12 m in most cases. classedas "other" rock typesand includered matrix daciteand highly alteredrock typesin- Acknowledgments corporatedinto the laharduring flow, probably The authorswould like to thank the staff of the as it passedover the upperreaches of the main CascadesVolcano Observatory Tirnberline-agedebris fan. and Dr. Paul Hammondof Portland StateUniversity for their critical reviewof both the manuscriptand the Depth of Old Mad age Lahar During Flow ideasthat wentinto it. Thanksgo alsoto David All Old Maid agelahar depositsare veneerson Yamaguchifor his help with the intricaciesof pre-existingterraces and representopen-reach- dendrochronology,and to Ellen Cameronfor her type deposition.Actual peak-flowheight was editorialexpertise.

LiteratureCited Crandell,D. R., 1980.Recenl eruptire hislory ofMount Hood, Oregon,and potential hazardsfrom future eruplions:U.S. Ceol. Surlev Bull. I492. Folson,M. M. 1970.Volcanic eruptions:lbe pioneers'allilude on the PacificCoast fron 1800to l8?5.The ore Bin 32:61,71. Janda,R. J., K. M. Scort,K. M. Nolan,and H. A. Martinson.1980. Lahar movenent, effects, and deposits: 1z The 1980eruprion of trlount St. Helens. U.S. Ceol. Survey Profesional Pap. 1250. Pp. 461-478. Priesr,G. R., and B. F. Vogt (eds.)1982. Ceologr and geothernalresources of rhe MountHood area,Oregon. 0regon State Departmentof Geologrand Mineral Industies,Special Pap. 14. Thwaites, R. G. (ed.) 1959. Original Journals of the Lewis and Clark Expedition. Anriquarian P.ess, 1t02 p. Wise, W. S. 1969.Ccology and petrology of the Mount Hood area: a sludy in High Cascadesvolcanism. Geol. Soc. Amer. Bull. 80r969-1006.

Receiredl0 Octoberl9B5 Acceptedfor publicat;onl7 January19B6

Post-GlacialLahars of the SandyRiver Basin,Mount Hood, Oregon 237