Bull Volcanol (1993) 55:176-189

Volc a ology 9 Springer-Verlag1993

Controls on accumulation of a volcaniclastic fan, Ruapehu composite volcano, New Zealand Beth A Palmer 1, Andrew M Purves 2 and Sue L D o n o g h u e ~

Department of Soil Science, MasseyUniversity, Palmerston North, New Zealand Received September 15, 1991/Accepted July 22, 1992 Abstract. The Whangaehu fan is the youngest sedimen-

tary component on the eastern ring plain surrounding Ruapehu volcano. Fan history comprises constructional (830--200 years BP) and dissectional (<200 years BP) phases. The constructional phase includes four aggradational periods associated with both syneruptive and inter-eruptive behavior. All four aggradational periods began when deposition by large lahars changed flow conditions on the fan from channelized to unchannelized. Subsequent behavior was a function of the rate of sediment influx to the fan. The rate of sediment influx, in turn, was controlled by frequency and magnitude of volcanic eruptions, short-term climate change, and the amount of sediment stored on the volcano flanks. Fanwide aggradation occurred when rates of sediment influx and deposition on the fan were high enough to maintain unchannelized flow conditions on the fan surface. Maintenance of an undissected surface required sedimentation from frequent and large lahars that prevented major dissection between events. These conditions were best met during major eruptive episodes when high frequency and magnitude eruptions blanketed the volcano flanks with tephra and rates of lahar initiation were high. During major eruptive episodes, volcanism is the primary control on sedimentation. Climatic variations do not influence sediment accumulation. Local aggradation occurred when lahars were too small to maintain unchannelized flow across the entire fan. In this case, only the major channel system received much sediment following the deposition from the initial lahar. This localized aggradation occurred if (1) the sediment reservoir on the flank was large enough for floods to bulk into debris flows and (2) sedimentation events were frequent enough to maintain sediment supply to only some parts of the fan. These conditions were met during Present addresses:

1 Department of Geology and Geol. Engineering University of Idaho, Moscow, ID 83843, USA 2 Wellington Regional Council, P. O. Box 41, Masterton, New Zealand 3 p. O. Box 7168, Te Ngae, Rotorua, New Zealand

both minor eruptive and inter-eruptive episodes. In both cases, a large sediment reservoir remained on the volcano flanks from previous major eruptive intervals. Periods of increased storm activity produced floods that bulked to relatively small debris flows. When the sediment reservoir was depleted, the fan entered the present dissectional phase. Syneruptive and noneruptive lahars are mostly channelized and sediment bypasses the fan. Fan deposits are rapidly reworked. This is the present case at Ruapehu, even though the volcano is in a minor eruptive episode and the climate favors generation of intense storm floods.

Introduction

Recent studies of reworked volcaniclastic deposits associated with stratovolcanoes show that the immediate sedimentary response to volcanic eruptions is aggradation within transport systems that carry sediment away from the volcano (Healy et al. 1978; Janda et al, 1981; Vessel and Davies 1981; Rodolfo 1989; Arguden and Rodolfo 1990). Syneruptive aggradational accumulations are characterized by stacked debris-flow and fluvial deposits representing high rates of lahar initiation and sediment transport during eruptive periods (Smith 1987a, 1987b, 1991; Palmer and Walton 1990; Waresback and Turbeville 1990; Palmer 1991). Inter-eruptive sedimentation is characterized by an initial period of dissection or landscape stability. Such studies document the control of volcanic activity on sedimentation. Additionally, the volcaniclastic systems described in many of these studies display, or are interpreted to display, a relatively straightforward resporise to major eruptions: aggradation during eruptions and initial dissection during quiescence. Inter-eruptive behavior varies, reflecting the action of a diverse suite of sedimentation controls. In this paper, we describe a volcaniclastic system with a complex response to volcanism. The system is the Whangaehu fan, a modern volcanogenic fan on the eastern ring plain of Ruapehu volca-

177 I

.

45

Rangipo Desert 0 I

Tongariro National Park

2 km I

~

t

Lake 10

I

@ 19

%0

/ 1100y" 0

100 km (

~.,a~o

Volcanic Zone

I

i0

2

\

g Ruapehu ~ "~ /

............ / / Wellington ~ , /

9 Measured Section 9 Check Point

k 00

,=.r0 Contours in Meters

Fig. 1. Map of the lower eastern flank of Ruapehu volcano and the Whangaehu fan showing location of data and cross sections. Check points are places where exposures were photographed and correlated. In many places sections are marked as a unit because

they are too closely spaced to mark individually. Two important locations on the fan are the 'chute' and the major gully to the north, 'Scorpion gully'

no, North Island, New Zealand (Fig. 1). As in other volcanic settings, deposits of the Whangaehu fan record a history marked by discrete periods of aggradation. We show that fan growth occurred during both syneruptive and inter-eruptive periods and that different sedimentation styles resulted from variations in eruption frequency and magnitude, climate, sediment storage on the volcano, and the degree of dissection of the fan.

T a u p e Volcanic Zone (Fig. 1). Ruapehu comprises a volcanic cone of mostly sheet lava and autobreccia surrounded by a ring plain of reworked volcaniclastic deposits (Hackett and Houghton 1989). The cone itself is 2797 m high with the active vent, Crater Lake, on the southern part of a broad summit plateau (Fig. 1). The Whangaehu River is the major outlet of the lake. Eruptions through Crater Lake, and collapse of the ash barrier a t the outlet generate lahars in the Wangaehu catchment (Healy 1954; Nairn et aL 1979). The eastern ring plain comprises fluvial, debris-avalanche, and laharic deposits interbedded with tephras from Ruapehu and volcanic centers to the north. A major stratigraphic marker on the eastern ring plain is the

Ruapehu volcano Ruapehu volcano is one of the predominantly andesitic composite volcanoes at the southernmost end of the

178

Fig. 2. Photograph of the Whangaehu fan. Arrows, from north to south (left to right), mark the Whangaehu River, 'Scorpion gully', and the 'chute'. Scale is 1 cm = 270 m. Photograph courtesy of the New Zealand Geological Survey

rhyolitic Taupo Pumice Formation, erupted from the Taupo Volcanic Centre ca. 1800 years Be (Wilson and Walker 1985). The Taupo Pumice Formation, by chance, marks an important change from fluvial to lahar-dominated sedimentation on the eastern ring plain. This change in sedimentation represents initiation of the vent on the southern summit of Ruapehu and development of Crater Lake (Donoghue et al. 1988; Palmer 1990). The post-Taupo deposits are interbedded with a sequence of at least 18 tephras that we call the Tufa Trig formation (informal designation). These are dark gray to black, coarse ash to lapilli-grade tephras erupted from Ruapehu volcano between ca. 1800 years Be and the present. Radiocarbon dating shows that most tephras are younger than ca. 830 years BP.

Methods The study area is in the Rangipo Desert on the eastern ring plain of Ruapehu volcano (Figs. 1 and 2). The stratigraphic framework of the study is 66 sections measured mostly along cut banks of modern channels and pits dug into the fan surface (Fig. 1). Army bomb craters also provided some exposures. Correlation among sections is based on (1) tephra stratigraphy, (2) identification of diamicton marker beds, and (3) direct correlation. Marker beds are diamictons with distinctive clast assemblages. The best marker is a sequence of diamictons containing a large percentage of scoriaceous clasts (Figs. 3, 4, and 5).

Whangaehu fan The Wangaehu fan is the dominant landscape feature on the eastern ring plain (Fig. 2). The fan forms where the Whangaehu River flows from a deeply incised flank valley onto the relatively unconfined ring plain. The Whangaehu fan is 6.0 km long an 4.5 km wide (Fig. 1; maximum dimensions). Gradient of the fan changes from 0.04 m / m at the head of the fan to 0.03 m / m at the distal margin where a fault scarp (40-60 m high) truncates the fan (Figs. 1 and 2). The Whangaehu fan is dissected by two major channels: the Whangaehu River to the north, and a southern channel we call the 'chute' (Figs. 1 and 2). During normal discharge, the Whangaehu River is confined to a single channel. During floods, however, the braided pattern of the river becomes apparent. The number of channels increases downfan to the fault scarp, where flow from different channels is recombined and diverted southward. The 'chute', in contrast, is remarkably straight and deeply dissected. Between the apex and intersection point at midfan, the channel is over 15 m deep and 5 0 - 8 0 m wide (Fig. 2). Two braided channels emerge from the incised area (Figs. 1 and 2). We call the northern channel system 'Scorpion gully'. Like the Whangaehu, the number of channels in the 'chute' system increases downfan. This channel pattern differs from classic mountainfront alluvial fans where only the head of the fan is incised and flow becomes unchannelized downfan. Construction of the Whangaehu fan, however, is from lahars and fluvial floods that are often too large to be confined to the channels. These large events produce

179 18

15

19

L'~"

22

8

L4 (~o P3 ~.

i-' - ~

I~'

9 ~

21

20

P3

Hff

".,

/

\

~;."

.~..o o.

Io;;

9

;@31~o/ ~ - \ ~1/~

'2

"~ .

.

"2

\

i P2

\

2 2

3

~ ~

~

~-_-'--i-~-~

' "-~-~-~"---~i

4

5

I --"--~

|

L~

. .-'~--~---~

~~

that can be correlated across the fan. The datum is the surface of the fan. Hff is hyperconcentrated flow

oj-

~://

"~--~-

C5~

are packages of sediment

~

9

8

6

~--e ~ ...

\

Fig. 3. Cross section along axis of a major channel, Sc marks the interval of scoria-bearing diamictons. Only the major diamictons L1-L4 are correlated. P1-P3

7

~,~.-~L4

/

J

, _.>...__

~olianSand

9

I

\

I..--.I

I.'~.1

2<:

P,

9 TaupoPumice

Fig. 4. Cross section transverse to flow through the southern part of the fan. Sections are the distal-most exposures. Abrupt changes in thickness mark the position of major channels. Tufa Trig members are numbered and not to scale

14

,.~_ /

.........

~ ~ / ~~5 , \~~

~

,, ",,

~,,,

15

J

~

16

.

'..

~'.,

NIL~/~~~ 'd~ ~~

K

<

~

. .

Fig. 5. Cross section transverse to flow across the medial fan. Same symbols and explanation as Fig. 4

P1

180 patterns of flow divergence and loss of channelization that characterize alluvial fans. The Whangaehu River and the 'chute' are the m a j o r transport paths in the m o d e r n sediment dispersal system. These channels are the primary routes taken by lahars as they cross the fan. Flow in the minor channels depends on the size of the flood events. Overbank flow f r o m the Whangaehu channel delivers sediment to the northern part of the fan. When flood depths in Whangaehu River exceed 2 0 m , flow is diverted into the 'chute' and sediment is distributed along the southern part of the fan (Fig. 1).

Lithofacies and processes

Lahars are the most important sedimentation events on the Whangaehu fan. Lahars are rapidly flowing mixtures of rock debris and water (other than normal stream floods) f r o m a volcano (Smith and Fritz 1989). Individual events can include b o t h debris-flow and hyperconcentrated-flow processes (Pierson and Scott 1985; Scott 1988; Rodolfo 1989). Debris-flow deposits dominate the Whangaehu fan (Dcm-Dmg, Table 1). Deposits of hyperconcentrated flow (Hff, Table 1) are more abundant 4-40 k m downstream in the valley of the Whangaehu River. Deposits of fluvial and aeolian processes are also important components of the depositonal framework.

Table 1. Abundance of lithofacies Related to lahars*

Lithofacies

Diamictons Dcm Dci Dram Dmi Drag Sand Hff

% Abundance

Number of beds

Thickness of beds

19 5 34 2 38

36 8 30 3 22

2

1

* Diamicton lithofacies after Shultz (1984). See Table 2 for descriptions

Table 2. Description of diamicton lithofacies of the Whangaehu

fan Lithofacies

Description

Thickness m

Dcm

Dci

Debris-flow deposits Debris flows produce five diamicton lithofacies that differ in abundance of clasts and style o f grading (Table 2). O f these five, lithofacies Dcm, Dram, and Drag are most abundant (Table 1; Fig. 6). Lithofacies Dcm is mostly confined to channels, but is also found in overbank areas. Conversely, lithofacies D m m and Dmg are most c o m m o n in overbank settings, but also fill channels. Thick, coarse-grained deposits of Dcm and D m m represent deposition f r o m large, high-competence debris flows. These large flows overtopped channels and spread blankets of coarse-grained diamicton over large areas of the fan surface. Boulders in these deposits f o r m linear clusters in areas adjacent to large channels. The boulder clusters are coarse sediment stranded as debris flows flooded overbank areas (Rodolfo 1989). Thin deposits of D m m and D m g represent deposition f r o m small, more dilute debris flows. Lithofacies H f f (Table 3) is most abundant in minor channels and overbank settings where it is associated with lithofacies Dram and Drag. Lithofacies H f f represents transition to hyperconcentrated flow along the margins of debris flows. LithofaCies and grain-size change across the fan is a function of the size of a debris flow. Deposits of large debris flows display no regular change in lithofacies and little change in clast size with increasing distance f r o m source (Figs. 3 :and 7). These relationships suggest that the debris flows were too large to develop changes in

% Abundance

Dram

Dmi Dmg

Clast-supported, ungraded. Some with basal matrix-supported inversely graded zone, rarely more than 0.2 m thick. Poorly sorted matrix, silt + sand + granules. Matrix in some has clay. Clasts to 3 m. Mostly lenticular with scoured base. Some tabular with non-scoured base. Clast-supported, inversely graded. Basal matrix-supported zone. Sandy matrix as in Dcm. Clasts to 1.5 m. Lenticular beds with scoured base or tabular beds with nonscoured base. Matrix-supported, ungraded. Some with basal inyersely graded zone. Sandy and muddy matrix as Dcm. Clasts to 1.2 m. Mostly tabular with non-erosive base. Some lenticular deposits. Matrix-supported, inversely graded. Sandy matrix as in Dcm. Clasts to 0.6 m. Tabular beds with non-scoured base. Matrix supported, variable styles of grading. 20% of have only a thin inversely graded base, 7~ inversely graded, 7% normally graded, and 4~ inverse to normally graded. Sandy matrix as in Dcm or without granules. Clasts <0.1 m, most 1-5 cm. Tabular beds with non-scoured to irregular bases.

0.1-2.5

0.2-I .4

0.2-2.2

0.3-0.6 0.1-0.6

flow characteristics over the short 6 k m distance o f the f a n . Deposits of smaller flows show a small decrease in clast size with increasing distance f r o m source and, in some units, a change in lithofacies f r o m D m m to Dmg (Figs. 3 and 7). Overall, deposits of the Whangaehu fan do not show the abrupt, downfan changes in lithofacies and clast size that characterize alluvial fans in volcanic and non-volcanic settings (Gloppen and Steel 1981; Rust and Koster 1984; Smith 1987a, 1991; Waresback and Turbeville

181 Table 3. Gravel, sand, and muddy lithofacies of the Whangaehu fan Lithofacies*

Description

Gm

Clast-supported, ungraded gravel. Clasts to 1.5 m. Tightly packed. Matrix of moderately sorted finecoarse sand with few granules. Low-angle cross-bedded sand. Moderately sorted, medium to coarse sand. Usually single sets, less than 0.1 m thick. Poorly sorted medium to coarse sand plus gravei. Some beds of alternating relatively coarse and finegrained sediment with well-developed layers. Fills shallow scour surfaces with less than 0.1 m of relief. Moderately well to well-sorted fine to medium sand. Structureless. Scattered pumice granules. Moderately sorted fine-medium sand and granules. Poorly developed bedding of alternating coarse and fine layers. Grades laterally into Drag. Moderately sorted sandy loam to loamy sand. Rootlets and rhizomorphs present.

S1 Ss

Sm Hff C

* Lithofacies codes after Miall 1978)

2

E

2 0

--

1'2

1'4

Distance From Source (km)

Fig. 7. Change in maximum clast size within selected diamictons. Largest clast within 10 m of the line of section is plotted. Sections are along the major axis of flow. Circles are from deposits of lahar2; squares are lahar3. The triangles are from the sequence of scoria-bearing diamictons (see Figs. 3-5). 'The chart begins at 8 kin, the distance from Crater Lake to the most proximal fan section

Fig. 6A-C. Debris-flow deposits in the fan. A Lithofacies Dcm and Dmg filling a major paleochannel in the central part of the fan (section 15, Fig. 1). The section includes deposits from sediment package Pz and P3. The upper united marked Dcm unit is L3. Jacob staff is 1.5 m tall. B and C Lithofacies Dmm and Dmg filling a minor paleochannel south of 'Scorpion gully'. Contrast the thickness and grain size of diamictons with those in (a)

1990). L a c k o f the typical alluvial-fan pattern resuks f r o m three f a c t o r s . The W h a n g a e h u fan has been dissected during m u c h o f its history. Lahars are channelized a n d travel farther before depositing sediment. Additonally, large variations in the size o f lahars obscures any patterns in grain size o f resulting lithofacies. Final-

ly, the scarp prevents complete development o f a distal f a n setting. Some volcanogenic fans are m u c h larger than the W h a n g a e h u fan (10--15 km; Kesel a n d Lowe 1987; W a r e s b a c k and Turbeville 1990). The lack o f typical alluvial-fan patterns in clast-size and lithofacies could, therefore, be a function o f premature truncation o f the lithofacies tract by the scarp.

Fluvial deposits A n u m b e r o f lithofacies deposited by fluvial processes are present o n the fan (Table 3). I n sandy deposits, s t r a tification r a n g e s f r o m well-developed, alternating coarse and fine layers to low-angle crossbeds t h a t drape shal-

182

Table 4. Facies associations in the sediment dispersal system Channel system Major Minor Margin*

Lithofacies Dcm

Dmm

Dmg

Hff

F

A

C

T

39 12 1

23 22 10

17 19 8

4 3 0

12 36 24

4 8 34

<1 <1 13

<1 <1 8

* Measurements are based on deposit thickness. Abundance of fluvial versus aeolian deposits is underestimated because more sections were measured in dunes than in widespread fluvial areas, such as the northern area of the fan

Another distinctive lithofacies is G m (Tables 2 and 3). This lithofacies underlies terraces in m o d e r n channels, caps Dcm, or fills paleochannels (Fig. 8b). The size of clasts and relationship to diamictons indicate that some deposits result f r o m fluvial reworking of debrisflow deposits during floods. These floods erode the original diamicton matrix and replace it with better sorted sand, producing caps of lithofacies Gm. Gravel transport during relatively large floods produces braid-bar complexes (Miall 1978; Kesel and Lowe 1987).

Aeolian deposits and paleosols

Fig. 8A-C. Fluvial and aeolian deposits on the Whangaehu fan. A Interval of poorly sorted sand and gravel with low-relief scour and fill deposition that is typical of deposits on.the margin of the fan. Jacob staff (1.5 m) is marked at 10 cm intervals. Lithofacies Gm cap on Dcm. Deposits fill a major paleochannel in the area of 'Scorpion gully'. The original diamicton matrix is preserved at the base of the Jacob staff. C Well-developed colony of vegetation on dunes overlying an interchannel section adjacent to 'Scorpion gully'. Large clasts at the top are probably lahar4. Jacob staff (1.5 m) for scale

low scour surfaces (Fig. 8a). Lithofacies S1 and Ss are present in well-defined channels and f o r m widespread sheets that can be traced over 1 km transverse to flow direction. Channel margins are difficult to recognize in these sheets.

Aeolian sand and paleosols are minor components of the fan deposits. Lithofacies Sm (Table 3) represents accumulation of dunes on the fluvial and debris-flow surfaces of the fan (Fig. 8c). The dune sediment contains pumice reworked f r o m the T a u p o Pumice Formation and is interstratified with tephras f r o m the T u f a Trig formation. The dunes do not show any perferential orientation relative to dominant wind direction. These features record growth on stationary sand piles by vertical accretion. Aeolian reworking of fluvial and debrisflow deposits produces a surface lag and downwind dune fields. Dune buildup requires vegetation, both to entrap and stabilize wind-blown sediment. Dune growth continues for as long as the vegetation cover exists. Loss of the vegetation cover leads to rapid erosion and destruction of the dunes. The dunes can be extremely stable: some are over 750 years old (Purves 1990). Paleosols are developed in the aeolian sands. Paleosols record periods of landscape stability when the supply of sand to the dunes was low.

Depositional framework

Sediment dispersal system The channel system of the Whangaehu fan controlled lithofacies distribution across the fan by partitioning flow through a system of major, minor and marginal channels (Table 4). Major channels were the primary lahar paths o n the fan and, therefore, contain a more complete sedimentary record. The deposits are dominated by lenticular, coarse-grained lithofacies D c m (Table 4; Fig.

183 Fan Margin Major Channel

'. ~.~ - Dmg

g~'~ Gm

Minor Channel

Dmm

~ __'Dmm ~ D m g ~.'~ ~oom ~,..,c~,,~Dcrn ~ D m g

Channel Interchannel

2..~

Dmg Dmg Dlllg

Drnm ~ D m m

Dmm ~

Dmg

a'~

Channel

Interchannel

Stream

4, sections 4 and 5; Figs. 6a and 9). Overbank lithofacies are tubular sheets of Dmm and Dmg that are thinner and finer grained than channel deposits (Fig. 8c). Fluvial sands are thin (<0.3m), forming discontinuous lenses within both channel and overbank deposits. Aeolian deposits are not common. Minor channels represent secondary lahar pathways. Fewer lahars entered this part of the system, and the resulting sedimentary record is incomplete. Deposits reflect the secondary nature of these channels. Diamictons are not as abundant and are generally thinner and finergrained (<0.7 m thick, clasts < 1.0 m) than those in areas with major paleochannels (Fig. 5, sections 16 and 10; Figs. 6b and 9). Fluvial deposits are thick (> 1.0 m), and interchannel deposits are commonly interbedded with aeolian sand and the Tufa Trig formation. Channels at the margin of the fan were rarely paths for lahars. Only the largest debris flows entered these areas. Thick, tabular deposits of fluvial and aeolian sand dominate the sections (Fig. 5, section 17; Figs. 8a and 9). Debris-flow deposits are mostly lithofacies Dmg. Beds are thin (< 0.5 m) and clasts are typically less than 0.2 m in length.

Architecture of fan deposits Five distinct packages (P1-Ps) of sediment bounded by basal erosion surfaces are recognized in the fan deposits (Figs. 4 and 5). We define three different architectural styles within the sediment packages, based on the number of diamicton units in major channels, abundance of fluvial and aeolian units, and deposit geometry (Fig. 10). Style-1 deposits are characterized by stacked diamictons and fluvial sand in major and minor channels and adjacent overbank areas (Figs. 6 and 10). Equivalent deposits in marginal areas are thick sheets of fluvial sand. Basal contacts of diamictons within the stack are nonscoured or small-scale ( < 0 . 3 m ) erosion surfaces. Stacked diamictons with little evidence of erosion between events and associated sheets of fluvial sand indicate rapid aggradation during style-1 accumulation. Older deposits are l~nticular, reflecting the dissected nature of the old fan surface. Upper deposits are tabular,

Fig. 9, Lithofacies associations found in different parts of the sediment dispersal system

Dune

Style 1 Drng .~

~

o.~

-.

.

Dmm ~ ~ ~,~Q.~o

o

-...

", , ~ ~ . 2 . ~ q . ~ ".~ ~ # o ~ ' Q . ~ ~ . _ L - ~

Channel BetweenE v e n t s ~ ~ c r n MajorChannel ~ MoreRecordHere

in Channel

Style 2 Fluvial Sand Sheet

~orked C Gm on Dcm

a

p

~

One Deposit Dominates

Style 3

Dunes

LagSurface

-- ~ Terrace Z" New ~ /'\ Cha~e~ion~~Gm & S~ k ~ ~ O i d BetweenEvents

/"

/-~=

Channel

Fig. 10. Diagram showing important features of the three types of accumulation style producing laterally extensive, planar surfaces on the fan. Style-2 deposits are dominated by a single, thick debris-flow deposit (Fig. 10). In this case, major channels are nearly filled with lithofacies Dcm. Small debris-flow deposits (Dmg) or fluvial sand complete the channel fill. Aggradation from deposits of multiple debris flows is confined to major channels. Minor channels usually

184

contain the deposit of one event. Deposits in the marginal areas are sheets of fluvial sand. Fluvial deposits are more abundant in style-2 than in style-.1 accumulation. In channels, fluvial floods produce lithofacies Gm by stream transport and by reworking debris-flow deposits into caps on lithofacies Dcm (Fig. 8b). Aeolian reworking in overbank areas produces granule and boulder lags. Dunes accumulate as sand is trapped around boulders, Despite rather intense reworking in some areas, fan behavior is aggradational in localized areas. Style-3 deposits, in contrast to style 1 and 2, are dominated by fluvical lithofacies (Fig. 10). Deposits are lithofacies Gin, S1, Ss with lithofacies Gm restricted to major and minor channels. Channels in marginal areas of the fan are predominantly S1 and Ss. Overbank deposits were reworked by wind, producing a complex association of boulders, dunes and lag surfaces. Paleosols are common in the dunes, recording long periods of landscape stability. Style-3 accumulation occurs when lahars are small, sporadic, and mostly confined to channels in a dissected setting. Lack of debris-flow deposits from recent lahars in the channels indicates that diamictons are rapidly reworked by fluvial floods into lithofacies Gm. The sedimentary record of style-3 accumulation is extremely difficult to unravel because the preservation potential of diamictons in Channels and interchannel areas is low.

History of fan accumulation Sediment packages P1-P5 represent major chapters in the history of fan development (Fig. 11). Information on timing of these crucial chapters of fan growth comes from tephra stratigraphy and radiocarbon dating. We recognize two major evolutionary phases in the history of the Whangaehu fan marked by overall differences in fan behavior, fan construction (P1-P4) and fan dissection (Ps).

Fan construction

The constructional phase comprised four discrete periods of aggradation during the period ca. 850 to 200 years B~, (Fig. 11). Sediment packages produced during the aggradational periods comprise the bulk of the modern fan. Each aggradationaI period began with emplacement of an exceptionally coarse-grained and laterally extensive diamicton. The lahars that deposited these diamictons are designated L1-L4 (Fig. 11). Ring-plain chronology indicates that these events occurred ca. 850 (L0, post-850 (L2), post-650 (L3), and post-490 (L4) years BP. Two of the large diamictons correlate with deposits described by Campbell (1973) in the valley of the Whangaehu River over 60 km downstream from the study area. We correlate our L2 and L4 with deposits of Campbell's 800- and 400-year events, providing the dates for initiation of package Pz and P4 (Table 5). Aggradation was punctuated by periods of dissection, leading to lateral shifts in the location of major

yr B.P. Dunes

Fan

Tfl 8' P5

200--

P, 9 [-4 O tg.

A

[-3

Tflo - ~600~Tf 8 - - I

r P2 0

"Ifu - - ~

830--~ Tf 5 1800

[-2

~ - L 1

Taupo Pumice

9

9

Fig. 11. Composite stratigraphy for fan and dune deposits. Timing of events is from radiocarbon dating. Dates,are 830---60 years BP under Tf5 (WK 1489, Donoghue 1991), 490_+65 years BP below P4, and a <200 years BP for Tf~5 (Wk 1487 and NZ 601, Purves

1990). Timing for lahar3 is based on an estimated age of ca. 600 years for Tf8 (Donoghue 1991). The scoria-bearing diamictons overlying Tf8 form a compound unit characterized by weaklydeveloped, non-erosional bedding planes. These features indicate accumulation from pulses in a single event, or a series of closely timed events, so the 600-year date is a good estimate. The stratigraphic position of Tf14 is based on our interpretation of lahar4 as eruption induced

5. Data used to calculate eruption frequency and lahar periodicity

Table

Period

Timing Years BP

Tufa r e c o r d

Diamictons Number

P5 P4 P3 P2 P~

200 400 600 800 830

Tf~5-Tfls Tf]4 Tfl~-Tf13 Tf6-Tflo Tf5 and Tfu

-2 3 8 3

to to to to to

Present 200 400 600 800

channel systems. Analysis of the sediment dispersal system for sediment packages P1 to P3 indicates that the major channel of the Whangaehu River was in the area of 'Scorpion gully' during construction, with only small lateral shifts in channel position (Figs. 1, 4, and 5). Absence of P1 below Pz and P3 on the southern fluvial margin of the fan (Fig. 4), suggests that erosion of the 'chute' began approximately 800 years ago. Coarsegrained lithofacies Dcm and Gm (boulders > 1 m) of L4 are coneentrated in the 'chute' area, and lithofacies Drag is present along the southern area of the fan. Deposits north of the 'chute' are widespread fields of clustered boulders that display little evidence of confinement to major channels. The ancestral Whangaehu Riv-

185 er was apparently choked with sediment at this time, leaving the 'chute' as the major sediment transport route during accumulation of package P4.

Fan dissection

Over the last 200 years, the fan has been in a dissectional regime, with fanhead and throughfan trenching dominating fan behavior. The dissectional regime includes brief periods of aggradation within the stream channels, caused by relatively small and infrequent lahars. These infrequent events and subsequent reworking have produced deposits of sediment package Ps. Package P5 is the sole representative of style-3 accumulation on the fan. Style-3 accumulation could also have occurred during dissectional intervals of the constructional phase, but preservation potential of these deposits is low and none are recognized in the older deposits. The sediment dispersal system changed dramatically during the dissectional phase, with avulsion of the ancestral Whangaehu River. The locus of debris-flow ac, cumulation is now in the northern part of the fan (Figs. 1 and 2). Absence of thick deposits of L4 in the northern region of the fan indicates that avulsion occurred after accumulation of package P4.

Sedimentation and volcanism

Sediment packages of the Whangaehu fan represent discrete periods of aggradation and dissection recording an overall history of fan construction between ca. 850 to 200 years ago, and fan dissection from ca. 200 years ago to the present. Contemporaneous explosive eruptive activity of Ruapehu volcano is recorded by the Tufa Trig formation. The eruptive behavior of Ruapehu can be divided into three regimes: major eruptive, minor eruptive, and inter-eruptive episodes~ based on variations in frequency and intensity of volcanic eruptions. The frequency of volcanic eruptions associated with each sediment package is calculated from the Tufa Trig formation (Table 5). One problem with this calculation concerns the nature of the Tufa record. The question of preservation potential relative to the size of a tephra de-

Table 6. Eruptive regimes of Mount Rua-

pehu and associated sedimentation

Package

Major eruptions P1 P2 Minor eruptions P3 Inter-eruptive P4 Minor eruptions P5

posit has yet to be addressed. Since 1945, at least 40 eruptions have been recorded at Ruapehu volcano, more than twice the total Tufa record. Most eruptions are too small to leave a tephra record on either the ring plain or the cone. The fossil record is biased toward larger eruptions. Additionally, lava and dome eruptions are not recorded in the tephra. The tephra record, therefore, is not a complete record of all eruptive activity. Another problem is the low preservation potential of pyroclastic deposits on the cone. Deposits from eruptions as recent as 1975 are not preserved on the cone (Hackett and Houghton 1989). The extremely stable dunes in Rangipo Desert provide a good environment for preservation of tephra deposits. The prevailing wind direction is from the west, so the most complete record of tephra accumulation is on the eastern ring plain. Although imperfect, the Tufa Trig formation is the best information available concerning explosive eruptive behavior. Data used to calculate eruption frequency are in Table 5. We calculated eruption frequency during the P5 interval using only the Tufa Trig formation rather than incorporating the historic record. This ensures congruity of the eruption record over the entire history of fan accumulation. Eruption intensity is also difficult to evaluate. Tufa Trig members Tfs, Tf6, and Tfl4 are the thickest tephras in the formation, ranging from 45-75 mm thick at the type section (Donoghue 1991). These members are laterally continuous in the study area. Thickness and lateral extent suggests emplacement during larger eruptions. Other members of the Tufa Trig formation are less than 25 mm thick, and most form discontinuous pockets of tephra. These features suggest emplacement during smaller eruptions. Major eruptive episodes occurred between 850-800 and 800-600 years BP (Fig. 11, Table 6). During these periods, eruption frequency was high (one event every 15 years; one event every 40 years respectively) and relatively large eruptions emplaced thick layers of tephra over the ring plain (Tfs, Tf6, Tf8). Style-1 accumulation characterized fan behavior during this time. Two periods of minor eruptive activity are recorded 600-400 years BP and 200 years to present, and corresponding fan sedimentation was style-2 and style-3 respectively. Eruptions were less frequent (one every 67 years; one ev-

Ruapehu

Whangaehu fan

Eruption frequency

Big events

Lahar periodicity

Style

1 every 15 years 1 every 40 years

Ts, T6 T8

1 every 10 years 1 every 25 years

1 1

1 every 67 years

--

1 every 67 years

2

1 every200 years

T14

1 every 100 years

2

1 every 50 years

--

--

3

186 Ruapehu Whangaehu Fan

Yr B:P.

Grant

Hubbard & Neall

McFagden

Unstable

Minor

Unstable

Dissection (PS)

- Erosion

Unstable Style3

Unstable

200-'

Inter- Aggradation (P4) eruption Style 2

400

Unstable

-

Unstable Minor

Style 2

/

600

Aggradation (P3)

Unstable

"I 9

800-1

Major

,

Aggradation

(P2)

Unstable

-

Erosion

Style 1

Major Aggradation

(P1)

Fig. 12. Correlation of Whangaehu events with eruption activity

at Ruapehu and periods of increased storm activity identified by Grant (1985), McFagden (1985), and Hubbard and Neall (1980) ery 50 years respectively) and of lower intensity than in the major eruptive regime (Table 6). Ruapehu entered a period of relative quiescence ca. 400 years BP when eruption frequency dropped to one event every 200 years and eruptions were small, interrupted, perhaps, by a single large eruption (Tf14). This inter-eruptive interval was marked by style-2 aggradation on the fan. The relationship between explosive eruptive regimes and sedimentation behavior of the Whangaehu fan is not straightforward (Table 6, Fig. 12). For example, packages P1 and Pz display the expected relationship: major aggradation during major eruptive episodes. Packages P3 and Ps, in contrast, record periods of fan aggradation and dissection during minor eruptive episodes. Lack of a clear relationship between volcanic eruptions and fan behavior could suggest that the Tufa Trig formation is an unreliable index of lahar generation. Some lahar-producing eruptions do not appear in the tephra record. Additionally, failure of Crater Lake can occur without an eruption (1954 Tangiwai lahar; Healy 1954). These two facts, however, do not explain why the fan displayed different behavior during minor eruptive episodes, especially since the Tangiwai lahar occurred during the dissectional regime. Variations in fan behavior record the operation of a set of sedimentation controls.

C o n t r o l s on a c c u m u l a t i o n

Factors

The major control on overall behavior of the Whangaehu fan is the rate of sediment influx. The fan aggraded

when sediment influx was greater than the capacity of the sediment dispersal system to transport material across the fan. Dissection occurred when sediment influx decreased causing streams to erode channels and sediment to bypass the fan system. Many factors influence sediment discharge. Those causing changes in fan behavior, however, must be factors that change over the short time period involved in fan sedimentation. Factors such as tectonic setting, drainage basin size, lithology, and erodibility of sediment cover are not likely to have changed over the short time interval of interest here. Long-term climate change was not an important factor controlling Whangaehu sedimentation. The New Zealand climate has changed little since the Pleistocene and any small changes have had little recognizable influence on landscape development (Malloy 1969). W e identify four time-dependent factors that influenced sedimentation on the Whangaehu fan: (1) frequency and intensity of eruptions; (2) short-term climate change; (3) sediment storage on the volcano flanks; and (4) degree of dissection of the fan surface. Periods of increased storms have influenced landscape stability, and therefore, the erosion and sedimentation history of central North Island ranges (Hubbard and Neall 1980; Grant 1985) and along the coast (McFagden 1985). The signature of periods marked by more frequent tropical storms is erosion and resedimentation, indicating landscape instability (Fig. 12). The signature of less stormy periods is soil development, recording landscape stability. Relative timing of periods of stability and instability reported in the ranges and along the coast differ because of localization of storms. Additionally, McFagden (1985) suggests that the coastal record is not as sensitive as that in the ranges, so that information on periods of landscape stability could be lacking in the coastal record during the 850-665 year interval. Documented periods of increased storm activity cannot be applied directly to the Ruapehu record. At times, however, the records of landscape stability in the ranges and along the coast coincide, suggesting climatic conditions that influenced the country as a whole (Fig. 12). Geomorphic thresholds are important features in sedimentation behavior (Schumm 1977; Dorn et al. 1987; Harvey and Wells 1987; Derbyshire and Owen 1990; Harvey 1990). An important threshold influencing erosive versus constructional behavior of the fan is sandsized sediment stored in valleys on the flanks of Ruapehu. Remnants of sediment deposited during P2, P3, and an earlier interval of fluvial aggradation, are preserved in the upper reaches of the Whangaehu catchment (Fig. 13). Debris flows form from floods by erosion and incorporation of this sediment into the flow, a process called bulking (Scott 1988). Flank sediment stores can influence the bulking behavior of floods. Finally, the degree of fan dissection influenced sedimentation behavior. All four aggradational episodes began with deposition from a large debris-flow that filled the major and minor channels with sediment. These sedimentationevents changed fan geomorphology from a dissected to a nearly undissected condition. Overbank

187

Fig. 13. Aggradationalinterval of sediment packages Pz and P3 plastered against the pre-P1 (?) fluvial deposits exposed on the lower volcano flank. Upper set of arrows mark the scoured contact between P2 and P3. A small terrace (arrow) underlain by P5 deposits is near the present level of the Whangaehu River

flow occurred more frequently in subsequent debris flows, resulting in deposition o f sediment on the fan, rather than transport through it.

The Whangaehu system

During major eruptive intervals, the rate of lahar initiation was high (one event every 10 years; one event every 25 years), resulting in delivery of large amounts of sediment to the fan (Table 6, Fig. 12). Style-1 accumulation of packages P1 and P2 reflects the rapid delivery of sediment to the fan by lahars with little time for reworking between events. The basal diamictons in both sediment packages are much thicker and more laterally extensive than overlying diamictons. This records a critical value for sediment accumulation on the volcano flanks. Generation of large debris flows occurred only after this critical value was reached. Deposits of the initial large events (L1 and Lz) drowned the fan in sediment and set the geomorphic stage for widespread aggradation as eruptions continued to feed sediment into the fan system. These packages are typical examples of syneruptive aggradation described in other volcaniclastic systems (Smith 1987a, 1987b, 1991; Palmer and Walton 1990; Waresback and Turbeville 1990). Other evidence of volcanic control on sedimentation by during accumulation of P1 and P2 is contemporaneous aggradation in parts of the Whakapapa catchment, the system that receives flow from Crater Lake eruptions on the northwestern ring plain. In the upper Whakapapa catchment two sediment packages underlie an interval interbedded Tufa Trig formation and an ash tentatively identified as the Kaharoa Tephra (650 years BP, Lawlor 1980; Palmer 1991, unpublished data). If this identification is correct, then the two Whakapapa packages occupy the same stratigraphic position as packages P1 and P2~on the Whangaehu fan.

Climate does not appear to have been an important sedimentation control during major eruptive periods. The climatic record in the ranges between 850-600 years sP indicates that P1 and Pz aggradation occurred during a period characterized by, predominantly, landscape stability (Fig. 12; Hubbard and Neall 1981; Grant 1985). Sediment package P4 is an exception to the general rule that dissection is the characteristic response of sediment-transport systems in inter-eruptive periods (Table 6). Aggradation was initiated by the large sediment load delivered by L4. L4 was probably generated by the large eruption that emplaced Tufa Trig member Tf14 on the ring plain (Fig. 11). Deposits of L4 represent one of the largest events in Ruapehu's recent history. A noneruptive trigger for this event is difficult to imagine. Additionally, an extremely coarse-grained (1-3 m boulders) deposit of bars and levees in the upper reaches of the Whakapapa catchment occupies about the same stratigraphic position as L4 (Palmer, unpublished data). Again, contemporaneous events in the Whangaehu and Whakapapa catchments support our interpretation of an eruption-driven event. Accumulation of P4 coincided with a period of increased storm activity throughout New Zealand (Fig. 12). Floods, together with a large sediment reservoir produced during eruption of Tf14 and previous activity, produced the sediment discharge necessary to drive aggradation in an inter-eruptive interval. Accumulation, however, was style 2. Lahar periodicity was too low (one event every 100 years) to sustain major construction over the entire fan. Contrasting styles of sedimentation occurred on the fan during minor eruptive periods (Table 6). Aggradation occurred between 600-400 years ago (P3) whereas the fan entered the present dissectional phase ca. 200 years ago. This contrast in style indicates that factors other than volcanism controlled fan behavior. Accumulation of sediment package P3 opened during a period of increased storm activity that is recorded in the ranges of Central North Island and along the coast (Grant 1985; McFagden 1985). Additionally, sediment accumulated in the upper catchment of the Whangaehu River during the two preceding aggradational periods. The combination of increased rainfall during storms and an adequate sediment reservoir set the stage for transformation of fluvial floods into debris flows by erosion of sediment from the flanks. Lahar periodicity and sediment influx were relatively low (one event every 67 years), however, and style-2 accumulation resulted (Table 6). Present conditions at Ruapehu favor style-3 (Ps). Debris flows are generated during eruptions or from drainage of Crater Lake, but these events have not been large, or frequent, enough to fill channels and promote fan growth. Diamictons form terraces in channels and are prone to reworking. Average rainfall on the volcano at 1100 m ranges from 1059-2851 mm (Robertson 1964; Thompson 1984). Rainfall in the Rangipo Desert at 1100 m averages 2400 mm (NZ Meteorological Survey 1973). Most l~recipitation occurs as short, intense storms. Storms with more than 100 mm of precipitation

188 produce floods on the ring plain. Storm floods, however, do not bulk to debris flows because the supply of sand-sized sediment is low in flank valleys. The net result is erosion of the fan and reworking of debris-flow deposits with little net sediment influx to the ring plain.

porting this project. Comments by Gary Smith, John McPherson, Bruce Houghton, and a Bulletin reviewer improved this manuscript substantially.

References Conclusions The rate of sediment infux is the major factor regulating aggradational versus dissectional behavior of the Whangaehu fan. Sediment influx, however, responds to a complex system of controls that include frequency and magnitude of volcanic eruptions, short-term climate change, and sediment reservoir on the flanks of Ruapehu. Large-scale, fan-wide aggradation requires a high rate of sediment delivery and accumulation on the fan. Sediment influx must be rapid and continuous enough to preclude the major dissection between sedimentation events that would result in sediment bypassing the fan system. Maintenance of an unconfined channel system across the fan is crucial to production of fan-wide aggradation. Major aggradational periods (P1 and P2) on the Whangaehu fan began with sedimentation from large events (L1 and L2) that filled channels with sediment, producing an undissected fan surface. Sedimentation from subsequent lahars was rapid enough to prevent substantial dissection. This sedimentation history requires a combination of large sediment stores on the flank and relatively continuous initiation of lahars. These conditions are best met during major eruptive episodes when sediment supply to the flanks is high and higher frequency eruptions provide the initial sedimentation event. Fan construction also occurred during minor eruptive and inter-eruptive episodes (P3 and P4 respectively). As in P1 and P2, aggradation began with sedimentation from a large debris-flow (L3 and t4). In P3 and P4, however, aggradation was spatially limited because sediment influx was not continuous enough to supply the entire fan with sediment. Aggradation continued locally, because storm-induced floods were able to erode sediment from flanks. Fan dissection characterizes the present minor eruptive interval. Stores of sand-sized sediment in flank valleys were depleted during previous fan construction and have not been replaced. Debris-flows are mostly channelized and transport sediment through the fan system. Storm-induced stream floods erode and rework debrisflow deposits in channels and the wind reworks overbank areas, promoting dune growth.

Acknowledgements. This study is the result of MS work of AM Purves, Ph D work of SL Donoghue, and postdoctoral work of BA Palmer. We thank the New Zealand Army for permission to work in the NZ Army Training Area and for logistical support. Support for AM Purves was from the EL Hellaby Indigenous Grasslands Trust and the Robert Bruce Memorial Trust. Support for SL Donoghue was from the NZ Department of Conservation. BA Palmer thanks the Soil Science Department at Massey for sup-

Arguden AT, Rodolfo KS (1990) Sedimentologic and dynamic differences between hot and cold laharic debris flows of Mayon Volcano, Philippines. GSA Bull 102:865-864 Campbell IB (1973) Recent aggradation in Whangaehu Valley, Central North Island, New Zealand. N Z J Geol Geophys 16: 643-649 Derbyshire E, Owen LA (1990) Quaternary alluvial fans in the Karakoram Mountains. In: Rachocki AH, Church M (eds) Alluvial fans, a field approach. John Wiley & Sons, Chichester pp 27-54 Donoghue SL (1991) Late Quaternary volcanic stratigraphy of the southeastern sector of Mount Ruapehu ring plain, New Zealand. Unpub Ph D Thesis, Massey University, Palmerston North, New Zealand Donoghue SL, Neall VE, Palmer AS (1988) Holocene geology of the upper Whangaehu River, Mt. Ruapehu. N Z Geol Soc Misc Pub 41a:61 Dorn RI, DeNiro MJ, Ajie HO (1987) Isotopic evidence for climatic influence on alluvial-fan development in Death Valley, California. Geology 15 : 108-110 Gloppen TG, Steel RJ (1981) The deposits, internal structure and geometry, in six alluvial fans - fan delta bodies (DevonianNorway) - A study in the significance of bedding sequences in conglomerates. In: Ethridge FG, Flores RM (eds) Recent and ancient nonmarine depositional environments: models for exploration. SEPM Spec Pub 31 : 49-70 Grant PJ (1985) Major periods of erosion and alluvial sedimentation in New Zealand during the late Holocene. Roy Soc N Z J 15:67-121 Hackett WR, Houghton BF (1989) A facies model for a Quaternary andesitic volcano: Ruapehu, New Zealand. Bull Volcanol: 51 : 51-68 Harvey AM (1990) Factors influencing Quaternary alluvial fan development in southeast Spain. In: Rachocki AH, Church M (eds) Alluvial fans, a field approach. John Wiley & Sons, Chichester, pp 247-270 Harvey AM, Wells SG (987) Response of Quaternary fluvial systems to differential epeiric uplift: Aquas and Feos river systems, southwest Spain. Geology 15 : 689-694 Healy J (1954) Origin of flood, in Tangiwai Railway Disaster: Report of Board of Inquiry. N Z Gov Printer, Wellington, pp 6-8 Healy J, Lloyd EF, Rishworth DEH (1978) The eruption of Ruapehu, New Zealand on 22 June 1969. N Z Dept Sci and Ind Res Bull 224:1-80 Hubbard CB, Neall VE (1980) A reconstruction of late Quaternary erosional events in the west Tamaki River catchment, southern Ruahine Range, North Island, New Zealand. N Z J Geol Geophys 23 : 587-593 Janda R, Scott KM, Nolan KM, Martinson HA (1981) Lahar movement, effects and deposits. In: Lipman PW, Mullineaux DR (eds) The 1980 eruptions of Mount St. Helens, Washington. USGS Prof-Pap 1250:461-478 Kesel RH, Lowe DR (1987) Geomorphology and sedimentology of the Toro Amarillo alluvial fan in a humid tropical environment, Costa Rica. Geografiska Annaler 69 A:85-99 Lawlor I (1980) Radiocarbon dates from Kohika Swamp PA (N68/104), Bay of Plenty. N Z Arch Assoc Newsletter 23 : 256267 Malloy BP (1969) Evidence for post-glacial climate change in New Zealand. N Z J Hydrol 8:56-63 McFagden BG (1985) Late Holocene stratigraphy of coastal de-

189 posits from Auckland to Dunedin, New Zealand. Roy Soc N Z J 15 : 27-65 Miall AD (1978) Lithofacies types and vertical profile models in braided river deposits: a summary. In: Miall AD (ed) Fluvial sedimentology. Can Soc Petrol Geologists Memoir 5 : 597-639 Nairn IA, Wood CP, Hewson CAY (1979) Phreatic eruptions of Ruapehu: April, 1975. N Z J Geol Geophys 22:155-173 New Zealand Meteorological Service (1973) Rainfall normals for New Zealand, 1941-1970. N Z Meteorol Serv Misc Pub 145 Palmer BP (1991) Holocene lahar deposits in the Whakapapa catchment, northwestern ring plain, Ruapehu Volcano (North Island, New Zealand). N Z J Geol Geophys 34:177-190 Palmer BP, Walton AW (1990) Accumulation of volcaniclastic aprons in the Mount Dutton Formation (Oligocene-Miocene), Marysvale volcanic field, Utah. GSA Bull 102:734-748 Purves AM (1990) Landscape ecology of the Rangipo Desert. Unpub MS thesis, Massey Univ, Palmerston North, New Zealand, 1-277 Pierson TC, Scott KM (1985) Downstream dilution of a lahar: transition from debris flow to hyperconcentrated streamflow. Water Resources Research 21 : 1511-1524 Robertson NG (1964) The frequency of tiigh intensity rainfalls in New Zealand. N Z Meteorol Serv Misc Pub 18 : 1-48 Rodolfo KS (1989) Origin and early evolution of lahar channel at Mabinit, Mayon Volcano, Philippines. GSA Bull 101:414426 Rust BR, Koster EH (1984) Coarse alluvial deposits. In: Walker RG (ed) Facies models. Geol Assoc Can Reprint Series 1:5370, second ed Schumm SA (1977) The fluvial system. Wiley, New York Scott KM (1988) Origin, behavior, and sedimentology of lahars and lahar-runout flows in the Toutle-Cowlitz river system. USGS Prof Pap 1447-A: 1-74

Shultz AW (1984) Subaerial debris-flow deposition in the Upper Paleozoic Cutler Formation, western Colorado. J Sed Pet 54: 759-772 Smith GA (1986) Coarse-grained nonmarine *olcaniclastic sediment: terminology and depositional processes. GSA Bull 97 : 1-10 Smith QA (1987a) Sedimentology of volcanism-induced aggradation in fluvial basins: examples from the Pacific Northwest, USA. In: Ethridge FG, Flores RM, Harvey MD (eds) Recent developments in fluvial sedimentology. SEPM Spec Pub 39 : 217-228 Smith GA (1987b) The influence of explosive volcanism on fluvial sedimentation: the Deschutes Formation (Neogene) in central Oregon. JSP 57 : 613-629 Smith GA (1991) Facies sequences and geometries in continental volcaniclastic sediments. In: Smith GA, Fisher RV (eds) Sedimentation in volcanic settings. SEPM Spec Pub 45 : 109-121 Smith GA, Fritz WJ (1989) Penrose Conference report: volcanic influences on terrestrial sedimentation. Geology 17:375-367 Thompson CS (1984) The weather and climate of the Tongariro Region. N Z Meteorol Serv Misc Pub 115:1-35 Vessel RK, Davies DK (1981) Nonmarine sedimentation in an active fore arc basin. In: Ethridge FG, Flores RM (eds) Recent and ancient nonmarine depositional environments: models for exploration. SEPM Spec Pub 31 : 31-48 Waresback DB, Turbeville BN (1990) Evolution of a Plio-Pleistocene volcanogenic-alluvial fan: the Puye Formation, Jemez Mountains, New Mexico. GSA Bull 102: 298-314 Wilson CJN, Walker GPL (1985) The Taupo eruption, New Zealand (1). General Aspects Phil Trans Roy Soc London A 314:199-228 Editorial responsibility: B. F. Houghton