Geomorphology 132 (2011) 327–338

Contents lists available at ScienceDirect

Geomorphology

j ourna l homepage: www.e lsev ie r.com/ locate /geomorph

Effects of rock avalanches on glacier behaviour and moraine formation

Natalya V. Reznichenko a,⁎, Tim R.H. Davies a, David J. Alexander b

a Department of Geological Sciences, University of Canterbury, Private Bag 4800, Christchurch, New Zealandb School of Geography, Planning and Environmental Management, University of Queensland, QLD 4072, Brisbane, Australia

⁎ Corresponding author. Tel.: +64 3 364 2987x7697;E-mail address: natalya.reznichenko@pg.canterbury.

0169-555X/$ – see front matter © 2011 Elsevier B.V. Adoi:10.1016/j.geomorph.2011.05.019

a b s t r a c t

a r t i c l e i n f o

Article history:Received 10 January 2011Received in revised form 24 May 2011Accepted 25 May 2011Available online 1 June 2011

Keywords:Rock avalanchesSupraglacial debrisIce-surface ablationGlacier advancesTerminal morainesPaleoclimatology and hazards

Although large rock avalanches are infrequent, sediment production in active orogens is dominated by suchevents, which may strongly influence geomorphic processes. Where rock avalanches fall onto glaciers, theymay affect glacier behaviour andmoraine formation.We outline the processes of rock avalanche initiation andmotion, and show that supraglacial deposits of rock avalanche debris have distinct sedimentological andthermal properties. Laboratory experiments on the effects of such debris on ice ablation are supplemented byfield data from two rock avalanches in the Southern Alps, New Zealand. Their effects are compared with thoseof the thinner supraglacial debris that results from small rockfalls and melt-out of englacial debris.Implications of rock–avalanche debris cover for glacier behaviour are explored using a mass-balance model ofthe Franz Josef Glacier in New Zealand, demonstrating a likely supraglacial rock avalanche origin for theWaiho Loop moraine, and considering the potential hazard of a large rock avalanche onto the present-dayglacier.

fax: +64 3 364 2769.ac.nz (N.V. Reznichenko).

ll rights reserved.

© 2011 Elsevier B.V. All rights reserved.

1. Introduction

In this paper we examine the effects of the deposits of large rockavalanches on glacier behaviour and moraine formation. Supraglacialrock avalanche deposits on the glacier ablation zone reduce ice-surface ablation, and this can significantly alter the overall glaciermass balance. With a sufficient proportion of the ablation zonecovered by rock avalanche debris the glacier mass balance willbecome significantly less negative, and, as a result, the terminus mayadvance. This advance may generate a terminal moraine that has noclimate-related cause. The much-debated Late-Glacial Waiho Loopmoraine, the Franz Josef Glacier, New Zealand (Barrows et al., 2007;Tovar et al., 2008), appears to be the result of such a sequence ofevents and we use it as an example to demonstrate how a large rockavalanche deposit can reduce ablation causing a glacier to advanceand form a prominent frontal moraine.

Rock avalanche deposits on glaciers and their effects have receivedparticular attention only recently. The increasing frequency ofrecognised historical and prehistorical catastrophic rock avalanchedeposits in high mountains areas worldwide (e.g., Post, 1967;Whitehouse and Griffiths, 1983; Hewitt, 1998, 2009; Hermannset al., 2001; Geertsema et al., 2006; Allen et al., 2011), together withwell-defined magnitude–frequency relationships that indicate thatvery large events contributemore volume over time than domedium-sized ones (Malamud et al., 2004; Korup and Clague, 2009) has

established the importance of such events in the sediment deliveryand geomorphology of active orogens. Where glaciers are present inmountain valleys, some rock avalanches inevitably fall onto them;however, supraglacial rock avalanche deposits appear to be lessfrequently reported than those that do not fall onto glaciers. This ispartly because of the limited extent of present-day glaciers and also,significantly, because supraglacial rock avalanche deposits are rapidlyreworked by glaciers and deposited as moraines whose potentialrock–avalanche origins have rarely been considered in the past. Bycontrast, a rock avalanche depositing into a glacier-free river valley,though progressively removed by river erosion, often leaves long-term recognisable traces of its initial extent in the form of remnantlarge angular boulders and vestigial high terraces (Chevalier et al.,2009; Lee et al., 2009). Thus, more likely that the frequency ofsupraglacial rock avalanches and, hence, their significance to glacierdynamics and behaviour have in the past been underestimated.

Notably, many recently identified rock avalanche deposits, forexample those studied by Hewitt (2009) in the Karakoram Himalaya,had earlier been interpreted as moraines and corresponding (incor-rect) paleoclimatic inferences drawn (see also McColl and Davies,2011; Kariya et al., 2011). One of the conclusions of the present workis that prominent moraines can be generated by rock–avalanche-driven glacier advances; these cannot presently be distinguished frommoraines reflecting climatically driven advances, and their inclusionin moraine age data may also have led to incorrect paleoclimateinterpretations.

In this work we highlight recently-published information aboutthe properties of rock avalanche deposit, their effects on glacierablation, and glacier response to reduced ablation. We report an

328 N.V. Reznichenko et al. / Geomorphology 132 (2011) 327–338

investigation of the current state and effects of two recent rockavalanche deposits on glaciers in Southern Alps of New Zealand, anddevelop a conceptual picture of the response of mountain valleyglaciers to emplacement of extensive rock avalanche debris on theablation zone. Mass-balance modelling of the Franz Josef Glacierresponse to severely reduced ablation also demonstrates the potentialof a large rock avalanche to substantially advance the present-dayterminus of this glacier, with following significant disruption tosociety and tourism.

2. Rock avalanches onto glaciers

2.1. Causes

A rock avalanche is a large-scale rock slope failure (of the order of amillion cubicmetres ormore) that leads to amass of fragmenting rocktravelling across a landscape at velocities up to 100 m/s (Voight andPariseau, 1978; Fort et al., 2009; Hewitt, 2009). Failures of this scaleare frequently triggered by major earthquakes, but a seismic trigger isnot necessary; for example, the 1991 Aoraki/Mt. Cook, 1991 Mt.Fletcher, 1999 Mt. Adams, and 2007 Young River landslides in theSouthern Alps of New Zealand between 1991 and 2007 were all about107 m3 in volume and neither seismically- nor rainfall-triggered(McSaveney, 2002; Hancox et al., 2005; Allen et al., 2011). Relativelyfew large-scale rock-slope failures appear to be triggered by rainfall(Melosh, 1987; Soldati et al., 2004). Contrary to convention, recentwork suggests that loss of ice buttressing of valley slopes during andfollowing deglaciation may have been overemphasised as a cause ofthe apparently high frequency of large rock avalanches in the earlyHolocene (McColl et al., 2010).

2.2. Processes

A rock avalanche usually initiates as the detachment of a largemass of rock from the upper part of a mountain edifice by completionof a through-going failure surface. To generate the highly-fragmenteddebris described below, the volume needs to be in the order of 106m3

or more (Davies and McSaveney, 2009). The detached mass initiallyslides under gravity down the mountainside, progressively collapsingas it does so into joint-controlled blocks. However, at an early stage,progressively smaller intact (i.e. uncracked) rocks will start tocomminute (or fragment; e.g., Davies and McSaveney, 2009) underthe high stresses of the fall, generating fragments of all sizes to

Fig. 1. The Sherman Glacier, Alaska, as seen on 27 August 1963 (left)

submicron diameter; large quantities of fine particles are expelledfrom the surface to form a growing aerial dust cloud. By the time itreaches the glacier surface, the mass is pervasively fragmenting andtravelling at high velocity. The debris then moves rapidly across theice surface, spreading both laterally and longitudinally as it does soand becoming thinner in the process. A rock avalanche travellingacross a glacier continues to comminute until it halts, as demonstratedby the frequent presence of distally located, shattered but undisag-gregated clasts; these are the result of fragmentation and, if formedmore proximal, would have been disaggregated by the shear strainfield in the avalanche (e.g., McSaveney, 1978). Often evidence is foundthat the ice-surface has been eroded by the rock avalanche passage,which contributes mass to the emplaced deposit (e.g., McSaveney,1978, 2002).

2.3. Deposits

Supraglacial rock avalanche deposits are usually much thinnerrelative to their volume than their nonsupraglacial counterparts, andthus larger in area. This is the result of the lower basal friction of rockavalanches that travel over snow and ice (Post, 1967; McSaveney,1975, 2002; Gordon et al., 1978; Evans and Clague, 1988; Jibson et al.,2004, 2006; Lipovsky et al., 2008; Deline, 2009; Hewitt, 2009), whichallows a given volume to spread over a greater area. For example, theaverage thickness of the 40×106m3 Sherman Glacier rock avalancheof 1964 (Fig. 1) was about 1.5 m (McSaveney, 1978), whereas thethickness of the majority of reported nonsupraglacial rock avalanchesthicknesses varies between 10 m and 60 m or more (Smith et al.,2006). This characteristic allows a rock avalanche to cover a very largearea of glacier with a debris carpet some metres thick.

As a result of the intense fragmentation that occurs during fall andrunout, the debris of rock avalanches is predominantly extremely fine.For example, McSaveney and Davies (2007) found that 90% by weight(N99% by number) of the surface debris of the 1991 Aoraki/Mt. Cookrock avalanche, New Zealand, was b10−5 m in diameter. The coarser(metre-scale) “carapace” layer is usually at the debris surface (causedby less intense fragmentation at the surface because of lowerconfining pressure); so the surface appearance is often blocky, butmost of the underlying debris is extremely fine, with larger clasts of allsizes forming a fractal grain size distribution (Fig. 2). The bulk densityof the subsurface debris is high caused by the wide grain sizedistribution, and the voids ratio and permeability are correspondinglylow, compared to those of other types of supraglacial debris. As a

and with rock avalanche deposit on 24 August 1964 (Post, 1965).

Fig. 2. Examples of rock avalanche deposits: (A) basal exposure of the nonsupraglacial 1929 Falling Mountain Rock avalanche, NZ; and (B) rock avalanche debris about 2 m thick onthe terminal surface of the Mueller Glacier, Southern Alps, NZ.

329N.V. Reznichenko et al. / Geomorphology 132 (2011) 327–338

result, rock avalanche debris has high thermal mass and inertia that is,it requires a large quantity of heat to increase its temperature, so itheats up slowly compared to debris with large volumes of air-filledvoids. This gives rock avalanche debris high insulation capacity,compared to the thin melt-out debris common on valley glaciers, aswe now describe.

3. Supraglacial debris: Effects on ablation

Themajor direct effect of a rock avalanche deposit on a glacier is toprotect the ice surface beneath the debris from solar radiation,resulting in reduced surface ablation beneath the deposit (Østrem,1965; McSaveney, 1975; Lundstrom et al., 1993; Mattson et al., 1993;Nakawo and Rana, 1999; Brock et al., 2010; Reid and Brock, 2010;Reznichenko et al., 2010). The heat flux through the debris layerdepends on its thickness and physical properties (lithology, bulkdensity or porosity, albedo, moisture content, and permeability; e.g.,Nakawo and Young, 1981, 1982; Kayastha et al., 2000; Nicholson andBenn, 2006), and determines the ice-surface melt rate. Where a largerock avalanche deposit covers a sufficient proportion of the ablationzone with debris of the order of metres thick, surface-ice ablation willalmost completely cease and, as a result, the dynamics of the coveredglacier will be significantly affected (Shulmeister et al., 2009).Ablation within or beneath the ice will be unaffected by surficialdebris, but this is usually a small component of total ablation (of theorder of 10%; Alexander et al., in review).

Fig. 3. The debris-covered lower Tasman Glacier, Southern Alps, NZ. (A) The rough topogrdeveloped glacier ‘karst’; the Hochstetter icefall is the steep tributary glacier in the backgroustream flows into the Tasman Glacier. Note the general thinness of the debris cover.

3.1. Melt-out debris

The ablation zones of many mountain valley glaciers are coveredwithextensive supraglacial debris. This resultsmostly fromthemelt-outof englacially or subglacially transported debris originating fromsubglacial or extraglacial sources (Rogerson et al., 1986; Winkler,2009); it is commonly several to a few tens of centimeters in thicknessand characterized by angular, coarse clasts with a small proportion offines and, as a result, high permeability (Drewry, 1986), low bulkdensity and relatively low thermal inertia. Melt-out debris increases inthickness toward the terminus from lateral displacement and continuedemergence of englacial debris and if it achieves sufficient thickness canreverse the increasing rate of clear-ice ablation toward the terminuscausedby reduction inglacier surfaceelevation. Thehighdebris porosityfavours air and water circulation, reducing the insulation effect of thedebris; for example, Sharp (1949) found that coarse debris 15 cm deephas the same insulating effect on ice as sand 3 cm deep. Debris-coveredglacier snouts usually are very rough, so a several-centimetres-thickdebris layer forms a mosaic with thin layers of debris, clear patches ofthe ice, and exposed ice cliffs. Differential melting amplifies thetopographic unevenness of the glacier and complicates the estimationof total melting rates under this cover (Fig. 3).

3.2. Rock avalanche debris

The effect of rock avalanche debris on underlying glacier ice isdistinctly different from that of melt-out debris owing to its greater

aphy of the debris-covered glacier about 8 km from the modern terminus with well-nd. (B) Differential ablation from discontinuous debris cover where the Hochstetter ice

330 N.V. Reznichenko et al. / Geomorphology 132 (2011) 327–338

thickness (meters) and thermal inertia (because of the fractal grainsize distribution and high proportion of very fine grains). The effect ofhigh thermal mass is to require a large quantity of heat to be inputbefore the mass can achieve steady-state heat transfer, so this takes along time. If this time exceeds the diurnal radiation cycle time, steady-state heat transfer to the underlying ice-surface is never achieved(Reznichenko et al., 2010). The major effect of a thick rock avalanchedeposit on a glacier ablation zone is thus to dramatically reduce thesurface ablation beneath the debris by comparison with adjacent bareor thinly-debris-covered ice (see next section). This leads tothickening of ice (in fact, a reduction of ice thinning) under thedeposit relative to neighbouring clear ice, which begins immediatelyand is often clearly evident a year after emplacement of the deposit. Ifthe deposit thickness exceeds 2–3 m, the underlying ice-surfaceablation reduces effectively to zero (Bozhinskiy et al., 1986). In thissituation the rate of relative ice thickening beneath the deposit equalsthe surface ablation rate of clear ice at that location.

A rock avalanche deposit also contributes a large mass of sedimentto the debris transportation system of the glacier almost instanta-neously, whereas the gradual exhumation of melt-out debris covertakes place slowly and quasicontinuously without adding mass to thesystem.

If the glacier surface is crevassed, part of the rock avalanchedeposit may enter the sub- and englacial transportation systems ofthe glacier (Hewitt, 2009). However, most of the debris remainssupraglacial and will eventually be carried to the terminus by glacierflow (e.g., the Sherman Glacier, Alaska; Figs. 1 and 4).

3.3. Insulation mechanisms

In order to study the role of the debris cover in insulating theunderlying ice frommelting, we conducted laboratory experiments inwhich ice blocks with clear surfaces and identical sizes covered witheither medium sand (representing typical highly permeable melt-outdebris) or rock avalanche debris were exposed to identical radiation,both steady and diurnally-cyclic (Reznichenko et al., 2010). Theseexperiments showed that the insulating effect of debris on icedepends on the occurrence of cyclic radiation input. For example,under specific laboratory conditions with continuous, steady-stateradiation, sand cover greater than 5 cm thick delays the onset of ice-surface melting by more than 12 h as the sand layer warms up, butafter that heat transfer through the debris is steady and ablation ratesbecome similar to those for bare ice. By contrast, when realisticdiurnal cycles of radiation are imposed the debris cover continues toreduce the rate of ice melt over the long term (Fig. 5A).

Fig. 4. The 1964 Sherman Glacier rock avalanche deposit 44 years after emplacement on 2terminus by glacier flow, where it is dumped to form a moraine (photos by M.J. McSavene

Our study showed that sandy supraglacial cover 5 cm thick(representing melt-out debris) reduced melting to 50% of that withbare ice under same conditions, and 9 cm of sand reduced ablation to25% (Fig. 5A). Using rock avalanche debris instead of sand underdiurnally-cyclic radiation (Reznichenko et al., 2010) increased theinsulating effect substantially. The presence of rock avalanche debrisalso significantly reduced rainfall-induced ice-surface ablation,whereas this effect did not occur with sand (Fig. 5B).

In summary, rock avalanche debris cover on a glacier ablation zonedramatically reduces total ablation beneath the debris; the reductionis much less under normal (melt-out) debris cover.

3.4. Field studies

In order to compare the laboratory results of Reznichenko et al.(2010) with effects in the field, we surveyed rock avalanche depositsfrom 1991 Aoraki/Mt. Cook and 2004 Mt. Beatrice on the Tasman andHooker Glaciers, respectively (Southern Alps of New Zealand, Fig. 6).Both avalanches were typical of aseismic slope mass failures in theSouthern Alps (McSaveney, 2002; Cox et al., 2008).

Ground penetrating radar (GPR) surveys of these deposits werecarried out to measure the effect of rock avalanche debris on thesurface-ice ablation. GPR images of two perpendicular profiles alongthe rock avalanche deposit and adjacent clear or debris-covered iceprovided information on the thickness of the rock avalanche deposits,the elevation of the underlying ice surfaces, the modification of thedeposits since emplacement, and the interaction of the rock avalanchedeposits with adjacent clear or debris-covered ice.

3.4.1. Mt. Beatrice rock avalanche depositThe 2004 rock avalanche from Mt. Beatrice (2528 m) was

emplaced onto the debris-free part of the ablation zone of the HookerGlacier, close to an earlier deposit that presumably originated fromthe same source (Fig. 6A). Satellite images show that between 2004and 2009 the deposit moved about 800 m down the valley, indicatingan average ice flow velocity of about 160 m/y at that location, which issimilar to previously estimated velocities at the same altitudes(higher than 1200 msl.) on the Tasman Glacier (Quincey and Glasser,2009).

In December 2009 we obtained the longitudinal and cross-profilesof the Mt. Beatrice deposit shown in Fig. 6A. Profiles revealed that fiveyears after emplacement an elevated ice platform, varying between 20and 30 m above the adjacent upstream and downstream ice surfaces,respectively, was present under the 3–7 m thick rock avalanchedeposit (Figs. 6 and 7), indicating an average clear-ice ablation rate of

008: (A) the rock–avalanche cover travels supraglacially; and (B) and is carried to they).

Fig. 5. (A) Example of surface lowering of bare ice and ice under 1, 5, 9, and 13 cm of debris cover under diurnal-cycle conditions. (B) Comparison of the surface lowering of the bareice, ice under debris cover of sand and rock avalanche material 9 cm deep with diurnal-cycle conditions and diurnal rainfall effect (Reznichenko et al., 2010).

Fig. 6. Location of the two rock avalanche deposits where GPR studies were conducted (red lines indicate GPR profiles, black arrows indicate ice flow directions, satellite image of 2001,LandSat): (A)Mt. Beatrice (2004) rock avalanche deposit on the Hooker Glacier, 2007 (aerial photo by S.Winkler); and (B) Aoraki/Mt. Cook rock avalanche (1991) deposit on the TasmanGlacier with Hochstetter and Tasman ice indicated (satellite image 2009, ASTER).

331N.V. Reznichenko et al. / Geomorphology 132 (2011) 327–338

Fig. 7. GPR profile along the 2004 Mt. Beatrice rock avalanche in December 2009; the rock avalanche deposit thickness reaches about 7 m, and relative ice thickening is about 20 and35 m at the upstream and downstream sides of the deposit, respectively. (A) The Mt. Beatrice rock avalanche on the Hooker Glacier deposit 5 years after emplacement; and (B)The~20-m-high upvalley slope of the deposit looking downvalley.

332 N.V. Reznichenko et al. / Geomorphology 132 (2011) 327–338

4–6 m/y. The average debris thickness we estimated to be 3–5 m,giving a volume of 1.4–2×105m3. However, as the debris is carrieddownglacier, the increasing relative ice thickness causes steep slopesto form along the sides of the deposit. These slopes grow in height andbackwaste from melting over time, redistributing the deposit on theedges. As a result, the elevated area of the avalanche deposit coverreduces with time, and the estimated volume decreases as material isshed to the adjacent glacier surface. The GPR data also indicateshearing of the thickened underlying ice owing to the differential iceflow (Fig. 7).

3.4.2. Aoraki/Mt. Cook rock avalanche depositThe 1991 Aoraki/Mt. Cook rock avalanche, whose volume was

originally estimated at 11.8±2.4×106m3 (McSaveney, 2002), trav-elled across the Grand Plateau, down the Hochstetter icefall, andspread onto and across the Tasman Glacier (Fig. 6B). Themotion of theavalanche deposit following emplacement was complicated. Shortlyafter deposition, the Hochstetter ice stream deformed the originalavalanche deposit into twomain parts: that on the Hochstetter ice hadmoved more than 2200 m down the valley by 2009, whereas that onTasman ice had moved only 500 m (Fig. 6B). On the latter deposit weobtained two perpendicular GPR profiles through the avalanche andthe adjacent upglacier, debris-covered ice.

The Aoraki/Mt. Cook rock avalanche deposit (Fig. 8) is up to 10 min thickness and a 25-m-high debris-covered ice ridge has formed atthe upstream edge of the deposit; this was not present immediatelyafter emplacement (McSaveney, 2002). The GPR survey showedmeltwater channels under the deposit, which can cause localcollapses. Therefore, caused by differential ice flow velocities in thedeposition area, complications from tributary ice flow and thesemeltwater channels, the relative ice thickening from reduced ablationis hard to estimate.

Similar ice thickening has been reported on other glaciers, with theformation of platforms up to 30 m above surrounding ice surfacesfrom reduced ice-surface ablation (Post, 1968; Molnia, 2008). Hewitt(2009) reported that the rock avalanche deposit onto the BualtarGlacier, Karakoram, increased the surface elevation by 5–15 m in oneyear by comparison with adjacent ice. Field data, thus, confirm theexperimental result that rock avalanche debris dramatically reducesice-surface ablation.

4. Effects of rock avalanches on glacier dynamics

4.1. Mass balance

The net mass balance of a glacier is determined by ice-mass gain inthe accumulation zone and ice-mass loss in the ablation zone. Apositive net mass balance over a certain period will cause glacierthickening and frontal advance, while negative net mass balance (i.e.ablation exceeds accumulation) will usually cause the glacier to thinalong its profile and the terminus may retreat. Both advance andretreat have response times specific to each glacier. However, in somecircumstances (e.g., a long, low-gradient glacier tongue withextensive debris cover), negative net mass balance can cause verticalsurface lowering while the terminus remains stationary, as with thedownwasting Tasman Glacier (Kirkbride, 1993).

Two main changes in the glacier system result from theemplacement of a rock avalanche deposit on the ablation zonesurface:

(i) the net mass balance becomes less negative or more positive byreduction of ice-surface ablation in the ablation zone, assumingthere is no change in accumulation rate; and

(ii) additional mass is provided (a) immediately, by the rockavalanche deposit, and (b) gradually, by the increasing relativeice thickness caused by reduced ablation under the rockavalanche deposit.

Rock avalanche debris deposited in the glacier accumulation zonerapidly becomes buried by snow and ice (and entrained in theenglacial transport pathway), so it only affects the ablation rate for ashort time. It will most likely emerge in the ablation zone as melt-outenglacial or subglacial debris after being transported and dispersed inthe en/subglacial transport pathways.

The reduction of total ablation under a rock avalanche deposit canbe large (e.g., Reid, 1969; McSaveney, 1975; Reznichenko et al., 2010;Alexander et al., 2011). Ice-surface ablation usually reduces to close tozero beneath a rock avalanche deposit. While sub- and englacialablation by meltwater heat flow, heat conductivity and otherprocesses will be unaffected by the surface debris; they contributeonly in a minor way (~ 10–20%; Alexander et al., in review) to totalablation. Thus, if a sufficient area of the ablation zone is covered, theglacier net mass balance can be significantly altered. Because the

Fig. 8. GPR profile along the 1991 Aoraki/Mt. Cook rock avalanche and adjacent debris-covered glacier surface in February 2009, the Tasman Glacier, Southern Alps, NZ. The thicknessof the rock avalanche deposit is up to 10 m, and the relative ice thickening at the upglacier edge of the deposit is about 25 m. (A) The ridge formed at the upvalley edge of the depositwith the Hochstetter icefall in the left background; and (B) GPR measurements on the upvalley slope of the Aoraki/Mt. Cook rock avalanche deposit.

333N.V. Reznichenko et al. / Geomorphology 132 (2011) 327–338

clear-ice ablation rate increases toward the terminus (Schytt, 1967)and the rock avalanche deposit will be progressively carried down-glacier, the reduction in surface ablation per unit area will increase asthe debris approaches the terminus. At the same, time the upglacierlimit of the debris cover is moving downvalley and clear-ice ablationthereby increasing, so the net variation of mass balance with time isnot simple. For the present we note only that when the rear of thedebris cover has advected downvalley of the original terminusposition, the glacier mass balance is re-established and the debris-covered ice downvalley thereafter behaves independently of theglacier.

4.2. Glacier response to positive mass balance

A rock avalanche covered 50% of the ablation zone of the ShermanGlacier to an average depth of 1.5 m during the Alaska earthquake of1964, and ablation beneath the deposit was reduced by about 80%(McSaveney, 1975). The resulting alteration in the net mass balancechanged previous glacier retreat to steady-state conditions and, aftersome years, a slow advance became evident (Post, 1968; McSaveney,1975), which still continued in 2009 (M.J. McSaveney, GNS Science,PO Box 30368 Lower Hutt, NZ, pers. comm. 2009). A rock avalancheonto the Bualtar Glacier, Karakoram, in 1986, caused substantialrelative ice thickening under rock avalanche debris covering only 15%of the ablation zone, but the positive impact on mass balance wasequivalent to a 20% increase in annual accumulation (Hewitt, 2009).As a result, the Bualtar Glacier (which had been retreating for severaldecades) surged during first 2 years after the rock avalanche andcontinued to advance during the following 12 years.

The mass added to the glacier both immediately (rock avalanchedebris) and gradually (relative ice thickening) will tend to increasethe glacier flow velocity. Firstly, the rock avalanche material itself isequivalent to about double the weight of the same volume of ice

(Shulmeister et al., 2009). Secondly, the relative thickening of iceunder the deposit is added to the normal (ice) mass of the glacier. Forexample, the area of glacier covered by the Mt. Beatrice rockavalanche 5 years after emplacement has the equivalent of about2×(1.4–2×105)m3 additional ice weight from the rock avalanchedeposit plus about 13×105m3 of ice (if the average additional icethickness is 27 m), which is increasing annually. Totally it isequivalent of an extra 17×105m3 of ice on an area of about 5 ha, or34 m of ice depth. This additional mass must alter the glacierdynamics by increasing basal sliding, and this will occur in proportionto the percentage of the ablation zone affected and in inverseproportion to the original ice depth.

The sliding velocity of a warm-based valley glacier is proportionalto the square of the bed shear stress, which depends on the ice-surfaceslope and the mass of the ice per unit area of bed (Paterson, 1994).Thus 10% additional ice depth should increase flow velocity by 20% ata given location. In addition, extra sediment finding its way to theglacier base via crevasses and moulins may affect the subglacialtopography and drainage and hence the basal water pressure, whichin turn affects sliding (Turnbull and Davies, 2002; Davies and Smart,2007); however, the magnitude of this effect is difficult to determine.This effect was first observed after the St. Elias earthquake by Tarr andMartin (1914), who proposed the “earthquake advance theory”whenseven glaciers were surging and advancing 10 years after the event.The advances were attributed to added mass from rock, snow, and iceavalanches, and also to the uplift of the St. Elias Mountains by 14 mduring the earthquake and the corresponding increase in glaciergradients (McSaveney, 1975).

The addition of sufficient rock avalanche debris to a glacierevidently can have significant effects on bothmass balance and glaciermotion (Shulmeister et al., 2009). An equilibrium glacier mayadvance; a retreating glacier may retreat more slowly, or come intoequilibrium, or advance; and an already advancing glacier may

334 N.V. Reznichenko et al. / Geomorphology 132 (2011) 327–338

advance more rapidly. A surge-type glacier may surge prematurely(Hewitt, 1988).

In case of the Mt. Beatrice rock avalanche onto the Hooker Glacier,no changes in glacier behaviour have been observed, because the areacovered by the deposit was small proportionally to the ablation area.Shulmeister et al. (2009) estimated that glacier behaviour would bealtered by a rock avalanche deposit that covered more than about 10%of the ablation zone area. With smaller deposits, the glacier mayexperience local effects from the additional supraglacial load, but theoverall mass balancewill not be significantly affected. TheMt. Beatricerock avalanche covers about 0.05 km2 of the roughly 7 km2 of theablation zone for the Hooker Glacier, which is proportionallyb1% ofthe area (Fig. 6A). The Aoraki/Mt. Cook avalanche deposit coveredabout 1.7 km2, b4% of the ablation zone of the Tasman Glacier (about45 km2; Kirkbride, 1989). The deposit has since been redistributedand compressed and now covers an even smaller area (Fig. 6B).

Fig. 9. An example of the modelled rock avalanche debris cover on the Franz Josef Glacier,terminus position at the Waiho Loop, assuming temperatures 2 °C less than the present-daablation reflect the accumulation of large quantities of debris toward the terminus, while inhas moved downvalley. Location of the Franz Josef névé is also shown in Fig. 6.

Therefore, these two deposits are not expected to noticeably affect theoverall behaviour of their glaciers.

4.3. Modelling

The effects of rock avalanche debris on glacier behaviour can bepredicted quantitatively. Alexander et al. (2011) developed a simplesteady-state mass balance model, calibrated on the modern (1999–2010) quasisteady-state conditions of the Franz Josef Glacier andsuccessfully proven against its Little Ice Age terminal morainepositions and trim lines. This was used to test the hypothesis ofTovar et al. (2008) that a catastrophic rock avalanche could havecaused the advance of the Franz Josef Glacier (western Southern Alps,New Zealand; Fig. 9) to the location of the Late Pleistocene/EarlyHolocene Waiho Loop moraine; and thus have formed the moraine.The modelling shows that a steady-state terminus position at the

NZ, showing percentage of ablation reduction, which is staggered below the ELA to ay, with an ELA of 1480 masl (satellite image of 2001, LandSat). The staggered levels ofthe upper-ablation zone, ablation rates will return to those of clean ice once the debris

Fig. 10. Response of a glacier to rock avalanche debris covering the ablation zone:(A) glacier in equilibrium prior to rock avalanche; (B) rock avalanche emplaced; totalablation reduced by~70%; (C) lack of melting increases ice depth and sliding velocity;rock-covered reach reduces in length caused by relative velocity and ice thickens;(D) glacier now debris-free to original terminus; debris-covered ice continues to flowand slowly melt, forming terminal moraine by dumping RA deposit; and (E) all icemelted, terminal moraine remains with ablation moraine.

335N.V. Reznichenko et al. / Geomorphology 132 (2011) 327–338

Waiho Loop requires mean temperatures about 4–5 °C cooler thanpresent-day. Anderson andMackintosh (2006), using an ice flow/massbalance model, found that an advance to the Waiho Loop required3–4 °C cooling from present-day. The results from both studiescontradict proxy data, which indicate temperatures no more than~2.5 °C cooler than present-day during the last glacial-interglacialtransition (LGIT) about 12,000–13,000 YBP (Carter et al., 2003; Turneyet al., 2003; Barrows et al., 2007). Direct dating of the Waiho Loopmoraine by Barrows et al. (2007) yielded an age of about 10,500 YBP,corresponding to significantly warmer temperatures than those thatoccurred during the LGIT and making a climatically driven advanceeven less likely.

To explain the advance of the Franz Josef Glacier from anequilibrium position close to Canavan's Knob to the Waiho Loopposition during a time without any cooling (but ~2° cooler thanpresent), Tovar et al. (2008) suggested, based on the sedimentology ofthe Loop moraine, that the ablation zone of the glacier was coveredwith debris by a rock avalanche. Alexander et al. (2011) showed that65% reduction in total ablation (or a gradual reduction from 20 to 80%along the glacier) is needed to cause the required 3 km of advancefrom near Canavan's Knob to an equilibrium position at the WaihoLoop moraine (Fig. 9), assuming constant climate; based on the abovedata this appears reasonable. A coseismic rock avalanche of therequired volume in the Waiho valley is certainly possible; Tovar et al.(2008) provided evidence of sufficiently large deep-seated, source-area scars in the correct location to match the dominant Looplithology. Similar events have been reported elsewhere (e.g., Post,1967; McSaveney, 1975, 2002; Gordon et al., 1978; Evans and Clague,1988; Orombelli and Porter, 1988; Gardner and Hewitt, 1990; Jibsonet al., 2004, 2006; Lipovsky et al., 2008; Deline, 2009; Hewitt, 2009).

Vacco et al. (2010) recently presented a one-dimensionalnumerical model of the Franz Josef Glacier. Though approximate, itconfirmed that a debris driven advance could reach the Waiho Loopunder certain circumstances and, significantly, that the resultingdeposition would have a prominent, high end moraine with anextensive blanket of ablation moraine to its rear, as appears to be thecase at the Waiho Loop based on available evidence (Shulmeisteret al., 2010).

In summary, there is substantial evidence that a rock avalanchedeposit covering a sufficient proportion of an ablation zone can causeglaciers mass balance fluctuations and its nonclimatically drivenadvance.

5. Consequences

5.1. Terminal moraine formation

Here we describe theoretically how a rock avalanche may causeglacial advance and frontal moraine formation without any hin-drances. A rock–avalanche-driven glacier advance will be short-lived;perhaps in the order of decades, compared to centuries to amillennium in the case of an advance driven by climate such as theYounger Dryas or the Little Ice Age. As the rear of the debris deposit iscarried downvalley by ice flow, ice-surface ablation will increase andmass balance will gradually become less positive. It will return tonormal conditions, or the mass balance as before the event, when allthe debris has been carried past the original terminus. Then theadvance of the detached debris-covered ice will eventually slow andstop, during which time substantial quantities of rock avalanchedebris will reach the new terminus and form amoraine. The sequenceof events is illustrated conceptually in Fig. 10; ice covered with rockavalanche deposit reduces in elevation less slowly because ofsuppressed melting than the previous clean or thinly-covered ice, soit flows more rapidly. In addition, ice at the terminus has compressiveflow which is less rapid than that farther upvalley, so the length ofdebris-covered ice reduces and its thickness increases accordingly.

Eventually the original glacier ablation zone is debris-free because allthe debris has been carried beyond the original terminus. The debris-covered ice downvalley is now independent of the glacier andcontinues to flow slowly under its own surface gradient while slowlymelting, forming a terminal moraine; it eventually becomes stagnantand slowly melts to form a sheet of ablation moraine to the rear of theterminal moraine. This form is similar to that generated by the modelof Vacco et al. (2010), and is strikingly similar to that displayed by thesmall Classen terminal moraine in New Zealand shown in Fig. 11.

Supraglacial rock avalanche debris is normally between 1 and10 m thick; this allows estimation of the size of a rock avalancherequired to cover a specified proportion of the ablation zone of a givenglacier and suppress ablation accordingly. For example, the Franz JosefGlacier ablation zone area is about 6 km2, so a rock avalanche of theorder of 106m3 in volume would be needed to cover 10% of theablation zone to a depth of N1 m and cause a significant change inglacier behaviour. Such a volume of rock avalanche deposit arriving atthe approximately 0.4 km wide terminus could form a moraine ofcross-sectional area ~103–104 m2; this would certainly be a promi-nent feature in the landscape.

How moraines are formed by climatic glacier advances is alsoincompletely understood. Bulldozing is able to generate small “push”moraines of the order of several metres high (Molnia, 2008; Winklerand Matthews, 2010), while sluicing of debris by subglacial waterflow generates characteristic stratified “outwash” moraines (Hyatt,2010); however, “dumping” of supraglacial debris, which can buildvery high and steep lateral moraines (Benn and Evans, 2010) seems tobe themost likely potential cause of very large, steep-fronted terminalmoraines such as the Waiho Loop (Vacco et al., 2010), and therequirement for large volumes of supraglacial debris to do thissuggests that a rock avalanche may be the most likely source of suchmoraines.

Fig. 11. Classen moraines, Godley Valley, NZ, 2009; the upper terminal moraine has theform indicated by Vacco et al. (2010) and by Fig. 10 (Photo by S. Winkler).

336 N.V. Reznichenko et al. / Geomorphology 132 (2011) 327–338

Given the potential for rock avalanches to generate terminalmoraines, improved understanding of moraine formation processesby rock–avalanche-driven advances is an urgent requirement.

5.2. Paleoclimatology

Terminal moraines are commonly used as paleoclimatic indicators.This method is based on the assumption that all moraines owe theirexistence to climatically driven advance-retreat sequences, but ourwork above clearly shows this assumption to be potentially in error.Any purely rock–avalanche-generated moraine whose date isassigned paleoclimatic significance generates an error in the inferredpaleoclimate record. Of course, a rock avalanche can fall onto a glacierthat is in the course of a climatically driven advance-retreat sequence,in which case the moraine age will have paleoclimatic relevance.

Terminal moraines form when a glacier reaches an advancedposition; when it retreats again the moraine records the previousmaximum extent (Benn and Evans, 2010). If a glacier terminusundergoes an advance-retreat sequence and is supplied withsediment by any process, a terminal moraine is likely to form thatwill remain after retreat has begun. The greater the volume ofsediment able to be delivered to the moraine, the larger it will be.Thus, a rock–avalanche-driven advancewill be capable of generating asubstantial terminal moraine because it involves a very large volumeof debris. Such a moraine will have no necessary association with anyclimatic change (Larsen et al., 2005; Tovar et al., 2008; Shulmeister etal., 2009). Even if a rock avalanche deposit does not cause a glacier toadvance, the eventual arrival of a large volume of supraglacial debrisat the terminus can form a large moraine in a short time, whereas toform the same size of moraine without rock avalanche debris wouldrequire the terminus to remain in one position for a relatively longtime.

In the same way, moraines left by surge-type and tidewaterglaciers may record nonclimatic ice-front fluctuations (Yde andPaasche, 2010). Therefore the terminal moraines that are used aspaleoclimatic indicators must be known to be generated by climaticvariation rather than by nonclimatic processes. To date, however, nosuch investigations of this topic have been reported, and all terminalmoraines are assumed de facto to result from climatic variations.Given the crucial contemporary importance of paleoclimatology inunderstanding climate change, this is potentially a serious deficiency.The various difficulties outlined herein can only be resolved when theprocesses of formation of both climatically driven and rock–avalanche-driven moraines are much better understood, so thatcriteria can be developed for distinguishing them in the field.

5.3. Effects on society

The potential for large rock avalanches to affect some present-dayglaciers may resulted in significant impacts on society and business,because of the increasing level of tourism-related human activity inthe vicinity of accessible glaciers (which are often the larger ones thatextend to low altitudes). A good example is the Franz Josef Glacierbecause of its location in a tectonically active setting that frequentlyexperiences intense seismicity (Shulmeister et al., 2009) and becauseof its popularity as a tourist destination owing to its easily accessibleterminus (at ~300 masl).

Catastrophic landslides have previously been caused by earth-quakes on the Alpine fault, a 700-km-long plate boundary faultrunning along the western rangefront of the Southern Alps of NewZealand (e.g., Chevalier et al., 2009; Dufresne et al., 2010). This faultgenerates earthquakes of M~8 several times per millennium onaverage, and the last one was in 1717 according to paleoseismicinformation (Rhoades and Van Dissen, 2003). The large number ofdeep-seated landslide source-area depressions are observed in theSouthern Alps, whose morphology indicates that their origin wasmost likely coseismic (Tovar et al., 2008). Shulmeister et al. (2009)suggested that a large rock avalanche could be expected in the orderof once per millennium on average in the Franz Josef Glaciercatchment. However, given that since 1999 six rock avalanches haveoccurred in the Southern Alps, all of the order of 107m3, and nonewere triggered by either earthquake or rain (Aoraki/Mt. Cook and Mt.Fletcher I and II; McSaveney, 2002; Mt. Adams; Hancox et al., 2005;Young River; unreferenced; Joe River; unreferenced), the probabilityof any kind of rock avalanche in any specific mountain valley is clearlynontrivial.

The ever-present and immediate hazard from a rock avalancheonto the Franz Josef Glacier (and perhaps running out along theproglacial upper Waiho valley) is to the lives of people on the glacierand in the valley (several hundred visitors at any one time on a fineday), since warning of such an event may only be a minute or so forpeople at the terminus and less for people on the glacier. Thistherefore represents a not insignificant threat to peoples’ safety, giventhe ~250,000 tourists that visit the valley, terminus, and ice surfaceannually (Corbett, 2001).

To estimate the potential effect of a rock avalanche on the present-day glacier, Alexander et al. (2011) assumed that debris from a rockavalanche in themid-upper catchment is voluminous enough to coverthe entire present-day ablation zone (area about 6 km2) to metre-scale depth . This would require a volume of about 107m3. If totalablation were reduced by 80% as a result, the modelling indicates thatthe Franz Josef Glacier would advance almost as far as the township(about 6 km; Fig. 9) if the present climate conditions and accumu-lation rates prevailed. The speed of a rock–avalanche-driven terminusadvance can be estimated from the known rate of ice accumulation inthe catchment, and appears to be of the order of 1 km/y (Tovar et al.,2008). Apparently, the glacier could possibly advance to the vicinity ofFranz Josef township in perhaps a decade; human life would not bethreatened by such an advance, but the main highway along the westcoast of South Island would be. The consequent effects on thebehaviour of the already troublesome Waiho and Callery rivers(McSaveney and Davies, 1998; Davies and McSaveney, 2001) wouldalso need to be investigated—when the Franz Josef Glacier advancedabout 1.5 km between 1982 and 1999, the proglacial area aggraded ofthe order of tens of metres. Clearly, there is the potential for majordisruption of tourism and other business in this iconic touristdestination.

This scenario has not to date been part of any assessment carriedout for the area, and, therefore, no management strategies aredeveloped. A similar threat could exist for the locality of the FoxGlacier, 20 km south of the Franz Josef Glacier, whose glacial andgeomorphological characteristics are similar. Evidently, if a rock–

337N.V. Reznichenko et al. / Geomorphology 132 (2011) 327–338

avalanche-generated glacier advance were to approach Franz Joseftownship, the carefully planned, ordered, and structured relocation ofthe settlement and of the highway are required. No active mitigationmethods appear to be practicable. The requirement for furtherinvestigation of this scenario is clear.

6. Conclusions

(i) Large rock avalanches falling onto glaciers can affect glacierbehaviour significantly by reducing the total ablation beneaththe deposit by up to 90%, causing substantial changes in glacierbehaviour such as advances and surging.

(ii) Rock–avalanche-driven glacier advances can form prominentterminal moraines.

(iii) Terminal moraines generated by rock–avalanche-driven ad-vances have no necessary climatic association; therefore, thevalidity of paleoclimatic inferences from moraines depends onidentifying the origin of the moraine.

(iv) Further investigation is required into the processes of terminalmoraine formation caused by rock–avalanche-driven andclimatically driven advances and into criteria for distinguishingmoraines formed by these processes.

(v) Rock–avalanche-driven glacier advances have the potential tocause serious disruption to society and business over decadaltimescales.

Acknowledgements

Stefan Winkler's comments on an early draft, together with thoseof three anonymous reviewers, led to substantial improvements of themanuscript, for which we are grateful.

References

Alexander, D.J., Shulmeister, J., Davies, T.R.H., 2011. A steady-state mass-balance modelfor the Franz Josef glacier, New Zealand: testing and application. GeografiskaAnnaler A 93, 41–54.

Alexander, D.J., Shulmeister, J., Davies, T.R.H., in review. High basal melting rates withinhigh-precipitation temperate glaciers. Journal of Glaciology.

Allen, S.K., Cox, S.C., Owens, I.F., 2011. Rock avalanches and other landslides in thecentral Southern Alps of New Zealand: a regional study considering possibleclimate change impacts. Landslides 8 (1), 33–48.

Anderson, B., Mackintosh, A., 2006. Temperature is the major driver of lateglacial andHolocene glacier fluctuations in New Zealand. Geology 34 (2), 121–124.

Barrows, T.T., Lehman, S.J., Fifield, K.L., De Deckker, P., 2007. Absence of cooling in NewZealand and the adjacent ocean during the Younger Dryas chronozone. Science 318,86–89.

Benn, D.I., Evans, D.J.A., 2010. Glaciers and Glaciation, 2nd ed. Hodder Education,London. 802 pp.

Bozhinskiy, A.N., Drass, M.S., Popovin, V.V., 1986. Role of debris cover in the thermalphysics of glaciers. Journal of Glaciology 32, 255–266.

Brock, B.W., Mihalcea, C., Kirkbride, M.P., Diolaiuti, G., Cutler, M.E., Smiraglia, C., 2010.Meteorology and surface energy fluxes in the 2005–2007 ablation seasons at Miagedebris-covered glacier, Mont Blanc Massif, Italian Alps. Journal of GeophysicalResearch 115, D09106. doi:10.1029/2009JD013224.

Carter, L., Manighetti, B., Neil, H., 2003. From icebergs to pongas: Antarctica's ocean linkwith New Zealand. Water and Atmosphere online 11 (3), 30–31.

Chevalier, G.G., Davies, T.R.H., McSaveney, M.J., 2009. The prehistoric Mt. Wilberg rockavalanche, Westland, New Zealand. Landslides 6, 253–262. doi:10.1007/s10346-009-0156-5.

Corbett, R., 2001. Social impact issues among visitors to Franz Josef Glacier, WestlandNational Park. Science and Research Internal Report 186, Department ofConversation, Wellington, NZ. 34 pp.

Cox, S.C., Ferris, B.G., Allen, S., 2008. Vampire Rock Avalanches, Aoraki/Mount CookNational Park, New Zealand. GNS Science Report 2008/10, Institute of Geologicaland Nuclear Sciences Limited, Lower Hutt, NZ. 34 pp.

Davies, T.R.H., McSaveney, M.J., 2001. Anthropogenic fanhead aggradation, WaihoRiver, Westland, New Zealand. In: Mosley, M.P. (Ed.), Gravel-Bed Rivers V. : NewZealand Hydrological Society. Wellington, NZ, pp. 531–553.

Davies, T.R.H., McSaveney, M.J., 2009. The role of dynamic rock fragmentation inreducing frictional resistance to large landslides. Engineering Geology 109, 67–79.doi:10.1016/j.enggeo.2008.11.004.

Davies, T.R.H., Smart, C.C., 2007. Obstruction of subglacial conduits by bedload sediment—implications for alpine glacier motion. Journal of Hydrology NZ 46 (2), 51–62.

Deline, P., 2009. Interactions between rock avalanches and glaciers in the Mont Blancmassif during the late Holocene. Quaternary Science Reviews 28, 1070–1083.

Drewry, D., 1986. Glacial Geologic Processes. Edward Arnold, London. 276 pp.Dufresne, A.M., Davies, T.R.H., McSaveney, M.J., 2010. Influence of runout path material

on the emplacement of the Round Top rock avalanche, New Zealand. Earth SurfaceProcesses Landforms 35, 190–201. doi:10.1002/esp. 1900.

Evans, S.G., Clague, J.J., 1988. Catastrophic rock avalanches in glacial environments.Proc. Fifth Int. Symp. on Landslides 2, 1153–1158.

Fort, M., Cossart, E., Delin, P., Dzikowski, M., Nicoud, G., Ravanel, L., Schoeneich, P.,Wassmer, P., 2009. Geomorphic impact of large and rapid mass movements: areview. Geomorphology: relief, processus, environment 1, 47–64.

Gardner, J.S., Hewitt, K., 1990. A surge of Bualtar Glacier, Karakoram Range, Pakistan: apossible landslide trigger. Journal of Glaciology 36 (123), 159–162.

Geertsema, M., Clague, J.J., Schwab, J.W., Evans, S.G., 2006. An overview of recent largecatastrophic landslides in northern British Columbia, Canada. Engineering Geology83, 120–143.

Gordon, J.E., Birne, R.V., Timmis, R., 1978. A major rockfall and debris slide on the LyellGlacier, South Georgia. Arctic and Alpine Research 10 (1), 49–60.

Hancox, G.T., McSaveney, M.J., Manville, V.R., Davies, T.R.H., 2005. The October 1999Mt.Adams rock avalanche and subsequent landslide dam-break flood and effects inPoerua River, Westland, New Zealand. NZ Journal of Geology & Geophysics 48,683–705.

Hermanns, R.L., Niedermann, S., Garcia, A.V., Gomez, J.S., Strecker, M.R., 2001.Neotectonics and catastrophic failure of mountain fronts in the southern intra-Andean Puna Plateau, Argentina. Geology 29, 619–622.

Hewitt, K., 1988. Catastrophic landslide deposits in the Karakoram Himalaya. Science242 (4875), 64–67.

Hewitt, K., 1998. Catastrophic landslides and the Upper Indus streams, KarakoramHimalaya, northern Pakistan. Geomorphology 26, 47–80.

Hewitt, K., 2009. Rock avalanches that travel onto glacier and related developments,Karakoram Himalaya, Inner Asia. Geomorphology, Elsevier 103 (1), 66–79.

Hyatt, O., 2010. Insights into New Zealand Glacial Processes from Studies of GlacialGeomorphology and Sedimentology in Rakaia and Other South Island Valleys.University of Canterbury, Christchurch, NZ, PhD Dissertation.

Jibson, R., Harp, E., Schulz, W., Keefer, D.K., 2004. Landslides triggered by the 2002Denali Fault, Alaska, earthquake and the inferred nature of the strong shaking.Earthquake Spectra 20 (3), 669–691.

Jibson, R.W., Harp, E.L., Schulz,W., Keefer, D.K., 2006. Large rock avalanches triggered bythe M 7.9 Denali Fault, Alaska, earthquake of 3 November 2002. EngineeringGeology 83, 144–160.

Kariya, Y., Sato, G., Komori, J., 2011. Landslide-induced terminal moraine-like landformson the east side of Mount Shiroumadake, Northern Japanese Alps. Geomorphology127, 156–165.

Kayastha, R.B., Takeuchi, Y., Nakawo, M., Ageta, Y., 2000. Practical prediction of the icemelting beneath various thickness of debris cover on Khumbu Glacier, Nepal usinga positive degree-day factor. IAHS Publ. 264 (Symposium at Kathmandu, Nepal1992—Snow and Glacier Hydrology), pp. 71–81.

Kirkbride, M.P., 1989. The influence of sediment budget on the geomorphic activity ofthe Tasman Glacier, Mount Cook National Park. New Zealand. Unpublished Ph.D,Dissertation, University of Canterbury, Christchurch, NZ.

Kirkbride, M.P., 1993. The temporal significance of transitions from melting to calvingat glaciers in the central Southern Alps of New Zealand. Holocene 3, 232–240.

Korup, O., Clague, J.J., 2009. Natural hazards, extreme events, and mountaintopography. Quaternary Science Reviews 28, 977–990.

Larsen, S., Davies, T.R.H., McSaveney, M.J., 2005. A possible coseismic landslide origin oflate Holocene moraines of the Southern Alps, New Zealand. New Zealand Journal ofGeology and Geophysics 48, 311–314.

Lee, J., Davies, T.R., Bell, D.H., 2009. Successive Holocene rock avalanches at LakeColeridge, Canterbury, New Zealand. Landslides 6, 287–297.

Lipovsky, P.S., Evans, S.G., Clague, J.J., Hopkinson, C., Couture, R., Bobrowsky, P., Ekström,G., Demuth, M.N., Delaney, K.B., Roberts, N.J., Clarke, G., Schaeffer, A., 2008. The July2007 rock and ice avalanches at Mount Steele, St. Elias Mountains, Yukon, Canada.Landslides 5 (4), 445–455. doi:10.1007/s10346-008-0133-4.

Lundstrom, S.C., McCaferty, A.E., Coe, J.A., 1993. Photogrammetric analysis of 1984–89surface altitude change of the partially debris-covered Eliot glacier, Mount Hood,Oregon, USA. Annals of Glaciology 17, 167–170.

Malamud, B.D., Turcotte, D.L., Guzzetti, F., Reichenbach, P., 2004. Landslide inventoriesand their statistical properties. Earth Surface Processes and Landforms 29, 687–711.

Mattson, L.E., Gardner, J.S., Young, G.J., 1993. Ablation on debris-covered glaciers: anexample from the Rakhiot Glacier, Punjab, Himalaya. IAHS Publ. 218 (Symposiumat Kathmandu, Nepal 1992—Snow and Glacier Hydrology), pp. 289–296.

McColl, S.T., Davies, T.R.H., 2011. Evidence for a rock–avalanche origin for The Hillocksmoraine, Otago, New Zealand. Geomorphology 127 (3–4), 216–224.

McColl, S.T., Davies, T.R.H., McSaveney, M.J., 2010. Glacier retreat and rock-slopestability: debunking debuttressing. In: Williams, A.L., Pinches, G.M., Chin, C.Y.,McMorran, T.J., Massey, C.I. (Eds.), Geologically Active. Proceedings of the 11thIAEG Congress. Auckland, NZ.

McSaveney, M.J., 1975. The Sherman glacier rock avalanche of 1964: its emplacementand subsequent effects on the glacier beneath it. Dissertation, Ohio StateUniversity, Columbus, Ohio, Ph.D. 403 pp.

McSaveney, M.J., 1978. Sherman glacier rock avalanche, Alaska, U.S.A. In: Voight, B.(Ed.), Rockslides and avalanches 1. Elsevier, NY, pp. 197–258.

McSaveney, M.J., 2002. Recent rockfalls and rock avalanches in Mount Cook NationalPark, New Zealand. In: Evans, S.G., DeGraff, J.V. (Eds.), Catastrophic landslides:effects, occurrence, and mechanisms: Geological Society of America Reviews inEngineering Geology, 15, pp. 35–70.

338 N.V. Reznichenko et al. / Geomorphology 132 (2011) 327–338

McSaveney, M.J., Davies, T.R.H., 1998. Natural Hazard Assessment for the township ofFranz Josef Glacier and its Environs. Client Report 43714B.10, Institute of Geologicaland Nuclear Sciences, Lower Hutt, NZ. 58 pp.

McSaveney, M.J., Davies, T.R.H., 2007. Rockslides and their motion. In: Sassa, K.,Fukuoka, H., Wang, F., Wang, G. (Eds.), Progress in Landslide Science Springer-Verlag, pp. 113–134.

Melosh, H.J., 1987. The mechanics of large rock avalanches. Geological Society ofAmerica Reviews in Engineering Geology 7, 41–49.

Molnia, B.F., 2008. Glaciers of North America. Glaciers of Alaska. In: Williams Jr., R.S.,Ferrigno, J.G. (Eds.), Satellite image atlas of glaciers of the world: U.S. GeologicalSurvey Professional Paper 1386-K. 525 pp.

Nakawo, M., Rana, B., 1999. Estimate of ablation rate of glacier ice under a supraglacialdebris layer. Geogr. Ann. 81A (4), 695–701.

Nakawo, M., Young, G.J., 1981. Field experiments to determine the effect of a debrislayer on ablation of glacier ice. Annals of Glaciology 2, 85–91.

Nakawo, M., Young, G.J., 1982. Estimate of glacier ablation under a debris layer fromsurface temperature and meteorological variables. Journal of Glaciology 28 (98),29–34.

Nicholson, L., Benn, D.I., 2006. Calculating ice melt beneath a debris layer usingmeteorological data. Journal of Glaciology 52 (178), 463–470.

Orombelli, G., Porter, S.C., 1988. Boulder deposit of upper Val Ferret (Courmayeur, Aostavalley): deposit of a historic giant rockfall and debris avalanche or a late-glacialmoraine? Eclogae geol Helv. 82 (2), 365–371.

Østrem, G., 1965. Problem of dating ice-cored moraines. Geogr. Ann. 47A (1), 1–38.Paterson, W.S.B., 1994. The Physics of Glaciers, 3rd ed. Pergammon, NY. 480 pp.Post, A.S., 1965. Alaskan glaciers: recent observations in respect to the earthquake-

advance theory. Science 148 (3668), 366–368.Post, A.S., 1967. Effects of the March 1964 Alaska earthquake on glaciers. United States

Geological Survey Professional Paper 544-D. . 42 pp.Post, A.S., 1968. Effects on Glaciers: in the great Alaska earthquake of 1964: Hydrology,

Pt. A. Nat. Acad. Sci. Pub. 1603, 266–308.Quincey, D.J., Glasser, N.F., 2009. Morphological and ice-dynamical changes on the

Tasman Glacier, New Zealand, 1990–2007. Global and Planetary Change 68,185–197.

Reid, J.R., 1969. Effects of a debris slide on “Sioux Glacier”, south-central Alaska. Journalof Glaciology 8 (54), 353–367.

Reid, T.D., Brock, B.W., 2010. An energy-balance model for debris-covered glaciersincluding heat conduction through the debris layer. Journal of Glaciology 56,903–916.

Reznichenko, N.V., Davies, T.R.H., Shulmeister, J., McSaveney, M.J., 2010. Effects ofdebris on ice-surface melting rates: an experimental study. Journal of Glaciology56, 385–394.

Rhoades, D., Van Dissen, R.J., 2003. Estimates of the time varying hazard of rupture ofthe Alpine Fault, New Zealand, allowing for uncertainties. New Zealand Journal ofGeology and Geophysics 46, 479–488.

Rogerson, R.J., Olson, M.E., Branson, D., 1986. Medial moraines and surface melt onglaciers of the TorngatMountains, Northern Labrador, Canada. Journal of Glaciology32 (112), 350–354.

Schytt, A., 1967. A study of ablation gradient. Geogr. Ann. 49A, 327–332.Sharp, R.P., 1949. Studies of superglacial debris on valley glacier. American Journal of

Science 22 (86), 43–52.Shulmeister, J., Davies, T.R.H., Evans, D.J.A., Hyatt, O.M., Tovar, D.S., 2009. Catastrophic

landslides, glacier behaviour and moraine formation—a view from an active platemargin. Quaternary Science Reviews 28, 1085–1096.

Shulmeister, J., Davies, T.R.H., Reznichenko, N.V., Alexander, D.J., 2010. Comment on glacialadvance and stagnation caused by rock avalanches, by Vacco DA. Alley RB, Pollard DEarth and Planetary Science Letters 297, 700–701. doi:10.1016/j.epsl.2010.06.046.

Smith, G.M., Davies, T.R., McSaveney, M.J., Bell, D.H., 2006. The Acheron rock avalanche,Canterbury, New Zealand—morphology and dynamics. Landslides 3, 62–72.

Soldati, M., Corsini, A., Pasuto, A., 2004. Landslides and climate change in the ItalianDolomites since the Late glacial. Catena 55 (2), 141–161.

Tarr, R.S., Martin, L., 1914. Alaskan Glacier Studies. National Geographical Society,Washington. 498 pp.

Tovar, D.S., Shulmeister, J., Davies, T.R.H., 2008. Evidence for a landslide origin of NewZealand's Waiho Loop Moraine. Nature Geosciences 1 (8), 524–526.

Turnbull, J.M., Davies, T.R.H., 2002. Subglacial sediment accumulation and basalsmoothing—a possible mechanism for initiating glacier surging. Journal ofHydrology NZ 41 (2), 105–123.

Turney, C.S.M., McGlone, M.S., Wilmshurst, J.M., 2003. Asynchronous climate changebetween New Zealand and the North Atlantic during the last deglaciation. Geology31 (3), 223–226.

Vacco, D.A., Alley, R.B., Pollard, D., 2010. Glacier advance and stagnation caused by rockavalanches. Earth and Planetary Science Letters 294, 123–130.

Voight, B., Pariseau, W.G., 1978. Rockslides and avalanches: an introduction. Rockslideand avalanches 1 Natural Phenomena. Elsevier, NY. 843 pp.

Whitehouse, I., Griffiths, G., 1983. Frequency and hazard of large rock avalanches in thecentral Southern Alps, New Zealand. Geology 11, 331–334.

Winkler, S., 2009. Gletscher und ihre Landschaften. Wissenschaftliche Buchgesellschaft,Darmstadt. 183 pp.

Winkler, S., Matthews, J.A., 2010. Observations on terminal moraine-ridge formation duringrecent advances of southern Norwegian glaciers. Geomorphology 116 (1–2), 87–106.

Yde, J.C., Paasche, Ø., 2010. Reconstructing climate change: not all glaciers suitable. Eos,Transactions of the American Geophysical Union 91 (21, 25), 189–196.