• -* I

Stripping of Keanakakoi tephra on Kilauea Volcano,

MICHAEL C. MALIN Department of Geology, Arizona State University, Tempe, Arizona 85287 DANIEL DZURISIN U.S. Geological Survey, Cascade Volcano Observatory, Vancouver, Washington 98661 ROBERT P. SHARP Division of Geological and Planetary Sciences, California Institute of Technology, Pasadena, California 91125

ABSTRACT

Morphological characteristics, erosional processes, and effects of burial and exhuma­ tion by debris mantles on basaltic volcanic landforms have been evaluated through field study of the Keanakakoi Formation, a I basaltic tephra formed in 1790 by phreato­ km magmatic eruptions from Kilauea caldera, Hawaii. The upper coarse lithic, interme­ N I diate fine vitric, and lower mixed members of this formation play different roles in the creation of micro-terrain elements during 1 stripping of tephra from the underlying I bedrock. Of the seven micro-terrain ele­ ments defined, bedrock, scabby upland surfaces, and lag gravels are the most dis­ tinctive and widely distributed. Different proportions and combinations of micro­ f j terrain elements define five zones of pro­ l gressive deterioration of the Keanakakoi tephra blanket southwestward from Kilauea caldera into the Kau Desert. Fluvial and eolian processes operate on different time scales and at different locations, governed by blanket thickness, debris caliber, and the formation of case-hardened crusts. Strip­ ping of an entire mantle is probably not v possible; however, materials trapped within depressions form the only clearly discerni­ SYMBOLS: Benchmarks ble morphological expression of previously /--' Fault scarps, bar and ball more extensive debris blankets. on downthrown side ~ Cracks 0 Conspicuous flows INTRODUCTION Figure 1. Sketch map of Kau Desert, Kilauea Volcano, Hawaii, showing five zones of Burial of surface terrains by a friable degradation of the 1790 tephra blanket. I. Complete cover by tephra. II. Areas of small debris blanket and their subsequent exhu­ bedrock outcrops. III. Areas of continuous but fractional bedrock outcrop. IV. Bedrock mation may have been widespread on Mars surface with small insets of debris. V. Bedrock and reworked accumulations and dunes. (Soderblom and others, 1973; Malin, 1976), Abbreviations: FPT = Footprint Trail, CP = Cone Peak, SH = Sand Hill, H = Halemau­ with important effects on the appearance of mau, HVO = Hawaiian Volcano Observatory, VH = Volcano House, KI = Kilauea Iki, K features perceived from orbit. Martian Keanakakoi (), AK = Ahua Kamokukolau, MIT = Mauna Iki Trail.

Geological Society of America Bulletin, v. 94, p. 1148-1158, II figs., October 1983. 1148 I STRIPPING OF TEPHRA, KILAUEA VOLCANO, HAWAII 1149

blanketing material may include weathering Kau Desert is a 350-km2 wedge apexing at the entire unit (> 5 m) to the 1790 event. His products, rocks comminuted by meteorite Kilauea caldera with a longitudinal axis interpretation was supported by recently impact, and tephra that was formed by vol­ along the Southwest Rift Zone (Fig. I). Sur­ acquired 14C dates which indicated a basal canic eruptions and emplaced (Binder and face materials are mostly thin, young basalt age younger than 350 ± 60 yr (Kelley and - others, 1977; Mutch and others, 1977) and flows, bedded tephra, and deposits of others, 1979). partly removed by wind (McCauley, 1973). reworked tephra debris. After nearly a century of lava-lake activ­ Objectives of the present study were (I) to A major phreatomagmatic eruption ity, Halemaumau erupted explosively on 91125 examine morphological characteristics of a within Kilauea caldera in 1790 A.D. dis­ May II, 1924, and continued to eject frag­ partly eroded tephra blanket on a fresh vol­ tributed tephra over the Kau Desert, mental lithic material and copious steam for canic landscape, (2) to identify the erosional creating the Keanakakoi Formation of 14 days. During the peak of the 1924 erup­ processes, and (3) to determine the effects of Wentworth (1938). A less violent phreatic tion, rhythmic explosions ejected blocks burial and exhumation on the surficial eruption within Halemaumau during 1924 weighing several metric tons, ash-laden appearance of the underlying . Lavas formed a thin tephra layer at the caldera steam clouds rose at least 2 km, and more in the Kau Desert are primarily pahoehoe, rim. Reports concerning the 1790 eruption than 30 cm of lithic tephra accumulated and so attention necessarily focuses upon consist of sketchy retrospective narratives locally on the south rim of Halemaumau. A that type of surface, although tephra has by natives as related to missionaries. An few centimetres of finer lithic debris barely locally mantled aa lavas, and such relation­ eruptive column several kilometres high was cleared the southwest edge of Kilauea cal­ ships are treated briefly. visible, and a group of Hawaiian warriors dera (Powers, 1948, PI. 3D; Macdonald, were suffocated 9 km downwind from the 1949, p. 72; Macdonald and Abbott, 1970, Setting caldera. This tragedy and many charact&is­ p. 315). An eyewitness account by Stearns tics of the deposits have been interpreted to (1925) recorded a heavy fall of ash at Kilauea is an active basaltic shield vol­ support the hypothesis that base surge Pahala, 34 km southwest, and of sharp, cano constituting the southeast part of played a significant role in emplacement of angular, sand-sized fragments, as large as I Hawaii Island. Major orographic features Keanakakoi tephra (Swanson and Christ­ mm, at nearby Wood Valley. Of especial include the 3 by 5 km summit caldera iansen, 1973). Powers (1948) regarded only interest are showers of accretionary lapilli with its nested pit crater, Halemaumau, the uppermost layers « I m) of the Keana­ (also called pisolites) reported by Stearns and prominent east-west-trending and kakoi Formation as a proQuct of the 1790 (1925, p. 202), 2.5 km southwest of Hale­ southwest-trending radial rift zones. The eruption, but Christiansen (1979) attributed maumau 2 hr after an explosion.

" KEANAKAKOI TEPHRA BLANKET

Stratigraphy

Keanakakoi Formation consists of ash, lapilli, and blocks of accessory material as well as essential vitric ash and pumice (Wentworth, 1938, p. 92-102). In Kau Intermediate Desert, this tephra blanket is heterogeneous, Vitric although well stratified. R. L Christiansen Unit (U.S. Geological Survey, 1979, written commun.) has studied in detail the stratig­ raphy, emplacement, mode of origin, and age of these deposits. For our purposes, a Lower simplified threefold definition of the forma­ Mixed tion is adopted: (I) an upper, coarse lithic Unit unit; (2) an intermediate, finer, and richly vitric unit; and (3) a lower, mostly fine, mixed lithic and vitric unit (Fig. 2). The Lava Flows upper unit is predominantly lithic, but IWS the intermediate and lower units contain admixtures of vitric, lithic, and crystal zones of o 2 meters fragments. [';-!',.,I Pumice I I of small Upper Lithic Unit. This is the most Bedrock UC - Unconformity distinctive unit of the three, ranging from Id dunes. I mm to several metres thick and consisting ialemau­ Figure 2. Simplified section of the three major units of the Keanakakoi Formation: an almost wholly of sand- to block-sized angu­

a Iki, K 0 upper lithic unit principally sand- to block-sized angular fragments of accessory rock, an lar fragments of dense, nonvesicular acces­ Intermediate vitric unit ofsilt- to sand-sized volcanic ash fragments (with minor amounts of sory rocks, largely olivine basalt, picrite lithic, crystal, and pumiceous debris), and a lower mixed unit similarly composed of sand­ basalt, diabase, and gabbro (Macdonald, and silt-sized vitric, lithic, and crystalline materials. Unconformities between each unit. 1949, p. 65). Scattered blocks as much as &

1150 MALIN AND OTHERS

I m in size are seen near the caldera rim, orifices around Kilauea, deposit sulfates loose, angular lithic fragments, for the most fragments of 20 to 40 em in diameter are and opaline silica (Naughton and others, part about 0.5 em in diameter, interrupt the abundant within I km of the caldera rim, 1976), and similar substances might be sequences of finer beds. Occasional lithic and particles from 0.5 to 1.5 em prevail derived from seepage waters. Crust forma­ blocks as large as 10 em are seen in such throughout. This unit contains a few thin tion is a continuing process, for it occurs on layers. Pumice fragments, 0.5 to 2 em in (2-4 em) layers of brown, silt-sized fine vol­ erosional facets of various ages cutting diameter, are particularly abundant in the canic ash locally rich in accretionary lapilli. across bedding and upon the walls of recent lower half of the intermediate unit, and Bedding, although crude, irregular, and gullies and cracks. A thinner, but somewhat some layers as much as 10 em thick are discontinuous, is nonetheless prevailing. less coherent, crust has already developed nearly pure pumice. The most obvious crys­ Primary dips to 15° are abundant owing to on 1924 tephra, indicating rapid formation. tal fragments are olivine, although feldspar scour fillings and to draping over an ir­ Intermediate Vitric Unit. This unit can be also may be present. regular surface. Cross-bedding occurs in as much as 2 m thick near the caldera rim, In many exposures, the uppermost part some layers, on a 10- to 20-cm scale, and in although extremely variable and sometimes of the intermediate unit is a massive, yellow­ I· channel fillings near the caldera rim. entirely missing. Fine grain size, high vitric brown, silt-sized fine ash, rich in accretion­ The upper lithic unit rests unconformably content, and a khaki brown color are char­ ary lapilli in its uppermost part. Tracing 01 upon the intermediate unit, as shown by acteristic. Fine lamination on a millimetre strata through the Kau Desert to the Foot­ local angular truncation of bedding (Fig. 2). scale is common, and layering is generally print locality (Swanson and Christiansen. or1 Near the caldera rim, this contact is scoured well defined throughout. Coherent layers of 1973), 9 km southwest of the caldera rim, ('( by U-shaped channels, many I m deep and 3 yellowish-brown, silt-sized fine ash (infor­ shows that the lower, desiccation-cracked, Ilr to 4 m wide. One 3-m-deep scour nearly cuts mally termed "mud") are generally thicker, "muddy" pisolitic layer, bearing fossil IIII thorugh the entire underlying section to as much as 20 em, and more abundant than human footprints, is this uppermost "mud" Il. ~ bedrock. Farther out in Kau Desert, thin­ in other parts of the Keanakakoi Forma­ of the intermediate unit. lin ness or absence of the intermediate unit tion. These thicker ashes are relatively mas­ Lower Mixed Unit. This unit is character­ suggests considerable erosion before em­ sive and usually contain accretionary lapilli. ized by layers of relatively well sorted, 1111 placement of the upper lithic unit. Scours 0.5 m deep within this unit occur black, sand-sized coarse ash composed 01 .... w; The surface of the upper lithic unit has a near the caldera rim, and there, at least, the angular lithic, vitric, and crystal fragments, Ito nearly ubiquitous case-hardened crust (Fig. intermediate unit clearly lies upon a scoured loosely packed and free running on exposed \\-'C: 3) and a distinctive reddish-brown color surface cut into the lower unit. faces. A few thin layers of yellow-brown 0.'1 provided by a sand-silt matrix. Deposition The intermediate unit contains lithic and silt-sized fine ash interrupt the coarse I of mineral matter from drifting solfataric crystal fragments in addition to glass and ashes, and the unit is finer toward the bas, fumes or from fluids sweated out of the pumice. Although fine silt- and sand-sized of sections approaching 2 m in thickne" deposits is a likely cause of the crust. grains predominate, well-sorted layers, as near the caldera rim. Sections I to 1.5 111 ICll Solfataric fumes, emanating from many much as several centimetres thick, of clean, thick, several kilometres southwest into >'111 Kau Desert, are predominantly black, sand .. lilt sized coarse ash. "pr The top of the lower unit is marked by " ItlW layer of shiny black, sand-sized coarse ash, .111<1 especially rich in large olivine fragmenh hlTI Fragments of pumice, some as much as ; em, are sparsely scattered throughout. Bcdo, I It'd of coarser lithic fragments, chiefly vesicuJ;11 IIlg basalt as much as several centimetres IIi hI' t· diameter, are occasionally seen. Siderom,· 1'17 _ lane fragments are particularly abunda III I., ( near the base, and some beds, several C,'li ., Iso timetres thick, are composed wholly of thl' lt,tI ( brownish glass. I' I', ('I General Relationships li'lll( 'po

The existing integrated tephra blanket III 'III) the Kau Desert covers approximately 'ill tlC\l'1 2 km , but remnant patches suggest an orii" ,IIUI nal cover of at least 200 km2• The blanket" "10, coarsest and thickest near the caldera 11111. !11lf'al exceeding 6 m on the southwest (KIIII ".1 \t' Desert) side and 10 m on the south rim Wnl ',phr Figure 3. Case-hardened crust on upper lithic unit erodes to form a "scabby" appearance. of Keanakakoi crater (Wentworth, I<)\S: (ldJilr Some stones are firmly held in place by crust; others, loosened by disintegration of crust, p. 93). The tephra blanket is elongated form lag gravel. Pocketknife (lower center) is 9 em long. southwestward, but a section reported to I" iil,III\' j STRIPPING OF TEPHRA, KILAUEA VOLCANO, HAWAII J151 for the most Pahoehoe Pahoehoe smooth and nonvesicular in some places, nterrupt the Lower Tephra Units but fractured or vesiculated in others. lional lithic Weathering produces surficial fracturing ~en in such and spalling that varies with flowage and to 2 cm in location and usually results in greater sur­ dant in the face roughness, except insofar as ropy struc­ : unit, and tures are removed by spalling. Aa bedrock 1 thick are terrain is treated in a subsequent section. lvious crys- gh feldspar Scabby Upland Terrain o 5m rmost part I I One of the more distinctive and areally extensive micro-terrain elements in Kau ve, yellow­ Figure 4. Sketch illustrating onlap unconformity within tephra-filled swale on pahoehoe. accretion­ Desert is formed by a durable, case­ Tracing of hardened crust on the surface of the upper , the Foot­ lithic unit. Undermining of near-horizontal ristiansen, only 15 cm thick (Powers, 1948, p. 288) near wind-blown sand approaching or exceeding resistant layers within this crusted surface Idera rim, Cone Peak, 2 km southwest of the caldera I m in thickness lie below, between, and by weathering, wind, and water produces a n-cracked, rim, is probably not representative, for sec­ above the two "muddy," pisolitic fine ash centimetre-scale, cliff-bench micro-topog­ ing fossil tions 0.5 to 2 m thick were measured within layers bearing fossil footprints (Powers, raphy of frayed and scabby appearance ost "mud" D.5 km east and west of Cone Peak. Local 1948, p. 289; Swanson and Christiansen, (Fig. 3). Further small-scale surface rough­ Impressions of a thinner cover result from 1973, p. 86; Cruikshank, 1974, p. 226). ness results from weathering, rain-beat, and character­ an onlap unconformity of the upper lithic There seems little question that the pisolitic wind action, which etch angular rock frag­ II sorted, unit onto bedrock (Fig. 4). Excavations in ashes represent primary tephra, especially in ments, firmly held in the crust, into positive lposed of swales along Mauna Iki Trail, I km west view of Stearns' (1925, p. 202) observation relief. Some stones rise on small pedestals I ragments, Irom Hilina Pali Road and 5.5 km south­ of mud rains and pisolite falls in Kau Desert or 2 cm high, and interstone areas are 1 exposed wcst of the caldera rim, reveal as much as during the 1924 eruption. Furthermore, a pocked with hollows formed by weathering w-brown, D.5 m of tephra where surface exposures thin layer of primary upper lithic material and by removal of stones formerly em­ : coarser suggest only a few centimetres, owing to the overlies the uppermost footprint-bearing bedded in the matrix. the base onlap relationship. bed. thickness The thickness of material removed from Bedding-Plane Terrain to 1.5 m remnants of the tephra blanket is not MICRO-TERRAIN ELEMENTS lest into k£lawn, but it was probably not great, con­ Small exposures of coherent sand- and ck, sand- Sidering the coarseness and coherence of the Stripping of tephra from a subaerial sur­ silt-sized ash layers of the intermediate and upper lithic unit. Reworking of tephra face is a function of time, initial thickness, lower Keanakakoi units are sparsely scat­ ked by a lower in the section during intervals within constitution (particularly particle size), in­ tered through areas of extensively stripped arse ash, and between eruptive episodes may have duration, substrate configuration, and the pahoehoe. Owing to draping in the lower­ 19ments. heen more significant. Trade winds were nature and power of the stripping processes. most layers, Wentworth's (1926, p. 25; 1938, lch as J lntainly competent to move freshly depos­ Differential stripping of Keanakakoi tephra p. 30) "mantle bedding," bedding-plane ut. Beds lied fine ash and pumice during and follow­ from lavas in the Kau Desert since 1790 has exposures are inclined in all directions at lesicular 1111( eruptive episodes, and blasts generated been controlled principally by thickness and angles approaching 15° and occasionally etres in hI' base surges (Swanson and Christiansen, particle size, as demonstrated by the in­ attaining 28° . Cracks and small depressions jerome­ 1'173) may have redistributed primary teph­ creased degree of stripping outward toward within ropes, festoons, and other irregulari­ )undant Iii Contemporaneous fluvial transport was the margins of the blanket where initial ties of pahoehoe surfaces harbor deposits of ral cen­ ,riso probably active, judging from torren­ thickness was least, and debris was finest. these materials. Nearly all exhumed rock I of this llill downpours witnessed by Stearns (1925, The effects of stripping are conveniently surfaces are stained a distinctive faint mus­ p 196) during the 1924 eruption. described in terms of resulting micro-terrain tard brown. Considerable dust probably has been elements, each displaying different but rea­ Irrl10ved from Kau Desert and largely sonably consistent characteristics. The fol­ Lag Gravel Terrain f ~ ported from the island. Stearns (1925, p. lowing seven micro-terrain elements have nket in 10/ ) described dust clouds rising from the been defined within that part of the Kau Patches of residually accumulated pea- to tely 50 .hert during and after the 1924 eruption, Desert formerly blanketed by the Keanaka­ walnut-sized lithic fragments derived from 10rigi Mid similar events probably occurred in koi Formation. the upper lithic unit constitute an areally mket" I7

1152 MALIN AND OTHERS

five zones of degradation displaying various combinations of the seven micro-terrain elements just described (Fig. I). Usually, only three or four elements occur within a single zone, but zone II, east of Cone Peak, has six.

Zone I

Zone I consists of a nearly continuous, surficially eroded tephra blanket forming a band of 1.5 km wide immediately southwest of Kilauea caldera. Topographically, this zone features a broad surface sloping about I° southwestward, which is sparsely dis­ sected by dendritic gullies as much as 4 m deep and riven by open linear cracks of the Southwest Rift Zone, bearing generally S6QoW. Figure 5. Lag gravel derived from upper lithic unit, capping finer tephra within swale in Micro-terrain elements distinguished pahoehoe. Abundance of pea- to walnut-sized fragments suggests significant removal of within zone I in order of decreasing area fine material presumably by eolian processes. are: (I) scabby upland, (2) lag gravel. (3) fluvial deposits, (4) colluvial mantle, and (5) bedrock. The predominating feature of closing depression: crudely circular or vious owing to crusting on the upper lithic this zone is the cemented crust developed on oblate, elongated, or irregular. Individual unit, thus allowing runoff. Segments of flu­ the upper lithic unit. It forms a broad. patches are more abundant than integrated vial channels elsewhere extend for a limited gently sloping scab~y surface between rills. complexes. A flat-floored swale, enclosed distance downslope from impervious areas, gullies, and cracks and constitutes fully 90% by swells of pahoehoe, with a discontinuous chiefly bedrock. Most channel floors are of the total area. This upland displays a gen­ border of upper lithic tephra around a cen­ mantled by coarse fluvial sands and fine tle, irregular, swell and swale configuration tral patch of residual gravel, is a common gravels. Broad sand bars form where chan­ of about 0.5 m relief, presumably reflecting arrangement. The gravel forms an armor of nels widen, and sand flats of many square the morphology of underlying pahoehoe single-stone thickness overlying as much as hectometres in extent occur where channels lava flows. A succession of crude wave-like 50 em of fine primary tephra. debouch into areas of low relief. forms, with 15 to 20 m separation and 10 to Fragments in gravel patches range from 50 em amplitude, also ruffles the surface for scattered, with a remarkably uniform sepa­ Eolian Deposits Terrain a few hundred metres outward from thc ration of 2 to 4 em, to densely packed, with caldera rim (Swanson and Christiansen. all fragments in contact or nearly so. An Deposits of wind-blown sand increase 1973). In the walls of cracks and gullies. impressive uniformity of particle size in a outward within Kau Desert and become a these waves are seen to consist ofconforma­ patch may reflect a predominating size in significant micro-terrain element near the ble accumulations of upper lithic materiaL the parent material and sorting by the outer edge of the once tephra-blanketed They are probably bed forms developed by transport mechanisms-impact creep, rain­ area. Minor leeside accumulations occur base surges. beat, and sheet wash. throughout the desert, and larger deposits Shallow swales and hollows scattered have formed in the lee of major topographic over the upland are floored by accumula­ Colluvial Terrain features-Cone Peak, for example-and tions of lag gravel, approaching desert downwind from wide flats of fluvial sand. pavement in sorting and close packing 01 Colluvium forms on the walls of gullies Lithic, crystal, and vitric particles from the fragments. These accumulations are onc­ and cracks cutting through the tephra. It intermediate and lower Keanakakoi units, stone thick, with an infilling of youngcl consists principally of various grades of and 's hair from younger eruptions, are Pele's hair and other small vitric fragments sand from the intermediate and lower units. the principal constituents of the eolian de­ sifted downward between gravel fragments Locally, cliffs in the upper lithic unit shed posits, which are dark colored, well sorted, Floors of fluvial washes are extensively coarse lithic fragments into the colluvium, and locally cross-bedded. The largest ac­ mantled by coarse sand and fine gravel and cemented blocks of upper lithic beds cumulations are dune ridges several metres reworked from tephra. Colluvial material. I, slump onto the colluvial slopes. high, usually vegetated, and modified by consisting of loose sand, lithic fragments. secondary blowout activity. and slumped slabs of well-cemented UppCI Fluvial Material Terrain lithic layers, mantles the steep walls of gul­ ZONES OF DEGRADATION lies and cracks below a topping cliff mad" Fluvial channels are most abundant near by the upper lithic unit. Bedrock constitutl', the caldera, where rainfall is greatest, the Progressive deterioration of the Keana­ only I% or 2% of the total area in zone I. deposits are sufficiently thick to permit kakoi tephra blanket southwestward from chiefly as narrow strips along cracks and channels, and the surface is most imper- Kilauea caldera is treated by definition of gully floors. At the caldera rim, a few knohs sa $ "

STRIPPING OF TEPHRA, KILAUEA VOLCANO, HAWAII 1153 rious :rraID lally, hin a W~(~~:':1 g- Fluvial sand and gravel -5% Peak, IE i f - Colluvium -7% ';J..... t,... uous, <-( "- e- Pumice - Trace ling a ilwest ~gl d - Lag gravel -8% • this tbout c _ Scabby upland of upper lithic ~ unit -65% , dis­ ~~~~~ • + + + + + b_ Bedding - Planes of intermediate s 4 m [2]++++++ )f the and lower tephra units -3% erally .... , - ...... 5~~-f; 0- Bedrock (Pahoehoe) -12% ished , area ravel, ~, anel o 5 10m lfe of I I I ,ed on ,road, I rills, y90% a gen­ Figure 6. Sketch distribution map and :ation section of micro-terrain elements within zone ~cting II, near Cone Peak, Kau Desert. Fraction of oehoe surface covered by each element given as 'e-like percentage. 10 to ce for n the Insen, of bedrock project through the western edge gravel cover areas as large as several square areas of surficial lag gravel. With the possi­ ullies. of the tephra blanket. hectometres, and similar materials line ble exception of ill-defined zone V, this is orma­ major fluvial channels. Lag gravel partly the widest zone, at about 2.5 km. terial. I,one II covers the floors of most swales and hollows Micro-terrain elements within zone III led by on both the upper lithic unit and bedrock. are essentially those of zone II, but with dif­ Zone II is the narrowest zone (0.5 km) Bedrock, as irregular knobs or linear expo­ ferent proportions. Zone III consists of 50% ttereel hut the most varied, containing the greatest sures along cracks and channels, makes up or more exposed bedrock, approximately mula­ number of terrain elements (Fig. 6). It is as little as 10% to as much as 50% of the 25% residual gravel, 20% tephra, and 5% desert characterized by islands of bedrock, princi­ surface. Bedding-plane outcrops of fine lay­ fluvial sand and gravel. Residual patches of ing 01 pally pahoehoe swells and tumuli, project­ ers in lower Keanakakoi units and local the upper lithic unit resting directly on bed­ : one­ IIlg I to 3 m above a sea of tephra or accumulations of wind-blown debris are the rock are abundant, and small exposures of lunge I secondary debris. The dominant terrain remaining terrain elements. oxidized layers of lower Keanakakoi units ments clement (40% to 50%) is the scabby upland, are also seen. Fluvial deposits are extensive nents like that of zone I. Along the east margin of Zone III where streams from zone II debouch. Isively the tephra lobe in this zone, a loose lithic Downwind from such areas are small lee­ gravel j[ravel mantles as much as 65% of the sur­ The tephra-bedrock pattern of zone III side accumulations of eolian origin. tterial. filce. This gravel appears to be upper lithic reverses that of zone II. Here, residual nents. material, which escaped cementation, from patches of tephra are scattered across an Zone IV upper which the fine matrix has been driven by integrated bedrock surface. One can walk a )f gul- wind, rain, and percolating water. Loose continuous, albeit irregular, course wholly Zone IV is the simplest zone, consisting maele !travel extends outward nearly to Ahua on bedrock in zone III; in zone II, the con­ of 80% to 90% stripped lava that harbors titutes Kamokukolau, 2 km southeast of the cal­ tinuous path is on tephra. Typical pahoehoe scattered remnants of tephra and of lag .one I. dera rim, and makes up about 20% of the swell and swale topography dominates. gravel in swales and pockets. Small, scab­ ~s and total area of zone II (Fig. 7). Many swales are partly filled with tephra, like patches of the upper lithic unit (scabby knohs I.ocal bars and flats of fluvial sand and which outcrops peripherally around central upland) remain on some lava swells, and a • $ • Mi. it • e 'lUi l a ;sa 222 Q £J bi x dU la

1154 MALIN AND OTHERS thi ne sal de ~g- Fluvial sand and mt gravel 2% gn hri IOoBI d - Lag gravel 65% up eXI Bedding-planes of res intermediate and lower be, tephra units 20% grc del 0- Bedrock (Pahoehoe) 13% A( OF

I Ke: o 5 10 15m Wir I 1 I I 1al Figure 7. Sketch distribution map ily and section of micro-terrain elements within zone II, near Ahua Kamoku­ \Va kolau. Fraction of surface covered by each element given as percentage. len An KiL tephra layers (bedding-plane terrain) form and discontinuous marginal bands around flats par in secondarily filled swales (Fig. 8). Fluvial tior sands line stream channels and locally form gen more extensive bars and flats. Much of the t-:va stripped pahoehoe in this zone displays a nor faint yellowish-brown, oxide coating, in­ wat ferred to have been imposed by tephra bur­ Kil, ial, as swales from which a tephra filling has aha recently been partly removed show a "high­ Ilan water" line of darker oxidation extending ~ to the upper level of the former filling. 1111S1 Exposures of yellowish-brown, oxidized, do\\ compact, sand-sized coarse ash beds of <1111' lower tephra units are larger and more lIort abundant here than in zone III. Small \\'CSl deposits of eolian sand occur mostly as I hi~ leeside accumulations. .dly

Zone V "I'W Zone V is primarily an area of reaccumu­ Hail lation, displaying a pattern resembling that III hi of zone III with islands of bedrock rising illto above a sea of predominantly reworked lllbi materials. The primary tephra blanket (that lit hi, is, that emplaced directly by volcanic proc­ n esses) was probably never thick here, and (tillS much of the material filling swales and pror \11('( Figure 8. Eroded swale filling of tephra, zone IV. about 6 km southwest of Kilauea creating dunes 3 to 4 m high is wind-blown caldera. Zone IV. Resistant tephra layers ring accumulation of lag gravel. sand (that is, of secondary origin). Much of I STRIPPING OF TEPHRA, KILAUEA VOLCANO, HAWAII 1155

this sand was emplaced penecontempora­ that characterize much of the upper Kau tephra from closed depressions and, in the neously with the tephra. Deposits of fluvial Desert. absence of lateral erosion by streams, may sand, approaching I m in thickness, are Ephemeral streams clearly erode the be a major means of removing such material clearly post-tephra. Bedrock constitutes as tephra. The nearly continuous blanket of from wide, gently sloping interfluvial sur­ much as 30% to 50% of the zone, and lag zone I is dissected by a system of dendritic, faces. After the upper lithic layer has been gravels, fluvial sands, and wind-blown de­ ephemeral, rill, and stream channels. As this breached, much of the underlying material nris are abundant. A thin layer of granular is the area of heaviest rainfall, and 80% to is susceptible to eolian entrainment. upper lithic tephra caps many of the tephra 90% of the surface is crusted, the runoff is Local granule ripples, composed of lithic exposures and also composes small scabs great. Scattered larger channels, as much as fragments as much as 4 mm in diameter, resting directly on the bedrock. Vegetation 30 m wide, which have ~ut to bedrock, are demonstrate the ability of wind to move necomes more prevalent owing to favorable still, for the most part, too young to have coarse debris by saltation impact (Bagnold, growing conditions provided by finer effected significant bedrock erosion, and 1941, p. 180), and lithic fragments in debris. they are not likely to, owing to the permea­ patches of residual gravel may have been bility of the bedrock lavas. Areas formerly moved down adjacent slopes by this proc­ AGENTS AND PROCESSES blanketed by tephra probably also had an ess. Stable surface stones show modest OF DEGRADATION integrated system of dendritic gullies and effects of sandblasting (polish, pitting, flut­ rills, and at that time fluvial erosion was ing) on their northeast sides. Principal processes involved in stripping more effective in removing tephra in those Wind erosion presumably is less signifi­ Keanakakoi tephra are wind-driven rain, areas than at present. As the tephra blanket cant now than formerly, because material wind-driven sand, soil-water sapping, fluv­ disintegrated, these gullies disappeared. susceptible to eolian entrainment is largely lit! runoff, and particle creep (not necessar­ Soil-water sapping appears to be an effec­ gone, and much of the surface is protected ily in that order of effectiveness). tive erosive process in Kau Desert. Except by crusts and lag gravels. Major currently for crusted surfaces, infiltration capacity of renewable sources of wind-blown sand are Water tephra materials is high, and even the crusts the floors of fluvial washes and bars and are locally cracked, permitting some pas­ flats subject to flooding. These areas be­ Kau is a desert more in the sense of bar­ sage of water into underlying deposits. come armored within a few months when renness than in that of lack of moisture. These deposits contain pervious beds that subject to the current wind regime, but each Annual rainfall averages 130 cm near conduct water laterally, particularly where episode of flooding creates a new supply of Kilauea summit but diminishes to 30 cm interlayered with less pervious fines. Upon sand. form and less southwestward. Precipitation is not emerging on a sloping face, this water flats particularly effective in supporting vegeta­ causes undermining and creation of es­ Mass Movements uvial lion or in causing fluvial erosion, owing to carpments of collapse. Steep walls of gullies form venerally high permeability ofthe substrate. and cracks display steep-head amphithea­ The most obvious products of mass )f the Ivaporation is also high, because prevailing ters, hanging tributaries, and waterfalls, movement within tephra-mantled areas are ays a northeast trade winds, having lost much forms characteristically produced by seep­ slump blocks and tilted slabs of cemented ~, 10­ \\ater on the windward (northeast) side of age sapping. Sapping should be particularly upper lithic layers on the walls of gullies and I bur­ Kilauea, absorb moisture as they warm kat­ effective along the tephra-lava contact, cracks. The 5- to 15-mm lithic fragments, Ig has .Ihatically upon descending the southwest where the relative difference in permeability composing lag-gravel patches, may move 'high­ llank of the volcano. concentrates percolating water. downslope by particle creep, but this proc­ nding Much Kau Desert precipitation occurs as ess is probably secondary to rain-beat and illing. Inlst and gentle drizzle, but torrential Wind saltation impact. dized, dtlwnpours do take place. Most heavy pre­ ds oj 'Ipitation is driven by strong winds from the Strong, northeasterly trade winds, occa­ STRIPPING OF AA LAVA more ntlrtheast (trades) or the south and south­ sional southwesterly storm winds, and lack Small IITSt (storm-generated, "Kona" winds). of sheltering by marked topographic relief Aa lava is less abundant than pahoehoe tly as I II is makes rain-beat (Ellison, 1945) unusu­ or dense vegetation favor eolian erosion in in Kau Desert, but some opportunity exists ,till' powerful and effective in this barren Kau Desert. Dust storms rising from the for comparison of stripping from these two legion. Testimony is provided by ubiqui­ desert following the 1924 phreatic eruption strikingly different surfaces. The roughness IIIIIS earth fingers, 2 to 4 cm long, pointing of Halemaumau demonstrated the efficacy and inhomogeneity of an aa surface ob­ I1pwind in areas of soft, coherent sediment. of eolian processes in removing fine mate­ viously make stripping more difficult. cumll f(;lIn-beat may playa role in concentrating rial (Stearns and Clark, 1930, p. 59), and The penetrability of fine tephra into aa is 19 thai "lllie fragments, 0.5 to I cm in diameter, sand dunes in southwest Kau Desert indi­ highly variable. Aa surfaces composed prin­ rislll}' ,!II tl lag gravel patches, and it probably con­ cate effective eolian traction and saltation. cipally of pebble- to small cobble-sized vorked 'flhutes to etching of the crusted upper Impact by wind-blown, saltating sand fragments can be sufficiently impenetrable :t (thai 'lillie surface. grains has probably played a significant role for a continuous mantle of fines to collect ~ prol' Ihe relatively impervious, almost ubiqui­ in etching the surface of cemented crusts on and remain on its surface, but coarse, 'e, and '

Figure 9. "Gunite" tephra forcefully emplaced onto rough aa lava, 3 km west of Kilauea caldera. Thickness and distribution of coating on knobs suggest forceful emplacement from the east-northeast, and adhesion suggests moist material.

though adjacent to aa with tephra rem­ nants. Only by standing on a high prom­ inence and looking down upon the coarse aa can one see that fine tephra chokes sub­ surface openings between blocks. Blanketing and subsequent partial remo­ val of tephra to a zone II condition on aa o 1m ! I results in numerous, randomly distributed, small lava knobs projecting above an inte­ blocks. Penetration may occur directly dur­ cial gunite, and their thicknesses and distri­ grated surface of tephra and reworked de­ ing initial emplacement of fine, dry tephra, bution on individual knobs suggest em­ posits. One can easily walk between thesc or it may be caused secondarily by rain-beat placement from an easterly or northeasterly bedrock islands, remaining always on teph­ and rain-water percolation. Although pene­ direction, possibly by trade winds or base ra materials. In aa of zone II, scabs 01 tration probably should not be classed as a surges from Kilauea. Similar deposits were crusted tephra adhering to favorable sites form of stripping, its effect on surface reported in association with the Taal erup­ on rough lava knobs are modestly abun­ appearance can be similar. tions of 1965 (Moore, 1967). Sand-sized dant, and crusted tephra also forms flanges Heavy rains (Stearns, 1925, p. 198, 200) tephra and accretionary lapilli, penetrating around the bases of the knobs (Fig. 10). Thc during the 1924 phreatic eruption suggest deeply into niches and cavernous openings filling in swales usually consists of a basal that fine tephra may, on occasion, be with overhanging lips and tortuous configu­ section of primary tephra overlain by emplaced in a moist state. If it were wet rations, suggest that forcible emplacement, reworked tephra debris, commonly lag enough to form "mud," significant initial possibly by a base surge, was involved. gravels underlain by finer sand and silt penetration might not occur, even on The surface rubble of aa flows is ob­ derived largely from the matrix of the uppel blocky surfaces. However, if water is in viously highly pervious to water, with the lithic unit. excess, inwashing and penetration would be result that runoff from unmantled aa sur­ A zone III aa development featurcs expected. Coherent coatings of fine tephra, faces is essentially nil. This condition con­ separate pools and small irregular insets 01 many centimetres thick, adhering to near­ tinues until subsurface interstices become tephra within an integrated lava surface vertical faces of bedrock knobs and cliffs choked with fine debris. Channelized fluvial (Fig. I I). Infilled swales contain primary within aa flows near the northwest margin erosion is thus both minimized and delayed. tephra topped by a thick secondary layci of Kau Desert, indicate that fine tephra was The inhomogeneity of aa flows as to consisting of fine, massive, structureless deposited there in a coherent, presumably penetrability, configuration, roughness, and debris intimately penetrated by vegetatioll moist, state, allowing it to cling to steep fragment size results in great differences in rootlets. Lag gravels include some spalled faces and to resist subsequent removal stripping. To casual view, many areas of fragments and chips from surroundin): (Fig. 9). These coatings resemble commer- coarse, blocky aa appear unmantled, al- lavas, supplementing the usual upper lithic unit particles. Scabs of cemented tephr

gh aa Filled swale Tephra Scabs IItion n the

rem- )rom- 0 5 10m :oarse I I I ; sub- Figure 11. Sketch offeatures within zone ilIon aa, showing separated insets of tephra within an otherwise continuous lava surface. remo­ on aa underlying rock is significant. Voids within exposed or eroded first by other, principally widths. Zone I, characterized by the thickest )uted, aa can become fully choked with fine mate­ fluvial, processes. Only water is capable of deposits and most stable surfaces owing to I inte­ rial to depths approaching I m. moving the coarser lithic fraction of the crusts, would appear to be the form origi­ :d de- Wind is the only agent seemingly capable tephra; the absence of throughgoing drain­ nally taken by the deposit, and probably the these of exporting tephra from aa. Its effective­ age, owing to infiltration, modest precipita­ longest lasting, least affected. Zone II, teph­ ness is hampered through armoring by tion, and high evaporation, restricts the showing the onset of stripping as noted by Lbs of coarse fragments, lack of saltating sand transportation of coarse material and al­ the increase in bedrock outcrop, may reflect e sites grains to dislodge fine particles, and, ulti­ lows the creation of residual concentrations not only a thinner initial deposit, but also a abun­ mately, vegetation. It may be possible for of pebbles and cobbles, as well as lag grav­ deposit at a later time in its evolution, when langes wind to remove some fine sand and dust els. These "armors" can defeat further strip­ erosional processes have breached the crust I). The owing to wind turbulence generated by ping. Intimately related to fluvial and eolian and removed some but not all of the tephra. , basal increased surface roughness on aa flows. In action are the crusts formed on exposed sur­ Zone II is much narrower than zone III, m by any case, sand does not travel far over aa by faces of the Keanakakoi debris. The crust despite their morphologic kinship, and this Iy lag saltation; it is too easily trapped in voids on the upper lithic unit firmly holds many may reflect a more rapid evolution in time ld silt and interstices. However, notable accumu­ coarse fragments and protects all underly­ from zone II to zone III than from zone I to :upper lations of eolian sand on aa occur near ing materials from erosion. It is this crust zone II (that is, little evolution occurs until places where wind has access to such a that now dictates the rate and location of crusts are breached; then it proceeds rapidly :atures copious source of material, such as fluvial erosion, by focusing water into a few chan­ at first and then slows). Zone IV may lsets of channels or flats, that sand floods the sur­ nels (promoting fluvial erosion) and, by its represent the final state, and one also long­ mrface face. The greater difficulty of stripping absence, permitting eolian action. lived. Thus, the sequence of evolution may rimary tephra from aa, compared to pahoehoe, Two aspects ofthe morphology and plan­ be envisioned not only as a monotonic y layer produces about a one-zone difference on imetric configuration of the partly eroded decrease of thickness of deposit with time, .ureless adjoining areas of these lavas, with zone II Keanakakoi tephra blanket are worth not­ but also as a "wave" propagating from the etation on aa lying alongside zone III on pahoehoe. ing: the characterization of states of degra­ distal to the proximal deposits. With the spalied dation by mixtures of micromorphological crusts acting to prevent significant lowering .unding DISCUSSION criteria, and the potential time implications of the present surfaces, it may be the case :r lithic of these spatial relationships. Of the seven that lateral erosion now is more important tephra Stripping of the Keanakakoi tephra has different micromorphological surface forms than deposit thinning. Prior to the estab­ bs arc heen controlled by several factors: thickness used to characterize five zones of significant lishment of crusts and residual concentra­ in aa of the deposit, size of the component mate­ difference, the occurrences of scabby up­ tions of stones, deposit thinning may have rtals, and relative effectiveness of erosional land, lag gravels, bedrock (pahoehoe or aa), been more important. The existence ofdune .hoehoe processes. Areas farther from the vent and bedding planes of lower units of the sands in zone V, interbedded with some st only within Kilauea caldera received less mate­ Keanakakoi tephra are most diagnostic. primary tephra, argues strongly for con­ La flow. rial to begin with, and these thinner deposits Bedrock is the principal element; the three temporaneous erosion and deposition. The a vIew have been nearly totally removed. The size proximal zones can be said to consist of nature of the erosional processes, the time illing 01 of the component materials and the relative "no" bedrock outcrop, "some" outcrop, and scales over which they operated, the distri­ effectiveness of stripping processes are in "largely" bedrock. These zones are most bution of crusts, and the rate at which they le same ~ome ways interrelated. In areas of fine easily further divided by relationships of formed are thus critical to understanding :gree 01 material, significant eolian deflation has mantling materials (primary tephra or sec­ the early, postemplacement history of Kea­ noff on occurred; the absence of the intermediate ondary, reworked debris) to bedrock; the nakakoi tephra. ld local IIlld lower units of the Keanakakoi in distal variation from largely mantle surrounding There are many ways in which the tradi­ be first portions of the tephra blanket may reflect "islands" of bedrock to largely bedrock with tional erosional agents, water and wind, act gravity­ early removal by eolian processes. Nearer insets of mantle is the clearest change. on the deposits in the Kau Desert. Water is tents on the caldera, where the upper lithic unit is It is tempting, although perhaps not probably the most important agent, where it condary hoth coarser and thicker, finer debris is pro­ appropriate, to associate time relationships can operate. It is restricted, however, by the nmgs til tected from wind action except where with the spatial relationships of zone peculiarities of Kilauea hydrology. Inte- aUUM e

1158 MALIN AND OTHERS grated drainages are found wholly and only roughness of the aa flow. Without stream ACKNOWLEDGMENTS within the debris materials-no drainage drainage, and with infiltration and wind the superposition onto underlying basaltic only effective transporting mechanisms, We have benefited greatly from generous flows occurs. Although alluvial deposits material trapped on aa flows has a much discussions with R. L. Christiansen con­ have formed at the mouths of drainages, greater residence time. It is likely that an aa cerning his largely unpublished data on these do not provide a substrate suitable for flow retains most of a mantle, after it is stratigraphic characteristics of the Keana­ further drainage development. Stream load covered. However, owing to its initially kokoi Formation. Personnel of the Hawaii is limited to the few ephemeral streams and rough surface and steep margins, only a Volcano Observatory (USGS/HVO), too is derived principally by bank scour, sap­ thick deposit will create a significantly numerous to specify, were most helpful. ping, and mass movements. In-washing by altered view of the flow from an overhead This research was supported by funds rainfall rather than stream flow is most perspective. The difference between aa and from the National Aeronautics and Space important in modifying tephra deposits on pahoehoe flows in retention of mantles, and Administration's Planetary Geology Pro­ aa flows, although this is not necessarily a the ability to discriminate these mantles, gram Office granted to the USGS/HVO "stripping" process. appear for the most part linked to the (Grant W 13.709), to the Jet Propulsion Wind is more effective over a wider area planimetric scale of depressions upon their Laboratory (where it represents one phase than is water and, although limited by the surfaces. of study carried out under NASA contract size of materials it can move, appears to Finally, there are the general implications NAS7-100), and to Arizona State Univer­ have played a significant role during at least of these observations, and their importance sity (Grant NAGW-I). three periods in the evolution of the deposit. to places such as Mars, where observations Evidence in zones IV and V clearly points to are necessarily limited to a distant view. It is REFERENCES CITED eolian processes active during eruptions and clear that mantles of debris, like the Keana­ Bagnold. R. A., 1941, The physics of blown sand and desert dunes during periods between eruptions. Today, kakoi tephra, can mask original morphol­ london, Methuen and Company, 265 p. Binder, A. B.. Arvidson. R. E.. Guinness. E. A., Jones K. L.. Morris. wind removes fine material from fluvial ogy. The degree of masking, the relief of the E. C, Mutch. T. A.• Pieri. D. C, and Sagan, c., 1977, Th( geology of the Viking Lander I site: Journal of Geophysical deposits and continues to work surface morphology, and the deposit thickness are Research. v. 82. p. 4439-4451. Christensen. P. R.. 1982. Martian dust mantling and surface composi­ exposures of other loose materials. intimately related. Exhumed surfaces often tion: Interpretation of thermophysical properties: Journal 01 The effects of burial on the bedrock itself show only faint suggestions of prior burial, Geophysical Research, v. 87, p. 9985-9998. Christiansen. R. l., 1979, Explosive eruption of Kilauea Volcano ill are also worth noting. A thick cover on principally in the form of lag gravels and 1790 [abs.]: Hawaii Symposium on Intraplate Volcanism and Submarine Volcanism, Hilo, HI. July 16-22, 1979, p. 158. pahoehoe flows easily masks their already other materials too deeply intercalated Cruikshank. D. P.• 1974, Mauna Iki and the K.au Desert. Kilauea Vol­ subdued morphology. A thinner cover leads within surface depressions to be removed. cano, Hawaii. in Greeley, R.• Geologic guide to the island oj Hawaii: National Aeronautics and Space Agency, U.S. Govern to "islands" of bedrock (typically tumuli) in The stripping (or infiltration) of debris is ment Printing Office, p. 218-232. Ellison, W. D., 1945, Some effects of raindrops and surface-flow on soil a sea of flat surfaces, and a "totally" dependent upon the processes active, the erosion and infiltration: American Geophysical Union Trans3' stripped surface still retains as much as 20% caliber of material that these processes can tions. v. 26, p. 415-429. Greeley. R.• Leach. R.• Williams. S. H., White, B. R.• Pollack. J. R.. by area of primary and reworked debris, for move, and the size of the materials actually Krinsley, D. K., and Marshall. J. P.• 1982. Rate of wind abr'l­ sion on Mars: Journal of Geophysical Research, v. K7 the most part filling depressions within the available. In Hawaii, several processes have p. 10009-10024. pahoehoe flow surface. Bedrock so stripped Kelley. M. l., Spiker. E. c.. Lipman, P. W.• Lockwood, J. P (: worked in concert to create the surfaces Holcomb. R. T., and Rubin, M., 1979. U.S. Geological Survc~ displays remarkably little physical evidence seen there. In the absence of significant flu­ Reston. Virginia. Radiocarbon Dates X V: anll Kilauea Volcanoes, Hawaii. p. 317. .1 of having once been buried. The principal vial action on Mars except, perhaps, early in Macdonald. G. A.. 1949. Petrography of the island of Hawaii: U.S change, as viewed from above, is the its history, critical factors concerning the Geological Survey Professional Paper 214-0. p. 51-96. Ii Macdonald. G. A., and Abbott, A. T.• 1970. Volcanoes in the S(';I "mottled" appearance resulting from the in­ stripping of debris blankets are the caliber Monolueu, Hawaii. University of Hawaii Press. p. 441. Malin, M. C, 1976. Nature and origin of intercrater plains on M'lI' sets of swale-filling mantle materials, often of the debris; the effectiveness of other [Ph.D. dissert.]; Pasadena, California, California Institute ,,[ d Technology, 180 p. surfaced by lag gravels with albedos that mechanisms, principally wind, in trans­ McCauley. J. F.• 1973, Mariner 9 evidence for" wind erosion in th, differ from the pahoehoe flow. Morphol­ porting materials; and the detailed form equatorial and mid-latitude regions of Mars: Journal of (lCll physical Research, v. 78. p. 4123-4137. I) ogy of pahoehoe flows is well preserved. of morphological variations. Nonvisible Moore. J. G.. 1967, Base surge in recent volcanic eruptions: Bulklill Volcanologique. v. 30. p. 337-363. II' Aa flows show a different response to remote sensing, particularly Viking Orbiter Mutch. T. A.• Arvidson. R. A., Binder. A. 8., Guinness, E. A.. <.Jlill burial. Although it is possible to mask an aa Infrared Thermal Mapper observations, Morris, E. C. 1977. The geology of the Viking Lander 2 Sill .11 Journal of Geophysical Research. v. 82, p. 4452-4467. flow by burial, the thickness must be great may address caliber (Christensen, 1982). Naughton. J. J., Greenberg. V. A.• and Gognel, R., 1976, Incrustation', I, and fumarolic condensateS at Kilauea: Journal of Volcanol0i'" II because of the scale of morphology to be The effectiveness of wind transport can be and Geothermal Research, v. I, p. 141-165. Powers. H. A., 1948, A chronology of the explosive eruptionli III covered. The topography on the surface of treated by laboratory simulation, theoreti­ Kilauea: Pacific Science, v. 2. p. 278-292. an aa flow can be very complex, with ridges cal modeling, and field study (Greeley and Sharp. R. P.• and Malin. M. c.. 1984. Geology from the Viking Landn, on Mars: Geological Society of America Bulletin (in press). and troughs several metres in relief, or it can others, 1982). Surface micromorphology Soderblom. L. A.• Kreidler. T. J., and Masursky. H.• 1973. Latitudin.LI distribution of a debris mantle on the Martian surface: Journal ,01 be relatively subdued, consisting of closely can be examined in Viking Lander images Geophysical Research. v. 78. p. 4117-4122. packed, approximately spherical clinkers. (Mutch and others, 1977; Sharp and Malin, Stearns. H. T.• 1925. The explosive phase of Kilauea Volcano, Haw,lIl III in 1924: Bulletin Volcanologique, v. 2. p. 193-209. However, to mask an aa flow from view 1984) and compared with terrestrial ana­ Stearns. H. T., and Clark. W.O.• 1930. Geology and water resourct:~ <'I '" requires that its flow margins be covered, logues such as those presented here. The the Kau District, Hawaii: U.S. Geological Survey Water·Supph Illl Paper 616, p. 29-191. . " and this necessitates a thick deposit. For synthesis of these diverse forms of study Swanson, D. A., and Christiansen. R. L.. 1973. Tragic base surge lit 1790 at Kilauea Volcano: Geology, v. I. p. 83-86. thinner deposits, spines, boulders, and with the high-resolution Viking Orbiter Wentworth. C. K.• 1926. Pyroclastic geology of : Honolllll' Hawaii. Bernice P. Bishop Museum. Bulletin 30, 121 p. ridges of aa clinkers may emerge upon images represents the most advanced stage --- 1938, Ash formations of the isle of Hawaii: Hawaiian Vokall" stripping. When they do, they may appear of analysis currently possible. The ultimate Observatory Third Special Report, p. 183. more mantled than adjacent pahoehoe test of the nature and evolution of martian flows, because of the difficulty in removing debris mantles likely will require in situ field MANllSCRIPT RECEIVED BV THE SOCIETV AUGUST 6. 1982 REVISED MANllSCRIPT RECEIVED SEPTEMBER 24. 1982 j .(', material trapped within the greater surface observations. MANlJSCRIPT ACCEPTED OCTOBER 13, 1982 Printed in U.S. /\ I