SEDIMENT TRANSPORT AND ACCUMULATION IN A FJORD BASIN, GLACIER BAY, ALASKA1,2 CHARLESM. HOSKIN3 AND DAVID C. BURRELL Department of Geology and Institute of Marine Science; and Institute of Marine Science, University of Alaska, College, Alaska 99701 ABSTRACT Sediment transport and accumulation were studied over several years to define the sedimentary environment in a fan-terminated, active valley glacier and adjacent fjord at Queen Inlet, Glacier Bay National Monument, Alaska. Recently exposed ice-contact sediment at the glacier snout is entrained during summer and early fall by meltwater streams crossing a subaerial-intertidal outwash fan. Suspended sediment in the meltwater (sediment loads exceed 1 g/liter at the inlet head) subsequently yields a cold, surface sediment plume overlying warmer saline water in the fjord basin. Mixing of these layers produces flocculation and settling occurs in discrete layers; each layer is considered to represent one tidal cycle. No gravel is present in fjord floor sediment, although both colluvium and outwash fan sediments contain gravel. This gravel distribution eliminates sliding, slumping of the outwash fan, and ice rafting as major contributors to fjord basin sediment in Queen Inlet. Piston cores of muddy fjord basin sediment show cyclical, thin, black horizons, thought to be annually produced. Spacing of these marker horizons indicates that sediment accumulation may exceed 1 m/year at the inlet head. The fjord floor is cut by sinuous valleys with natural levees, terraces, and sandy silt axial sediment. These valleys are believed to be the result of sediment-transporting bottom currents. INTRODUCTION The modern valley glacier-fjord regimen is a microcosm of the detrital phase of the sedimentary cycle. Source, transport, and accumulation are contained in a natural package of logistically manageable dimensions in a virtually closed system. This paper examines the processes and products of accumulationin the fjord basin of Queen Inlet, Glacier Bay National Monument, in southeastern Alaska (fig. 1). We have monitored Queen Inlet for over 5 years, commencing in the summer of 1966. The regimen of Queen Inlet is particularly well suited for sediment transport studies because (a) there is only one volumetrically important source of sedimentary particles (ice- contact deposits at the glacier snout) with the tremendousadvantageof having no well1 Manuscript received April 27, 1971; revised March 28, 1972. 2 Contribution no. 148 from the Institute of Marine Science, University of Alaska. 3 Present address: Box 5345, Southern Station, University of Southern Mississippi, Hattiesburg, Mississippi 39401. [JOURNAL OF GEOLOGY, 1972, Vol. 80, p. 539-551] @ 1972. The University of Chicago. All rights reserved. developed grain-size modes (Slatt 1971), (b) transport (meltwater streams) resembles an on-or-off mechanism with maximum flux in summer and zero flux in winter, which supplies two grain-size populations (suspended load and bedload), and (c) separate bodies of sediment, recognizable by grain-size composition, accumulate in the fjord basin. For detailed results of grain-size analyses for 32 Shipek grabs and three piston cores, see Hoskin and Burrell (1968, table 1, p. 102-104). PHYSIOGRAPHY GENERAL Queen Inlet (fig. 2) is a typical fjord with high, steep walls, a valley glacier, and an outwash fan and meltwater streams which connect CarrollGlacier-via an intertidal zone-to the marine portion of the fjord. The present long-term climatic conditions are causing most glaciers to retreat in Glacier Bay (Goldthwait 1966), yet Carroll Glacier, now the principal glacier of Queen Inlet, advanced by as much as 1 mile during the period 1964-1968 (Field 1969), and this advance has since continued. Less than 1 mile now (1970) separates 539 This content downloaded from 192.148.225.018 on October 27, 2017 06:57:38 AM All use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c). 540 CHARLESM. HOSKIN AND DAVID C. BURRELL 1360 136° 1320 FIG. 1.-Location of Queen Inlet within Glacier Bay National Monument, Southeast Alaska the glacier snout and tidewater. The physiographic map of Queen Inlet (fig. 2) was made from bathymetric (PDR) profiles taken in the summer of 1967. Bedrock (mostly foliated granitic rocks and carbonates; MacKevett et al. 1971, pi. 1) forms steeply sloping subaerial and submarine walls which continue at the inlet floor beneath a sediment fill of unknown thickness. Both Triangle and Composite Islands are outcrops of bedrock, and some rough bottom topography at the inlet head may be submarine outcrops of the same material. Lateral moraines are subaerially exposed on the northwest inlet wall just above sea level. The sill which restricts the southwest entrance (figs. 2, 5) has not been studied in detail. The entrance southeast of Composite Island has no such restriction. From seismic reflection profiles, Von Huene (1966, p. 300) considered the baymouth sill in Nuka Bay (Kenai Peninsula, Alaska) to be glacial drift and sills within Nuka Bay to be basement rock (Von Huene 1966, p. 293; p. 294, fig. 2). Holtedahl (1967, p. 190) favors a bedrock origin for This content downloaded from 192.148.225.018 on October 27, 2017 06:57:38 AM All use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c). 541 SEDIMENT TRANSPORT AND ACCUMULATION IN A FJORD QUEEN INLET East West -0 20 40 S60 S80 -100 fMS 0 0.5 1.0 n.m. FIG. 3.-Precision depth recorderprofiles for traversesshownin fig. 2. Channeldepthsbelow the FIG.2.-Physiographic representationof Queen Inlet showingthe inlet-floorvalleys. The intertidal mean low water in fathoms: n.l. = natural levee, zone is stippled; numberedtraversesare the PDR t = terrace. profiles of fig. 3; * = locations of the analyzed channel sediment samples of fig. 11; X = major channels of surface streams is not yet clear. lateralstreamdeltas. Two streams have built gravelly deltas (labeled X, fig. 2) into the fjord basin, but the sills in Hardangerfjord of western Nor- careful echo sounding around each delta way. An extensive terminal moraine and has revealed no channels, and therefore no outwash fan have developed in front of connection is believed to exist between Carroll Glacier. these two surface streams and the inletfloor valleys. Meltwater streams crossing INLET-FLOOR VALLEYS the outwash fan in the vicinity of Triangle There are at least two inlet-long depres- Island could connect with the inlet-floor sions in the floor of Queen Inlet which are valleys, but lacking the necessary data from here termed "inlet-floor valleys." Two val- side-looking echo sounders, their relationleys are clearly evident in transverse sec- ship is not known (fig. 2, question marks). tions north of profile 6 (figs. 2, 3). Only Most transverse profiles show terraces one valley appears in the next profile south within the valleys and natural levees where (7) and it is believed that the two valleys the valleys join the adjacent flat fjord merge as shown in figure 2. The valleys are floor. Buffington (1952) has noted that subsomewhat sinuous in plan, and less than marine valleys with levees are perfectly 30 m wide at the inlet head, increasing to analogous to their subaerial counterparts. about 120 m near the inlet mouth. The valley walls are steep (see fig. 3), but actual Such valleys have not been previously reslopes may not be determined from the corded from fjords and are of considerable wide-beam echo soundings obtained to date. interest because they are believed to be The valleys are between 3.7 and 12.7 m evidence for sediment-transporting bottom deep. The relationship between the head currents. Echo-sounding studies in adjacent (north) of the inlet-floor valleys and the Glacier Bay fjords (predominantly with This content downloaded from 192.148.225.018 on October 27, 2017 06:57:38 AM All use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c). CHARLES M. HOSKIN AND DAVID C. BURRELL 542 tidal glaciers) have revealed no inlet-floor There is no freshwaterinflow from Carroll Glacier during the long subarctic winter. valleys. Thermohaline mixing of the fjord waters HYDROGRAPHY is induced by surface cooling and, during Hydrographic and suspended-sediment this period, the fjord is essentially unstratiprofiles have been obtained at the head of fied and homogeneous. Spring insolation Queen Inlet at various seasons over a 5- effects upon the upper water column is year period. Figure 4 illustrates the seasonshown by the April profile of figure 4. A al vertical temperature distribution of the strong, shallow thermohaline (figs. 4, 6) reference locality, station Ql-10 (fig. 5). is developed with the glacial meltwater inflow during the summer and early fall, and surface warming of the marine water helps stabilize this surface stratification. Runoff peaks in early September with the combinedcontributionsof maximumglacial melt and enhanced precipitation. Carroll Glacier terminates prior to the intertidal area (fig. 5), so that meltwater enters the fjord as a cold surface prism. More commonly (in southeast Alaska), the valley glaciers are in direct contact with the marine water; these are the "iceberg fjords" of Pickard (1967). For this latter class of fjord, fresh water enters the fjord in a more 3 5 4 2 6 7 diffused fashion and the resulting circulaT C tion patterns, at least adjacent to the glaFIG. 4.-Annual low-water temperature districier face, are poorly understood. bution at station QI-10 of fig. 5. Figure 6 plots salinity values concomitant with the temperatureprofiles of figure 4. These illustrate the classic fjord hydrogQUEEN INLET raphy, with primary summer mixing and ACONA 099 SEPT. 70 099 JULY 70 076 SEPT. 68 038 APRIL67 S Ol-10 LW 4 8 12 16 20 24 28 32 S %. FIG. 5.-Location of reference station Ql-10 and longitudinal profile of fig. 7. FIG. 6.-Annual low-water salinity distribution at station QI-10 of fig. 5. This content downloaded from 192.148.225.018 on October 27, 2017 06:57:38 AM All use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c). SEDIMENT TRANSPORT AND ACCUMULATION IN A FJORD circulation restricted to the upper layers (e.g., Cameronand Pritchard 1963, p. 313). Data obtained for both tidal velocities and freshwater inflow in Queen Inlet to date are somewhat crude. However, it would appear that the surface-layer water circulation is more strongly influenced by tidal mixing and marine entrainment during the early part of the summer season, becoming progressively river-dominated toward the fall; that is, a progressionfrom a standard type-B estuary toward type A (Pritchard 1955). In September the Queen Inlet halocline is restricted to the surface 10-15 m (fig. 6). At this time, saline entrainment, and hence the nontidal component of the balancing return flow of subsurface marine water up the inlet, would be minimal. McAlister et al. (1959) have given depths of 0-6 m and 6-35 m, respectively, for the surface outflow and inflow (from mouth to head) for the Silver Bay fjord (fig. 1, near Sitka) at an equivalent time when river runoff predominatedover tidal mixing. The July halocline of figure 6, when surface tidal mixing is of relatively enhanced importance, extends to a depth greater than 30 m, and the "base" of the upper fjord FIG. 7.-Temperature in July. 543 "circulation cell" lies deeper also. The net nontidal velocities are, however, well below the September values because of the reduced river flow. Surface wind mixing, due to intense foehn-type winds from the glacier, might well be sporadically important but has not been observed in this particular inlet. Figure 7 gives the July low water-early flood temperature structure for the longitudinal section shown in figure 5 (stations 99-7-99-2). The profile to the left of the vertical broken line is a "backwater"region adjacent to the sill. At the maximum flood stage, the fresh water is backed up north of stations 99-2 and 99-1 (QI-10). This latter station (location of the profiles given in figs. 4, 6, 9, and 10) therefore "sees" the surface prism each tidal cycle. Figure 7 shows warmer (> 5.7° C) and more saline (> 30.7%0) Glacier Bay water partially penetrating the density structure. At the head of the inlet (i.e., in the vicinity of station QI-10), the subhalocline water is essentially homogenousand "marine."With the characteristic fjord depth/width ratio, and the absence of a complete entrance sill, deep tidal water movements are very slight, distribution along the profile shown in fig. 5. Low-water-early flood tidal stage This content downloaded from 192.148.225.018 on October 27, 2017 06:57:38 AM All use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c). 544 CHARLES M. HOSKIN AND DAVID C. BURRELL and the principal nontidal circulation is confined to the surface waters as noted. However, there is a progressive slight dilution and warming of the deep waters at the inlet head during the summer (figs. 4, 6). This is a reflection of the summer-fall subsurface trends within the contiguous Glacier Bay (which is itself a complete fjord system); the summer hydrographic patterns developed within Glacier Bay have been discussed by Quinlan (1970). It should be noted that the subsurface water column of "iceberg"fjords is far less stable, as might be expected, than for Queen Inlet where fresh water is added only on the surface. Pickard (1967), for example, has recorded considerably higher dissolved oxygen contents at depth for the former type. This enhanced vertical diffusivity seems to occur regardlessof the fjord geometry. SUSPENDED-SEDIMENT SURFACE REGIME PLUME During late August-September, glacier meltwater covers large areas of the mudflats (fig. 8) in ill-defined streams. In early summer this water is confined to several major courses (principally between Triangle Island and the east shore), but these have changed over the years with the advance of Carroll Glacier. It is impossible to monitor quantitatively the freshwater inflow with any degree of accuracy, but the contribution from Carroll Glacier certainly far exceeds that from any lateral stream. This runoff transports considerable quantities of sediment from the glacier, moraine, and intertidal mud-flat areas into the fjord; thus, the surface prism described above is, in fact, a distinctive sediment-ladenplume, which may maintain its identity (at low water) for several miles down the inlet. Figure 8 illustrates a quasi-synoptic distribution of the plume at this tidal stage at the head of Queen Inlet in September.The suspended-sediment load (fig. 8, S.L.) is always in excess of 1 g/liter at the mudflat-marine boundary; the sample given in figure 8 is for just one of the many streams emptying into the fjord. The major glacial streams at this time were located on either side of Triangle Island, with the principal plume transport concentrated on the west side of the fjord, as is the case in highlatitude (N) estuaries. FLOCCULATION FIG. 8.-Quasi-synoptic distribution of the sediment plume at low water in September. See text for discussion of boundaries A and B. The temperature (T), salinity (Sal), and suspended-sediment load (S.L.) data are for the surface 5-10 cm. SITE Boundary A of figure 8 is the highly distinctive flocculation front of the plume. This has been shown by laboratory experiment to occur with mixing to approximately 4 %o salinity. We concur with Schubel (1971) that there is much ignorance concerning the mechanism of agglomeration of suspended sediments in estuaries, but the active fjord environment may well be one of the few examples where use of the term "flocculation" per se is correct. The exceedingly high sediment concentrations within the freshwater plume ensure that flocculationproceedsrapidly, once initiated, with increase in ionic strength of the water; hence the sharp leading edge (fig. 8, A) of the plume. There is (as noted also by Schubel) no established method for determining the degree of such sediment agglomeration. Our work (see below) utiliz- This content downloaded from 192.148.225.018 on October 27, 2017 06:57:38 AM All use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c). 545 SEDIMENT TRANSPORT AND ACCUMULATION IN A FJORD ing electronic (Coulter Counter) procedures appears to yield size spectra of the primary particles only. Settling rates of the flocs under shipboardlaboratoryconditions indicated that most were in excess of 100 4, that is, fine sand-silt size-range. Electron micrographs of undisturbed bottom-Sediment samples show typical, randomlyorientated floc textures (N. R. O'Brien, personal communication, 1970). Between boundaries A and B of figure 8, the settling sediment is close to the surface and still visible, so that load samples from this zone and the primary plume area are not dissimilar. Boundary B in reality marks an indistinct zone when visual contact with the sediment is finally lost. The samples of figure 8 were collected by small boat, and the boundary positions fixed by radar ranging from the research vessel. ment-load values obtained contemporaneously) were taken at station QI-10 at a time of maximum freshwater runoff. These data required approximately 30 min for collection, bracketingthe tidal stages shown. It is believed that each of these layers represents an individual tidal flocculation; that is, for any given locality at the head of the inlet, a flocculated sediment layer is produced every tidal sweep of the surface plume approximately on a 12-hour cycle. The suspended-sediment band spacing is much increased earlier in the summer (and later in the fall) compared with the highrunoffperiod data of figure 9. For example, figure 10 gives the load distribution for July 1970. Only one major subsurfaceband is apparent.If the formationof these layers is in phase with the tidal cycle, then net vertical settling rates from 1.0 to 5.0 m/ hour, respectively, for the September and SETTLING OF FLOCCULES July conditions are indicated (figs. 9, 10). Vertical profiles of suspended-sediment Such seasonal variations in the band spacdistribution at stations within the area oc- ings are consistent with the hydrographic cupied by the plume at low, but not high, conditions discussed above. During periods water (e.g., station Q1-10) show that this material is present within the water column (ppm) 8 SAL.(%o) VOLUMECONCENTRATION 20 50 in discrete layers. Figure 9 is illustrative 30 10 40 in this respect; here transmissivity data (which corresponds very closely to sedi- (m) 20 50 100 60 100 140 180 220 LOAD (mg/1) RELATIVE % TRANSMISSION FIG. 9.-Distribution patterns of particulate sediment over tidal cycle at station Q1-10 in September. FIG. 10.-Particulate sediment-load distribution (closed circles) and volume concentration (crosses; see text) at low water at station Ql-10 in July; open circles = salinity profile. This content downloaded from 192.148.225.018 on October 27, 2017 06:57:38 AM All use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c). 546 CHARLES M. HOSKIN AND DAVID C. BURRELL of low freshwater inflow, when the upper part of the inlet has the characteristics of a partially mixed estuary, the rapidly settling sediment is mixed and dispersed more vertically. Conversely, at the time of maximum runoff, the net downward velocity of the sediment is decreased because of the greatly increased surface outflow, but the surface hydrography is that of river-dominated flow. This means that the circulation cell is shallow, vertical mixing is minimal, and the marine reservoir beneath the halocline is relatively stagnant. Under these conditions, the discreteness of the floc layers may be maintained through the bulk of the water column. It should be noted that there are no anomalous density stratifications beneath the halocline. The relevant velocity profiles for the comparable Silver Bay fjord have been given by Rattray (1967, p. 55-56). Under high-runoff conditions in Queen Inlet, the subhalocline current velocities obtained to date (of the order of 5 cm/sec or less) are outside the operating range of the measuring system used. FLOCCULE SIZE The sediment volume-concentration profile given in figure 10 was obtained by a Coulter Counter electronic sizing technique (Burrell and Hadley 1971). This procedure is applicable only to the silt-size fraction (4-64 /i) of the suspended sediment. The covariance of the volume and load plots thus indicates that this size component is representative of the total sediment. The density range given by these values is 2.73.5 g/cm3. If we assume the true density of this silicate material (the organic carbon content of the deposited sediment at station Q1-10 is < 0.1%) to be something less than 2.0, it would appear that less than 50% of this sediment has been rejected by the sizing technique and hence lies outside the silt-size range. Major silt-size modes for suspended sediment samples taken from the high-load bands within the water column lie in the range 7.0-7.6 p (7.8-5.2 /). This does not exactly coincide with the inletfloor major size mode range (see below), but this discrepancy may be due to the totally different sizing techniques used. It must be reemphasized that these data refer to primary particles and not floc agglomerates. ACCUMULATED SEDIMENTS SOURCE The size-frequency distribution of recently exposed ice-contact sediment (muddy sandy gravel; terminology of Folk 1954) at the snout of Carroll Glacier has been examined in detail by Slatt (1970, 1971). This material is the principal source of sediment now being redistributed and accumulating in other environments in Queen Inlet. Slatt (1970) has shown the grainsize composition of the ice-contact sediment to be gravel > sand > silt > clay, and he (Slatt 1971) has shown from pebble roundness measurements that some of this material may be reworked outwash. INTERTIDAL OUTWASH FAN Entrainment of the ice-contact sediment by meltwater results in strong grain-size fractionation. Boulder gravel is concentrated as lag material in the terminal moraine, and pebble gravel and sand are deposited from the meltwater streams, forming the bulk of the outwash fan (4 samples from the intertidal portion of the fan average 2 gravel, 91 sand, and 7 mud, % w/w); mud and some sand are transported across the fan into the marine environment. Temporary deposition and subsequent erosion are frequently occurring processes in the outwash fan. Silt and clay dispersed in meltwater streams flocculate as they mix with salt water in the fjord basin. Flood tides transport the suspended floccules onto the outwash fan, and the settling lag effect (Kuenen 1961) results in accumulation of mud in the intertidal zone during the meltwater season. A piston core taken from this mud deposit in September 1967 showed a thickness of 81 cm, with an average mud content of 94% w/w. In winter, meltwater stream flow ceases, and mud is no longer This content downloaded from 192.148.225.018 on October 27, 2017 06:57:38 AM All use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c). SEDIMENT TRANSPORT AND ACCUMULATION IN A FJORD available from that source. However, the tides continue to rework the outwash fan; by April 1968 mud was gone from the fan and the site where the core was taken 6 months earlier. In other environments,such as the Dutch tidal flats (Kuenen 1961) and some estuaries (Chesapeake Bay; Schubel 1971), mud is accumulated through the filter-feedingactivity of invertebrateorganisms. This is thought to be of little importance in Queen Inlet because so few living organisms have been observed; carbonate shells and shell fragments have not been seen in the 37 submarinesamples analyzed. FJORD BASIN General.-There is a remarkable partition between sandy silt (grain-size modes of 2.3 4, 205 4; 4.5 4, 44 p) in inlet-floor valleys and mud (grain-size mode of 6.5 4, 11 p) on the predominantlyflat floor of the inlet. This is most clearly seen in the form of size-frequencydistributionsfor these two sediments (fig. 11). Neither sediment has a simple Gaussian distribution; both are bi- or polymodal and fine skewed, with inlet-floor valley sandy silt and flat-bottom channelsedimentin inletfloorvalleys n=12 40 30 QUEEN INLET 20 mud on flat bottom 10 0 2 3 Sand 4 5 6 7 Mud 2 3 4 5 6 Sand 7 8 9 10 Mud FIG. 11.--Size-frequency distribution for sediment from the inlet-floor valleys (fig. 2) and for flatbottom samples. 547 mud being leptokurtic and mesokurtic, respectively. These grain-size distributions are open-ended, as no technique conveniently generates data for fractions finer than 10 4, 0.98 i, which amounts to an average of 37% w/w for the flat-bottom mud. It is suggested that the textural differences between flat-bottommud and inletfloor valley sandy silt indicate provenance by two distinct processes. The overflow and underflow concept developed by Gould (1951) for the Colorado River and Lake Mead may be applicable here. Overflows correspond to the surface-sediment plume (above). The inlet-floor valleys with their natural levees, terraces,and coarser-grained axial sediment are strong evidence for underflows, but attempts to measure current flow directly within the inlet-floor valleys have not provedsuccessful to date. Underflows.-It has been observed that many other Alaskan fjords with nontidal glaciers do not have the feature we term "inlet-floor valleys." We suggest that the inlet-floor valleys are formed and maintained by underflows related to meltwater and sediment discharge. Clearly, more data are needed, but as the geometry and dimensions of the outwash fan in Queen Inlet, to our knowledge, are not duplicated in other fjords of Glacier Bay, they may be key factors in underflowgeneration. If the sediment load of the outwash streams was great enough, the streams might continue to flow along the fjord bottom, but Bates (1953) has pointed out that few streams have sufficient load to sink beneath the usually denser seawater. Queen Inlet seawater, with a temperatureof 6° C and 31 %o salinity, has a density of 1.0243 g/cm3 (Zerbe and Taylor 1953). Measurements of suspended-sedimentload made in September 1969 in the largest meltwater stream in Queen Inlet (upstream from the high-tide mark) had a load range of 11.722-14.504, and an average load of 12.769 g/liter, at a water temperatureof 0.5° C (Slatt 1970, appendix C, p. 111). These suspended-sediment loads result in meltwater stream densities of This content downloaded from 192.148.225.018 on October 27, 2017 06:57:38 AM All use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c). 548 CHARLES M. HOSKIN AND DAVID C. BURRELL 1.0073-1.0090 g/cm3, with the average suspended-sediment load corresponding to a stream density of 1.0079 g/cm3. As calculations show that a suspended-sediment load of 39 g/liter is required to equal the density of local seawater, the meltwater stream cannot sink due to the suspendedsediment load in transit at that time, and thereforemust form an overflow.Suspended load in meltwater streams varies with time of day and season, as documentedby Rainwater and Guy (1961) for the Chamberlin Glacier in the eastern Brooks Range, Alaska. The Queen Inlet meltwater streams at other times might have greater suspendedsediment loads, but such data are not available. U.S. Geological Survey water-supply papers contain no suspended-sedimentload data for the Glacier Bay area. To show that suspended-sedimentloads in Alaskan meltwater rivers do exceed the density of Alaskan coastal seawater, we cite data from the Susitna River of south-central Alaska, which drains valley glaciers on the south slope of the Alaska Range. Suspended-sediment loads at two different stations on the Susitna River exceeded 72 g/liter on July 16, 1965 and 68 g/liter on September 6, 1965 (U.S. Department of the Interior 1970). As the inlet-floor valleys appear to begin where the meltwater streams leave the outwash fan and join the fjord basin, it is tempting to postulate that the stream flow separates, with the bedload part continuing to flow down the slope of the submarine outwash fan (which exceeds 2.50) and then along the fjord floor. This would occur if the bedload material were sufficient to exceed the density of local seawater; unfortunately, measurements of stream bedload are difficult to obtain and are not now available for Queen Inlet. The outwash fan (gravel and sand) has been built by meltwater streams,and therefore,bedload transport must be considerable. Another possibility is that meltwater stream-bed material continually accumulates just seaward of the stream mouth. Periodically, this material may slide down the slope of the out- wash fan into the fjord basin, thereby generating an underflow. The latter may be analogous to the underwater sand falls of Southern California submarine canyons as illustrated by Shepard and Dill (1966, fig. 55). Holtedahl (1965) has clearly documented turbidity currentsas important sediment transporters in Hardangerfjord, Norway. It is uncertain whether turbidity currents are active in Queen Inlet, as graded beds have not been seen in the few short (3-4 m) cores available from the inlet-floor valleys. Also, the floor of Hardangerfjordis flat (Holtedahl 1965, figs. 37 and 39), whereas the inlet-floor valleys in Queen Inlet may suggest activity of some process other than turbidity currents. It is not yet known whether the sediment-transporting bottom flows in Queen Inlet are continuous or sporadic; most probably they are not continuous.However, as mud floccules sediment over the entire fjord basin during the time of meltwater flow, and as only sandy silt is found in the inlet-floor valleys, the flows must occur fairly frequently. Slumps.-Slumping in the ordinary sense (and as distinct from sliding, above) is not believed to be an important process in Queen Inlet. Slumpingin other fjords forms mounds of sediment at the juncture between the fjord floor and wall (Von Huene 1966, in Nuka Bay, Alaska; and Cone et al. 1963, in Hardangerfjord, Norway); these mounds have not been seen in Queen Inlet. It is also important to note that slump deposits in Hardangerfjordare not located near river mouths (Holtedahl 1965, p. 129). A further argument against slumping in Queen Inlet concerns the distribution of gravel (particles > 2,000 p). Gravel is abundant in subaerial inlet wall colluvium (average 49% w/w) and the intertidal portion of the outwash fan (average 2% w/w). Gravel has not been found in 14 grabs of sandy silt from inlet-floor valleys or from eight grabs and seven cores of mud from the flat fjord floor. Lack of gravel in the fjord basin argues for an absence of slumping. This content downloaded from 192.148.225.018 on October 27, 2017 06:57:38 AM All use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c). SEDIMENT TRANSPORT AND ACCUMULATION IN A FJORD Ice rafting.-Major sediment transport by ice rafting (identified by gravel supply) either from Carroll Glacier within Queen Inlet or from other inlets into Queen Inlet may be discounted by lack of gravel in the fjord basin. Ice rafting is important in areas adjacent to Queen Inlet, since gravelrich bottom sediment (70-80% w/w) has been recovered from upper Glacier Bay inlets. Ovenshine (1970) has provided additional documentationof this process. Sedimentation rates.-Exceedingly high rates of sediment accumulationat the heads of many Alaskan fjords have been proposed, but direct rate measurements are generally not available. If sediment thickness data from seismic reflection profiles is used, accumulation is assumed to have occurred during the period 10,000 years B.P. to today (Von Huene 1966, fig. 2, and p. 295), the accumulationrate for the basin at the head of Nuka Bay is of the order of 0.45 m/year. Jordan (1962) interpreted changing bathymetric profiles within several Alaskan fjords in terms of accumulation of glacial sediment. Caution must accompany such interpretations, however, as changing bathymetry may partially reflect movement of the glacier and, possibly, the different sounding and navigation techniques used. Sediment accumulationin the fjord basin by settling from the surface plume occurs mainly through the summer-fall seasons (our field observations, 1966-1970), but this periodicity is not obviously apparent from grain-size characteristics of the bottom-sediment column. Three-meter piston cores of fjord bottom mud contain essentially uniform sediment (1 sand, 66 silt, 33 clay, % w/w), with thin and nonsystematic sand layers a few grain diameters thick. Interestingly, all mud cores do contain thin layers (few millimeters) of intensely black mud-sized particles. We have found these black layers in other inlets in Glacier Bay, and in Endicott Arm (fig. 1; D. E. Buckley, personal communication, 1968). Grid coring of Queen Inlet (fig. 12) has shown that the black layers are re- 549 FIG. 12.-Locations of black marker horizons in piston-core profiles, indexed in meters from sediment surface. Dotted lines show position of inletfloor valleys. peated on an approximatelyregularpattern within each core; we believe that these black layers are produced on some seasonal pattern. Regular annual production, a priori, would seem most likely. Spacing between black layers, accepting the seasonal origin, would indicate sediment accumulation rates in excess of 1 m/year at the head of Queen Inlet-high values, but possible. Spacing between black layers gradually decreases toward the mouth of Queen Inlet (fig. 12), and this is interpreted to mean that sediment is supplied by the outwash streams at the inlet head. It should be noted also that, at the head of the inlet, the band spacing is greater for the western profile; increased deposition on this side of the fjord would be consistent with the sediment-plume distribution discussed above. Lack of increased spacing between black layers in core 2634 (fig. 12), mile west of a gravelly delta built by a surface stream, shows no important sediment input via underflowsto the fjord basin from this stream. Some sediment escapes into Glacier Bay, This content downloaded from 192.148.225.018 on October 27, 2017 06:57:38 AM All use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c). 550 CHARLES M. HOSKIN AND DAVID C. BURRELL as we have observed the surface-sediment plume south of Composite Island at low tide. Suspended sediment may be transported into Queen Inlet by the tides, but no data are available on this point. The origin of the black layers is not known at present, and the layers do not correlate with any other discernable properties of the cores. Chemically reduced bands analogous to the varved deposits of British Columbia fjords 600 miles to the south (Gross et al. 1963) are not a likely interpretation because neither Queen Inlet (nor any other Alaskan fjord) has been shown to go anoxic. All sediment Eh measurements-admittedly crude in this poorly poised environment-have indicated an oxidizing environmentthrough at least the top meter of bottom mud. Organic carbon profiles (method of Menzel and Vaccaro 1964) also have shown no significant enhancement in the black horizons. However, the accuracy of these data is limited because of the high inorganic carbonate content (glacially abraded limestone and dolomite) and the thinness of the layers. Further chemical study is in progress. from the intertidal portion of the outwash fan. The bulk of Queen Inlet bottom sediment is texturally uniform mud, both areally and with depth-at least down to 3 m. Piston cores show thin, black marker horizons which are judged from spacing increments to be seasonally, and possibly annually, produced. Spacing between black layers increases toward the juncture of the meltwater streams and the fjord head, which suggests mud is most abundantly supplied by the meltwater streams. Spacing of black layers does not increase in the vicinity of other surface streams, showing their lack of significant sediment input. Accumulationrate of mud (uncorrectedfor water content and compaction) may exceed 1 m/year at the inlet head, decreasing to about 0.4 m/year near the inlet mouth. Sand deposited by meltwater streams has built an extensive outwash fan, which continues to grow seaward. Sand is also supplied to the fjord basin, perhaps by the underflow mechanism of Gould (1951). Sinuous inlet-floor valleys with natural levees, terraces, and sandy silt axial sediment are evidence for some form of sediment-transportingbottom flow. SUMMARY Gravel is present in the outwash fan and This study of sediment transport and subaerial inlet wall colluvium, but has not accumulation in an iceberg-free fjord has been recovered from the fjord floor, which shown that mud and sand are entrained discounts slumping, land sliding, and ice from glacier snout ice-cored deposits during rafting as sediment suppliers in Queen Inthe summer-fall season. Size sorting during let. transport in meltwater streams results in ACKNOWLEDGMENTS.-Some of these data separationof mud and sand. Mud is supplied to the fjord dispersed were presented initially at the American Geoand suspended as a thin surface plume over physical Union meeting, Portland, October the marine surface. Mixing with saline 1969. The work has been supported in part by U.S. Atomic Energy Commission contract water causes flocculation, and the floccules AT(04-3)-310, with ship time provided on the settle through the water column in discrete R/V Acona under NSF grant GP5515. We are layers. It is suggested that each of these grateful to our colleagues for many forms of layers represents one tidal cycle. Mud is help, and to Dr. O. T. Ovenshine for much also seasonally deposited and reworked appreciated critical comments. REFERENCES CITED BATES,C. C., 1953, Rational theory of delta formation: Am. Assoc. Petroleum Geologists Bull., v. 37, p. 2119-2162. BUFFINGTON,E. C., 1952, Submarine "natural levees": Jour. Geology, v. 60, p. 473-479. BURRELL,D. C., and HADLEY,R. S., 1971, Suspended sediment size fractionation within a fjord-estuary: application of electronic counting techniques (Abs.): Am. Geophys. Union Trans., v. 52, p. 853. This content downloaded from 192.148.225.018 on October 27, 2017 06:57:38 AM All use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c). SEDIMENT TRANSPORT AND ACCUMULATION IN A FJORD CAMERON, W. M., and PRITCHARD,D. W., 1963, Estuaries, in HILL, M. N. ed., The sea: New York, Interscience, v. 2, p. 306-324. N. S.; and KENYON,K. E., CONE,R. A.; NEIDELL, 1963, Studies of the deep-water sediments with the continuous seismic profiler. 5. The natural history of the Hardangerfjord: Sarsia, v. 14, p. 61-78. FIELD,W. 0., 1969, Current observations on three surges in Glacier Bay, Alaska, 1965-1968, in Seminar on the causes and mechanics of glacial surges, St. Halaire, Quebec, 1968: Canadian Jour. Earth Sci., v. 6, p. 821-839. FOLK,R. L., 1954, The distinction between grain size and mineral composition in sedimentaryrock nomenclature: Jour. Geology, v. 62, p. 344359. GOLDTHWAIT, R. P., 1966, Glacial history, part 1, in GOLDTHWAIT, R. P.; LOEWE,F.; UGOLINI, F. C.; DECKER,H. F.; DELONG,D. M.; TRAUTT. R., III; MAN,M. B.; GOOD,E. E.; MERRELL, and RUDOLPH,E. D., Soil development and ecological succession in a deglaciated area of Muir Inlet, Southeast Alaska: Inst. Polar Studies Rept. 20, p. 1-18. GOULD,H. R., 1951, Some quantitative aspects of Lake Mead turbidity currents, in Turbidity currents and the transportation of coarse sediments to deep water-a symposium: Soc. Econ. Paleontologists and Mineralogists, Spec. Pub. 2, p. 34-52. GROSS,M. B.; GUCLUER,S. M.; CREAGER, J. S.; and DAWSON,W. A., 1963, Varved marine sediments in a stagnant fjord: Science, v. 141, p. 918-919. HOLTEDAHL, HANS, 1965, Recent turbidities in the Hardangerfjord, Norway: Colston Research Soc. Proc., v. 17, p. 107-140. _--1967, Notes on the formation of fjords and fjord-valleys: Geografiska Annaler, v. 49, p. 188-203. D. C., 1968, InletHOSKIN,C. M, and BURRELL, floor valleys and sediment, Queen Inlet, Glacier Bay, Alaska, in BURRELL,D. C., and HOOD, D. W., eds., Clay-inorganic and organic-inorganic associations in aquatic environments: Univ. Alaska Inst. Marine Sci., Rept. R68-1, pt. 2, p. 89-112. JORDAN, G. F., 1962, Redistribution of sediments in Alaskan bays and inlets: Geog. Rev., v. 52, p. 548-558. KUENEN,P. H., 1961, Problems of epicontinental sedimentation, in SEARS,MARY,ed., Oceanography: Am. Assoc. Adv. Sci. Pub. 67, p. 475-495. M.; and BARNES, McALISTER,W. B.; RATTRAY, C. A., 1959, The dynamics of a fjord estuary, Silver Bay, Alaska: Univ. Washington Dept. Oceanogr., Tech. Rept. 62, 67 p. 551 MACKEVETT, E. M.; BREW,D. A.; HAWLEY,C. C.; LYMAN,C. H.; and SMITH,J. G., 1971, Mineral resources of Glacier Bay National Monument: U.S. Geol. Survey Prof. Paper 632, 90 p. MENZEL,D. W., and VACCARO, R. F., 1964, The measurement of dissolved organics and particulate carbon in seawater: Limnology and Oceanography, v. 9, p. 138-142. OVENSHINE, O. T., 1970, Observations of iceberg rafting in Glacier Bay, Alaska, and the identification of ancient ice-rafted deposits: Geol. Soc. America Bull., v. 81, p. 891-894. PICKARD, G. L., 1967, Some oceanographic characteristics of the larger inlets of southeast Alaska: Jour. Fisheries Research Board Canada, v. 24, p. 1475-1506. PRITCHARD,D. W., 1955, Estuarine circulation patterns: Am. Soc. Civil Engineers Proc., v. 81, p. 717/1-717/11. QUINLAN, A. V., 1970, Seasonal and spatial variations in the water mass characteristics of Muir Inlet, Glacier Bay, Alaska: Unpub. M.S. thesis, Univ. Alaska, 145 p. F. H., and GUY, H. P., 1961, Some RAINWATER, observations on the hydrochemistry and sedimentation of the Chamberlin Glacier area, Alaska: U.S. Geol. Survey Prof. Paper 414-C, p. ClC14. RATTRAY, M., 1967, Some aspects of the dynamics of circulation in fjords, in LAUFF, G. H., ed., Estuaries: Washington, D.C., Am. Assoc. Adv. Sci., p. 52-62. SCHUBEL, J. R., 1971, A few notes on the agglomeration of suspended sediment in estuaries, in SCHUBEL, J. R., convener, The estuarine environment, estuaries and estuarine sedimentation: Washington, D.C., Am. Geol. Inst., p. X1-X29. F. P., and DILL, R. F., 1966, Submarine SHEPARD, canyons and other sea valleys: Chicago, Rand McNally, 381 p. SLATT,R. M., 1970, Sedimentological and geochemical aspects of sediment and water from ten Alaskan valley glaciers: Unpub. Ph.D. dissertation, Univ. Alaska, 125 p. --1971, Texture of ice-cored deposits from ten Alaskan valley glaciers: Jour. Sed. Petrology, v. 41, p. 828-834. U.S. Department of the Interior, 1970, Quality of surface waters of the United States, 1965: U.S. Geol. Survey Water-Supply Paper 1966, pts. 1216, 462 p. VON HUENE, ROLAND, 1966, Glacial-marine geol- ogy of Nuka Bay, Alaska, and the adjacent continental shelf: Marine Geology, v. 4, p. 291-304. ZERBE,W. B., and TAYLOR, C. B., 1953, Sea water temperature and density reduction tables: U.S. Coast and Geod. Survey Spec. Pub. 298, 21 p. This content downloaded from 192.148.225.018 on October 27, 2017 06:57:38 AM All use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c).