Episodic Holocene alluviation in meltwater channels near Calgary, Alberta

Episodic Holocene alluviation in meltwater channels near Calgary, Alberta Can. J. Earth Sci. Downloaded from www.nrcresearchpress.com by MICHIGAN STA...
Author: Noel Harrison
5 downloads 0 Views 740KB Size
Episodic Holocene alluviation in meltwater channels near Calgary, Alberta

Can. J. Earth Sci. Downloaded from www.nrcresearchpress.com by MICHIGAN STATE UNIV on 01/28/17 For personal use only.

GERALDD. OSBORN Department of Geology, University of Calgary, Calgary, Alta., Canada T2N IN4 Received 30 December 1975 Revision accepted for publication 4 February 1977

Sheetlike layers of sediment consisting of coarse clasts in a fine-grained matrix occur at the bases of steep meltwater-channel slopes north and northwest of Calgary. Individual layers are separated by thin organic horizons which are probably accumulations of water-transported burned organic matter. The deposits are interpreted to be products of a series of unchannelled water-sediment slurries created whenever heavy rains fell on denuded or destroyed grass cover stripping the soil from the steep upperparts of the valley slopes. Fire was most likely responsible for periodically destroying the grass cover. At least 24 erosion-alluviation events occurred in one of the channels in the last 15 000 years. These alluviation events seem to be the main process by which the meltwater channels are being infilled. On peut observerdes couches en feuillets de sediments consistant en particules grossieres et en une matrice a grain fin B la base des pentes abruptes des chenaux d'eau de fonte au nord et au nord-ouest de Calgary. Les couches individuelles sont separees par de minces horizons organiques qui sont probablement des accumulations de matiere organique brDlee et transportee par I'eau. On interprete ces dep6ts comme les produits d'une sCrie de coulees boueuses contenant des kdiments et de I'eau non canalisie chaque fois que des pluies abondantes ont attaqui le sol denudi ou la couverture vk&tale ditruite pour eroder le sol dans les parties supirieures pIus raidesdes venants. Le feu est trks probahlernent responsable de la destruction *riodiques de la couvenure vtgktale. Au moins 24kpisodes d'emsion-xlluvionnement se sont pmduits dans un de ces chenaux au c o w s des 15 000 demitrcs annees. Ces e p i s d e s d'alluvionnement semblent itre le processus principal de rernplissage des chenaux d'eau de fonte. [Traduit par le journal]

Can. J. Earth SCI.,14, 1515-1520(1977)

Introduction Thin organic horizons separate layers of alluvium and colluvium in glacial meltwater channels north and northwest of Calgary. The deposits indicate that alternating periods of stability and erosion occurred on the slopes of the channels during the Holocene. Counts of the horizons enable calculation of the minimum average recurrence intervals of the erosional events, which were probably consequences of fire and intense rain. The deposits are of geomorphic interest because they suggest that fairly large quantities of sediment, including large pebbles, can be transported by unchanneled slope wash. The three channels in which the deposits are found are the valleys of Bighill Spring Creek, Beddington Creek, and Nose Creek (Fig. 1). They are all narrow, relatively deep (25-1 10 m) channels which were presumably related to the retreating late Wisconsin Laurentide ice sheet, and now contain small underfit streams (Fig. 2). The channels, generally concave-upward in cross section, are incised into sandstone of the Paleocene Porcupine Hills Formation; outcrops are

114

~d

FIG.I. Index map.

common on the steep upper parts of the channel slopes. The lower slopes are remarkably smooth and are interrupted only rarely by coulees and associated alluvial fans. The northwest-facing slope of the Bighill Spring channel is forested with conifers and aspen, while the southeast-

Can. J. Earth Sci. Downloaded from www.nrcresearchpress.com by MICHIGAN STATE UNIV on 01/28/17 For personal use only.

1516

CAN. J. EARTH SCI. VOL. 14. 1977

facing slope and all the slopes of the other channels support mostly grass, with scattered clumps of trees and shrubs. At the mouth of the Bighill Spring channel is a series of deltas that were deposited in glacial Lake Calgary in late Wisconsin time. Stalker (1968) estimates the age of these deltas as between 15 000 and 19 000 years. The channel presumably was being actively eroded until about that time. A similar history can be inferred for the other channels.

The Deposits The thickest deposits occur in the Bighill Spring channel. Exposures occur along the single dirt road running through the channel. Holocene deposits are present through the length of the channel but the best exposures occur on the southeast-facing slopes 2.5 km upstream from the mouth of the channel (site 1, Fig. 1). In this area the channel wall slopes at an angle of 30-35" near the top, grading to 10" above the roadcut (Fig. 3). The thickest exposed section is about 4.3 m thick. The deposits here, exemplified in Fig. 4, are unconsolidated but cohesive. They consist mainly of pebbly sandy silt and silty sand. Pebbles are of two types: angular Porcupine Hills sandstone and rounded quartzite. The former range in size from sand to 12 cm across, with rare larger clasts. Quartzite pebbles are from 1 to 5 cm in size. Pebbles, whether of sandstone or quartzite, are commonly concentrated into poorly sorted beds and lenses, 0.5 m to tens of metres in lateral extent and 3-30 cm thick (Fig. 4). In most cases pebble beds and lenses appear to lie conformably over underlying sediment but in a few cases wide, shallow cut-aid-fill structures are visible.

FIG. 2. A portion of the Bighill Spring channel, looking southwest. Underfit Bighill Spring Creek flows on left.

Outside of the beds and lenses, single pebbles are isolated within the finer-grained sediments. A few large (0.3-1.3 m), flat blocks of sandstone are scattered through the section (Fig. 4). They are invariably aligned in a roughly horizontal attitude. Dark gray organic-rich horizons separate the deposits into thin layers. Organic horizons are a few millimetres to 3 cm thick, have (usually) sharp upper and lower boundaries, and are spaced 1-5 cm apart over most of the exposure. In some cases two horizons gradually merge to become one and may gradually bifurcate again elsewhere. Horizons remain roughly horizontal along the axis of the valley, but gullies and artificial cuts show that the horizons dip slightly toward the axis of the valley, maintaining parallelism with the hillslope. Organic horizons are laterally continuous through the exposure, terminating only at the rare cut-and-fill structures. The horizons do not occur within pebble beds or lenses, but rather only at their tops or bottoms. Where pebble beds are absent, however, there is usually no discernible difference in the sediments above and below a particular organic horizon. The present soil, at the top of the section, consists of an Ah horizon (Canadian Department of Agriculture 1974) 10-15 cm thick. A volcanic ash layer, about 1 cm thick, overlies a thin organic horizon about 1 m below the top of the section. The ash is identified as Mazama, as described later. Other good exposures in the southeast- or east-facing slopes occur just north of Bighill Spring Provincial Park (site 2). Exposures are a maximum of 1 m in thickness; the thickest sections occur where the road cuts through small

FIG. 3. Photograph showing relationship of Holocene deposits (exposed by roadcut) to channel slopes in the Bighill Spring channel. Sandstone outcrops are visible on the upper slopes.

Can. J. Earth Sci. Downloaded from www.nrcresearchpress.com by MICHIGAN STATE UNIV on 01/28/17 For personal use only.

OSBORN

1517

FIG.4. Layered deposits and organic horizons in the Bighill Spring channel. Note concentrations of pebbles in lower part of exposure, and large sandstone slab above shovel. Arrow points to level of Mazama ash.

alluvial fans associated with minor coulees in the upper slopes. Slope angles are the same as in the previously described exposures, but sandstone outcrops on the upper slopes are much rarer. The deposits consist mainly of rounded quartzite pebbles and cobbles, up to 20 cm long, set in a silty sand matrix. Sandstone clasts are rare. Some of the quartzite stones are concentrated into poorly defined, discontinuous beds, while others are isolated and randomly distributed in the matrix. A maximum of 10 organic horizons are visible at any one site. They are 1-4 cm thick and are spaced 1-15 cm apart. Many of the horizons have rather diffuse upper and lower contacts. No ash is visible. Several other exposures occur in southeastfacing slopes elsewhere in the channel. The deposits are similar to those already described, although exposed sections are thinner and the number of exposed organic horizons is consequently fewer. Northeast-facing slopes were examined to see if the different slope aspect influenced the nature of the deposits. At site 3 thick forest cover

extends all the way down the slope to the creek. Stream bank exposures and pits dug 10 m up the slope from the creek reveal homogeneous clayey silt containing scattered quartzite and sandstone pebbles. There is no bedding and there are no organic horizons. The sediment is significantly finer, less compact, and less cohesive than that on the southeast-facing slopes. At site 4 the road cuts through a small fanshaped deposit whose surface slopes at an angle of 15". Grass and shrub cover extends upslope for 12 m, at which point aspen forest begins. The deposit consists of compact, unstratified clayey sandy silt containing isolated rare quartzite pebbles and common sandstone pebbles. Except for the lack of organic horizons, the sediment is similar to that on the southeast-facing slopes. A few hundred metres to the north forest cover becomes more sparse in places and road cuts reveal a few diffuse organic horizons in pebbly, sandy silt. The sediment here is similar to that across the channel, although pebbles are not quite as common.

Can. J. Earth Sci. Downloaded from www.nrcresearchpress.com by MICHIGAN STATE UNIV on 01/28/17 For personal use only.

1518

CAN. J. EARTH SC:I. VOL. 14. 1977

It is concluded from these observations that the process(es) responsible for the layered sediments and organic horizons is(are) not effective where trees cover most of the slope. Conversely, grass or grass and shrub cover is conducive to the process(es). Further evidence for such a relationship is the fact that forested slopes tend to be slightly more planar than grass slopes which tend to be concave (Fig. 2). This is to be expected if sediment is collecting at the bases of grass slopes but not of forested slopes. In the north-south trending channel of Nose Creek exposures are not common, but an excellent exposure occurs in the west-facing slope at site 5. Here a creek meander has cut into the base of the old channel wall to expose a cross section perpendicular to the channel axis. The slope angle varies from about 3" on the west end of the cut to 10" on the east. Slightly further east the slope rises steeply (30-35") to the rim of the channel. Sandstone crops out on the steep upper slopes. The exposed sediment consists of mostly isolated pebbles and slabs of sandstone, up to 10 cm long, in sandy silt matrix. Pebble beds are rare in the upper part of the 1-1.2 m thick exposure but become more common near the bottom. Quartzite pebbles are very rare throughout. At least 10 organic horizons, all rather diffuse, are visible; their slope parallels that of the ground surface. In the second (from the top) organic horizon a bison(?) rib fragment was found. Most of the modern Ah soil horizon at this site has been eroded away, probably by wind. In the Beddington Creek channel, road cut exposures of organic horizons are common north of latitude 51°15' (Fig. 1). The most informative section is on a west-facing slope at site 6 (Fig 5). The road here cuts through a slight bulge in the slope which appears to be an alluvial fan. The slope angle is 12" at the cut, increasing to 20" farther upslope. Sandstone crops out on the upper slopes. The deposits and organic horizons are similar to those already described; a maximum of 12 horizons occur at any one place in the 1.3 m thick exposed section. Of interest are the following. (I) Horizons are most numerous in the center of the bulge; they converge to the left and right as the sediment layers between horizons pinch out. (2) Two 1 cm thick organic horizons occur within the 23 cm thick Ah surface soil; they are barely distinguishable from the dark gray soil by their slightly darker color. The uppermost is 15 cm below the surface.

FIG.5. Organic horizons and isolated pebbles in the Beddington Creek channel.

Origin oft he Deposits A channel-slope origin for the deposits, rather than an up-channel fluvial origin, is indicated by (I) association with alluvial fans on the channel slopes, (2) large flat blocks that were not likely stream-carried, (3) angularity of the sandstone pebbles which suggests the pebbles moved very short distances, (4) cut-and-fill channels oriented roughly perpendicular to the channel axis, and (5) organic layers that dip downslope, perpendicular to the channel axis. Clasts of sandstone were derived from the sandstone cropping out on the upper slopes of the channels. Rounded quartzite pebbles were undoubtedly derived from till on the rims of the prairie surfaces into which the channels were cut. Such pebbles are common on the rim of the Bighill Spring channel (whose deposits contain numerous quartzite pebbles) but rare on the rim of the Nose Creek channel (whose deposits contain very few quartzite pebbles). The downslope movements responsible for the deposits were of two kinds. The large flat blocks of sandstone could not have been water-carried down the slope, nor is there any evidence to suggest they were borne on landslides. Thus they probably moved some distance by creep and are by definition colluvium. Sandstone slabs now littering the slopes below the outcrops are the present-day manifestation of the process. However, the fact that organic horizons were repeatedly buried indicates that periods of rapid downslope movement occurred, in addition to the creep that is always occurring. An exigent organic mat could not have been buried by creeping colluvium because the mat would be part of the creeping mass, not independent of it. Thus the bulk of the deposits is considered to be allu-

Can. J. Earth Sci. Downloaded from www.nrcresearchpress.com by MICHIGAN STATE UNIV on 01/28/17 For personal use only.

vium, deposited by slope runoff during circumstances of high erodibility and heavy rainfall. The thin but laterally extensive nature of the individual layers suggests they are products of slope wash (overland flow); the occasional shallow cut-and-fill structures indicate some concentration of water took place but the slope wash was largely unchannelled. In fact, dechannelling may have occurred in some places, such as at the Beddington Creek site described above, where water apparently spread out over an alluvial fan. In any event, by the time the running water reached the bottoms of the slopes it must have been very heavily charged with sediment, and some of this sediment was deposited on the gentle lower slopes of the channels. The moving medium could not strictly be termed a mudflow because the deposits consist mainly of silt and sand, yet the isolated pebbles surrounded by fines indicate a water-sediment slurry was involved. It should be pointed out that while unchannelled slope wash is known to have erosive capabilities (e.g. Emmett 1970; Pearce 1976), natural or artificial slope wash that could move large pebbles has apparently never been observed. In fact, natural slope wash of any kind has rarely been observed, and little is known about it (Emmett 1970). However, there is no question that the sediments described have come from the slopes above them, via largely unchannelled agencies. Presumably precipitation of unusually high intensity is necessary to produce such deposits. The original nature of the organic horizons is not obvious, but it is clear each must represent a past topographic surface. Stalker (1973) attributed similar buried horizons at a site in the Kananaskis Valley, Alberta, to in situ burnedover grass cover. This is a possibility for the horizons described herein; however, work of Dormaar and Lutwick (1975) in the North Saskatchewan River valley suggests that in situ fire-affectedhorizons should be brown or reddishbrown in color. Another possibility is that the horizons are accumulations of burned organic matter transported from the slopes above. The sharp lower contacts of some of the horizons support this hypothesis. If organic matter was indeed transported, it was by the same slope wash that buried the organic matter under alluvium. The erosion-alluviation events must have occurred when heavy rains periodically fell on de-

nuded or destroyed grass cover stripping the soil. Fire was likely involved in the erosion events because it renders the soil easily erodable, can cover fairly large areas at one time, and occurs periodically. Prairie fires were started during Holocene time by both lightning and people. The propensity of Plains Indians in historic times to light prairie fires for diversion or attraction of buffalo is cited by Bird (1961), Macoun (1883), and Hector (1857, in Spry (1968)). Presumably this practice went on for some unknown period in prehistoric time. Although archeological data specific to the Cochrane area is lacking, data from Waterton Park, 230 km to the south, indicate habitation in that area for at least 10 000 years (Reeves 1970). Chronology and Frequency of Events Carbon-14 dating of some of the organic horizons was tried, unsuccessfully. A horizon 0.5 m below the present Ah soil in the Bighill Spring channel was dated as 'modern,' indicating contamination by modern carbon has taken place. The bone fragment found at Nose Creek is unfortunately too small to provide a reasonably reliable I4C date. Thus the ash found in the Bighill Spring channel provides the only definite time marker. This ash was identified as Mazama on the basis of both refractive index analysis of the glass fraction and microprobe compositional analysis of the magnetite fraction. Details of the analytical methodology are given by Duford (1976). Magnetite results are given in Table 1. Mazama ash is considered to be about 6600 years old (Westgate et al. 1970). Each organic horizon represents one erosionalluviation event. A count of the horizons provides a minimum number of events, because some horizons could have been eroded away or otherwise rendered unrecognizable by subsequent TABLE 1. Comparison of magnetite oxide percentages in Bighill Spring channel ash to known values for Mazama ash* Percentage Oxide TiO, Fe2O3 MgO A1703

FeO

+

Mazama ash

Bighill Spring ash

8.63 86.45 2.07 2.11

8.38+0.25? 84.64k0.56 1.93k0.15 2.01+0.12

*Mamma values from Westgate et al. (1970). ?Error figures are standard deviations.

Can. J. Earth Sci. Downloaded from www.nrcresearchpress.com by MICHIGAN STATE UNIV on 01/28/17 For personal use only.

1520

CAN. .I.EARTH SCI.

events. (In addition, without intervening organic horizons it is usually not possible to distinguish deposits of separate alluvial events.) In the thickest exposed section (in the Bighill Spring channel) there are 7 organic horizons between the ground surface and the ash. There are 17 horizons below the ash. The maximum average interval between events over the last 6600 years is thus about 950 years. The maximum age of the lowermost deposits is probably about 15 000 years; all 24 horizons distributed over such a time span give a maximum average recurrence interval of 625 years. Fires, whether natural or man-caused or both, were probably more frequent than this; presumably only the fire that was followed by intense rainfall caused erosion and downslope alluviation to occur. Because the ash was not found at sites other than site 1 it unfortunately cannot be determined how synchronous events were from place to place. The existence of the present thick surface soil at most of the sites, in one case actually enveloping 2 thin organic horizons, suggests that the frequency of events has decreased in recent times. The present soil, if covered during another alluviation event, would leave a relatively thick buried soil. Yet there are no such buried soils in any of the sections. Of course, it is possible that buried soils could have been eroded away, but that would necessitate considerable erosion over long lateral distances. There is no obvious reason why the frequency of alluviation should have decreased in the last several hundred years, if indeed it has. Conclusions Sheetlike layers of sediment consisting of coarse clasts in a fine-grained matrix, and separated by thin organic horizons, occur at the bases of steep meltwater-channel slopes north and northwest of Calgary. The sediments were derived from channel slopes above. The thin, laterally-extensive nature of the layers indicates that the sediments were deposited from largely unchannelled slurries of slope wash and debris. The organic horizons probably represent washed-in deposits of burned organic matter from the slopes above. Fire and subsequent intense rainfall were most likely responsible

VOL.

14, 1977

for the erosion-alluviation events. At the best exposure of deposits 24 layers are visible, giving a maximum average recurrence interval at this site of about 625 years. These alluviation events are apparently the main process by which the meltwater channels are being infilled, although the frequency of events may have decreased in recent times.

Acknowledgements I thank L. V. Hills and Tom Oliver of the Geology Department, University of Calgary, for discussions of the problem; Diane Matt for assistance in the field; and Matthew Duford for analysis of ash samples. The work was supported by a grant from the National Research Council of Canada. BIRD,R. D. 1961. Ecology of the aspen parkland. Canada Department of Agriculture, Publication No. 1066. OF AGRICULTURE. 1974. The sysCANADA DEPARTMENT tem of soil classification for Canada. Canada Department of Agriculture, Publication No. 1455. DORMAAR, J. F. and LUTWICK,L. E. 1975. Pyrogenic evidence in paleosols along the North Saskatchewan River in the Rocky Mountains of Alberta. Canadian Journal of Earth Sciences, 12, pp. 1238-1244. DUFORD,J. M. 1976. Late Pleistocene and Holocene cirque glaciations in the Shuswap Highland Area, British Columbia. PhD thesis, University of Calgary, Calgary, Alta. EMMETT,W. W. 1970. The hydraulics of overland flow on hillslopes. United States Geological Survey, Professional Paper 662-A. MACOUN,J. 1883. Manitoba and the great North-west. Thomas C. Jack, London, England. PEARCE,A. J. 1976. Magnitude and frequency of erosion by Hortonian overland flow. Journal of Geology, 84, pp. 65-80. REEVES,B. 0. K. 1950. An archeological resource inventory of Waterton Lakes National Park. Department of Archeology, University of Calgary, Calgary, Alta. SPRY, I. M. (Editor). 1968. The papers of the Palliser Expedition 1857-1860. The Champlain Society, Toronto, Ont. STALKER, A. M. 1968. Geology of the terraces at Cochrane, Alberta. Canadian Journal of Earth Sciences, 5, pp. 1455-1466. 1973. Surficial geology of the Kananaskis Research Forest and Marmot Creek basin region of Alberta. Geological Survey of Canada, Paper 72-51. WESTGATE,J. A,, SMITH,D. G., and TOMLINSON, M. 1970. Late Quaternary tephra layers in southwestern Canada. In Early man and environments in northwest North America (R. A. Smith and J. W. Smith, Eds.). The Student's Press, Calgary, Alta. pp. 13-34.

Suggest Documents