Introduction

The Colville River Delta has long been recognized forits complex geomorphic environments and its abundantand diverse fish and wildlife populations. To help pro-vide information essential for engineering design andevaluation of potential environmental impacts in prepa-ration for oil development on the Colville River Delta,studies on soil stratigraphy and permafrost develop-ment were conducted during 1992Ð1996 to investigatethe nature and distribution of surficial deposits in thedelta.

The Colville River is the largest river on Alaska'sNorth Slope and is one of eight major rivers with signif-icant freshwater input to the Arctic Ocean (Walker1983). The river's volume and heavy sediment load pro-duces a dynamic deltaic system (665 km2) with diversegeomorphic, hydrologic, and ecological systems(Walker, 1983; Carter and Galloway, 1982, 1985;Reimnitz et al., 1985; Rawlinson, 1993; Jorgenson et al.,1997). Similar to other deltas around the world, theColville River Delta is a complex environment that ischaracterized by migrating distributary channels,waterbodies of various origins, natural levees, sanddunes, sandbars, and mudflats (Walker, 1976, 1983).Unlike temperate and tropical deltas, however, it isgreatly influenced by two other factors: (1) low temper-atures that prevent the movement of most of the annual

precipitation until spring breakup, and (2) the presenceof permafrost (Walker, 1976). Permafrost developmentcreates a variety of patterned-ground features, such asice wedges, ice-wedge polygons, frost mounds and pin-gos, and alters the character of river discharge and ero-sional processes on the delta (Walker, 1976). To assessgeomorphic patterns and processes on the delta, thisstudy investigated the stratigraphy of near-surfacematerials along numerous toposequences in the delta,classified and mapped terrain units, classified anddescribed cryostructures, and analyzed material accu-mulation rates.

Methods

Descriptions and samples were obtained from 40boreholes along 11 toposequences and from 8 bankexposures. Cryogenic structure (form, distribution, andvolume of ice) originally was classified in the field in1995 according to Russian (Katasonov, 1969) and NorthAmerican (Pihlainen and Johnston, 1963) systems andin 1996 according to Murton and French (1994). Afterthe fieldwork, we modified the Murton and Frenchclassification system to better differentiate the struc-tures that we observed in the delta (Shur andJorgenson, 1988). The terrain-unit classification systemthat we used was adapted from systems developed byKreig and Reger (1982) and Rawlinson (1993). For map-ping, we used a compound classification system that

Abstract

To help provide information essential for engineering design and evaluation of potential environmentalimpacts in preparation for oil development on the Colville River Delta, studies on soil stratigraphy and per-mafrost development were conducted during 1992Ð1996 to investigate the nature and distribution of surficialdeposits in the delta. The studies involved investigation of stratigraphy of near-surface materials along numer-ous toposequences in the delta, classification and mapping of terrain units, classification and description ofcryostructures, dating and analysis of material accumulation rates, and determination of erosion rates. Afterdetailed classification and analysis of the microscale and macroscale differences in soil properties across thiscomplex landscape, we synthesized the patterns and processes that we observed into a simplified conceptualmodel of the evolution of the deltaic landscape.

M. Torre Jorgenson, et al. 523

EVOLUTION OF A PERMAFROST-DOMINATED LANDSCAPE ON THECOLVILLE RIVER DELTA, NORTHERN ALASKA

M. Torre Jorgenson1, Yuri L. Shur2, H. Jesse Walker3

1. 2. ABR, Inc., PO Box 80410,Fairbanks, AK 997082 Presently with Harding Lawson Associates, Inc., Anchorage, AK

email: yshur@harding.com

3. Louisiana State University,Baton Rouge, LA, 70803email: hwalker@lsu.edu

The 7th International Permafrost Conference524

Figure 1. Map of terrain units in the central portion of the Colville River Delta, northern Alaska.

combined information about terrain units, surfaceforms, and vegetation into integrated terrain units.Only a portion of the terrain unit classification andmapping is presented here, but a more completedescription of the entire ecological land classificationsystem is presented in Jorgenson et al. (1997). Duringmapping, only the surface terrain unit was classifiedbecause of the lack of information about subsurface lay-ers. After analysis of the microscale and macroscale dif-ferences in soil properties across this complex land-scape, we synthesized the patterns and processes thatwe observed into a simplified conceptual model of theevolution of the deltaic landscape.

Results and discussion

ACCUMULATION OF DELTAIC DEPOSITS

The texture and structure of sediments (mineral sedi-ments, organic matter, and ice) on the delta were classi-fied into 15 lithofacies (repeating assemblages of textureand structure), 8 primary cryostructures (ice structuresbased on continuity and shape), and 12 terrain units(three-dimensional architectural elements related todepositional processes) (Figure 1). Large differences inorganic-matter accumulation, ice volume, and overallmaterial-accumulation rates were found among terrainunits (Figure 2), as well as in particle-size distribution,salinity, active-layer depths, and cryostructures.

CONCEPTUAL MODEL OF FLOODPLAIN EVOLUTION ON THE DELTA

By examining similarities in sediment characteristics,organic matter accumulation, and ice aggradation alongterrain sequences, we have identified several generalpatterns and processes that affect floodplain evolution

(Figure 3). The deltaic deposits were formed by fourprocesses: (1) fluvial deposition of mineral material, (2)eolian deposition of mineral material blown off barrenriverbars, (3) accumulation of organic material derivedfrom partially decomposed plants, and (4) accumula-tion of ice. Erosion of these materials by thermal andhydraulic mechanism contributes to degradation ofthese materials and resetting the depositional sequenceof the deltaic landscape. The relative importance ofthese processes in the development of delta-floodplaindeposits changes during the various phases of flood-plain evolution from riverbed/sandbar deposits toabandoned-floodplain deposits. Eolian sand deposits,which are prominent features on the western side ofmost distributary channels, are not included in thismodel, however, because they are not a direct conse-quence of fluvial deposition.

Of particular importance is the formation of syngenet-ic permafrost caused by the addition of new material atthe soil surface, a decrease in active-layer thickness, andby the accumulation of ice below the active layer (Shurand Jorgenson, 1998). Over the course of floodplainevolution, new material is added to the top of the activelayer through deposition of fluvial sediment on the soilsurface and accumulation of organic matter. The accu-mulation of organic material, increased saturation ofthe active layer, and changes in vegetation structurealter the thermal regime of the soils, causing the thick-ness of the active layer to decrease. This addition ofnew material and the decrease in active-layer thicknessresults in new mineral and organic material beingincorporated in the top of the permafrost. At the sametime, ice is formed at the bottom of the active layer,

M. Torre Jorgenson, et al. 525

Figure 2. Comparison of mean (±SD) organic matter accumulation, ice volume, overall accumulation rates (sediments, organics, and ice), relative height (rela-tive to inactive floodplain), and thaw depths among terrain units, Colville River Delta, northern Alaska.

because water freezes to the top of the cold permafrostduring refreezing of the active layer in the fall. Duringsome summers, not all of the ice produced the preced-ing fall thaws; over time, therefore, ice accumulates inthe underlying sediments. This accumulation of sedi-ments, organic matter, and ice causes the permafrostsurface to rise over time altering the hydrologic regimeand sensitivity to thermal degradation.

In the following discussion, we relate the trends thatwe observed to landscape position (relative elevationand distance from river) and we then synthesize theseinto a conceptual model of floodplain evolution on theColville delta that emphasizes the terrain units associa-ted with each distinctive phase of floodplain evolution.

DELTA RIVERBED/RIVERBAR DEPOSITS

Delta Riverbed/Riverbar Deposits, which are situatedin and along the margins of active channels, frequentlyare flooded (every 1Ð2 yr), resulting in scouring andfairly rapid rates of sediment accumulation. The sedi-ments are usually composed of rippled sands or fines,which are typical of lateral accretion deposits, andunderlying massive or inclined (including cross-bedded) sands, which are typical of sandy bedforms.The inclusion of thin detrital peat layers, which becomestranded on the surface by receding floodwaters, is aunique characteristic of these sediments. The coarse tex-ture of the sediments and lack of vegetative cover,result in deep thaw layers (mean depth = 0.73 m).Below the active layer, cryostructures typically aredominated by non-visible pore ice (structureless); ice

contents are low (40Ð50%) because of the sandy textureof the sediments. In addition, the rapid rates of deposi-tion move the active layer upward before much ice canbe formed at the bottom of the active layer. Ice wedgesare absent. Due to the frequent scouring and sedimentdeposition, the surface either is barren or has pionee-ring herbaceous vegetation.

DELTA ACTIVE-FLOODPLAIN COVER DEPOSITS

Delta Active-floodplain Cover Deposits occur onlevel, nonpatterned areas in narrow strips immediatelyadjacent to channels. Flooding is fairly frequent (every3Ð4 yr), but the accumulation rate (2.6 mm/yr) of mate-rial (mostly mineral sediment) is lower than that ofRiverbed/Riverbar Deposits. The sediments generallyhave layered (horizontally stratified silts and very finesands) or massive fines overlying rippled fines withinterbedded silts, sands, and detrital peat layers associ-ated with lateral accretion deposits. The active layer isrelatively deep (0.63 m) and well-drained. With increa-sing distance from channels, the deposits cross a transi-tion zone where flood deposition is less frequent andthin (<1 cm) layers of moss occasionally form at the sur-face. The active layer frequently is saturated at depth(>0.3 m) and the soils are mottled, indicating morereducing conditions. Below the active layer are layeredfines and rippled fines with detrital organics that aredominated by lenticular cryostructures, but which alsoinclude minor amounts of vertically oriented vein ice.Contraction cracks occasionally are evident on theground surface, although polygon rims are not, indicat-ing that ice wedges are in the initial stage of develop-

The 7th International Permafrost Conference526

Figure 3. A conceptual model of changes in lithofacies, cryostructures, and active-layer thickness during evolution of floodplain deposits of the Colville RiverDelta, Alaska.

ment. Well-drained soils near channels support vigo-rous growth of tall willows (Salix alaxensis), whereasseasonally saturated soils farther from the river supportgrowth of low willows (Salix lanata).

During the active-floodplain phase, approximately0.5Ð1.5 m of silty sediment can accumulate on top of theriverbed deposits, which in turn reduces the frequencyof flooding. The accumulation of fine-grained sedi-ments provides the primary material for ice aggrada-tion during the next evolutionary phase. Eventually, theadded sediments and migration of channels reduces thefrequency of flooding to a point where peat starts toaccumulate. Based on material accumulation rates (2.6 mm/yr) and the lack of change in these depositsbetween 1955 and 1992 (Jorgenson et al., 1995), we esti-mate that this phase may persist for 100Ð300 yr. Afterthis early phase of floodplain evolution, a large transi-tion in permafrost development occurs on Inactive-floodplain Cover Deposits.

DELTA INACTIVE-FLOODPLAIN COVER DEPOSITS

Delta Inactive-floodplain Cover Deposits occur imme-diately adjacent to the river along higher cutbanks andat greater distances from the river along point bars.Flood frequency (approximately every 5Ð25 yr) andsedimentation rates are substantially lower than onactive-floodplain cover deposits, thereby allowingbuild-up of organic material and creating distinctiveinterbedded layers of alluvial silt and autochthonouspeat. Layered organics (mean thickness = 0.24 m) ge-nerally are contained within the active layer but some-times extend into the permafrost. The active layerremains saturated throughout the summer, resulting inanaerobic conditions and gleyed soils. Thaw depths(mean = 0.43 m) decrease substantially from the pre-vious phase in response to changes in vegetation com-position and changes in thermal properties of the soil.The sediments in the active layer, which were depositedduring the previous phase, slowly join the permafrost.The upper layer of permafrost, which forms during thisphase from the active layer of the previous phase, isextremely ice-rich (65Ð85% volume) and has distinctivecryostructures (see Shur and Jorgenson, 1998), includ-ing layered, vein, reticulate, and ataxitic ice (see Shurand Jorgenson 1998). The ataxitic and la-yered ice mayrepresent periods when the active layer was static (withlittle sediment accumulation or change in active-layerthickness) and ice accumulated at the same place. Thiscombination of distinctive cryostructures in the upperhorizon of permafrost, which is evident in both theRussian and North American Arctic, has been termedthe "intermediate layer" (Shur, 1988). When layeredorganics extend into the permafrost, organic-matrix ice(an amorphous type of ice generally visible within theloose, fibric organic matrix) is the dominant cryostruc-ture. The permanently saturated soils usually supportwet sedge-willow meadows.

Accumulation of ice in Inactive-floodplain CoverDeposits also includes other types of ice development,such as the formation of ice wedges and sheet ice. Theformation of ice wedges begins near the end of theactive-floodplain phase, and they eventually developinto large bodies of massive ice that form a continuous,low-density network of ice-wedge polygons in Inactive-floodplain cover Deposits. At this latter stage, icewedges occupy approximately 5% of the volume of thetop 2 m of permafrost. In addition, massive formationsof sheet ice (Shur and Jorgenson, 1995) occasionally arefound near the bottom of Inactive-floodplain CoverDeposits. The genesis of this ice is unknown but may berelated to refreezing of sediments in unfrozen drained-lakes or to freezing at the bottom of deep, water-filledpolygons.

The decrease in thaw depth, change to saturated con-ditions, and reduced sedimentation occurring duringthis phase all contribute to the accumulation of ice atthe top of the permafrost. The accumulation rate ofmaterials (mostly ice, with some organics and sedi-ments) is substantial (1.3 mm/yr). The total thickness ofInactive-floodplain Cover Deposits ranges from 0.7 mon new deposits to 3 m on older deposits. As a result ofice accumulation, the thickness of the intermediatelayer can become two or more times greater than thethickness of the active-floodplain cover deposit fromwhich the intermediate layer formed (Shur, 1988).Eventually, Inactive-floodplain Cover Deposits accumu-late sufficient ice and organic material, along withminor amounts of fluvial and eolian material, so thatthe ground surface rarely is flooded. At this stage, thedeposit can be considered an Abandoned-floodplainCover Deposit. The time required to reach this transi-tion is about 1500Ð2500 yr.

DELTA ABANDONED-FLOODPLAIN COVER DEPOSITS

Delta Abandoned-floodplain Cover Deposits, whichrepresent the oldest portions of the delta landscape(with the exception of a few isolated old alluvial ter-races), typically have deep (1Ð3 m) accumulations ofmassive and layered organics that frequently have beendeformed or turbated by the formation of large ice-wedges. Eolian material also is present occasionally.Due to the accumulation of organic material and shal-low thaw depths (mean = 0.36 m), the active layerbecomes almost entirely organic. Below the active layer,the massive organics usually have organic-matrix ice,while turbated, layered organics are dominated byorganic-matrix ice and minor amounts of lenticular, lay-ered, vein, reticulate, and ataxitic ice. The organic-richsediments have extremely high ice contents (80Ð90%).Accumulation rates (0.4 mm/yr) of surface materials(mostly organics, ice, and trace amounts of eolian siltand sand) decrease sharply, however. The continueddevelopment of massive ice wedges creates a network

M. Torre Jorgenson, et al. 527

of high-density, high-relief, low-centered polygons, inwhich the ice wedges occupy approximately 20% of thevolume of the top 2 m of permafrost. Because of theirregular topography resulting from polygon develop-ment, the surface typically supports a complex mosaicof wet sedge-willow and moist sedge-shrub meadows.At the latest stage of development, the centers of thepolygons are raised sufficiently by organic matter andice accumulation to become high-centered polygonssupporting moist sedge-shrub meadows. Based onradiocarbon dating of basal peat layers (e.g., 2160 ybpat 2.9 m, 2440 at 2.7 m, 2780 at 2.4 m, 3210 at 3.0 m, 3290at 2.7 m, 3960 at 3.1 m, and 4090 at 3.2 m from separateprofiles), we estimate that this phase occurs 2000Ð4000yr after the active-floodplain cover phase.

DELTA THAW BASIN DEPOSITS

By the time the floodplain has evolved to the aban-doned-floodplain stage, so much ice has accumulatedin the sediments that the deposits become susceptible tothermal degradation and collapse, as indicated by thehigh areal extent of thaw lakes on inactive and aban-doned floodplains. Some of these thaw lakes becometapped and drained by river channels (Walker, 1978).Due to breaching by channels and the resulting lowerelevation of the exposed lake bottom, sediment deposi-tion from floodwater again becomes frequent. At thispoint the whole process begins again. Due to stilling ofsediment-laden floodwater in the basins, sediments arecomprised of massive, clay-rich fines. The unusuallyhigh ice contents in young, ice-poor thaw basins wasattributed to the high clay contents. There are very fewlocations in the delta, however, where tapped lakeshave evolved all the way back to Inactive-floodplain

Cover Deposits. It appears that rates of channel migra-tion usually prevent completion of a thaw-lake cycle.Jorgenson et al. (1995) found that most of the delta isreworked by erosion and deposition over a period ofapproximately 2,000 years with only scattered terrainpatches persisting up to 4,000 years.

EXTERNAL FACTORS

Complicating this analysis of evolutionary trends arethe effects of sea level rise. Sea level has risen about 5 msince 5,000 yr ago, for an average rise of 1.0 mm/yr(Hopkins, 1982). This increase in sea level is evident insoil profiles: the surface elevations of new riverbeddeposits and of organic horizons in inactive floodplainsare considerably higher than the elevations at whichthese deposits are found in older soil profiles. This dif-ference in elevation of organic deposits on the deltaindicates that sea level was approximately 1.7 m lower2900 years ago (0.6 mm/yr). The current rise in sealevel (2.0 mm/yr) probably is increasing the frequencyof flooding on the higher floodplain steps, because therate of sea level rise is faster than the rate of materialaccumulation on abandoned and inactive floodplains(0.4-1.3 mm/yr).

Acknowledgments

This research was funded by ARCO Alaska, Inc.,Anchorage, AK and we appreciate the guidance provid-ed by John Eldred, Mike Joyce, and Stan Pavlos ofARCO. We appreciate the helpful review comments byS. Murphy, F. Wright, and F. M. Nixon.

The 7th International Permafrost Conference528

References

Carter, D.L. and Galloway, J.P. (1982). Terraces of the ColvilleRiver Delta region, Alaska. In Coonrad, W.L. (ed.), The USGeological Survey in Alaska: Accomplishments during1980. U.S. Geological Survey USGS Circular , 884, pp. 49-52.

Carter, D.L. and Galloway, J.P. (1985). Engineering-geologicmaps of northern Alaska, Harrison Bay Quadrangle. U.S.Geological Survey Open File Report, 85Ð256, 47 p.

Hopkins, D.M. (1982). Aspects of the paleogeography ofBeringia during the late Pleistocene. In Hopkins, D.M. etal., (eds.), Paleoecology of Beringia. Academic Press, NewYork, pp. 3-28.

Jorgenson, M.T., Aldrich, J.W., Pullman, E.R., Ray, S.R.,Shur, Y., Smith, M.D., Stickney, A.A., Walker, H.J. (1995).Geomorphology and hydrology of the Colville River Delta,Alaska, 1995. Unpublished rep. prepared for ARCO AlaskaInc., Anchorage, AK, by ABR, Inc., Fairbanks, AK. 136 pp.

Jorgenson, M.T., Roth, J.E., Pullman, E.R., Raynolds, M.,Stickney, A.A., Smith, M.D. and Zimmer, T. (1997). Anecological land survey for the Colville River Delta, Alaska, 1996.Unpubl. rep., for ARCO Alaska, Inc., Anchorage, AK, byABR, Inc., Fairbanks, AK.

Katasonov, E.M. (1969). Composition and cryogenic structure

of permafrost. In Permafrost investigations in the field.National Research Council of Canada, OT. Tech. Trans.1358, pp. 25-36.

Kreig, R.A. and Reger, R.D. (1982). Air-photo analysis andsummary of land-form soil properties along the route ofthe Trans-Alaska pipeline system. Alaska Division ofGeology and Geophysical Surveys, Geological Report, 66. 149 p.

Murton, J.B. and French, H.M. (1994). Cryostructures in per-mafrost, Tuktoyaktuk coastlands, western arctic Canada.Canadian Journal of Earth Science , 31, 737Ð747.

Pihlainen, J.A. and Johnston, G.H. (1963). Guide to a fielddescription of permafrost. Associate Committee on Soil andSnow Mechanics. National Research Council, OT. Tech.Memo. (79 pp).

Rawlinson, S.E. (1993). Surficial geology and morphology of theAlaskan Central Arctic Coastal Plain. Alaska Division ofGeology and Geophysical Survey, Fairbanks, AK. Rep. ofInvest. 93Ð1. (172 pp).

Reimnitz, E., Graves, S.M. and Barnes, P.W. (1985). BeaufortSea coastal erosion, shoreline evolution, and sediment flux. U.S.Geological Survey, Open-File Rep. 85Ð380. (66 pp).

M. Torre Jorgenson, et al. 529

Shur, Y.L. (1988). The upper horizon of permafrost soil. InSenneset, K. (ed.), Proceedings 5th International Conferenceon Permafrost. Tapir Publishers, Trondheim, Norway, pp.867-871.

Shur, Y. and Jorgenson, M.T. (1998). Cryostructure develop-ment on the floodplain of the Colville River Delta, north-ern Alaska. In Proceedings 7th International Conference OnPermafrost, Yellowknife, Canada. Nordicana (this volume).

Walker, H.J. (1976). Depositional environments in the ColvilleRiver Delta. In Moller, T.P. (ed.), Recent and AncientSedimentary Environments in Alaska. Alaska GeologicalSociety, Anchorage, pp. C1-C22.

Walker, H.J. (1978). Lake tapping in the Colville River Delta.In Proceedings 3rd International Conference on Permafrost.National Research Council of Canada, OT, Canada, pp.233-238.

Walker, H.J. (1983). Colville River Delta, Alaska, guidebook topermafrost and related features. Alaska Division ofGeology and Geophysical Survey, Fairbanks, AK.Guidebook 2 for Fourth International Conference on Permafrost.(34 pp).