SAMPLE COLLECTION

The Cottonwood Creek poplar grove contains 11 stands mapped from compass bearings and pacing; an uncorrected aerial photograph (approximately 1:1500 scale) was also used to adjust tree measurements to an aerial basis. After counting the total number of ramets in each stand, we randomly selected 2 trees from each of the 11 stands for measuring diameter at breast height (1.38 m) and total height (m). Although this sample is admittedly small (2% of all trees on the site), there was minimal variation in diameter (average difference = 3.6 cm, or 8%) and height (average difference = 1.3 m, or 5.6%) between the 2 trees within a stand, likely because they were ramets from the same parent. Increment cores were taken at 15 cm above the ground for the determination of radial increment and stand age.

Stembark conditions, including “cracked” (thin, vertical frost-induced cracks), scarred (broader abrasion from moose grazing and flooding), and undamaged, were determined from a random survey of at least 20 trees per stand. Seedlings (including root suckers) were tallied in 5 randomly located, 4-m² plots in each stand. A stem analysis (Husch et al., 1972) was performed on a single tree in August 1997. Tissue samples of poplar branches, current twigs, current buds, and foliage were collected from the upper, middle, and lower crown. Disks were collected every 1 m along the stem for a total distance of 12 m, and the proportions of bark, sapwood, and heartwood were determined.

Variation in soil properties and soil temperature was examined along a north-south transect perpendicular to the river and along an east-west transect in forested and nonforested areas parallel to the channel. The north-south transect extended 200 m south from the river at 20-m intervals, passing through two of the largest stands. A 40-cm-deep soil pit was excavated at each sampling point.

Thirteen soil pits in alternating poplar stands and shrubby areas were excavated along the 1000-m-long east-west transect parallel to the river. Soils were described along both transects using the Soil Survey Division Staff (1993) protocol. Samples were taken from each horizon >2 cm in thickness. Soil temperature was measured at a depth of 40 cm with a digital thermometer.

Samples of bulk precipitation, poplar throughfall, and stream water were collected during a heavy precipitation event on 18 August 1998. An additional sample of stream water was collected on 14 August 1997.

LABORATORY ANALYSIS

Soil samples were returned to the University of Wisconsin–Madison for the determination of pH and gravimetric moisture content (dried at 70°C for 48 hr). Oven-dried soils were passed through a 2-mm sieve. Ninety-three soil samples were sent to the University of Missouri Soil Characterization Laboratory, where analyses were performed using methods established by the Soil Survey Staff (1996). These analyses include easily oxidizable organic C by dichromate digestion (6A1a); total N by semimicro Kjeldahl (6B2a); extractable P by the Olsen method (6S3); cation-exchange capacity by saturation with ammonium at pH 7 (5A8); exchangeable acidity (6Ha, 6H5); extractable Al (6G1); and particle-size distribution with sand and silt fractionation by sieving and pipette, respectively (3A, 3A1). Base cations were extracted with ammonium acetate (pH 8.2).

The University of Wisconsin Soil and Plant Analysis Laboratory (SPAL) ( http://riprock.soils.wisc.edu) analyzed stream water, precipitation, and throughfall solutions for cations with an inductively coupled plasma optical-emission spectrometer (ICP-OES) and anions by ion chromatography. Bicarbonate was determined using the Gran titration method (Gran, 1952). Tissue samples were analyzed by SPAL for total nutrients (ICP-OES) and total Kjeldahl nitrogen following grinding in a laboratory mill.

Annual radial growth of poplar trees was determined from increment cores using a binocular zoom microscope (7–30×). Because Armillaria mellea (butt rot) and Fomes igniarius (heart rot) stem decay resulted in incomplete cores, the total age of 21 of the 22 sampled trees was estimated from the outermost 10 years of radial growth projected to the pith of the tree. Stand basal area was estimated stem density and basal area of the two sample trees. Finally, a wood fragment collected from a soil pit in stand 10 at the 52-cm depth was split; one subsample was sent to the U.S. Forest Products Laboratory for identification and another to the Lawrence Livermore Laboratory for radiocarbon dating.

STATISTICAL ANALYSES

Six of the stands were located within 100 m of the river and 5 of the stands more than 100 m away from the river. Differences in mean values between stands in proximity to the river were compared using an unpaired t-test, assuming equal variances.

STAND CHARACTERISTICS

There were no significant differences in tree height, age, number of seedlings, the incidence of scarred and cracked trees, and basal area with distance from the river (Table 1). However, diameter was significantly greater (p = 0.01) and density was significantly lower (p = 0.006) within 100 m of the river.

Average radial increment was 1.1 mm yr⁻¹ and height increment was 7.6 cm yr⁻¹ (data not shown). Based on dendrochronological studies, a single tree experienced the greatest growth rates in 1940, 1970, 1959, 1885, 1887, and 1974, while poor growth occurred in 1875, 1889, and 1980 (data not shown). The radiocarbon date of the wood fragment found at a soil depth of 52 cm in stand 10 was 650 ± 50 BP. This date calibrates to between 550 and 670 cal yr BP at 2-σ (Stuiver et al., 2000).

TISSUE ANALYSES

Calcium, nitrogen, and potassium are the most abundant macronutrients in tissues of the sampled balsam poplar, followed by Mg, P, and S (Table 2). The micronutrients are arrayed Zn > Na > Fe > Mn > B > Al > Cu, except in sapwood, where Na was abundant, followed by Al, Fe, Zn, Mn, B, and Cu. In perennial tissues, such as heartwood, sapwood, bole bark, live branches, and current twigs/buds, the mobile elements N, K, and P increased toward the apex of the tree. In contrast, the immobile element Ca was more concentrated toward the base of the tree or live crown.

SOIL PROPERTIES

Soils of the study area are poorly developed and contain a sequence of alluvium and buried organic horizons (Table 3). Lithologic discontinuities, consisting of overbank silts over river cobbles, were found at depths ranging from 40 to 43 cm in 5 of the 13 pits along the east-west transect and in 7 of the 11 pits along the north-south transect.

There were no trends in soil morphological or chemical properties with distance from the river or in relation to vegetation type (i.e., balsam poplar stand vs. shrubs) along the river. Silt concentration of the surface mineral horizon was lower away from the river, but the differences were not statistically significant (Table 1). In general, soils within 100 m of the river were silt loams with an average of 76% silt, 100% base saturation primarily by Ca, and a pH averaging 7.7.

SOLUTION ANALYSIS

The distribution of cations in precipitation, on a charge basis in descending order, was Ca > Al > Na > Mg > K > Mn; anions were distributed: NO₃ > Cl > SO₄ > HCO₃ (Table 4). Throughfall cations can be ranked K > Ca > Mg > Al > Na > Mn, and anions HCO₃ > Cl > NO₃ > SO₄. Streamwater cations can be ranked Ca > Mg > Na > Al > K > Mn and anions SO₄ > HCO₃ > NO₃ > Cl. All solutions were weakly acidic to neutral, with a pH ranging from 6.3 to 7.7.

GROWTH AND ORIGIN OF THE COTTONWOOD CREEK BALSAM POPLAR STAND

The diameter and height growth rates of balsam poplar at Cottonwood Creek are comparable to those reported elsewhere in arctic Alaska and in areas immediately south of the Brooks Range. Radial increment ranged from 0.5 to 2.2 mm yr⁻¹ (data not shown), which is comparable to the 0.1 to 2.2 mm yr⁻¹ values reported by Edwards and Dunwiddie (1985) and Lev (1987).

Morphological features of the stands were examined using the criteria of Gom and Rood (1999). The 11 stands investigated at Cottonwood Creek likely represent individual clones in that they were separated by distances ranging from 50 to 250 m, vary considerably in age, and contain distinct differences in phenological properties such as stem form, branching habit, bark color and texture, and leaf size and shape. The stands are of seed origin; the seeds may have been carried over the Brooks Range by wind and deposited throughout the southern Arctic Foothills. Because of the unique thermal and moisture conditions along the rivers draining the northern Brooks Range, balsam poplar becomes established only in specific localities. Alternatively, the seeds may be carried by water birds favoring riparian areas. The clones have persisted for at least 90 to 200 yr and possibly as long as 670 yr, based on a radiocarbon-dated wood fragment found at a depth of 52 cm in one of the stands.

There were no obvious differences in soil properties along the river or with distance from the river other than silt content, thus providing no distinct criteria for site selection for individual stands. However, several factors appear to promote the establishment and growth of poplar groves: the soil thermal, moisture, and nutrient regimes all appear to be more favorable near the river. Soil temperatures taken at the 40-cm depth along both the north-south and east-west transects in mid-August 1998 were 10–12°C, in contrast with soil temperatures of 0–0.5°C reported at similar depths and time frame in arctic tundra (Nelson et al., 1997). The continuous subsurface flow of the river, even during the harsh winter months, depresses the permafrost table near and beneath stream channels (Corbin and Benson, 1983). A depression in the permafrost table results in higher near-surface soil temperatures; less heat energy is needed to thaw frozen ground and is, therefore, available to increase soil enthalpy.

A second factor that may influence the distribution of balsam poplar along Cottonwood Creek is the effect of lithologic discontinuities on soil moisture relations. Lithologic discontinuities strongly affect water retention and movement in soil parent materials (Kung, 1993). In the alluvial soils along Cottonwood Creek, silt-rich materials over cobbly loamy sand materials enhance water retention and availability for poplar roots by suspending water in the unsaturated zone. The widespread occurrence of lithologic discontinuities along transects, and their absence elsewhere on the arctic tundra, suggest that they provide a unique environment for rooting of balsam poplar. The effective rooting depth of balsam poplar ranged from 13 to 42 cm and averaged 29 cm.

A third factor investigated relative to the distribution of balsam poplar along Cottonwood Creek is the nutrient regime. Natural waters (bulk precipitation, throughfall, and streamwater), soils, and tree tissues at the Cottonwood Creek site are particularly enriched in calcium. Dissolved calcium in river water likely originates from the groundwater-fed stream system originating in the Brooks Range, which is supersaturated with respect to calcite (Table 4). The soils may be enriched in calcium from weathering of the calcareous alluvium or more likely from dry deposition of nearby aufeis deposits, which is discussed in the next section. The foliage, branch, and heartwood tissues of North Slope balsam poplar (Table 2) all contain as much as two times the concentration of calcium as Populus species outside Alaska (Bartos and Johnson, 1978; Pastor and Bockheim, 1984).

Although the tissues are enriched in calcium, it could be argued that the high level of calcium may not be essential for poplar growth. For example, crop and pasture plants grown in calcium-rich solutions showed little additional gain in biomass as calcium concentrations increased (Loneragan and Snowball, 1969). Moreover, calcareous parent materials occur elsewhere on the North Slope in the absence of balsam poplar (Bockheim et al., 1998). Therefore, Ca enrichment adjacent to the river may be a cofactor influencing the occurrence and growth of balsam poplar beyond its normal range.

SUBSURFACE WATER FLOW

Springs emerging from the Brooks Range provide a continual source of water at temperatures around 0°C that inhibits permafrost beneath rivers and their adjacent floodplains (Corbin and Benson, 1983). The surface temperature of North Slope streams may be lower than the temperature of the surrounding ground because of surface reflectance and the high specific heat of water. The heat pulse penetrates deeper under the stream than in the adjacent floodplain due to the channel topography and higher thermal conductivity of saturated channel sediments. The net effect is to lower the permafrost table to depths of 6 m beneath the stream and about 4 m on the adjacent floodplain (Corbin and Benson, 1983). These findings have implications for the thermal regime of Cottonwood Creek.

AUFEIS DEVELOPMENT

Aufeis or overflow deposits that form during the winter may cause a depression in the permafrost table. As the freezing front penetrates downward into the thaw bulb, aufeis (icings, naleds) can develop. Subsurface water flowing in the coarse alluvium beneath the stream channel is under increasing hydrostatic pressure due to the 9% volumetric expansion associated with the phase change (Hall, 1980a). During freezing of the water, ions are excluded from the ice crystal lattice, and at high concentrations they may further depress the freezing point. Eventually, the frozen overburden will deform upward and crack, allowing the water to rupture to the surface, where it quickly freezes. This process is repeated many times during the winter, and over time the aufeis aggrades upward. Ice at the top of the deposit is enriched in ions relative to the lower layers. The icings may reach sufficient thickness to prevent complete thermomechanical ablation in summer.

Aufeis remnants persisting well into the summer or year-round may contribute to river braiding by obstructing floodwater and causing channels to widen or new channels to form (Harden et al., 1977). Aufeis characteristically forms at the mountain front, where stream gradients decrease as rivers leave constricting valleys, pouring onto extensive outwash plains (Carey, 1973).

Corbin and Benson (1983) related soil temperature and stream water temperature along the North Slope to the latent heat of fusion of icing deposits during the winter and penetration of the summer heat pulse. Although icings enhance heat loss by removal of the insulating snow cover, their net thermal effect is the addition of heat to the system through the release of latent heat during freezing. The heat available from freezing a 20-cm-thick aufeis overflow is more than twice the amount required to bring thick accumulations of aufeis to their melting point (Corbin and Benson, 1983). The percolation of water downward along ice fractures advects heat more efficiently than conduction, which is the only other mechanism for transporting heat to the soil surface.

INFLUENCE OF FAULTS

A system of faults across the Brooks Range–Arctic Foothills contact contributes to the groundwater flow network and may indirectly influence the distribution of balsam poplar (Fig. 2). Whereas faults running perpendicular to the generally north-flowing rivers have limited aufeis development and are not associated with known poplar groves, faults parallel to the rivers have aufeis and according to our map are often accompanied by balsam poplar.

Spring water emerging from the north flank of the Brooks Range carries a large dissolved load of calcium carbonate and calcium sulfate (Childers et al., 1977; Hall, 1980b; Himes, 1980). The bedrock of the Brooks Range is composed of a complex sequence of calcareous limestone, sandstone, and shale (Sloan, 1987; Moore et al., 1994). The Lisburne Limestone, Endicott Sandstone, and Sadlerochit Sandy Limestone aquifers direct groundwater from recharge areas high in the Brooks Range along faults and solution cavities to discharge areas in the Arctic Foothills and Coastal Plain (Carey, 1973; Childers et al., 1977; Hall, 1980b; Sloan, 1987; Moore et al., 1994). Freezing of streamwater during the winter forces precipitation of this dissolved load, providing a continually renewed source of base cations to soils along reaches with aufeis.

The impact of faults on the groundwater flow system is not completely understood, but the transport time and distance traveled from the recharge area appear to be important (Carey, 1973; Childers et al., 1977). The groundwater residence time may increase due to impeded flow through pervasive permafrost, enhancing dissolution of the calcareous bedrock (Childers et al., 1975; Himes, 1980; Sloan, 1987). The presence of a thaw bulb adjacent to the river increases drainage of water through the soil profile, reduces moisture, increases soil temperatures, and provides a deeper rooting zone than in areas away from the thaw bulb (Carey, 1973; Corbin and Benson, 1983).

The estimated annual precipitation of 150 mm and an average of 25 mm during the summer months (June–August) would appear to be insufficient to support poplar growth, particularly considering that the seasonal thaw layer as determined from this study is more than 2 m thick. However, lithologic discontinuities in stratified fluvial sediments may suspend water in the unsaturated zone, which plants can readily use.

QUATERNARY HISTORY OF POPLARS ON THE NORTH SLOPE

It is not known when balsam poplar first appeared on the North Slope. The poplar grove examined at Cottonwood Creek contained clones ranging from 90 to 200 yr in age; a wood fragment from 52 cm below the modern soil surface was radiocarbon dated at 650 ± 50 BP. These data suggest that balsam poplar may have been on the North Slope for at least the past 650 yr. However, poplars may have existed at specific locations on the North Slope for thousands of years. Buried poplar logs along the north-flowing Sagavanirktok River and the Nigu River in the Brooks Range and in the floodplain of the Putuligayuk River near Prudhoe Bay were radiocarbon dated at 8400 ± 300 yr BP (Detterman, 1970), 11,100 BP (Hopkins et al., 1981), and 35,600 ± 550 BP (Hopkins et al., 1981), respectively, suggesting that poplar may even have been present during the mid-Wisconsin.

Therefore, it is likely that poplars on the North Slope of Alaska and in northwest Canada have occurred as groves or clumps on suitable sites throughout the Holocene (Mott, 1978; Hopkins et al., 1981). The natural pioneering characteristics of Populus and its ability to withstand severe climatic conditions make it adaptable to the harsh climatic conditions of the region (Ritchie and Hare, 1971). Whether or not the modern stands are relicts cannot be determined from this study. However, warm summers in Beringia during the early Holocene may have permitted poplar expansion (Hopkins et al., 1981). These warmer climatic conditions were associated with greater amounts of exposed land from marine regressions, creating a more continental climate.

The warming of arctic Alaska during the past three decades has resulted in increased shrub abundance (Sturm et al., 2001). This increase in shrub cover will alter the partitioning of energy in the summer, trapping of snow cover in the winter, and ecosystem carbon distribution. A major question is whether warming of arctic Alaska has affected the growth and distribution of balsam poplar.