LESSONS FROM THE BENCH GLACIER, CHUGACH RANGE, ALASKA: SLIDING SPEEDS AND EROSIONAL OUTPUT REFLECT SEASONAL EVOLUTION OF THE SUBGLACIAL HYDROLOGIC SYSTEM
ANDERSON, ROBERT S. UC Santa Cruz.
MacGregor, Kelly R. UC Santa Cruz.
Riihimaki, Catherine A. UC Santa Cruz.
Anderson, Suzanne P. UC Santa Cruz.
Waddington, Ed D. Univeristy of Washington.
Basal sliding paces erosion by both abrasion and quarrying. Any model of glacial modification of a landscape must therefore include at a minimum the temporal and spatial pattern of sliding. As part of a project designed to explore the long term evolution of alpine valley longitudinal profiles in the face of glacial cycles, we have studied the 7 by 1 km, 180 m thick Bench Glacier near Valdez in the Chugach Range of Alaska, for two melt seasons. One of our principal targets was the pattern of sliding velocity over the length of the glacier and its change through the summer melt season. We have documented the glacier shape, the meteorological forcing of the glacier, the hydrological balance, the surface velocity field, and the erosional output of both sediment and solutes.
Glacial thickness. The Bench Glacier was chosen for its simple geometry (it has no tributaries), its moderate size, and its accessibility. Fourteen cross-glacier radar profiles revealed a thickness profile with a maximum thickness of 180 m, and relatively symmetrical U-shaped cross sections. Present day basal shear stress in the ablation zone averages roughly 70000 Pa when adjusted appropriately with a shape factor that reflects the valley cross section.
Meteorolology. We established each summer a met station in the upper ablation zone measuring air temperature, incoming radiation, wind speed, precipitation and snow thickness at 15-minute intervals. In addition, we deployed shielded Hobo data loggers to record air temperature at the same intervals at 8 ablation stakes on the glacier. Summer rain and snow were minimal in these two years. The spatial pattern of mean air temperatures was essentially uniform along the glacier centerline, showing no simple lapse rate over the 600 m elevation gain from terminus to headwall. On-glacier temperatures were very different from those measured on the glacier foreland and on the adjoining ridge. Wind was consistently down-glacier, and was well established by mid-day. This katabatic wind presumably homogenizes the air temperatures. Snow thickness changes at rates dictated by both the daily and longer period storm cycles.
Balance. As the glacier has receded by several km from its Little Ice Age moraines, it has been in long term negative balance. Retreat rates have averaged 25 m/year since 1950. Snow probing upon arrival revealed the accumulation pattern. In both years the snow thickness increased two-fold from terminus to headwall, but was much thicker in 2000 than in 1999. Snowpits in the early melt season documented uniform profiles of snow densities of 480-520 kg/m^3, allowing conversion of snow thickness changes to melt water equivalent. Frequent measurement of snow depths at 12 to 14 ablation stakes along the centerline of the glacier revealed the pattern of snow melt through the melt seasons. Melt rates consistently increased toward the terminus despite the essentially uniform air temperatures and the expected uniformity of the incoming radiation. This implies that the evolution of the albedo field plays a central role in establishing the pattern of melt generation. As the snow is thinner on the terminus, low albedo ice crops out earlier on the terminus than higher on the glacier, promoting higher melt toward the terminus. The pattern of specific balance derived from combining accumulation and melt patterns is close to linear with elevation. High accumulation in 2000 translated into a positive net balance of roughly one million cubic m, while 1999 was highly negative; at –6 million cubic m, it was typical of the last 50 years. Measurement of both meltwater inputs and water outputs at the exit stream revealed a picture of glacial storage through the melt season. Each summer, this storage peaks in the early melt season, maximum storage corresponding to tens of cm of water thickness.
Glacial response over the melt season. We employed both optical and GPS methods to document the spatial pattern of surface motion throughout both melt seasons. Both summers yielded similar surface speed evolution. A steady background surface speed field of roughly 2-3 cm/day is interrupted only once per year by an up-glacier propagating wave of elevated velocities. This anomaly initiates at the terminus, propagates at a speed of 250 m/day, and involves roughly 1-km of the glacier at a time. Surface speeds reach up to an order of magnitude above background, reaching 30 cm/day; peak speeds increase up-glacier. A high resolution time series or speeds from GPS-derived 4-hour positions at a stake in the upper ablation zone reveals that the arrival of the horizontal speed anomaly at a site is associated with a vertical speed anomaly, indicating divergence of the motion from bed-parallel. Maximum divergence suggests bed separation of order 15 cm. Bed-parallel motion is re-achieved within 4 days after the abrupt slowing of the horizontal speed back to background values. We interpret this speed anomaly to reflect sliding at the glacier bed. The period over which the sliding wave propagates upglacier corresponds to the time of high storage of water within the glacier. We interpret the background speeds, both late summer averages and winter averages, as being related to internal deformation of the ice. Using the pattern of surface speed, ice thickness and surface slope, we conclude that the rheology of the relatively thin ice in the ablation zone is dominated by a linear (n=1) relation between stress and strain rate, with an ice viscosity of roughly 2 x1013 Pa-s. This corroborates recent results form a large scale experiment on the nearby Worthington Glacier, in which the rheology changes from n=1 to n=3 at a critical stress. Given the surface slopes in the ablation zone of the Bench Glacier, this corresponds to roughly 120 m depths.
Erosional output. Sediment. We have monitored both the water and sediment discharge from the glacier at a gage site only 500 m from the terminus. We have converted the 15- minute time series of turbidity to surface (suspended) sediment concentrations using a rating curve derived from simultaneously collected surface water samples. In 1999 we also attempted to document bedload transport using an acoustic method, taping one minute of hydrophone noise per hour. This was calibrated with measurements of bedload in which the entire flow of the stream was sent through a coarse mesh fence. In 2000 we performed bedload measurements using a Helley-Smith sampler. The relationship between sediment discharge and water discharge shows significant hysteresis. No simple rating curve can be established. The bedload in particular appears to become depleted over the summer. The proportion of bedload is many tens of percent of the total sediment yield. We interpret the hysteresis to reflect the exhaustion of the sediment source at the glacier bed. In the case of bedload, this implies that sediment has been effectively removed from an area of the bed corresponding to the footprint of the subglacial channel system. Integration of the sediment discharge through the melt season suggests a mean subglacial sediment evacuation rate of roughly 1 mm/yr.
Solutes. We also monitored the chemistry of the outlet waters. Samples were collected twice daily, allowing calibration of a 15-minute time series of conductivity. The conductivity (hence TDS) showed considerable variation over the melt season, with several periods of distinct behavior. Swings in TDS indicate mixing between a consistent low concentration end-member and an evolving high concentration end-member. This high end-member becomes progressively more dilute over the summer, suggesting that subglacial residence times decline. Breaks in the behavior correspond in many instances with major step changes in water discharges.
We are currently attempting to synthesize this information into a model of glacial motion and erosion over an annual cycle. The sliding pulse, the sediment discharge, and solutes reflect the evolution of the subglacial drainage system. We envision the following seasonal cycle. If sliding is inversely proportional to the effective pressure at the bed, the sliding pulse becomes a probe of the evolution of the subglacial pressure field. The pressure field increases as melt is released from the snowpack and makes its way toward the bed. In the early melt season, melt water inputs exceed outputs, reflecting the inefficiency of the subglacial drainage system; the channel system and the connections between cavities have collapsed over the winter. Sliding initiates at the terminus where melt is most prolific. Sliding terminates when the pressures decline, due to either 1) the generation of subglacial storage space by the opening of cavities by sliding, or 2) the establishment of more efficient drainage due to connections between cavities or upglacier insertion of a channel system. This sliding termination propagates upglacier as the water must be drained toward the low pressure point maintained at the exit. Water with long contact times with subglacial sediment, and the sediment itself, is exhausted from the subglacial system through the melt season, yielding strong hysteresis in the sediment – water discharge relationship, and strong seasonal trends in the high end-member solute concentrations.
In summary, this small alpine glacier achieves all of its annual sliding in a short pulse in the early melt season, the upglacier propagation of which must reflect the evolution of the subglacial drainage system. Even this small glacier is capable of eroding its bed at a high rate of about 1 mm/yr. The simultaneous collection of meteorological, glaciological, and sediment discharge data provides strong constraint on all the components of the system that are relevant to the construction of models of the generation of the long term glacial signature in alpine valleys.
Previous | Abstract Index | Next
Copyright © 2001 INSTAAR, Univ. of Colorado