Darrel A Swift

Department of Geographical and Earth Sciences, University of Glasgow, Glasgow G12 8QQ, UK

Picture courtesy of Peter W. Nienow

At Haut Glacier d'Arolla, Switzerland (Fig. 3.1), suspended sediment transport during the 1998 melt season demonstrates the importance of subglacial drainage system morphology for basal sediment evacuation because it influences both the capacity of meltwater to transport basal sediment and the mechanisms by which sediment is accessed and entrained.

Early in the melt season, surface runoff enters a distributed subglacial drainage system (Nienow et al., 1998) during which extreme increases in runoff stimulate periods of rapid glacier motion (termed 'spring events'; Mair et al., 2003). Later in the season, removal of the surface snowpack from the ablation area results in increasingly peaked diurnal runoff cycles that promote the up-glacier extension of a hydraulically efficient network of subglacial channels (Nienow et al., 1998). During 1998, two spring events occurred during steep rises in catchment discharge (subperiods 2 and 4, Fig. 3.2a & b; Mair et al., 2003), the first coinciding with intense rainfall and the second with both widespread thinning of the snowpack and a rapid increase in the efficiency of meltwater routing to the glacier terminus (Swift et al., 2005a). Dye tracer investigation demonstrated predominantly channelized subglacial drainage beneath the ablation area by late July, indicating rapid up-glacier extension of the channel network during subperiods 4 to 6 (Swift et al., 2005a; cf. Fig. 3.2a).

Suspended sediment transport during 1998 was monitored in the proglacial stream draining the western subglacial catchment (Fig. 3.1), into which extraglacial sediment contributions were negligible. Hourly mean suspended sediment concentration (SSC) was obtained from a continuous record of proglacial stream turbidity calibrated using 1159 point-collected water samples (Swift et al., 2005b). Catchment suspended sediment load

France j J^many j Switzerland ' T-f-\ Italy

Grande Dixenc Intake

France j J^many

Bouquetins ' 3838

^^ Permanent ice and snow ^ Bare rock and scree ^ Moraine and proglacial sediments H Supraglacial moraine

Mont Brulé 3585

Mont Collonii>S?/:

3637X^5"

Bouquetins ' 3838

^^ Permanent ice and snow ^ Bare rock and scree ^ Moraine and proglacial sediments H Supraglacial moraine

Mont Brulé 3585

1 km

Figure 3.1 Haut Glacier d'Arolla, showing the supraglacial divide between the eastern and western subglacial catchments during 1998 (dotted line) and velocity stakes used in Fig. 3.2b (crosses). Contours (dashed lines) and elevations are in metres.

Figure 3.2 Catchment discharge, glacier velocity and sediment transport at Haut Glacier d'Arolla during 1998. (a) Hourly mean catchment discharge (Q) and subperiods of the melt season (numbered) used in Fig. 3.3; (b) glacier velocity at stakes 102-701 (Fig. 3.1; Mair et al., 2003); (c) daily catchment suspended sediment load (SSL); (d) residual SSL from a log-linear relationship between SSL and Q; (e) residuals from a log-linear relationship between 'diurnal' SSL (see text) and Q diurnal amplitude.

Day during 1998

Figure 3.2 Catchment discharge, glacier velocity and sediment transport at Haut Glacier d'Arolla during 1998. (a) Hourly mean catchment discharge (Q) and subperiods of the melt season (numbered) used in Fig. 3.3; (b) glacier velocity at stakes 102-701 (Fig. 3.1; Mair et al., 2003); (c) daily catchment suspended sediment load (SSL); (d) residual SSL from a log-linear relationship between SSL and Q; (e) residuals from a log-linear relationship between 'diurnal' SSL (see text) and Q diurnal amplitude.

(SSL; Fig. 3.2c) was calculated from SSC and hourly mean catchment discharge (Q) measured at the Grande Dixence S.A. intake structure (Fig. 3.1). Following log-transformation, SSL was highly correlated with Q (r2 = 0.93); however, the residuals from this relationship (Fig. 3.2d) demonstrated subseasonal changes in the rate of sediment evacuation compared with discharge. Notably, the efficiency of sediment evacuation appeared to be lowest during subperiod 5 (Fig. 3.2d) when growth of the subglacial channel network was most rapid (Swift et al., 2005a) and reached a maximum during subperiod 8 (Fig. 3.2d) when up-glacier extension of the network had largely ceased.

Relationships between SSC and Q (Fig. 3.3) for shorter periods of the melt season (cf. Fig. 3.2a) demonstrate that changes in the efficiency of basal sediment evacuation were controlled by both subglacial drainage system morphology and sediment availability (Swift et al., 2005b). The gradient of the SSC versus Q graph (Fig.

3.3) is determined by the relationship between the sediment-transporting capacity of the flow, which scales linearly with flow velocity, and discharge. During subperiods 1-4, SSC ~ Q~L3 and therefore SSL ~ Q~23 under predominantly distributed subglacial drainage conditions. However, during subperiods 5-8, SSC ~ Q~2'2 and therefore SSL ~ Q~3'2 under flow predominantly through hydraulically efficient channels. Flow velocity is therefore inferred to have increased more rapidly with discharge during subperiods 5-8, probably as a consequence of rapid discharge variation within subglacial channels under increasingly diurnally peaked surface runoff cycles (Swift et al., 2005a,b). Relationship intercepts (Fig. 3.3) demonstrate changes in sediment availability under both distributed and channelized conditions, most notably a relative increase in availability between subperiods 5 and 8.

The limited availability of basal sediment during subperiod 5 (Figs 3.2d & 3.3) suggests that channelization confined meltwa-

Figure 3.3 Relationships between SSC and Q plotted over the range of discharge observed for individual subperiods of the melt season (cf. Fig. 3.2); subperiod 7 has been excluded due to potentially high extraglacial sediment contributions during heavy rainfall. The most efficient evacuation of basal sediment occurred at flows >4m3s-1 during subperiods 6 and 8.

1 10

m3s"1

Figure 3.3 Relationships between SSC and Q plotted over the range of discharge observed for individual subperiods of the melt season (cf. Fig. 3.2); subperiod 7 has been excluded due to potentially high extraglacial sediment contributions during heavy rainfall. The most efficient evacuation of basal sediment occurred at flows >4m3s-1 during subperiods 6 and 8.

ter to areas of the bed from which sediment was rapidly exhausted. Thereafter, increasing sediment availability (Fig. 3.3), coupled with a strong increase in flow velocity with discharge, appears to have resulted in highly efficient sediment evacuation during the peak of the melt season (Fig. 3.2d). Strong increases within subglacial channels imply that water pressures also increased rapidly with discharge (Swift et al., 2005b), suggesting that increasingly strong diurnal discharge variation may have increased sediment availability by encouraging local ice-bed separation, leading to extrachannel flow excursions, and/or a strong diurnally reversing hydraulic gradient between channels and the surrounding distributed system (Hubbard et al., 1995). The importance of water pressure variation is supported by the absence of significant trends in the residuals from a relationship between 'diurnal' SSL (i.e. SSL calculated between diurnal discharge minima) and Q diurnal amplitude (Fig. 3.2e; Swift et al., 2005b).

Whereas previous studies have generally emphasized declining sediment availability, these results demonstrate highly efficient sediment evacuation under channelized subglacial drainage conditions. Efficient flushing of basal sediment is critical in order to sustain the direct ice-bed interaction that is necessary for erosion (Alley et al., 2003), and strong diurnal water pressure variation within subglacial channels may locally enhance glacier sliding. As a result, seasonal establishment of channelized drainage beneath temperate glaciers or ice caps has the potential to considerably elevate erosional capacity. Importantly, annual and glacier-to-glacier changes in the pattern and timing of subglacial drainage system evolution are likely to contribute significantly to variability in glacial sediment yield. Meaningful relationships between glacial sediment yield and surrogate indicators of erosional capacity are therefore unlikely to be found without explicit consideration of the hydraulics of subglacial drainage.

Was this article helpful?

0 0

Post a comment