Well-developed surface channels on Taylor Glacier, McMurdo Dry Valleys, Antarctica, begin as medial moraines incised as shallow, narrow surface depressions, and retain this geometry for tens of km. Over a distance of 1100 m, the channel geometry dramatically changes, reaching depths >20m and widths >100 m. After rapidly enlarging, the channels appear to evolve toward a new equilibrium geometry. Compared to the glacier surface, the air temperature in the channels is warmer by ∼1.7˚C, wind speed is reduced by ∼2.4ms–1 and net shortwave radiation is greater by ∼14Wm–2. The microclimate in the channel shifts the energy balance towards enhanced melt. Field evidence and energy-balance modeling indicate ablation in the deep channels is ∼4.5 times greater than the local horizontal glacier surface and that melt accounts for ∼99% of the summer ablation, compared to ∼75% on the adjacent horizontal glacier surface. Melt in these channels supplies 65% of the unaccounted water discharge into the neighboring lake. In large part, the channels generate the water they carry, rather than merely route water generated elsewhere.

Sufficiently deep water-filled fractures can penetrate even cold ice-sheet ice, but glaciogenic stresses are typically smaller than needed to propagate water-filled fractures that are less than a few tens of meters deep, as shown by our simplified analytical treatment based on analogous models of magmatic processes. However, water-filled fractures are inferred to reach the bed of Greenland through >1 km of ice and then collapse to form moulins, which are observed. Supraglacial lakes appear especially important among possible crack ‘nucleation’ mechanisms, because lakes can warm ice, supply water, and increase the pressure driving water flow and ice cracking.

Variations in the chemical composition of subsurface runoff within Canada Glacier, Antarctica, are used to identify the main source waters, which are vertical surfaces, and melt from solar-heated debris within channels, cryoconite holes and pools. The main flow paths are subsurface connections between cryoconite holes, pools and riffles. The latter may become partially disconnected during hard freeze. The chemical composition of runoff at the outlet of Canada Glacier during January 2000 was dominated by Ca2+, HCO3– and sea salt (Na+ and Cl–), and became depleted in sea-salt and non-sea-salt (*) SO42– as the subsurface drainage system in a frozen pool-and-riffle system was flushed and the melting ice surface became depleted of overwinter dry deposited salts. Only during 2 days of hard freeze did sea salt and *SO42– increase in concentration together. Otherwise, sea salt and *SO42– declined while *Ca2+ and HCO3– increased. The latter ions are derived from the chemical weathering of sediment in frozen-topped pools, channels and cryoconite holes. It is inferred that the hydrochemical processes which occur in the vestigial, subsurface drainage system are the elution of ions from ice melt, dilution of these ions downstream by ice melt from vertical surfaces and the dissolution of dust, in subsurface pools, channels and/or cryoconite holes.

Digital elevation models of the bed and surface of Unteraargletscher, Switzerland, are used to reconstruct the theoretical pattern of basal water drainage for the years 1927, 1947, 1961 and 1997, during which period the glacier was thinning and receding. The theoretical drainage pattern for 1997 compares well, in a broad sense, with the locations of active moulins and the hydraulic connection status of boreholes drilled to the glacier bed. Changes in the basal water-flow pattern over the period 1927–97 that are revealed by the theoretical reconstructions of the subglacial drainage system structure are likely to have resulted from changes in glacier geometry. Concurrent with the retreat and thinning of the glacier, the height of medial moraines increased, probably due to the insulating effect of the debris cover reducing the melt of the underlying ice. This increase of moraine heights has led to the formation of hydraulic barriers at the glacier bed such that water flow has become channelized beneath the ice along drainage axes that parallel the course of the medial moraines on the glacier surface.

To test models of the hydraulics and geometry of englacial conduits, 48 holes (3900m of ice) were drilled into Storglaciaren, Sweden, in search of conduits. About 79% of the holes intersected a hydraulically connected englacial feature. A video camera was used to examine the features and measure local water-flow rates. Because of the extremely clear ice that surrounded most features, their geometry could not be discerned. Of the remaining features, 80% (36) were fracture-like, 16%(6) were of complex geometry, and 4%(2) exhibited a conduit-like geometry. The fracture-like features exhibited steep plunges (∼70˚), narrow openings (∼40mm) and slow water-flow speeds (∼10mms–1). We argue that these fracture-like features are indeed englacial fractures of unknown origin. The depth to fractures intersection varied from near the glacier surface to 96% of local ice depth, with a maximum depth of 131 m. Few hydraulically connected fractures exhibited water motion, indicating some preferential flow pathways exist. We found one ‘traditional’ englacial conduit after an intentional search in a field of moulins. These results suggest that englacial water flow is conveyed through a ubiquitous network of fractures and that conduits are relatively rare.

Observations from along the length of Bench Glacier, Alaska, USA, show that the subglacial water-pressure field undergoes a multiphase transition from a winter mode to a summer mode. Data were collected at the glacier surface, the outlet stream, and in a network of 47 boreholes spanning the length of the 7 km long glacier. The winter pressure field was near overburden, with low-magnitude (centimeter to meter scale) and long-period (days to weeks) variations. During a spring speed-up event, boreholes showed synchronous variations and a slight pressure drop from prior winter values. Diurnal pressure variations followed the speed-up, with their onset associated with a glacier-wide pressure drop and flood at the terminus stream. Diurnal variations with swings of up to 80% of overburden pressure were typical of mid-summer. Several characteristics of our observations contradict common conceptions about the seasonal development of the subglacial drainage system and the linkages between subglacial hydrology and basal sliding: (1) increased water pressure did not accompany high sliding rates; (2) the drainage system showed activity characteristic of the spring season long before abundant water was available on the glacier surface; (3) the onset of both spring activity and diurnal variations of the drainage system did not show a spatial progression along the length of the glacier.

We have conducted short-term pump experiments with pump pressures exceeding ice overburden to study the seasonality of the subglacial hydraulic system of Engabreen, Norway. Data were collected from load cells installed flush with the ice–bedrock interface and pressure transducers installed in boreholes leading from bedrock tunnels underneath the glacier to the ice–bedrock interface. The water-pressure recordings, seen in relation with the load-cell record, show the existence of hydraulically connected vs unconnected bed areas. Monitored boreholes have been used to inject water at high pressures. Each experiment led to the growth of a high-pressure water cavity whose spatial extent could be inferred from load-cell and pressure transducer records. Post-pump pressures were low after summer pump tests and close to ice-overburden level after winter pump experiments. We conclude that drainage takes place in a fast-flow, low-pressure, channel-based drainage system during summer, and a low-flow, high-pressure, linked-cavity drainage system during winter.

The base of the ice in the transition zone between an ice stream and an ice shelf is likely to be well lubricated by broadly distributed water, a condition which should permit fast sliding motion. It has been observed that motion takes place not smoothly but by localized stick–slip events that propagate in the downstream direction towards the ice shelf and at velocities approximately that expected for shear wave velocity of the basal till. Thus slip packets of gliding edge dislocations are likely to move at the base. I show here that subsonic dislocations should move upstream, rather than downstream, if frictional resistance is determined by normal traction stress change at the base. Transonic dislocations are expected to move in the downstream direction. However, if frictional resistance is lowered by hydrostatic pressure reduction at the base, the subsonic dislocation should move downstream.

We analyzed the ice flow of McCall Glacier, Alaska, USA, by numerical glacier modeling and radio-echo sounding (RES). Model experiments were carried out with a higher-order numerical ice-flow model, and results were validated with measurements of annual ice velocities and compared with previous estimates of ice-flow dynamics. During the 2003 summer campaign, detailed RES measurements were carried out along the central flowline of the ablation area with a 5 MHz (central frequency) ice-penetrating radar, where 10m ice temperatures are approximately –7.5˚C. The bed reflection power (BRP) beneath this central flowline abruptly increases at one location area, followed by a slow decrease down-glacier. The model experiments show that basal sliding (<50%) is necessary to match the observed annual mean surface velocities in the area that is characterized by high BRP values. However, when thermomechanical effects are taken into account, a temperate basal ice layer is apparent in the ablation area, which locally softens the ice and can explain to a certain extent the anomalous flow field. The model results confirm that the present temperature field is a remnant of a larger glacier geometry that was near steady state before the onset of enhanced surface thinning in the 1970s.

We investigate the relationships between meteorological, hydrological and glaciological data collected at Haig Glacier, Alberta, Canada, for the 2002 and 2003 ablation seasons. Correlation, lag cross-correlation and multivariate regression analyses are employed to assess the seasonal evolution of relationships between temperature, temperature residuals, total daily radiation, albedo, accumulation-area ratio (AAR) and total daily discharge (Qi). Early-season melt is temperature-dependent, when AAR remains high and the hydraulic properties of the snowpack limit both diurnal discharge variability and a rapid hydrologic response. As the melt season progresses, a decreasing AAR and ripening of the snowpack induce a glacier-wide decrease in albedo, and a structured radiation–discharge response is observed. Radiation-detrended temperature values offer modest improvements over physical temperature values in multivariate regression models estimating daily discharge values. Using a detrended-temperature indexed melt model, we assess the transport efficiency of the glacial hydrologic system through a comparison of total modelled daily melt and observed discharge. Transport efficiency values support the notion of a purge effect during freezing events and at the end of the ablation season, and suggest that it is the evolution of the supraglacial drainage system that controls diurnal discharge variability.

Many glaciers are subject to melting due to high summer air temperatures. Here, the presence of meltwater in the subsurface layers of the glacier bulk, and its subsequent percolation and refreezing are implemented in the calibration of a paleothermometer. Accounting for the melt feature index and the measured oxygen-isotope ratio allows for calibration of the paleothermometer and comparison of different climatic proxies. The results of reconstructions agree with previous reconstructions at the depth of attenuation of the seasonal climate signals, which supports the validity of the paleothermometer calibration. The sensitivity of the reconstruction to variations of the model parameters was also studied. It was found that most likely snow–firn sequence and temperature fields were subjected to significant change due to current warming. Temperature changes in the snow–firn thickness of Akademii Nauk (Severnaya Zemlya, Russian High Arctic) and Austfonna (Svalbard) ice caps exceed by ∼6˚C the average Arctic temperature anomalies for the last 150 years. The reconstruction of the past surface temperatures and the parameters of the subsurface heat source due to refreezing of meltwater lead to the conclusion that meltwater spreads inside two to four annual layers for Akademii Nauk and Austfonna ice caps, respectively.

Periods of storage and release of surface runoff that is routed englacially and subglacially are identified from a time series of cumulative water balance between 31 July and 11 September 1999 at Findelengletscher, Switzerland. The influence of subglacial hydrology on water routing within the glacier, and therefore on trends of water storage and release, is determined through comparisons of phase relationships between daily maxima, daily minima and diurnal ranges of borehole water levels, supraglacial runoff and proglacial discharge. Variations of water levels in 21 boreholes in the ablation zone suggest that although subglacial drainage is spatially dynamic, hydrologically efficient tunnel–conduit-style drainage dominates diurnal cycles of water transfer through the glacier. Over longer periods, however, storage and release of subglacially routed water is greatly influenced by the coexistence of, and temporary interconnections between, hydrologically inefficient distributed drainage and the tunnel–conduit network. Water levels in three boreholes indicate that after water storage increases in distributed drainage, hydraulic gradients between the different drainage systems may increase sufficiently to cause connections that initiate release of water when: (1) low maximum daily surface runoff causes low water pressures in the tunnel–conduit system; or (2) reorganization within distributed drainage causes spatially localized increases in water pressure.

We model runoff from glaciers in the McMurdo Dry Valleys, Antarctica, with summer (December–January) average air temperatures from 1990 to 2002 for the purpose of estimating decades- to millennial-scale glacial runoff into Lakes Fryxell, Hoare and Bonney. The relationship between summer temperatures and melt is found to be exponential near the melting temperature. We propose a variety of simple models that are calibrated using measured discharge from a number of streams draining from ten glaciers in Taylor Valley. The surface melting rate is constrained by mass-balance measurements from four of the glaciers. A model based solely on temperature produced good results (coefficient of determination, r2 = 0.71) for the south-facing glaciers, but poor results for the north-facing glaciers (r2 < 0). The inclusion of a solar radiation index increased the modeled melt from the north-facing glaciers and thus improved the results (r2 = 0.73) for the north-facing glaciers, with little change from the south-facing glaciers. Including a wind index did not improve the correlation between modeled and measured runoff.

Vestari-Hagafellsjökull is a surge-type outlet glacier from the Langjökull ice cap, Iceland. Intensive hydrological investigations were carried out during non-surge conditions in the summers of 1999 and 2000, and 14 boreholes were drilled using pressurized hot water over an area 800 m from the margin and approximately 5000 m2 in size, where ice thickness ranged from 60 to 70 m. Initial investigations showed that a large fraction of the boreholes drilled to the bed did not drain and were assumed not to connect to the subglacial drainage system. Subsequently, we investigated the hypothesis that boreholes which remain full may do so as a consequence of a balance between englacial inflow and basal drainage rather than the standard assumption that such boreholes are simply unconnected. In testing this hypothesis, we developed a new technique for measuring water motion within the borehole by monitoring the passage of a saline solution down the borehole’s water column. The technique allows rates of motion to be established, as well as allowing the quantification of net addition and loss of water from the borehole. Observations based on the motion of saline plumes within the boreholes lead us to the conclusion that some boreholes do indeed remain full as a consequence of a balance between englacial inflow and subglacial drainage. The abrupt dilution that occurs at the top of these boreholes suggests inflow from a near-surface englacial water source, while the descent of the saline plumes implies that water is being lost at the base to the subglacial system. The system appears to be driven by excess water head in the boreholes over flotation and implies that the borehole/bedrock interface can be ‘leaky’.

In winter 2002/03 and 2003/04, thermistors were installed in the ice on two shallow ponds in central Alaska, USA, in order to obtain data on ice temperatures and their response to increasing and decreasing air temperatures, and flooding and snow-ice formation. Snow depth and density, and ice thickness were also measured in order to understand how they affected and were affected by ice temperature variability. The lowest ice temperature (–15.5˚C) and steepest temperature gradient (–39.8˚Cm–1) occurred during a 9week period in autumn 2002/03 when there was no snow on the ice. With snow on the ice, temperature gradients were more typically in the range –20 to –5˚Cm–1. Average ice temperatures were lower during the warmer, first winter, and higher during the cooler, second winter because of differences in the depth and duration of the snow cover. Isothermal ice near the freezing point resulted from flooding and snow-ice formation, and brief episodes of warm weather with freezing rain. Under these circumstances, congelation-ice growth at the bottom of the ice cover was interrupted, even reversed. It is suggested that the patterns in temperatures brought about by the snow-ice formation and rain events may become more prevalent due to the increase in frequency of these events in central Alaska if temperature and precipitation change as predicted by Arctic climate models.

Accurate modeling of ice sheets requires proper information on boundary conditions, including the geothermal heat flow (or heat-flow density (HFD)). Traditionally, one uniform HFD value is adopted for the entire modeled domain. We have calculated a distributed, high-resolution HFD dataset for an approximate core area (Sweden and Finland) of the Scandinavian ice sheet, and imbedded this within lower-resolution data published for surrounding regions. Within the Last Glacial Maximum ice margin, HFD varies with a factor of as much as 2.8 (HFD values ranging between 30 and 83 mWm–2), with an average of 49 mWm–2. This average value is 17% higher than 42 mWm–2, a common uniform value used in ice-sheet modeling studies of Fennoscandia. Using this new distributed dataset on HFD, instead of a traditional uniform value of 42 mWm–2, yields a 1.4 times larger total basal meltwater production for the last glacial cycle. Furthermore, using the new dataset in high-resolution modeling results in increased spatial thermal gradients at the bed. This enhances and introduces new local and regional effects on basal ice temperatures and melt rates. We observed significant strengthening of local ‘ice streaming’, which in one case correlates to an ice-flow event previously interpreted from geomorphology. Regional to local variations in geothermal heat flow need to be considered for proper identification and treatment of thermal and hydraulic bed conditions, most likely also when studying Laurentide, Greenland and Antarctic ice sheets.

Observations from basal water-pressure sensors along the length of Bench Glacier, Alaska, USA, show that diurnal fluctuations of water pressure are seasonal and restricted to summer. Most notable about these fluctuations is their disappearance in the late summer and early autumn, long before the seasonal end of diurnal meltwater input. Here we present data documenting the end of diurnal water-pressure fluctuations during the 2002 and 2003 melt seasons. The end of diurnal fluctuations occurred abruptly in multiple boreholes spaced meters to kilometers apart. There was no obvious spatial progression of termination events, and a clear correlation with meteorological forcing or discharge in the outlet stream was not apparent. After diurnal pressure fluctuations ended, basal water pressure returned to a high, generally steady, value either in an irregular pattern or by a distinct increase. This high water pressure was interrupted by episodic, acyclic events throughout the autumn before becoming stable and high in winter.

Measurements of electrical conductivity (EC) and deuterium isotopes in glacier run-off provide the basis for the application of simple mixing models that separate hydrographs into four constitutive components: subglacial, englacial, melt and rainfall-derived waters. Volumes of these components are determined from the models in two adjacent drainage basins within the glacier. Peak arrival times of both EC and isotopes during discharge events on short-term time-scales (days to weeks) differ in each terminus stream by as much as a factor of 5. Englacial water storage determined from the model varied greatly (98%) between neighboring basins within the glacier. Estimates of basal water volumes expressed as a layer thickness at the bed of the glacier differed by 50% (5 and 10mm each). Other results suggest that a greater percentage of water is stored at the glacier bed during rainfall events, and exceeds the storage capacity found within the seasonal snow and englacial zones combined.

The Canadian Lake Ice Model (CLIMo) is used to simulate the effects of climate variability and change on lake ice phenology (freeze-up (FU), break-up (BU), ice-cover duration) and total thickness and composition (snow ice, congelation ice) in central Alaska, USA. The model results suggest that, due to the Pacific Decadal Oscillation shift that occurred in 1976, the ice-cover duration is shorter (4 days) due to earlier BU in spring, and thinner (0.05 m) due to a reduction in the formation of snow ice. Sensitivity tests indicate that air-temperature changes cause the duration change, and snow-depth changes cause the total ice-thickness and composition change. The role of annual and monthly air-temperature and snow-depth changes is examined further in a series of experiments designed to elucidate the possible effects of future climate change. Air temperature is varied in 1˚C increments between –4 and +4˚C, and snow depth is varied in 25% increments between –100% and +100% of the long-term averages for 1952–75 and 1977–2000. The FU and BU dates (ice duration) are most affected by annual air-temperature change. Total ice thickness and composition are most affected by annual and monthly snow-depth change. A key finding is that snow-depth increases cause total ice thickness to increase as a consequence of increased snow-ice formation on top of the ice cover at the expense of congelation-ice formation at the bottom of the ice cover.