Mollers All 120 8000 Aarhus C Denmark

Northwestern Germany experienced several ice advances and retreats during the Last Glacial Maximum. This part of the Scandinavian Ice Sheet, located at the margin of the Baltic Sea basin, was dominated by a land-based but highly dynamic Baltic Ice Stream that terminated about 30 km southwest of the study area at its maximum extent. Abundant tunnel valleys, drumlins, and low preconsolidation ratios of tills suggest fast ice flow caused by some combination of enhanced basal sliding and deformation of soft sediment in this and adjacent areas (Piotrowski & Kraus, 1997; Piotrowski & Tulaczyk, 1999), both indicative of subglacial water pressure elevated to the vicinity of ice flotation point. This study was conducted to evaluate the capacity of the glacier bed

Figure 10.1 Study site at the Eckernforde Bay in northwestern Germany.

around the Eckernforde Bay (Fig. 10.1) to drain basal meltwater as groundwater flow, and constrain conditions favourable for generation of tunnel valleys, which are the most prominent features of the glacial landscape in northwestern Germany.

A hydrogeological model of the Eckernforde Bay was made using the available borehole data, geological maps at the scale of 1: 25,000 and field mapping of coastal cliff sections. The model comprises the entire Pleistocene succession and a part of the Tertiary, down to the first widespread aquitard considered impermeable and thus hydrologically inactive. The sediment succession was generalized into five major units as follows (Fig. 10.2): unit 1, Miocene lignite sands; unit 2, Miocene mica clays; unit 3, Pleistocene lower till complex; unit 4, Pleistocene sand horizon; unit 5, Pleistocene upper till horizon. These units represent two aquifers (1 and 4) and three aquitards (2, 3 and 5). Owing to the discontinuous distribution of the aquitards, in some places the two aquifers are in direct hydraulic contact.

Hydraulic conductivities were estimated from material description of the borehole data or directly derived from pumping tests. They range from ca. 5 x 10-5 to 5 x 10-4ms-1 for aquifers and from ca. 5 x 10-8 to 1 x 10-6ms-1 for aquitards. Spatial interpolation of thicknesses and conductivities of the hydrogeological units was carried out with a fuzzy-kriging procedure (Piotrowski et al., 1996; Marczinek & Piotrowski, 2002).

A three-dimensional, steady-state groundwater flow model for the catchment area of the Eckernforde Bay was created for the modern day conditions with the Finite Difference code MODFLOW (Plate 10.1A & C). The model was validated using water level data from 63 wells screened in both aquifers with a satisfactory result corresponding to the model of Kaleris et al. (2002), which shows a confluent discharge pattern from the surrounding aquifers into the bay, a regional groundwater sink at present. Subsequently, steady-state groundwater flow was simulated for full glacial conditions with the ice margin situated at a still-stand line several kilometres southwest of the study area. The lateral boundaries of the model running parallel to the ice-flow direction were taken as no-flow boundaries, the up-ice and down-ice boundaries were open to facilitate flow, and a prescribed head boundary was assigned to the uppermost layer, with head values corresponding to the maximum water pressure determined by the ice thickness. The ice thickness was calculated using a parabolic formula adjusted for soft beds and warm ice (cf. Piotrowski & Tulaczyk, 1999), in accord with palaeoglaciological data and theoretical considerations. The calculated ice profile rises from southwest to northeast and the ice thickness reaches about 250 m in the northeast corner of the model area.

The simulation under the ice sheet cover shows a fundamental reorganization of the groundwater flow field as compared with the modern-day (interglacial) situation. Driven by heads diminishing towards the ice margin, the groundwater in both aquifers discharges in the opposite direction, i.e. away from the Eckernforde Bay and the Baltic Sea basin (Plate 10.1B & D). Under glacier coverage, hydraulic heads are between ca. 85 and 185 m, whereas at present they are between ca. 2 and 14 m. Subglacial groundwater flow is faster by a factor of 30 than under non-glacial conditions, and the discharge along the down-ice boundary is ca. 4m3s-1.

In consecutive runs of the model we considered scenarios for both frozen and unfrozen ice margins, as well as channelized drainage along the ice-bed interface (Table 10.1). A single subglacial channel with water pressure just slightly below the glacier flotation pressure would increase the total discharge by about one-third. Permafrost under the ice margin would reduce the discharge by about 8% whereby the bulk drainage would occur through the Miocene aquifer. If both permafrost and the channel are considered (an unlikely scenario), then the discharge at the ice margin would increase by about one-quarter with respect to the basic model.

A conservative assumption of basal melting (36mmyr-1, Piotrowski, 1997a) as the only source of water at the ice sole gives recharge of ca. 14m3s-1 in the model area and its upstream catchment area. This is several times greater than the discharge at the down-ice boundary in all scenarios, which shows that the hydraulic capacity of the substratum was by far insufficient to evacuate all the basal meltwater to the ice margin as groundwa-ter flow, similar to the situation in other parts of northwestern

0 0

Post a comment