Glacial Geology of the
Stony Brook-Setauket-Port Jefferson Area

Gilbert N. Hanson


 

Home Glaciotectonics Tunnel Valleys Evidence References

Last updated on June 03, 2008

Tunnel Valleys

Subglacial stream valleys or tunnel valleys are remnants of  tunnel channels. 

"Tunnel channel: A large cavern, partly in the ice and partly scoured into the bed, extending into and under the ice from the ice margin, from which subglacial water drains. The remnant of a tunnel channel is a scoured, dry valley (tunnel valley), usually half-filled with outwash sand and gravel. The filling with sand and gravel often gives the valley a flat floor. Several examples in Wisconsin and Michigan have scoured valleys hundreds of feet deep (with maybe all but 150 to 200 feet filled with sediment), a quarter to a half mile across, and run several miles in length. The tunnel channel roof was probably as high up into the ice as the valley floor was scoured deep. Another peculiar feature of tunnel channels is that the floor of the channel may go up and down in elevation (a.k.a. up-and-down long profiles), unlike a scoured river valley with a steadily sloping floor. At times, the whole  cross-sectional area of the tunnel may have delivered torrents of water. Sugden and John (1976) point out, "up-and-down long profiles are most easily explained by water flowing under hydrostatic pressure in enclosed conduits." Tunnel channels typically form where the ice margin was frozen to the bed. The channels may have provided the only outlet for trapped subglacial water." 

Description from Geology 104 and 151 Ice Ages: Past and Future and  Ice Ages: Past and Present  at the University of Michigan taught by John R. Hoagland

'O Cofaigh hasreviewed the models for the origins of tunnel valleys. Characteristics of tunnel valleys (from C.'O Cofaigh, 1996):

  • Valleys are elongate depressions with steep often asymmentric sides which can be greater than 100 km long and 4 km wide

  • Over deepened areas cut into bedrock or unconsolidated sediment

  • Valleys are frequently sinuous and form anastomosing networks as opposed to dendritic patterns (See Fig. 1 below.)

  • May include relatively straight individual segments independent of each other

  • Till, if present, occurs along valley walls

  • Valleys typically have concave or undulating long profiles

  • Valleys commonly trend oblique to modern drainage pattern

  • Valleys may terminate in outwash fans at ice-marginal positions

  • Valleys have variable widths

  • Drumlins, eskers and transverse ridges may be preserved on valley floors

  • There may be smaller hanging tributary valleys

 

Fig. 1 Plan view showing an anastamosing tunnel valley network in northern New York and southern Ontario on the eastern shore of Lake Ontario from Shaw and Gilbert (1990).


 

 

The three main theories for the origins of tunnel valleys according to 'O Cofaigh, (1996)  are:

1) Subglacial stream erosion with the deformed sediments creeping to the site of erosion and being removed by the subglacial stream. The sediment surface under the glacier is thus lowered to form the valley. The The actual size of the tunnel valley is much larger than the channel of the stream. In some cases eskers are found on the floor of the tunnel valley which may represent the actual size of the channel.

2) The valley are formed by subglacial streams during ice retreat (deglaciation) at or close to the margin of the glacier and the valleys become younger up stream (time transgressive). The subglacial streams may be associated with Jokulhlaups (catastrophic subglacial meltwater floods).

3) Tunnel valley systems formed simultaneously by catastrophic subglacial meltwater floods. As opposed to the time transgressive origin in 2) above, a simultaneous catastrophic event is suggested by the large scale of the integrated anastimosing pattern of tunnel valley networks. 

Why might tunnel valleys be common along the north shore of Long Island?

Pietrowski, 1997, in a study of the subglacial hydrology in northwestern Germany gives a model that may be applicable to the north shore of Long Island. In this model of a warm (wet) based glacier water derived from melting of the glacier generally travels to the front of the glacier through the underlying sediments and as a water film along the ice/bed interface. When the front of the glacier was along the north shore of Long Island, the bottom surface of the glacier was rising toward the front of the glacier to the the south. Generally in a warm based glacier the front few kilometers of the glacier is cold based, that is the underlying sediments are in permafrost. These frozen sediments near the front of the glacier impede shallow groundwater flow. Pietrowski suggests that in an area where the slope of the interface between the glacier and the underlying sediments rises toward the front of the glacier that there is not a sufficient hydraulic gradient to discharge the melt water along the ice/bed interface or through the underlying sediments. As a result the melt water ponds under the ice sheet in the basin forming subglacial lakes. As the amount of water and water pressure builds  up open channels form to evacuate the excess water from the system. The flow the water through the channels are characterized by high discharge rates and short durations, on the order of days to weeks. The amount of melt water produced from the glacier would not be adequate to maintain the flow. Water pressure drops and tunnel valleys close. This hydrogical cycle would then be repeated. 

In the vicinity of Port Jefferson Harbor a subglacial tunnel developed which eroded the underlying sediment and redistributed into the outwash fan to the south of Port Jefferson. The potentiometric surface marks the top of the water table within the glacier and the top of the surficial stream. The arrows mark the possible path of water through a conduit within the glacier and through a subglacial tunnel. The water in the subglacial tunnel can have a high hydrostatic head.  As a result the water would have little difficulty flowing up from the harbor to exit from the glacier with a high velocity carrying a large load of sediment. 

 

 

Fig. 2 from Gustafson and Boyd (1987) shows the hydrology of the Malaspina glacier, Alaska. They suggest that this temperate glacier is an analog for the southeastern margin of the Laurentide ice sheet that deposited glacial sediments on Long Island. They suggest that most of the stratified sediment on Long Island had its source in subglacial streams. Subglacial streams eroded the underlying sediments and transported them to the front of the glacier. They reject the idea that the sediments in front of the glacier were derived dominantly from reworked till near the base of the glacier which was carried to the surface of the glacier and then reworked by streams as proposed for example by Koteff and Pessl (1981).

 

One characteristic of a tunnel valley is that there will be till draping the walls of the valley (C.'O Cofaigh, 1996). Several construction sites exposing the walls of valleys in the Stony Brook-Setauket-Port Jefferson area have been investigated showing till draping the valley walls. Go to this page showing till overlying sands and gravels in valleys in the Stony Brook-Setauket-Port Jefferson area.

 

Fig. 3 a The dark layer is till. If a stream valley forms sub aerially after the glacier has deposited the till (top image), the till layer will be cut and not drape the valley (bottom image).

Fig. 3b. If the valley is formed subglacially (in top image, blue is ice, dark blue is water in tunnel and in  bottom image white is air) after the tunnel valley forms , the advancing ice of the glacier is likely to deposit till along the walls of the glacier (bottom image).