Report on the joint BGS/Engineering Group of the Geological Society meeting 'Glacial geology in engineering' presented by Professor Geoff Boulton, University of Edinburgh and Neil Trenter, consulting geotechnical engineer, (formerly Sir William Halcrow & Partners) held on 11 February, by James Milner-Smith of Mott MacDonald.
A joint meeting in February of the BGS and the EGGS at Great George Street, London focused on engineering in glacial geologies. The two main speakers approached the problems from differing, but not opposite views.
Professor Boulton, from the University of Edinburgh, focused on the development of geological formations that affect engineering behaviour of tills in practice. Neil Trenter, formerly of Sir William Halcrow, looked at the practical guidance of engineering design in tills, which is the subject of a soon to be published CIRIA report.
Both presenters agreed on the highly unpredictable nature of glacial soils and hence the difficulty of engineering in them.
Behaviour of tills in engineering practice -Professor Geoff Boulton
Professor Boulton's focus was on the products of till deposition and the processes that led to these. He began by highlighting that till behaviour is controlled by three factors - granulometry and mineralogy, sediment body geometry and drainage.
These govern not only the key properties of past and present formation of tills, but for the future too. Spatial attributes of order and scale cannot be ignored, and temporal attributes of event frequency and variability he suggests are also significant factors in controlling the often highly unpredictable nature of tills.
If the engineer can understand these factors then he is in a strong position to make an informed judgment on the likely ground conditions that may be encountered.
The granulometry and mineralogy of a till is a function of the geology from which the till is derived - glaciers which flow over crystalline substrata produce tills of larger grain sizes than those which have advanced over sedimentary facies. Professor Boulton's emphasis was on appreciating the landsystem under which deposition of the till occurred. In assuming one of three predictable situations for glaciation, predictable depositional environments are similarly described. These settings are the highland, the supra-glacial and the sub-glacial, (Figure 1). [Also Fookes, 1997]
The highland (glaciated valley) setting leads to the deposition of lateral moraines, which the professor described as producing a frequency of sedimentary type changes that is bewildering. Not
an encouraging prediction for the engineer working in such form-
The supra-glacial setting which gives rise to a distinctive topography which, if interpreted appropriately, can give a clear indication of the distribution of the underlying sediments, mainly sands and gravels, interfingered with tills.
The majority of tills, it was explained, are deposited sub-glacially. This situation gives rise to two significant processes that operate on the tills. Under a moving glacier, where the rate of advance is typically greater than the rate of drainage of the sediments, undrained loading occurs. This frequently generates very high sub-glacial water pressures leading to liquefaction and piping, producing strong dilatancy, with a void ratio of 0.6, compared with a void ratio of 0.4 at greater depths. Hence the granulometry and mineralogy of tills will be profoundly affected by the system under which they are deposited.
Professor Boulton was keen to demonstrate how strain within the sub- glacial setting controls till geometry leading to folding as sediments are forced up and over boulders or stiff sediments which retard movement, (Figure 2). The motion of the glacier leads to shear strain attenuating these folds - and refolding is common. He cited examples from Iceland where strain markers show major refolding episodes. In some instances sands have become so mixed with tills as a result of this process that they are effectively totally homogenized to produce massive tills. Hence, at most scales tills are highly unpredictable. The glacial system clearly produces organisation, but one to which we do not yet have the key.
Eskers are the deposits of sub-glacial rivers. These rivers would have been zones of reduced groundwater head, part of a dynamic system within the glacier. Models of Quaternary ice sheets of Scotland show a good correlation between esker distribution and predicted heads within the glacier. However Professor Boulton emphasised that because a glacier is always melting basally, large potential gradients tend to be produced and these fluctuate strongly. A further example from an Icelandic glacier was of a water head fluctuation of 250kPa over a 17 day period. This type of behaviour leads to non-linear strain and deformation.
Fluctuations of water head beneath glaciers not only affect the consolidation of till deposits but can also lead to hydrofracturing across interbedded silts and sands, which may lead to liquefaction of sand layers. In some instances overpressurisation beneath glaciers generate such phenomenon as hydraulic jacking of fractured bedrock, which result in very uneven and unpredictable rock heads.
Professor Boulton's main conclusions were to emphasise that the processes operating within glacial systems and terrains led to ground conditions which might appear to be unpredictable, but which do have an underlying order. It is for the engineer to be aware of the potential for complexity and for order, and to use this knowledge in investigating ground conditions. There is real progress being made in understanding the geological aspects of till sequences on a ground scale.
CIRIA report on engineering in tills -Neil Trenter
The new CIRIA report on engineering in tills will be issued by early summer. As ever, explained Neil Trenter, the CIRIA guide aims to be a reference for the engineer and geologist alike in focusing experience and professional wisdom into a handy document. The report will be a comprehensive account of tills, in terms of their origins, properties, distribution and engineering classification, as well as an account of engineering problems they can produce.
In his talk Trenter chose to focus on four of these engineering situations - earthworks, cuttings, tunnels and piling. He was keen to emphasise that glacial tills are only one of several types of glaciogenic sediments, along with fluvioglacial deposits and lacustrine deposits. Of these three the tills are the most variable in composition, grain size and distribution, though are often found in association with both or either of the other two.
With respect to earthworks, the CIRIA report focused on four. The first was the problem of boulders being misidentified as bedrock. The speaker emphasised that this is not only a contractor's problem but a consultant's problem too, and they should both take site investigation very seriously to ensure that these problems were minimised.
The second issue was in heterogeneity of tills. Often within what was thought to be mainly clay units granular layers or lenses were encountered. Since most clay tills have low plasticity indexes, and are workable over only small ranges, any influx of perched water can lead to slurrying making earthworks engineering difficult.
Thirdly, the experience of handling mixtures of cohesive and granular soils is that the two should be worked separately if at all possible, as the result of adding clay at its plastic limit to saturated sand would be a mixture having a reduced plasticity index, thus increasing the problem of slurrying further.
Trenter also wished to impress the distinction between acceptability and control, in regard to earthworks in tills. He stressed that these are two very separate processes. Acceptability is the identification of suitable materials, while control is the process of monitoring to ensure that the specification is met. It is important for the engineer to be aware of this and to enforce it.
With till cuttings drainage, till fabric and its relation to cut orientation and the presence of glaciolacustrine clays and old failure surfaces must all be considered. In heterogeneous tills, drainage is essential to ensure that instability of the cut face does not arise. Fissures and foliation must be checked in tills as unfavourable orientation can lead to instability if the dip of the fissures or foliations is out of the face and their strike is parallel to the orientation of the cut.
Trenter again emphasised the importance of site investigation to engineering in tills. He gave the illustration of Chandler's stability curve for intact till slopes in north east England (Figure 3) to
show a valuable relationship between slope height and angle; however fissures or foliation could not be ignored. Lacustrine silty clays show lower effective strength parameters than for tills and so it is imperative for the engineer to pay close attention to trial pits and borehole information in assessing design values. The engineer should also be aware of the chance of old failure surfaces, from solifluction
Difficulties of tunnelling within tills may arise from four main areas: mixed face conditions; varying till thickness and depth to rockhead; possibility of nests of cobbles and boulders; and varying groundwater conditions, including artesian pressures.
The speaker used case histories to show what can go wrong when some of the above are encountered. For example:
During the excavation of the Bishopsriggs section of the Kelvinside Sewer, through sandstone, the tunnel was inundated by some 300m3 of glacial deposits that had been encountered due to an irregular rockhead level.
The St Clair railway tunnel in Canada encountered difficulties when a nest of cobbles and boulders was struck during excavation [Finch 1996].
In Denmark water ingress of 3m3 per second resulted in the world's largest underwater dewatering at the Storeblt crossing project. Again, good site investigation is essential if potential problems are to be identified in good time.
For piling in tills Trenter discussed two issues raised by the CIRIA report: the design depth and how it can be achieved and the design approach to mixed successions. Founding a pile at shallow depth in granular soils may secure vertical compressive loading but would be quite inadequate for other load cases involving a strong horizontal load component. Pile types may be considered with respect to the landsystem within which they are to be placed (Figure 4). However the speaker also stressed the importance of a pre-construction strategy much on the theme developed by Tomlinson for driven piles. A minimum accept- able blowcount and depth over which it should apply is arrived at and, if this cannot be achieved a decision should be made in advance as to which con- struction seq- uence should be followed.
The speaker concluded that when it came to engineering in tills the mes- sage needs to be 'site invest- igation, site investigation, site investigation!'
Bob Seedhouse of Neath and Port Talbot Council, wish- ed to know whether or not the CIRIA report would provide any information regarding the accuracy of engineering as BS 5930 did not seem to.
Neil Trenter replied that the report did not elaborate on engineering descriptions but it did review the use of trial pits in getting a fuller understanding of the ground conditions and added that rotary core drilling, although expensive, is often a more appropriate method of investigation than cable percussive drilling.
Another question addressed the problems of shear planes at residual strength in lacustrine clays. Professor Boulton highlighted that discontinuities should be considered at all scales, and that laboratory tests are often inappropriate at determining the characteristics of bulk samples. Stability can be seriously affected by such shear planes, as was seen at Walton's Wood.
Professor Peter Vaughan of Imperial College (who described himself as 'retired') hypothesised that the void ratios of the highly heterogeneous tills of northern England are a result of variable amounts of water in the sediments at the time of deposition. He wished to know if this was credible, and was void ratio controlled from the time of deposition. In response Professor Boulton pointed out that under modern glaciers consolidation of till deposits varied greatly within short distances. He reiterated the problem of large changes in pressure heads and was interested in the professor's hypothesis.
The question of the irregular nature of permeabilities within tills was raised. Professor Boulton used the example of poor site investigation techniques in tills giving test values of permeability at 10m-11m/s, when it was known by experience to be five or six orders of magnitude higher. Neil Trenter added that the CIRIA report addressed this issue, and noted how fissure orientation affected permeabilities. Professor Boulton added that the permeability of a sediment below glaciers is not intrinsic to the material but to the system that creates it. The system will tend to organise itself to be hydrologically efficient.
Following the open discussion there were a further three short presentations. The first of these, by Mike Winter of TRL, looked at some recent research into the granulometry of tills. The effect of large particles on acceptability determination for earthworks compaction. The stone content at which the test ceases to represent the bulk excavated material appears to be around 45% to 50%. Below this value matrix determines compaction behaviour. Above this value the stones determine the compaction behaviour.
The next short presentation was by Greg Field of Fugro, regarding some glaciogenic soils offshore, west of Shetland. The sediments on the continental slope, between 450m and 550m deep are related to the Otter Bank end moraine sequence of the West Shetland shelf. These are understood to have been deposited in the absence of loading under an ice shelf and are normally consolidated. However tests showed that although the soils showed a normally consolidated soil profile (relative to North Sea ones) these sediments had lower moisture contents and higher unit weights.
The last speaker was Fin Jardine of CIRIA who extended his gratitude to the Department of the Environment and to the Highways Agency which had helped fund the project and to the invaluable input and hard work from the steering committee.
He reminded the audience that the next topics to be looked into would be the Mercia Mudstone and the Woolwich & Reading Beds, and CIRIA would welcome contributions towards these similarly valuable studies.