South East England and the Ice Ages

The northern hemisphere has been in the grip of a series of “ice ages” during the last 2.5 My. Britain has been near or at the southern edge of successive glaciations, though evidence for the early phases has mostly been obliterated by subsequent advances of ice sheets. Discussion of these phenomena is thus generally limited to the last half million years of glacial advance and retreat. Analysis of ice cores from Antarctica and Greenland, and of ocean sediments, has enabled the climatic history over this period to be determined. A summary of the temperature variation, particularly of the northern hemisphere is shown in Fig 1. The broken line indicates the present day temperature.

Fig 1: Schematic of Northern Hemisphere Temperatures from Oxygen Isotope Data.

The furthest advance of an ice sheet over the UK occurred during the Anglian Stadial (cold period) from 300 to 250 Kya (thousands of years ago), as shown by the lower line in Fig 2. It came as far as the north of London and is believed to have caused the River Thames to have changed its course to the more southerly current position. The most recent “ice age” was the Devensian, from about 80 to 13 Kya, the extent of which is shown by the upper blue line in the figure. Between these episodes was the Walstonian Stadial, for which only very limited evidence on the ground is available. As with the Devensian, that ice sheet did not come further south than the Midlands.

The south east of England was thus never glaciated. It was, however, a periglacial landscape for most of the last half million years. That means the land was frozen hard (permafrost) to a considerable depth, probably hundreds of metres, and for thousands of years at a time, broken only by the seasonal summer thaw of a thin top layer. It was only within this surface “active layer” that mass wastage (movement of erosion débris) occurred, thus creating the characteristic elements of periglacial scenery.


The influence of that environment is still evident in today’s scenery and geomorphology. From this last “ice age” we are only just emerging, in geological terms.

Fig 2: Extent of Anglian and Devensian Glaciations.

Climate and Landscape around the Mole Valley

Up till the 1950s it was thought that the uplift of the Wealden Anticline (dome) was a relatively rapid phenomenon occurring in the Oligocene (23 Mya). It created an area of high relief which has been subject to erosion ever since, both sub-aerial and marine. Large amounts of observational data on the geology and geomorphology of the area were accumulated from the late 18th Century onwards. A comprehensive synthesis of these data was made in the late 1930s by Wooldridge and Linton. This work was finally published in 1955 (Ref 6). It was considered definitive until stratigraphic and mineralogical research in the 1980s and onwards suggested that their interpretation of the data was wrong.

The newer paradigm is of ‘dynamic equilibrium’ in the evolution of the landscape. The Wealden uplift occurred only slowly and intermittently over more than 30 My. The tectonics were a mix of Atlantic spreading (Mid-Atlantic Ridge opening) and the fringe effects of the Alpine Orogeny.

By the late 1960s the physics of glacial erosion and of humid temperate erosion processes was well understood: but periglacial processes were only investigated in detail from around 1970 onwards, when access to arctic northern Canada and Siberia became easier.


Since then a large corpus of work has accumulated, which has led to two significant generalisations:-

  1. 1  During the last 500,000 years, it has been much colder than the present for over 80% of the time.

  2. 2  Periglacial land forming processes are at least an order of magnitude more effective at modifying landscapes than are the humid temperate processes currently acting around us.

The present conditions in which we live are thus anomalous, short-lived, and rare. We are in a warm interstadial. Frequency analysis of the last million years of ice advances suggests that this benign period is already overdue for ending.

Though the drainage pattern of the Weald and the southern London Basin is typically consequent and subsequent, having evolved more or less in equilibrium with the slow uplift of the Wealden Anticline, it is now realised that much of the present relief of the region is a relict (ie “fossilised”) periglacial landscape from the Devensian glaciation.

Because of the slow pace of landscape evolution, relatively sudden and fast moving changes leave the landscape in severe disequilibrium with the climatic régime above. Thus our present scenery is at variance with the current climate. Our countryside is barely yet affected by the sudden climatic amelioration of the last 10-12000 yrs.

The area of the Mole Valley would have been mainly tundra (Fig 3) during periods of ice sheet advance, with episodes of boreal (coniferous) or mixed forest during the milder interstadials. The most recent (though probably not the last) of these cycles ended about 10,000 yrs ago.

Fig 3: Northern tundra landscape

It is worth trying to see this landscape through “new eyes”. The present view north from Box Hill is shown in Fig 4. Something of how it might have looked around 270,000 yrs ago is shown in Fig 5. This was at the maximum extent of the Anglian ice sheet, the edge of which would possibly have been visible on a good day from any high point on the North Downs.


Fig 4: Looking north from Box Hill at the present day

Fig 5: Vatnajökull Icecap in Iceland, taken about 20 miles (32km) from the edge

Erosion Processes and Landscape Evolution

There are several geomorphological features characteristic of periglacial landscapes, mainly caused by mass wasting processes. Some of these are evident in the Mole Valley, while others are found in other parts of the Chalk Downs ridge and areas close by.


1 Dry Valleys

The Chalk Downs are dissected by many dry valleys. Fig 6 shows a typical dry valley (with a tributary coming from the right of the picture} on the northern dip slope of Box Hill.

Fig 6: Dry valleys at Box Hill

Dry valleys around the Mole Gap are shown schematically in Fig 7 as thin black lines. Their interpretation is still a matter of debate. Originally they were believed to have been created by stream erosion at a time when the chalk ridge and the water table within it were higher. The fact that there are three dry valleys that cross the Downs completely, at Gomshall, Betchworth and Godstone, suggested that these were once substantial consequent streams whose upstream ends were removed as the escarpment face migrated northwards and was lowered by erosion. The Mole Gap could be seen as an extreme example of this process, the river eroding its valley more successfully than the others nearby by virtue of its larger catchment area.

Fig 7: The Dry Valley system either side of the Mole Gap 5

This concept has now been superseded by the idea that on a frozen (permafrost) surface percolation of water into the chalk is blocked, and thus more surface runoff from snow-melt is available for fluvial erosion. Alternatively, it has been suggested that they were eroded by movement of frost shattered chalk rubble down the slopes (solifluction or viscous creep – see section 2 below).

Fig 8: Asymmetrical dry valley south of Bookham, looking north

Fig 8 shows a dry valley which demonstrates both the smooth sweep of its line with no meandering (see end of this section), and also asymmetry; the right bank, covered in trees, is much steeper than the other slope. There is evidence that such a difference may have developed by enhanced freeze/thaw solifluction on the W/SW facing slope due to greater insolation than on the opposite bank (see Ref 2).

The drainage density (ie the kilometres of river per square kilometre of land area) in this area is currently around 0.1km/sq km. The dry valleys of the Chalk have a very high drainage density of 2.3 km/sq km. This is said to reflect the rapid erosion of a soft rock over an impermeable surface such as underlying permafrost.

As Fig 7 shows, the valleys confined to the dip slope have in many cases been captured by other dry valleys developed along the strike of the chalk (ie more or less in an east-west direction). The significance of these strike valleys is not yet understood. There is a row of Eocene outliers resting unconformably on the Chalk dip slope, from Nower Wood near Headley north-eastwards to Banstead. It is possible that these strike valleys are relict subsequent drainage from a time when the Eocene outliers formed a continuous escarpment. The analogous feature is the Tillingbourne, and also the Pipp Brook, running between the Chalk and the Lower Greensand. Fig 9 assists in visualising this phenomenon; it shows the valley system as seen from over Bookham, looking SSE. The far horizon is the High Weald; the summit ridge of the Chalk runs from upper left to upper middle right of the picture.


Fig 9: Dry valleys on the Chalk dip slope in the Mole Gap area, looking SSE. OS Map Visualisation, vertical scale X6 [copied under license, Memory-Map]

2 Solifluction

Solifluction, or more specifically gelifluction, is the process of slow mass movement down permafrost slopes caused by repeated cycles of frost heave and thaw. It operates on slopes as gentle as a few degrees, and can have rates from a few mm/yr to several metres/yr. Prolonged frost shattering of chalk, either on an annual or a diurnal cycle, reduces it to a mix of rubble (clasts) and fines which during thaw constitute a viscous mud. Gelifluction movement of this broken chalk down dry valley slopes accounts for the deposits of “head” or “Coombe Rock” found in almost all dry chalk valleys. Often the larger clasts in the head show a distinct down-slope orientation indicating the direction of flow. This movement process tends to smooth out the relief of the chalk dip slope by moving material from the interfluves into the valley bottoms. An example can be seen south of Bookham, where ploughing shows the soil to be very thin between the valleys, and thicker (richer colour) in the depressions, even on very gentle slopes (See Fig 10). Though this differentiation was originally caused by gelifluction, it is now maintained by rain splash movement of soil particles in spite of the mixing effect of annual ploughing.


Fig 10: Variation of soil thickness originally caused by gelifluction

The effects of gelifluction are best seen where a valley is truncated, as in Fig 11. The brown head/soil mixture is clearly differentiated from the chalk rock.

Fig11: Cross section of head-filled dry valley near Seaford, East Sussex


Intense gelifluction can lead to accumulation of head as a deposition fan at the end of a valley. The River Mole swings away from the valley mouths of both the Polesden Lacey and the Headley Valleys. It has been speculated that this is due to deflection around such fans at the ends of the valleys. The Fredley estate sits on a noticeable rise in the surface which deflects the Headley Valley mouth towards the north. This may be such a deposit; it is shown as chalk on the geology map, but no borehole data exist to confirm it as solid rock or head.

Most dry valleys tend to be relatively straight or gently curved, typically like the Polesden Lacey Valley, or as in Fig 8. The Headley Valley is anomalous in that it is quite sinuous (see Fig 7). This suggests it was not formed primarily by gelifluction, but rather by earlier stream erosion, and has not been severely modified since. Its morphology is not fully understood.

3 River Profile

During the twentieth century river terraces, particularly of the River Thames, were much studied and even modelled mathematically. Tributary rivers such as the Mole had their terrace levels manipulated to correlate with the definitive Thames scheme

These correlations are now considered to be ill-founded. It is presently believed that the upstream terraces are controlled much more by local climate rather than by grading of the profile to a distant base level – in this case the confluence with the Thames at Molesey. On this hypothesis, only the lower reaches of a river are likely to be base level controlled.

However, this is again a matter of some debate, since two knick points in the river profile were identified early on in field studies, one close to Cobham Mill, and another in the upper reaches of the Mole at Meath Green. In both cases there is a distinct gorge developed below the knick point (see fig 12). Meath Green is very far above the Mole Gap, and any signs of conventional river grading may be a legacy from much earlier drainage development.

Fig 12: Gorge of River Mole near Meath Green; approx. 12 ft deeper than floodplain further upstream


The climate control hypothesis implies that during stadials aggradation occurs. Because of intense frost shattering of rock during these conditions and solifluction of the resulting rubble downslope, periglacial rivers tend to carry boulders and coarser material which are only moved during transient periods of high flow rate while thaw and snow-melt are occurring. These high peak loads ensure that the river beds are then clogged with gravel and sand. This forms a new floodplain, and development of braided channels across it ensues. During warm interstadials much slower normal fluvial erosion occurs, leading to lowering of the local river bed and meander formation as the floodplain is dissected. Fig 13 shows relict braiding of the Mole channel near Fetcham at the north end of the Gap; the same is also evident near Leatherhead Bridge. Above Leatherhead the channel is to a large extent a typical meandering river (but see Section 4 below).

Fig 13: Braided channel of River Mole near Leatherhead

A cross-river profile (Fig 14) midway through the Gap shows the present channel bottom at about 110 ft OD (33m), and a terrace at or just under 200 ft (61m) on either side of the river. Outside of the valley at this latitude the dip slope surface is between 350 and 450 ft OD (107-137m).

Fig 14: Cross-section of Mole Gap [OS copied under license, Memory-Map] 10


In the view in Fig 15 the thickly wooded strip in the middle distance is the transition slope between the 200 ft terrace and the general dip slope height of around 400 ft. The 200 ft terrace is taken to be the pre-glacial valley bottom. This is the result of incision into the Chalk of the originally consequent river up till about 1.5My ago. The lowering by about another 100 ft (30m) is ascribed to periglacial valley development.

Fig 15: Aerial view looking westwards from above Juniper Hall

Fig 16 shows the 200 ft terrace south of Druids Grove. Norbury House is on the horizon. The gentle slope up to the 400 ft level suggests this valley was wide and that the River Mole was mature and well-adjusted to it. The drop to the present river (at 110 ft OD) is a steep cliff just off picture to the right, where a meander is actively undercutting the bank.

Fig 16: The 200 ft terrace from the top of Ham Bank, looking north 11

So far only the vertical profile of the Mole has been discussed. In its mapped course within the Gap, the river appears to exhibit a mature meandering course – but see Section 2. Below Ham Bank (the left bank opposite the Fredley Estate) there is a highly unusual sharp bend, about 110 degrees deflection in about 30 yards (27m) of river. This is explained as the result of the river being captured by a fault or joint weakness, possibly when the chalk was first exposed during Wealden erosion. Research on several northern tributaries of the Thames flowing south on Tertiary deposits has shown that their sometimes anomalous alignments often follow deep seated faults in the unexposed Chalk. The Mole is the only southern tributary to have this disturbance among the other normally consequent rivers such as the Lodden, Wey and Darent.

4 Swallow Holes

During particularly prolonged dry weather the River Mole flow is insufficient to fill the main channel, and it flows through subterranean channels within the valley bottom. Parts of the river bed are then exposed. Such a drying out is shown in Fig 17.

Fig 17: A dry River Mole, Sep 2011 [courtesy Surrey Mirror]

Entry into these channels is down swallow holes (dolines, sink holes). In marginally dry periods the river only vanishes along short lengths of bed, but during prolonged drought the flow remains largely subterranean as far as the Fetcham Pond just west of Leatherhead. In still conditions the resulting underwater spring can be seen emerging from the pond bed as a small “chalk volcano”. Fig 18 shows an aerial view of the pond in 2003, with the spring visible as a ring of chalk slurry on the bed. Another circular feature close to the eastern end of the pond may be a second vent.


Fig 18: Fetcham Pond [courtesy Google Earth]

Fig 19 shows the location of several swallow holes. They generally occur as small holes in the river bank; obviously they are usually below the normal water level, and they are difficult to locate except in periods of low water flow. A survey in 1958 identified 25 active swallow holes between Dorking and Mickleham. Some swallow holes are very large, one near Burford Bridge being recorded at about 80 ft (24m) diameter (now capped). The Mole Gap also has several fossil swallow holes, where earlier meanders of the river may have flowed (see Fig 20).


Fig 19: Swallow Holes in the Mole Gap [courtesy BGS]

Fig 20: Dark spots visible during dry weather may be fossil swallow holes {courtesy Google Earth]

This phenomenon is not unique to the Mole, and they are found in some Chiltern rivers. It is not known whether the subterranean flow is along enlarged joint planes, or in a system of small caves. No inspections down swallow holes have so far been recorded. The size as mentioned above suggests some sizeable cavities probably exist.


It has been suggested that the Mole Gap itself is an unusually narrow gorge for a Chalk river valley, and may have resulted from subterranean solutional erosion and collapse of the river bed into the cavities (see next section).

5 The Whites

The Whites is an area of unstable cliff above the right bank, close to Burford Bridge (see Fig 21)

Fig 21: The Whites

It is an example of active erosion, at present due to sapping by the river at its base, though its earlier history seems open to speculation. The Whites is above the outside of a meander. Slumping movement is evident from the inclination of trees downslope and at the river bank (see Fig 22). However the remainder of the cliff is much less active, even though it is still being undercut by the river. An explanation for this may possibly be found at the top of The Whites. The cliff edge shows a distinct step profile over some distance along the edge (see Fig 23) and there are also examples of small scale local movements (see Fig 24).


Fig 22: Trees moving downslope below The Whites

Fig 23: Stepped surface above The Whites


Fig 24: Local slumping above The Whites

Chalk is known to support this type of slabwise slumping in other locations, leading to stepped profiles (See Fig 25). However the inclination of The Whites is much less steep than a sea cliff. It is possible that this is a relict periglacial landslip or débris flow from late in the Devensian stadial when freeze/thaw activity was becoming more pronounced. Current river erosion at the base will probably steepen the cliff slowly as poorly consolidated rubble is removed during periods of high flow rate. The river bed in this reach contains a particularly large quantity of coarse chalk clasts, which supports this hypothesis (see Fig 26).

Fig 25: Step-profiled landslip NE of Beachy Head


Fig 26: Coarse chalk rubble in river bed below The Whites

Just recently (2014) the author discovered a small hole in the Chalk a few feet from the cliff edge (see Fig 27). When sounded, a depth of 13 ft (4m) was noted without touching bottom. This might be due to erosion opening a natural joint in the rock, or else working along a residual tension crack in the rock formed at the time of the original slip.

Fig 27: Opening into a cavity in the Chalk above The Whites


It seems that the River Mole is not the simple country stream that it first appears to be. It has many interesting features, some unique to it, and several as yet not fully understood.

6 Further Reading

For a general appreciation of the Tertiary and Quaternary development of Great Britain, chapters 12 and 13 of Ref 1 are useful. Refs 2 and 3 cover all aspects of periglaciation; in particular ref 3 covers the effects across the UK in detail, giving separate treatments of lowland Britain and its mountainous areas. Ref 4 is a technical treatment of glaciation, its history and its effects in the Northern Hemisphere. Ref 5 covers comprehensively the dynamics of landscape formation, including specifically fluvial and periglacial processes. Ref 6 is the classical work on the geomorphology of south east England: though much of its interpretation has been rewritten in the light of subsequent research, it contains a wealth of observational data. It is also very readable.

  1. 1  Toghill, P; The Geology of Britain – an Introduction (2000); pub Airlife

  2. 2  French, HM; The Periglacial Environment (3rd Edn 2007); pub Wiley

  3. 3  Ballantyne, CK and Harris, C; The Periglaciation of Great Britain (1994); pub Cambridge University

  4. 4  Dawson, AG; Ice Age Earth (1992); pub Routledge

  5. 5  Ritter, DF, Kochel, RC & Miller, JR; Process Geomorphology, 4th Ed (2002); pub McGraw-Hill

  6. 6  Wooldridge, SW and Linton, DL; Surface Structure & Drainage in South-east England (1955); pub George Philip

[All photos and diagrams by D E Lester, except where indicated]