INTERNATIONAL SOCIETY FOR SOIL MECHANICS AND GEOTECHNICAL ENGINEERING

This paper was downloaded from the Online Library of the International Society for Soil Mechanics and Geotechnical Engineering (ISSMGE). The library is available here: https://www.issmge.org/publications/online-library

This is an open-access database that archives thousands of papers published under the Auspices of the ISSMGE and maintained by the Innovation and Development Committee of ISSMGE. A FIELD STUDY OF FACTORS RESPONSIBLE FOR QUICK CLAY SLIDES. ETUDE SUR PLACE DES FACTEURS RESPON SABLES DES GLISSEM ENTS EN ARGILES TRES SENSITIVES

L. BJERRUM , T: LOKEN , S. HEIBERG, R. FOSTER. Norwegian Geotechnical Institute, Oslo, Norway. The City University, London, U. K.

SYNOPSIS The paper describes the result of a field study of a 45 km^ large m arine clay area 40 km nort h­ east of Oslo. The scope of the study was to isolate , describe, and evaluate the relative im portance of the various factors and processes responsible for the o ccurrence of quick clay slides. The m ost important finding of the study is probably the establishm ent of a zone of aggression within which the risk of qu ick clay slides is greatest. It is concluded that the possib ility of preventing quick clay slides m ay exist by the con­ struction of thresholds in the stream s to reduce fu rther erosion.

INTRODUCTION

Within N orw ay's 40000 km^ of m arine clay area, HAUFRSFTFR KLÛFTA quick clay slides present a serious problem , each 220 century leading to loss of lives and property. In the study of this problem the Norwegian G eotechni­ cal Institute has hitherto concentrated its activit y * 180 6 on detailed investigations of individual slides. Th e (A y ...... Marin« clay real solution to the slide problem is not, however, t 140 TVS- to explain a slide which has already occurred, but 2 B e d ro c k to find som e m eans of predicting and preventing 100 ___ I______I______I potential slides. To do this, it is necessary to 10 15 20 relate our understanding of the m echanics of a slid e K ilo m e tr e s to the forces at w ork in nature which are the cause s of instability. As the first phase of such a study, Fig. 1 G eological cross section illustrating the Institute in 1967 initiated a regional m apping deposition of the m arine clay 9800 years ago. study of a 45 km^ area in the R om erike district (Fig. 2) situated about 40 km NE of Oslo. QUARTERNARY GEOLOGY

This area was selected due to the fact that (I) it In com m on with the rest of Norway the area of has been the scene of num erous disastrous quick Rom erike was heavily eroded by the ice during the clay slides of which the U llensaker slide in 1953 glaciated periods. The bedrock is a granitic gneiss has been studied in great detail, (II) the quarter- of Pre-C am brian age with a tectonic structure in nary geology is w ell known and the originally sea the NE to SW direction which has been accentuated bottom plain is still intact adjacent to the northe rn by ice erosion. This has resulted in the form ation and western boundaries of the area, and (III) the of valleys and oblong knolls in the bedrock, and lower stream in the southern part of the area h'ts these can be seen clearly in the outcropping rock reached a grade of equilibrium , w hereas those in form ations. the northern part showed active erosion. After the final glaciation the m arine clays at R om e ­ In the study of the selected area all scientific rike w ere deposited in the sea in front of the disciplines relevant to the quick clay slide proble m retreating ice m asses during deglaciation about were used, nam ely geology, geom orphology, and 9800 years ago. The ice front stopped tem porarily hydrology which included erosion studies. The or even advanced slightly at H auerseter just north techniques of air and terrestrial photogram m etry of the research area, building up an outwash plain were applied. Seism ic and geoelectrical methods of sand and gravel (illustrated in Fig. 1). The w ere em ployed as well as conventional geochem ical finer particles w ere carried further south and and geotechnical methods and procedures. sedim ented as m arine clays interspersed by a few nearly horizontal 6ilt strata, which probably The study is certainly not com plete. This brief resulted from shorter periods of greater discharge review of the results to date is presented, however , of m elting water. The m arine lim it in the area has in the belief that they will dem onstrate som e of th e been established at 205 m above present sea level m ethods by which such problem s m ust be attacked. by Holtedahl (1924). Due to the isostatic land

531 BJERRUM, LOKEN, HEIBERG and FOSTER

NORWAY

Osloi ('

QENM ARKA EEEk UFA f

1475

KLOFTA

Fi g. 2 Map showi ng the r esear ch area. 1. Bedr ock area, 2. Ol d sea bot t om plain, 3. Qui ck cl ay sl i de scar s 4. Redeposi t ed qui ck cl ay sl i de masses, 5. Out croppi ng bedrock, 6. ” Fr ont of aggressi on” .

532 QUICK CLAY SUDES.

Fig. 3 Exam ple on stream erosion in old sea bottom plain from an area next to the 1967 quick clay slide in TrOgstad.

uplift follow ing the withdrawal of the glaciers QUICK CLAY SLIDES (Kenney» 1964) the m arine clay deposits which now have a surface elevation of 170- 185 m etres above Detailed descriptions in English of typical quick sea level cam e above the water level about 9500 clay slides have been published by G. H olm sen years ago. ( 1929), G. Holm sen (1946), P. Holmsen (1953), Bjerrum (1955b), Kenney (1967b) and Drury (1968). At the present tim e the clay deposits have consoli­ The slide always starts with a relatively shallow dated under their own weight, and, in addition, the initial slide in the slope towards the stream valle y, upper layers have been changed to a m ore or less frequently caused by brook erosion, and occurring stiff crust by drying and weathering. If a boring i s in clay which is generally w eathered and not quick. m ade in the old sea bottom plain, first a 5-7 m From this initial slip the slide develops backwards thick crust of stiff fissured clay is encountered. into the quick clay, towards the rear slope, sim ul­ Below this clay layer the norm ally consolidated taneously widening in all directions, see Fig. 4. soft m arine clay is found, extending to depths whic h The m ost characteristic feature of the slide is the can exceed 70 m . The clays which were laid down change of consistency of the clay in the process of in m arine water with a salt content of 35 g /l are sliding. As the clay becom es involved in the slide m ainly com posed of hydrous m ica, chlorite, quartz m ovem ents, and therefore rem oulded, it changes and feldspar. A description of their geotechnical to a viscous liquid. Through the opening form ed by properties has been given by Bjerrum (1954a). At the initial slide the clay slurry m oves from the som e locations where leaching has occurred and the cavity and descends down the valley carrying with i t salt concentration in the pore water has been re­ flakes of the upper dry crust. The speed with which duced, the clays have been transform ed into quick these slides occur can be illustrated by the Ullen- clays, which are characterised by the peculiar saker slide of 1953, during which about 200000 property that, when rem oulded, their consistency of liquid clay flow ed away from the slide within a changes to that of a liquid. few minutes. The number and magnitude of the quick clay slides occurring in the m arine areas can Since the m arine clay area rose above sea water be illustrated by the fact that an estim ated 50 m il l, level about 9500 years ago, at which tim e it appear ­ cu. m etres of quick clay have been rem oved by the ed as a wide, nearly horizontal, plain, it has been slides which have occurred within 25 km^ of the re­ subjected to a dram atic period of denudation lead­ search area. ing to the highly dissected landscape found today, as illustrated in Fig. 3. In the developm ent of thi s To guide the study of the factors and processes landscape the following processes have been of responsible for the quick clay slides the following governing im portance : consolidation, weathering, analysis of the conditions for their occurrence has seepage and leaching, frost action, erosion, slope been m ade : failures and slides, redeposition of slide m asses, vegetation, human activity. The occurrence of 1. The first requirem ent is obviously the existence disastrous quick clay slides, the scars of which ca n of a body of quick clay near and active erosion be seen in great num ber within the area, depended feature. As the quick clays are form ed by entirely on the relative rate at which som e or all of leaching with fresh water of m arine clays, the these processes were working. An engineering- occurrence of quick clay deposits is governed geological evaluation of the problem of quick clay by the seepage conditions. slides presum es therefore a thorough understand­ ing of these processes and the factors governing 2. A second prerequisite is active erosion of their rate. valleys by stream s, so that the stability of the

533 MERRUM, LOKEN, HEIBERG and FOSTER

Fig. 5 Terrace of redeposited slide m ass from Fig. 4 The Ullensaker quick clay slide, 1953. the 1765 quick clay slide.

valley slopes is being reduced and initial slides the m ost im portant bedrock outcrops occur in the are being precipitated. As m entioned above, it southern part w here the brooks have eroded their is these slides which trigger off the flow slides. valleys to the greatest depth, see Fig. 2. In the stream beds these rock outcrops form thresholds 3. A third requirem ent, generally satisfied by the resulting in sm all waterfalls. From the point of norm al pattern of stream erosion, is a topo­ view of the developm ent of the stream s these out­ graphy which allows the liquid clay to flow away crops are im portant as they form fixed points from the slide. governing the depth of erosion in the upstream part of the valley. A few rock exposures have also been 4. It is a further condition that the rate at which observed in the scars of som e of the old quick clay the clay behind the valley slopes is being slides. changed to a stiff crust by weathering should be sm all com pared* with the rate at which the Old sea b_ottom_plain_. The oldest and still easily valleys are cut down by erosion. Should the recognized feature of the sedim entary deposits in rate of weathering be faster than the rate of the landscape of the research area is the rem ains erosion, the slope failures would not cut deep of the wide nearly horizontal plain which represent s enough to expose a soft unweathered clay and the old fjord bottom . Bordering the bedrock area thus to start a quick clay slide. in the northern part of the research area where developm ent of the stream valleys i s still in pro­ The relative im portance of the factors and process­ gress, the horizontal plain dom inates the landscape . es governing the occurrence of the quick'clay slide s The sam e is the case in the w estern part of the are a will clearly vary from place to place. It is the pu r­ w here the village KlOfta is located. pose of the study of the research area presented below to obtain a picture of this com plicated In the central part of the area where the develop­ problem . m ent of the stream valley system has reached a further stage of m aturity and quick clay slides hav e intensified the denudation, the rem ains of the old GEOMORPHOLOGY plain are found as isolated areas in between the valleys and the slide scars. The old plain is today The area of study is m ainly used for farm ing, the m ainly cultivated farm land and nearly all farm flat plateau being cultivated for crop and the slop es buildings and houses are built on this plain. of the valley being used as pasture. M an-m ade work is very lim ited. Str eam_valley_s_._ The m ost dom inating feature in the landscape is the system of deep valleys cut down Based on a detailed field study of the research are a into the old sea bottom plain as a result of erosio n the various features of the landscape have been by the stream s and brooks draining the area. W ith­ classified, mapped and described. In Fig. 2 the in the area studied the stream valleys tend to run various types of landform s are shown, and the 5 in the north-south direction, see Fig. 2, possibly m ain categories of land features will be described following a system of old buried rock valleys. The below . young valleys in the northern part of the area show depths of 10-15 m . They have a distinct ” V -shaped” P.edr_o_ck_§rea_and rock, exposures^ Along the nort h­ form with sharp angles of intersection with the hor i­ east boundary of the research area the landscape is zontal plain into which they are cut down; see for characterised by the outcrop of bedrock, rising up illustration Fig. 3. The slope angle varies between to 80 m above the plain of the sea bottom . This 20 and 25°. The m ore m ature valleys in the area appears today as an undulating forest-covered central and southern part of the area m ay be up to rock area dom inated by hills reaching heights of up 25 m deep and they have a m ore rounded shape, and to 265 m above sea level. Within the research area their topography is frequently com plicated by

534 QUICK CLAY SLIDES. stream m eandering and slide activity. the area is still flowing on redeposited slide m ass es originating from the slides which occurred in the Scars_of_quick _clay_s.lj.deSj. One of the m ain res u lts eighteenth century. of the detailed m apping and field study of the area was the discovery of a large num ber of scars from hitherto unknown sm all and large quick clay slides. LEACHING The characteristic features of these scars are the wide bow l-shaped depression and the nearly hori­ It is w ell established (R osenqvist, 1946; Bjerrum zontal bottom having a very slight inclination to­ and R osenqvist, i^57‘) that the N orw egian quick wards the brook. M ost of them showed the typical clays have been form ed by leaching of the m arine bottle-neck opening towards the stream valley. The clays during the period that they have been above rem ains of the oldest quick clay slides appear toda y sea water level. The available evidence indicates as concave wide hanging valleys dem onstrating that that in this process thé effect of the downward per ­ they occurred at a tim e w here the stream valleys colation of surface water is negligible and that th e had not reached their present depths. Fig. 4 shows form ation of quick clays in alm ost all cases is due the 1953 slide. Behind this slide, between its uppe r to leaching caused by an upward flow of fresh left corner and the sinall barn, the indistinct sca r groundw ater. The artesian condition creating the of an old, still iarger quick clay slide can be dis ­ upward flow originates in general in the cracks and tinguished. fissures in the bedrock which com m unicate with a "free groundw ater” in fissures at higher elevations . In Fig. 2 all the identified scars of quick clay sl ides The consequence is that w here the depth to rock is have been plotted. W here the date of the slide is sm all the rate of flow and the chances of encounter ­ known it is shown, but m ost of them are undated. ing quick clay w ill be greatest. In the southern and central parts of the area the stream valleys are bordered by a closely spaced During the leaching of a m arine clay the reduction system of scars of quick clay slides, clearly dem on ­ in salt concentration in the pore water from the or i­ strating how the developm ent of the stream valleys ginal 35 g /l to about 5-10 g /l w ill have little eff ect in this area is accom panied by the system atic on the sensitivity, which is of course a m easure of occurrence of quick clay slides. In the northern the degree of instability of the clay structure. It is part of the area w here the brooks are still in the only when the concentration drops below 5-10 g /l process of cutting their valleys into the old sea that the increase in sensitivity occurs. In the ex­ bottom plain, no quick clay slides have occurred. trem ely quick clays the salt concentration is gener ­ It has thus been possible to draw a border line ally low er than 1 g /l. This non-linearity in the which in Fig. 2 has been called the "fron t of leaching process explains why quick clays in gener­ aggression” and which separates the northern intact al are found as isolated lenses or layers surrounde d area from the southern area of quick clay slides. by m arine clay with a norm al low sensitivity; in The nam e has been chosen to draw attention to the spite of a reduced salt concentration, the m arine significant fact that all the youngest slides occur red clay still contains m ore than the critical 5-10 g /l . just south of this front. Evidence of this process within the research area

J_e_r_r^-_c_e.s_ J _ e^_eP.P_sjted _slid e m a sse s. In th e m ain is given by the correlation which can be establishe d stream valleys of the upper central part of the are a between the occurrence of quick clay slides and the a num ber of terraces are found, which have been depth to rock. As seen from Fig. 1, m ost of the identified and m apped on Fig. 2 as redeposited slid e quick clay slides are concentrated within a zone m asses. Such a terrace deposit is shown in Fig. 5. running N orth-South, through the central part of th e These slide m asses originated from quick clay area. Available borings, rock outcrops and seis­ slides which have occurred higher up in the valley. m ic profiling indicate that within the zone of clos e­ ly spaced quick clay slides the depths to bedrock

Each quick clay slide was accom panied by the flow below the original horizontal plain w ere 15-35 m , of rem oulded slide m asses descending like a liquid com pared with depths of m ore than 70 m in the area down the stream valleys until they finally m et an im m ediately west of this zone w here few scars of obstruction and stopped, form ing a huge lake of quick clay slides have been found. liquid clay. W ith tim e these redeposited slide m asses consolidated to a relatively stiff clay. It is The results of a detailed study of the 1953 slide i n this clay w hich is -found as terraces in the valley s U llensaker (B jerrum , 1954b; Kenney, 1967a) also below the younger quick clay slides, see Fig. 5. confirm the postulated corrélation between depth to There w ere no problem s involved in identifying the rock and occurrence of quick clays. In the cross 6 redeposited slide m asses from the 1953 slide, see section in F ig. are plotted the results of 20 bor­ Fig. 4, and the terraces in the m ain stream valley ings which dem onstrate clearly that the zone of could be identified as belonging to the 1765, 1736, quick clay, and the slide, occurred exactly where and possibly the 1475 slides , see Fig. 7b. the depths to rock w ere sm allest. At both sides of the buried " rock h ill” the clay still has a low se n­

In spite of their lim ited extent these redeposited sitivity, proving that the leaching has not proceed ­ slide m asses have payed an im portant role in the ed as far as above the hill top w here the gradient rate of downcutting of the stream valleys. The fill ­ is highest. ing up of the valley with slide m asses obviously m eans a delay of upstream erosion until the stream has cut a new channel through the slide m asses. Because the redeposited slide m asses are m ore resistant against erosion than the virgin sedim ent, this delay can be considerable. It is thus a m atter of fact that the m ain stream in the central part of

535 BJERRUM, LOKEN, HEIBERG and FOSTER

m ents accom panying the upward leaching and this causes a reduction in sensitivity. For instance, under the quick clay form ation in which the 1953 E A S T WEST slide at Ullensaker took place, there was found a clay of low sensitivity which showed a low concen­ Surface before sliding S t re a m Old sea bottom tration of salt in the pore water. A detailed chem i ­ After sliding cal analysis of this clay is not available, butitis be­ lieved that its low sensitivity can be explained by the chem ical analysis of a sim ilar clay from Dram m en. In Dramm en the chem ical analysis showed a change in com position of the pore water. The concentration of M agnesium ions was as high as 150 m g/l com pared with 20 m g/l in the corre­ sponding quick clay. This increase, which probab­ ly results from a disintegration of the chlorite Fig. 6 Schem atic cross section through the m inerals, is accom panied by an increase in liquid U llensaker quick clay slide, 1953. lim it and a decrease in sensitivity. This im prove­ 1. W eathered crust m ent in properties of the clay thus results from 2. Marine clay the exchange of absorbed cations rather than the 3. M arine clay, leached until quick cem enting effect dom inating in the surface w eather­ 4. M arine clay, leached until quick, then ing process. weathered.

WEATHERING STREAM EROSION

As mentioned above, where exposed, the upper 4-6 The only agents of erosion of real concern are the m of the m arine blue soft clay have by drying, free z­ stream s draining the research area. It should ing, leaching and oxidation been changed to a crust im m ediately be m entioned that these stream s are of brownish fissured stiff clay. The geotechnical sm all. The average discharge of the stream where properties of this crust have been discussed by a m easuring weir was established, see Fig. 2, is Bj errum ( 1 954b), M oum and Rosenqvis t (1957 ) an d about 180 litres per second and in the fall of 1968 a DiBiagio and Bjerrum (1957). peak of about 1500 litres per second was observed, and in m ost of the dry period the discharge was

As a result of surface water penetrating through th e less than 5 litres per second. That this is, howeve r, system of fissures the, upper crust nas been leache d a significant discharge is indicated by the am ount of and the salt concentration in the pore water is low . m aterial transported out of the area per year, whic h The crust is, however, not quick because the has been estim ated on the basis of observation of weathering processes accom panying the leaching sedim ent concentration at the m easuring weir. The with oxigen-rich water lead to an increase in stren gth loss of m aterial from the 6 km^ catchm ent area is and plasticity of the clay which m ore than offset t he estim ated to be 500-1000 m^ per year and by far the sim ultaneously occurring effect of the salt leachin g. greatest transport occurs during the peak discharge . By weathering is here understood a disintegration o f This m aterial consists not only of soil rem oved by the clay m inerals occurring as a result of the low pH the active erosion of the stream beds, but also soi l and the high oxidation capacity of the surface wate r carried down to the stream s by sheet wash, frost which has replaced the pore water. This disintegra­ action, etc. tion of the clay particles leads to a release of ca tions which partly are rem oved by leaching, but which In order to describe the erosion occurring it is us e­ partly also are absorbed at the surface of the clay ful to follow the stream Hynna from the outlet in particles, thereby increasing the plasticity of the the m ain river, Rttmua, to the watershed at the clay, or precipitate as nearly in soluble com pounds northern part of the area. In this description it i s acting as cem enting agents (Rosenqvist, 1955; possible to distinguish between three significantly LOken, 1967 ; M oum et al, 1968). Recent research different reaches : indicates that iron and alum inium ions dom inate in this process (M oum , 1968) (1) The first 6 km, the zero point being the outlet in the River ROmua, is the lower reach, see

The rate at which the weathering process proceeds Fig. 7 a. This stretch of the stream years ago at depths depends on the rate at which oxigen-rich reached a stable elevation governed by num er­ surface water w ill penetrate and this again depends ous rock outcrops. In between the rock out­ on such factors as depth of fissures and roots of crops the stream flows in virgin m arine clay vegetation. As far as these factors can be evaluate d with very sm all gradients, 1 : 300 to 1 : 500. Of it is believed that the rate of weathering is very high the total head loss of 27 m , 15 m are lost by im m ediately after the exposure of a clay surface - water falling over rock outcrops. W hereas the as can be observed on the valley slopes - but is ex ­ further downcutting of the valley has stopped, trem ely sm all when the thickness of the crust has at sharp bends in the stream there is under­ reached the depth at which chem ical influence is cutting and m inor slope failures. The slope in negligible. this reach have, however, in general com e into equilibrium and this reach is considered to have reached m aturity, so that the danger W hereas the weathering proceeding from the sur­ face is a well studied phenom enon, it has only of quick clay slides is sm all. recently been appreciated that a sim ilar chem ical phenom enon takes place from the bottom of the sedi-

536 QUICK CLAY SLIDES.

cut itself through these m asse 9 is of the order of 4m per century. Further upstream , stream A and stream B both flow through slide m asses which originated from the 1765 slide. As long as the stream s are eroding in slide m asses, the slope failures which occur are lim ited to slips in redeposited clays, and the danger of quick clay slides is elim inated. From the standpoint of potential quick clay slides, these areas can be thought of as tem porarily neutral­ ized. This applies to alm ost the whole of the interm ediate reach, the only exception being at a few locations w here the stream has cut al­ m ost through the edge of the 9lide m asses and active erosion m ay occur in the virgin clay. At a later date, when the stream has cut through the slide m asses, it is not unlikely that active erosion m ay be resum ed in the upstream end of the reach, w hereas it is m ost likely that equi­ librium had already been reached in the down­ stream end prior to the deposition of the slide m ass es.

(3) The upper reach of the stream s includes the reaches upstream of km 9.3 for stream A and Distance in km 8.6 for stream B. This reach is characterised by the fact that the stream beds are in virgin Fig. 7 Longitudinal profiles of the m ain stream clay and active erosion is occurring. Hynna, (a) ” lower reach” , (b) "inter­ mediate reach” and "upper reach” . The active erosion is partly a vertical down- cutting of the stream bed, but in addition "h ori­ (2) The interm ediate reach starts approxim ately at zontal erosion” is occurring at num erous locat­ km 6.0. The longitudinal section through the ions, for instance at bends, or w here the stream stream s and their tributaries north of this point is obstructed by tree roots, etc. In addition to is shown in F ig. 7b. At km 7. 7 the tributary the stream erosion there is a considerable stream B leaves the m ain stxeam A. The inter­ am ount of clay brought down to the stream by m ediate reach continues in stream A up to km such agents as frost, sheet wash, slope failure, 9. 3 and in stream B to km 8. 6. cittle, and hum an activity . The contributing processes are very m uch m ore im portant in the The interm ediate reach of the stream is defined upper than in the low er reaches. as the portion w here stream s are flowing alm ost entirely on redeposited slide m aterial. At an The "fron t of aggression” was defined above as earlier stage the stream beds w ere low er than the upstream boundary of the occurrence of they are today, and generally speaking the quick clay slides. The redeposited m aterial stream s are in the process of working them ­ resulting from such slides w ill, by definition, selves back to their old levels. introduce the stream regim e referred to as the interm ediate reach. H ence, the front of ag­ The thickness of the redeposited slide m asses gression coincides very nearly with the bound­ upon which the stream s are flow ing varies over ary between the interm ediate and upper reaches. the length of the reach depending on w here the slide m asses from the various quick clay slides cam e to rest. At locations w here the slide At, or just above the front of aggression, at the m asses w ere redeposited the new conditions led low er end of the upper reach, the stream w ill to a reduction in the erosion when the stream have cut m ost deeply into the virgin clay, and hence the probability of the slopes com ing into passed over the slide m asses and an increase in erosion w here the stream cut through the down­ close proxim ity to a zone of quick clay is high­ stream part of the slide m asses and continued est. It is, therefore, logical to expect quick clay slides to be concentrated at these points in the old stream bed. in the tributary stream s.

F or instance in Fig. 7b at km 6. 0 th ere is a The front of aggression w ill not advance at a sharp drop in the stream bed. At this location uniform rate up the stream . After each slide, the stream flow s over the dow nstream end of re- the redeposited slide m asses from a step in the deposited slide m asses originating from a slide profile, and the rate of erosion in the upper which occurred about 20 years ago. Upstream reach w ill be reduced. H owever, when the of this drop, w here it flow s through 2 km of re- slide m asses have been cut through, there w ill deposited clay, the gradient of the stream has begin a rather rapid and intense period of eros­ been reduced from the original 1 : 200 to 1 : 400. ion on the virgin clay of the upper reach, con­ centrated initially at the front of aggression. U pstream of km 7. 5 the stream has alm ost ero­ ded its bed through the slide m asses originating This rapid erosion w ill probably occur at a from the slide of 1736, the terraces of which m uch faster rate than the counter process of are clearly observed at both sides of the brook w eathering, thus intensifying the possibility of at elevation 145. The rate at which the stream quick clay slides. This non-uniform rate of

537 BJERRUM. LOKEN. HEIBERG and FOSTER

active erosion of the virgin clay m eans that the occurrence of slides is inclined to be periodic. CATCHMENT AREA, KM2

In the above description no distinction has been made between stream s A, B and C. In the upper reach there is, however, a distinct difference be­ tween the stream s. W hereas stream s A and B are fed m ainly by surface run-off and norm al drainage of the slopes, stream C has cut its bed so close to the bedrock that an artesian set of ground water conditions prevails. Occasionally the stream flows through mud volcanos, and the softening effect of the upward artesian flow has also greatly influenc­ ed the stability of the valley slopes.

To obtain a com plete picture of the stream erosion Fig. 8 Correlation between catchm ent area, stream in the various reaches it is necessary to appreciat e gradient and erosion that the discharge of the stream increases consider ­ ably from the upper to the lower reaches. The abi­ lity of a stream to cause erosion depends not only on its gradient, but also on its discharge which again depends on its catchm ent area. In order to obtain an im pression of the relative im portance of these two factors, data was obtained for a num ber of lengths of stream , and in Fig. 8 they are plotte d in a diagram showing gradient as a function of catc h­ ment area. Each length of stream was designated as either aggressive, at grade or slack, depending on whether erosion is occurring, whether the stream bed is just equilibrium , or whether the stream has reached ultim ate m aturity. Fig. 9 Longitudinal profiles of stream s A and B Based on this classification a curve can be drawn with theoretical equilibrium grades. through the points in Fig. 8, separating the length s with active erosion fr6m those which are in equili­ which can trigger off quick clay slides is illustra ted brium or slack. by the 1953 slide.

The relation between the equilibrium grade and catchm ent area shown in Fig. 8 can be used to pre­ SLOPE STABILITY dict how far erosion of a stream will proceed be­ fore equilibrium conditions are reached. As an The field studies included a survey of the valley exam ple are shown in Fig. 9 longitudinal .sections of slopes in the interm ediate and upper reaches of the stream s A and B covering the interm ediate grades stream s (see below ). All slides and failures and using the northern-m ost rock outcrop at km 5.3 in representative stable slopes, w ere mapped, des­ Fig. 7a as a basic fix point. From this figure it c an cribed, and classified. A study of a num ber of be learned that in the upper reach of stream B a stable slopes in virgin m arine clay in the upper further vertical erosion of the order of 5 m can be reach of stream A, showed that the slope cam e into expected and that very aggressive erosion is occur­ equilibrium with a regular plane surface. M ost of ring in a zone at km 8. 6. In brook A aggressive the stable slopes in this area appear to stand at a n erosion is at present occurring at km 8.0 in the angle of 20-21°, independent of height. H owever, slide m asses of the 1765 slide and at km 9. 2 in th e- drainage conditions and difference in properties of rem ains of an undated quick clay slide. Just up­ the soil m ay cause deviation from this general set stream of the boundary between the interm ediate of conditions. The equilibrium shape and inclinat­ and the upper reach of stream A a further down- ion of the slope indicate that the stability is gov ern­ cutting of the order of 3 km can be expected when ed by the properties of the upper weathered crust the front of aggressive erosion eventually reaches m ore than by those of the lower soft clay. The this point. Further erosion on stream A is, how­ upper crust possesses, as mentioned, a system of ever, lim ited to the distance from km 9.2 to 10.0. fissures prom oting seepage parallel to the surface, North of km 10.0 the stream is already at equlibri- which is the only set of pore-pressure conditions um grade. which is consistent with inclinations as steep as 2 0 . The fissures also explain why the stiff clay in spi te The above description of the various reaches con­ of a high strength behaves like a cohesionless m at­ cerns the main system of stream s. These stream s erial. To surface parallel seepage and an inclinat­ have, however, a number of sm all tributaries and ion of the slopes of 20° there corresponds an angle for each of these tributaries a sim ilar set of cond it­ of shearing resistance of about 38°. ions will exist, as described for the main stream s. The tributaries have each an upper reach w here The following types of slope instability were erosion is occurring, but as the catchm ent area of regonized. m ost of them is sm all, the rate of erosion is low. Ultim ately, the tributaries will reach equilibrium (1) Deep rotational shear failures with gradients corresponding to the catchm ent area, see Fig. 8. That the tributaries m ay cause erosion These are failures extending up into the valley

538 QUICK CLAY SLIDES.

slopes. They occur only where there is active happened and what is at present happening in the re ­ stream erosion. The stream channel is deep and search area, and it can also be of help in the pre­ undercutting of the toe is proceeding. This type of diction of what is likely to happen in future. failure dom inates in the interm ediate reach w here the stream s are eroding redeposited slide m asses. The m ost significant result of this study is the in ter­ H ow ever, a num ber of these slides w ere caused by action between the rate of erosion and the occurren ce excavations in connection with construction of the of quick clay slides. Each tim e a quick clay slide local roads. occurs the liquid slide m asses flow ing out of the slide scar w ill be redeposited further down the The largest deep rotational slide which was found valley. These redeposited slide m asses w ill form occurred in the upper reach of stream C in connect­ a threshold in the stream which w ill delay tem pora­ ion with an artesian ground-w ater condition. At thi s rily erosion upstream of the slide m asses. The location the bottom of the valley was flat and soft , consequence of this interaction is that the quick c lay and the artesian condition m anifested itself by slides are occurring along a front of aggression ad ­ springs w here soft mud was brought to the surface. vancing stepw ise in an upstream direction.

(2) Sm all bank or slo^e failures The danger of quick clay slides is thus dependent o n the rate of erosion occurring in the virgin clay ne xt By far the greatest num ber of failures w ere to the front of aggression. The danger w ill, howeve r, rather sm all bank failures caused by m eandering or depend on whether the rate of downcutting of the downcutting of the stream . The depths of these fail ­ stream exceeds the rate at which the soft m arine ures are about 1-2 m only. In the sam e category be­ clay in the valley slopes is transferred to a stiff clay long the shallow flake-type surface failures which by w eathering. If the rate of w eathering exceeds th e are usually caused by frost action, but m ay also be rate of erosion, the slides accom pnaying the erosio n a result of surface seepage. w ill be sm all and shallow and the risk of the slide s exposing soft sensitive clay is therefore sm all. (3) Surface erosion Another result of the study is the determ ination of In addition to the above described slope fa il­ the equilibrium gradients for the stream s from ures a very considerable contribution to the denuda t­ which it is possible to predict future erosion. ion is the gradual downhill transport of soil by su r­ Applied on one of the m ain drainage stream s it is face.erosion. Am ong the various types of surface predicted that the in the upper reach north of the erosion the sheet wash by precipitation accounts fo r "fron t of aggression” the stream w ill in future cut the greatest transport, m ost of all in cultivated itself a further 5 m into the clay. It is obvious t hat areas. The erosion by sheet wash is very much aji erosion of this order of m agnitude w ill cause accelerated by cattle breaking up the grass cover a nd num erous slope failures of which one or several rem oulding the upper clay. m ay release quick clay slides. There is thus every reason to believe that the num ber of quick clay Of less im nortance is the downhill creep caused by slides which w ill occur north of the line of aggres s­ frost action. An interesting type of surface erosio n ion in the centuries to com e w ill be as large as by sheet flow is occurring in the valley slopes of the those which have already occurred south of the line . upper reach w here new tributaries are in develop­ m ent. W here this occurs, a depression is observed The study has also shown that it is not unlikely th at in the upper part of the slope and a bulge near the the danger of future quick clay slides triggered of f toe, the m inim um width being in the low er half. by further erosion can be reduced at reasonable cos t. The discharge of the stream s causing the erosion is sm all. The possibilities of constructing sm all arti ­ CONCLUSIONS ficial thresholds in the stream s, such as low cost w eirs, to prevent further erosion look, therefore, The investigations of the past 15 years into the pr o­ rather prom ising. A study of how this can be done perties of m arine clays and the detailed studies of a and the fu ll-scale testing of such structures in th e num ber of quick clay slides have provided a basic research area w ill be the next stage in this study. knowledge of the various factors governing the occu r­ ence of quick clay slides. This knowledge has now Finally, the geotechnical studies carried out have been utilised in a regional engineering geological dem onstrated that the leaching causing the clays to study of a m arine clay area. The initial purpose of be quick ultim ately w ill cause a w eathering of the this study has been to describe, classify, and m ap quick clays, proceeding upwards from the bedrock. all factors and processes of im portance for the This w eathering w ill gradually reduce the sensitivi ty occurence of quick clay slides. and increase the shear strength of the clay. There­ fore, if it should prove possible to prevent future The conclusions of greatest interest are obviously erosion and thus to "freeze” the topography, it is those which have a bearing on the possible prevent­ believed that with tim e the stability of the slopes ion of the disastrous quick clay slides endangering w ill increase and thus the danger of future quick life and property in the 40000 km^ m arine clay area clay slides w ill be practically elim inated. in N orw ay. It is still too early to present final r e­ sults of the present study. H owever, a few im port­ ant findings m ay be m entioned.

The study has provided a picture of how a m arine clay area gradually is eroded down by stream s, slope failures, quick clay slides, etc. This pictur e seem s to be consistent, both with regard to what ha s

539 BJERRUM. LOKEN. HEIBERG and FOSTER

LIST OF REFERENCES Kenney, T.C . (1964) Sea-level m ovem ents and the geologic histories of Bjerrum , L. (1954) a. the post-glacial m arine soils at Boston, N icolet, G eotechnical properties of Norwegian m arine clays. Ottawa and O slo. G Sotechnique, V ol. 4, N o. 2, p. 49-69. G eotechnique, V ol. 14, No. 3, p. 203-230. A lso published as N orw egian G eotechnical Institute. Publ. , 4. Kenney, T.C . (1967) a. Shearing resistance of natural quick clays. Ph.D . Bjerrum , L. (1954) b. Thesis subm itted to U niversity of London. Stability of natural slopes in quick clay. O slo, N orwegian G eotechnical Institute. 138 p. European C onference on Stability of Earth Slopes.

Stockholm 1954. Proceedings, Vol. 2, p. 16-40. Kenney, T.C . (1967)b. A lso published in G eotechnique, V ol. 5, N o. 1, 1955 , Slide behaviour and shear resistance of a quick cla y p. 101-119, and in N orw egian G eotechnical Institute . determ ined from a study of the landslide at Seines, Publ. , 10. N orway. G eotechnical C onference O slo 1967 on Shear Strength Bjerrum , L. and I. Th. R osenqvist (1957) Properties of Natural Soils and R ocks. Proceedings, N orske leirskred og deres geoteknikk. V ol. 1, p. 57-64. Naturen, V ol. 81, No. 2, p. 108-121. A lso published in N orw egian G eotechnical Institute. LOken, T. (1967) Publ., 15. R ecent research at N orw egian G eotechnical Institute concerning the influence of chem ical additions on DiBiagio, E. and L. Bjerrum (1957) quick clay. Lecture presented at the Sw edishSociety Earth pressure m easurem ents in a trench excavated for Clay R esearch. M eeting. Oslo 1967. in stiff m arine clay. To be printed in G eologiska FOreningen i Stockholm . International C onference on Soil M echanics and F Or h an dlin gar. Foundation Engineering, 4. London 1957. Proceed­ ings, V ol. 2, p. 190-202. M oum , J. and I. Th. Rosenqvist (1957) A lso published in N orw egian G eotechnical Institute. On the weathering of young m arine clay. Publ., 26. International C onference on Soil M echanics and Foundation Engineering, 4. London 1957. Proceed­ Drury, P. (1968) ings, V ol. 1, p. 77-79. Th$ H ekseberg landslide, M arch 1967 ; an illustrat­ A lso published in N orw egian G eotechnical Institute. ion of the R om erike clay landslide problem . Publ., 26. N orwegian G eotechnical Institute. Publ., 75, p. 27- 31. Moum, J. (1968) Chem ical environm ent com pared with m echanical Holmsen, G. (1929) behaviour. Lerfaldene ved Kokstad, Gretnes og Braa. G eotechnical C onference O slo 1967 on Shear Strength O slo. 45 p. Properties of Natural Soils and R ocks. Proceedings N orges geologiske unders0kelse, No. 132. V ol. 2, p. 125-126.

Holmsen, G. (1946). M oum , J. , O. I. Sopp [ and] T. LOken (1968) L eirfalltyper. Stabilization of undisturbed quick clay by salt w el ls. Statens vegvesen. Veglaboratoriet. M eddelelse, 4, Vag- och vattenbyggaren, V ol. 14, No. 8, p. 23-29.

P -1-3. A lso published in N orw egian G eotechnical Institute. Publ., 81. Holmsen, P. (1953) Landslips in N orw egian quick-clays. Rosenqvist, I. Th. (1946) G eotechnique, V ol. 3, No. 5, p. 187-194. Om leirers kvikkaktighet. A lso published in N orw egian G eotechnical Institute. Statens vegvesen. Veglaboratoriet. M eddelelse, 4, Publ., 2. p. 5-12. Holtedahl, O. (1924) Studier over isrand-terrassene syd for de store 0st- Rosenqvist, I. Th. (1955) landske sj 0er. Investigations in the clay-electrolyte-w ater system . K ristiania, (D ybwad). 110 p. O slo. 125 p. N orske vitenskapsakadem i i O slo. Skrifter. M at.- N orw egian G eotechnical Institute, Publ. , 9. naturvitenskapelig klasse, 1924, N o. 14.

540