Analysis of the occurrence and importance of slot development (grykes) within shallow dolomite zones...

NICOLE TROLLIP, LOUIS VAN ROOY AND PATRICK ERIKSSON 333

IntroductionAn estimated 2.5 million people live on dolomite and inexcess of 1.2 billion South African Rands of propertydamage has been observed to date (Buttrick et al., 1995)and in excess of 800 sinkholes have occurred in theCenturion area alone (Department Public Works, 2003).A better understanding of the geology and dolomitestability of sites to be developed is hence crucial in thequest for releasing land for development that is deemedsafe from a dolomite risk perspective.

A geological setting often considered to beparticularly problematic is shallow dolomite. This ismainly due to the fact that the overburden is thin andsmaller quantities of water are required to mobilize thesoil and result in instability at surface. Furthermore,shallow dolomite often outcrops at surface, sometimesas actual rock masses, but often as isolated pinnaclesand large floaters (detached from main bedrock). This may be deceptive as structures may be founded onloose boulders and mistaken to be stable rock.Associated with dolomite is the insoluble weatheringproduct of dolomite rock, referred to as WAD in theSouth African context. This material is highlycompressible and often highly mobilizable (can erodevertically) and presents particularly difficult conditionsto found on. This paper presents an analysis of theoccurrence and importance of slot development

(the South African terminology refers to these features asgrykes) within shallow dolomite in a carefully selectedtype area on the Eccles Formation.

Karst developmentKarst development, in particular on the Eccles Formationof the Malmani Subgroup of the Chuniespoort Group,Transvaal Supergroup, has lead to a highly variabledolomite and chert bedrock topography (Figure 1), withdeep weathering having taken place along linearfeatures such as fractures.

Bedrock is generally deep on the gravity lowanomalies and shallow on the gravity high anomalies.Although the presence of thick and continuous Karoomaterial was initially thought to exist, boreholeinformation together with the bedrock topography mapand careful construction of sections such as these, ruledout the presence of large valleys in-filled with Karoo agematerial and instead present infill material more likely tobe much older (compact chert, locally brecciated).Initially the shallow dolomite in the south-western partof the study area was considered to be a near-continuous mass of rock. More detailed work revealed this only to be the case in isolated pockets,weathering most likely having been influenced greatlyby joint patterns. Where fracturing of the rock is moreintense, deeper leaching is likely to have occurred.

Analysis of the occurrence and importance of slot development(grykes) within shallow dolomite zones in a selected

type area on the Eccles Formation

Nicole TrollipCouncil for Geoscience, 280 Pretoria Road, Silverton, Pretoria, 0001, South Africa

Present address: VGIConsult, P.O. Box 604 Fourways, Johannesburg, 2055, South Africae-mail: [email protected]

Louis van Rooy and Patrick ErikssonDepartment of Geology, University of Pretoria, Pretoria, 0002, South Africa

e-mail: [email protected]; [email protected]

© 2008 September Geological Society of South Africa

ABSTRACT

A better understanding of the geology and dolomite stability of sites to be developed is crucial in the quest for releasing land for

development that is deemed safe from a dolomite risk perspective. A geological setting often considered to be particularly

problematic is shallow dolomite bedrock. This paper presents an analysis of the occurrence and importance of slot development

(grykes) within shallow dolomite in a selected type area.

Geophysical exploration forms the initial phase of a ground stability investigation. A Bouguer gravity anomaly contour map

was produced from the data set for the type area. The interpretation of the residual gravity map was used as the basis for planning

the subsequent phase of investigation, in which further rotary percussion borehole drilling was carried out. Over 750 boreholes

and trench-mapping results were available to test the interpretation against. Various attempts showed just how problematic it is to

produce a gravity map that represents a good estimated depth to dolomite bedrock. The shallow dolomite areas were firstly

delineated using only the residual gravity results and incorporate the gravity high plateau where the average bedrock head is

predicted to be shallower than 8 m. An analysis is given of the conditions related to these shallow bedrock areas.

Shallow dolomite areas associated with some stratigraphic horizons are notorious for instability, however, to the layperson the

presence of rock on a site signifies stable conditions to build on. Upon closer investigation dolomitic rock is found to present

problematic founding conditions, which are not always easy to overcome by engineering design.

SOUTH AFRICAN JOURNAL OF GEOLOGY, 2008, VOLUME 111 PAGE 333-344

doi:10.2113/gssajg.111.2-3.333

SOUTH AFRICAN JOURNAL OF GEOLOGY

ANALYSIS OF THE OCCURRENCE AND IMPORTANCE OF SLOT DEVELOPMENT, ECCLES FORMATION334

Figure 1. Idealized Cross-Sections constructed using borehole information and observations at surface of the study area. Note the

correlation between bedrock topography and gravity anomaly in Section A-B.

NICOLE TROLLIP, LOUIS VAN ROOY AND PATRICK ERIKSSON


335

Fig

ure

2: C

once

ptua

l dia

gram

of t

ypic

al k

arst

land

scap

e in

Sou

th A

fric

a (a

fter

Wal

tham

and

Foo

ks; 2

003)

.

Figure 2. Conceptual diagram of typical karst landscape in South Africa (after Waltham and Fookes; 2003).

Co

nce

ptu

al d

iag

ram

of

kars

t la

nd

scap

e (a

fter

Wal

tham

an

d F

oo

ks)

Cross-sections E-F and C-D present idealized profiles inrelation to inherent risk classes deemed appropriate,based on individual boreholes.

The weathering process of dolomite is well-summarised in the Guideline for engineering geologicalcharacterisation and development of dolomite land(Council for Geoscience; 2003):

Rain water (H2O) takes up carbon dioxide (CO2) inthe atmosphere and soil (where the concentration of thisgas may be up to 90 times greater than in theatmosphere) to form a weak carbonic acid (H2CO3).The weakly-acidic groundwater circulating along tensionfractures, faults and joints in the dolomitic successioncauses leaching of the carbonate minerals. The solubilityof dolomite is high in comparison to other rocks, butsignificant solution cannot be observed over shortperiods (months and years).

This process may be represented as follows:

CaMg(CO3)2 + 2 H2CO3 Æ Ca(HCO3)2 + Mg(HCO3)2

The process of dissolution progresses slowly in the slightly acidic groundwater (above and at thegroundwater level). The resultant bicarbonate-rich wateremerges at springs and is carried away.

This process of dissolution has resulted in a verticallyzoned succession of residual products, which in turn aregenerally overlain by geologically younger formations orsoils (Figure 2).

Hard, unweathered dolomitic bedrock is overlain byslightly weathered jointed bedrock and thereafter,through a sudden, dramatic transition, passes upwardsto totally weathered and low strength, insoluble residualmaterial consisting of mainly manganese oxides (wad),



Figure 3. 1996 Bouguer anomaly map of the study area. The red areas represent areas of gravity high where dolomite bedrock is shallow,

i.e. bedrock at and near surface, the blue areas represent areas of gravity low where dolomite bedrock is deep and the yellow and green

areas represent transitional areas where bedrock depths vary, steepest bedrock gradients are anticipated where the contour lines are closely

spaced. (Van der Merwe, 1996).



337

chert and iron oxides, that reflect the original insolublematrix structure. Depending upon the local subsurfacestructure, this very low strength, porous and permeablehorizon may in certain locations be up to several tens ofmetres thick but is generally less than 10 m thick. Withthe passage of geological time, concurrently with thedownward progression of the intense weathering of thedolomitic bedrock, compaction by the mass of theoverlying materials has resulted in a progressivedensification of these low strength materials.Consequently, the vertical succession of the residualproducts of weathering, reflect an upward increase instrength and a decrease in porosity and permeability.This process results in a decrease in overburden qualitywith depth, which in turn leads to higher rates ofpenetration, so often noted in drilling investigations,when the dolomitic bedrock is approached. Infiltratingwater from leaking services or surface accumulationsacting on this low-density material, results in a loss ofsupport through slumping or subsurface erosion.

Given sufficient time and the correct triggeringmechanisms, instability may occur naturally but isexpedited many orders of magnitude by man’s activities.The primary triggering mechanisms in such instancesinclude the ingress of water from leaking water-bearingservices, poorly managed surface water drainage andgroundwater level drawdown. Instability can occur inthe form of sinkholes and dolines (local term denotingsurface subsidence due to compaction, under load orlowering of groundwater, of low density residualdolomite). Topography and drainage, the naturalthickness and origin of the transported soils andresiduum, the nature and topography of the underlyingstrata, the depth and expected fluctuations of thegroundwater level, and the presence of structuralfeatures such as faults, fractures and dykes are all factors which influence the risk of subsidence takingplace.

Figure 4. Residual gravity field calculated by subtracting the third order polynomial surface.

Waltham and Fookes karst typesWaltham and Fookes (2003) proposed an engineeringclassification of karst based on the assessment of karstworld-wide. They were of the opinion that aclassification that identifies certain essential parametersthat influence dolomitic ground conditions and thedegree to which these are present, is useful to the civilengineer when faced with the task of recommendingengineering solutions. The classification has as itsparameters: sinkhole frequency, rockhead variability (or bedrock topography) and sizes of undergroundcaves. Five classes are defined: juvenile, youthful,mature, complex and extreme karst. Of interest in termsof bedrock topography is that areas where pinnacleddolomite exists with a relief of 5 to 20 m is termedcomplex karst, and areas where pinnacled dolomiteexists with a relief of greater than 20 m with loose pillarsundercut between deep soil fissures, are termed extremekarst. Such karst is often encountered in Gauteng,

especially on the Eccles Formation. The challenge inapplying such classification in South Africa is todetermine which karst types are truly present on a site,as almost all local karst is mantled and masked by soilhorizons. Some adaptations are needed to themorphological features depicted in the extreme andcomplex karst type schematic diagrams to better reflectgeneral local conditions in Gauteng. This will involvemuch further study prior to adopting such classificationterminology in the South African context. Figure 2 showsan attempted and generalised adaptation for the typicalmorphological features that can be expected.

Investigation of dolomitic landGeophysical exploration forms the initial phase of aground stability investigation. No reliable geophysicaltechnique exists that can accurately gauge the presenceand size of voids in the bedrock or overburden, so the purpose of this survey is primarily aimed at



Figure 5. Difference between the depth to bedrock estimated from the polynomial surface fitting method and actual depth to dolomite.

Points in red and orange show discrepancies of more than 21 m (between 21 and 99 m).



339

determination of the dolomite bedrock topography andthickness and density of overburden. The gravitymethod is the most successful geophysical tool for thispurpose.

Gravity surveys are relatively easy to carry out andare cost effective. The interpretation of gravity data ismore difficult because different mass distributions canmatch a single anomaly. For this reason it is essential toinvestigate the site by drilling, as this information can beused to better constrain the gravity model.

In this method, readings are taken with a gravimeterat each station of a grid surveyed for the site. The gridspacing is a function of the type of anomaly expected,the depth to the source of the anomaly, and the size ofthe investigated area.

Bouguer gravity map A Bouguer gravity anomaly contour map was producedfrom the data set for the type area (Figure 3).

This map was carefully studied to determinepositions that should be investigated further by rotarypercussion boreholes. The purpose of the exercise wastwofold:a. to determine the overburden and bedrock conditions

within gravity-high, gravity-low and -transition zonesand

b. to better define and refine the gravity map. The latterwas achieved by targeting gravity high anomaliesand using the depth to dolomite bedrock to producea residual gravity map (Figure 4).

Residual gravity mapOnce depth to dolomite bedrock on the Bouguer gravityhigh anomalies could be confirmed by drilling, theBouguer field was adjusted by subtracting a regionalfield so that the map becomes a better representation ofdepth to dolomite bedrock. An abundance of boreholeinformation in this type area enabled the determinationof quite a detailed regional field.

The interpretation of the residual gravity map wasused as the basis for planning the subsequent phase ofinvestigation, in which further rotary percussionborehole drilling was carried out. Areas of relativelyshallow bedrock are represented as gravity highs, andareas of relatively deep bedrock are represented asgravity lows. Where dolomite is shallow, suddenchanges in bedrock topography such as grykes(solution-enlarged vertical joints which form slots in thebedrock) and pinnacles, may have a profound effect onground stability conditions. These relatively smallfeatures, however, are very difficult to identify using thegravity method, and they are often not discovered.

Karoo material

Chert andminor wad

BH 1=e.g. risk class 3(a)

BH 2=e.g. risk class 6

BH 3=e.g. risk class 3(b)/6

BH 4=e.g. risk class 6/7

BH 5=e.g. Risk class 4

Figure 6. Diagrammatic representation of gravity anomalies associated with type geological settings and possible risk classes.

Interpretation of the gravity surveyOver 750 boreholes and trench-mapping results wereavailable to test the interpretation against. Variousattempts showed just how problematic it is to produce agravity map that represents a good estimated depth todolomite bedrock (i.e. one where the actual andestimated depth to bedrock depths are very similar orrarely differ greatly). The residual gravity map presentsa north-west - south-east trending gravity low with apronounced signature. Within the shallow dolomiteareas of the northern half of the site three shallowertroughs at the same orientation can be discerned.The shallow dolomite found in the western part isrevealed as a gravity high area.

Determination of a regional gravity field is a veryambiguous exercise because it is based on variousassumptions. Cole (2005) indicates that the polynomialsurface method assumes that the regional field changessmoothly over the study area. Cole (2005) calculated

depth to bedrock using the results and compared theseto actual depths to dolomite intersected by a number ofboreholes in the study area. Figure 5 shows thedifferences in metres between the estimated and actualdepths to bedrock.

Upon closer inspection of the points in Figure 5, it becomes apparent that points that show the greatestdiscrepancies (red and orange dots) are revealed asboreholes that have been drilled on steep gravitygradients. Here depths to bedrock ranges are anticipatedto be great and hence it is not necessarily an indicationthat the survey is inaccurate. Furthermore, it is possibleto get large variations in depth in shallow bedrock areasas boreholes may be drilled into grykes, which are toonarrow to manifest as an anomaly (narrower than thestation spacing, in which case the gravity meter cannotdiscern a difference in density). Figure 6 represents amodel of the dolomitic bedrock within the gravityfeatures to illustrate this.



Areas excluded due

to poor spread of

borehole information

Smaller areas of shallow

dolomite excluded in the

final zonation

Figure 7. Preliminary Zonation map based on isopach map of estimated depth to bedrock. Bedrock anticipated to be shallower than 8 m

in blue and deeper than 8 m in red.

NICOLE TROLLIP, LOUIS VAN ROOY AND PATRICK ERIKSSON 341


Figure 8. Idealized diagram depicting typical shallow dolomite area incised with grykes and photograph as an example.

Phote of an old sinkhole in a gryke within a shallow dolomite area

Diagram depicting shallow dolomite incised by grykes



Do

lo

mite

b

ed

ro

ck

So

il fille

d g

ryke

Figure 9. Detailed mapping of a property exhibiting shallow dolomite with extensive outcrop. (Courtesy Relly Milner and Shedden).



343

Delineation and characterisation of shallowdolomite areasThe shallow dolomite areas were firstly delineated usingonly the residual gravity results and incorporate thegravity high plateau where the average bedrock head ispredicted to be shallower than 8 m (Figure 7).

The distinction between deeper than and shallowerthan 8 m bedrock depths has been useful in identifyingthe potential shallow dolomite bedrock areas. It wasdecided to ignore smaller zones of shallow dolomite.Two areas presented too few boreholes to confirm thepresence of shallow bedrock and were also excised.

The bedrock head is described as the rock plateau orplane, which can be incised by grykes. Certainboreholes on the gravity high plateau revealextraordinarily deep dolomite bedrock (20 m or more).These are interpreted to represent relatively narrowgrykes (Figure 1). These features would not normally bediscernable on the gravity map and are probably linearfeatures representing fractures and faults in the bedrockalong which preferential dissolution took place. The dimensions of these features are almost impossibleto determine and in all probability vary over short lateraland vertical distances. As these features are never thesubject of investigations, in many instances when poorconditions were struck, the investigator elected toterminate drilling (such a location is deemed unsuitablefor placement of a structure and there is then no pointin drilling further). This reality makes it impossible forthe shallow dolomite areas to be relatively accuratelymodelled (in order to define bedrock topography).Often when poor conditions are encountered, a highrisk of any size sinkhole is assigned to the borehole.However, this is not necessarily true as the size is almostexclusively constrained by the width of the gryke in veryshallow dolomite areas (certainly where bedrock is lessthan 2 m deep). This is illustrated in Figure 8.

In some shallow dolomite areas, for example wherebedrock outcrops extensively at surface, it is possible bymeans of an excavator to excavate the soil in thesuspected grykes and map the dimensions of the grykes.An example of such an exercise is presented in Figure 9.

Careful inspection of the orientation (dip and strike)of the bedrock ensures that floaters can be distinguishedfrom true bedrock.

Two cross-sections were drawn (Figure 1) on areasof known shallow dolomite in the selected type area,one in an east-west direction and another in a north-south direction. These cross-sections clearly show thepresence of grykes, although their widths and depths aresometimes speculative. Figure 8 presents a diagrammaticrepresentation of a shallow dolomite area, drawn frominformation obtained from the cross-sections as well asvisual observations of sinkholes that have occurred insuch areas. A photo of a sinkhole which exposed thedimensions of a gryke has been included for clarity anda three dimensional perspective. The cross-sectionsseem to suggest that grykes do not have one preferredorientation, in fact there are at least two preferredorientations discernable.

Many boreholes intersected dolomite rock at shallowdepth (<8 m) with an accompanying thin overburden. In some instances the overburden is so thin (<2 m) thatlimited excavation can remove the entire overburdenand reveals dolomite bedrock as slabs, floaters andpinnacle heads. Shallow dolomite areas in specificstratigraphic horizons, such as the Eccles Formation, arenotorious for instability. This is mainly due to twofactors- only a limited quantity of mobilization agent isneeded to mobilize the thin overburden and triggerinstability; the presence of grykes infilled by high-mobilization material, in turn blanketed by a thin veneerof overburden, also presents conditions conducive toinstability. The overburden typically consists of residualdolomite of varying wad content, fines with varyingparticle sizes (although rarely with high percentages ofclay) and chert fragments also of varying sizes.The upper part of the overburden seems to be morecompact than that nearer to dolomite bedrock (as deduced from drill penetration times). The conditions above dolomite bedrock are also oftenpoor, exhibiting more porous conditions as reflected by percussion drilling penetration times of under15 seconds per metre as well as sample loss and air loss.

Table 1. Typical site conditions for inherent risk low, medium and high (Buttrick et al., 1995)

Inherent risk Typical site conditions

Low The profile displays no voids. No air loss or sample loss is recorded during drilling operations. Either a very shallow

water table or a substantial horizon of materials with a low potential susceptibility to mobilisation may be present

within the blanketing layer (e.g. continuous intrusive features or shale material). Depth to potential receptacle is

typically great and the nature of the blanketing layer is not conducive to mobilisation.

Medium This type of profile is characterised by an absence of substantial ‘protective’ horizon and has a blanketing layer of

materials potentially susceptible to mobilisation by extraneous mobilisation agencies. The water table is below the

blanketing layer.

High The blanketing layer of the high risk profile reflects a great susceptibility to mobilisation. A void may be present and

is interpreted to be very likely, within the potential development space, indicating that the process of sinkhole

formation has already started. Boreholes may register large cavities, sample loss, air loss, etc. Convincing evidence

exists of cavernous subsurface conditions which will act as receptacles. The water table is below the blanketing layer.

In a dewatering situation, the lowering of a shallow groundwater level would obviously increase the risk of

mobilisation.

Where the so-called poor conditions thicken (>4 m) itappears that grykes containing dolomite residuum havebeen intersected (Figure 1). This interpretation is basedon the deduction that the residuum has been isolatedwithin the gryke from the elements by the surroundinghard rock and has not been reworked and compacted.This type of overburden is thus characterised by anabsence of a substantial ‘protective’ horizon andcontains materials susceptible to mobilisation bymobilisation agencies. Low-density material, sometimesaccompanied by cavernous conditions in theoverburden reflects a great susceptibility to mobilizationand indicates that the process of sinkhole formationcould already be underway. Convincing evidence oftenexists of cavernous subsurface conditions which will actas receptacles. The overburden is hence characterized as having a medium to high mobilization potential(Table 1).

ConclusionTo the layperson the presence of rock on a site signifiesstable conditions to build on. However upon closerinvestigation dolomitic rock presents problematicfounding conditions which are not always easy toovercome by engineering design. A better understandingof the weathering and karstification process of dolomite,together with detailed site investigation by means ofdrilling and careful evaluation of the results can aidgreatly in ensuring that development is placed ongeologically stable areas. Where structures cannot avoid

shallow dolomite areas, grykes can be identified andappropriate ground improvement proposed and suitablefoundation and structural solutions implemented.

ReferencesButtrick, D.B., and van Schalkwyk A. (1995). The method of scenario

supposition for stability evaluation of sites on dolomitic land in South

Africa. South African Institution of Civil Engineering Journal,. Fourth

Quarter 1995, 9-14.

Council For Geoscience/South African Institute of Engineering and

Environmental Geologists (2003). Guideline for engineering-geological

characterisation and development of dolomitic land. Council for

Geoscience, 66pp.

Council For Geoscience (2005). Gravity Interpretation of the Cornwall Hill

gravity survey information. Council for Geoscience Internal Report.

Council For Geoscience (2006). Determining a regional field for gravity data

collected at Cornwall Hill, Pretoria. Council for Geoscience Internal Report

2006-0178, 10pp.

Department of Public Works (2003). Appropriate Development of

Infrastructure on dolomite: Guidelines for Consultants., Department

of Public Works of South Africa, PW344. 163pp.

Trollip, N.Y.G. (2006). The geology of an area south of Pretoria with specific

reference to dolomite stability. Unpublished MSc dissertation. University of

Pretoria, South Africa, 87pp.

Van der Merwe, J. (1996). Report on the dolomitic stability and foundation

investigation carried out for the proposed Cornwall Hill development to be

situated on portions 1 and 5 Doornkloof 391-JR, Centurion, Gauteng.

(Unpublished).

Waltjam, A.C. and Fookes, P.G. (2003). Engineering Classification of karst

ground conditions. Quarterly Journal of Engineering Geology and

Hydrogeology. 36, 101-118.

Editorial handling: A Bumby



Analysis of the occurrence and importance of slot development (grykes) within shallow dolomite zones...

Documents

Transcript of Analysis of the occurrence and importance of slot development (grykes) within shallow dolomite zones...