Figure 6.28 Spectra-Physics EL-1 (Spectra-Physics)

this case forms a 360° horizontal plane which is detected by aportable sensing device also shown in Figure 6.28. The laser unitautomatically corrects for any error in level of the instrument,providing it has been roughly levelled to within 8° of the vertical.An accuracy of ± 5 to 6 mm per 100 m up to a maximum rangeof 300 m can be achieved with this type of instrument.

6.2.3.5 Collimation error

So far the assumption has been made that once the standing axisof the level has been set truly vertical, then the line of collima-tion will be horizontal. This may not always be the case.

If this condition does not occur, then a collimation error issaid to exist. This is illustrated in Figure 6.29. If accuratelevelling is to be achieved, it is essential that a regular testingprocedure is established in order to check the magnitude of anycollimation error that may exist.

Figure 6.30 Two-peg test

the difference between the two readings calculated. Thisvalue represents the true difference in height between A andB. Any collimation error which exists will have an equaleffect on both readings and, hence, will not affect thedifference between the readings. In this case the difference inheight is 1.415-0.932 = 0.483m.

(3) The instrument is now moved to a point C close to the staffat B (about 3 to 5 m away), as in Figure 6.30(b). The readingon staff B is recorded (1.301). If no collimation error exists,the reading on staff A should be equal to the reading on staffB ± the true difference in height as established in (2), i.e.1.301+0.483= 1.784m.

(4) The actual observed reading on staff A is now recorded(1.794). Any discrepancy between this value and that der-ived previously in (3) indicates the magnitude and directionof any collimation error. For example, in this case, the errorwould be 1.794- 1.784= 10mm per 50m.

An error of up to 2 to 3 mm over this distance would beacceptable. If, however, the error is greater than this, theinstrument should be adjusted. Unlike theodolite adjustments,this type of adjustment can normally be performed without anygreat difficulty by the engineer and the procedure is as follows.

For the dumpy and automatic level: alter the position of thecross-hairs until the centre cross-hair is reading the value whichshould have been observed from step (3) above. This is achievedby loosening the small screws around the eyepiece which controlthe position of the cross-hairs.

For the tilting level: again alter the position of the centrecross-hair until it is reading the value previously determined in(3), in this case by tilting the telescope using the tilting-screw.Unfortunately, this will displace the bubble. The bubble must,therefore, be centralized by means of the bubble-adjustingscrew.

6.3 Surveying methods

6.3.1 Horizontal control surveys

Any engineering survey or setting-out project, regardless of itssize, requires a control framework of known co-ordinatedpoints. Several different control methods are available as des-

Figure 6.29 Collimation error

A common field procedure which can be used to test a level isknown as the 'two-peg test'. The procedure is as follows:

(1) Set out two points A and B approximately 50m apart, asshown in Figure 6.30(a). The level is set up at the mid-pointof AB and levelled as in section 6.2.1.3.

(2) A reading is taken on to a staff held at points A and B and

Line of collimationHorizontal plane

Collimation error

cribed below. The choice of which method to use depends onmany factors, e.g. the purpose for which the control is required,the accuracy required, the density of control points which isrequired, the type of equipment and computing facilities whichare available and, lastly, the physical nature of the ground.

6.3.1.1 Triangulation

Triangulation is the oldest, and in the past was the mostcommon, method of control for large civil engineering projects.The principles are well known and essentially involve. theestablishment of a measured baseline from which a network oftriangles is formed, all of the angles of the triangle beingmeasured. The development of EDM has, however, led to theestablishment of several alternative control methods, such astrilateration and traversing. The introduction of EDM hastherefore tended to make 'classical triangulation' obsolete as amethod of control.

63.1.2 Trilateration

Trilateration is a method of establishing control which involvesthe direct measurement, normally using EDM, of all the sides ofa network of triangles, in contrast to triangulation whichinvolves the measurement of angles. Although the method hasbeen used in this 'classical form', it does not offer any significantadvantages over the method of triangulation. The method hastherefore not become particularly common.

6.3.1.3 Traversing

A traverse is a method of establishing control by measuring thedistance between successive points and also the horizontal anglebetween adjacent stations, as shown in Figure 6.31.

Figure 6.31 Traversing

The method is very popular for several reasons. Firstly, it is amore flexible method than triangulation. In the case of triangu-lation, the positions of the control stations must be chosen sothat not only are they intervisible, but also that the trianglesformed are well conditioned. For this reason the reconnaissancestage in triangulation projects is extremely important and oftenvery time-consuming. In contrast, with traversing, much lessattention has to be paid to the reconnaissance stage, since it isonly required that adjacent stations be intervisible. This allowsthe surveyor much greater flexibility in the choice of controlstation positions. The stations can then be positioned in areasclose to the detail to be picked up, or close to the project forwhich they are required. This can be of enormous benefit inareas which are either very flat or, alternatively, heavily for-ested. A second reason for the popularity of traversing is thecomputational simplicity, both of determining provisional co-ordinates and also of adjusting any misclosure which may exist.It should, however, be mentioned that in recent years morerigorous techniques of adjustment, based on the principle ofleast squares, have become more common for the adjustment oftraverse.

Traversing does, however, suffer from one serious drawback:the lack of redundant observational data. As a consequence, theeffect of small errors of measurement is not only difficult todetect but is also cumulative in nature. To counteract thisproblem, additional angle and distance observations are oftentaken in order to strengthen the control framework. In the past,this additional information tended to be used solely for thedetection of gross errors. Nowadays, by using a suitable adjust-ment technique, these additional observations can be used toimprove the precision of the coordinates.

The normal procedure adopted for traverse adjustment isfirstly to determine and adjust the angular misclosure and,secondly, to determine and adjust the misclosure in easting andnorthing. The angular misclosure is determined, in the case of aclosed polygon, by summing the internal angles. These shouldtotal (2n - 4) right angles, where n is the number of traversesides. A misclosure of >(20'VW)» for example, would not beacceptable for site traverses.

The question of which method to use for the adjustment ofany misclosure in eastings and northings is a matter which hasbeen examined by Schofield.19 Traditionally, the Bowditch andTransit methods have been used.

Bowditch: zf£c=ME-^and ANc=M»~j

Transit: AE , ... AN*E=M*L\JE\Md AN< = M"Z\AN\ (6.22)

where ME, A/N, is the misclosure in easting, northing, d is thelength of a traverse leg, Z\AE\, Z\AN\ is the absolute sum ofthe provisional AE, AN, and AEC, ANC is the correction to beapplied to the provisional AE, AN.

For most small site traverses observed using a theodolite andsteel tape both techniques will give acceptable results. Scho-field,19 however, also discussed the problems which arise whensemi-rigorous methods of adjustment such as Bowditch andTransit are used to adjust modern EDM traverses. The mainconclusion reached is that both methods are based on assump-tions which are not applicable to EDM. For example, it isassumed that the expected error in an EDM measurement isproportional to the distance measured, which is clearly not true.It is suggested that all EDM-based traverses should be adjustedby a rigorous method such as variation of coordinates.

6.3.1.4 Survey networks

The combination of angle, distance and orientation measure-ments to form a control framework is now commonly referredto as a survey network. The advantages of such an approach areconsiderable. Firstly, the scale and orientation errors associatedwith classical triangulation can be reduced by the inclusion ofadditional distance and azimuth measurements. Secondly, thecontrol framework does not suffer from the serious propagationof errors which can occur in traversing. Thirdly, the optimumnumber of observations (angular and distance) can be deter-mined before the fieldwork commences, using computer simula-tion techniques. Finally, by using the method of variation ofcoordinates a least squares procedure can be used to determinethe most probable values of the coordinates.

The principle of least squares states that the most probablevalue of a sample is that for which the sum of the squares of theresiduals is a minimum. Variation of coordinates is a computa-tional method, based on this principle, which is used for thedetermination of coordinates in a survey network.

Figure 6.32 Adjustment of a braced quadrilateral by variation ofcoordinates

Consider Figure 6.32, which illustrates a small network inwhich all the eight angles a°l to 8 and all distances /0I to 6 havebeen observed. Point A is fixed, the bearing to point D is fixed inorder to orient the network and the a priori standard errors ofthe observations are estimated. There are, therefore, fifteenobservations and six unknowns to be determined, i.e. theeastings and northings of points B, C and D. The method ofsolution is as follows: (1) assume provisional coordinates for allpoints. These may either be scaled graphically from a plan orcomputed using a selection of the measurements; (2) using theprovisional coordinates, compute values for the angles O6I to 8and the distance /cl to 6; and (3) set up an observation equationfor each individual observation.For a distance ^:

- cos ^dTV4 - sin fc.AEt + cos p.£N-} + sin P1-AE-= (/;-/S>+v (6-23)

or

PdTV1 - QAE- + RdTVj + Sd^ = (/; - /§) + v (6.24)

For bearing

^N-^dE-^dNi+^dEsAj Aj Aj Aj

= foS-tfj) + v (6.25)

For angle <xijk:

< j

k

/sinA.rinM + /-00.A + Cg^A d£

\ Ak Aj / V Ak Aj /

+ sin41 COS^1 smA^ + cos^

k k 'ik 'ik

-(<%- «y+ v (6.26)

For a position

ANi = (Nf-N?)+ v (6.27)

AE1 = (Ef-E?)+v (6.28)

where p;j = direction of line ij, dN{, d^E^the unknowns, thecorrections to the provisional coordinates, and v = residual.

These observation equations can be expressed in matrix form:

A X = b + V (6.29)

F Matrix of ~| * |~ Vector of 1'T w 4 f T [ ~ v , r 1'coefficients unknowns Vector of Vector of I

L J L J L o-c terms J L residuals J15 6 15 15

The solution is found by forming the normal equations.

A*WA\ = A*Wb (6.30)

where W is the weight matrix, the diagonal elements of whichare equal to 1/cr2 where o refers to the a priori standard error ofthe observation.

The normal equations may then be solved using a Choleski'striangular decomposition method or by matrix inversion.

\ = (A*WA)-} AJWb (6.31)

The column vector X therefore contains the corrections (AN-,AE) to the provisional coordinates of the unknown points.Using these new values for the coordinates, the entirecomputational procedure is repeated until there is no furtherchange in the coordinates.

Usually when this method is used a complete error analysis ofthe results is carried out. By this process information about theprecision and reliability of the coordinates can be obtained. Forfurther reading on this aspect of the method see Ashkenazi,20'21

Ashkenazi et a/.,22-23 and section 3.4.2.

6.3.2 Detail surveys

After the main control survey has been observed and computed,a detail survey is carried out in order to locate the positions offeatures which are to be presented on the map.

6.3.2,1 Tacheometry

The most common method of carrying out a detail survey byground methods at scales smaller than 1:500 is that of stadiatacheometry. The basic principles have been outlined in section6.2.2.2.

In practice, the observational and booking procedure shouldbe as follows. The theodolite should be set up and levelled overthe point from which observations are to be taken. This pointmay coincide with one of the main control survey stations, ormore commonly be one of a series of subsidiary control stationswhich have been established by traversing from the maincontrol. By this process not only are any errors contained withinthe control network, but also the positions of the subsidiarycontrol will be closer to the detail which is to be surveyed.

The circle of the theodolite should then be oriented until ahorizontal circle reading of 0° coincides with the direction to anadjacent control point. This is known as the reference object(RO). This procedure simplifies the subsequent plotting of thefield observations. The height of instrument should be recorded.

The staffman is then directed to the various points which areto be mapped. The staff is held vertical over the point and thestadia readings, vertical angle and horizontal circle reading tothe point are observed from the theodolite. Several techniquesfor speeding-up the field recording of the stadia hair readings

have been advocated.24 Assuming that a pocket calculator,preferably programmable, will be used to reduce the observa-tions, the most efficient field method is to set the bottom hair tothe 1 m or 2 m point of the staff. Using this method, the mentalcalculation of the stadia intercept becomes much quicker.

The booking procedure for stadia observations is illustratedby Table 6.3. It is also essential to draw, whilst in the field, agood sketch-map of the area being surveyed. This is invaluablewhen the results are being plotted.

6.3.2.2 Chaining

This is the simplest form of detail surveying. The methodinvolves measuring the lengths of the sides of a series of trianglesor braced quadrilaterals. Points of detail are then picked-up bymeasuring offsets from these lines. The procedure is illustratedin Figure 6.33.

Figure 6.33 Chaining

In this case, two traverse stations from an existing controlsurvey are within the boundary of the site which is to besurveyed.

The first stage in the survey involves breaking-down thisexisting control into either well-conditioned triangles or bracedquadrilaterals. In this case, a braced quadrilateral is chosen. Theinclusion of diagonal measurements provides an independentfield check on the measurements. These 'chain lines' are nor-mally measured with a steel band.

The second stage in the survey process involves the measure-ment of offset or ties from these chain lines in order to pick-up

Figure 6.34 Offsets and ties

the survey detail. Figure 6.34 illustrates the distinction betweenoffsets and ties.

Ties, which involve two measurements to one point, arenormally employed when the offset distance is long and, hence,the accuracy with which the perpendicular to the chain line canbe set out is low. Offset and tie measurements are normallymade with a fibreglass tape.

A standard method of booking is conventionally adopted forchain surveying. Details of this and other points relating tochain surveying can be found in any of the standard surveyingtextbooks listed in the bibliography.

6.3.3 Vertical control surveys

In general, all civil engineering projects require not only plani-metric control, but also vertical or height control points. Thetwo most common methods of obtaining this height control areby levelling and by trigonometrical heighting.

6.3.3.1 Levelling

Levelling is the name given to the process of determining, bymeans of a surveyor's level, the height of a point above somedatum, normally mean sea-level.

As is the case with horizontal control, it is important to workfrom 'the whole to the part'. It is therefore normal practice todesign a levelling control framework in a hierarchical manner,in order to contain small errors. Typical maximum allowablemisclosures for level loops in one such hierarchical design areshown in Table 6.4.

The basic principle of levelling involves taking horizontalbacksight (BS), foresight (FS) and intermediate sight (IS) read-ings, as defined by the line of collimation of the level, on tovertical staves as shown in Figure 6.35. The difference betweensuccessive readings indicates the difference in height betweenpoints. By this process it is therefore possible to determine thereduced level (RL) of a series of points.

Offsets

Ties

Table 6.3 Booking and reduction of tacheometric readings

Point

A(RO)1

Observer:Booker:Date:

Horizontalcircle

O0OO'10° 30'

TJMKDRG22/9/85

Staff readingsUML

1.4301.2151.000

Staff intercept(S)

0.430

Station BHeight of Instrument: 1 .690 mReduced Level: 58.35m

Vertical circle

910OO'

Horizontaldistance (Z)H)(100 cos2 6)

42.99

V(50^sin 2(9)

0.75

Reduced level

59.58

OBMRL = 51 m

Figure 6.35 Levelling

The reduced level of a point is defined as its height abovesome datum. In the UK, the fundamental datum established bythe Ordnance Survey is mean sea-level at Newlyn, Cornwall,known as Ordnance Datum. A further series of points of knownheight have been established throughout the UK and these areknown as Ordnance bench-marks (OBM).

Table 6.5 Rise and fall method of booking

The results of a levelling operation are, by convention,booked in a standard manner. Two methods of booking areused. The rise and fall method (Table 6.5) is usually employedwhen running lines of levels between bench-marks in order toestablish additional supplementary height points. The height ofcollimation method (Table 6.6), on the other hand, is moresuitable for tasks such as recording cross-sectional information,in which many intermediate sights have been taken.

Contouring. A contour is an imaginary line joining points ofequal elevation. The level can be used in a variety of ways forcontouring. One of the simplest methods is by means of gridlevelling.

With this technique, the area to be contoured is covered by animaginary grid of lines forming squares of 10, 20 or 30m. Thelevel is set-up in a central position and levels are then taken to asite temporary bench-mark (TBM) and at the intersections ofthe grid lines, as shown in Figure 6.36.

Contours may then be interpolated either graphically ormathematically from the grid of levels.

Quantity determination. The method of grid levelling providesa convenient means of determining earthwork quantities of, forexample, borrow pits.

The depth of cut (h) at each intersection point is established.This will be equal to the difference between the ground-level andthe proposed formation level of the borrow pit. Figure 6.37illustrates a simple example for a 20-m grid of levels.

The volume of excavation consists of a series of rectangularprisms each having a base area (A), in this case equal to 400 m2.The total volume (V) for the general case is therefore:

V=^(Zh, + 2Zh2 + 3Zh, + 4Zh4) (6.32)

Table 6.4 Accuracy of levelling (K is the distance in kilometres)

Order Maximum allowable misclosure (m)

1st 0.004V^2nd 0.008V#3rd 0.012 ^JK4th 0.024V/:

Checks: ZBS-ZFS = 3.313m ZRISE-TFALL = 3.313m.First RL-last RL = 3.313m

Table 6.6 Height of collimation method of booking

Checks: 27BS-ZFS = 3.313m First RL-last RL = 3.313m*Height of collimation = reduced level + backsightRL = HC-FS or IS

Z

Backsight

2.3451.935

4.280

Intermediatesight

1.213

Foresight

0.632

0.335

0.967

Rise

1.7130.7220.878

3.313

Fall

0.000

Reducedlevel

51.00052.71353.43554.313

Remarks

OBMBC

OBM

Z

Backsight

2.3451.935

4.280

Intermediatesight

1.213

Foresight

0.632

0.335

0.967

*Height ofCollimation(HC)

53.34554.648

Reducedlevel(RL)

51.00052.71353.43554.313

Remarks

OBMBC

OBM

Figure 6.37 Quantity determination using a grid of levels

where /r, = depth of cut used once, H2 = depth of cut used twice,etc.

The volume determination assumes that the ground slopebetween grid intersections is constant. Clearly, therefore, byreducing the size of the grid, the accuracy of the quantitydetermination will be increased.

6.3.3.2 Trigonometrical levelling

For projects where the acceptable accuracy requirements of theheight control are lower than would be obtained by levelling, themethod of trigonometrical heighting is normally used.

The method is based on the measurement of the vertical anglebetween two points by means of a theodolite, together witheither the slope or horizontal distance. By simple geometry, thedifference in height can therefore be calculated. As the distancebetween the points increases, however, the effects of two pheno-mena, Earth curvature and refraction, become more significant.Figures 6.38 and 6.39 illustrate the highly exaggerated effect ofthese phenomena.

In Figure 6.38, a distant point B' appears too low by anamount Ac. A positive Earth curvature is therefore necessary. Itcan also be seen in Figure 6.39 that the effect of refraction is torefract the line of sight of the telescope so that a distant point C

Figure 6.39 Trigonometrical heighting: atmospheric refraction

appears too high by an amount Ar. A negative refractioncorrection is therefore required.

It can be shown that the difference in height between A and Ccan be determined by the equation:

J/*AC = rftan0 + Cdc-Jr) (6.33)

where Ac-*, = *££& (6 34)

and AT= coefficient of refractions 0.07, R = radius ofEarth = 6378 km, d= horizontal distance, and 9 = vertical angle.

If k = 0.07 and dis in kilometres, then Ac- Ar = OMlSd2m.The results obtained using this simple approach can be

increased significantly by observing vertical angles from bothends of a line. Reciprocal observations of this type, particularlyif they are observed simultaneously, can eliminate completelythe necessity for Earth curvature and refraction corrections. Inthis case, the difference in height between A and C can becomputed by the general expression

Rangingpoles

Figure 6.36 Grid levelling

Ranging poles

Horizontalline of sight

Centre of Earth

Figure 6.38 Trigonometrical heighting: Earth curvature

A. _ d tan 0A + 0C.(/*SA- hTc) -(H^- frjJ*AC~ 2 + 2 (6.35)

where 0A = vertical angle from A to C, Oc = vertical angle from Cto A, /*?A = height of signal at A, A^ = height of signal at C,/ITA = height of theodolite at A, /^0 = height of theodolite at C

6.3.4 Deformation monitoring surveys

Deformation monitoring surveys using conventional land-sur-veying methods, photogrammetric survey methods or specialistgeotechnical methods are used to quantify the amount by whichan engineering structure has moved both vertically and horizon-tally over specific periods of time.

The information which surveys of this type provide can be ofcritical importance and may be used to indicate that either: (1)the structure is stable and consequently safe; (2) the structure isexperiencing small random movements which are not imposingsignificant forces on the structure; (3) the structure is experienc-ing small localized systematic deformations, e.g. as caused byseasonal effects, which may or may not be significant; (4) lastlyand most significantly, the structure is experiencing deformationwhich is increasing as a function of time. This may indicate thatremedial measures have to be taken, in some cases immediately,in order to avoid catastrophic consequences.

For a variety of reasons, surveys of this type have beenincreasingly more important in recent years. Several interrelatedfactors have been responsible for this increase in interest.Firstly, for many types of structures it is now a mandatoryelement of the civil engineering process that a monitoring surveybe commissioned. Notable in this respect in this country are, forexample, reservoirs, which now have to be monitored followingthe implementation of the 1975 Reservoirs Act. A furtherexample, from Switzerland, is the requirement by the SwissFederal Government for all dams with a height greater than15 m and a cubic capacity greater than 50 000 m3 to be moni-tored.25 A second factor which has led to this increase in interesthas been the speed of development both in the manufacture ofprecise survey instrumentation and also in high-speed, low-costcomputing facilities. This, in conjunction with improvements invery elegant and highly sophisticated software for the designand analysis of surveying observations, now provides the sur-veyor and engineer with a highly accurate measurement system.Thirdly, the tolerances to which civil engineers are now design-ing and constructing many modern structures necessitates amuch higher order of accuracy in the initial dimensional controland also in the subsequent deformation monitoring. Indeed, thisfactor may have acted as a 'springboard' for many of thedevelopments in software and instrument design. It is alsohoped that to some extent civil engineers are now more aware ofthe possibilities offered by modern survey techniques and aretherefore requesting this type of survey more frequently thanwas the case in the past.

Most of the instruments which are used for deformationmonitoring projects have been discussed in several of theprevious sections of this chapter. It will therefore be the aim inthis section to concentrate on two other important aspects ofdeformation monitoring surveys. The first is the design ofsuitable reference and monitoring points. Clearly, when verysmall displacements are being measured it is crucial that thepoints from which measurements are being recorded are notsubject to movement. The second aspect which will be discussedis the computational processes associated with the horizontalcontrol networks which are often used to quantify the extent ofany structural deformation. Finally, it should be noted thatdeformation monitoring can often be carried out, in some casesmore efficiently, using close-range photogrammetric techniques.Reference should be made to Chapter 7 for further details.

6.3.4.1 Design of reference and monitoring points

The full accuracy potential currently offered by modern survey-ing equipment for measuring small structural displacements canonly be fully realized if care and attention is paid to the design ofthe reference points from which observations will be taken.Equally important is the need to design appropriate monitoringpoints to be placed on the structure under investigation.

Reference points. Two distinct types of survey reference pointcan be identified. The first, typically a survey pillar, forms thereference framework for the horizontal and vertical controlmeasurements, whilst the second, a steel or concrete pile, formsthe datum for levelling observations.

Scammonden reference monument(b) Survey pillar (Penman and Charles26}

Figure 6.40 Survey pillar designs

CoverKern centring plateSteel casing(450 mm dia.)Reinforcing cage

Reference monument

(a) Survey pillar (Penman and Charles26)

Reinforcement

0.75 m dia.boreholefilled withconcrete

0.46 m i.d.steel tubeBackfill

Gap filled withpetroleum jelly

O.40 m i.d.steel tube

Starter bars

TopsoilWeathered chalkwith flints

WeatheredchalkIntactchalk

Three distinct designs of survey pillar are illustrated in Figure6.40. The main requirement in siting pillars is the need to ensurethat they are founded on stable ground outside the zone ofinfluence of the structure under investigation. It is, however,often difficult to assess whether this is the case before observa-tions are recorded. It is therefore important to incorporate intothe design an insulating gap which ensures that the central pillaris not in contact with the surrounding ground. This shouldensure that the effects of diurnal or seasonal earth movementsare minimized.

A further common design feature is the incorporation of somesystem of forced centring. The two most common forcedcentring systems for deformation monitoring are the Wild ballcentring system, and the Kern system. Both designs are illus-trated in Figure 6.41. Use of either system should ensure thaterrors from this source do not exceed ± 0.1 mm.

This concept of an insulating sleeve around the referencepoint is also evident in the levelling datum designs illustrated inFigure 6.42. Figure 6.42(a) can be seen to consist of a steel footdriven into the ground at the bottom of a 5 m deep borehole.This steel foot is connected to a central rod which extends

Figure 6.41 Pillar centering systems (a) Kern system; (b) Wildsystem

(c) Cross section of Base Line Pillar (Deeth et a/27)Figure 6.40 (Continued) Survey pillar designs

Kern 196 base platePVC pipeAir gap

DrainConcrete walkway

Backfill

Cage reinforcement

Sulphate-resistingconcrete

0.7m

1.20 m approx.(dependent onground-level)

1.20m

1.50m

0.5m0.4m

almost to ground-level. In order to minimize the potential effectof ground movement influencing the datum, the central rod issurrounded by a telescopic sleeve. An alternative design isillustrated in Figure 6.42(b). In this case, the inner datum rodconsists of a 10 mm bore galvanized tube surrounded by a guardtube of 25 mm bore. The guard tube is driven into a 150 mmdiameter hole which has been bored to a depth of 5.5 m. Thedatum point consists of a dome-shaped steel ball about 0.1 mbelow ground-level. The ball is covered by a protective cap andby a sealing plate (possibly a manhole cover), at ground-level.Further details relating to the installation procedure may befound in Cheney.28 It is also important to install a sufficientnumber of datums to enable any settlement or uplift of thedatums to be detected. Ideally, three should be installed.

Monitoring points. The main requirement of a survey monitor-ing point is that it can be either permanently affixed to, orprecisely relocated, on the structure being investigated. Again,two distinct types of monitoring point can be identified, those towhich angle/distance measurements will be taken and thosewhich will be used as precise levelling settlement points.

The measurement of angles and distances to monitoringpoints on the structure will normally require that the targetpoints be designed so that they are capable of accepting bothconventional survey targets and also corner cube reflectors. Thesimplest approach, and that commonly used for dam deforma-tion work, is to build a series of pillars on the structure. Thepillars are fitted with a suitable forced centring system whichenables both survey targets and corner cube reflectors to beinterchanged very accurately. An example of such a system isillustrated in Figure 6.43.29 An alternative approach may beused in situations where the distances to be measured are short.In these cases it may be possible to permanently affix targets andreflectors to the structure. Details of one particular arrangementwhich involves the use of reflex acrylic reflectors is reported byKennie.13 Cheney30 also discusses the use of permanentlymounted reflectors, in this instance for use with the KernMekometer ME 3000.

For settlement measurements, one of the most commondesigns of monitoring point is that which has been designed bythe Building Research Establishment (BRE). The componentsof the settlement system are illustrated in Figure 6.44. It can beseen that the system consists of four components: (1) a stainlesssteel socket 65 mm long by 22 mm diameter which is groutedinto the structure; (2) a detachable settlement bolt; (3) aprotective Perspex cap; and (4) an alloy wrench to allowremoval of the cap. The main advantages of such a system are,firstly, a high degree of accuracy in relocating the settlementpoint; Cheney28 states that the settlement bolts may be reposi-tioned in the socket to within 0.03 mm. The second advantage isthe level of protection offered to the settlement point fromaccidental or other damage.

6.3.4.2 Computational processes

Horizontal control surveys for deformation monitoring pur-poses are essentially no different from control surveys for otherpurposes (e.g. national control surveys). Whilst they may differin size (generally much smaller), and accuracy requirements(generally much higher), the basic principles in terms of obser-vational and computational techniques are common. It is notsurprising to note, therefore, that in recent years the type ofcontrol survey generally adopted for deformation purposesconsists of a mixed set of angle and distance observations or a'survey network'. The advantages of such an approach inreducing scale and orientation errors have been discussed pre-viously. However, the primary advantage of this approachoccurs when the data are subjected to a rigorous least squares

Ground levelCover box

12 mm dia.steel ballCentral rod

Telescopingtube100 mm dia.borehole

Steel foot

Datum point (installed position)(a) Datum (Penman and Charles26)

5m

Datumlevel

Groundsurface

Ground surfaceSaltglazed

" Pipe

Sealing plate

(Site protection forthe installed beam)

Pitbottom

Pit bottom

Borehole

Protective capDatum levelling domeDome coupling

Datum baseGuard tubeShoe

Datum rod

(c) (Component detail)

(Before driving)

(Installedposition)

(b) Datum (Cheney28)

Figure 6.42 Levelling datum designs (from Penman andCharles,26 and Cheney28)

Figure 6.44 Building Research Establishment settlement boltsystem

adjustment using the method of variation of coordinates (seesection 6.3.1.4). This not only enables the most probable valuesof the positions to be arrived at, but also statistical data aboutthe precision and reliability of the network to be determined.Furthermore, by interpreting these statistical indices and thenthe results of a network adjustment from two different epochs itis possible to evaluate the statistical significance of any changeswhich are observed. The various stages in the computationalprocess are shown in Figure 6.45.

The various statistical indices which can be derived from a

variation of coordinates adjustment is well documented in theliterature, e.g. Cooper.1 It is therefore intended to make onlybrief mention of the primary features of each statistical index.

Three distinct types of statistical indices can be identified: (1)those which are concerned with network precision; (2) networkreliability; and (3) for deformation analysis.

(i) Network precision.

ResidualsThe residuals (v) of the observation equations are obtainedfrom:

v = AX-b (6.36)

where A, X and b are as defined in Equation (6.29). In caseswhere a large number of observations exist, the ratio v/cr, whereCT refers to the standard error of the observation, may be used inorder to reject suspect observations.' For example, if v/a isgreater than 2.5, this indicates that there is only a 1.2%probability that this variation is caused by random observa-tional effects and is more likely to be indicative of a poorobservation. On this basis, therefore, the observation should berejected.

Unit varianceThe quantity aj is computed from the expression:

2_\TWv°Q m-n (6.37)

Figure 6.43 Pillar target point (from Egger29)

:400ffi$ €l№ BOOT

;*flifc

Figure 6.45 Stages of computation: deformation monitoring surveys

where m is the number of observations, n the number ofunknowns and W is the weight matrix. It can be shown that,theoretically, cr0 (the standard error of unit weight) should equal1. A large departure from unity is usually indicative of somegross blunder, whilst a small departure may "give an indicationwhether the initial a priori standard errors of the observationshave been correctly estimated. In certain circumstances it ispossible to correct the initial standard errors by a 'trial anderror' process in order to bring the value of (T0 closer to unity.

Variance - covariance matrixThe variance - covariance matrix (axx), is an extremely useful

by-product of the adjustment process. It is computed from thefollowing expression:

^ = Oj(A1WA)-' (6.38)

The diagonal elements of the matrix (a\, erj,., . • • ), refer to thevariances or squares of the standard errors associated with theunknowns E-, N{, etc. It is therefore an extremely useful means ofmeasuring the precision of the network. The off-diagonal ele-ments, or covariances (crE.,N, . . . ) may be used in order toconstruct error ellipses.

Error ellipsesThe 'absolute error ellipse' of a calculated point is an ellipse

defined by crmin and <7max with the semimajor axis fl = crmax and thesemiminor axis b = omm. When drawn around a point the ellipseis generally interpreted as depicting the region within whichthere is 39% confidence that it contains the position of the point.The values of crmax and crmin are determined by the followingexpressions:

ff™ = № + ON) + Ii(Oi ~ O + OeJ"2 (6.39)

»-. = ««1 + «?,) - IKoI - «W + *BN]"2 (6.40)

Two drawbacks exist with error ellipses of this type. Firstlythe size of the error ellipse is not 'invariant' and depends on theposition of the point which is chosen as the origin. Secondly, noinformation is given about any inter-station correlation whichmay exist. An alternative type of error ellipse which overcomesthe first drawback is the 'relative error ellipse'. In this case theellipse is computed on the basis of the variances and covariancesof the differences in coordinates between points. It is therefore ameasure of the relative positional accuracy between the points.

A posteriori standard errors

The a posteriori or post adjustment standard errors of theadjusted quantities are a further set of 'invariant' statisticswhich can be computed. For example, the a posteriori standarderror associated with the adjusted distance between any twopoints in the network is given by:

~] |~ P "I (6.41)a? = [PQ RS] 4 x 4 Sub-matrix Q

of<J x x RL J L s _

where P, Q, R and S are the trigonometric functions defined inEquation (6.24).

(U) Network reliability. Network reliability in this context isdefined as the ability of the network to detect gross errors in theobserved quantities. Ashkenazi21 has suggested that the follow-ing statistic should be used in order to assess the reliability of thenetwork:

.. .... _ a posteriori standard error of the observation1 1 y a priori standard error of the observation

(6.42)

Networkconfiguration

Precision ofobservations

Network ofobservations

Mathematicalmodel

Adjustment Results

Networkprecision

Deformationanalysis

Networkreliability

Deformation?

Statisticalanalysis

Theoretically the greater the number of observations indir-ectly affecting a particular observation, the lower the ratioshould be. Thus low ratios are indicative of high reliability,whilst a ratio of one would indicate complete unreliability.Ashkenazi suggests that the maximum acceptable value of thisratio should be 0.9

(in) Deformation analysis. The statistical indices discussed sofar relate to a single set of network observations. A much morecommon problem in deformation monitoring work, however, isthe situation where two or more sets of observations of the samenetwork, taken at different times, exist. In these circumstances,further analysis is required in order to determine whether anydifferences in coordinates which exist are statistically significant.

Currently this particular problem is the subject of consider-able research interest. Following the second FIG symposium onDeformation Measurements in Bonn, 1978, a committee wasestablished in order to investigate different approaches into theanalysis of deformation measurements using the same measure-ment data. The provisional results of this committee's workwere reported by Chrzanowski.31 The primary problems whichthis group, and others, have been examining are, firstly, what isthe most appropriate means of eliminating systematic effectsfrom the data sets, and secondly, what are the most appropriatetypes of statistical test to apply to the recorded deformations inorder to test their significance.

The first problem, that of eliminating systematic effects, suchas a scale bias in EDM measurements, has been discussed byAshkenazi et 0/.23 and Ashkenazi.21 In cases where a systematicbias is considered to be present in one set of data, the systematicbias parameters (scale, orientation and translation) may bedetermined by means of a least squares four-parameter Helmerttransformation. By the use of this procedure it is thereforepossible to eliminate from one data set a series of systematiceffects which, if not eliminated, may be mistaken as systematicdeformations.

The main dangers associated with the application of thistechnique to networks is that it may lead to the removal of agenuine systematic deformation under the mistaken believe thatit is caused by observational error. It has therefore beensuggested by Dodson,32 that it is preferable for any systematiceffects to be eliminated by careful instrument calibration, or bydesigning the network so that 'the influence of such effects doesnot influence the acceptance or rejection of any hypothesis ofdeformation'.

The second problem, that of devising the most appropriatestatistical tests, has also been examined by Ashkenazi andDodson.33 The approach which they have adopted involvescomparing the detected coordinate differences with the corres-ponding coordinate standard errors. The basis of this approachcan be illustrated by considering two points P and Q in afictitious network. If the position of P in the network is stronger(i.e. more accurately determined because of a good geometricalconfiguration and higher precision observations) than point Q,then quite clearly a smaller difference between two successivevalues of the coordinates of P, may be more significant than alarger difference corresponding to point Q. Thus periodic coor-dinate differences may only be considered to be significant if thestandard error associated with the deformation at that point isalso examined. Therefore, if:

A2= (AE2+ AN*)2 (6.43)

where A is the computed deformation based on planimetriccoordinate differences AE, and AN, and the standard error in Ais given by:

*i = (*i£ + *i*)/2 (6.44)

where aA£ « v'2oE and aAN « >/20W (6.45)

then, the ratio AIo may then be used in order to test thesignificance of any deformation. For example, if the ratio AJa is2 (95% probability level) then there is only a slight chance (19:1against) that the movement has been caused by random obser-vational errors and is much more likely to have been caused bymovement of the point. In contrast, a ratio of 0.7 would indicatethat there was an even chance (50% probability level) that thedifference was caused by movement, it being equally likely thatthe variation was simply a manifestation of a random observa-tional error. Ashkenazi22'23 discusses the application of thistechnique to sets of data observed at Cattleshaw reservoir.

6.4 Computers in surveying

Computers have, throughout their development, been extremelyimportant in the fields of surveying and mapping. Initially, theiruse was almost exclusively restricted to the 'number crunching'requirements of large organizations carrying out geodetic com-putations or the adjustment of major control frameworks.Operations of this type were carried out on large mainframecomputers in batch mode. Whilst slow and cumbersome tooperate by modern computing standards, these early computersoffered enormous benefits to the surveyor: their computationalspeed, and their ability to deal with the application of rigorouscomputational techniques, such as the method of least squares,on a much greater scale than was possible previously. In recentyears, the trend, in computational terms, has been towards thedevelopment of more 'user friendly' software which can beoperated on smaller computers, particularly microcomputers(see Milne34 and Walker and Whiting35).

Whilst the use of computers for solving mathematical prob-lems in surveying continues to develop, the main thrust area ofinterest in recent years has been in the development of interac-tive graphics systems for processing and displaying surveyingand mapping data. Thus, the emphasis is becoming concen-trated more on the capability of the computer to store, search,retrieve and display digital map data, than on computationalspeed. Such systems offer considerable benefits both in terms ofspeed of access and flexibility of use, e.g. the ability to be able todisplay selectively map data at a user-defined scale. Indeed,there is little doubt that over the next decade the storage of mapdata in digital form will continue to expand at an ever-increasing speed as the costs of computer memory drop and theavailability of data in digital form increases. The engineer willtherefore come into increasing contact with these types of data,primarily at the design stage of a project, using large-scaledigital maps and ground-modelling systems, but also at theproject planning stage using data from land/geographic infor-mation systems.

6.4.1 Digital mapping and ground-modelling systems

In the UK, responsibility for national mapping lies with theOrdnance Survey. The Ordnance Survey are also involved in theproduction of digital maps (Thompson36 and Logan37), andactively considering the possibility of producing a nationaldigital topographic database (Thompson38 and Rhind39). Thetask of compiling large-scale digital maps to cover the entirecountry is an enormous task and one which requires substantialresources if it is to be produced within a reasonable time-scale.To date, the success of the Ordnance Survey in providing large-scale digital mapping coverage has been very limited (10 to 15%coverage). It is hoped that as technology develops and, perhapsmore significantly, as funding becomes more available, that thissituation will improve. Nevertheless, in spite of the lack of

Ordnance Survey-derived data, many organizations areinvolved in the acquisition, processing and sale of digital surveydata in the commercial sector. Several sophisticated suites ofsoftware have also evolved for the manipulation of this digitalsurvey data, particularly in conjunction with civil engineeringdesign information. Two examples which are representative ofthis range of software are MOSS and the Eclipse InteractiveGround Modelling System.

6.4.Ll MOSS system

MOSS is a combined surveying and engineering design systemwhich records information in the form of three-dimensional'strings'. These 'strings' consist of a series of linked coordinatedpoints representing features such as kerbs, roads, railway lines,or ground level detail, such as contours. Each string is given alabel and by covering an area with a sufficient number of stringsand/or point information it is possible to represent the terrain indigital form in the form of a MOSS 'model'. A wide variety ofsurvey methods can be used in order to create a string ground-model, and Figure 6.46 illustrates a few of the options which areavailable. The design features within the system also enable amodel of any proposed works to be generated. By comparisonof both models it is therefore possible to generate other datasuch as earthwork quantities.

The MOSS system was initially launched in the mid 1970s bya consortium of several county councils within the UK.Although it was not initially developed with the production oflarge-scale plans as its primary aim, it has been upgraded andenhanced since its inception and is now widely used for large-scale survey purposes. Figure 6.47 is an example of the use ofMOSS in this mode. MOSS is, however, much more widely usedas an integrated survey and design system, rather than as astand-alone survey system, and it has been used in this mode fora wide variety of projects including: highway design (Houg-ham40 and Fawcett41), railway design (Bedingfield and Craine42)and the design of land reclamation schemes (Wilson43). A newinteractive version of MOSS has recently been launched byMOSS Systems Ltd, a company which was formed by severalmembers of the original consortium in 1983. This version willoffer considerable benefits to its users, particularly in terms ofspeed and ease of use.

OrdnanceSurveydigital data

Aerial survey Independentdigital data

MOSSground mode

Conventionalground survey

Ground surveyusingelectronicdata capture

Figure 6.46 Alternative methods of creating a MOSS groundmodel

Figure 6.47 MOSS model taken from a graphics VDU showingan Ordnance Survey digital map, updated with a small roadimprovement, bridge detail, proposed river diversion, and culvert.(MOSS Systems Ltd, model courtesy of West Sussex CountyCouncil.)

Pike Shoot Cottage

Coultershaw FarmHouse

Coultershaw Bridge

Sluice

6.4.1.2 Eclipse system

The Eclipse interactive ground modelling system is a furtherexample of an integrated system for survey processing, digitalmapping and engineering design. In common with the MOSSsystem the software is designed so that it can be installed on avariety of different computer systems. However, unlike MOSSwhich is designed for use with a variety of mainframe (IBM,ICL, etc.), minicomputers (DEC/VAX, Prime, etc.) and en-gineering workstations (Sun, Apollo, etc.), the Eclipse systemhas been designed primarily for Wang minicomputers and,recently, for the Wang PC and the IBM PC/AT.

The program modules within the system include software forprocessing and editing - interactively - survey control and detailobservations. Data input can be by manual keyboard entry, orautomatically by transfer from one of the electronic datacollectors currently on the market. The design software includesfacilities for road design, drainage design, building layout designand land reclamation design. Throughout, facilities exist forderiving additional information such as areas and earthworkquantities. For high-quality output a series of options within thedraughting software enable the user to enhance cartographicallythe screen image in order to produce final contract drawings ifnecessary.

Other systems currently available which perform operationssimilar to those outlined include the Wild System-9, Intergraph,HASP, AXIS and ProSurveyor.

6.4.2 Land information systems

Land information systems (LIS), are currently one of the majorgrowth areas in computing as applied to surveying and map-ping. They are of particular relevance to engineers involved withfeasibility and planning studies. The concept of a LIS can bestbe described by considering the definition offered by Anders-son:44

A land information system is a tool for legal, administrativeand economic decision making, and as an aid for planningand development which consists, on the one hand, of adatabase containing spatially referenced land related dataand, on the other hand, of procedures and techniques for thesystematic collection, updating, processing and distributionof the data. The base of a LIS is a uniform spatial referencingsystem for the data in the system which also facilitates thelinking of data within the system with the other land relateddata.

Two other terms, Geographic Information System (GIS)(Hallam45) and Urban Information System (UIS) (Parker andBray46), are also used to describe systems which operate in asimilar manner but within a regional or local urban arearespectively.

All of these systems are very much in their infancy at present.However, many organizations concerned with, for example,public utilities (gas, electricity, waste water disposal) and landtaxation/valuation/registration are already investing significantsums of money in the development of such systems. A review ofthe history and future possibilities for LIS in the UK is providedby Dale.47 The LIS will have enormous impact on the amount ofinformation available to the engineer and planner. The manage-ment and efficient use of this data is the challenge which has tobe faced in the future.

6.5 Acknowledgements

The author would like to thank the following persons andorganizations who helped in the preparation of this chapter:

Geotronics (UK) Limited, Moss Systems Limited, SpectraPhysics (UK) Limited, and Wild Heerbrugg Limited for per-mission to reproduce photographs of their products. Also, toMrs Veronica Brown for her valuable comments on the initialdraft.

References

1 Cooper, M. A. R. (1987) Control surveys in civil engineering.Collins.

2 Mikhail, E. M. and Grade, G. (1981) Analysis and adjustment ofsurvey measurements, Van Nostrand Reinhold.

3 Cooper, M. A. R. (1982) Modern theodolites and levels, 2nd edn.Granada.

4 Miller, R. M. (1969) 'Accuracy of measurement with steel tapes',Building Research paper CP5J/69, Building Research Station,Watford.

5 Hodges, D. J. and Greenwood, J. B. (1971) Optical distancemeasurement. Butterworths.

6 Smith, J. R. (1970) Optical distance measurement, CrosbyLockwood Staples.

7 Jaakola, M. (1971) 'Survey with the Tellurometer MA-IOO',Survey Review, 159, 29-34.

8 Froome, K. D. (1971) 'Mekometer: EDM with submillimetreresolution, Survey Review, 161, 98-112.

9 Meir-Hirmer, B. (1978) 'Mekometer ME3000. Theoretical aspects,frequency calibration, field tests', Proc. Int. Symp. on EDM andthe influence of atmospheric refraction. IAGG, Wageningen, May.

10 Murmane, A. B. (1982) The use of the Mekometer ME3000 in theMelbourne and Metropolitan Board of Works', Proc. of the 3rdInt. Symp. on deformation measurements by geodetic methods.Budapest.

11 Burnside, C. D. (1982) Electromagnetic distance measurement, 2ndedn. Granada.

12 Kennie, T. J. M. (1983) 'Some tests of retroreflective materials forelectro-optical distance measurement', Survey Review, 207, 3-12.

13 Kennie, T. J. M. (1984) 'The use of acrylic retroreflectors formonitoring the deformation of a bridge abutment, CivilEngineering Surveyor, 9, 6, 10-15.

14 Schwendener, H. R. (1972) 'Electronic distances for short range:accuracy and checking procedures', Survey Review, 164, 273-281.

15 Ashkenazi, V. and Dodson, A. H. (1975) The Nottinghammulti-pillar baseline', Proc. of the 26th General Assembly of theInt. Ass. of Geodesy, Grenoble.

16 Sprent, A. and Zwart, P. R. (1978) 'EDM calibration - a scenario',Australian Surveyor, 29, 3, 157-169.

17 Ministry of Defence (1978) Military engineering Part 2: FieldSurvey, 29, 1-8.

18 Murray, G. A. (1980) 'Lasers and Dinorwic', Civil EngineeringSurveyor, 5, 6, 6-11.

19 Schofield, W. (1979) The effect of various adjustment procedureson traverse networks', Civil Engineering Surveyor, 4, 4, 13-19.

20 Ashkenazi, V. (1968) The solution and analysis of large geodeticnetworks', Survey Review, 146, 166-173.

21 Ashkenazi, V. (1981) 'Least square adjustment: signal or justnoise', Chartered Land Surveyor, 3, 42—49.

22 Ashkenazi, V., Dodson, A. H. and Crane, S. A. (198Oa)'Monitoring deformations to millimetre accuracy', Proc. of theFIG Commission 6 Symp. on engineering surveying. London.

23 Ashkenazi, V., Dodson, A. H., Skyes, R. M. and Crane, S. A.(198Ob) 'Remote measurement of ground movements by surveyingtechniques', Civil Engineering Surveyor, 5, 4, 15-22.

24 Redmond, F. A. (1951) Tacheometry. The Technical Press.25 Egger, K. (1983) 'Geodetic measurement and the unusual behavior

of the Zeuzier arch dam', Land and Minerals Surveying, 1, 10,15-21.

26 Penman, A. D. M. and Charles, J. A. (1971) Measuringmovements of engineering structures, Building Research StationPublication No. CP 32/71.

27 Deeth, C. P., Dodson, A. H. and Ashkenazi, V. (1979) 'EDM:accuracy and calibration', Symposium on EDM, Polytechnic ofthe South Bank. London.

28 Cheney, J. E. (1973) Techniques and equipment using thesurveyor's level for the accurate measurement of building

movements, Proc. of the British Geotechnical Society Symposiumon Field Instrumentation. London.

29 Egger, K. (1970) Precision measurement with special reference todeformation of dams, Kern Instrument Company.

30 Cheney, J. E. (1980) 'Some requirements and developments insurveying instrumentation for civil engineering monitoring', Proc.of the FIG Commission 6 Symp. on Engineering Surveying.London.

31 Chrzanowski, A. (1981) 'A comparison of different approachesinto the analysis of deformation measurements', Proc. of the FIGConference. Montreaux.

32 Dodson, A. H. (1984) 'Pre-analysis and design of a measurementscheme', Land and Minerals Surveying, 2, 1, 13-19.

33 Ashkenazi, V. and Dodson, A. H. (1978) Measuring deformationsby surveying techniques. Seminar, University of Nottingham.

34 Milne, P. H. (1984) Basic programs for land surveying. Spon.35 Walker, A. S. and Whiting, B. M. (1983) 'Multitudinous micros

and micros in mapping', Land and Minerals Surveying, 1, 1, 34—42.36 Thompson, C. N. (1978) 'Digital mapping in the Ordnance Survey

1968-78. Proc. of the ISP Commission 4 Symp. on 'NewTechnology for Mapping"1. Ottowa.

37 Logan, I. T. (1981) 'Ordnance Survey digital mapping', Proc. ofthe 1st UK National Conf. on Land Surveying and Mapping. PaperG4.

38 Thompson, C. N. (1979) The need for a large-scale topographicdatabase', Proc. of the Conf. of Commonwealth Surveyors.Cambridge.

39 Rhind, D. W. (1981) 'Digital data banks and digital mapping',Proc. of the 1st UK National Land Surveying and MappingConference. Paper G2.

40 Hougham, P. (1980) The application of MOSS to roadworks in ashire county', Proc. 2nd Int. MOSS Conference. Bournemouth,Paper 9.

41 Fawcett, D. S. (1980) The design of a complex minorinterchange', Proc. 2nd Int. MOSS Conference. Bournemouth,Paper 8.

42 Bedingfield, P. G. and Craine, G. S. (1981) 'An automated surveyand integrated design system', Proc. of the 1st UK National LandSurveying and Mapping Conference. Paper F3.

43 Wilson, P. (1980) 'Land reclamation using MOSS', Proc. 2nd Int.MOSS Conference. Bournemouth, Paper 4.

44 Anderson, S. (1981) 'LIS-what is that?', Proc. of the FIGCongress, Montreux, Paper 301.1.

45 Hallam, C. A. (1979) The USGS geographic information,retrieval and analysis system: overview', Proc. of the 39th ACSMmeeting. Washington DC, 229-246.

46 Parker, D. and Bray, D. (1983) The surveyor and urbaninformation systems', Chartered Land Surveyor, 4, 3, 4-15.

47 Dale, P. (1984) 'Land information systems - which way in theUK?' Land and Minerals Surveying, 2, 6, 313-316.

Bibliography

Bannister, A. and Raymond, S. (1986) Surveying, 6th edn. Pitman,510pp.

Methley, B. O. F. (1986) Computational models in surveying andphotogrammetry. Blackie.

Olliver, J. G. and Clendinning, J. (1978) Principles of surveying, VoIs Iand II, 4th edn. Van Nostrand Reinhold.

Schofield, W. (1984) Engineering surveying 2, 2nd edn. Butterworths,276pp.

Shepherd, F. A. (1981) Advanced engineering surveying, EdwardArnold, 276pp.

Shepherd, F. A. (1984) Engineering surveying, 2nd edn, EdwardArnold, 370pp.

Uren, J. and Price, W. F. (1978) Surveying for engineers. Macmillan,298pp.

Uren, J. and Price, W. F. (1984) Calculations for engineering surveying.Van Nostrand Reinhold, 309pp.