December 2015

The influence of drill hole deviation on Resource estimates in steep dip deposits

  • By Jon Barber MAusIMM(CP), Principal Consultant, Jon Barber Mining Consultants and Stuart Whyte MAusIMM(CP), Geology and Exploration Superintendent, Yarrabee Coal Company

Addressing knowledge gaps in steeply dipping coal geological models and the implications for Resource and Reserve estimates

Author audits of steeply dipping coal geological models in the recent boom time highlighted knowledge gaps in the area of true and apparent thickness and flow-on effects to Resource and Reserve estimates. Discussion with metals geologists suggests that this issue is well understood in their domain; the profusion of relevant information in the oil and gas literature shows that it is well understood and applied in that industry.

Why the knowledge gap in coal? It could, as Arnott (2015) generously terms, be due to ‘indiscretions’ that arose from the high staff turnover and less traditional oversight of the boom time. The authors consider that the knowledge gap is partly related to the boom, but also reflects coal’s geological environment. Most Australian coal mining and exploration is in relatively flat-lying shallow seams where nominally vertical drilling remains vertical and true and apparent thickness are virtually identical.

The base and precious metals geological environment is more varied, with steep dipping and deep deposits necessitating inclined drilling to obtain orebody intersections. The typical oil and gas industry well is deep, with unintentional and intentional deviation to maximise reservoir intersections. The metals and oil and gas industries expect and plan for hole or well deviation. In contrast, coal exploration in flat-lying coal deposits plans for and delivers vertical holes. This has perhaps led to an assumption that holes are vertical in steep dip deposits. However, this assumption is false and will underestimate resources.

In dipping strata, holes naturally deviate up-dip and intersections with the coal seam are often inclined rather than vertical. Modelling using these inclined intersections will always underestimate the coal Resource and Reserve. This issue is most prevalent in steep dipping deposits that characterise Indonesian, Mongolian and Colombian coalfields. Steep dip is less common in Australian, South African and North American coalfields. The authors consider that this flat-lying depositional environment, coupled with the boom time growth, has led to this knowledge gap.

The Yancoal Yarrabee deposit in the Bowen Basin is characterised by seam dips of 0-90°. Using these steep areas of Yarrabee, it is shown that not correcting for deviation results in a coal Resource underestimation of up to seven per cent and a reduction in the coal Reserve of more than ten per cent. This example confirms the rationale for the Joint Ore Resource Committee (JORC) Code 2012 Table 1 entries related to mineralisation width and reinforces the importance of the term estimate in a JORC Resource statement.

Holes deviate up-dip

Coal exploration data is mainly gathered from nominally vertical drill holes. However, holes deviate up-dip, where the seam dip is less than 60°, and down-dip, where the seam dip exceeds 60°.

Many theories are proposed as to why holes deviate from vertical (Brown et al, 1981). In one theory, the chip created when drilling laminated strata is postulated to form by cracking through to the next lamination and the smaller wedge on the up-dip side is easily removed. This increases down-dip lateral bit forces that deflect the bit up-dip. Penetration rate, bit type, the use of stabilisers and the stabilising effect of a core barrel also influence deviation (Brown et al, 1981).


Deviation causes holes to intersect the seam at various angles and provide a mix of thickness observations. This is illustrated in Figure 1, with hole X deviating into the seam and hole Y deviating away from the seam. Figure 1 uses the oil and gas naming conventions of:

MLT = measured logged thickness, from drill hole log picks
TVT = true vertical thickness
TST = true stratigraphic thickness, often referred to as true thickness.

For hole X, MLT exceeds TST but understates TVT. For hole Y, MLT exceeds both TST and TVT.
Setchell’s equation (Evenick, 2008) calculates TVT as:

TVT = MLT × [Cos(β) – {Sin(β) ×
ψb-ψs) × Tan (φs)}]

β is drill hole dip at the point of entry to the seam measured from vertical ψb is the azimuth of drill hole direction (measured from north) ψs is the azimuth of seam dip direction (measured from north) φs is seam dip (measured from horizontal).

In applying Setchell’s equation, seam dip and azimuth at the drill hole are extracted from the floor or roof surface model using the three point method, and hole dip and azimuth are taken from the downhole deviation survey.

Why the focus on TVT? The coal volume in Figure 1 can be estimated as TST times its length, thus TST × L1. Alternatively, the volume can be estimated from TVT and plan area, thus TVT × L2. Most geological and mining software packages operate in Cartesian XY coordinates with grids or block models based on a horizontal XY grid pattern. Thus, volume is estimated as vertical thickness times plan area, or
TVT × L2.

With area defined in the XY plane, the modeller must use TVT not MLT. As shown in Figure 1, MLT underestimates TVT for hole X. The use of MLT in steep dip models will generally underestimate resources. For flat-lying seams, holes remain near vertical and MLT and TVT are virtually identical. Using MLT is also easy; it is observed from the drill picks and, without adjustment, provides a reliable thickness estimate. However, transferring this comfort in MLT to steep dip deposits is fraught with danger. Unfortunately, modellers experienced in flat-lying deposits may underestimate resources in steeper deposits.

Yarrabee mine case study

Yancoal Australia Pty Ltd operates seven mines in Australia, including the Yarrabee open cut mine in Queensland. Yarrabee Mine produces 3.8 Mt/a of PCI coal from an open cut on the eastern edge of the Bowen Basin. Compressional strain oriented east–west has resulted in a complex geological environment with over-thrust faulting and syncline/anticline structures. Seam dips range from 0-90°. Yarrabee exploration holes exhibit the up-dip deviation characteristic of bedding dips below 60°; however, a small number of holes deviate down-dip. Figure 2 illustrates a section of the Yarrabee deposit and highlights this up-dip hole deviation.


All holes shown are rotary chip of 120 mm diameter and are logged for gamma, density, calliper and verticality (deviation) using electronic multi-shot systems. Figure 2 shows the Cancer, Aries and Pollux seams. The Castor Upper and Castor Lower seams are not displayed for clarity. The Pollux seam is approximately 4 m thick.

Table 1 lists details for holes A and B shown in Figure 2. In both holes, MLT underestimates TVT. By difference, this underestimation is expressed as
a thickness error or TE% = (1-MLT/TVT) × 100.


In steep dip deposits, differences of ten per cent between MLT and TVT are common. In the extreme case of a 45° seam dip and a 45° hole dip, the potential thickness error is -30 per cent.

Figure 2 highlights an additional issue. On this cross-section, no hole reaches the syncline axis as they all deviate away. The collar separation between holes C and A of 50 m increases to 100 m at the Pollux seam. At Yarrabee, the model is enhanced with strings along the syncline axis. These strings are located using the seam dip and drill hole deviations.

The Resource underestimation created by modelling with MLT rather than TVT is illustrated using part of the Yarrabee deposit. The thickness error is expressed as (1-MLT/TVT) × 100. Figure 3 shows this error, the observation points and the 10 m Pollux floor contours. The floor contours are indicative of the seam dip. As expected with up-dip deviation, the errors are negative and correlate with seam dip. In the steeper areas, MLT understates TVT by 7-10 per cent. In the flat-lying areas, MLT is within ±2 per cent of TVT. In the horseshoe-shaped area at the lower right of Figure 3, MLT understates TVT by 6.8 per cent.


For seam dips below 60°, holes deviate into the stratigraphy and the deviation increases with depth. A shallow hole (>50 m) in a steep dip area has insufficient length to deviate and MLT will approximate TVT. This is confirmed in Figure 3, where steep dip crop line areas have low errors.

Grouping the errors into 10° dip classes and 40 m depth classes confirms this point. The classed data is shown in Table 2 and Figure 4, with the size of the circles proportional to the thickness error. As expected, the underestimation increases with increasing seam dip and depth.



The positive errors at shallow depth (0-40 m) all occur in steep dip sub-crop areas where the hole has deviated slightly down-dip. The six deep holes with a mean positive error of +1 per cent are all located in the flat area of the deposit. In flat areas, minor deviation caused by drill rig setup or that is strata induced is not self-correcting. The bit forces provided by a wedge chip in steep strata don’t prevail in flat areas. Deviation in a flat area will always result in a slight positive error. A caveat to Table 2 is the small sample size at the extremes of the population.

True vertical thickness in waste units

The potential for coal estimation errors caused by deviation applies equally to waste or interburden thickness. Many modelling packages extend the depth of holes by estimating the position of deep seams in the shallow holes using interburden thickness from surrounding holes. This provides stratigraphic completeness at each hole and enhances modelling. In flat-lying deposits, these extension estimates use MLT, which is virtually identical to TVT. However, in dipping strata, the extension estimates may use a mix of TVT (from straight holes) and MLT (from deviated holes) to extend the shorter holes.

A second issue is the projected trace of the short hole. Should short holes be extended vertically or mimic the deviation of nearby deep holes? A vertical projection should be estimated with TVT and a deviated projection should be estimated with MLT. This mix of complete and incomplete holes can result in software-induced seam floor bumps. In severe cases, these may be falsely interpreted
as faults.

Another issue arises when old unsurveyed holes are mixed with surveyed holes. Given the older holes weren’t surveyed, it is probable that they are more deviated than recent holes. Older unsurveyed holes may need an average deviation survey applied to improve their utility in the modelling process.

JORC implications

The JORC Code 2012 requires the Competent Person to provide deposit information that covers:

  • orientation of data in relation to geological structure
  • relationship between mineralisation widths and intercept lengths.

These two issues are well illustrated in Figure 2. The holes shown penetrate the seam at a variety of angles and MLT does not always provide mineralisation widths. The Competent Person must address these issues within JORC Table 1 on an ‘if not why not’ basis.

As modelling packages estimate volumes in XY Cartesian space, the use of TVT in steep deposits is necessary for the accurate estimation of Resources and Reserves. In the horseshoe area shown in Figure 3, the Resource is understated by 6.8 per cent if MLT is used as a proxy for TVT.

Any Resource error flows onto the Reserves. Reserve limits are based on break-even or marginal economics. Assuming constant costs with depth, a -10 per cent model thickness error at a break-even depth of 100 m will reduce the economic depth to 90 m. Assuming constant seam dip, this equates to a ten per cent reduction in the pit footprint or area. At the pit crest, the thickness error will be zero and at depth the error is ten per cent, giving a five per cent average error. The revised economic Reserves ratio can be estimated as thickness × area or 0.855 (0.95 × 0.9). Thus, a ten per cent thickness error can reduce the economic Reserve by 14.5 per cent. The Reserve reduction will be more severe where a hurdle tonnage or NPV is required to justify an investment.

The authors have also noted that production reconciliations from steep dip deposits can exceed 100 per cent of the model tonnage. Such high reconciliations may indicate the model has been underestimated by using MLT instead
of TVT.


Drilling at Yarrabee in steep dip areas illustrates the general up-dip deviation of holes in steep dip deposits. MLT taken from deviated holes in dipping deposits should be corrected to TVT for model generation. The use of MLT as a proxy for TVT will lead to Resource underestimation in steep deposits. In the authors’ experience, modellers with a background in horizontal deposits can fall into the trap of using MLT as a proxy for TVT. In parts of the Yarrabee deposit, the Resource is underestimated by 6.8 per cent if MLT is used for TVT. In deposits that mix older, unsurveyed holes with recent, surveyed holes, modellers should consider applying estimated surveys to the older holes.

Mining software packages provide methods for converting MLT to TVT and TST. However, software is simply a tool and modellers should understand the capabilities and limitations of the software in relation to deviated drill holes.

JORC Table 1 requires commentary regarding mineralisation thickness, and JORC emphasises, for good reason, that a Resource be expressed as an estimate. The Competent Person needs to understand the issues raised herein when estimating Resources for steep dip deposits. 


Arnott D, 2015. Update on the Australian Coal Guidelines, The AusIMM Bulletin, August:40–43.

Brown E T, Green S J and Sinha K P, 1981. The influence of rock anisotropy on hole deviation in rotary drilling – a review, International Journal of Rock Mechanics, 18:387–401.

Evenick, J C, 2008. Introduction to Well Logs and Subsurface Maps, Tulsa, Pennwell Books.

JORC, 2012. Australasian Code for Reporting of Exploration Results, Mineral Resources and Ore Reserves (The JORC Code).

Sindle T G, Mason I M, Hargreaves J H and Cloete J H, 2006. Adding value to exploration drillholes by improving trajectory survey accuracy, Australian Mining Technology Conference.

Tearprock D J and Bischke R E, 2003. Applied Subsurface Geological Mapping with Structural Methods, New Jersey, Prentice Hall.

Share This Article