1 Applicability of Classifications for Tunnelling – Valuable for Improving Insight, but Problematic for Contractual Support Definition or Final Design. T. G. Carter 1 1 Golder Associates, Toronto, Canada Rock Mass Classifications have been applied worldwide since the 1940’s and nowadays form the back- bone for empirical design approaches for rock tunnelling [1]. Singh and Goel [2] in the introduction to their book on classifications quote Lord Kelvin’s philosophy as almost a justification for why classifications developed and why they are still so necessary in rock engineering – viz. "When you can measure what you are speaking about, and express it in numbers, you know something about it, but when you cannot measure it, when you cannot express it in numbers, your knowledge is of a meagre and unsatisfactory kind; it may be the beginning of knowledge, but you have scarcely in your thoughts, advanced to the stage of science." This statement, however, lies at the heart of the continuing problem that has plagued applicability of rockmass classifications to tunnelling – that geological conditions vary and thus cannot be reliably and repeatably quantified in numeric terms. Hence, because of this imprecision there continues to be a disconnect between prediction and reality (Figure 1). Figure 1: Characteristic mismatch between classification-based predictions and as-encountered conditions. Tunnel support class predicted and mapped data on the left, and design strength estimate predictions and field data on the right, (9 classifications, 11 observers – from Edelbro et al., 2007) [3] . This disconnect is not solely related to “uncertainty” and “inaccuracy” in the degree to which prescriptive characterization of natural variability in actual geology can be achieved. Some uncertainty arises because of ambiguous parameter definition. This to some degree is tractable by statistical evaluation [4]. Some though arises due to differences in observer perspective, and this is much more difficult to resolve (particularly when classification values have contractual significance). This “divergence of opinion” oftentimes is taken as true differences in fact, and frequently then forms the basis for claims. Partly as a result of this, tunnel contracting has moved progressively towards use of Geotechnical Baseline Reports (GBR’s) and risk sharing contract language [5]. GBR’s however have not proved to be a panacea to solving the fundamental disconnect. As a consequence there still remains a need to solve this problem – (i) ideally, by completely divorcing classifications from being tied directly to contract payments and (ii) by returning them to their original and most useful role – as aids to preliminary design – where they have real value – for improving insight and understanding of likely tunnelling rockmass conditions and variability. Proc. World Tunnelling Conference (WTC 2010), 36th ITA Congress. Vancouver, Paper 00401, Session 6c