## Optimum Level of Expenditure

An incomplete appreciation of the ground conditions, which is subsequently presented at the construction tender stage, can only cause the Contractor problems, which will almost inevitably increase the final cost to the Client. The adage, 'you pay for a site investigation whether you have one or not' (Anon, 1991), is backed up by many publications, for example, in UK by Rowe (1972), Clayton, Simons and Mathews (1983) and by Attewell and Norgrove (1984).

These publications and many others have established that one of the largest elements of technical and financial risk in civil engineering projects is unforeseen ground conditions. Experience has shown that a modest increase in ground investigation expenditure at the outset could have been repaid many times in reduced project cost over-runs. It is also conceivable, at least in principle, that expenditure on ground investigation might be excessive, far outweighing any possible savings for the project. Thus, in between these two positions, there must be an optimum level of expenditure on ground investigation.

To arrive at this optimum level of expenditure, Goldsworthy (1999), based on theoretical possibilities and assumptions, considers the total project cost, T, as given by:

 total project cost (T) = Planned project expenditure (P) + Ground investigation expenditure (G) + Cost of unforeseen conditions (E) = k * P

T = P + G + E

The cost of unforeseen conditions (E) will be some proportion, k, of P.
Assuming that this proportion, k, depends on the ratio of project expenditure to Ground investigation expenditure (P/G) gives
k = SP/G, where S is a parameter which reflects the sensitivity of a project to ground conditions.

Hence:

T = P +G + SPý/G

and, in terms of the ratio of total project cost to planned expenditure,

T/P = 1 + G/P + SP/G

If R is the ratio of ground investigation expenditure G to the total project cost P,

G = RP and T/P = 1 + R + S/R

Illustrative plots of this last relationship for three typical projects indicate that values of the predicted optimum expenditure G, as a percentage of P are, for:

a low sensitivity (S = 0.0005) building project, around 1.5%

a typical (S = 0.002) civil engineering project, about 4%

a high sensitivity (S = 0.005) tunnel project, between 6% and 8%.

Goldsworthy speculates that 'regional differences relating to typically encountered ground conditions and to contractual practices may affect the sensitivity factor distributions'.

The above estimates can be compared with current levels of expenditure on ground investigation, which are reported to be well below 1% of the total construction costs (Institution of Civil Engineers, 1991 ; Littlejohn, et al., 1994). The cost of claims due to unforeseen ground conditions is usually far in excess of this, let alone the extra costs of over design to cover risk of unknown ground conditions. Note, however, that we believe that the scope of investigation should be sufficient to answer the important questions, and not simply be determined as a percentage of the project costs, although they can give some guidance.