Scottish Government

9. Conclusions and recommendations

1. There are doubts as to whether the Welsh and North East Scotland data are sufficiently comparable to justify combination into a unified data matrix. This seems to be due to the different sampling strategies employed in the two studies.
2. Although weak in many cases, the correlation with rainfall made it difficult to validate the quantified risk assessment scoring systems. The most likely reason for this is that the 'noise' in the data which is produced by rainfall (a non-risk-related factor) is so great that it masks the 'signal' derived from the risks indexed in the two investigations.
3. To effect better calibration of a quantitative risk index, it is likely that additional, and greater resolution, sampling would be required which was specifically designed to discriminate between the 'rainfall effect' and catchment risk factors. With such enhanced data it is possible that more sophisticated statistical analyses may be able to identify other contributing factors.
4. The present regulatory regime, particularly for Category 1F supplies, is clearly unsatisfactory and a risk scoring system would certainly be preferable to the status-quo even without firm empirical data fully to calibrate a set of score weights and derive a fully quantitative risk assessment.
5. Movement towards a scoring system which could be subject to subsequent re-calibration if suitable data are acquired would be prudent. In the interim, such a system would be a valuable tool for indicating weaknesses in the supply system that can be improved.