This paper examines first a statistical modelling of the number of navigational accidents (collisons, groundings and strikings) involving tankers and freighters in Canadian waters; it then looks at a direct method of modelling oil spill frequency and magnitude, given a navigational accident. The former is used to examine various hypotheses in the literature regarding the impact of various factors on navigational risk. The latter is based on a direct probability tree analysis which allows the estimation of detailed oil spill impacts by facilitating the examination of representative spill sizes.

This paper examines the frequency distributions of the ranges at which vessels take action to avoid collision. Published data from simulator exercises are employed. Both give-way and stand-on vessels are considered. Lognormal distributions fitted to the data are shown to be superior to those previously published. Underlying ‘causal’ models are developed which are capable of generating the observed distributions.

The present-day emphasis on the need to analyse and identify risk situations in ports is hardly a new phenomenon. In fact, port managements have, as a matter of course, been very conscious of this need but over the years, with the ever-increasing transport by sea of hazardous hydrocarbon, chemical as well as nuclear-related products, this need has simply become much more pronounced. Moreover, present-day public awareness and concern about the possible incident impact on the environment in the form of pollution or personal injury has even further increased the awareness of the corporate responsibility and the need to assess the operational risk levels of ports. Of equal importance, of course, is the need and the capability to predict operational risk levels for new port developments or modification projects or of the possible risks associated with the introduction of hazardous cargo-handling facilities at existing or new port terminals.

Criteria and a methodology for evaluating the relative effectiveness of different navigation aids in a coastal region, taking account of the ship-borne equipment, human factors and the needs of the navigator, have been discussed in the Nautical Institute' recently-published book on Pilotage and Shiphandling. The method is based on the product of the probabilities that the equipment will be fitted, that it will be available (in reliability terms), that it will be properly used and that it will provide the required information.