Groundings and collisions: risk and uncertainty - case studies from the Gulf of Finland on chemical tankers

Uncertainty is an integral part of risk itself, affecting e.g. to what extent and how risk analysis results can be used in decision making. Despite this, the uncertainty in maritime risk analysis is rarely discussed or analyzed. This thesis addresses this uncertainty as well as selected aspects of quantitative grounding and collision consequence analysis. The thesis links these results with general risk theory through uncertainty and its effects on risk management.The spill frequency and sizes of chemical tankers sailing in the Gulf of Finland are modelled. As the current level of understanding is lower for modelling groundings than collisions the former is investigated more in-depth: Grounding damage models are investigated both on a general level and in depth with regard to the effects on grounding damage, depending on the rock model and its differences from real sea bottom data. One of the key uncertainty issues is the actual sea bottom shape. Currently, the rock models used for this purpose are idealized basic geometrical shapes such as cones and polynomials. Their actual similarity to real sea bottom shapes is unknown but is presumed to be low. This means that the uncertainty regarding the results of such analyses is high. The same can be said of many other aspects of quantitative maritime risk analysis.This thesis introduces a systematic approach to mathematically modelling sea bottom. The rock model damage results are compared to results obtained using the real sea bottom shape. The results were inconsistent in terms of equal damage using the sea bottom data and the model results, even with statistically well-fitting models. This currently makes grounding models that require precise input regarding sea bottom shape uncertain to use. Regional grounding damage modelling is still possible with models that do not need this input but these models lose some precision compared to more detailed methods. Furthermore, the uncertainty of all the models utilized in the thesis is shown to be medium to high. This affects the usability of such risk analysis results in risk management. Therefore uncertainty should be systematically analyzed and explicitly communicated in maritime risk studies. Further recommendations for future research include i.e. more in-depth sea bottom shape analysis.