A risk analysis of mortality as a result of the sinking of Norwegian merchant ships during World War 2
Background. To what degree would it be possible to use carefully recorded historical material in modern day risk assessments? I wanted to find out if it was possible to use the merchant marine records from the Second World War as a basis for gathering statistical data on possible risk factors to survival in maritime disasters. Furthermore I wanted to find out if any single factor could pose a higher risk of mortality than others. My main hypothesis was that an orderly conducted evacuation of the ship, as opposed to a less orderly one, would significantly reduce lethality amongst the crew members during the sinking of a merchant vessel. Furthermore I hypothesised that both fire onboard and a short time from attack to complete sinking, would greatly increase the death rate amongst the crew members, to a higher degree than the other supposed risk factors.
Material and methods. I chose 100 well documented sinking reports from the Of these 40 were randomly selected. For each of these the following variables where registered: whether the ship was sunk during daytime or at night. If lifeboat manouvers where held prior to the sinking or not. Was the ship sunk while sailing in a convoy, or was it sailing alone. Was it sunk in summertime or during wintertime. What the weather was like at the time of sinking. What kind of cargo the ship was carrying, divided into high or low risk cargo. How fast the ship sunk. If or not during the sinking, fire broke out. And finally, if the ship was evacuated in an orderly fashion or not, with a complete or partial collapse of command structure. After this the absolute numbers of casualties for each sinking was registered and these numbers where recalculated and presented as average percent values for each of the variables. Calculations were then made to establish a possible significant relationship between assumed risk factors and actual mortality.
Results. It was found that the average percent of casualties for evacuations that were performed in an orderly fashion numbered 8 %, while those that did not had an average mortality rate of 43 %. This was found to be a significant difference, and the other assumed risk factors could not in a substantial manner account for the differences in mortality after being corrected for through a regression analysis. Furthermore it was indicated that amongst the other assumed risk factors, only a night time sinking and sailing in a convoy seemed to significantly increase the risk of mortality.
Interpretation. The state of evacuation, independent of other assumed risk factors, seems to greatly influence the mortality rate during a sinking. Also being sunk whilst sailing in a convoy or being sunk during night time, can be assumed to increase the mortality rate amongst crew members.