Assessing Risk from Meteorological Phenomena Using Limited and Biased Databases

The assessment of risk attributable to many phenomena relies on the analysis of past history. In the ideal situation, statistics derived from these data should reveal probabilities and trends in the occurrence of significant events. For more dangerous meteorological events like Tropical Cyclones and Severe Thunderstorms, the number of recorded events is somewhat limited. Changes in the nature of information gathering, and technology have biased these limited observations. We need to consider these factors when using the data to assess future risk.