For the 2008-2012 predictions, we include an estimate of the 2007 season in the data used by our models. Since the 2007 expert elicitation took place in October, and the 2007 season was not complete at that time, the numbers used for the 2007 prediction were calculated by taking the number of storms that had occurred up to October 15, 2007 and adding a correction factor for the rest of the season. The correction factor is based on the historical ratio of the number of storms after and before October 15. This ratio is shown in Table 1 along side the number of storms before October 15th, 2007, and the number of storms in each category for the year 2007 that were used for the 2008-2012 predictions.
We note that for the 2006-2010 and 2007-2011 predictions, the models used full seasons for the years preceding the predictions.
Was this article helpful?