Aparticularly powerful hurricane's hurricane force winds extended 120 miles from its center. this hurricane was a big storm, and that affects how we think about the prediction errors. suppose we add 120 miles to each error to get an idea of how far from the predicted track we might still find damaging winds. explain what would happen to the correlation between prediction error and year, and why.

The correlation between "prediction error" and "year" would be  affected, as the proposed additional of 120 miles extent for possible wind damage should lead to reconsideration on the tracking process. The tracking is done over statistics regarding - actual and past - sustained wind speed of the hurricane, measures in climate factors, similar past occurrences in the same region. It is all linked, see?

A big event is not occasional, so, necessary is to amplify the time scope for analysys to get a trustworthy prediction.

Explanation:  Statistical factor crossing models such as the Poisson Distribution (it tells the chance of an event to repeat in a time period), SHIPS and GFLD (that apply adapted equations as the hurricanes goes) models have helped a lot in general forecasting prevention, although still lacking accuracy as desired. Calculations, nevertheless, are made over a given geographical area, a given time period, wind speed persistance, similar event matching and other current climate issues. So, event "size" and previous years occurrences really matter for prediction on possible damage.