An Approach to Verifying Prognoses of Line Phenomena and its Application to Fronts over Southeastern Australia

Abstract

An approach to verifying position forecasts of line phenomena was developed. The technique, which tests for occurrence/nonoccurrence and the time of arrival of line phenomena, was applied to frontal position prognoses over southeastern Australia. It was found that operational manual prognoses exhibited no skill in cases of frontogenesis or total frontolysis. In other cases, accuracy varied seasonally and geographically, with summer apparently being the most difficult period for frontal prognosis. Frontal position forecasts were successful for 70 percent of frontal crossings in the 27-month trial period, but two of every five forecast events were false alarms. The Hanssen and Kuipers skill discriminant for a 24-hour Yes/No forecast of a frontal crossing was new 0.6.