Regime-based persistence probability of ceiling and visibility

Several ceiling and visibility (C&V) objective forecast guidance systems use persistence as one input among several methods, in addition to using it as a benchmark for verification. For best results, these systems need to determine when persistence has a high probability of failure and adjust the crossover time to other methods appropriately.

This paper examines regime-based performance of persistence in the context of operational C&V forecasting. Based on analyses of long-term records of hourly observations at airports in Canada, it has been found that under specific types of forecast situations, or weather regimes, the expected duration of persistence can be significantly different than the average. A regime-based approach highlights the need to account for variability in the observations and shows promise in improving the ability to forecast change. There are indications that using persistence as a benchmark without factoring in regime may at times be inappropriate.

Supplementary webpage: http://collaboration.cmc.ec.gc.ca/science/arma/aram2010