Abstract
Canadian weather offices have recently begun to receive probability of precipitation information estimated from numerical weather prediction guidance, but this information is not given in the public forecast. This paper uses a well-known procedure to estimate the increased economic value of these probability forecasts over equivalent categorical forecasts for Toronto. The cost-loss matrix given by Epstein (1969) is generalized and applied in economic value calculations of a Probability of Precipitation Amount (POPA) procedure for the same Toronto data set.
The increased economic value of these probability forecasts is a substantial fraction of the value of categorical forecasts and appears to justify the expense involved in a marketing/education program necessary for the public to understand these forecasts.
