•Relies on calibration of
subjective forecasts of probability.
•Need rain/no rain probability
•Uses conditional
exceedence fractiles:
–The X50,
or amount where there is an equal chance of getting more or less
precipitation that that number
–the X25, amount the forecaster thinks there is a 25% percent
chance of exceeding that value.
•Can
then use curve to set probabilities for any amount.