Weather forecasts now and forever will be coupled with some degree of uncertainty. This is based on fundamental science (Chaos Theory) and is not predicated on any lack of understanding the physics and dynamics of the atmosphere nor technical and computer limitations (though this adds additional uncertainty). The degree of uncertainty in forecasts is not constant - it varies as a function of season, location, parameter, and circulation regime. And, everything else being equal, it varies from one day to the next. There exist operational tools now for reliably estimating the uncertainty or, if you like, the confidence in forecasts a priori, i.e., before the forecast valid time. They include statistical post processing of model output based on past error statistics and ensemble prediction techniques.

The issues at hand are, first, educating the public (and your bosses) that forecast uncertainties are inevitable and should be addresed and, second, how broadcasters might convey uncertainties to viewers in a meaningful and comprehensible way. It is believed that audiences at large will be receptive to such information and recognize the advantages of having broadcasts which reliably descriminate between, for example, high and low confidence situations and provide a "heads up" of alternatives to the current forecast that might involve a sgnificant weather event. This paper will present several ideas on these subjects. Hopefully too it will initiate a continuing dialogue within the broadcast community of how best to deal with what is an inescapable fact of life - uncertainties in forecasts.