Algorithm p-type bias – both events combined
Notes:
The data on this slide indicates whether or not the algorithm output from the models exhibited any kind of bias. For each forecast from the algorithms that were incorrect, a bias was assigned using the following convention: If the algorithm predicted a precipitation type that was “colder” than the observed precipitation, the forecast was counted as being cold biased. (For example, if the algorithm predicted snow, and freezing rain was observed, this was counted as a cold biased forecast). Likewise, if the algorithm predicted a precipitation type that was “warmer” than the observed precipitation, the forecast was counted as being warm biased. (For example, if the algorithm predicted rain, and sleet was observed, this was counted as a warm biased forecast).The results of this study indicated a cold bias for 3 of the 4 algorithms. By contrast, the GFS Bourgoin algorithm appeared to exhibit a warm bias.
Slide 25