Interesting. Logically it makes sense that a "cone of certainty" or confidence narrows as you approach the start of a weather event but it's interesting to consider what contributes to that confidence at longer lead times.
I recall a time or two last winter when you had a high level of confidence in something materializing at long lead times - 10-15 or so days out. I don't remember the finer details of the storm, but do remember one of them blossomed into a region-wide event.
With respect to model improvements in recent years, is there a way to quantify that?
For example, is it in terms of "Model X is correct XX% of the time" or is it more nuanced than that?
I ask because it seems like there can be drastic fluctuations from run to run. One run showing an event in SNE only to have it move 100's of miles north in the next run, etc. Although, maybe pinpointing specific areas (IMBYism) is more of an individual human response rather than a testament to the reliability of a model output...? I.e.: it's going to rain in a general area and it could be SNE, CNE or some combination thereof.