Why do we forecast snow so far in advance

A comment yesterday is a good point for discussion especially with this next storm headed our way. A reader asks why we forecast snowfall better then 36 hours in advance. Because the models are not perfect and the track changes sometimes. I think this is a good point to discuss here on a post instead of just posting a reply in the comment section. So JQ here it goes: First off let me say that I have always had a hard and fast rule not to forecast any snowfall totals more than 36 hours in advance. As you know our forecast models get much less gridded information past 48 hours so past that we are forecasting on the track more than qpf. Let me back off a little and tell you that the model development has been on hyper speed especially over the last ten years. Because of the advances made in computer power and speed the development of models has accelerated as well The accuracy has jumped expedientialy as well. So to say anything about what has happened 'over the years' would not be a fair assessment of what has happened recently. Because of speed and accuracy it has allowed us to give much more advance warning of impending storms. But the most difficult part of a storm to forecast is always the moisture. That's a tricky proposition even when the storm is in progress across the area. If you go back and look at forecasts, very few times are there total 'busts' where we will forecast heavy snowfall and nothing happens. What normally happens is that a narrow band of very heavy snowfall is across an area and light amounts around it. The biggest problem is run into with viewers is that they believe a forecast is a bust if they don't see the heaviest snowfall in their backyard. Twenty miles away we may see the heaviest snowfall, but because it didn't fall at their house then the forecast was wrong. Microcasting has now become an art, with taking weather the global to your backyard. So I think the advance warning is worth it, even if the heaviest snow may be 20 to 50 miles different than the actual forecast.