15 / 10
      13 / 8
      12 / 4

      First snow looking more likely on...

      Fall colors in Marquette Township 10-11-13 at sunrise
      So with each passing day, the first snow later this week looks more likely, so let's talk about it. I'm going to walk you through my extended forecast.

      Forecasting the weather becomes less certain with each additional day into the future. With modern technology, we can normally forecast with great accuracy up to a few days out, and sometimes as much as a week out, but when you get to the 5-6 days and beyond, there is immense uncertainty that increases with each day.

      But in this case, models have been telling us for several days that sometime later this week we'll be getting cold enough for frozen precipitation.

      For long-term forecasting beyond a few days, I mainly use three models: the Canadian, European, and the American GFS (Global Forecasting System) models.

      As of Sunday morning, the GFS is predicting snow-friendly temperatures Thursday night all the way to Monday or Tuesday night. It's also predicting precipitation each day. Temperatures will likely still be 40s to low 50s in the daytime, meaning just rain, but at night, a mix of snow is possible inland according to this model.

      The Canadian somewhat agrees. It's predicting snow-friendly temperatures Thursday night to only Sunday night before a warm up for early next week. It is also predicting some precip each day.

      The European is coldest of the three. It's predicting snow-friendly temps Friday morning to Sunday night, but with the least amount of precip each day. The European is saying that the freezing temperatures will dip as far south as Oklahoma and bring a winter storm to the central plains, which will mean the first snow for many people if it pans out.

      Not any one of these is a guarantee. I examine all of them to determine a middle-ground solution based on probabilities.

      So as a meteorologist, what do I take from this? First off, it's not worth forecasting precip amounts (either rain or snow). It's too far away for any certainty. Some of the forecast implies that we could see purely snow some of these nights away from the Great Lakes, but what I think is most likely is that temperatures will be cold enough for just a mix of rain and snow, probably Friday and Saturday night when it seems to be the coldest. Rain/snow mixes can still lead to accumulation, but it will likely melt quickly in the following day when highs seem likely to be in the 40s or warmer. I'm thinking parts like Marquette may not see any frozen precipitation being so close to the still relatively warm lakes.

      I hope this wasn't boring. :)

      The first snow of the season is not something lightly forecasted. I wanted to communicate what goes into such a forecast and the uncertainty of forecasting something so many days in advance. This forecast is still subject to great change, whether it be warmer or colder. After all, it only takes a few degrees to make it be more rain or more snow.

      Rest assured, we'll be keeping a close eye on it as the days approach.

      For the rest of your forecast, check out our weather page. Tuesday and Wednesday are looking to be rainy (and just rainy, not snowy).

      For other weather news, join me on Facebook and Twitter. Fall colors are pretty much peaking right now. I took the attached picture Friday in Marquette Township at sunrise. The sunrise really brings out the colors! Send us your pics through Facebook and Twitter!