With the first big snowfall of the year bringing with it closed roads in Saskatchewan, a bit of forewarning went a long way.

The snowfall warning sent out by Environment Canada was part of their forecast over the past few days, which had been calling for a good snowfall.

In order to make that prediction, they've collated and figured out how to interpret all of the readings they get from models regularly.

Environment Canada Meteorologist Natalie Hasel says that those aren't foolproof, and the work is a bit more involved than just reading from one model.

"Every model will be wrong in some way, some are more useful than others. so for things that area little bit longer term, looking at our range of ideas or suggestions, or solutions for the weather question is usually more useful than just using one model."

Those are called ensembles, enabling teams to combine the strengths and cover the weaknesses of varying weather models.

"In the public forecast as well we've used ensembles," said Hasel, "Which are systems that use multiple models to come up with an average answer. Those tend to do pretty well in the extended forecast."

While the programs do tend to make a good forecast, they also get a helping hand from humans when the forecast is in the near future.

"Our public forecast definitely has human involvement for the first two days. After that it's computer model driven," said Hasel, "There are very few instances where we'll actually have something fiddled with by a person past the day after tomorrow."