From what I've read, the determination of a colour and alert type is now quite subjective and can be triggered by a wide range of largely unmeasurable criteria. This may increase confusion in understanding what a particular alert means in terms of actual weather.

I believe it's just being referred to as the "experimental" GDPS. It's planned to replace the current operational GDPS next year once the supercomputers are migrated. More info here, along with links to data access: eccc-msc.github.io/open-data/ms...

Yes, that is a bigger concern than ever. If we want to train AI models straight from observations, we'll need...observations!

Thus, how well they handle the details of high-impact weather events remains an open question.

One final note on accuracy. On paper, AI models are more accurate than traditional (physics-based) weather models. However, meteorologists still have limited experience using them to produce forecasts.

Although, I say "pure" AI model in quotes because there are no truly pure AI models. All AI models still rely on traditional modelling for some aspects of the forecast process.

Unlike most AI weather models, this model is a "hybrid" type. It uses an AI model to "correct" the traditional physics-based weather model, improving its accuracy. A key advantage to this approach is that it's higher resolution and has more outputs than "pure" AI models.

The farther you are from downtown the more likely you already had frost earlier in the month. Winnipeg is large enough that there is a noticeable urban heat island effect - especially in the shoulder seasons.

A growing season of 175 days is comparable to normal growing seasons in:
- Toronto, ON (174 days)
- Yarmouth, NS (174 days)
- Abbotsford, BC (181 days)

The growing season at The Forks was 23 days longer than at the airport (152 days) this year. It was also 54 days longer than the normal growing season length at the airport of 121 days.

It's definitely one of the cold spots around the city, but it accurately represents its location which is semi-rural. For that reason it's a good location for Winnipeg because it more closely matches the historical records, which would've also been in a much less urbanized environment.

However, The Forks only recorded a low of 2.4 C, meaning urban Winnipeg largely escaped the freeze. Due to Winnipeg's urban heat island, the growing season in the urban core is almost one month longer than at the airport.

You are likely seeing the line of storms that is about 20 km east along highway 12.