It’s happening again. A Facebook picture is circulating showing record snowfall for this next weekend across eastern Iowa. This is probably the 5th one I’ve seen this year for a potential storm 7 days out. As your local meteorologist, I want to give you the facts about why you should NOT believe or share these posts.
First, here are a few facts:
- Models tend to over predict snow totals 7+ days out. The two main extended range models are the GFS and European. The GFS is notorious for overshooting snowfall forecasts 5+ days out. It typically narrows down totals the closer we get. The Euro is guilty of overdoing totals too.
- Storms that are 5 days away (or more) typically haven’t developed yet or are way off in the Pacific. Models do an excellent job of predicting the development of this low but final placement across Iowa can be WAY off.
- Models typically don’t come into agreement until, at the most, 48 hours out. When it comes to narrow, heavy snow bands, it may be within 6 hours.
- Snow forecasting is hard and dependent on many, many factors. Here is a blog post that touches a bit on that: Snow Ratios
Now, let me show you an example of how, just recently, these posts have failed.
The image is from a GFS model run posted February 25th. It doesn’t show which model run (there are four a day) and which parameter is being used. It does state that the dates are from March 1-5 and this is the snow total forecast.
Below is an image from a website used for forecasting. This is from the GFS model run started at midnight, February 24. This is how much snow the GFS (same model as above) was forecasting based on the 10:1 ratio (see previous blog post). So according to this GFS model run 7 days out, we should have woken up to 4-7+ inches of snow across eastern Iowa Saturday morning.
Did that happen? No.
Below shows how much snow fell in that 24-hour time frame:
Below shows the reason why we wait until closer to an event to forecast snow totals. The GFS had a much better handle on the system 24 hours out. Throughout the week, it lowered totals and shifted the system north.
You may be seeing posts circulating for this upcoming weekends storm. The satellite image below shows the low really hasn’t even developed yet. It’s still located well west of Hawaii.
Facebook posts are circulating that show 14+” snow amounts with little information. Below are how many times we’ve actually seen 10+” snow totals in Waterloo since 1895:
Here is what I can tell you confidently about this weekend’s storm.
- There will be a storm to impact the Midwest. Both long range models agree on this.
- Accumulating snow will happen…somewhere. Neither model agrees on amounts or whether we’ll see rain or snow.
- The KWWL Storm Track 7 team will keep you updated.