The psychology of snowfall forecasting

One of the most challenging weather events to forecast is snowfall. The amount of snow that falls during a certain period of time is one thing, and the spacial distribution of that snowfall is another. In most snow-producing weather situations here in southeast Wisconsin, we get an unequal distribution of snowfall, i.e. some areas get more snow than others.

We depend heavily on computer forecast models to give us an idea of the amount of moisture available to be converted to snowfall as well as the vertical motion at various layers of the atmosphere. Some of our heaviest snowfalls occur when the greatest upward motion in the atmosphere occurs in the best dendrite snow growth region. This region is between -12° C. and -16° C. where dendrite crystals are the largest. This is an efficient way to grow snowflakes.

Another factor is the temperature in the layer with best snow growth. This determines the snow to water ratio of the storm system. The most common snow to water ratio we hear is 10:1, meaning 1 inch of rain forms 10 inches of snow. But this is a general average. Each weather system is usually different, and the ratios can change during the passage of the storm system.

An initial surge of lift and moisture can create a heavy, wet snowfall with ratios of 7 to 1. When cold air begins to infiltrate the lower levels of the atmosphere, the ratios may increase to 20 to 1, resulting in a transition to a lighter, fluffier snowfall. These changing snow to water ratios can change the total amount of snow that falls from the storm.

I've given you a quick summary of some of the meteorological factors that we look at when forecasting snow totals. Now comes the human factor: we must consider how the snow forecast will be perceived by our viewers.

If I forecast a snowfall of 3" to 6", most people only hear one number: 6. In their minds, this becomes the snow forecast. In the same way, if I forecast 1" to 3" south of a Milwaukee-Madison line and 3" to 6" north of a Milwaukee-Madison line, most people only hear one number: 6". And if The Weather Channel puts a map on their air that shows up to 12" of snow falling across Wisconsin, we hear it from many people that 12" will fall here in our corner of the state, even if the 12" is forecast to fall in northern Wisconsin.

Because of this "high number bias", as I call it, I find myself tending to forecast lower snowfall totals. An example would be if I see the potential for 6" of snowfall in some parts of our viewing area, but 2" to 4" will fall in most of the area, I hesitate to even mention the 6" potential. If the accumulation map I show on TV has that number 6 on it, that is all most people need to see to think all of us will see a half a foot of snowfall.

Forecast snow totals can also change and modify our forecast as a storm system approaches. This is common since moisture content, areas of maximum vertical motion, and surface/upper air storm track can change. The atmosphere is dynamic. Think of it as a giant blob of jello that is constantly changing appearance.

I cringe when I hear a forecaster say "later this week we could see a big snowstorm with 6 inches or more of snow." That statement stays in a viewer's mind for the next 4 or 5 days until the storm arrives. By that time the storm track may have turned 50 miles farther north and we get rain and no snow. People remembering what was said earlier in the week will never really hear the subsequent forecast updates. They expect snow and no rain. In their minds the forecast was a bust.

When I have the possibility for snow in the 4th, 5th, or 6th day of our 6-day outlook, you will often hear me say "the next storm later this week could bring us some accumulating snow, but at this point it is too soon to tell. We'll update you as the storm system gets closer." That's not hedging my forecast, that's reality.