How accurate are forecast models 1 to 2 weeks ahead?

How accurate are the forecast models one to two weeks in advance? That’s one of the more interesting challenges weather forecasters face.

The science of numerical weather prediction has made great strides in the past 30 years. Significant improvements in forecast accuracy is a modern science success story. Today’s five-day forecast is now as accurate and the three-day forecast was about 25 years ago.

The first seven days of model predictions show significant skill. Forecasts up to seven days out usually provide actionable weather forecast information.

But it’s far more challenging for current weather forecast models to paint a clear picture between seven and 14 days in advance.

A seven-day forecast can accurately predict the weather about 80 percent of the time and a five-day forecast can accurately predict the weather approximately 90 percent of the time. However, a 10-day — or longer — forecast is only right about half the time.

Broad brush strokes: Long and short waves

The broader brushstrokes of medium-range forecasting involve accurately identifying wave positions in the upper air pattern. If the models get the mean positions of the longer waves right in the one-to-two-week range, forecast trends are much more accurate. Longer waves are called Rossby waves.

And temperature trends are more accurate than day-to-day precipitation trends. That’s because long-wave patterns have a larger-scale impact on temperatures over a region.

It’s the short waves that drive local precipitation. And they’re much harder to pinpoint one to two weeks in advance. The map below shows short waves embedded within the overall long wave trough.

Subzero by mid-January?

Medium-range upper air forecast patterns the past few days favor a deep trough diving into the central United States from Canada around mid-January.

If that pattern verifies, it would shove potentially subzero air mass south from the Arctic Circle. NOAA’s latest GFS 16-day temperature outlook suggests a subzero shot around mid-January that bottoms out at minus 23 for the Twin Cities.

Experienced medium-range forecasters don’t buy into absolute temperature forecast numbers one or two weeks out. But medium-range temperature trends are often more reliable. So, the notion of colder, subzero air in the next two weeks is had to ignore.

But forecast models have been pushing that event a few days further out for the past few runs. When forecast see that, we identify it as a lower confidence trend than if the models were consistent from day-to-day.

So, we can credibly say the one-to-two-week outlook favors some subzero air in Minnesota. But the magnitude and duration of that cold are still unclear.

As we say in the weather biz, stay tuned.