Picking weather models: An inside look at choosing a snow forecast, what it means for next storm

Not all computing is the same

An radar image of Michigan at 7:41 p.m. on Feb. 4, 2021 (WDIV)

DETROIT – The ears of Detroiters perk up when snow is in the future. That’s when it’s up to the Local4Casters to provide on a top notch weather forecast. Computer models are a significant forecast for analyzing any prediction.

As of Tuesday evening, snow is over a day and a half away. That may be soon to some. It is an eternity when it comes to numerical models and their digital prognostications. Meteorologists take their education and experience to decipher which computer models are most relevant when figuring out the time snow will make its appearance the day after tomorrow.

Related: Wind, flood advisories, then winter storm watch in Metro Detroit: Significant snow coming Thursday

Here is a breakdown of this week’s computer models and decision to go with one versus another or combining elements of different models to help Southeast Michiganders get prepared not scared:

A Tale of 2 Sets of Models

Two computer models, the Global Environmental Multiscale System (GEM) model and the European Centre for Medium Range Weather Forecast (ECMWF) model have a cold front developing and racing through the Motor City region Thursday morning. As it does, steady rain showers quickly change over to steady snow showers with not much sleet or ice. Snow lasts through much of the afternoon with this model as temperatures plummet. The rapidly sinking mercury and persistently falling snowflakes combine to provide a significant amount of snow for most of Southeast Michigan from northern Oakland and Macomb Counties south through Detroit and Ann Arbor to the Michigan-Ohio Border. By the end of Thursday evening, these models predict 3 to 8 inches of snow with much of it falling south of 8 Mile Road.

Two other computer models, the North American Mesoscale (NAM) model and the Global Forecast System (GFS) model have the same cold front but take its origin (called frontogenesis) and trajectory in a very different way that results in drastically different snowy outcome (or not so snowy). These two models have the cold front forming and remaining farther west and north. It has the same drop in temperatures at the surface where people, cars, trucks and houses are but temperatures aloft (inside clouds) either fluctuate near or are above freezing. This results in a longer period of transition from rain to snow Thursday with a greater chance of freezing rain and sleet. This means the full transition to snow occurs later rather than sooner in Southeast Michigan. In fact, it snows sooner and longer not just north of 8 Mile Road but north of M-59 and into the Saginaw Valley. So, greater amounts of snow that are 6 to 12 inches are possible in The Thumb and the Saginaw Valley according to these models. The projected prolonging of wet and icy conditions reduces the snow totals in Southeast Michigan to 1 to 4 inches.

The Choice

Here’s what goes into choosing which models may make more sense for any given time.

Initialization

In real estate it’s location, location, location. In synoptic meteorology and forecasting its initialization, initialization, initialization. Initialization is the starting time for any model when it is produced. Numerical models for weather are a simulation. They simulate how the weather may be in the future. At their initial point, they are also simulating current conditions. For many weather forecasters, a key indicator for whether a weather model may make a good prediction is how well the match is between that model’s simulation of current conditions and reality. The greater the match, the better that model may be at matching the future state of weather.

As of now, the GEM and ECMWF models start with environmental conditions that are a better match to conditions on the ground at the time of their production. The NAM and GFS do not match up as well.

Consistency

Consistency is another factor. Computer models are produced not just once a day. Many are produced as much as two, four or eight times a day. The High Resolution Rapid-Refresh (HRRR) model is produced hourly (up to 24 times a day). Each production is called a “run”.

QUICK ASIDE: It takes a tremendous amount of computing power and time for each run of a weather model. Generally, the less frequent the run the further out in time a model can predict. Conversely, an hourly model like the HRRR can not look ahead as far out in time as the ECMWF, GEM, GFS or NAM models. Their runs occur two to four times a day. Hence, they can peer two to ten days in the future. The HRRR looks just 24 to 36 hours into the future and stops.

Usually, a weather model that produces similar forecasts for the same time period with each run can engender more confidence. In this week’s case, all four models and their progressive runs have been quite consistent. The ECMWF and GEM keep higher snow totals more south into Southeast Michigan. The NAM and GFS keep greater snow accumulations north and west of Southeast Michigan.

The conclusion and the caveat

Therefore, with consistency being equal and initialization as the distinctive factor, there is greater confidence in the GEM and ECMWF models that have a faster transition from rain to snow over Detroit and Southeast Michigan Thursday resulting in 3 to 8 inches of accumulation.

That is a conclusion but only for now. The caveat is that there remains plenty of time between now and Thursday. The frontal system that will produce mild weather and rain changing over to cold weather and snow and ice has not even formed, yet. Once it does and as subsequent model runs are completed, a clearer picture is formed for when and how much snow will fall.

The newest model runs arrive when many are asleep, and we’ll go over them with a fine-toothed comb and have our latest forecast for you on TV and our digital platforms before the alarm clock sounds.

So, as with all things in media, stay tuned. Whatever the outcome, we are here for you. We will provide the latest information with our most up-to-date forecast to help you and your family. Our goal is to make sure everyone is prepared and not scared.


About the Author:

Andrew Humphrey is an Emmy Award winning meteorologist, and also an AMS Certified Broadcast Meteorologist (CBM). He has a BSE in Meteorology from the University of Michigan and an MS in Meteorology from the Massachusetts Institute of Technology, where he wrote his thesis on "The Behavior of the Total Mass of the Atmosphere."