How snow forecasts work
Meteorologists forecast total accumulation, but the physics that determines the number involves cloud microphysics, freezing levels, and rain-snow line movements that models still get wrong.
The two hardest parts
The models
- GFS — global reference. Runs 4x daily. Snow ok, not great.
- NAM — North American Mesoscale, 12 km CONUS. Reasonable snow forecasts.
- HRRR — 3 km convection-allowing. Great for snow squall detail.
- ECMWF — European model. Often the best for medium-range winter forecasts.
- Canadian GEM — sometimes best for northern US winter storms.
- Ensembles (GEFS, EPS) — probabilistic. Show forecast spread.
Why timing matters more than amount
- A 6-inch snowfall during rush hour cripples a city.
- The same 6 inches overnight barely disrupts anything.
- Meteorologists know they'll be judged on the AMOUNT, but timing is often more useful.
- When school districts read forecasts, they need timing (7 AM vs 3 PM) more than exact inches.
The rain-snow line problem
The most common forecast bust: rain-snow line ends up 20 miles east or west of forecast.
- A ridge in temperature can push the line 50+ miles overnight.
- Coastal storms are especially prone to this because ocean warmth pushes the line inland.
- Nor'easters famous for this — Philly can get 2 inches or 20 inches depending on the exact track.
- Ensemble spread quantifies this uncertainty.
The snow ratio problem
The 10:1 ratio taught in intro meteorology is a rule of thumb, not a law.
- Warmer storms (30°F+): 5-8:1. Heavy wet snow.
- Standard storms (20-30°F): 10-15:1.
- Cold Alberta clippers (10-20°F): 15-20:1. Fluffy.
- Extreme cold (below 10°F): 20+:1. Extreme fluff.
- Dendritic growth zone (10-15,000 ft) temperature drives this.
- Forecasters look at model soundings for that layer.
Why some storms overperform
- Convective banding — thin heavy snow bands can drop 3+ inches/hr.
- Lake-effect enhancement.
- Frontogenesis — atmospheric front strengthening produces sudden bursts.
- Isentropic lift — gently rising air on strong warm air overrunning.
Why some storms underperform
- Warm layer aloft turns snow to sleet.
- Ocean warmth from storm track being 50 miles farther east.
- Model overestimated moisture.
- Dry slot arrives faster than forecast.
- Precipitation missed the metro entirely, hit rural areas.
Reading a winter storm forecast
When forecasters "hype"
Local TV meteorologists sometimes get accused of hyping storms for ratings. But most operational meteorologists take their responsibility seriously. When a forecast says HIGH SNOW, and you get 3 inches, the forecast wasn't hype — the storm shifted 40 miles.
On social media, weather hype is different — screenshots of model runs 8 days out get shared as forecasts. Those are not forecasts; they're model output.