Lake-effect snow forms when cold air crosses a large, relatively warm, unfrozen lake, picks up heat and moisture, and releases it as snow downwind. In the United States it is most associated with the Great Lakes, where it is common in late fall and winter.
The ingredients
The first ingredient is instability over the water. A widely used guide calls for the lake surface to be at least 13 C (about 23 F) warmer than the air at the 850 millibar level, which roughly corresponds to a dry-adiabatic lapse rate.
Fetch, the distance the wind travels over open water, controls how much heat and moisture reaches the lowest layer of air. Forecasters look for setups that maximize it.
The height and strength of the inversion above the lake cap how tall the lake clouds can grow. In a National Weather Service (NWS) Burlington study of Lake Champlain, a subsidence inversion forecast to lower to about 2,000 feet was judged the main limiting factor in two cases that produced little more than flurries.
Why wind direction matters
Wind direction decides which shoreline gets the snow and sets the fetch. Winds also need to be fairly uniform with height. COMET training notes that directional shear under 30 degrees favors a single, well-organized band, and the Burlington study cited less than 40 degrees below the inversion for heavy lake-effect snow on Lake Champlain.
What the bands produce
The NWS notes that lake-effect clouds organize into narrow bands that can produce 2 to 3 inches of snow per hour or more. Because the bands are narrow, heavy snow can fall in one place while the sun shines a mile or two away.
Where the rules of thumb bend
Thresholds vary by lake and setup. The 13 C figure is common, while research summarized for the smaller Lake Champlain pointed to a difference of at least 15 C, and in one case a fetch only 12 percent of the lake's length was not enough for organized lake-effect snow.
The same process works beyond the Great Lakes. The American Meteorological Society notes the term is also used around the Great Salt Lake.
