Challenges Forecasting the Snow Level

The snow level is a critical variable for mountain weather forecasting, but it is both difficult to define and predict.  

In fact, there isn't even an agreed upon definition of snow level amongst meteorologists or the public.  One might say it is the level above which precipitation is snow and below which precipitation is rain, but its just not that simple.  Precipitation doesn't just change from snow to rain at a given altitude.  Snow takes time to melt as it falls.  Different types of snow crystals melt faster than others.  Sometimes it is snowing but not accumulating on the ground due to melt.  What do you do when there is a mixture of snow and rain or drizzle?  Where do you put the snow level? 

To begin, perhaps it is easier to begin with a description of the transition zone, the layer in which snow is warming, melting, and turning to rain as it falls.  

The schematic below provides a simple conceptual model of the transition zone.  The freezing level marks the top of the transition zone and is the highest level at which the temperature is 0°C.  Temperatures above this level are below 0°C.  

The transition zone (h/t Ron Stewart)

The freezing level and the snow level are not the same because snow does not melt instantly when it falls below the freezing level.  Instead, it begins a transition as it falls, gradually changing from dry snow to wet snow to slush (part ice part water) and finally rain.  The distance over which this occurs depends on many factors including the profiles of temperature and relative humidity below the freezing level, type and size of the snow crystals, and fall speed of the snow crystals. 

For example, if the relative humidity is low, a given snow crystal can penetrate farther below the freezing level since the crystal will warm more slowly because it is also cooled by sublimation.  A dense graupel particle can generally penetrate farther below the freezing level because they have a high fall speed and take longer to warm and melt fully compared to a small dendrite.  

In addition, because the melting of snow cools the atmosphere to the melting point (0C), there is often an isothermal 0C layer in the transition zone, as depicted above.  This layer can be quite deep, especially if precipitation rates are high.  Because of this effect, the snow level may lower some when the precipitation rate increases and rise some when the precipitation rate decreases. 

In the above schematic, I have denoted the level at which the precipitation is half snow and half slush as the snow level, but that may not conform to what you think is the snow level.  Perhaps it should be lower and at the level where precipitation is all rain or all rain or slush.  Alternatively, perhaps it should be higher and at the level where precipitation is all snow?  

Or, perhaps instead of focusing on the type of precipitation falling, we focus instead on the level at which snow is accumulating on the ground.  That's sometimes called the snow line.  But if we do that, what exactly is the ground?  Grassy surfaces?  Asphalt surfaces?  

There are many ambiguities.

As discussed in Van Cleave et al. 2011, about 15 years ago the National Weather Service Western Region actually established a definition of snow level for their applications:

"The snow level is the elevation above which snow will fall, and below which rain will fall. A mix of rain and snow may be observed at elevations within a few hundred feet of the snow level. Snow will not accumulate on the ground below the snow level and may not even accumulate at elevations above the snow level."

Even that definition is somewhat ambiguous, but they also defined techniques for calculating and forecasting the snow level and identified the wet bulb 0.5°C level as the best proxy for snow level in the western region. That said, there can be variations in precipitation type and accumulations on the ground around this level.  

One situation that can be particularly challenging involves stable layers in which wet-bulb temperatures are very near 0°C through a deep layer.  An example is provided by last night's GFS forecast for 1000 UTC 01 January 2026 (3 AM MST New Year's Day).  At that time, a precipitating cloud layer has spread over northern Utah in southwesterly flow aloft, but stable conditions remain at low levels as the cold air currently in place over the Salt Lake Valley is slow to mix out.  The freezing level in this case is located just above 750 mb, or near about 2500 m MSL (~8200 ft MSL).  Because the atmosphere is saturated at that level, the freezing level in this case also represents the wet-bulb zero level. However, below that level, the layer is very close to 0°C down to almost 800 mb.  This might enable snow to penetrate farther below the freezing level than if temperatures were increasing with decreasing altitude below the freezing level.  

Layers in which temperatures are relatively constant with height are called isothermal, which means having a constant temperature. Having a deep isothermal layer near 0°C is one way for the snow level to penetrate to low elevations.  One needs to keep a close eye on these layers as they can make a big difference for snow level forecasts.