Snow Melting in Freezing Conditions

By Karel van der Voort

Oegstgeest, Holland.  June 2013

Melting and (re)freezing of snow can occur even if the temperature is continuously below zero deg C. For instance, if the top surface of the ice sheet is pushed down by the weight of snow to the extent that water can flow up through cracks, pores and holes in the ice and soak the snow.   If refrozen, these water-soaked layers of snow can form a new ice layer or path of smooth ice.

Sometimes it is possible to skate completely around a snow covered lake or sea area along the shore or across a lake when water has flown through a crack or cracks which have formed a skateable path of frozen slush formed in this way.

A refrozen shore lead (mote) along the edge of a lake can provide a smooth skating path.

Snow that wicked up water from a wet crack.  The slush then froze making a hard, smooth passageway through snow covered ice.

This can provide skateable ice without hindrance from snow even in the absence of a thaw.

As is well known, snow wicks water due to capillary action driven by the hygroscopic surface of the snow crystals , open structure of the snow.

Submerging an ice sheet with snow:
How much snow is needed to push a sheet of ice under allowing water to get to the upper surface of the ice sheet? Perhaps one could have some idea by the fact that the bulk density of solid ice is about 0,9 kg/l, and the bulk density of newly fallen snow is generally about 0,1 kg/l. If so, 10 cm of ice floats 1 cm above the water surface. And it needs 10 cm of snow to press down the ice that much. That ought to be less as the snow becomes compacted overtime.

Also, the less solid or dense the ice (from air pockets from snow structure or caused by internal thawing), the less snow (weight) is needed to push the ice under.

Other ways snow turns to ice:

Practice based on large numbers of observations has shown, that the 'pushing under' of ice is not the only factor that can saturate snow with water at air temperatures below freezing. Other factors include:

  • sources/isolators of warmth or shadow areas, like rows of trees or rocks alongside the lake, buildings, bridges etc.
  • The snow itself; it insulates the ice reducing the amount of freezing of the slush layer significantly. Perhaps one has noticed these wet, soggy patches of snow on ice when the temperature is low and wondered why these areas are not frozen solid, even at -10 deg C or below,
  • The thickness/quality of the ice plays another role - the thinner the ice, the faster the snow on it will melt. Particularly solid or black ice is not a very good insulator for the water under the ice. Spots of thawed snow on a in snowy ice surface are usually an indication of very thin ice. The faster freshly fallen snow melts on ice at air temperatures below zero, the thinner the ice and the less likely that it will be strong enough.
  • Open spots in the ice, you will find that the wind causes waves in an open area with water, pushing some of it to come onto the ice. The same thing happens with waves in big lakes or sea areas that are partially frozen.
  • Strong currents in rivers and narrow areas
  • Long days/high angle sun exposure
  • Salt in sea/brackish water areas that can cause the snow to melt on the surface at a few degrees below freezing.

White ice is usually either snow ice (frozen slush) or splash-out ice.  Both can be an excellent skating surface when they are cold and smooth. In come cases a thick crust of saturated snow ice over partially saturated snow ice can be skated on as well.  Snow ice usually gets soft thaw conditions faster than black ice.

 Water on ice

It is a very common misperception that rain is a negative factor for ice, and that it causes significant thawing. On the contrary, a uniform layer of water on ice has, to a certain extent, a preservative effect. A water layer on the ice will be cooled to near freezing by the underlying ice.  Additionally water can evaporate and cool the water further if the dew point of the air is below freezing.

Under certain circumstances, such as sufficient wind, low humidity and, in particular, a mostly clear sky, it is possible that at temperatures above freezing so much evaprative cooling and IR radiation from the water occurs that a thin water layer will freeze solid. There have also been numerous cases of melting snow on ice in thawing weather where the ice slush froze more or less solid under a clear sky.


 A layer of water keeps the ice surface cool and if it contains snow slush, the open pores get filled, leading to a very smooth ice surface. Generally, ice under a universal or consistent skim of water remains surprisingly hard in thawing conditions, even if a slushy surface is expected and after days of rain.

Also, a consistent layer of water is an indication that the ice hasn't become very porous which is a feature of ice in an advanced thawing condition. That generally means, it has retained most of its strength.

Fog from high humidity in overcast conditions

Thawed Ice:

Thawed and porous ice with a dry surface as any melt-water flows down through the pores in the ice. The depressions are from leaves that absorb the strong sun creating local melting.  They typically melt to below the lake water level. 

This is at the end of a period of excellent skating ice on the Mirowsee, near Mirow in northeastern Germany.  The ice is about 12 cm (4-3/4") thick.  It went out a few days later in a windy, warm storm that lasted a couple days.

Partially thawed small grain ice.  The gain boundaries are visible because of bubbles from dissolved air in the boundary layer ice.  The reflection of Karel provides some scale.

Partially thawed large grain ice.  Melting takes place at the grain boundaries first. The bigger crystals have dimensions of several inches.

A dusting of snow that fell on porous ice (probably with a thin 'overnight ice' skin where the snow has not melted).  Areas where the snow on an ice sheet has melted are, in some circumstances, thinner or weaker.


About the Author

Karel on excellent skating ice near the boarder of Sweden and Norway

  He is 47 and has been skating since the age of 10, starting on a canal near his house.  He spends the ice season finding good skating ice in Holland and elsewhere around Europe.  He, like many of us, is trying to make practical sense out of the complexity of ice.






North American Fatality Report for the 2013 Season


Open water in the east end of Rangeley Lake (Maine) from Rt 4 on January 1, 2013. A couple days before this, five snowmobilers, riding at night, drove onto the open water. One rider survived.

There were at least 53 ice fatalities in North America last winter.  The following are links (in red text) to a summary report and short reviews of each incident (by month).  The circumstances that lead to all this misfortune tells the ground truth about ice safety.  The Summary Report, in particular, is recommended reading.

Summary Report

November and December 2012

January 2013

February 2013

March and April 2013


Ice Shove Season Is Not Quite Over

There have been a number of spectacular ice shove events in the midwest this winter.   

Mille Lacs Lake, WI- May11, 2013:  Darla Johnson's video of an ice sheet coming in is a great video of where the ice meets structures. 

Mille Lacs Lake WI Aftermath May 12, 2013:  A video look at the same ice the next day. 

Dauphin Lake, Manitoba--May 10, 2013:  The same windy weather seen at Mille Lacks Lake drove an even more distructive push. 13+ vacation homes were distroyed.   

Lake Winnebago, WI: January 19, 2013:  An unusual shove in minimally thawed, 12" thick, mid winter ice.  Usually ice this strong resists the wind.  The shove looks like it took place as the wind picked up after a cold front passed.  Cold ice shoves typically have blocks that have dimenstions 2-5 times the thickness.  Warm ice shoves have blocks that are 1-2 times the thickness. 

Lake Winnebago, WI, April 10, 2013:  A warm ice shove driven by North-east and east winds affected the west shore (usually the east shore of the lake sees most of the shoves). 






More Maquam Bay Ice Shove

The following pictures are from the day after the push or a week later.

The next day (before much decay has taken place)

The same pile a week later. The top of the pile is about 5 feet lower.



Piled ice a day after the push

 A week later most of the ice is gone. Maybe half of the volume went out in the previous 12 hours as high winds pushed 3 foot waves along the shore line. ( the picture was taken a few feet to the right)

A week after the event the partial melting of the piles has make their structure a bit easier to see

Another example of an unstable vertical scarp in the ice pile at the water's edge (a week after the push). This scarp is roughly 8-9 ft tall. A day later 3 ft waves washed most of ice pile away (see 3 pictures above).




Late Season Ice-Shove Event

Ice-shove is a relatively uncommon (but not rare) phenomena on big lakes like Lake Champlain (VT),  Lake Winnebago (WI) and Utah Lake (UT).  They are also  called ice-push although that generally refers to the slow, rachetting, thermal pushing of ice into shore or objects in the lake .   The ice is shoved by strong winds and/or collisions of wind blown or current driven ice sheets.   Ice shoves often result in ice piles The pile formation process is quick (on the order of 15 minutes for a big pile) so it rarely witnessed.  We had a particularly spectacular ice-shove event on April 7 when forty+ mph winds drove a large plate of well thawed ice back into Maquam Bay.    The following are a couple of pictures and a map.  More will be posted in the next few days.

The base of the pile is about 6 feet above the lake. Total height roughly 15 ft. Rain on the way

Looking north toward the back of the bay.

The ice in Maquam Bay had been pushed south over the previous week and on Saturday it was in position to be pushed back north by the strong winds coming Sunday. As it rubbed against the east side of Hog Island the ice overrode itself, forming piles.

More on this ice shove a week later