I am confused as to how a lake actually freezes. It appears to be dependent on air temperature. So the does the ice on top continue to build and freeze and push the ice deeper into the lake? Or, does the ice underneath freeze as it interacts with the unfrozen water, with the air temperature moving through the already frozen ice? Is the water temperature under the ice a constant?
Top down. The air is colder than the earth below.
An ice film forms on the water surface. The colder the air abover that ice, the more cold is “transmited to” the bottom of the ice, causing the water below it to freeze in successively deeper layers.
(The cold is not actually “transmitted,” rather, more heat is extracted from the ice and adjacent water by the colder air above.)
The temperature of the lake wil generally remain stable through the winter at some temperature above freezing. At some point, the ice, itself, acts as a barrier to the ability of the surface air to extract heat from the water below. This, of course, is subject to variables such as the size and shape of the lake (deeper is warmer, more concentrated shape is warmer) as well as the temperature and duration of the cold, above. There are Antarctic lakes that are sall enough and have been cold enough long enough to be frozen solid.
so as stated above, lakes freeze top down. Fish then can survive in their little water pocket below the ice.
-Lil
Other explanations are partly right and partly wrong.
Water has its minimum density at -4 deg. Hence the heavier warmer water is at the bottom of the pond. Cold water and ice at the top.
Cold air flows across the surface of the pond, cools the water by convection to establish a film of ice.
Heat moves upward throughout the water by convection and conduction, to build more ice on the bottom surface.
The heat then moves upward through the ice to the top by conduction to be removed by convection from the top surface.
This page has a more detailed explanation of the seasonal cycle of freshwater lakes and the semi-annual phenomenon of turn-over of a deep lake’s water mass.
If memory serves me correctly,
- The initial pond temperature is assumed to be >> 0 °C
- The air temperature (above the pond) is assumed to be < 0 °C.
- A pond loses heat via the top surface.
- Water has its maximum density at approximately +4 °C.
As the pond is cooling down to +4 °C, the colder water near the surface drops to the bottom (because it is more dense), and is replaced by warmer water that was previously hanging out at the bottom.
Eventually the entire depth of the pond is around +4 °C. But the pond’s top surface continues to become colder and colder. This water at the top surface does not drop to the bottom. This is because it is less dense than the +4 °C at the bottom of the pond.
Eventually the top of the pond is at 0 °C, while the bottom is somewhere between 0 °C and +4 °C; it thus starts to freeze. And because the air temperature is < 0 °C, the ice that forms on top of the pond will eventually be < 0 °C.
Not always. In exceedingly harsh winters, the ice can remain too long and the oxygen is depleted. The fish die.
The simple version:
If lakes froze from the bottom up, ice fishing would not be very productive. Skating, on the other hand, would not have any “thin ice” problems. 
To be more precise: When snow cover is sufficient to limit sunlight penetration, photosynthesis doesn’t occur.
This bears repeating:
It helps immensely that water is one of a very few substances that expands when it freezes - otherwise chunks of ice forming in the lake would sink, and many lakes would probably freeze solid. Most liquids shrink, and the solid state is denser than the liquid state. Background on why this is, which also touches on why water has a maximum density at 4 C, as mentioned:
http://www.iapws.org/faq1/freeze.htm
Many SF writers have touched on the fact that a planet containing, say, liquid methane, would likely have lakes that froze upwards. I don’t know if this phenomenon is expected to be confirmed on Titan (which may have methane lakes) or not.
As others have said, as water chills below 4 C. it becomes less dense and stays at the top where it solidifies as ice. Ice is also less dense than water and floats at the top of the pond.
This is a somewhat unusual property - normally liquids become denser right down to the freezing point, and the solid is more dense than the liquid. The solid will sink to the bottom.
But not ice, if it did, some have theorized that the oceans would freeze from the bottom up, killing the eco-system that thrives at the lowest levels. Even in the tropics there would be bottom layers that would not thaw. Earth would be a cold, dead planet.
This is true. I only meant to say that the fish may be able to survive. I thought that was obvious with the word “can” instead of “will”. And I swear I’m not trying to sound pissy.
-Lil
No sweat; just offering some additional information in the spirit of discussion.
Xurando- Where do you live that you’ve never seen any outdoor body of water freeze?
I’m quite intrigued.
Xurando, may, indeed, live where lakes never freeze. However, I read the question diffferently and interpreted the options as freeze from the bottom or accumulate ice (in the manner of frost) until the weight of the increasingly thick ice pushes down into the lake–a view that a person who had seen lakes freeze might hold if he or she had never actually looked at them over the course of a few very dry cold days.
I’m not being sarcastic in my comment (for a change). I’m just curious as to the reason for the question. I live somewhere where lakes, ponds, puddles, gutters, and rivers do freeze from time to time, and to be honest, I don’t think I’d miss the cold weather all that much if I either moved, or global warming got it’s ass in gear.
Um, er, lakes FREEZING?
Oh, yeah, ice cubes do float to the top, don’t they?
Never mind.