It is my understanding that within a given medium warmer substances tend to rise and cooler substances fall, i.e. warmer air above cooler.
Assuming the same is true with water then in a lake the cooler water should fall to the bottom while warmer water rises to the top, correct? If so, I would expect that lakes would freeze from the bottom up. Can you please explain why this isn’t the case?
Thanks for your explanation.
In: 23
The ELI5 version is that the whole lake has to get to 0C first before any ice can form at all (with the top water getting cooler, falling down, warmer water rising, cooling, falling down, etc), then once the entire lake is sufficiently cold, ice crystals will start to form and rise to the top as they are less dense (and the cold air/wind at the top help to wick away heat).
For a really cool book about this topic that explains it in a ELI18-ish way, check out [Lakes: Their Birth, Life, and Death](https://www.amazon.com/Lakes-Their-Birth-Life-Death/dp/1643260480)
Edit: and the reason the whole lake doesn’t freeze is that the top layer of ice provides an insulating effect
It’s not that warmer things rise, it’s that warmer things tend to “spread out” as the molecules get more giggly, and that makes them less dense. Denser objects fall. Less dense rise. That’s why a helium balloon rises without being cooler. Helium is less dense than nitrogen and oxygen, which push the helium upward as they fall below it.
Ok so what about ice? Water has a particularly weird, rare property that it actually is less dense when it’s frozen. This because it’s crystalline structure is less more widely packed than when it’s liquid. This is why ice cubes float. Also why, when it freezes on top (the area exposed to cold air) the ice doesn’t sink to the bottom. Luckily.
Because if it did sink, then more would freeze and sink. And more. And eventually it would all be frozen and never melt. And earth would be a giant ice ball. With no life.
Read a book called Cat’s Cradle. It’s about people who invent something called Ice 9 that basically works this way.
>Assuming the same is true with water then in a lake the cooler water should fall to the bottom while warmer water rises to the top, correct?
Yep, that does happen. The topmost layers of water get cooled first, but they don’t cool all the way to freezing point. They cool enough for their density to increase enough to sink to the bottom, but it hasn’t frozen yet. When that ‘layer’ sinks, the water just below it rises to the top, and the same thing happens again – it cools down enough to sink, but not until freezing point. This happens over and over until eventually, all of the water body reaches a temperature close to freezing point. Then, the uppermost layer of water becomes cold enough to go to the freezing point (naturally, since it’s directly in contact with the air to which it’s losing heat). So the uppermost layer freezes. Now for the slightly more confusing part – normally solids tend to be denser than the corresponding liquids, but in the case of water and a few other liquids, ice is actually *less dense* than water. As a result of this, the ice floats on the water, not sinking to the bottom. Ice is also a great insulator of heat, meaning that it doesn’t allow the transfer of heat very easily. So the water below the ice (which is close to freezing point), actually remains liquid without freezing, because the ice above doesn’t allow the remaining heat to flow out. So unless it gets *really* cold, normally you just have a layer of ice on top and just water beneath it. That’s why aquatic animals like fish can survive cold winters in lakes and ponds.
There’s a *bit* more nuance involved about *why* ice is less dense than water and about the exact temperatures which the water body is at, but that’s the basic explanation for why it happens the way it does.
The air temperature can drop before the water temperature, and as a result, the surface of the water can freeze first where it is in contact with the cold air. Although warm water has a tendency to rise, once it freezes and turns into ice, it has a lower density compared to liquid water and therefore, it floats on the surface.
Water is at its densest at 4°c and gets less dense as you approach 0°c leading to the odd effect of bodies of water freezing from the top down as opposed to bottom up as that denser 4°c water settles at the bottom and the cooler 0°c water ends up at the top.
As water freezes, it expands. As it expands it takes up more volume with the same mass, so ice floats. So even if ice formed at the bottom it would float and collect at the top then freeze solid.
The top of the water has the most opportunities to shed heat compared to further down, so it’s the most likely place to freeze as well. But for the most part it’s really just that ice floats.
A lot of answers attribute it to the liquid density anomaly of water, that it is densest at 4°C. This is not the reason at all though. What somewhat matters is that ice does not sink, but even that is not that important as I will explain below. The same freezing from top would happen with almost any liquid and I have seen it with a variety of molten metals as well! You can actually compare it to sauces getting a skin layer on top first, if you want.
Reminder for the American people reading this, just to keep the following understandable: 0°C is where water freezes.
So what is actually going on? The **air is what gets below 0°C first**, both water and ground take a long time to catch up; in most areas the ground never really freezes below a few centimeters in, actually. The **cold air then cools the surface** of the pond/lake, while everything else is warmer than that.
It is important that **water can only freeze where it is colder than 0°C**, as freezing produces energy; water left at 0°C and perfectly isolated will never turn into ice (you could even go a bit below 0°C).
So where does it freeze? Well, only where it is cold enough: at the top. And unless the frozen liquid sinks down (it won’t for water, as ice floats; but even many other ones won’t as the surface tension and the rigidity of the solid version can keep it on top) **nothing will bring anything colder than 0°C downwards**. And as mentioned above, **liquid water cannot just spontaneously freeze at 0°C**, as that would produce energy and thus heat!
Even if water would sink at 0°C to the bottom, all that does is gradually cooling the entire lake to 0°C, then the above process starts. In actuality, the density anomaly causes a nice and cozy 4°C bed at the bottom for life to survive (this is why the anomaly is really important!), and then freezing at the top begins. The ice layer even makes a reasonable insulator towards that cold air. Regardless, only when this circulation cannot keep up with distributing the temperature anymore is where freezing starts.
The very same happens in any liquid. If the solid version does not float on top, it could however happen that it forms small clumps that then sink down and accumulate at the bottom, effectively filling it up. There is still some liquid in-between the clumps until the freezing form the top gets there.
Until now I have blissfully ignored the ground as if it never freezes. It may happen, though. This causes one mayor change: now also the bottom of the pond (for a full-on lake, this is extremely unlikely or just impossible) gets below 0°C. This causes freezing from the sides and bottom as well. And as mentioned for non-water liquids above, even with the density change it often will just stick there. In this case, the pond freezes from the outsides inwards.
**tl;dr**: unless the ground is rather heat conductive (compared to the liquid), that lake could just as well be alcohol, ammonia, liquid copper or liquid nitrogen; it would still freeze from the top.
A lot of answers attribute it to the liquid density anomaly of water, that it is densest at 4°C. This is not the reason at all though. What somewhat matters is that ice does not sink, but even that is not that important as I will explain below. The same freezing from top would happen with almost any liquid and I have seen it with a variety of molten metals as well! You can actually compare it to sauces getting a skin layer on top first, if you want.
Reminder for the American people reading this, just to keep the following understandable: 0°C is where water freezes.
So what is actually going on? The **air is what gets below 0°C first**, both water and ground take a long time to catch up; in most areas the ground never really freezes below a few centimeters in, actually. The **cold air then cools the surface** of the pond/lake, while everything else is warmer than that.
It is important that **water can only freeze where it is colder than 0°C**, as freezing produces energy; water left at 0°C and perfectly isolated will never turn into ice (you could even go a bit below 0°C).
So where does it freeze? Well, only where it is cold enough: at the top. And unless the frozen liquid sinks down (it won’t for water, as ice floats; but even many other ones won’t as the surface tension and the rigidity of the solid version can keep it on top) **nothing will bring anything colder than 0°C downwards**. And as mentioned above, **liquid water cannot just spontaneously freeze at 0°C**, as that would produce energy and thus heat!
Even if water would sink at 0°C to the bottom, all that does is gradually cooling the entire lake to 0°C, then the above process starts. In actuality, the density anomaly causes a nice and cozy 4°C bed at the bottom for life to survive (this is why the anomaly is really important!), and then freezing at the top begins. The ice layer even makes a reasonable insulator towards that cold air. Regardless, only when this circulation cannot keep up with distributing the temperature anymore is where freezing starts.
The very same happens in any liquid. If the solid version does not float on top, it could however happen that it forms small clumps that then sink down and accumulate at the bottom, effectively filling it up. There is still some liquid in-between the clumps until the freezing form the top gets there.
Until now I have blissfully ignored the ground as if it never freezes. It may happen, though. This causes one mayor change: now also the bottom of the pond (for a full-on lake, this is extremely unlikely or just impossible) gets below 0°C. This causes freezing from the sides and bottom as well. And as mentioned for non-water liquids above, even with the density change it often will just stick there. In this case, the pond freezes from the outsides inwards.
**tl;dr**: unless the ground is rather heat conductive (compared to the liquid), that lake could just as well be alcohol, ammonia, liquid copper or liquid nitrogen; it would still freeze from the top.
>cooler water should fall to the bottom while warmer water rises to the top, correct?
Not correct. While it’s true in general that substances expand as heated, it’s not always so straightforward near phase change boundaries. Ice is less dense than water, that’s why it floats. And liquid water is [not at it’s densest at 0 degrees, but at around 4 degrees](https://www.open.edu/openlearn/pluginfile.php/471504/mod_oucontent/oucontent/18543/ddd2c58c/6625d7f1/s206_blk1_part5_f3_01.eps.jpg).
>cooler water should fall to the bottom while warmer water rises to the top, correct?
Not correct. While it’s true in general that substances expand as heated, it’s not always so straightforward near phase change boundaries. Ice is less dense than water, that’s why it floats. And liquid water is [not at it’s densest at 0 degrees, but at around 4 degrees](https://www.open.edu/openlearn/pluginfile.php/471504/mod_oucontent/oucontent/18543/ddd2c58c/6625d7f1/s206_blk1_part5_f3_01.eps.jpg).
Latest Answers