In the situation of an ice cube placed in water and eventually melting, it is supposed that natural convection is the process in which heat transfer from water to ice occurs.

However, natural convection in this situation requires a continuous replacement of water particles near the ice cube to transfer energy to the ice particles, as the water particles that transfer energy - lose their energy, and in the process increase in density. Or so, I thought.

But actually since the density of ice is less than water’s, this means that the energy depleted particles near the ice cube actually decreases in density.

So my question is, how is natural convection possible in this case?

Do the water particles move up in this case instead of sinking?

  • $\begingroup$ Can I see a link or simulation data of a melting of an ice cube in water? $\endgroup$ Apr 13, 2019 at 20:48
  • $\begingroup$ Heat transfer from liquid to ice does not depend directly on convection. See my answer to your previous question about the mechanism of heat transfer.. $\endgroup$
    – GiorgioP
    Apr 13, 2019 at 21:02
  • 1
    $\begingroup$ Ubaid, when the water molecules next to the ice get to 4 deg C, they become as dense as they can, and they sink to the bottom of the container. Warmer water takes their place. Also, note that all three forms of heat transfer always take place at the same time: convection, conduction, and radiation. $\endgroup$ Jun 1, 2019 at 2:19

2 Answers 2


Water is most dense at $T = 4\mathrm{^\circ C}$.

  1. If the water is colder than $4\mathrm{^\circ C}$, the small water volumes, that cool down till $0\mathrm{^\circ C}$ around the ice, expand, and their density decrease and go beyond the surrounding water's density, therefore, they move upwards.
  2. If the water is warmer than $4\mathrm{^\circ C}$, the small water volumes, that cool down till $0\mathrm{^\circ C}$, can be denser or sparser, depending on how warm and sparse the water around. But the water volumes at $0\mathrm{^\circ C}$ will warm up to $4\mathrm{^\circ C}$, meaning they are the most dense volumes, and start sinking.

In both cases there is a thin layer of water at $0\mathrm{^\circ C}$ that can move upwards. But in the second case, another, outer layer can move downwards, beyond the ice cube, providing the leading effect.

  • $\begingroup$ Well that’s that I guess, did you get the information of waters varying density at different temperatures from an equation by the way? $\endgroup$
    – Hisham
    Apr 13, 2019 at 20:07
  • 1
    $\begingroup$ It is a hardcore physical-chemical calculation to prove that water is most dense at $4^\circ C$, and it even relies on deeper empirical parameters. I don't think it is proven ab initio, that is, from the very basis of quantum mechanics. (Btw, it is science, so instead of guessing I should give you references.) $\endgroup$ Apr 13, 2019 at 20:10
  • $\begingroup$ An estimate of 4 degrees C is good enough, but for other stuff on water this is a great online reference: www1.lsbu.ac.uk/water/density_anomalies.html#density $\endgroup$
    – user137289
    Apr 13, 2019 at 22:21

For this question to make sense, the ice has to be floating on top of the water. The density differences that are important are between a cold liquid water layer immediately below the floating ice and the bulk liquid hot water below that layer. The cold liquid is denser than the hot liquid. Since it is above (in gravity), it will fall. The hot liquid will flow in behind. The hot water will interact by conduction (molecular collisions) with the cold ice. The hot water will become colder (because it loses heat) and the ice will melt (because it gains heat). The cycle of convective fluid flow will continue.

See my answer at this post for comments about the concerns when the ice would happen to be placed below the water.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.