This is an example of impedance mismatch. Acoustic impedance - a measure of the way a sound wave interacts with the medium it is passing through - varies depending on the medium concerned. When a wave encounters a medium with a different impedance from the one it is in, most of its energy will be reflected at the boundary.

Most people will be familiar with ultrasound scans of a developing fetus. For these scans, a gel is applied to the expectant mother's skin to reduce the impedance mismatch between her body and the transducer and so maximise the transfer of acoustic energy.

Despite the acoustic mismatch between air and water, our everyday experience is that sound passing through the air manages to reach receptors immersed in the fluid of the cochlea within the inner ear. For this, we can thank the design of the middle ear for impedance-matching the sound.

Vibrations are passed from the eardrum via the auditory ossicles - the three smallest bones in the human body - to a membrane in front of the cochlea called the oval window. The eardrum is attached to the first of the ossicles, called the malleus (or hammer). This is pivoted about the second bone, the incus (or anvil), which in turn is fused to the third bone, the stapes (or stirrup). The stapes drums on the oval window.

The eardrum has an area about 15 times that of the oval window. As the sound energy striking the eardrum is concentrated at the oval window, the amplitude of the sound vibrations is increased. Lever action further boosts this amplification. Without the middle ear ossicles, barely 0.1 per cent of the energy arriving at the eardrum would reach the inner ear.

Mike Follows, Willenhall, West Midlands, UK

With her underwater and you out, most of her sound energy is bouncing off the under-surface of the water and is not coupling into the air--that's what the moderator means by the mysterious and undefined phrase "impedance mismatch".

In contrast to the air, water is a nearly incompressible fluid and so sound transmits through it very efficiently, but with only a small amplitude that, once it reaches the surface, does not jiggle the air very efficiently.  You'll only hear a faint sound. 

  This is simply attributed to the difference in acoustic impedance between air and water.  Acoustic impedance is a material property related to density that quantifies the resistance to sound wave propagation through the medium.  Acoustic impedance is inversely proportional to density because the closer the particles of a substance are together the shorter distance they have to travel to bump into the adjacent particle and thus propagate the sound wave. When dealing interfacing mediums, such as the barrier between water and air, the amount of sound energy transmitted and the amount reflected is a ratio of the relative acoustic impedances of the two mediums.  In this case liquid water is significantly denser than air and as a result has a much lower acoustic impedance than air (very high impedance medium).  As a result of this high impedance ratio the bulk of the sound energy from the scream is reflected back towards the bottom of the pool.  When you were submerged for the scream this medium barrier was not present, and therefore you heard it loud and clear, sort of.

The human ear is not built to hear sound transmitted in a liquid.  Notice that sounds under water sound much higher pitch than those above the surface.  This is again related to the density of the medium, which is a large factor in the speed of sound propagation.  In air sound travels much slower than it does in water.  Take a sound source with a fixed wavelength, your daughter’s voice, and put it in two mediums of different densities (water and air).  In air her voice sounds normal, but under water the sound travels faster and thus has a higher frequency.  

I came across this in an anthology which may provide an answer.

A very strange thing indeed - poem by B. Michael James

It was a very strange thing indeed.

I became aware of the silence.

The comforting murmur

Of the central heating boiler had ceased.

So I sat there

On the loo, in the quiet.

It was a very strange thing indeed.

It wasn't loneliness I felt

I was not facing another day on my own:

Soon, I would be seeing my friend.

A very strange thing indeed.

This silence.

I have no idea what it meant.

This not unpleasant quietness.

Could death be like that?

Taversing an unknown nothingness

To a new awakening.

It was a very strange thing indeed.

Sound travels as a result of vibrating particles. This is why in space,you cannot hear any sounds if you didn't have a spacesuit (not recommended). In water, the particles are quite spaced out whereas in air,the particles are very spaced out. When producing a sound underwater means that a water-air boundary would have to be passed however this means that most of the sound waves are actually reflected back at the boundary since of their acoustic impedance does not match, this explains your question.