Is it possible to measure temperature using sound? Is it possible to measure the temperature of something using sound, by blasting sound waves at solid objects, liquids and gasses, then measuring how much sound is reflected or absorbed, to give an accurate atmospheric measurement of temperature?
 A: 
Is it possible to measure the temperature of something using sound...?

Yes, it is not only possible, it is available commercially.  It is especially useful in harsh environments where conventional temperature probes might not survive.
For example, TMT makes an acoustic system for measuring 2-D temperature distributions in blast furnaces:

The acoustic gas temperature measuring system TMT SOMA™ (SOMA stands
  for Sonic Mapping) as described hereafter is able to provide a
  complete and continuous 2D temperature distribution of the top gas
  inside the blast furnace above the burden. The system is based on the
  principle of acoustic gas temperature measurement sometimes also
  referred to as "sonic pyrometry"…

TMT claims accuracy of  ± 2.5%.
Another company making acoustic temperature measurement systems is SEI.  Here is a schematic of their approach showing the ability to measure 2-D temperature profiles:

SEI claims an accuracy of better than 0.5%.
If you want to design your own system, Linear Technology Corporation makes integrated circuits to support your efforts.   Their Application Note 131 describes acoustic temperature measurement as follows:

Acoustic thermometry is an arcane, elegant temperature measurement
  technique. It utilizes sound’s temperature dependent transit time in a
  medium to measure temperature. The medium may be a solid, liquid or
  gas. Acoustic thermometers function in environments that conventional
  sensors cannot tolerate.

Linear Technology quotes accuracies of 1°F or better.
A: In principle, you could detect changes on the speed of sound when passing between different zones of a fluid which have different thermodynamic conditions (i.e. different temperature and pressure as well, or density). Indeed sound propagates differently in the same fluid when different temperature, pressure or density, or any combination of them is present. 
But when passing through interfaces, i.e. gas to liquid or gas/liquid to solid, I would say the reflection of sound is much more intense and would be harder to understand.
Now the question if all this is really feasible, I guess you would get it if you try, maybe there is some way but it sure would require some study. There is normally a big difference between the theoretical possibility, and the actual possibility of understanding, classifying and quantifying some magnitudes from how other influence them.
