Andrew, my intuition here was simply that the CMBR temperature fits with blackbody radiation assuming an expanding universe resulting in adiabatic cooling. I guess you are saying that the cooling is greater than this mechanism could predict (is that right?) so maybe this could be a result of the newly discovered accelerated expansion (??).

The rate of cooling exactly agrees with the prediction you'd make from adiabatic expansion. The difference is that adiabatic cooling refers to the situation where you have a load of particles in a sealed box with a movable piston and allow that piston to move slowly. In such a case, the loss of energy of the particles is compensated by the kinetic energy delivered to the piston. In other words, the energy is transferred outside the system of particles, but is conserved overall.

In the case of the Universe, there is no external system to transfer the energy to, so you come back to the question: where did the energy go? (With the same answer as ever: nowhere.)

If so then you can't discount the effects of dark energy, though not affecting the CMBR directly, but being possibly the ultimate cause of greater cooling than expected.

Dark energy does indeed affect the cooling rate indirectly, since it accelerates the expansion.

(Incidentally there is also a second, smaller, more subtle effect of dark energy on the CMBR. This is called the integrated sachs-wolfe (ISW) effect, as alluded to by tommya in the post above. This ISW effect wouldn't exist in an exactly uniformly spread-out (homogeneous) universe, whereas the other effects do.)

Indeed, GR would have conservation of energy-momentum but energy can also said to be conserved. I think, rather as you observe in your post, this is a matter of interpretation. In flat (Minkowski space) energy is conserved. It could be argued that the FLRW metric this is not a bad approximation though has a few issues with boundary conditions.

In a global sense the FLRW metric is very different from Minkowski space, and it is exactly this difference that means energy conservation goes out the window. (Specifically, it is the lack of time symmetry of the FLRW metric.)

Even then it could be said the energy is transformed into gravitational rather than not conserved. If you think of light emerging from a gravity well, the energy put in is in a different frame from when it is received. Observers in both frames will only agree on the total energy used if the gravitational effects are taken into account, including time dilation, and I don't think there will energy lost; it just depends on the frame. At least I think so.

You're right about this particular case, but it is just one case.

If you attempt to follow this into a formal definition of 'what is conserved in every situation', you'll end up with exactly the energy-momentum tensor. In many simple situations, that could be translated back into the language of gravitational potential energies -- but not always.

To follow a discussion in another thread, do you know the formula for gravitational redshift from inside a massive sphere. i.e. that uses Gravitational potential as a function of radius. All the references I can find consider redshift outwards from the outside of a sphere and use field.

In the case of weak-field gravity where matter is moving at non-relativistic speeds (like on, inside or around the Earth or Sun, for instance), you can always use the Newtonian potential without worry. In that case you'll find the gravitational redshift is always equal to ΔΦ/c^2 where ΔΦ is the difference in Newtonian potential between emitter and observer and c is the speed of light. Rather neatly, this result arises directly from the equivalence principle without even needing to go via GR itself!

The cosmological redshift effect can indeed be thought of consistently in this kind of way, as 'the photon continually climbing out of a gravitational field'. But in that case, the Newtonian expression doesn't work directly (the weak-field limit doesn't apply globally).

Monthly astronomy podcasts - newbielink:http://www.thenakedscientists.com/astronomy [nonactive]