Messy? I'd say it's probably impossible, just like trying to determine an "average" temperature of the earth. How do you define an equilibrium in a chaotic system with limitless variables?
You have a good point. However, if we can measure solar + cosmic influx, earth’s outflux and estimate the relatively small amount of heat being generated on the surface and escaping from the core to the surface, we should have a pretty good idea of the rate of change of averages of earth’s surface or atmospheric temperature. It doesn’t tell you the weather in Boise next week, but it does tell you whether or not it’s getting warmer.