You gotta remember, I'm not advocating worst-case IPCC scenarios. I think we're going to see a ~2.5 C rise in global temperature in the next CENTURY. Which (if linear, which it won't be) would be 0.25 C in 10 years. That's 0.4 F in 10 years. So if you'd be willing to bet on a 0.3 F (0.18 C) degree rise in global temperature as determined by the National Climatic Data Center over the next decade, I'd take it (for a 12-pack of Blue Ridge Amber Lager, not for $10K). And I win if for any year the yearly global average temperature is 0.18 C higher than the yearly global average temperature for 2002. This year could have a moderate El Nino in it, so that's not a bad starting point.
I don't trust anybody's sea-level rise data. The difference between U.S. analysis of TOPEX/Poseidon altimetry data and European analysis of the same data is different by like 5 mm. So here's an alternative: the timing of spring thaw on northern lakes and rivers will be earlier, on average, by more than 1 day over the period 2002-2012. We will consult Dr. John Magnuson of the University of Wisconsin or whoever inherits his data set should he become unavailable.
I don't think a single year would work very well as the yearly values vary too greatly and it would be very likely one year would be higher. We could set the starting point at 54.2F for 2001 and if the data for 2002-2012 produce a best fit line with a tempreture greater than 54.5F in 2012, you win.