Well, if one was to bet on a subject like this, one would have be quite specific about the terms, and the determination of whether or not a predicted outcome had been realized. Now, you are proposing to wager that none of the "worst-worst-worst case computer simulations" will turn out to be close to correct. I think that's a pretty safe bet! Thus, I (or anyone) would be pretty dumb to take you up on it.
Now, here's a clearer statement of my prediction. If you can figure out how to quantify it, then I'd put a 12-pack of Blue Ridge Amber Lager on the line. To make it fun.
"A variety of indicators exhibit trends as of year 2002 that if continued would support the observtion of a basic decline in global environmental quality. By the year 2012, these trends will be clearly more indicative of declining environmental quality than in 2002. While a few of these trends may be reversible via concerted action, the majority of the trends will not reverse by 2012."
I'd suggest proposing 15 indicators that have a defined trend over 1992-2002. I'd predict that by 2012 a majority (8) of these indicators will have a more pronounced trend in the same direction as current.
So, if one wanted to use the Hudson Bay polar bears as an example, we'd need to have data showing a decreasing trend in average weight or average fat rating (the article linked describes that assessment). I'd make two predictions that could be assessed on this indicator: a) that the population of Hudson Bay polar bears will decline 2002-2012; and b) that the remaining population will have a lower average weight and fat content in 2012 than in 2002.
I would propose 2, if either one happens you win. The IPCC predicts about a 4-10 degree F rise in temps over the next 100 years. I would say lets place the limit at 0.5 degree F rise over the next 10 years as measured by US thermometer data. The IPCC is predicting a 5mm rise per year in sea levels. That's about 2.5 inches over 10 years, not sure who I trust on this one, but I am sure we can agree on someone.