Bayesians and Climate Science

As a committed Bayesian, I have to take new data, and update old information.  How should I do that.

A while back, I posted that there are on the order of 7 questions to ask about Global Warming:

1. Is the globe warming?
2. Is human activity causing it?
3. Is the net effect expected to be bad or good?
4. Can anything meaningful be done (technically) to slow/reverse this?
5. Can anything meaningful be done (politically) to make the changes happen?
6. Is the net cost/benefit to society more positive if we solve global warming, allow global warming, some intermediate, or some alternative?
7. What is the opportunity cost of the results of F.

My priors were: 

1. Probably warming p=70%
2. Perhaps contributing, not driving p=60%
3. Probably good net effects from warming p =75%
4. Probably can do stuff to fix...p=80%
1. probably the current emissions reductions stuff won't fix...p=80%
2. probably nothing short of new technology will fix...p=70%
3. probably cutting carbon emissions to 0 worldwide would not fix p = 60% (see 2)
   
5. probably not able to do anything anyhow (b/c India + China)  p=80%
6. If GW is bad, you should delay solving it because of tech growth and exponential $ growth: p=90%
7. If you try to solve it now, you are spending huge $ that are better spent many other places p=90%

ClimateGate revises the first numbers substantially:

1. Probably we're not warming p = 60%
2. Not certain if human contribution matters at all.  p=50%
   

 Reasoning is:

A.  Data + models shown to be a total mess.
B.  Political activity clearly on side of AGW folks.
C.  There is attempt to suppress dissent/cook books:  mostly this only matters if your case is weak.
D.  With A + B, there should be pretty strong evidence for AGW. 
E.  There isn't pretty strong evidence.
F.  As per Bryan Caplan, the absence of evidence is evidence of absence...if someone's been looking enough.