Just as analytic solutions to Newtonian physics work well in most cases while ignoring the effects of relativity, many computer models work well without including the effects of every possible variable. It's simple if one variable accounts for 90 % of the observed data and another accounts for 6 % etc etc, leaving out a variable with a 0.1% potential effect is not likely to have a major impact.

Simulators must focus on prime variables in order to make the research and computer run times manageable. Just as my checking account analogy demonstrated that net effects can be more important than short term variability, I don't think there is any reason to believe daily cloud cover is an important variable unless the average amount of cloud cover changes over time (quite possible but that didn't seem to be what you were referring to).

In regard to comments regarding volcanos and impacts, we have good reason to believe already these are short term effects based on our experience. Sure climate model predictions will be wrong when these events happen, but we are talking about long term trend predictions not the climate in a specific year. A major impact may have longer effects but these happen at far less frequency than the time the effects of global warming are estimated to require.

Nuclear power heat emmissions are certainly negligable compared to the heat input from the sun.

Regards,
Mark