That's a good question, abailart. It guess it comes from the assumption that we are aware of every feedback loop, every cause and effect, every climate driver, every influence on every climate driver, every process and rate of the absorption of CO2 in the carbon cycle, every process and rate of absorption in the water cycle, the understanding and dynamics of ocean currents, all historic environmental conditions that have ever existed, the complete understanding of the Sun's activities and it's influence on climate, clouds, ocean currents, etc. and the complete understanding of the contribution of all geologic factors involoved in the climate.

The fact is that models clearly show that it is ...

That's a good question, abailart. It guess it comes from the assumption that we are aware of every feedback loop, every cause and effect, every climate driver, every influence on every climate driver, every process and rate of the absorption of CO2 in the carbon cycle, every process and rate of absorption in the water cycle, the understanding and dynamics of ocean currents, all historic environmental conditions that have ever existed, the complete understanding of the Sun's activities and it's influence on climate, clouds, ocean currents, etc. and the complete understanding of the contribution of all geologic factors involoved in the climate.

The fact is that models clearly show that it is quite probable that the number is much higher than 90%. Maybe 99%.

If the smartest
climatologist in the world knew 100 variables
(and there are perhaps thousands), each one to an accuracy
of 99 percent, the probability of a correct predictable model on
climate change would be 36.6 percent. Remember, that is IF a) the
scientist knew ALL the variables, b) knew all the feedback loops, c)
knew all the variables of the feedback loops to an accuracy of 99 %.

Also, statistically, you can make all kinds of assumptions about what percentage of something you may not know. As this scientist can attest.