This often trips up engineers, who associate computer modeling with the concept of a known system. But climate models are scientific, not engineering. They exist to explore uncertainty.
The theory of anthropogenic climate change does not depend on global climate models. If they did not exist at all, we would have the same concerns about climate change.
The models are based on the theory, not the other way around.
How do we know that the plants won't absorb all of the excess CO2 without studying the local effects? The world is full of positive and negative feedback loops, discussing linear effects on the earth without taking into account these loops won't get us very far.
A lot of the biggest questions about how climate change will play out can't be well-modeled yet--like cloud formation, methane deposits in deep ocean or permafrost, etc. We need to learn more, and as we do, we feed that into the models and they improve.
> How do we know that the plants won't absorb all of the excess CO2 without studying the local effects?
We know that plants aren't absorbing all the excess CO2 because we directly measure CO2 concentrations in the atmosphere, and they are going up.
Are we magically absorbing more heat than 200 years ago? Getting closer to the Sun? Axial tilt shift? Gremlins?
Wake up.