Let’s assume that you had doubts about global warming. Some people do and, truthfully, we utterly lack the ability to accurately model how much warming will happen, where and by when. Emphasis on accurately. The reason is not for lack of trying or continual learning and model refinement, but centers on the complexity of the task.
Even seemingly simple systems that are actually complex are impossible to predictively model. An example is a pile of sand growing grain by grain that will finally slump at some unpredictable time and in an unpredictable way. You would think that such a simple system could be accurately modeled, but that’s not the case. Exactly when the pile will finally slump is unpredictable. Exactly how large the resulting slump will be is also unpredictable. The “when” and the “how much” are unknowable (using current modeling techniques).
All that can be calculated for certain is that a higher pile with steeper sides/areas (a.k.a. “fingers of instability”) is more likely to slump sooner and more catastrophically.