[Title]
Toward Answering Some Leading Climate Science Questions by Constraining Models with Observations


[Abstract] 
Much of climate research is built around climate models for understandable reasons. Yet, as a model developer and user, I see a common fallacy of putting blind trust in models, or in other words, taking model results simply at face value. Climate models are bound to be uncertain in many aspects for good, physical reasons. Furthermore, when models give different answers, the most common response is that we need more models, and let us average them up. Although there is indeed value in ensemble means, I personally do not think that they are the way out especially when many models are closely related. To me, the key to answering many leading climate questions lies in using observations to constrain models. I will use a few of my recently published works to substantiate the claim.