The satellite observation graph does not even give a reason to think there is a linear upward trend at all. However, despite valid points about the miscalibrat

The satellite observation graph does not even give a reason to think there is a linear upward trend at all. However, despite valid points about the miscalibration of the models for the surface anomaly graph (those models overpredict effect on anomaly), the observed data still do _obviously_ support a strong positive linear effect in raising temperature over the past 20 years. It may be that climate models are involving nonlinear effects that experience difficult to explain lags in the real atmosphere. But either way I’d say the dramatic _trend_ in the observations is the most salient thing, and the 95% envelope is really not too important for the model’s inference goals. I’d like to see more Bayesian models applied to this.