Steve Hsu, who I trust*, interviewed Tim Palmer of Oxford and the Royal Society, who I know nothing about, about how reliable climate models are. Palmer seems very reasonable. His short answer is that WRT the basics they are very good, as it is just 19th C physics, but the deeper structures still have too much variance. However, as the models improve they are slowly pointing toward worse outcomes rather than better.
They started by discussing how to get top physicists into fields of practical application, as most want to study the (sexier) theoretical issues and take their PhD's and post docs there. Palmer notices a trend in that group in their late twenties of deciding later to want to move to physics of practical import, but they have difficulty transitioning to what are essentially new fields.
He asserted quite blandly the high level of certainty that some recent warming (about 1 degree Celsius) has happened and high certainty it is caused largely by greenhouse gases and human activity. Because there is a known heightened warming effect from increased water vapor, that is an intensifier and it isn't going away. Such things are frequently asserted by journalists and politicians, and even from scientists who have a high drive to activism, but my experience has been that the other scientists are more cautious and measured. Yet as Palmer went on to say that it is the subsequent questions that are more complicated and difficult, and this is sometimes obscured, it seemed to me that this was a clearer distinction than what I have been making.
One particular difficulty is that "cloud cover" is a term thrown around with imprecision, and two aspects of it in particular have very different effects. Low lying clouds reflect sunlight back into the atmosphere and off the planet, so an increase in water vapor at that level is mitigating effect on warming. But increased cloud cover higher in the atmosphere is an intensifier of warming. Modeling of this has been difficult and shows high variance. Nearly all models show a worrisome increase in upper level cloud cover. Palmer thinks the question is so large and so important that we should be spending much more money on studying it.
There is uncertainty whether the moisture at high levels is ice or supercooled water droplets. The later would be worse in terms of intensifying warming. Recent observations and calculations are indicating that there is more supercooled water and less ice, and models are being adjusted accordingly. The projections from this change point to strong possibilities of much higher warming and greater variance in extreme weather.
Interestingly, Palmer focuses on the secondary effects of warming. It may be that there will be more hurricanes and droughts and a greater rise in sea levels, but most of these are manageable (though very expensive) in developed areas. People shrug that things being a little warmer in Scotland might not be a bad thing, but Palmer notes that decreased sunlight will mean more depression and alcoholism, and worse cardiovascular outcomes. However, what really worries him are the places that are becoming unliveable, such as parts of Iraq and into Central Asia, and especially much of central Africa. This latter is about the only place where population is still growing, and when people find a place unlivable, they historically just move. In large numbers. And they are impossible to stop. He notes this is already happening and doesn't believe Europe has quite come to grips with this.
However, a different point of view comes from Hsu discussing Steve Koonin's book Unsttled: What Climate Science Tells Us, What It Doesn't, and Why It Matters, and the WSJ article about him. Koonin was, of all things, chief scientist for Obama's Energy Department. Perhaps it is not that different a POV after all, as he believes that the climate science we are doing is pretty good, but the reporting and posturing about it are dangerous. I appreciated the use of the word "overegged," which one hardly sees anymore. Koonin also believes that because even a slight chance of catastrophe is too much, we need to assess this risk mfar better than we have been.
*I very much trust in science. Professor of Theoretical Physics and Computational Mathematics at Michigan State. He was VP of for research there but stepped down because of political controversies around human genetics. Basically, he asserted that the data says what it says and would not back down. He was the one who got me into a study where they did a full genome on me at a time when those things cost about 10K. It has never been useful to me, however, as I don't have the tools to read it. It just sits on my computer now. I don't know his other politics.