I think the basic problem with climate science is in the assumption that climate is stable and you've pretty accurately described it here. To understand what the climate is, all we have to do is to wait long enough that the attractor has explored all of its phase space (in a statistical sense).
What's recently becoming clear is the importance of the oceans. It started with the El Nino oscillation which lasts a few years. Now they're exploring the consequences of longer time period oscillations like the "Pacific Decadal Oscillation", "Atlantic Multi-decadal Oscillation" and "North Atlantic Oscillation" which have periods of about 50 years. Here's a recent Nature article:
Ocean impact on decadal Atlantic climate variability revealed by sea-level observations
McCarthy, Haigh, Mirschi, Grist & Smeed
Nature 521, 508-510 (28 May 2015)
...
In the past 90 years, the AMO has undergone three major transitions: warming in the mid-1990s and 1920s, and a cooling in the 1960s.
...
http://www.nature.com/nature/journal/v521/n7553/full/nature14491.html
The consequence of these oscillations is that some of the temperature rise seen in the 1990s is natural variation. The good news is that as each year goes by, we understand these better and better. Note that the above article has only been out for a few weeks. The bad news is that the alarmists are digging in their heels and ignoring the new science. They want to attribute *all* of the warming of the 1990s to CO2. The result is that their estimates for future warming are deeply biased to the high side. They are beginning to take these effects into account and the warming bias is slowly decreasing but their continued funding requires that they alarm the public so they have no motivation to share the good news.
These long period oscillations imply that observations cannot distinguish the effect of CO2 and the effect of these oscillations until they are integrated out. What's going on is that the attractor has some very long memory built into it so it takes a very long time to fully explore its phase space. But even when you integrate over periods of a thousand years you still find massive cycles going on in the climate. For example:
Reconstruction of Prehistoric Landfall Frequencies of Catastrophic Hurricanes in Northwestern Florida from Lake Sediment Records
Liu & Fearn
Quaternary Research, Volume 54, Issue 2, September 2000, pp 238-45
Sediment cores from Western Lake provide a 7000-yr record of coastal environmental changes and catastrophic hurricane landfalls along the Gulf Coast of the Florida Panhandle. Using Hurricane Opal as a modern analog, we infer that overwash sand layers occurring near the center of the lake were caused by catastrophic hurricanes of category 4 or 5 intensity.
...
The most likely explanation of the abrupt stratigraphic change above 1.60 m is that there was a remarkable increase in hurricane frequency and intensity affecting the Florida Panhandle and the Gulf Coast after 3400 14 C yr B.P. as a result of a continental-scale shift in circulation patterns. Based on the chronology of eolian activity and sand dune deposition in the central United States, Forman et al. (1995) suggested that during the mid-Holocene thermal maximum (ca. 6000 14 C yr B.P.), the jet stream and the Bermuda High were situated to the north and northeast of their present positions, respectively. As a result of the anticyclonic flow around the southern and western flanks of the Bermuda High, moist air from the south Atlantic was pumped northward along the Atlantic coast of North America. However, by 3000 14 C yr B.P., Neoglacial cooling had caused the jet stream to shift south and the Bermuda High southwest from their ca. 6000 14 C yr B.P. positions, thereby pumping more moisture from the Gulf of Mexico and the Caribbean into the central plains of the United States (Forman et al., 1995). The more southwesterly position of the Bermuda High after 3000 14 C yr B.P. would also result in more hurricanes making landfall on the Gulf of Mexico coast instead of the Atlantic coast (Fig. 7).
...
http://www.sciencedirect.com/science/article/pii/S0033589400921665
Of course the above reads directly on claims of "climate change" or "climate variability". We don't know what the climate has changed in the past. As futurecurrents says, we don't even know what the temperature has been in the most recent 15 years. So of course we don't have a hope in hell of distinguishing the effect of CO2 from the natural variation in the climate.
And when you integrate out the climate effects that last a thousand years or so you quickly run into longer cycles, that is, the ice ages themselves. This is the inherent difficulty in estimating the effect of CO2 in the absence of a working model of the ocean / atmospheric system.
I expect that a working model of the ocean / atmospheric system will come together over the next 20 years and we will then know how much the CO2 is effecting the system. We will know when that model arrives because the model will be *extremely* useful to society in areas other than theoretical climate science. For example, farmers will plan their crops out a few years in advance. Right now, the alarmists are famous for making predictions that are random.
The best argument the alarmists had was that recent changes in temperature are much faster than changes seen in the historic record but the geologists do not agree with this. Instead, the past has examples of much faster temperature changes than the mild rise seen in the last 50 or 150 years. In fact, most of the bad science done by the alarmists has been their attempts to deny temperature change in the past.
It's a fact of life that when an idiot analyzes ancient temperature proxies they will use bad statistical analysis. And since bad statistics tends to destroy signals, the natural result of incompetent statistics is to deny ancient temperature changes. Bad statistics also creates false signals, but when you average a number of such false signals, again you will get no signal, not even a false one. Since those same incompetent statistics are never applied to the modern period where temperatures are measured by accurate thermometers, the natural result is "hockey-stick" graphs of temperatures. The handle has little temperature signal because incompetent statistics has eliminated any signal, while the blade is provided by modern temperatures which are not washed out. Hence the alarmists deny the existence of previous warm periods such as the Medieval Warm Period when England exported wine to France.
From a social point of view, when an incompetent scientist comes up with a hockey stick graph, he's treated as a hero by the people who hate industry and want to believe that CO2 is dooming us. They promote his work without looking at it carefully. In psychology, this is known as the "confirmation bias". Then the skeptic community (with the opposite confirmation bias) tears him apart. The alarmists end up divided in either supporting the old crap or moving on to something new. Either way it's not very convincing to neutral observers. This was all exposed in their climategate emails. In private they admitted that Mann's hockey stick was bogus but they never admitted this to the public because; to do so was inconvenient for their continued funding (i.e. it might result in unnecessary disagreements with people who would rule on their obtaining grants in the future, or convince the government that there isn't a problem and the money would be better spent on human health or real pollution problems).
"
.png)
