All of the graphs below are adapted, or in one case used unchanged, from an article at the Foresight Institute by J.Storrs Hall. The only value-add in my part is to highlight the areas on each graph which correspond to the succeeding one. Click to enlarge them.
That is the record of the Vostok Ice Sheet thickness. Continental drift over the period is relatively small, and that's the only thing we may have to correct for. Otherwise the raw data can be used to give an accurate measurement of the temperature at that point on the Earth's surface, with no corrections, normalisations, or other data massaging needed. It shows a repetitive pattern corresponding to periodic ice ages, a sharp temperature rise, slowly decreasing, in a staggered sawtooth pattern.
So let's look at the highlighted bit. And go to the Greenland Ice sheet thickness on the other side of the world, for a more detailed view of what's happened in the last 15,000 years or so.
This time, we see what happened after the last Ice Age ended. Bumpy, but with a slight downward trend. Again, this doesn't measure global temperatures, but at least gives us a clue as to polar temperatures at that particular spot on the Earth's surface, and not requiring a great deal of massaging to get adjacent data points to be comparable. OK, so let's look at the highlighted part. The bit from about 3000 BCE onwards - human-recorded history.
It looks like things are a bit cooler now than when the battle of Khadesh was fought, and when the Sumerian civilisation was at its height. Cooler by far than the Mycenean period 1500 years later, when Troy fell. At least, they were in Greenland. We must be careful not to make too many assumptions about the correlation between global temperature and polar temperature, or we step from firm ground with a very high degree of confidence indeed, to increasingly risky and uncertain territory.
So on to the highlighted area once more. From 800 AD onwards, about the time when Greenland was first explored by a people who recorded what they saw there.
Sure enough, those who called it "Greenland" weren't lying. There's the MWP, the Medieval Warming Period, blatantly existing at the northern polar region. Quite close to where most of our scripted records of the period were written. Also recorded in the ice are various cold periods that would have led to severe hardship in nearby Western Europe. The first, and probably the most shocking, around 1250. Another around 1425. The "Little Ice Age" around 1660, while a lowpoint in recorded history, seems to be just a relatively minor dip in the long cold time between 1600 and 1820. Again though, we can't automatically assume a high degree of correlation between Greenland and Global temperature, but a relatively high degree between Greenland and nearby Iceland, Scotland, and Newfoundland is more certain.
Now let's look at the highlighted bit again. The age of Political Climatology.
Is the rise at the end due to the Industrial Revolution? It correlates with it rather well, though seems to precede it a bit. What is clear though is that while Humans have had a huge impact on atmospheric CO2 content in the atmosphere, the actual temperature change in Greenland has been relatively minor. If it exists at all. Because variations in temperature far greater than that have been found throughout recorded history, at least in Greenland. In fact, from the looks of it, it may well be just noise.
And the variations in temperature in recorded history truly are noise in the regular staggered sawtooth pattern of periodic Ice Ages in the last 400,000 years. It's been far hotter than this in the relatively recent past - and a heck of a lot colder too.
What seems to be the case is that temperatures (in Greenland, remember, not necessarily elsewhere) are slowly dropping over the last 10,000 years; that we're in a relatively cold patch in the general variation over that period; but that's just part of the long, slow descent to the next Ice Age. Peak warmth in this interglacial has passed.
I have little faith - actually, none whatsoever - in the results from various climate models where the data is extremely sparse, heterogenous, and subject to truly huge uncertainties. Having looked at the Climate Research Unit's initial data set, I don't see how anyone could extract any meaningful conclusions from it with a reasonable degree of confidence. Far too many normalisations, corrections, and guesses to overcome problems of missing calibrated measurements have had to be made.
A thermometer at one station may be moved 200 metres or more in altitude between one year and the next, with no period of parallel running to calibrate the two so that a single set of measurements, like with like, can be formed. Then the thermometer may be discarded, and the "reading" at that point substituted by the results of interpolations from nearby stations - some of whom may in turn just be "readings" that are interpolations from other stations further away. Or even from the previous year's very uncertain results from the thermometer that was first moved, then deleted!
Outlying figures that are obviously wrong, due to local environmental changes, malfunctioning instruments and so on have to be weeded out in any such exercise.
One problem though with the CRU work. There is a consistent pattern of the "obviously wrong" figures that disagree with the model's broad results being from the most reliable data sources. Readings every year from actual thermometers that have not been relocated, and where the local environment has remained undisturbed.