It is an accessible presentation, but not dumbed down too far: the focus is very much on the science, with just a little history of science and no biographical or anecdotal material. Archer includes some equations, some chemistry, and a lot of detail about the carbon cycle and climate history. Anyone without a paleoclimatology background could learn quite a bit from it.
Part one, "The Present", looks back about a century and forward a similar distance.
Archer begins with a basic explanation of the greenhouse effect, along with an outline of how it was discovered and has steadily risen in prominence. He then looks at the observational evidence from the last hundred years or so, confirming the theory. And he presents the forecast for the next century, for temperatures, rainfall changes, sea level rises, and floods and storms.
As well as explaining the basics of greenhouse forcing and feedbacks, this includes a bit about how climate models work, some background on ice sheets, and so forth. These first fifty pages would be a fine read by themselves for anyone after a basic explanation of the science behind global warming.
Part two, "The Past", turns to paleoclimatology, to the reconstruction of past climates.
Archer begins by looking at climate changes on millennial scales: the Little Ice Age and the Medieval Optimum, evidence from pollen and tree rings, the 8.2k event ("a sudden cooling and a global tendency for drought that lasted for several centuries", 8200 years ago), the Younger Dryas event, and Dansgaard-Oeschger events (which occur about every 1500 years) and their grouping into Bond cycles and Heinrich events. Some of these involved quite rapid transitions, most plausibly driven by shifts in ice sheets and ocean currents.
Moving further back in time, Archer describes the historical discovery of ice ages and the modern use of oxygen isotope ratios in ice and CaCO₃ to track the evolution of ice sheets. There are cycles from orbital precession, obliquity and eccentricity around 20, 100 and 400 millennia. And bubbles of air trapped in ice reveal past changes in CO₂ concentrations.
On an even longer-term scale, the carbon cycle is driven by the geology of weathering and subduction. Archer doesn't dig too deeply into the full range of proxies we have for temperatures and CO₂ concentrations going back hundreds of millions of years, but touches on benthic foraminifera, CaCO₃ deposits from desert soils, and stomatal densities.
How does what we are doing to the climate compare with this history?
"Mankind is becoming a force in climate comparable to the orbital variations that drive the glacial cycles. ...
the global warming climate event is not unprecedented in Earth history. Climate changes through the glacial cycles were probably as severe as global warming has the potential to be. The Earth and the biosphere will survive.
Viewed in the same time perspective, however, human civilization is also totally unprecedented in Earth history. ... Civilized humanity has never seen a climate change as severe as global warming."
Perhaps the most interesting comparison is with the Paleocene Eocene thermal maximum (PETM), 55 million years ago.
"It is unclear how much carbon was released in the PETM, but typical guesses are similar to the amount of fossil fuel coal available for us to burn. The deep ocean warmed by 5–8°C, and probably the land surface warmed about that same amount, too. This is considerably more than the warming projected for the year 2100, for two reasons. One is that the climate of the Earth takes a few centuries to warm. If it warms 3°C by 2100, there could be another 2°C of warming 'in the pipeline,' based on the CO₂ concentration of the atmosphere at 2100. Also, all of the coal will probably not be burned by the year 2100."
Part three, "The Future", turns to the implications of the past for what lies ahead.
Two chapters explain the long-term carbon cycle and the key role the ocean plays in it. Archer goes into some detail here, explaining ocean carbonate cycling, atmospheric equilibriation with the ocean and land, ocean acidification, and the neutralisation of CaCO₃. The central result of this is that carbon dioxide residence time in the atmosphere is quite long.
"The sequence of events, then, is this. CO₂ is released to the atmosphere. On a timescale of centuries, most of it invades the ocean, leaving 15% or 30% in the atmosphere. The invading CO₂ acidifies the ocean, provoking an imbalance in the CaCO₃ cycle, which acts to neutralize the acid. The pH chemistry of the ocean recovers on a timescale of maybe 2-10 millennia. After this time, the models predict 10% or so of the fossil fuel CO₂ to remain in the atmosphere for hundreds of millennia into the future."
Following the PETM it took over 100,000 years for ocean temperatures and CO₂ levels to return to pre-event values. "The Earth has the ability to look after its own climate, but only if we are prepared to wait a few hundred thousand years."
These ocean processes tend to dampen the effects of fossil fuel CO₂. Unfortunately paleoclimatology suggests there are other feedbacks which involve increasing CO₂ release with increasing temperature: "Climate records of the past give reason to fear that the carbon cycle could eventually act as an amplifier of human-induced climate change ... rather than damping it down as it is doing at the moment."
At higher temperatures the ocean simply holds less CO₂. Less predictably, there is much carbon in land storage, mostly in permafrost at high latitudes, and there are methane hydrate deposits frozen in ocean sediments, which at thousands of gigatons of carbon equivalent dwarf even fossil fuel deposits. There is a lot of uncertainty here, but some calculations suggest these stores might release as much carbon as fossil fuel emissions, doubling their long-term impact.
Turning to sea level, Archer presents a key plot of sea level against temperature at different points in the geologic past. The IPCC Report suggests 2100 might see a 3° increase in temperature and about half a metre of sea level rise. But the paleoclimate comparisons suggest that such a temperature rise would match a fifty metre sea level rise — the difference between the transient and equilibrium responses here is huge. How long such a temperature rise would need to be sustained to produce such a rise depends on how fast the Earth's ice can melt, and understanding that requires an understanding of ice sheet dynamics... which were explicitly excluded from the scope of the IPCC Fourth Assessment Report.
A final chapter considers the long-term interplay of greenhouse warming with orbital changes. It may stop the "trigger northern hemisphere sunlight intensity" being reached, thus preventing the next ice age.
Archer leaves policy considerations to a brief epilogue. Uncertainties about ice sheets mean it is difficult to know what a "safe" CO₂ concentration might be. Stabilisation scenarios can be built around "wedges", technological or social changes capable of saving 1 gigatons of carbon a year by 2050. "Ultimately, the future of Earth's climate comes down to decisions about coal."
There are two ideas that Archer doesn't explicitly raise, but which stand out as implications. The long-term risks from anthropogenic global warming are significantly different to the short-term ones, and deserve policy consideration in their own right, at least if we are concerned about "tail" risks to the viability of human civilization. And if we had no climate simulations — or, the common skeptical claim, no confidence in them — then we would be more worried about global warming rather than less, since the direct paleoclimate comparisons are not at all reassuring.
The Long Thaw lacks full references, but there's no original research in it and the details can easily be found. There is a further reading section with a few suggestions for each chapter.