As I pointed out at the beginning of this article, the last two millennia are the most pertinent to today's problem of measuring the human influence on global warming. The really long-term climate cycles, based on Earth's orbit and plate tectonics, show negligible change over this stretch of time. That leaves the other natural and human factors mentioned previously. Let's look at the two records, actual climate and climate forcings.
The Climate Record
One basic statistic, the global surface temperature, has recently been reconstructed by different groups of researchers for the last 2000 years. (One recent effort appeared in 2008 in the National Academy of Science Proceedings.) Note that this is for the world as a whole; it averages out night and day, all of the seasons and all fluctuations on the regional and continental scale. Taken as an undifferentiated whole, the world's surface temperature has fluctuated rather little since the zero year (which of course was actually 1 BCE). Individual years vary within a range less than 1.5 degrees C; decades vary by less than 1.0 degree.
Until the 20th century, global average surface temperature appears to have dropped very slightly since year zero. No century has differed from the other centuries by more than 0.2 degree, with one exception: the 20th century. And the 20th century has shown a rise of between 0.6 and 0.9 degree. The 20th century jumped the bounds and is by far the warmest century of the last two millennia, probably much longer.
Other factors such as precipitation and the various cycles such as ENSO cannot yet be considered at this global scale over these 2000 years. We have only a few fragments of the record in a few regions. One thing we can tell, though, is that the world's climate did not behave very differently before today.
The Forcings Record
We have global records that go back 2000 years of the four most important forcings: solar output, volcanism, levels of greenhouse gases and sulfate aerosols. They too show limited variation over the centuries, although volcanism appears to have been relatively quiet between the years 1000 and 1500 compared to today. Greenhouse gas forcing began its well-known climb starting around 1800, and sulfate aerosol rose starting around 1900.
There is no getting around it: the 20th century shows an outsized, unprecedented climb in global average temperature at the same time that greenhouse gases have exceeded any level known in the past half-million years (as far back as we can see as of today). Sulfate aerosol counters the effect of the rise in greenhouse gases by about one-third, while solar variation is a far smaller effect. Volcanism is always a wild card, but even the greatest eruptions do not affect the global average surface temperature for more than a few years at a time.
Climate Modeling
All of this data, past and present, can be fed into computer models of the global climate called GCMs, not for "global climate model" but for "general circulation model." GCMs simulate the circulation of the atmosphere and the oceans, plus the interactions or "couplings" between them. News stories always call these models "sophisticated," and indeed they are, but the leading models pixelate the world into large cells, 5 degrees of latitude and longitude in size. Individual cells in a GCM correspond poorly with actual climate records in their boundaries—this is no surprise when you look at a map and consider how complex the world is inside a 5-degree box.
GCMs are useful for analyzing the climate system. Researchers tweak them, and the data they feed on, in many different ways to weigh the influence of forcing factors, and to estimate how the internal climatic variations would respond to different forcings. Many of the parameters, the numbers used to represent various factors in the complex real world, are poorly constrained by the data and are being studied in elaborate experiments.
Where We Are Now
Imperfect as they are, GCMs are our best tool for helping us assess the reliability of the climate record. And they are our only tool for modeling the possible climates of the future. The picture they draw is a hazy one, but it points firmly toward further warming for several centuries to come. And it points squarely at human input of greenhouse gases as the basic cause.
This knowledge does not automatically translate into a set of policy choices. Climate scientists can help in assessing the possible choices, but they cannot determine those choices. Only political systems can do that. There is a wide range of legitimate debate, but questioning the science is no longer legitimate.
I have gained a great deal of clarity on the science of this topic from the rigorous and thorough review by Jones and Mann in the June 2004 Reviews of Geophysics. I was once a typical skeptic, but am not one now.
