High CO2 worlds
There is evidence that at certain times in the geologic past, the natural climate state was warmer than modern and that atmospheric CO2 was correspondingly higher. The causes of the onset of such warm times are not perfectly understood, but proxy evidence for temperature, atmospheric CO2, sea level, ocean acidity, etc., all provide important insights into how the Earth system responds to significant warming. The three high CO2 periods highlighted by the IPCC are the Paleocene-Eocene Thermal Maximum (PETM; 55.5–55.3 million years ago), the Early Eocene Climatic Optimum (EECO; ~52–50 million years ago), and the Mid-Pliocene Warm Period (MPWP; ~3.3–3.0 million years ago); key data from these episodes are summarized in the table to the right. Note that climate reconstructions for these periods of the distant past are very tentative—evidence is indirect, has large error bars, and is very limited in its spatial distribution.
Glacial-Interglacial Cycles
Starting around 2.6 million years ago, the Earth has been in an ice age marked by the establishment of ice sheets (on Antarctica, Greenland, and elsewhere at times) and cyclical fluctuations between cooler glacial periods and warmer interglacial periods. The Holocene is one such interglacial period. Glacial periods occurred regularly, at first every ~40,000, then every ~100,000 years (Figure 5.3 panels e–f), and were driven primarily by orbital variations (Figure 5.3 panels a–c). However, it is well established that atmospheric CO2 has played an important role in these temperature fluctuations (Figure 5.3 panel d), as it acted as a strong positive feedback mechanism.
The Last Glacial Maximum (LGM; ~21,000 to 19,000 years ago)—the maximum extent of ice growth during the Last Glaciation—was marked by cold temperatures and low atmospheric CO2 of around 200 ppm. Temperature reconstructions show very different cooling patterns from region to region, but global estimates of surface air temperature range from 3.1ºC to 8.3ºC below pre-industrial values. Ice cores from central Greenland, on the other hand, indicate 21ºC to 25ºC cooling; this high-latitude amplification of global temperature trends is typical of all climate events, including modern warming (known as polar amplification), and is often underestimated by paleoclimate models. The LGM is a useful case study for estimating climate sensitivity—the temperature response to a doubling of atmospheric CO2. Various approaches yield climate sensitivity values in the range of 1ºC to 5ºC per doubling CO2.
The LGM was followed by relatively rapid warming of 3 to 8 ºC during the Last Glacial Termination (19,000 to 11,650 years ago). The most rapid warming during that time reached likely reached 1 to 1.5ºC per thousand years. Reconstructions suggest that warming in the Southern Hemisphere preceded warming in the Northern Hemisphere and was tightly coupled to a rise in atmospheric CO2.
Climate change of the last 2,000 years
Temperature changes of lower magnitude have occurred during the last two millennia, and proxy reconstructions of these changes offer the opportunity to compare the modern instrumental record with natural variability over a period with very similar tectonic, orbital, and other parameters that may otherwise affect climate. The two climate events that the IPCC focuses on are the relatively warm Medieval Climate Anomaly (MCA; 950–1250 C.E.) and the relatively cool Little Ice Age (LIA; 1450–1850). Figure 5.7 shows various temperature reconstructions of the last 2000 years, including both the MCA and LIA excursions.
Although there is significant spread between reconstructions, the IPCC asserts that “there is medium confidence that the last 30 years were likely the warmest 30-year period of the last 1400 years” for the Northern Hemisphere. There is much more limited proxy evidence for paleoclimate in the Southern Hemisphere, but records there suggest that 20th century warmth exceeds any 30- or 50-year period in the last four centuries.
The same models that are used to predict future climate change can also be applied to the past to reconstruct paleoclimate using known variations in the Earth’s orbit and axis, changes in atmospheric composition, shifts in tectonic configuration and vegetation cover, etc. Then, these paleoclimate model outputs can be compared to (or adjusted based on) paleoclimate data. One major limitation in paleoclimate modeling is that the highly complex models used to project future climate (which include fully interactive atmosphere and ocean) are extremely computationally expensive to run beyond the century-scale. Relevant timescales for paleoclimate are typically thousands to millions of years, and running fully coupled models for these timescales is infeasible. Still, scaled-back/idealized models or short snapshots of past climate are possible, and provide useful insight into the functioning of the climate system in different states. Figure 5.8 illustrates a comparison of paleoclimate model output and proxy data reconstructions for the last 2,000 years. For this global reconstruction, there is good agreement—but some models miss certain spatial patterns, including polar amplification.
Figure 5.3: Orbital and climate parameters for the glacial-interglacial cycles of the last 800,000 years. (a)-(c): orbital cycles related to the shape of the Earth's orbit and angle of its axis; (d) atmospheric CO2; (e) tropical sea-surface temperature anomalies; (f) Antarctic temperature anomalies; (g) oxygen isotope records from bottom-dwelling marine organisms; (h) sea level relative to present.
Figure 5.6: A comparison of proxy-generated temperature data (a) and modeled temperature (b) from the Last Interglacial period. Note that warming at high latitudes is greater in the data than the model output .
Figure 5.7: Various temperature reconstructions for the last 2000 years for the Northern Hemisphere (a), Southern Hemisphere (b), and globe (c). Thick black line is the instrumental record for the most recent part of this period. Temperatures are reported as anomalies relative to the 1881-1980 mean.
Figure 5.8a: A comparison of temperature reconstructions from proxy data (grey) and modeled temperatures (red/blue) for the last millennium. Temperatures are reported as anomalies relative to their 1500-1850 mean. MCA = Medieval Climate Anomaly; LIA = Little Ice Age; 20C = 20th Century.
Earth System Responses and Feedbacks
The Last Interglacial (LIG; 129,000 to 116,000 years ago) appears to have been ~1ºC to 2ºC warmer than pre-industrial values, and at least 2ºC warmer at high latitudes. Figure 5.6 shows temperature estimates from both proxy data (panel a) and model output (panel b) and highlights the regional differences in temperature change at this time. The LIG was also marked by high atmospheric CO2 and high global sea level (since ice sheets were at a relative minimum, more of the Earth’s water was contained in the oceans).
Past extremes in climate provide important insights into how the Earth responds to various climate forcings. A key feature of the climate system is the presence of feedback mechanisms--a change in one part of the system causes a second change that either amplifies or reduces the original change (see glossary for examples). The magnitude of climate change ultimately depends on the strength and sign (whether positive or negative) of internal feedbacks in the system. The following section summarizes the climatic response to forcings for several important paleoclimate events in Earth's history.