The timing of interglacials

By Andy May

P. C. Tzedakis and co-authors have just published a new paper in the February 23, 2017 issue of Nature entitled “A simple rule to determine which insolation cycles lead to interglacials.” The paper introduces new rules for defining interglacial periods in the geological record. They come up with the same interglacial periods that Javier identified in his post Nature Unbound I: The Glacial Cycle.

The Earth has been in an ice age for the last 2.6 million years, Javier defined an ice age as:

“… any period when there are extensive ice sheets over vast land regions, as we see now.”

Tzedakis, et al. note that

“The fundamental property that underlies the concept of an interglacial is high sea-level.”

The higher sea-level is a result of melting a significant amount of land-ice during the interglacial. We are currently in the “Quaternary Ice Age,” which is either the coldest or the second coldest period in the last 500 million years as can be seen in figures 1 and 2. These are the most popular temperature reconstructions of the past 540 million years. Ice ages (or a collection of closely spaced continental glacial periods) have occurred in the geological record roughly every 150 million years in the Phanerozoic. The cause of these cold periods is not known, but we are clearly in one now.

Figure 1, source Veizer, et al., 1999 and Wikipedia

Figure 2, Phanerozoic temperatures, source Geocraft

The current (Quaternary) ice age is punctuated by warm periods, called interglacials. These warm periods are identified in the geological record by rising sea level. They persist for about 15,000 years on average and are typically 4° to 5°C warmer than the preceding glacial period, with the difference much larger at the poles than at the equator. Glacial periods are much longer than interglacials, and are the norm for the Quaternary, the warm interglacials are the anomaly. As discussed in Nature Unbound I and in Tzedakis, et al., 2017, we have had 13 interglacial periods in the past one million years. These are identified with red bars in Figure 3 (Javier’s figure 12).

Figure 3, Orbital obliquity increases, which correlate to July insolation peaks at 65°N, are colored. Red identifies successful interglacials and blue identifies a failure. The labels are MIS numbers. Low late-glacial temperatures (red circles below the blue dashed line) stimulate interglacials. High insolation at 65°N, the green circles above the green dashed line also stimulate interglacials. MIS 13 is an anomaly. Source Nature Unbound I.

The same interglacials are identified, with slightly different nomenclature, in figure 2 (our figure 4) of Tzedakis et al. The numbers in figure 3 and across the top of figure 4 are the Marine Isotope Stage (MIS) number, the odd numbers refer to “interstadials” which are warmer periods, separating the even numbered “stadials” or cooler periods. Notice that both Tzedakis et al. and Javier find more than one interglacial in MIS 7 and 15. We are currently living in MIS 1. Some interstadials are significant enough (as judged by the rise in sea level) to be labeled interglacials and some are not. One of the problems in Quaternary geology is how to objectively tell a true interglacial period from a common interstadial. Javier and Tzedakis, et al. have different criteria, but come to very similar conclusions.

Figure 4, Obliquity peaks are shaded in gray, the black line is the caloric summer half-year insolation at 65°N, the red circles are insolation maxima nearest the onset of interglacials, black diamonds are continued interglacials, light blue triangles are failed interstadials. The orange line is the δ¹⁸O stack representing temperature. The upper numbers are MIS numbers for interglacials and the lower are kyrs (thousands of years) before present or the number of a continued interglacial or a failed interstadial. The “Mid-Pleistocene Transition” toward lower-frequency higher-amplitude glacial cycles is apparent near MIS 38/37. Source Tzedakis, et al., Nature, 2017.

Javier’s methodology for identifying interglacials begins with locating every period of rising obliquity which creates a window that can initiate an interglacial. Fewer than half of these periods results in an interglacial. Next, he looks for the periods where summer insolation at 65°N exceeds 550 W/m² and where the temperature of the preceding glacial period is below 4.55 0/00 δ¹⁸O. δ¹⁸O is a common proxy for atmospheric temperature because the colder it gets, the less ¹⁸O is found in glacier ice . The boundaries and the resulting classification are shown in figure 3.

Tzedakis (2017) uses a different methodology that results in the same set of interglacials for the past one million years. The methodology is summarized in figure 5.

Figure 5: Temperature peaks for the last 2.6 million years separated into successful interglacials (red dots), failed interglacials (blue diamonds), continued interglacials (black diamonds) and uncertain assignments (open symbols). The dashed black line separates successful interglacials from unsuccessful interstadials with only two misclassifications (59 and 63). The ramp in the dashed line is the “mid-Pleistocene transition.” Source: Tzedakis, et al., 2017.

Figure 5 plots effective energy required to cause an interglacial versus time. As can be seen more effective energy is required to initiate an interglacial over the past 600,000 years than before 1.5 million years. In figure 4, interglacials (red dots) were more frequent and more regular before 1.5 million years ago, when they corresponded to the obliquity cycle of 41,000 years. Peak summer solstice insolation at 65°N is a function of the 21,000-year precession cycle. But, rising obliquity enhances the “caloric half-year insolation at 65°N” which is more relevant to ice loss. Prior to 1.5 million years ago, every other insolation peak at 65°N was boosted by increasing obliquity and an interglacial would occur. The idea of “caloric summer half-year insolation” originated with Milanković.

More recent interglacials occur about 100,000 years apart, meaning more insolation peaks are skipped now than before 1.5 million years ago. Thus, recent glacial periods are longer now and average ice volume is larger today than in the past. The ramp between the two horizontal lines is the mid-Pleistocene transition (MPT). Effective energy is computed using equation one from Tzedakis, et al., 2017. It is computed using the caloric summer half-year insolation peak at 65°N in (GJ/m²) and the time since the previous interglacial period. Tzedakis, et al. explain including the time since the previous interglacial in terms of ice stability. That is, the longer the ice has existed and the thicker it is the more unstable it is.

Why current interglacials require more effective energy to initiate is not known. Tzedakis, et al. list several possible reasons, but do not offer a preferred theory. Why current glacial periods are more severe today than prior to 1.5 million years ago, is also not known.

Clark, et al. 2006 have noted that the severity of glacial periods and the total land-ice volume increased dramatically after the mid-Pleistocene transition. The additional land-ice present now, versus before the MPT, represents a decrease of 50 meters of sea-level equivalent. While land-ice volume increased after the MPT, the area covered with ice did not, suggesting that average land-ice thickness increased. Clark, et al. (2006) also estimate a decrease in in global deep-water ocean temperature of 1.2°C currently, relative to the pre-MPT period of 41,000 year glaciations. Thus, we are not only in a major ice-age, we are also in the coldest part of the current ice age.

So, although Javier and Tzedakis, et al. used different criteria they did identify the same interglacials for the past million years. Tzedakis et al.’s method is able to classify all but two interglacials correctly for the past 2.6 million years and their method only uses orbital forcing and elapsed time as input. This last point is important as they found no need to incorporate either CO₂ concentration or δ¹⁸O records. This suggests that glaciations are caused solely by astronomical forcing, although the reason for the MPT is unclear. Tzedakis, et al. is also important because they seem to have resolved most, if not all, outstanding problems with the original Milanković theory.