Guest Post By Renee Hannon
In the mid-1900’s many scientists were suggesting the Earth was cooling. Now scientists are forecasting global warming. Indeed, instrumental data shows global temperatures warmed by approximately 1-degree C during the past 165+ years. With warming rates of 0.5 to over 1.3 degrees C per century this has caused considerable alarm for many. This recent warming is commonly attributed to increasing greenhouse gases, primarily CO2.
This post examines natural paleoclimate trends and simple characteristics of past and present climate cycles at different time scales. Data suggests distinct differences between short-term climate variability and longer-term climate change. This is important because short-term climate variability can be misinterpreted as underlying climate change resulting in poor science and potentially worse policy decisions. This post compares modern instrumental trends to paleoclimate trends. This comparison reveals modern warming has characteristics of natural short-term climate variability and not long-term climate change.
Comparing modern instrumental measurements to long-term paleoclimate data is not a simple task. They are vastly different types of datasets. Ice core paleoclimate isotope data are indirect indications of temperature (proxies) over millions of years compared to instrumental temperature measurements with high resolution of hours, days and decades. However, paleoclimate data cannot be ignored or dismissed when trying to understand present-day temperature trends. Paleoclimate characteristics and trends provide the overarching framework and climate history to better understand centennial temperature fluctuations and potential future global temperature tipping points. Earth’s natural baseline of historical climate must be established prior to any attempt to assign potential human impacts.
“Weekly or daily weather patterns tell you nothing about longer-term climate change (and that goes for the warm days too). Climate is defined as the statistical properties of the atmosphere: averages, extremes, frequency of occurrence, deviations from normal, and so forth.” Shepherd.
Climate timeframes and relative scales observed in paleoclimate data are illustrated in Figure 1. The glacial cycle repeats approximately every 100,000 years and consists of an interglacial and glacial period. Cold glacial conditions predominate 70% of the time while warmer interglacial conditions occur about 30% of the 100,000 years. This entire cycle has repeated four times over the past 400,000 years. The glacial cycle and occurrence of interglacial warm periods are commonly accepted as being influenced by the Milankovitch astronomical processes.
Figure 1: The climate framework using EPICA Dome C ice core temperature proxies. a) a glacial cycle over 100,000 years with warm interglacial periods in red and the long glacial period in between. Termination events and onset of the interglacial period are labeled. GM refers to glacial maximum. b) zoom in of the interglacial period (MIS 5e). Each interglacial period consists of a warming, plateau, and cooling segment, medium term events such as climate optimums and intervening cool events, and the smallest detectable climate variations in black that last only hundreds of years.
The glacial and interglacial periods are very different climate cycles and are composed of different patterns and events. The interglacial framework and short-term events superimposed on all segments of the interglacial period are examined below.
Interglacial Periods Define the Long-Term Underlying Trends
There are approximately five interglacial periods ranging in duration from ten to thirty thousand years during the past 500,000 years. The marine isotope stage (MIS) terminology is used in this post. Figure 2 shows the correlation of four of the interglacial warm periods. MIS 7e was omitted from this correlation due to its unique and unusual character of several short interglacial periods. Further discussion of MIS 7e can be found here.
The entire cycle for the interglacial warm period is defined from the glacial maximum to the next significant low temperature minimum. This cycle is subdivided into warming onset, plateau, and cooling segments. Figure 2 shows the systematic and repeatable sequence of these three segments.
Figure 2: Correlation of interglacial warm periods over the past 400,000 years. Present day HadCrut data is shown in red on the Holocene MIS 1 temperature proxy curve. Three key segments are highlighted; red is the onset warming, yellow is the interglacial plateau, and blue is interglacial cooling. Multi-millennial events are labeled as optimum, Younger Dryas (YD), 8.2 kyr event, and corresponding intervening cool events in past interglacial periods. Dansgaard-Oeschger (D-O) events are labeled and are mostly associated with the glacial period. Note the high frequency temperature fluctuations superimposed on the various segments of all interglacial periods.
Amplitude and Duration: The most significant events are terminations of the glacial period and rapid onset of global warming of the interglacial period. These events are frequently referred to as Terminations I-V. The interglacial warming onset shows the largest temperature increases of 5-7 degrees C globally and up to 12 degrees C in the Antarctic dome C data. This dramatic increase in temperature occurs rapidly over 5,000 to 7,000 years as glacial sheets begin to decrease in size, sea levels rise and greenhouse gases increase. This warming process eventually reaches an interglacial optimum and plateau.
The interglacial plateau shows variations in temperatures of approximately 1-4 degrees C and lasts from 10,000 to greater than 20,000 years. MIS 7e plateau was an exception lasting only 6,000 years. Earth is currently within the Holocene interglacial plateau that has lasted for more than 11,000 years so far. During this time, sea level is greater than minus 20 meters relative to present day and the Northern Hemisphere (NH) is predominantly ice free except for the Greenland ice sheet (Berger et. al).
The interglacial plateau is followed by global cooling of 4-6 degrees C and up to 8 degrees C in the Antarctic data that takes 7,000 to 13,000 years to re-enter the next glacial period. Many scientists propose that decreasing obliquity, or Earth’s tilt, is responsible for initiating the cooling tipping point. The initiation and growth of ice sheets occurs, ocean temperatures begin to cool, sea level falls and greenhouse gases gradually decline during global cooling.
Rate of Change: The temperature rate of change was determined for the warming onset, plateau and cooling segments of the interglacial period. Trendlines are calculated using the linear regression analyses by the “least squares” method shown in Figure 3.
Figure 3: MIS 5e as an example for establishing trends and rate of change for the interglacial period segments. The warming and cooling trends have a strong correlation coefficient (R2), whereas the plateau tends to have a lower R2.
Figure 4 compares trendlines for each warming onset, plateau and cooling segment of the five interglacial periods. Note duration is in years. The starting points are pinned at zero, except for the global warming phase which was pinned at minus 10 degrees C. These simple trendlines provide visualization of the underlying long-term interglacial trends and removes medium and short-term internal variations as well as noise.
Figure 4: Rate of change or trendlines for the past five interglacial segments from EPICA Dome C temperature proxies. Initial starting points are pinned at zero except for the global warming trendline which is pinned at -10 degrees C. The length of the trendline approximates the duration of the interglacial segment. a) rate of change for the interglacial plateau segments. Marcott and May’s Holocene global reconstructions are included. b) rate of change for warm onsets also referred to as Terminations and c) rate of change for interglacial cooling segments.
Interglacial trends over the past 400,000 years exhibit steep warming onsets, slower cooling rates and nearly flat plateaus. Average warming onset rate of change is approximately 2.0 degrees C/millennium in the Antarctic with exceptionally strong correlation coefficients of 0.98. Average plateau rate of change is minus 0.01 degrees C/millennium (excludes MIS 7e) with weaker correlation coefficient of about 0.5. Interglacial cooling is less than 1.0 degree C/millennium with strong correlation coefficients of 0.95. Since the warming rate is twice as fast as the interglacial cooling rate, the typical interglacial period has an asymmetrical pattern suggesting Earth heats up due to natural processes more rapidly than when it cools.
During the interglacial plateau, trendlines are relatively flat. MIS 5e, 7e, and 9 have an early and strong climate optimum which resulted in an overall cooling trend for the plateau. The Holocene MIS 1 global temperature reconstructions over the past 11,000 years by Marcott and May show a slight overall cooling trend compared to the Antarctic Dome C MIS 1 temperature proxy. MIS 11 which has the longest plateau shows a slight warming trend due to a later second climate optimum.
The warming onsets are very consistent for MIS 5e and 9 as the trendlines practically overlie each other. MIS 7e has a rapid onset and a very short plateau, if any. MIS 11 has the slowest warming onset rate. The Holocene’s warming onset started out like MIS 5e and 9 but was interrupted by the Younger Dryas (YD) cooling event.
Interestingly, the most consistent trendlines are the interglacial cooling rates which demonstrate a narrow deviation for the past four interglacial periods. MIS 9 cooling trend was measured up to a stadial event which is why it appears shorter. Interglacial cooling rates demonstrate consistent, predictable trendlines suggesting Earth’s climate follows similar, repeatable processes such as ice growth rates and oceanic/atmospheric process interactions as it cools.
Holocene Past Millennia shows Cooling Trends
If Earth continues harmoniously in step with natural processes, the next significant tipping point will be the Holocene interglacial cooling. Scientists generally agree the Earth has been cooling over the past several thousand years at an average global rate of -0.20 degrees C/millennium.
An interesting paper by Stenni et. al. incorporated seven different Antarctic ice core regions consisting of 112 records and used four different reconstruction methods. They observed a Holocene cooling trend in the Antarctic of -0.26 to -0.40 degrees C/millennium for the past 1900 years prior to present day warming of the most recent 200 years. Figure 5 compares these recent trends to the past interglacial cooling trends.
Figure 5: Comparison of the rate of change or trend for the past 1900 years bracketed by the Holocene plateau trend and average interglacial cooling trend. The Holocene plateau and cooling trends are from EPICA Dome C temperature proxies and the past 1900-year trends are from Stenni’s Antarctic region reconstructions. Initial starting points are pinned at zero and projected out in time. Marcott and May’s Holocene global reconstructions for the interglacial plateaus trendlines are included for comparison.
It appears the past millennia Holocene cooling trends in the Antarctic are approximately half way between the 11,000-year Holocene plateau trend and approaching global cooling trends of the past four interglacial cycles. The average interglacial cooling trend from the Dome C data is approximately 0.7 degrees C/millennium and represents the next climate change tipping point.
Interglacial Short-term Events Display Steep Trendlines
Understanding short-term cycles is important. Unfortunately, they are not well defined on the ice core temperature proxies due to resolution difficulties and local latitude differences. Short-term events within the Holocene interglacial period include the Medieval Warm Period (MWP), Roman Warm Period (RWP), Little Ice Age (LIA), and other cool events such as 4.2, 5.9, 7.2 and 8.2 kyr events. Some of these events are not as obvious on the Dome C data. However, they are more pronounced on northern latitude ice core temperature proxies. Global temperature reconstructions also tend to dampen the amplitude of these smaller events.
Davis, et. al conducted an extensive study on the Vostok ice core data examining centennial events. This study concluded the sample resolution of the Vostok ice core data can detect centennial scale cycles. However, it is inadequate to detect decadal scale cycles. Spectrum frequency analyses over the past 12,000 years show centennial scale events on four different Antarctic temperature proxy records at 193, 318, 379 and 493 years. Several figures from Davis’ study are compiled in Figure 6.
Figure 6: a) spectral power density periodogram of Vostok temperature-proxy records over the Holocene for 12,000 years showing six peaks. b) example correlation of the centennial peaks and troughs over four isotope records; Vostok, EPIC Dome C (EDC), EPIC Dronning Maud Land (EDML), and Talos Dome (TD). c) table showing the statistics for the TOc350 cycles with the present day HadCrut and UAHv6 statistics in red.
Interestingly, Holocene temperature peaks in the above periodogram are similar to Holocene solar peaks in the Lomb-Scargle periodogram shown in Javier’s figure 62a. The de Vries cycle is approximately 208 years and the Eddy cycle is approximately 970 years. Davis has attributed the broad 1000-year peak to the Bond cycle.
Amplitude and Duration: Davis identified 650 individual cycles of Temperature-proxy Oscillation (TO-c350) cycles in the Vostok data over the past 220,000 years. At least 60 occurred within the Holocene that are correlative over four Antarctic ice records with an example shown in Figure 6b. The TO-c350 events show temperature amplitude averages of 0.7 degrees C and an average cycle duration of 350 years. Davis speculates these cycles are the result of oceanic oscillations which frequently operate on a centennial scale bases.
Rate of Change: These centennial cycles tend to have steep warming trends, or segments, followed by steep cooling segments. Warming and cooling rates are rapid averaging 0.43 degrees or more per century. They have a high frequency of occurrence, rise and fall within hundreds of years and are highly variable. These short-term events have a rapid peak turnaround in temperatures and lack a plateau or have an extremely brief plateau lasting several decades at most.
Modern instrumental temperatures and warming rates are added to the table in Figure 6c. Their significance is discussed below.
Modern Warming Displays Steep Trendlines
Recent temperature measurements over the past 165+ years based on satellite, marine and land instruments obtained and analyzed by HadCrut, GISS, and Berkeley indicate global temperatures have increased by approximately 1 degree C shown on Figure 7. This calculates into warming rates of 0.5 degrees C to 0.7 degrees C per century. The southern hemisphere trend for HadCrut shows a slightly lower trend of 0.47 degrees C per century. In the past 38 years, global average lower tropospheric temperature (TLT) anomalies show trends up to 1.3 degrees C per century from UAH datasets compiled by Spencer and Christy. This steeper trend can also be seen on the HadCrut, GISS and Berkeley plots from 1976 to present.
Figure 7: a, b d) plots of global temperature in degrees C since 1850 from Hadcrut, GISS, and Berkeley combined land and ocean datasets. Rate of change per Century and correlation coefficient shown on each plot. c) UAH global temperatures from TLT are from 1976 to present day, a much shorter period and shows the highest rate of change.
The instrumental data is a good example of how “…trends for short periods are uncertain and very sensitive to the start and end years…” as noted by the IPCC WG1AR5. As an example, since 1980 to present day, instrumental temperatures show an overall warming trend of 1.8 degrees C per century (Berkeley).
The timeframe of instrumental records needs to be mentioned. Recent instrumental data spans 165+ years during the past 11,000+ years of the Holocene interglacial warm period as shown on figure 2. Instrumental records represent a very small subset (1.4%) of the Holocene interglacial plateau and a tiny blip in geologic time.
Classification of Climate End-Members
By establishing natural trendlines for different levels of climate temperature oscillations it is possible to improve our understanding of how the past 165 years have deviated from the natural baseline. While ice core proxies and instrumental temperature measurements are different datasets, general characteristics and trends of climate oscillations do exist. Comparisons may not be quantitative, however qualitative trends and observations are frequently used with imperfect data to develop scientific hypotheses that can be tested.
Once a natural short-term climate variation from the natural paleoclimate baseline is established, only then can the potential Anthropogenic influence be further investigated.
Rate of Change: By comparing instrumental rate of change to the interglacial plateaus shown in Figure 8a, there is a significant difference in trendlines. Instrumental trends have much steeper slopes (0.6 degrees C/century) than the flat longer-term interglacial plateau (0.001 degrees C/century or 0.01 degrees C/millennium). Projecting the 165-year instrumental trends suggests within 500 years temperatures will reach 2.5 to 3.5 degrees C warmer than present day. Figure 8a demonstrates and amplifies the conundrum of the recent warming trend compared to long-term trends and to reconstructions. This was illustrated quite nicely by the “Mann hockey stick”. Reconstructions tend to dampen the short-term amplitudes. Marcott states that his reconstruction preserves variability for periods longer than 2000 years, only 50% at 1000-year periods, and no variability less than 300 years.
Figure 8: Comparison of modern warming trends with interglacial long-term trends. Starting points are pinned at zero temperatures and zero timeframe for the duration. Instrumental trends are projected for 500 years. Interglacial trendlines projections approximate their duration. a) present day warming trends and interglacial plateau trends. MIS 7e is omitted in the plot. May and Marcott’s global reconstructions rates are included. b) Comparison of instrumental trends with past Termination warming onset trends.
Comparing instrumental warming trends to the Termination onset warming slopes shown in figure 8b is highly informative. Terminations I through V are significant paleoclimate events where the termination of the glacial state occurs, and Earth begins to change into a warm interglacial state. These long-term trends show the largest increase in temperatures of any paleoclimate event during the past millions of years. Importantly, present day slopes are much steeper than the interglacial warm onsets even on the polar Antarctic datasets.
Here, modern warming trend is compared to both shorter-term events as well as long-term interglacial trends.
Figure 9 compares absolute warming and cooling tends against their durations for various scales. Included in the plot are the interglacial only events from the Vostok TOc350 dataset by Davis, slopes for the past 100 and 1900 years from the Antarctic region by Stenni et. al., and trends calculated for Dome C interglacial segments discussed in earlier in this post.
Figure 9: Plot of rate of change in degrees C/century and duration for individual warming, cooling and duration segments. Black dots are instrumental Berkley and UAHv6 rate of change. Blue dots are interglacial segments calcuated by this author, orange dots are Holocene warming rates from Davis, TOc350. Green octagons are from the Antarctica region for 100 and 1900 year durations by Stenni.
Decadal data cannot be extracted from ice core temperature proxies and is missing from this plot. Instrumental data (black dots) are plotted using both the 38-year rate of change of 1.3 degrees C/century and the 165-year rate of change of 0.6 degrees C/century.
Figure 9 shows two distinct classifications; one group is less than +/-500 years of duration and the other is greater than +/-700 years.
The group with less than 500+/- years duration are predominately short-term warming and cooling segments. They typically don’t have a plateau long enough to establish a trend. Warming and cooling rates range from 0.01 to >3.0 degrees per century. Interestingly, the small-scale events which have the shortest duration and smallest temperature amplitude tend to have the steepest trendlines. This is also visibly noted on Figure 1b where the black short-term oscillations are superimposed on all segments of the interglacial period underlying long-term cycle in red.
The long-term interglacial segments demonstrate a clear and distinct second group. Warming, cooling and plateau segments that last longer than 700+/- years consistently have rates of less than approximately 0.25 degrees per century. This includes the geologically significant Termination onsets. These long-term events also exhibit the largest temperature ranges up to 12 degrees C in the Antarctica and Greenland data.
Distinguishing Climate Variability from Underlying Climate Change
The IPCC has the following definitions for climate variability and climate change:
“Climate variability refers to variations in the mean state and other statistics (such as standard deviations, the occurrence of extremes, etc.) of the climate on all spatial and temporal scales beyond that of individual weather events.”
“Climate change refers to a change in the state of the climate identified (e.g., by using statistical tests) by changes in the mean and/or the variability of its properties and that persists for an extended period, typically decades or longer.”
Although the above definitions are qualitative in nature, Figure 9 distinguishes climate change from climate variability. As more data on multi-centennial events becomes available in the future it will help narrow the range. For now, multi-centennial warming and cooling trends less than +/-500 years in duration can have a wide variance in rates of change. However, warming and cooling trends greater than +/-700 years in duration which are the underlying long-term trends have rates of change less than approximately 0.25 degrees C per century based on Antarctic data. These are true climate change events. Climate variability best describes the shorter multi-centennial trends which are merely the overprinting internal oscillations on longer-term climate change.
Currently, instrumental temperature characteristics are consistent with natural climate variability of short-term events and steep warming trendlines.
Scientists studying climate change should entertain multiple working hypotheses. We’re all familiar with the CO2 hockey stick with ever increasing global temperature projections into the 21st century. Past natural climate characteristics suggest another hypothesis. One that Modern Global Warming is part of a natural warming segment of multi-centennial climate variability and the impact of CO2 is overestimated. The modern high rates of change are certainly in line with past natural short duration events. And in a couple of centuries, there may be a quick turnaround at its peak (no to minor plateau) and then a cooling multi-centennial segment will ensue.
Every several centuries Earth will experience these false alarms of high rates of temperature changes both positive and negative. Recognition of short-term cycles and resulting adaptation strategies should be different than for longer-term climate change. These multi-centennial cycles of climate variability will continue over the next several millennia until the true underlying Holocene interglacial long-term cooling begins to take Earth back into the next glacial period.
Characteristics in paleoclimate data occurring prior to industrial revolution cannot, by definition, be attributed to anthropogenic forcing. Paleoclimate data can establish a natural baseline for long, medium, and short-term climate cycles. The Holocene and past interglacial plateaus are characterized by high frequency multi-centennial climate variability that has different characteristics than the underlying longer-term multi-millennial climate change.
Climate Change tipping points and underlying long-term framework are the interglacial onset, plateau, and eventual cooling. These cycles demonstrate dramatic temperature changes of up to 12 degrees C. The next natural long-term tipping point for Earth will be a cooling at the end of the Holocene interglacial period. It will take thousands of years for Earth to enter and progress through this phase, providing Earth along with its ecosystems and inhabitants time to adapt as necessary.
Climate variability has a high frequency of occurrence, short durations of less than about 500 years and can demonstrate warming and cooling rates greater than 0.25 degrees C/century. Currently, instrumental and satellite temperature data exhibit rates of 0.5 to 1.8 degrees per century that are well within the assemblage of natural centennial events. Therefore, present modern warming is likely a natural multi-centennial warming segment that will soon begin a temperature turnaround.
This post compared modern instrumental trends to paleoclimate trends. The comparison reveals present day warming is not consistent with rates of change (temperature) observed in long-term climate change. The rate of change for modern warming is consistent with climate variability.
Acknowledgements: Special thanks to Andy May and Donald Ince for reviewing and editing this article.