Fig. 1 Vostok ice core data 420,000 year record of climate change.
Vostok Ice Core Data over a 420,000 year time span from Petit et al. Nature (1999).
The Vostok ice core samples are astonishing in the wealth of information they contain. Careful analysis of samples of ancient ice by literally hundreds of scientists has revealed earth's climate history over the past 420,000 years. Moreover, this extends our knowledge through four previous cycles of world-wide glaciation and interglacial warming.
It takes some effort to comprehend the amazing detailed data in Fig.1.
First, check the axes. The time axis of the graph starts at zero and proceeds into the past. The temperature data is plotted in the red color curve. The temperature of what? Analysis of the Vostok ice cores provides us with an average mean temperature of ancient near-surface ocean water.
Here are a few simple observations:
Examine the temperature record in the data plot. Notice that the temperature oscillates. That is, characteristic features in the temperature data repeat in time. The temperature has a periodic nature.
Temperature plotted at the zero of the time axis axis comes from analysis of the most recently formed ice. Ice samples recovered by the Vostok project from deeper deposits provide us with the geochronology of the ice. We get the properties of the earth's oceans extending back through 420,000 years. Four hundred and twenty thousand years!
Over this long period of time temperatures have varied in a narrow range of approximately 10 degrees Celsius or 18 degrees Fahrenheit. Further, the most recent temperatures in the plot are not unusually warm for this phase of the climate cycle. This can be seen by comparison to earlier main temperature peaks in previous climate cycles. Can you identify the main temperature peaks? There are five of them in Fig. 1 including the one at time zero. A previous main temperature peak occurs at approximately 125 thousand years ago. Another is at approximately 240 thousand years ago. What were the maximum temperatures in these previous climate cycles?
Keep in mind these are mean temperatures, averaged over many years. Daily and seasonal temperature oscillations are averaged-out to give the average mean temperature. This type of data is, therefore, well suited for the study of long term climate cycles.
How do you get the ocean temperature from ice samples?
This fascinating temperature data was obtained from hydrogen isotope concentration measurements in ancient frozen ice samples retrieved from deep under the surface of an antarctic glacier.
How does that work?
It turns out that hotter near-surface ocean temperatures tend to concentrate the amount of deuterium (D), the heavy hydrogen isotope, in near-surface ocean water. This results in higher D/H ratios in the ice. This is handy since the main constituent of oceans is water consisting of DHO and H2O molecules. When this ocean water is frozen and trapped in an ice sheet, the D/H ratio stops changing. Hence, careful measurement of D/H concentration ratios from the ice core gives temperature as a function of depth and time. Really a simple concept, but requiring lots and lots of very careful work by many scientists involved in the Vostock project. The result of their prodigious effort has produce the data plotted in Fig. 1, which we can now examine.
What else is plotted in Fig. 1?
What about CO2?
During the freezing process tiny bubbles of air are trapped in the ice. When ice core samples are examined as a function of depth, we have a time history of the composition of earth's atmosphere preserved in those air bubbles.
How did atmospheric CO2 concentration change over these climate cycles?
Examine the CO2 concentration history shown in the blue curve in Fig. 1.
Notice that atmospheric CO2 concentrations (ppmv: parts per million by volume) in past climate cycles are comparable, and in some cases larger, than present day values.
[We need to put in a reference for present day atmospheric CO2 concentrations, and near surface ocean temperatures from isotope analysis. Woods Hole guys may have this.]
When compared to previous climate cycles in Fig. 1, present day atmospheric CO2 concentrations do not appear to be anomalously high or low. Rather, they are comparable to CO2 levels near and after the onset of global cooling phases of previous climate cycles. This is inter-relationship is explained in more detail in the article by Gregory Fegel, excerpted above, and in the following section, "Earth's Oceans, Lakes, and Streams Regulate Atmospheric CO2."
Want more information on the Vostok ice core data?
What causes Climate Change?
The frequency of these temperature oscillations is predicted with remarkable accuracy by precise mathematical physics calculations of orbital mechanics of the Earth, Sun, Moon system. This calculation is credited in the above article and is called Milankovich Theory.
[The success of Milankovich theory in predicting the frequency of climate cycles (ice ages and warm interglacial periods etc.) is roughly equivalent to the success of quantum mechanics in predicting the optical line spectrum of the hydrogen atom. Such agreement is considered substantial confirmation of physics based theories.]
The Milankovich theory successfully predicts these climate cycles
without need of further assumptions. The beauty of the Milankovich theory is that it predicts the main features of climate cycles from a very simple understandable process. When the earth is exposed to more intense sunlight on average, it gets warmer.
One can embellish this picture by adding other effects. For example, the known physical changes in the light output of the sun, sunspot epochs and the like, or changes in the solar wind hitting the earth can be added to the mix. Further, one can include corrections to the earth's overall infrared emissivity factor caused by atmospheric carbon dioxide and changes in the reflectivity of the earths surface or albedo. Albedo is a measure of how deserts and ice sheets reflect more sunlight back into space than do forests. Deserts tend to cool the planet. All very interesting, however, these processes do not appear to be needed to explain the main features of climate cycles. Milankovich seems to have gotten it right.
Earth's Oceans, Lakes, and Streams Regulate Atmospheric CO2
Carbon dioxide (CO2) is present in gaseous form in the atmosphere as we have all heard. CO2 also exists in water as a solution. CO2 dissolves in ocean water like it does in a can of carbonated soda. There is a natural tendency for gaseous CO2 in the atmosphere to equilibrate with CO2 dissolved in earthly bodies of water. The rate of this equilibration is governed by solution chemistry and transport processes. Generally, increasing atmospheric CO2 drives more CO2 into solution, and visa-versa.
Ocean temperature plays an important role in this equilibration. Hotter ocean temperatures tend to drive more CO2 out of solution and into the air. That is, higher ocean temperatures
cause increased atmospheric CO2 concentrations by this mechanism. One can expect that long timescale variations in atmospheric CO2 are strongly coupled to ocean surface temperature. We can say that earth's oceans, lakes, and streams are huge reservoirs for storage and release of CO2. True for other gases too.
Atmospheric CO2 is also converted into non-gaseous molecules by chemical reactions including various bio-metabolic processes. Plant photosynthesis removes CO2 from the atmosphere. In oceans, shell formation by plankton and other shelled creatures sequesters carbonates permanently. For example, great geological deposits of chalk consist of shells from deceased microscopic sea creatures, accumulated over eons. Such shell formation can be viewed as a process that removes carbonates from the ocean permanently. Natural sequestration of carbonates.
The above qualitative discussion highlights some important geo/biochemical process that tend to regulate atmospheric CO2. However, atmospheric CO2 concentrations themselves do not seem to play a primary causative role in Earth's climate cycles. Climate cycles on Earth seem to be largely explained by the Milankovich Theory.
Also plotted in Fig. 1 are:
The methane concentration indicator (CH4), the Oxygen 18 indicator (An indicator of global ice mass? Some discussion on this below. Expert comments welcome.), and the solar radiation indicator roughly proportional to the total averaged power from the sun (mostly in the form of light) that was striking the earth at the given time period.
___________________________________________________________________________
Break Time and a Pop Quiz
Now for an entertaining thought experiment. In thought experiments we get to ask: "What happens if we could ....?"
Here's an interesting one. What happens if we could ....
Put the earth in a huge thermos bottle (Dewar flask for chemistry majors). No light or heat gets in from the sun, and nothing gets out of the bottle.
After the earth is in the bottle, we start monitoring the near surface ocean temperature.
Q1: At first the ocean temperature would (1) increase (2) decrease (3) stay the same.
Q2: In the long term the ocean temperature would eventually have to (1) increase (2) decrease (3) stay the same.
If you know the answer please leave a comment. Explain your choice.
___________________________________________________________________________
What about the past Five Million Years?
Further data and discussion on global temperature fluctuations, ice ages, warming ages.
The five million year history of global climate variation from Lisiecki and Raymo in the journal Paleoceanography (2005).
Fig. 2 Five Million Years of Climate Change from Analysis of Ocean Floor Sediment Cores [Lisiecki and Raymo in the journal Paleoceanography (2005)]
Here Delta-T is the temperature variation relative to the present era baseline. Delta-T is inferred from measured variations of Oxygen 18 concentration in deep sea floor core samples. The time scale in Fig. 2 covers the past 5 million years. The present era is at “year zero” on the time axis and time marches backward to the right. See below for details of data analysis.
A few observations about the data in Fig. 2 follow.
First, very near the year zero time, one can see a strong warming trend immediately after the most recent temperature minimum. The minimum occurred at roughly 0.05 million years before the present. This most recent temperature oscillation corresponds to the most recent ice age and the subsequent present day warming trend.
Going back further in time, one sees repetitive short term warming and cooling oscillations. This relatively rapid cycle repeats on a roughly 50 kyr to 100 kyr timescale. Over the past few million years this oscillation frequency is trending toward lower frequencies. Also evident in the Fig. 2 data is a long term cooling trend during the past 3 million years.
Mysterious magnetic pole migration and flipping.
Another geophysical curiosity is the phenomenon of polarity transitions in the earth's magnetic field. Paleo-oceanographic records from other research projects have shown the earth's magnetic field undergoes periodic transitions called pole-flips. During these pole-flips, the north and south magnetic poles move around and exchange positions. Magnetic pole-flips are generally believed to be the result of a magnetohydrodynamic dynamo instability. As yet there is no quantitative physical theory of this phenomenon. Interestingly, some planets like Mars and Venus have little or no geophysical magnetic field.
The most recent of Earth's magnetic pole-flips occurred about 780,000 years ago. Did that pole flip disturb the temperature oscillations in Fig. 2? On casual inspection I cannot see anything unusual at that time. Of course, the climate effects of the pole-flip may have been only a short term transient of duration less than 1000 years, maybe much less. It would be interesting to obtain ice cores at 780,000 years at depth of ~6000m (don't know if such ice deposits exist) and use Vostok like techniques to look for evidence of climate effects of the pole flip.
Here is a link to a detailed cartoon diagram of the Vostok region.
http://www.sale.scar.org/images/new_vostok_cartoon_high.jpg
___________________________________________________________________________
Side Bar: Why are the temperature fluctuations smaller one to five million ago?
As an expert at reading and analyzing time series data, but non-expert in paleo-oceanography, I want to mention an interesting feature of the data in Fig. 2. Namely, the amplitude of temperature oscillations seems to systematically decrease going back in time, over millions of years.
Here is a brief side bar on my personal thoughts on this apparent trend toward lower fluctuations evident in Fig. 2.
Is it real or maybe an artifact?
Maybe the climate cycles were less severe one to five million years ago, or maybe there is a systematic error smoothing out the delta-O18 oscillations in older, deeper sediments.
One can ask: What would be the affect on the data analysis if concentrations of heavy carbonate molecules, those carrying the O18 isotope, undergo some diffusion while buried in the sediment layers?
Diffusion of molecules would tend to smooth out O18 concentrations making the temperature oscillations look smaller. Moreover, there would be a greater smoothing effect in older sediments, in part because diffusion would have a longer time to act. Interestingly, such diffusion would not change the frequency of the oscillations much, just round them off and smooth out the bumps. The delta-O18 oscillations would tend to get smoother as one goes back in time. Which is what they seem to do in Fig. 2.
What factors affect diffusion rates of carbonate molecules in deeply buried sediments?
Diffusion rates are governed by the diffusive transport coefficient, D, of the material containing the O18 carbonate concentrations. Larger values of D make for more rapid smoothing of concentration gradients.
We can make some qualitative statements about factors affecting the value of D. Generally increases in D can be expected for sediments subject to (1) increased temperature as a function of depth (2) changes in the micro-structure and porousness of the compressed solid state, (3) increased pressure as a function of depth, (4) variations in composition. Overall, I would expect the transport coefficient D to be an increasing function of depth under the sea floor.
All of this is rather straightforward, and is probably well-known by experts in this field. Undoubtedly, the authors were aware of this. They may have determined that diffusion is negligible after all, or it may be that values of D are simply not available because of the uncertainties mentioned.
Perhaps the long term drift to smaller amplitude temperature oscillations going back over five million years, may in part be an artifact of this diffusion process. It seems at least plausible and worthy of more discussion.
Expert and Non-expert comments welcome on this topic.
___________________________________________________________________________
Now ... back to the analysis methods used to get Fig. 2.
What is delta 18-O and delta 18-O Benthic Carbonate?
In both Fig. 1 and Fig. 2 a quantity called “delta of oxygen 18” is plotted. These quantities in the data plots actually refer to somewhat different drilling and analysis methods used by authors of the two referenced papers. Below is a brief overview of this type of oxygen 18 analysis and theory.
In Fig. 1 the data comes from ice cores obtained by drilling into deep ancient ice in the Antarctic. The deeper the ice, the older the ice. In fact the top axis in Fig. 1 gives the depth of each ice sample in meters and the bottom axis gives the corresponding age of the ice deposit obtained from that depth.
In Fig. 2 drilling cores come from the sea floor and consist of deep layers of accumulated sediment. Chronology is again established from the-deeper-the-older concept. In both studies oxygen isotope analysis is used on material samples from the drill cores.
Here is a rough explanation of this isotope analysis.
Oxygen atoms come in two isotopes 16O and 18O having atomic mass 16 and 18 respectively. Any sample of oxygen from nature will contain both isotopes. The ratio of these concentrations is a measurable quantity called the concentration ratio, and it's average value (over many locations and samples) is called the natural isotopic abundance ratio. This is true for oxygen gas, dissolved oxygen, and for oxygen in molecules like carbonates. Importantly, there are isotope enrichment processes that occur in the ocean. These include physical processes such as diffusion and evaporation which tend to mobilize the lighter isotopes leading to small but measurable variations of the isotope abundance ratios. Moreover, these physical processes are strongly dependent on the ambient temperature. Hence, their effects, such as concentration changes, can be used as an indicator of temperature change. The laboratory techniques for isotope abundance measurements are amazingly accurate when done by experts.
Next, we need samples of material to analyze.
Scientists go "into the field" to collect these samples. Core drills similar to those use for oil exploration are used to extract cylindrical sections of subsurface material. Scientists keep track of the depth below surface giving the geochronology of the material. The data in Fig.1 were obtained from drill core samples from an expedition to a south polar ice sheet. The data in Fig. 2 is from drill core samples from the ocean floor. These samples are then taken back to the lab for analysis, like the CSI the guys on TV. Measurements of the oxygen-18 oxygen-16 isotope ratios can be done with great accuracy using sensitive mass spectrometers and thermal desorption techniques. As you can imagine it requires a huge amount of scientific work to get the data that is plotted in the above Figures.
In order to track the natural isotope enrichment history, one tracks the deviation in the enrichment from the natural isotope ratio. The deviation from the natural ratio is referred to as “delta 18O” and is the quantity plotted in the Figures.
What does "benthic carbonate" mean?
The term “benthic carbonate” refers to carbonate samples that originated in the benthic zone. The benthic zone is the ecological region at the lowest level of a body of water such as an ocean or a lake, including the sediment surface and some sub-surface layers. Benthos are organisms inhabiting the benthic zone. [Wikkipedia]
The plot, Fig. 2, of “delta 18O from benthic carbonate” refers to oxygen isotope analysis of calcium carbonate, CaCO3, deposits from drill core samples obtained from the sea floor. The calcium carbonate is believed to originate from skeletal remnants, shells, bones etc. of bottom dwelling foraminifera and from accumulated skeletal remains of other ocean dwelling creatures.
What are bottom dwelling foraminifera? (This lovely definition from Steven J. Gould, I think.)
"Bottom dwelling foraminifera are sea creatures that crawl, walk, or slither around on the sea floor."
____________________________________________________
Scientific quality variations in the field of climatology.
Examples of sloppy thinking and common misconceptions regarding the above data in a Wikipedia article from 2009
Popular discussions of these climate data range from thoughtful, to careless, to bordering on disinformation. The following example is from a current Wikkipedia entry (dated 2009) relating to Fig. 2. My comments and corrections are added in italics.
Wikipedia Article Text and Judgmental Comentary
Figure 2 above is from the paper of Lisiecki and Raymo (2005). It consists of data from combined measurements on 57 globally distributed deep sea sediment cores. The measured quantity is the oxygen isotope fractionation [sic] (delta 18-O) in benthic foraminifera [not clear such foraminifera are the sole source for delta 18O, some discussion needed here] , which serves as a proxy for the total global mass of glacial ice sheets. ["(delta 18O) a proxy for total global mass of glacial ice sheets"?? Lisiecki and Raymo use the delta-18O as an ocean temperature indicator, as is evident in Fig. 2. The main figure in their paper!]
[As far as I can see, the quantity "delta 18O" is an indicator of variations in ocean temperature. Considerably less clear, and likely to be nonsense, is the assertion that delta 18O may be a strong indicator, e.g. proportional to ... The mass of all the ice on the surface of the earth, as claimed by the Wikipedia person]
[More discussion and Expert comments welcome!]
Lisiecki and Raymo constructed this record by first applying a computer aided process of adjusting individual "wiggles" [sic] in each sediment core to have the same alignment (i.e. wiggle matching) [sic].
["wiggle matching" is not a scientific term. The author may be referring to timeseries data analysis techniques including, periodogram, spectral analysis, and other more sophisticated mathematical tools used for analysis of periodic and quasi-periodic data. "wiggle matching" sounds like nonsense and it is. It seems likely the wiki author has no understanding of the actual methodology used in the hard sciences to establish geochronology of the core samples in the referenced papers.]
Then the resulting stacked record is orbitally tuned [sic] by adjusting the positions of peaks and valleys to fall at times consistent with an orbitally driven ice model (see: Milankovitch cycles).
[Here the term ‘orbitally tuned’ is used by the wiki authors to describe well-known cycles of insolation caused by variations in the earth’s orbit and it’s slowly changing axis of rotation. These cycles can be calculated rather precisely using Newtonian mechanics and can be extended into the distant past, even back to 5 million years ago.]
Both sets of these adjustments are constrained to be within known uncertainties in sedimentation rates and consistent with independently dated tie points (if any). Constructions of this kind are common, however they presume that ice sheets are orbitally driven, and hence data such as this cannot be used in establishing the existence of such a relationship.
[If you find the above Wikipedia text confused and unclear, you would be correct. It is in fact poorly written.]
[It seems likely that the Wiki authors do not appreciate the mathematical precision of such calculations. They are not ’assumptions,’ rather they are mathematically calculated quantities stemming from basic Newtonian orbital mechanics.]
The observed isotope variations are very similar in shape to the temperature variations recorded at Vostok, Antarctica during the 420 kyr for which that record exists. Hence the right hand scale of the Figure was established by fitting the reported temperature variations at Vostok (Petit et al. 1999) to the observed isotope variations. Hence, this temperature scale should be regarded as approximate and its magnitude is only representative of Vostok changes. In particular, temperature changes at polar sites, such as Vostok, frequently exceed the changes observed in the tropics or in the global average. A horizontal line at 0 °C indicates modern temperatures (circa 1950). [This is peripheral criticism.]
[The data in Fig. 2 clearly show the variation in the temperature indicator over an extensive time interval with very nice time resolution. These features of the time series are independent of the absolute temperature calibration factor.]
[That is, the Fig. 2 data give us a clear picture of cyclic oscillations in temperature that is more important than their precise magnitude. Moreover, the evident quasi-periodicity of these data is in quantitative agreement with well known secular changes in the earth's orbit and rotation axis orientation.]
[The success of Milankovich theory in predicting the frequency of climate cycles (ice ages and warm interglacial periods etc.) is roughly equivalent to the success of quantum mechanics in predicting the optical line spectrum of the hydrogen atom. Such agreement is considered substantial confirmation of physics based theories.]
Labels are added to indicate regions where 100 kyr and 41 kyr cyclicity [sic] is observed. These periodicities match periodic changes in Earth's orbital eccentricity and obliquity respectively, and have been previously established by other studies (not relying on orbital tuning).
Wiki authors do not give a satisfactory discussion of the significance of their term "orbital tuning" in this context.
For discussion of how such orbital changes might drive climate change, see Milankovitch cycles.
[This Wiki article, from 2009, is an example of sloppy analysis and bad science. Sadly it seems rather typical of the low scientific and analytical standards of the literature promoted by Believers.]