Mass spectrometry and climate science. Part I: Determining past climates

by Roland Hirsch

Mass spectrometry is essential for research in climate science.

Understanding climate requires having sufficient knowledge about past climate and about the important factors that are influencing climate today, so that reliable models can be developed to predict future climate.

Analytical chemistry enables measurement of the chemical composition of materials, from the amounts of elements and their isotopes in a sample to the identity and concentrations of substances in the most complex biological organisms.

This two-part series covers the application of a powerful analytical chemistry technology — mass spectrometry — to two important areas in climate science:

  • Obtaining reliable information about past climate
  • Understanding composition and behavior of aerosols, which have a large impact on climate

The examples that are included for each topic were selected out of many published papers on the study of climate using mass spectrometry, partly because they feature a very wide range of types of these instruments. The authors were very helpful in providing me with information on their work.[1]

The technology described in this essay may at times be quite complicated! However, I hope that the results of each study will be understandable.

Part 1: Determining past climate

Information about past climate is quite limited. The atmospheric temperature records obtained using satellites and covering nearly all of the earth start only in 1979. Surface temperature records cover only a small portion of the earth, perhaps 15% going back to about 1900, and much less before that. The Argo buoys were deployed ~15 years ago and cover much of the oceans, which had minimal coverage before then. Information about aspects of past climate other than temperatures is even more limited.

Analytical chemistry is providing improved information about past climate, for example:

  • Temperatures, changes in climate, and extreme weather events
  • Concentrations of CO2 and other atmospheric components
  • Extent of sea ice and glaciers over time
  • Impact of geological events such as volcanic eruptions and earthquakes

The first step in studying a sample that was formed in the past (such as a fossil or a layer of sediment under the ocean materials that is more than a few hundred years old and have no attached information) is to determine its age.

The most common way to do this is using mass spectrometry to measure an elemental isotope ratio that is dependent on age. Carbon is often used for dating once-living samples, such as plants or artifacts made of wood, as its main isotopes, carbon-12 (12C) and carbon-13 (13C), are stable, while carbon-14 (14C) is radioactive, with a half-life of about 5700 years. The supply of 14C in the air is constantly replenished by cosmic rays hitting nitrogen-14, but once an organism dies the 14C fraction will steadily go down. [i]

Mass spectrometry is a preferred technique for measuring ratios of isotopes of elements, and reliably determines the fraction of 14C in the once-living sample and thus the age at the end of its life for up to about 20,000 years ago, and possibly somewhat earlier. For older samples other elements must be used, for the 14C fraction is too small to be reliably measured.

Several other elements have naturally-occurring long-lived radioactive isotopes that could be used to determine the age of an older sample. Potassium, for example, is a widely-distributed element with substantial concentrations throughout the earth. Potassium-40 (40K) is a very long-lived isotope (1.25 billion years) with (unusually) two forms of decay, one to stable argon-40 and the other to stable calcium-40.

Argon in the atmosphere contains the 40Ar produced over the lifetime of the earth. However, if the decay occurs in a solid that did not contain any air when it solidified and does not allow the argon produced by potassium decay to escape, then the amount of 40Ar reflects the date when the solidification took place.[ii]

The reverse also can be measured: the amount of 40Ar in air trapped in ice will reflect the age at which the ice formed. [iii] This mass spectrometric technique has recently been used to determine the age of samples at different depths in an Antarctic ice core (Figure 1).[iv] [v] The authors of these studies then measured the amounts of different atmospheric gases in those samples and plotted them as a function of sample age (Figure 2).

Figure 1: Age of samples taken at indicated depth below surface of ice core

Figure 2: Results for key atmospheric gases as function of ice core age (various techniques were used for measuring the gases)

Mass spectrometry is often used to study past temperatures. The ratio of the trace isotope oxygen-18 (18O) to the most common isotope oxygen-16 (16O) in a once living organism depends on the temperature of the air at the time that the oxygen was incorporated into the organism by metabolism. The ratio is measured using mass spectrometry. The higher the temperature, the lower the fraction of 18O in the once-living organism.

A representative study was carried out on mollusks found in layers of sediments near the northwest shore of Iceland, covering the period 350 B.C. to A.D. 1600. The layers in the shells reflected the year-round temperatures in which the mollusks had lived (Figure 3). From such data on many samples at different layers in the sediments the authors constructed a chart of temperatures over that time period. Notably, they were able to correlate these temperatures with historical records in Iceland from 865 to 1600 (Figure 4). The authors pointed out that “On the basis of δ18O data, reconstructed water temperatures for the Roman Warm Period in Iceland are higher than any temperatures recorded in modern times.”[vi]

Figure 3: Example of temperatures derived from a shell that lived through four summers (S) and three winters (W)

Figure 4: Variation of temperatures from the Roman Warm Period to ~AD 1800.Information from historical documents is at the top on the right side.


Another study using mass spectrometry to determine temperatures using oxygen isotope ratios was carried out over in fjords in Sweden, covering a 2500 year period. One of the fjords is on the north coast of Sweden (Atlantic Ocean) and the other on the southwest (North Sea near Denmark). The authors state:

“The record demonstrates a warming during the Roman Warm Period (~350 BCE – 450 CE), variable BWT [bottom water temperatures] during the Dark Ages (~450 – 850 CE), positive BWT anomalies during the Viking Age/Medieval Climate Anomaly (~850 – 1350 CE) and a long-term cooling with distinct multidecadal variability during the Little Ice Age (~1350 – 1850 CE). The fjord BWT record also picks up the contemporary warming of the 20th century (presented here until 1996), which does not stand out in the 2500-year perspective and is of the same magnitude as the Roman Warm Period and the Medieval Climate Anomaly.[vii]”

The authors of this study include a chart relating their temperature information with others in the North Atlantic (Figure 5).

Figure 5: Bottom water temperatures for different locations in the North Atlantic Ocean going back to about 350 BCE. Abbreviations shown at the top of the chart: RWP represents the Roman Warm Period, DA represents the Dark Ages, VA/MCA represents the Viking Age/Medieval Climate Anomaly and LIA represents the Little Ice Age.

The oxygen ratios thus allow estimating temperatures for once-living organisms. Importantly, they allow doing so year-round. Tree ring diameters, which are sometimes used for estimating past temperatures, primarily reflect temperatures during the growing season, and are also influenced by factors such as rainfall and location (such as above or below the treeline at a given time).

A recent study provides a different use of the oxygen isotope ratios for studying past climate: do volcanos have an impact on major climate factors such as the El Niño – Southern Oscillation (ENSO)? Some studies have suggested that major volcanic eruptions can impact the ENSO cycle, but only a few such events have relevant weather data. Fossil corals in the regions impacted by ENSO dating back centuries have well-defined monthly layers. They were dated by mass spectrometry using U/Th ratios. Then oxygen-18 measurements for these layers allow estimation of temperatures coinciding with major volcanos.

The results were combined with prior studies to produce a temperature record covering from ~1100 to ~2000 CE. Six major volcanos in this time period were charted against the temperatures (Figure 6). No evidence was found that the volcanos had caused an ENSO event. [viii]

Figure 6: Coral δ18O measurements of temperatures for six major volcanic eruptions in the last 900 years, with lines at the bottom showing the degree to which stratospheric aerosols reduced downwelling sunlight during the volcanos

Conclusion for Part One

This completes the first post on applications of mass spectrometry to climate science. The second post will focus on studies of factors that need to be understood to be able to develop reliable models of climate, with emphasis on research on aerosols.


[1] The figures and charts and other information from the papers are the property of the authors and publishers.



[iii] M.L. Bender, et al., “The contemporary degassing rate of 40Ar from the solid Earth”, PNAS (2008) 105, 8232-8237

[iv] J.A. Higgins, et al., “Atmospheric composition 1 million years ago from blue ice in the Allan Hills, Antarctica”, PNAS (2015) 112, 6887-6891

[v] Y. Yan , et al., “Two-million-year-old snapshots of atmospheric gases from Antarctic ice”, Nature (2019) 574, 663-663   and

[vi] W.P. Patterson, K.A. Dietrich, C. Holmden and J.T. Andrews, “Two millennia of North Atlantic seasonality and implications for Norse colonies” PNAS (2010) 107, 5306-5310

[vii] I.P. Asteman, H.L. Filipsson, and K. Nordberg, “Tracing winter temperatures over the last two millennia using a north-east Atlantic coastal record”, Climate of the Past (2018) 14, 1097–1118.

[viii] S.G. Dee, et al., “No consistent ENSO response to volcanic forcing over the last millennium”  Science (2020), 367, 1477-1481


Roland Hirsch has served the field of analytical chemistry in a 52-year career that spans teaching, research, and leadership at Seton Hall University, and 33 years of government service at the National Institutes of Health and the U.S. Department of Energy.  Roland has been a leader of the ACS Division of Analytical Chemistry, as Councilor for 25 years, as Division Secretary for 4 years, Chair-Elect, Program Chair, and Chair, and as its Web Editor for 22 years. Roland organized the 50th-anniversary celebration of the Division and 25 years later, wrote the definitive history of the first 75 years of the Division, published in Analytical Chemistry in 2013.  Roland has also been active in ACS Governance, including Chair of the Committee on International Activities, Secretary of the Committee on Nominations and Elections, Member of the Committee on Divisional Activities, Senior Chemists Task Force, Committee on Committees, and Liaison to the ACS Committee on Professional Training.

Based on a presentation prepared for the American Chemical Society National Meeting in Philadelphia in March 2020. It was to have been in the Division of Analytical Chemistry’s session “Advances in Mass Spectrometry”. The meeting was canceled, but this presentation was revised and made available on the web site for the meeting:

28 responses to “Mass spectrometry and climate science. Part I: Determining past climates

  1. Studying the shells of giant clams, Chinese researchers (Yan, et al., ‘Higher sea surface temperature in the northern South China Sea during the natural warm periods of late Holocene than recent decades’) used a common paleothermometer known as δ18O to measure strontium to calcium ratios and reconstruct sea surface temperatures (SST) to compare historical and current SSTs (see wiki, generally re delta-O-18 and Sr/Ca –e.g. the precise Sr/Ca ratio in the coral skeleton shows an inverse correlation with the seawater temperature during its biomineralization). The researchers found that both the Roman Period (RWP, about 2,000 years ago) and the Medieval Climate Anomaly (MCA, about 1,000 years ) were warmer than today (Modern), based on current sea surface temperatures.

    • Curious George

      We may be overestimating the accuracy of the δ18O as a temperature proxy. Remember the hockey stick graph, where a proxy paleo temperature from tree rings was spliced with modern thermometer measurements, to show a rise instead of a decline? Wikipedia gives a formula relating δ18O to temperature – dating from 1953. Did nobody repeat the calibration in the last 67 years?

  2. Any way of co.paring the Modern for years after 2005,say to 2015 or later?

  3. Reblogged this on uwerolandgross.

  4. Ireneusz Palmowski

    After production in the upper atmosphere, the carbon-14 atoms react rapidly to form mostly (about 93%) 14CO (carbon monoxide has been estimated to be roughly 12 to 16 years in the northern hemisphere. The transfer between the ocean shallow layer and the large reservoir of bicarbonates in the ocean depths occurs at a limited rate.[25] In 2009 the activity of 14C was 238 Bq per kg carbon of fresh terrestrial biomatter, close to the values before atmospheric nuclear testing (226 Bq/kg C; 1950).

    • Ireneusz Palmowski

      Nuclear tests in the stratosphere had a greater impact on changes in the chemical composition of the upper atmosphere than changes in galactic radiation.

  5. Ireneusz Palmowski

    This post presents evidence from Russian scientists describing how those same Cosmic Rays (GCR) have a dramatic top-down effect on atmospheric circulation by interacting with ozone in the stratosphere. The basic idea is that the climate effects from increasing cosmic rays vary according to Arctic polar vortex shifts from fast and strong, to weak and wavy, resulting in alternating climate epochs.

  6. Dave Winterflood

    I have printed whole article and tapped and printed each graph. Both only produce blurred and not sharp annotation and words. We need see the details of time line and data and sources. Please remedy and advise me on

  7. Dave Winterflood

    I have gone to etc. but they are restricting membership and access to your paper.
    I included the name of our company and got strange institutional names pop up.
    So no good there. Perhaps ask them their game plan.

  8. M.I.T. Professor Richard Lindzen: belief within climate science that CO2 essentially controls global temperature is recent, wierd and probably wrong:

  9. Geoff Sherrington

    Analytical chemistry was once my main work, so I gained some knowledge about strengths and weaknesses. In the middle of this learning, there were some papers about analysis of lunar rocks and soils done by the most venerated labs at the time, mostly government and academic rather than private. Mass spectrometers were expensive. The accuracy of the labs was compared and it was a shocker. (One author was George H Morrison). My lesson was that analytical uncertainty must be measured by experiments like round robins, standard samples etc and not by self-generated expressions of excellence. Too few analysts now abide by this lesson.
    The second comment here notes that no matter how good the chemical analysis is, it can be upset by wrong interpretation. From above “The higher the temperature, the lower the O-18 fraction in the once-living organism.” Here, that temperature is described as the air temperature at the time of growth of the organism. While this might be correct as a broad observation, one has to ask about the temperature. Is it the night time low temperature, the daily high, the average over a day or a month or a year, the average temperature of the region or the neighbourhood temperature where the organism grew (maybe unrepresentative, like atop a high mountain)? Or, if the organism was mobile and moved through various temperatures, which one applies? Is it a summer air temperature or a winter one for short-lived organisms? Is it the temperature of the air blown in on regular patterns like the roaring forties? You get the picture.
    Here, we meet accuracy again, or more accurately, uncertainty. How does one conduct the statistical mathematical procedures that valdly express the uncertainties inherent in assumptions like these? The short answer is that most authors do not even try in climate research, placing reliance on faith in beliefs rather than in hard science measurements.
    We have a crisis of reproducibility. You have just read of a major cause of the problem. Geoff S

  10. Pingback: Mass spectrometry and climate science. Part I: Determining past climates |

  11. Pingback: Mass spectrometry and climate science. Part I: Determining past climates – All My Daily News

  12. Roland

    This is a very good article of special interest to me as a Historical climatologist. I specialise in the English and Near European climate from around 1086 i.e. from the Domesday book onwards . The records earlier are often good but need to be sorted out from the religious and superstitious elements that were a large part of observations.

    Like others I had great difficulty in actually reading the charts and graphs as they have not come out clearly..
    You say;

    “Information about past climate is quite limited”

    I am noit sure I would agree with you. From Manorial Records, farmers accounts, those from monasteries, from town clerks, from diaries and from other sources,in England and Europe at least we have some very good and very detailed records for much of the past thousand years with the caveat of superstation and legends etc.

    As you know Hubert Lamb collected many records together as did Phil Jonbes and there have been a number of books in recent decades where records have been assiduously collected by such as Le Roy Ladurie William Rosen, Kington, Behringer and Fagan and many others, each of these very well referenced. The Byzantium records throw some light outside of the Roman Western Empire and extend oiur knowledge by hundreds of years.

    In part 2 will you be making any estimates of past climates by way of some simple to understand graphs? I look forward to reading it


  13. 97% Settled science update:
    “Dr Helen Cleugh from the CSIRO told the commission [global warming] was interacting with, and exacerbating, previous weather systems in a way never seen before.

    “Perhaps put more simply, [global warming] means that ** the past is no longer a guide to future climate-related impacts and risks.” **

    Wait. What?
    We dug up Australian weather records back to 1838 and found snow is falling less often

    Guess they didn’t get the meme-o.

  14. Is it (laser ablation) inductively coupled mass spectroscopy that is employed for this isotope dating?
    I’m surprised that mass spec is better than radiological measurement even for carbon-14.

  15. Pingback: Weekly Climate and Energy News Roundup #412 |

  16. Pingback: Weekly Climate and Energy News Roundup #412 – All My Daily News

  17. Roland F. Hirsch

    Note: If you cannot access the presentation at ,
    please go to
    where you can obtain an ACS ID at no charge. It should allow you to sign in to view the presentation.

  18. For those who are interested in palaeoclimatology should aquaint themselves with the exellent work on the topic being conducted by Christopher Scotese on paleaogeogrphy and plate techtonics.
    I agree that spectroscopy is an important tool although there is very much more to understand.