How X-Ray Fluorescence Methods Reveal Past Climate Changes


Geology tries to understand the past by looking at the contemporary world. There is the famous phrase of uniformitarianism “Present is the key to the past”, which can be understood as the “laws” of nature have acted more or less the same throughout the geological past. Therefore, in order to comprehend the marks of ancient glaciers on rocks, you better know or have seen a modern one. On the other hand, James Hutton (called as the founder of modern geology) in his famous article Theory of the Earth; or an investigation of the laws observable in the composition, dissolution, and restoration of land upon the globe, he states that:

In examining things present, we have data from which to reason with regard to what has been; and, from what has actually been, we have data for concluding with regard to that which is to happen thereafter.

Hence, extracting data today provides a mirror of the past, and allows us to forecast the future. On the other hand, we should put a note here. While geologists conduct research, most of the time their aim is not to understand the future, it is mostly an academic curiosity.

Nowadays, climate change is one of the hot topics through the society and scientific community, climate projections and meteorological forecasts are in our everyday life. Dynamical climate forecast models would be almost impossible to construct if there were no meteorological data to compare with training data. Meteorological forecasts would be impossible if we didn’t know the meteorological setting of the moment and a couple of hours before. Furthermore, it would be impossible to know what kind of climate we can encounter in the future without knowing the past.

Direct measurements of weather span not more than 200 years or so, where the older the data it has lower resolution and higher measurement errors. Now, let’s see, what may be the disadvantages of having that much short temporal data, in terms of geological timescales. Fig 1. is the stable oxygen isotope data from a Greenland drilling project. The reader can ignore the unit of the y-axis and read the data as temperature changes in Northern Hemisphere high latitudes.

Fig 1. The stable oxygen isotope data from a Greenland ice core. It can be read as temporal changes in temperature. From top to bottom, note the timespan of the data changes. See, how the dynamic structure of climate changes as well if you expand the timespan.

The upper curve shows the changes between 38 and 39 kyr BP (read as thousand years before present). According to this 1000 years of data, one may say that the temperatures are showing an increasing trend. However, if we expand the time scale (the middle curve that represents the changes between 30 and 60 kyr BP), the temperatures show a more complex pattern. One may state that it shows almost a periodic pattern. Now, let’s take a look at the whole Greenland data (lower curve), which spans the last 125 kyr BP. The record becomes much more sophisticated and now it is much harder to define a pattern. Therefore, it is essential to appreciate the past records to speak about future.

What are the methods to learn past changes? Although this topic is beyond the scope of this article, the major ones are paleontological parameters, geomorphological features and geochemical proxies gathered from marine and terrestrial deposits.

In the 19th century, James Croll (a janitor at the museum of the Andersonian University in Glasgow) formulated a hypothesis that Earth’s climate is affected by the changes in the parameters of the orbit of the Earth around the sun in geological timescales. However, his hypothesis has been abandoned at the early stages of the 20th century, since geomorphological features did not reflect the number of glaciations and their timings. In 1976, with the advancement of coring technology and usage of geochemical proxies made it possible to falsify the geomorphological tests which rejected Croll’s theory, and confirmed the hypothesis of James Croll. Since then, geochemical proxies are seen as one of the most reliable sources in paleoclimatic interpretations.

Some of the most famous geochemical proxies are stable isotopes (especially of oxygen, hydrogen, and carbon), XRD (X-ray diffraction) analyses and XRF (X-ray fluorescence) scans. Neither of them is a perfect and explanatory power of all these proxies depends on the location of the samples and their resolutions.

Fig 2. This is a photo of geologists on a platform sailing to Lake Salda to drill.

Drilling cores, and analyzing them is the most widely used techniques of the Quaternary (last 2.6 million years) paleoclimate studies. Sediment cores are interpreted as history books. Basically, the top of the core represents today, and the bottom represents a time in geological history. The core as a whole, except some conditions, gives almost a continuous story of the past environment of the drainage basin.

XRF scanning, by emitting X-ray radiation and ejecting electron etc. (the physics behind is beyond the scope), generates a spectrum of the sample, and by doing this, gives semi-quantitative counts of some elements to the researcher. Relatively new core scanning technologies provide a technique such that the whole core can be put under an XRF scanner.  Furthermore, this technology made it possible to scan the cores up to 0.2 mm resolution, which means it scans almost continuously. Therefore, XRF counts are valuable resources for the paleoenvironment studies.

Fig 3. This is an example photo of a core section (on the left) and some of the corresponding XRF elemental profiles, Rothwell, R. G. (2015).

XRF counts can be interpreted in many ways, according to the geography, drainage basin, or the basin’s characteristics. For example, a study uses Ti (titanium) count as an indicator of paleo-precipitation since Ti is an immobile element and has a relatively low affinity. But, another study may use K (potassium) normalized to Ca (K/Ca) as a paleo-precipitation indicator. Some other profiles may be used for erosion, organic content, redox conditions etc.

Other than basin based approaches, there also exist some statistical approaches to “decipher” the XRF data. One can use linear methods such as factor analysis to find out the governing elemental profiles, or principal component analysis to find orthogonal directions with maximal variance. However, in nature linearity is hardly achieved. In a recent study of ours, we offered a newer nonlinear method for paleoclimate studies. Independent component analysis, which is widely used in signal processing, telecommunications etc., rotates the axes of the XRF data to a new set of axes such that extracted data are statistically independent of each other. By doing this so, rather than looking at the sole data or normalizing them, the authors claimed to reveal the precipitation and temperature records of the studied site.

XRF data provide a rich and almost continuous set of data, suitable to apply statistical analyses. Therefore, in order to understand the extent of contemporary climate change by looking at the past, XRF data give invaluable information.

This study, Climate proxies for the last 17.3 ka from Lake Hazar (Eastern Anatolia), extracted by independent component analysis of μ-XRF data, was recently published in the journal Quaternary International. The Author, Z. Bora Ön is affiliated with Muğla Üniversitesi, Department of Geology and İstanbul Teknik Üniversitesi, Avrasya Earth Sciences Institute.


Rothwell, R. G. (2015). Twenty years of XRF core scanning marine sediments: What do geochemical proxies tell us?. In Micro-XRF Studies of Sediment Cores (pp. 25-102). Springer Netherlands.

Written By
More from Z. Bora Ön

The Performance Of A Vanadium Redox Flow Battery With Non-Uniformly Compressed Electrodes

Electrical energy storage technologies are an important solution for a mismatched problem...
Read More
Opinions expressed are solely the authors and do not express the views or opinions of Science Trends nor the author's institution.

Leave a Reply

Your email address will not be published. Required fields are marked *