Talk:Abundance of elements in Earth's crust

Data from WebElements source
The archived page at https://web.archive.org/web/20070309033534/http://www.webelements.com/webelements/properties/text/image-flash/abund-crust.html (reference 5) shows no data and the original page no longer exists as such there is no way to check the reference (unverifiable) should this data be removed?. — Preceding unsigned comment added by A wiki editor 42 (talk • contribs) 05:51, 22 February 2022 (UTC)
 * Indeed, if it can't be verified it may be deleted. That would be the whole column then.
 * BTW, column "Israel Science and Technology" source (reference 6) does not seem to be independent; just a reproduction of other source(s) if sourced at all. Delete? -DePiep (talk) 09:33, 22 February 2022 (UTC)
 * @DePiep I found https://webelements.com/hydrogen/geology.html (you can replace hydrogen in the URL with helium, lithium, and so on to get the rest of the elements) SVG-image-maker (talk) 13:42, 4 August 2023 (UTC)

There are problem with total %
CRC column does not give 100%, it gives more than 100%. So that this is some wrong numbers between first 13 elements. Just try to sums first 13 positions — Preceding unsigned comment added by 176.123.216.36 (talk) 16:03, 29 March 2022 (UTC)


 * The CRC column also doesn't sort properly; it seems like the table sorter is ordering entries lexicographically on that column rather than numerically, so for instance 585 (0.0585%) ranks as greater than 461,000 (46.1%); lastly, it's unclear why that column doesn't just list the percentages. 162.196.254.96 (talk) 08:34, 27 September 2023 (UTC)
 * OK, fixed the CRC sort. It seems that when you put in the value with %, it sorts wrong, but it also sorts wrong with the e template.
 * If there are more things to fix, I have a program to change them. Just ask. Oh, do we need the percentage for the CRC column? Gah4 (talk) 06:10, 28 September 2023 (UTC)
 * If there are more things to fix, I have a program to change them. Just ask. Oh, do we need the percentage for the CRC column? Gah4 (talk) 06:10, 28 September 2023 (UTC)

Problematic and confusing information
It seems from the archives of this page that the issue I'm about to flag has been raised before. Someone coming to this page wants to find out, quickly, the relative abundance of different chemicals. The first problem they have is that they encounter a table sorted in order, apparently listing the elements in order of abundance (so far so good). But on closer inspection, the table actually presents five attempts at a list from five different sources. Um, OK. Is one better or more reliable than the others? Why are there five? Why not one? Or fifty? Which one should be used by a school kid writing a report? If none of these is definitive or better than the others, how are we placed to present a table with rows apparently ranked in an order of abundance - if these sources disagree on the order (and they do, if you check the table around, say, nitrogen and cobalt). It's not even clear which of the sources has been used to come up with the numeric ranking of the table. In other words, wWhat does row 30 of the table actually indicate? It doesn't indicate that nitrogen is #30, because the sources seem to disagree on that. What we have is a table of apparently very accurate data (in the CRC column - to four decimal places) and yet wide disagreement - so the "accuracy" it seems to suggest is spurious? 45154james (talk) 11:47, 13 November 2022 (UTC)
 * The table is sortable on all columns. So, whichever column you sort on is the one you get. (It can also be sorted by name or Z.) If we try to say that one is better, that might be WP:OR.  I don't see ones with four digits, but many have three or two, which might be too much.  I am not sure that "crust" is defined accurately enough to say, and again suspect that one digit would be better. It is not required that we use the same number of digits as the source. I suspect that some are better at some elements, and worse at others. Gah4 (talk) 05:54, 14 November 2022 (UTC)
 * Thanks! Appreciate the clarification. But you see the problem. Suppose I am a school-kid writing my essay about vanadium. I come to this page, sort the table by each of the columns, and find vanadium ranks 19 (Darling), 20 (Barbalace), 18 (WE), 22 (IST) or 11 (CRC). So what do I write in my homework? :) Then I go to the article Vanadium and find "Vanadium is the 20th most abundant element in the earth's crust" - with an entirely different (random) source (none of the above). Yes, good point about WP:OR. Where the lead currently says "Estimates of elemental abundance are difficult because...", maybe it needs a few more words explaining that this is why sources differ, that there is no real answer...?? 45154james (talk) 11:14, 14 November 2022 (UTC)
 * I suppose so. But also, note, that as well as is needed by such student, the difference between them doesn't really mean much. Consider the season results for some sport. It is interesting to know which team came in first, second, and third, but doesn't matter much between 18th and 21st. Though school kids might not know that, yet. Reminds me of the students in high school, or even early college years, doing labs, and then wondering what the "right" answer is. There usually isn't one. Much of these abundances vary depending on where you look. And people haven't looked everywhere, and especially deeper into the crust. Now, how much does this article need to explain it? I suppose it could do some.
 * As far as I know, most often people want the abundance itself, more than the rank. Lately I have (in another forum) answered a few questions about the abundance of uranium. It is interesting how high it is, much higher than gold, and close to tin, which we commonly think of as a cheap metal. However, more important than abundance is the actual concentration where you are looking Many metals have places where they occur in much higher concentration. There aren't so many of those for uranium. Much of it is in low concentration well distributed. But the article also doesn't explain that. Gah4 (talk) 14:31, 15 November 2022 (UTC)
 * Thanks for your thoughts. I think you're convincing me that maybe a short paragraph at the start of the article outlining some of these very points might be a good idea. I might try to draft a few ideas later. Thanks again! 45154james (talk) 08:03, 16 November 2022 (UTC)
 * I suppose so. But also, note, that as well as is needed by such student, the difference between them doesn't really mean much. Consider the season results for some sport. It is interesting to know which team came in first, second, and third, but doesn't matter much between 18th and 21st. Though school kids might not know that, yet. Reminds me of the students in high school, or even early college years, doing labs, and then wondering what the "right" answer is. There usually isn't one. Much of these abundances vary depending on where you look. And people haven't looked everywhere, and especially deeper into the crust. Now, how much does this article need to explain it? I suppose it could do some.
 * As far as I know, most often people want the abundance itself, more than the rank. Lately I have (in another forum) answered a few questions about the abundance of uranium. It is interesting how high it is, much higher than gold, and close to tin, which we commonly think of as a cheap metal. However, more important than abundance is the actual concentration where you are looking Many metals have places where they occur in much higher concentration. There aren't so many of those for uranium. Much of it is in low concentration well distributed. But the article also doesn't explain that. Gah4 (talk) 14:31, 15 November 2022 (UTC)
 * Thanks for your thoughts. I think you're convincing me that maybe a short paragraph at the start of the article outlining some of these very points might be a good idea. I might try to draft a few ideas later. Thanks again! 45154james (talk) 08:03, 16 November 2022 (UTC)

Barbalance part 2
I propose to remove the Barbalace column, due to some wildly different values for B, Br and W. Also we cannot tell where Barbalace got its information from, so it is not actually a reliable source. If no one objects or does it before me, I will act later in April 2023. Graeme Bartlett (talk) 03:14, 25 March 2023 (UTC)


 * I agree. The column should be removed. The Boron abundance figure of 950 ppm sticks out like a sore thumb. It looks like a typo, given that the next column gives a figure of 8.7 ppm.
 * Also, and more importantly, the Barbalance column links to a website calling itself 'environmental chemistry' with a heavy bias to pollution issues. Even more importantly the link doesn't take us to a table with elemental abundance data, just a period table, and I couldn't easily find any such data on the site.92.12.82.119 (talk) 13:50, 25 March 2023 (UTC)
 * Well you have to click on each element and scroll down to find the info. But I don't hold the navigation against it, just the data and lack of providence. Graeme Bartlett (talk) 03:36, 26 March 2023 (UTC)
 * @Graeme Bartlett It's entirely possible that they meant 9.50, they just forgot the decimal point. You can try to contact Barbalace if it was just a typo. (the link is at the bottom of ) There's a section at  called "Periodic Table of Elements Bibliography" which is where the data comes from. In addition,  shows another source, namely N. N. Greenwood and A. Earnshaw's Chemistry of the Elements, 2nd Edition. SVG-image-maker (talk) 13:53, 4 August 2023 (UTC)
 * I have now removed the Barbalance column. Graeme Bartlett (talk) 10:28, 28 September 2023 (UTC)

Uranium
In some other site, I found the statement that the average (I suspect US) yard contains 2kg uranium. That is, to a depth of about 1m, for about 1000 tons of soil. A different way to look at abundance. Gah4 (talk) 06:17, 28 September 2023 (UTC)
 * That is a 2 ppm, consistent with the table. Graeme Bartlett (talk) 10:27, 28 September 2023 (UTC)
 * Yes, but in a different way than one usually thinks about it. If you ask just about anyone, whether they have 2kg of uranium in their yard, most will say no. Even people used to reading this table. Well, I seem to get a lot of Quora questions on the supply of uranium, which is not in short supply. But this makes it so obvious. Gah4 (talk) 00:15, 29 September 2023 (UTC)
 * A cubic meter of soil is around 2000 kg, so by my math 2kg/2000kg is 1000ppm. Or to put in the way proposed, 2000kg of soil contains 2g of uranium, not 2kg. Johnjbarton (talk) 02:30, 17 March 2024 (UTC)
 * But 2kg was for a whole yard containing 1000 tons of soil. Graeme Bartlett (talk) 10:09, 17 March 2024 (UTC)
 * Hah! 66'x66' a yard deep? Does not seem like a useful analogy to me. I've dug down a yard but most folks have not. Johnjbarton (talk) 14:28, 17 March 2024 (UTC)
 * Seems that soil might be density about 1.5. Not that you have to have dug that deep, but that you could, with appropriate machinery. Even more, you might believe that you own the soil that deep. Gah4 (talk) 07:17, 18 March 2024 (UTC)
 * Elements aren't uniformly distributed within the crust. Uranium in particular is fractionated in magmas, and is also concentrated in certain types of sediments. I don't know how this will effect the concentration in the average yard, but I find it conceivable that it is appreciably lower than the crustal average. Lavateraguy (talk) 16:53, 15 April 2024 (UTC)
 * Seems that soil might be density about 1.5. Not that you have to have dug that deep, but that you could, with appropriate machinery. Even more, you might believe that you own the soil that deep. Gah4 (talk) 07:17, 18 March 2024 (UTC)
 * Elements aren't uniformly distributed within the crust. Uranium in particular is fractionated in magmas, and is also concentrated in certain types of sediments. I don't know how this will effect the concentration in the average yard, but I find it conceivable that it is appreciably lower than the crustal average. Lavateraguy (talk) 16:53, 15 April 2024 (UTC)
 * Elements aren't uniformly distributed within the crust. Uranium in particular is fractionated in magmas, and is also concentrated in certain types of sediments. I don't know how this will effect the concentration in the average yard, but I find it conceivable that it is appreciably lower than the crustal average. Lavateraguy (talk) 16:53, 15 April 2024 (UTC)

Are the rankings in the references?
The table here gives references for the % abundance but not the rankings. Is that correct? Do any of the sources give the rankings directly? Johnjbarton (talk) 02:32, 17 March 2024 (UTC)

Proposal to remove the rankings for low abundance elements in the table
As far as I can tell, the ranking of the elements in the table is a (mild) form of WP:SYNTH. I propose to change the table to avoid these rankings.

There are sources for the rankings of the most abundant elements:
 * Lutgens, Frederick K., and Tarbuck, Edward J.. Essentials of Geology. United Kingdom, Prentice Hall PTR, 2000.
 * https://sites.tntech.edu/hwleimer/geol-1040/geol-1040-lecture/abundances-of-chemical-elements-in-the-earths-crust/
 * (I believe, based on some sites) CRC 97th (for top 10)

There are a few non-Wikipedia web pages that also rank elements, but they don't have sources either. These are basically the same concept AFAICT: list by ppm and sort.

In my opinion the rankings of the elements below the top 10 or so are not scientifically valid. Measurements of the low abundance elements are inherently imprecise and very likely inaccurate due simply to our limited ability to sample the Earths crust. You can see this in the table. For example, different sources give the ppm value for Zr as 130, 250, or 165ppm. Depending on which one you pick the ranking will be different.

A more valid presentation of this data would be a table for the top 10 (or more if we have a reliable source) and a second table with no ranking.

Beyond accuracy, my motivation is seeing IP users presistently adding rankings to eg Hafnium and Erbium.

Johnjbarton (talk) 15:50, 21 March 2024 (UTC)


 * A related idea would be to break down the table according to the Goldschmidt classification. So Top Ten, then other elements by Goldschmidt categories. This would give the table more value as geology knowledge. Johnjbarton (talk) 18:12, 21 March 2024 (UTC)
 * Why not a single sortable table? Goldschmidt classification could be one of the columns, as could 1-10 ranking with everything > 10 listed as 11+. Color code the rows for the Goldschmidt classification. YBG (talk) 01:40, 22 March 2024 (UTC)
 * Oh the color-code rows sounds cool. I was looking into adding a Goldschmidt column but coding the rows might be easier. I guess it would be done by styling the rows. (Sorry I don't what you mean by "single sorted table"? Is single-sorted a thing or do you mean "one table"?) Johnjbarton (talk) 01:54, 22 March 2024 (UTC)
 * Wikitable scan be sortable. I am suggesting having a single table with a Goldschmidt column then users can sort it by that column. One table can then serve many different purposes. YBG (talk) 02:09, 22 March 2024 (UTC)
 * The table is already sortable. If the table is split, you can no longer sort everything. YBG (talk) 02:12, 22 March 2024 (UTC)
 * I put the row coloring in. (And @YBG I deleted one of your edits which I didn't understand).
 * Please review. Johnjbarton (talk) 03:21, 22 March 2024 (UTC)
 * This looks good to me. I agree that inherent imprecision means that we can't really rank them outside the top few (and I suppose putting At at the bottom, but that's mostly trivia). Double sharp (talk) 03:56, 22 March 2024 (UTC)
 * @Johnjbarton- Some notes
 * Per MOS:COLOR, color should never be the sole indicator of any information. Adding a Goldschmidt classification column resolves this and additionally enables sorting by this characteristic.
 * In the edit you reverted, I combined the headers for element name and symbol so the symbol column didn’t take up so much room. Doing this by itself meant it was no longer possible to sort by symbol. To restore this capability, I added a row of empty cells. If I were to do it again, I would move the references to this row. With this explanation, would it be ok for me to re-combine the element headers for name and symbol?
 * Do we really need four sources? I’d say pick one to use consistently abd fill in any blanks with data from another source, with a note to indicate this fact. It might also be good to add notes in situations if significant differences between the sources.
 * Do we really need both % and ppm? Percentages are really only useful down to maybe 0.1%. I’d say either eliminate them entirely or at least eliminate them below 0.5%
 * YBG (talk) 06:20, 22 March 2024 (UTC)
 * Thanks I did not know about the MOS:COLOR.
 * Is combining name and symbol worth it? Your other suggestions will solve the width.
 * No, four sources add no value. We have no idea where the data from Darling comes from and the WebElements data is not specific. The Israel S&T numbers point to the Kring conf. proceeding but that only has oxides. The Kring paper has a long list of refs. I've not looked at the CRC edition cited but historically it has refs to its sources making it much more reliable. Two "sources" would have the advantage of altering the reader to intrinsic variability of these values.
 * I would go with cutting the % below some value, you pick.
 * Johnjbarton (talk) 15:50, 22 March 2024 (UTC)
 * @Johnjbarton: Is there a disadvantage to combining the name and symbol headers using the scheme I used that puts the sorting icons in a separate row? YBG (talk) 01:03, 23 March 2024 (UTC)
 * @YBG try it, let's see. Johnjbarton (talk) 01:35, 23 March 2024 (UTC)

After removing the rankings here, do we need to remove the rankings in all the element articles too? Which elements can stay? --Nucleus hydro elemon (talk) 14:55, 22 March 2024 (UTC)


 * We should remove any that are not referenced as rankings. I suggest a separate small table here in this article with the reference or we could embed the data in the big table. That way we'd have a quick way to check other articles. I'll post a ref and listing if I find one. Johnjbarton (talk) 15:54, 22 March 2024 (UTC)


 * The title of the table in this Wikipedia article claims it is for "Earth's (continental) crust" but the Darling", "Web Elements" and "Israel" sources are all for "Earth's crust". These are not the same - Earth also has oceanic crust. The first two websites are themselves unsourced and the third website cites a source for Earth's crust (published 100 years ago) and a source for continental crust.
 * The "Darling" website is self-published information by a man with a PhD in Astronomy (not geology or geochemistry) and whose employment history is software creator then long-time popular science writer. The website is basically a blog without sources. The "Darling" web page cited in the Wikipedia table even uses the graph from this Wikipedia article (without displaying copyright attribution - not a good sign; possible risks of WP:CIRCULAR?). The information on the website might be correct but there is no way to check this because it is unsourced. Therefore, the "Darling" column should be removed from the table because I think it is not a reliable enough source - see WP:SELFPUBLISH. (But, if the "Darling" column is kept for some reason, the cited web page has the abundance of tin as "trace", but in this Wikipedia table, it is 0. I think it would be preferable to replace this with blank).
 * I do not have access to a copy of the cited CRC source reference, but its publication details suggest it is a reliable source for element abundance in Earth's crust. (Google Books snippet view suggests that it does give data for Earth's crust rather than Earth's continental crust but I can't be sure). Some other reliable geochemistry source references exist that list element abundance for Earth's continental crust, Earth's oceanic crust, Earth's (primitive) mantle, and whole Earth, which I think would be more informative than the current 4 versions of probably incorrectly labelled "Earth's (continental) crust".
 * I suggest that the "Darling", "Web Elements" and "Israel" columns should be removed from the Wikipedia table because the data are unreliable and/or have been misrepresented on Wikipedia. GeoWriter (talk) 18:24, 22 March 2024 (UTC)
 * The question was raised in the discussion: whether we need deleting the abundance ranks on the pages of separate elememts. The answer was: to remain ranks that have a source. However, I think this is not sufficient. There may be several sources with different ranks for an elememt, while other elements may have their sources with controverse figures, that may conflict with the former ones. This lead to unnessesary edit warrings. My proposal: to specify the ranks of rare elements roughly. For instance in a range (min. and max., per two sources), or in simple approximation (e.g., the element is in the 6th dozen of the most abundant, per one source), or without any numbers at all, relatively to the ranking neighbors (element X is as rare as elements Y and Z, or element X is less rare than element W).
 * Tosha Langue (talk) 05:15, 15 April 2024 (UTC)
 * Regarding:
 * "The question was raised in the discussion: whether we need deleting the abundance ranks on the pages of separate elememts."
 * That broader question should be discussed on Wikipedia_talk:WikiProject_Geology since it impacts many pages. I opened a topic there, please join. Johnjbarton (talk) 16:22, 15 April 2024 (UTC)

How to represent Goldschmidt category in a column?
Per the discussion above the color alone should not be a value in the data table. A column can be added but what would it contain? A letter? L,A,C,S,T? Seems obscure but compact. The category name, Lithophile, Siderophile? Correct but takes a lot of space and is repetitive? I guess that would be the clearest. Johnjbarton (talk) 17:19, 22 March 2024 (UTC)


 * I’d say use the full name. YBG (talk) 00:58, 23 March 2024 (UTC)
 * ✅ Johnjbarton (talk) 22:27, 23 March 2024 (UTC)

Proposal to remove the "Darling", "Web Elements" and "Israel" columns from the table
Modifying the table is not trivial so I think we should get consensus on removing these three columns. Personally I would like to have a second column to give graphic evidence of variability, but we have a hard time even finding one solid source.

I agree with @GeoWriter, which I paraphrase as: these three source do not have a citation path back to reliable data sources. They are thus not secondary sources we should cite.

Please reply keep or delete. Requesting input from @YBG, @Nucleus hydro elemon, @Double sharp and anyone else of course. Johnjbarton (talk) 00:35, 23 March 2024 (UTC)


 * Delete YBG (talk) 00:53, 23 March 2024 (UTC)
 * Delete. BTW, there is another table with this data on p. 1294 of Greenwood and Earnshaw's Chemistry of the Elements (2nd ed.), derived and updated from W. S. Fyfe's Geochemistry; probably that one is more worth including. Double sharp (talk) 03:10, 23 March 2024 (UTC)
 * Delete. Also, the graph maybe has a different data source. I know H looks much more abundant because the graph is by atom, not by mass; but I have no idea how Pd is less abundant than Ru when the table says Pd is much more abundant than Ru, as their atomic masses don't differ that much. --Nucleus hydro elemon (talk) 05:51, 23 March 2024 (UTC)
 * @Double sharp: Are any of these data worth preserving in one of the element data pages? YBG (talk) 06:02, 23 March 2024 (UTC)
 * I don't think so, since they don't appear to go back to RS. Double sharp (talk) 10:37, 23 March 2024 (UTC)
 * Ok. Makes sense. YBG (talk) 13:08, 23 March 2024 (UTC)
 * ✅ Johnjbarton (talk) 22:27, 23 March 2024 (UTC)