You can also calculate it like this..

[(207Pb/235U age)/(206Pb/238U age)-1] x 100 (%).

Another approach (maybe one of the most common) is (206Pb/238U age /207Pb/206Pb age)*100. There are no consensus about how to report U-Pb data, but some groups are discussing this. Check https://sites.google.com/a/laserchron.org/laserchron/home/ and https://docs.google.com/a/laserchron.org/leaf?id=0B9ezu34P5h8eZmE3MTAxYzMtYzJjNy00Y2YzLWEyZDQtNTgyNmY5YTRkMDQ3&hl=en_US

In my opinion, using the difference between the 7/6 and 6/38 is better because, typically (at least by LA-ICPMS and SIMS), these are the ratios that are actually measured (i.e. the 7/35 is calculated from the 7/6, 6/38 and a natural U isotopic ratio). In addition, because the 7/35 is usually the least precise and accurate date. 207Pb is the least abundant isotope in zircon and is therefore is the most affected by even minor common lead so calculating discordance using the 7/35 results in fewer 'discordant' analyses being filtered.

thanks for the three answers.

I agree with John on the argument, but 207Pb/206Pb ages are also affected by the low concentration of 207Pb in young zircons. In my case, some young grains have "huge" degrees of discordance, though the error ellipse (which is controlled by the  206Pb/238U and 207Pb/235U ages) lies on the concordia. Besides, fot young zircons, the 206Pb/207Pb ages have often very large errors. Hence I am a bit confused and unsure what to use (my zircons are between 80 Ma and 3.2 Ga old.

Felipe, I cannot find the discussion in the links you gave, could you give me another hint where to look? thanks

When calculate the discordance, we need to select one age that thought to be correct, e.g., 6/8 ages for young zircon while 7/6 age for old (>1000Ma), and then calculate the degree of deviation of other ages. I think for young zircons (

Dear Ulrich, try the link at the end of this message. I do agree with John that 68 and 76 ratios are the most relevant because they are actually measured. For this reason, I should have made myself clear and highlighted exactly what you said about low Pb concentration grains, which are common in young rocks.

http://www.plasmage.org/recommendations

thanks Felipe, there is also an interesting link on that web-site:

http://www.plasmage.org/images/documents/Prague%20workshop%202011_08/Nemchin_-_Data_filtering_and_visualisation_in_U-Pb_geochronology_of_DZ_populations.pdf

Great, it seems to be very interesting. Look, if you are also interested in papers regarding the data presentation, especially for detrital zircon, you should read the papers from Vermeesch (check the link below).

https://www.ucl.ac.uk/~ucfbpve/

We use (206/238) / (207/206) for 'older' ages, and (206/238 / 207/235) for younger ages (i.e. where the 207/206 can become very imprecise). Just don't ask where the cut-off is!!

This is much in the same way that a detrital age spectra may switch from 206/238 ages to 207/206 ages at a certain age. 

Hey Nick, about the cut-off (rsrsrs), have you and your colleagues had this discussion? This is in our mind here for a long time.  

Apart from the methods proposed by Vermeesch' group, there are also somewhat more intuitively obvious ways of presenting detrital data that also take the uncertainty into account, as proposed in the attached publication. We found that the cross-over for the uncertainties from 6/38 to 7/6 ages was somewhere between 900 and 1000 Ma, but this is likely to depend on the equipment and method used to obtain the ages.

Nick, I think that is a very good proposal and I agree with Marlina that the cross-over of the procedures is at about 1 Ga.  Actually, in the Philippines, we I do most work, we have very few zircons in the "critical" range, most are younger than 400 Ma or older than 1.6 Ga. 

Also, thanks for sharing the paper, Marlina :-)

Dear Mr. Saha,

do you have a paper or something else, that i can use as citation for that formula of yours?

Best Regards

 I apologise for a late contribution to this debate, which I have only just seen. There are two separate arguments going on here - one on the circumstances under which we might choose the 206/238 age in preference to the 207/206 age (which is complex and depends on factors such as U content, common Pb correction, measurement precision) and one on how to calculate discordance. Returning to Ulrich Knittel's original question, I would say it depends on how you plan to use the data. Logically Ulrich was on the right lines - discordance between the two decay schemes, on first principles, would be the difference between 206/238 and 207/235 ages, and if you are basing your calculations on the Wetherill plot (either intersection ages or Ludwig 'Concordia' ages, that is surely the way to judge discordance). Subhojit Saha gives the simpler form of calculation. Some authors are still reporting % concordance figures rather than %discordance - = (100-%discordance).

If, as in most SIMS datasets, you are just choosing between 206/238 and 207/206 ages, then [1 - (206/238 age)/(207/206 age)] x 100% is a reasonable way of looking at the discordance to filter the data, as several of you have suggested. This still give positive discordance for points that would fall below concordia in the Wetherill plot.

But, as so often, the best plan is to state how you have calculated parameters in your published data tables!

Dear Robert, the formula you stated,  [1 - (206/238 age)/(207/206 age)] x 100%

would you please provide a reference for that so that I can cite it.

My zircons are between 60-70 Ma? Should i used (207/235 age)/(206/238 age) * 100 % or (206/238 age)/(207/206 age) *100 % for calculating the degree of discordance?

The problem with 7/6 ages in grains so young is that even a small amount of common lead will result in huge discordances. So I suggest you to calculate the discordance with 6/8 and 7/5 ages.

Thank you Felipe for suggesting and answering.

I recommend Nongmaithem's answer

Best Regards

Thanks to Ulrich Knittel for asking the question and the rest of you who contributed to give answers. I find the replies very useful.

Is there any reference for that eq. (1-(206Pb/238U age /207Pb/206Pb age)*100 Ulrich Knittel

I used method described by John Cottle in my work

Can u mention the paper Abdulgafar K Amuda

Osama K. Dessouky it's yet to be published.

@Osama: I used Hidaka, Shimizu, Adachi. "U–Pb geochronology and REE geochemistry of zircons from Palaeoproterozoic paragneiss clasts in the Mesozoic Kamiaso conglomerate, central Japan: evidence for an Archean provenance." Chemical Geology 187, (2002): 279-293 as source, these authors cite Biao 1996 as source but I could not obtain a copy of that paper

Ulrich Knittel Thanx, I sent you message.

https://doi.org/10.5194/gchron-3-247-2021

Thank you Yizhou, that is a very useful paper! Personally, I like to check whether or not the error ellipses of the analyses intersect the concordia - though I thus include some inprecise measurements.

Article On the treatment of discordant detrital zircon U–Pb data

On that topic.

In our tool for filtering dZr data and computing "best ages" we use several options:

(1) fixed limit (Ma): if 6/8 age < limit, then discordance is calculated between 75 and 68, else between 76 and 68.

(2) between 75 and 68

(3) between 76 and 68

(4) the lesser of two, in percent of the age.

Also, two options on how to discard discordant zircons: either by setting a discordant filter (e.g., +20%, -10%), or filtering out the analyzes that miss concordia by x sigmas.

Opinions welcome.

Article A review of detrital zircon data treatment, and launch of a ...

Thank you for posting this Monika, this is a useful paper by the "pope" of geostatistics, I had not yet found it

Thank you Vladislav for bringing this paper to our attention!

The 207Pb/206Pb age is used when it is older than 1000 Ma, while the 206Pb/238U age is used when it is younger than 1000 Ma (or 1200 Ma).