Pb-Pb 235U- 207Pb 238U- 206Pb 232Th-208Pb All Methods Score = 0 56 55 67 43 91 0 < Score < 0.50 40 47 52 13 316 0.50 ≤ Score < 1 35 37 47 14 305 Score = 1 363 347 324 80 286 Total Count 494 486 490 150 998 Average Score 0.80 0.80 0.76 0.62 0.57 % Concordant (Score = 1) 73% 71% 66% 53% 29% Table 5. The distribution of concordance scores for each U-Th-Pb method from the “U-Th-Pb Comparisons” dataset. Also includes the average score and the percentage of concordant records. Figure 6. The distribution of concordance scores for all 998 records in the National Geochronological Database with ages determined using three or more U-Th-Pb methods. This is the “U-Th-Pb Comparisons” dataset. One of the most striking features of our analysis of the USGS National Geochronological Database is how few of the 18,575 records included multiple age determinations using multiple methods. Only 4,875 (26.2%) of the records included two or more age determinations, only 998 (5.4%) included ages calculated using at least three of the four U-Th-Pb methods, and only 34 (0.2%) included age determinations using three or more methods where at least two were not U-Th-Pb. Of the 4,875 records with two or more age determinations, just over half (53.2%) had a concordance score of 1.00, meaning that all the age ranges overlapped, and 23.3% of the records had a concordance score of 0.00, meaning that none of the age ranges overlapped. Comparisons between two of the U-Th-Pb methods had concordance scores averaging 0.57, while comparisons between three of the methods averaged 0.64 and four of the methods averaged 0.46. Moreover, of the 34 records associated with three or more age determinations where at least two were not U-Th-Pb, only one (2.9%) had a concordance score of 1.00. Given how few records in this large database included multiple age determinations, there are clearly limits to what we can say with confidence about the prevalence of concordance. However, taken together, our results suggest that records with age determinations from more methods tend to have lower concordance scores, and this should at least temper the claims sometimes made in the literature concerning “the vast amount of concordance” between radioisotope age determinations (e.g. Isaac 2007, p. 144). One possible caveat concerns the way in which we treated each record as though it represented a separate rock unit, even though some rock units are represented in the database by multiple records and thus may have been dated by more radioisotope methods than is at first apparent. For instance, the Pikes Peak Granite of the central Front Range of Colorado is represented in the database by about 20 separate records with 90 age determinations between them. We are not certain what effect, if any, collating records that refer to the same rock units will have on our results, but this would seem to be an obvious next step. However, given that concordance scores tended to be lower when rock units were dated by multiple methods, we might expect that collating more age determinations for individual rock units would result in lower overall concordance scores. Some interesting patterns emerge from our analysis. In cases where a single radioisotope method is used multiple times on a single rock unit, there is often significant “internal discordance” within the results from that one method. In our study, average “internal concordBEACHY, KINARD AND GARNER How often do radioisotope ages agree? 2023 ICC 395
RkJQdWJsaXNoZXIy MTM4ODY=