Improving uncertainty estimates for radiogenic isotopic dates used in calibrating the geologic timescale
|Location||International Geological Congress,oslo 2008|
|Author||Villeneuve, Mike۱; Schmitz, Mark۲|
|Holding Date||08 September 2008|
The modern timescales represent the melding of an absolute-time chronostratigraphic scale with a relative-time biostratigraphic scale. Because a chronostratigraphic scale is a measured entity, it inherently contains a level of uncertainty associated with these measurements and GTS2004 marked the first rudimentary attempt to incorporate these into the construction of the timescale. Although trecognizing that the radiogenic isotopic-based geochronology that calibrates parts of the timescale does a thorough job of propagating analytical errors of a random nature, some external and systematic uncertainties are not incorporated in the ones published with the date. In order to account for this effect,, minimum uncertainty levels were applied to the GTS2004 data to prevent overweighting of analyses due to statistical anomalies that resulted in unwarranted precision. However, no attempt was made to assess the overall validity of the uncertainty assigned to the data points. As such, data that were statistically equivalent, but quite disparate in their robustness, carried the same weight within the chronostratigraphic regressions. For example, a discordant, but collinear array of 3 zircon fractions could quote superior analytical uncertainties relative to a multi-fraction cluster of zircons that overlap on concordia. Clearly, the latter case represents superior adherence to closed-system behaviour and should carry more weight than the former. Although semi-quantitative in their application, proposals for data-point weighting, based upon data reproducibility and analytical methodology are assessed, in order to improve upon the analysis presented in GTS2004. In general, 206Pb/238U dates are taken as the baseline ages. Those dates based upon multiple-overlapping, concordant, U-Pb fractions are weighted highest, and their allowed uncertainty can approach stated analytical uncertainties. Degraded uncertainties are associated with increasing evidence of non-reproducibility, discordance and finally those brought about by methodological limitations. While 40Ar-39Ar dates do not achieve the precision attainable by U-Pb dates, the new protocols incorporate the numerous studies conducted since GTS2004, thereby providing an enhanced intercalibration of the two isotopic systems. In particular, the conjoined basic parameters of apparent monitor age and 40K decay constant are newly incorporated in the basic dataset. This allows for the accurate use of undisrupted 40Ar-39Ar dates where no usable U-Pb data exists.