Assigning an age in million years before present (Ma) to maximum flooding surfaces (MFS, or any other isochronous sequence stratigraphic surface) can be a misleading practice unless the level of accuracy is adequately quantified.

Nonetheless, many authors (e.g. Haq et al., 1988; Sharland et al., 2001, 2004; Haq and Al-Qahtani, 2005; Simmons et al., 2007) have assigned precise ages to MFS and sequence boundaries using different vintages of the geological time scale (GTS). This practice remains useful because it provides very approximate constraints on rates of deposition, durations of sequences and their orders, and hiatuses. This note compares how age revisions in GTS affect the numerical age assignment of Arabian Plate MFS.

Sharland et al. (2001, 2004, Table 1) published the biozones, their corresponding stages and estimated age values of 65 Arabian Plate MFS based on the GTS 1996 (Gradstein and Ogg, 1996). These authors also correlated the ages of their Cenozoic and Mesozoic MFS to those of Haq et al. (1988). Haq and Al-Qahtani (2005) adopted the 63 Phanerozoic MFS of Sharland et al. (2001, 2004), but did not consider Neoproterozoic Pc10 and Pc20. They positioned the remaining 63 MFS (with minor modifications) in the stratigraphic column of Saudi Arabia, and revised their ages based on GTS 2004 (Gradstein et al., 2004; see GTS 2004 in GeoArabia, 2007, v. 12, no. 1, p. 208-209).

GTS 2004 has been ratified by the International Commission on Stratigraphy (ICS) and includes stage names that are recommended for global correlations (Gradstein et al., 2004; Gradstein and Finney, 2007). The estimated ages of the stage boundaries have been compared by the ICS across various vintages of the Geological Time Scale, starting with the GTS 1937 of Holmes and ending with GTS 2004. These comparisons were reprinted by permission of the ICS inAl-Husseini (2005, see Cenozoic Comparison Chart on p. 140; Mesozoic Comparison Chart on p. 142; and Paleozoic Comparison Chart on p. 144).

Most recently Simmons et al. (2007) revised the age estimates of the 63 Phanerozoic MFS of Sharland et al. (2001, 2004) using GTS 2004. Changes in the age estimate for an MFS mostly occur because the bounding ages of stages were revised due to better radiometric and other age-dating techniques. The dating of an MFS may also change because either (1) the reference section pick was revised (i.e. it is placed in a different body of rock belonging to a different biozone), or (2) the biostratigraphy of the reference section was revised by, for example, the discovery of new fossils offering a more correct biozonal assignment.

In Table 1 the 63 Phanerozoic MFS age estimates (Sharland et al., 2001, 2004) are compared to those of Haq et al. (1988), Haq and Al-Qahtani (2005) and Simmons et al. (2007). The comparison does not account for changes due to revised positions or biozones (stage) of the MFS, and mostly reflects GTS revisions. Column (1) of Table 2 compares the age differences between Haq et al. (1988) and Haq and Al-Qahtani (2005) for the 38 Mesozoic-Cenozoic MFS that were identified in Arabia. Of these 38 MFS, 13 have ages that shifted by less than 3.0 My, 11 shifted by 3.0–6.0 My and 14 shifted by more than 6.0 My.

Shifts in age of 3.0 My or more are significant because the duration of a third-order sequence is generally estimated to be 2.0, 2.4 or 2.8 My (Matthews and Frohlich, 2002). Age shifts of more than 3.0 My are substantial and reflect the inherent inaccuracy of GTSs from vintage-to-vintage.

Column (2) of Table 2 compares the age differences between Sharland et al. (2001, 2004) based on GTS 1996 and Simmons et al. (2007) based on GTS 2004 for the 63 Phanerozoic Arabian MFS. Nearly half (27) have changed ages by one-or-less million years (My), approximately one quarter (15) MFS shifted in age by 1.0–3.0 My. One-third (21) of the MFS shifted in age by more than 3.0 My. In some cases, several MFS were closely spaced in Sharland et al. (2001, 2004) and the large shifts imply a different order. For example, the Permian MFS P20, P30 and P40 (Khuff sequences) only spanned 3.5 My in Sharland et al. (2001), suggesting fourth-order MFS; whereas in Simmons et al. (2007) they span 13.0 My, thus implying they are third-order MFS.

The estimated ages of MFS can also differ when different authors interpret the same sequences using the same GTS. The magnitude, in this case, is much less but still significant. This can be illustrated when comparing the age estimates of Haq and Al-Qahtani (2005) to those of Simmons et al. (2007); both are based on Arabia’s sequence stratigraphy and GTS 2004 (Table 2, Column 3). In this comparison, 45 MFS have ages that differ by one or less million years. About a quarter (14) differ by 1.0–3.0 My and two differ by more than 3.0 My.

Any apparent convergence in estimated ages between different authors using the same GTS and/or sequences is not, however, an indication of chronostratigraphic accuracy. It reflects the level of precision in estimating age values using the same data. Therefore it is important not to refer to an MFS by its estimated age value alone. This is because an MFS (or isochronous surface), as defined in a particular study, has a well-defined position in the rock column and/or biozone. The absolute age value, however, should carry an error bar that reflects the age extent of the biozone in which the MFS occurs and the uncertainty of absolute age values for that interval of geological time.

Applying numerical error analysis to geological time may not be straightforward. Consider, for example, the cited one standard deviation for the age of the Aptian/Barremian boundary. It was recalibrated from 121.0 ± 1.4 Ma in GTS 1996 to 125.0 ± 1.0 Ma in GTS 2004; but a shift of 4.0 My (121.0 to 125.0 Ma) exceeds the standard deviation cited for both estimates (1.0 and 1.4 My).

The author thanks Joerg Mattner and Michael Simmons for their useful comments.