Geochronology, 1, 1–16, 2019 https://doi.org/10.5194/gchron-1-1-2019 © Author(s) 2019. This work is distributed under the Creative Commons Attribution 4.0 License. Chlorine-36=beryllium-10 burial dating of alluvial fan sediments associated with the Mission Creek strand of the San Andreas Fault system, California, USA Greg Balco1, Kimberly Blisniuk2, and Alan Hidy3 1Berkeley Geochronology Center, 2455 Ridge Road, Berkeley, CA 94550, USA 2Department of Geology, San Jose State University, San Jose, CA, USA 3Center for Accelerator Mass Spectrometry, Lawrence Livermore National Laboratory, Livermore, CA, USA Correspondence: Greg Balco ([email protected]) Received: 23 April 2019 – Discussion started: 30 April 2019 Revised: 3 July 2019 – Accepted: 5 July 2019 – Published: 18 July 2019 Abstract. We apply cosmogenic-nuclide burial dating using with stratigraphic age constraints as well as independent es- the 36Cl-in-K-feldspar=10Be-in-quartz pair in fluvially trans- timates of long-term fault slip rates, and it highlights the po- ported granitoid clasts to determine the age of alluvial sedi- tential usefulness of the 36Cl=10Be pair for dating Upper and ment displaced by the Mission Creek strand of the San An- Middle Pleistocene clastic sediments. dreas Fault in southern California. Because the half-lives of 36Cl and 10Be are more different than those of the commonly used 26Al=10Be pair, 36Cl=10Be burial dating should be ap- 36 10 plicable to sediments in the range ca. 0.2–0.5 Ma, which is 1 Introduction: Cl= Be burial dating too young to be accurately dated with the 26Al=10Be pair, and should be more precise for Middle and Late Pleistocene sed- In this paper we apply the method of cosmogenic-nuclide 36 iments in general. However, using the 36Cl=10Be pair is more burial dating (henceforth, “burial dating”) using the Cl-in- 10 36 10 complex because the 36Cl=10Be production ratio varies with K-feldspar= Be-in-quartz ( Cl= Be) nuclide pair to deter- the chemical composition of each sample. We use 36Cl=10Be mine the age of buried alluvial sediment exposed in a sed- measurements in samples of granodiorite exposed at the sur- imentary section that has been offset along a strand of the face at present to validate calculations of the 36Cl=10Be pro- San Andreas Fault in southern California. The concept of duction ratio in this lithology, and then we apply this infor- cosmogenic-nuclide burial dating relies on a pair of cosmic- mation to determine the burial age of alluvial clasts of the ray-produced nuclides that (i) are produced at a fixed ratio same lithology. This particular field area presents the addi- by cosmic-ray bombardment of surface rocks and minerals, tional obstacle to burial dating (which is not specific to the but (ii) have different decay constants. When sediment is ex- 36Cl=10Be pair, but would apply to any) that most buried al- posed to the surface cosmic-ray flux during erosion in its ini- luvial clasts are derived from extremely rapidly eroding parts tial source area and fluvial transport, concentrations of both of the San Bernardino Mountains and have correspondingly nuclides increase over time, but their concentration ratio con- extremely low nuclide concentrations, the majority of which forms to the production ratio. If the sediment is then shielded most likely derive from nucleogenic (for 36Cl) and post- from the surface cosmic-ray flux by burial in a sedimentary burial production. Although this precludes accurate burial deposit, nuclide production halts, and nuclide concentrations dating of many clasts, data from surface and subsurface sam- decrease due to radioactive decay. If the two nuclides have ples with higher nuclide concentrations, originating from different decay constants, the ratio of their concentrations lower-erosion-rate source areas, show that the age of upper will change over time. Accordingly, the ratio of their con- Cabezon Formation alluvium is 260 ka. This is consistent centrations can be used to estimate the duration of burial and hence the depositional age of the sediment. Burial dating has Published by Copernicus Publications on behalf of the European Geosciences Union. 2 G. Balco et al.: Chlorine-36=beryllium-10 burial dating been widely applied to date clastic sediments in caves, river only from analytical uncertainty on the measured nuclide ra- terraces, and sedimentary basins (see summaries in Dunai, tio, then the relative uncertainty in the burial age σ t=t (nondi- 2010; Granger, 2006), and it is potentially uniquely useful mensional) is for Quaternary clastic sediments that are difficult to date by σ t σ R other means. In contrast to many other dating techniques that D N ; (3) t λ R t require either fossil material or the formation of new miner- R N als at the time of sediment emplacement, it can be applied to where σ RN is the uncertainty in the normalized ratio and is any sediment that has experienced a period of surface expo- equal to σ RM=RP, and σ RM is the uncertainty in the mea- sure followed by burial. sured nuclide ratio. Thus, the uncertainty in the burial age Nearly all existing applications of burial dating have used is inversely proportional to the decay constant of the ratio 26 10 the Al= Be nuclide pair. These nuclides are produced in λR. Given equivalent precision on the ratio measurements, quartz at a ratio of 26Al=10Be D 6:75, and the half-lives of selecting a nuclide pair with a larger difference between de- 26 10 Al and Be are 0.7 and 1.4 Ma respectively, so the half-life cay constants, and thus a larger λR, results in a more pre- of the 26Al=10Be ratio is 1.43 Ma. The half-life of the ratio cise burial age. Chlorine-36 has a half-life of 300 000 years, 1 36 10 is defined as −ln /λR, where λR D λA − λB and λA and so the difference between the Cl and Be decay con- 2 D × −6 −1 stants (λR 1:8 10 , corresponding to a half-life of the λB are the decay constants (yr ) of the shorter- and longer- 26 26 10 ratio of 0.38 Ma) is substantially larger than for Al and lived nuclides, respectively. The Al= Be nuclide pair has − 10Be (λ D 4:8 × 10 7). As long as 36Cl production is pre- three important advantages for burial dating. First, both nu- R dominantly by spallation (see discussion below), uncertain- clides are produced within the same mineral, which ensures ties in 26Al, 10Be, and 36Cl production rates and decay con- that concentrations of both nuclides reflect the same exposure stants are similar, so a 36Cl=10Be burial age should be more history. Second, quartz is ubiquitous in clastic sedimentary precise than a 26Al=10Be burial age, as long as the burial du- deposits; third, the mineral chemical composition, and there- ration is not so long that the 36Cl concentration has decayed fore the 26Al=10Be production ratio, is invariant for quartz. to a level that cannot be precisely measured. However, the relatively long half-life of the 26Al=10Be ratio Figure1 shows the results of a more complex uncertainty limits the useful age range of this nuclide pair: it is best suited analysis (Balco and Shuster, 2009) that also includes a model to burial dating of sediments in the age range ∼ 0:7–4 Ma, for measurement uncertainty as a function of nuclide concen- and it is not well suited for Middle and Late Pleistocene sed- tration. For the majority of the Pleistocene, 36Cl=10Be burial iments younger than ∼ 0:5 Ma (Fig.1). In principle, one can ages are theoretically expected to be more precise than overcome this limitation by using a nuclide pair with a larger 26Al=10Be burial ages, and the 36Cl=10Be pair should pro- difference between half-lives. If the half-life of the nuclide vide useful (e.g., 20 % or better) precision for burial ages as ratio is shorter, the ratio in a buried sample diverges from the young as 0.15–0.2 Ma. production ratio faster, and, assuming that other uncertain- On the other hand, the 36Cl=10Be pair has several disad- ties in production ratios and concentration measurements are vantages in comparison to the 26Al=10Be pair. The primary equivalent, a burial age estimate based on the measured ratio disadvantage is that production of cosmogenic 36Cl does not is more precise at younger ages. We show this mathemati- occur in pure quartz (SiO ). Although in natural quartz there cally by simplifying the complete equations for burial dat- 2 is sometimes nonzero 36Cl production from impurities or ing (Granger, 2006) by assuming that a sample experiences a from Cl-bearing fluid inclusions (Bierman et al., 1995), the two-stage burial history consisting of (i) a period of surface production rate is very low and usually cannot be accurately exposure that is short enough that radioactive decay during estimated. Likewise, although 10Be is produced in other min- the exposure period can be disregarded, followed by (ii) in- erals that also have significant 36Cl production (e.g., pyrox- stantaneous burial at a depth large enough that post-burial ene; see Blard et al., 2008; Collins, 2015; Eaves et al., 2018), nuclide production is negligible. With these assumptions, 10Be extraction methods and production rates in these min- RM erals are not well established. Thus, measurement of both RN D D exp(−λRt); (1) 10Be and 36Cl in the same mineral is generally not feasi- RP ble. This means that 36Cl=10Be burial dating is only possible where RN (nondimensional) is the measured nuclide ratio in when (i) quartz occurs with another mineral that is suitable the sample RM normalized to the production ratio RP, λR is for 36Cl measurement, and (ii) both minerals have experi- −1 the decay constant of the ratio as defined above (yr ), and t enced the same exposure history.