This article is edited and drawn from:
Davies, B.J., 2022. Dating Glacial Landforms II: Radiometric Techniques, in: Haritashya, U. (Ed.), Treatise in Geomorphology (Second edition). Cryospheric Geomorphology. Elsevier, pp. 249-280. (link)
Introduction to Cosmogenic Nuclide Sampling Strategies
Sampling strategies for exposure-age dating will depend upon the research objectives, the rock type and target minerals available and resultant choice of isotope, and the local environment.
Researchers should in all cases aim to collect multiple replicate samples where possible, especially in the case where a particular landform (such as a moraine) is to be dated.
A fundamental knowledge of the principles of the six cosmogenic nuclides is required in order to design such a sampling strategy. Fundamentally, the sample should contain sufficient target mineral to obtain measurable quantities of the required nuclide.
Sample context
Careful geomorphological mapping is critical to understand the depositional processes affecting boulders and any post-depositional modification (e.g. Kelley et al., 2014; Koffman et al., 2017; Schaefer et al., 2009).
For example, it is important to understand the relationship between moraines, boulders, slope processes and any ice-dammed palaeolakes, as boulders deposited below lake water will have an exposure age that records the timing of lake-level fall, rather than deglaciation (Davies et al., 2020, 2018; Hein et al., 2010; Thorndycraft et al., 2019).
Signs of glacial transport
Boulders should only be sampled if the operator can be sure that they were glacially transported, and not deposited by rockfall or other geomorphic processes. Both bedrock and glacially transported boulders should show signs of glacial transport, abrasion, and erosion (faceting, striations, polish, edge-rounding, of a erratic lithology), so that any inherited nuclides have ideally been removed.
Replicate samples
Multiple boulders should be sampled per moraine, as geological scatter is a common issue in exposure-age dating. This is because moraines degrade over time, and boulders can be weathered away, exhumed, buried, or destroyed by erosion. They may have an inheritance, with cosmogenic isotopes dating from a prior exposure. This can lead to geological scatter (Applegate et al., 2012; Heyman et al., 2016, 2011).
Sampling 3 to 5 boulders for moraines dating to the Last Glacial Maximum or younger can help with identification of outliers (Putkonen and Swanson, 2003). Larger boulders more than 0.5 m above the ground height should be targeted in order to minimize the risk of exhumation and post-depositional processes causing geological scatter (cf. Heyman et al., 2016).
Boulders should only be sampled where they are in a stable position in the landscape or on the moraine and there is no possibility that they have rolled, rotated, or otherwise moved significantly since deposition (Figure 7). For moraines, ideally boulders should be situated on the moraine crest.
Sampling cobbles
Cobbles may be sampled where they lie on flat bedrock, and the possibility of cycling through the active layer as a result of periglacial processes can be excluded. Cobbles should be marked with the uppermost surface marked. If they are less than ca 5 cm thick, then attenuation through the sample and self-shielding can be effectively disregarded.
Cobbles should however be large enough that they are not moved due to strong winds. They should be sampled from exposed locations where snow cover is likely to be thin or inconsequential.
Cobbles should weigh ca 1 kg minimum, so as to increase the likelihood of the required amount of the nuclide. They should show signs of glacial transport (abrasion, striations, faceting) and ideally be of an erratic lithology, so that local production of the cobbles can be excluded.
Internal shielding
The ideal boulder, rock, landform surface or bedrock surface for sampling should be sufficiently extensive, flat, and horizontal. The angle of the surface will affect the shielding and so should be recorded. Flat surfaces require fewer corrections for shielding, and so have a greater precision and accuracy.
Samples should be collected at least 50 cm away from any edges (Gosse and Phillips, 2001). Edges are susceptible to ‘edge effects’, particularly for nuclides such as 36Cl that are produced by muon capture, as cosmic rays could penetrate from multiple directions.
Samples should ideally be taken from the flattest, central, uppermost surface of the boulder, but this can in reality be challenging unless a rock saw or small explosive charge is used. Shielding from snow cover and vegetation can be minimized by choosing boulders that are large and upstanding and above the local topography (e.g. Figure 8A, B, F), as they are more likely to be windswept (Gosse and Phillips, 2001).
Sample description
A thorough sample description is required in order to calculate shielding and production rate scaling. A hypothetical proforma to assist with this is presented in Figure 7, which follows best-practice guidelines (Darvill, 2013; Gosse and Phillips, 2001).
Samples should be sketched, photographed from all angles, and details of location (decimal degrees), elevation (m asl) and geomorphic context should be very carefully recorded.
The boulder’s characteristics should be carefully described, including dimensions and height above ground surface, signs and measurements of weathering or erosion (upstanding quartz veins, weathering pits, flaking), lithology, grain size and quartz content. Sample thickness should be recorded as well.
Calculating topographic shielding
In order to calculate the shielding, the dip and dip direction of the surface and the angle of elevation to the skyline should be recorded at regular intervals. This can be checked with a digital elevation model (DEM) in a geographic information system (GIS) if the horizon is not always visible (Codilean, 2006).
Sample collection: minimise geological vandalism
Samples should be collected with the aim of leaving as little permanent scarring as possible on the landscape. Sampling with a hammer and rock chisel or small charge can be best, as these methods take small flakes and leave little permanent visual impact.
Sampling with a rock saw can be easier and quicker, and allows the operator to choose more precisely where to sample, but if this method is used, some time should be spent afterwards to obscure and roughen the straight cuts.
Samples should only be taken with the permission of the landowner, and permits may be required in some localities.
Ideally, samples should be stored in a durable cloth bag, clearly marked with permanent marker pen.
The volume of sample required depends on the proportion of the target mineral, grain size of the target mineral, and required isotope. Around 1 mg of 10Be/9Be is required for AMS analysis. For 10Be dating of quartz-rich rocks (such as a typical granite, with 10% quartz), a minimum of 500 g but ideally around 1 kg of sample should be obtained. Samples from rocks with a younger exposure age should be larger, to obtain the required level of precision (Gosse and Phillips, 2001). In sample processing, material could be lost due to accidental chemistry, and offcuts may be required for thin section or duplicate chemistry. Larger samples will be needed in rocks with a lower quartz content. 5 kg of rock could be required, for example, for the analysis of fine-grained quartz-poor rocks (ibid).
Whole-rock analyses of 36Cl do not require mineral separates, and so smaller samples of 500 g may be appropriate.