Surface exposure dating with cosmogenic nuclides

In the last decades surface exposure dating using cosmogenic nuclides has emerged as a powerful tool in Quaternary geochronology and landscape evolution studies. Cosmogenic nuclides are produced in rocks and sediment due to reactions induced by cosmic rays. Landforms ranging in age from a few hundred years to tens of millions of years can be dated (depending on rock or landform weathering rates) by measuring nuclide concentrations. In this paper the history and theory of surface exposure dating are reviewed followed by an extensive outline of the fields of application of the method. Sampling strategies as well as information on individual nuclides are discussed in detail. The power of cosmogenic nuclide methods lies in the number of nuclides available (the radionuclides 10Be, 14C, 26Al, and 36Cl and the stable noble gases 3He and 21Ne), which allows almost every mineral and hence almost every lithology to be analyzed. As a result focus can shift to the geomorphic questions. It is important that obtained exposure ages are carefully scrutinized in the framework of detailed field studies, including local terrace or moraine stratigraphy and regional morphostratigraphic relationships; as well as in light of independent age constraints. [Oberflächenexpositionsdatierungen mittels kosmogener Nukliden] Kurzfassung: Im letzten Jahrzehnt hat sich die Methode der Oberflächendatierung mittels kosmogener Nuklide zu einer leistungsfähigen Methode in der Quartärchronologie und quantitativen Landschaftsanalyse entwickelt. Kosmogene Nuklide werden durch kosmische Strahlung in Festund Lockergestein gebildet. Die Konzentrationen der kosmogenen Nuklide kann mittels Massenspektrometrie ermittelt werden. Dies ermöglicht je nach Verwitterungssrate die Datierung von Landschaftselementen und Landschaftsformen mit Altern zwischen einigen 100 Jahren bis über 10 Millionen Jahren. Neben einem Abriss der historischen Entwicklung und Theorie der Oberflächendatierung mittels kosmogener Nuklide enthält dieser Artikel eine ausführliche Übersicht der zahlreichen Anwendungsgebiete dieser Methode. Probenahmestrategien und die Eigenheiten der einzelnen Nuklide werden im Detail besprochen. Die Vielzahl der mit dieser Methode in den verschiedensten Mineralien bestimmbaren Nuklide (Radionuklide 10Be, 14C, 26Al und 36Cl und Edelgase 3He und 21Ne) erlaubt die Beprobung und Analyse verschiedenster Lithologien. Der erreichte hohe Entwicklungsstand der Methode erlaubt es den Fokus auf die eigentlichen geomorphologischen Fragestellungen zu legen. Die Sensitivität der kosmogenen Oberflächendatierungsmethode muss trotzdem sorgfältig im Rahmen ausführlicher Feldstudien erfolgen, wie zum Beispiel durch die Analyse von lokalen und regionalen Terrassenoder Moränenstratigraphien oder durch den Vergleich mit anderen Datierungsmethoden.


Introduction and history
Cosmogenic nuclides build-up predictably with time in minerals exposed to cosmic rays.Therefore measuring their concentrations allows determination of how long rocks or sediment have been exposed at or near the surface of the Earth (LAL 1991;GOSSE & PHILLIPS 2001).At present the most commonly utilized nuclides are the radionuclides 10 Be, 14 C, 26  Al,  and 36 Cl and the stable noble gases 3 He and 21 Ne (Table 1).Because of the wide variety of nuclides available (with different half-lives or stable) and the fact that they can be measured in a variety of minerals a broad spectrum of geomorphological problems can be addressed (Fig. 1).By measuring the concentrations of cosmogenic nuclides rock surfaces themselves can be directly dated.This is a unique and powerful tool never before available to geomorphologists.
Bedrock landforms, fluvially-or glacially-polished bedrock surfaces, fault footwall faces, and landslide bedrock detachment surfaces can be sampled and dated directly.There is no other method where this is possible.Sedimentary units such as moraines, landslide deposits, fluvial terraces, debris flows or alluvial fans can be directly dated by sampling boulder surfaces or by taking samples made up of numerous clasts.Sites unsuitable for luminescence techniques may be dated with cosmogenic nuclide methods, for example sediments that have not been exposed to light long enough or coarsegrained material.In the past geomorphology has relied on radiocarbon for the indirect dating of landforms.The upper age limit for radiocarbon dating of organic material is about 50 ka, whereas under certain conditions landform ages on the order of tens of millions years can be measured with exposure dating (SCHÄFER et al. 1999;DUNAI et al. 2005).For radiocarbon showing the various landforms that can be dated and approaches for using cosmogenic nuclides to address questions of timing and rates of landscape change (see also BIERMAN & NICHOLS 2004).

SUSAN IVY-OCHS & FLORIAN KOBER
dating organic material must be present in the sediment, which is not the case in most high alpine early Holocene or Lateglacial deposits.Finally organic material included in sediments is not dating the landform itself, the relationship of the organic material to the landform can never be unequivocally established.Entrained wood fragments are simply older than the enclosing sediment.Because of the presence of higher concentrations (production rates are higher in space than they are on earth) cosmogenic nuclides were initially investigated in lunar and meteorite samples already beginning in the 1950's (FINKEL & SUTER 1993 and references therein).
Earliest attempts to measure cosmogenic nuclides in terrestrial rocks were made by DAVIS & SCHAEFFER (1955) and SRINIVASAN (1976).
DAVIS & SCHAEFFER (1955) used low-level decay counting to measure 36 Cl in a high-Cl phonolite from unglaciated high elevation sites in the Rocky Mountains.In 1976, Srinivasan analyzed the cosmogenic noble gas 126 Xe in barite from a sedimentary unit (SRINIVASAN 1976) and highlighted the potential of noble gases in surface exposure dating (NIEDERMANN 2002).Routine measurement of cosmogenic nuclides and use of cosmogenic nuclides for determi-nation of exposure histories and erosion rates only became possible after the development of accelerator mass spectrometry (AMS), and the construction of high sensitivity noble gas mass spectrometers between 1970 and 1980.These technical developments opened the way to measurements of exceedingly low nuclide concentrations.In the late 1970s first accelerator measurements were reported for 14 C (BENNETT et al. 1977;NELSON et al. 1977), 10 Be (RAISBECK et al. 1978), 26 Al (RAISBECK et al. 1979) and 36 Cl (ELMORE et al. 1979).Early measurements of cosmogenic nuclides in rock and sediment samples were made using 3 He (KURZ 1986), 10 Be and 26  Al (KLEIN et al. 1982), 21 Ne (MARTI & CRAIG 1987), 36 Cl (KUBIK et al. 1984;PHILLIPS et al. 1986), and 14 C (JULL et al. 1992).In 1991, LAL presented detailed terrestrial cosmic ray systematics, setting the standards for production rate and scaling formalisms as well as discussing potential applications and promising nuclide combinations (LAL 1991).GOSSE & PHILLIPS (2001) published a comprehensive review of cosmogenic nuclide methods in the Earth Sciences including the appropriate equations.Further summaries are given by NISHIIZUMI et al. (1993), CERLING & CRAIG (1994a), BIERMAN et al. (2002)

Production of cosmogenic nuclides
The Earth is constantly being bombarded by cosmic rays (primarily protons, alpha particles as well as other heavier nuclei).The primary cosmic ray flux consists of galactic and solar cosmic rays with the former clearly being more important for production of cosmogenic nuclides in minerals (LAL & PETERS 1967).Interactions of high-energy cosmic ray particles with nuclei in the Earth's atmosphere result in a cascade of secondary particles (especially neutrons).This means that in traversing the atmosphere, the flux of cosmic ray particles first increases (in the first few kilometers) then steadily decreases.Consequently, nuclide production rates in rocks at the surface of the earth are lower at lower altitude.Primary galactic cosmic rays (especially the lower energy part) are modulated by the Sun's magnetic field.High solar activity reduces the primary cosmic ray flux.Cosmic ray particles are deflected by the Earth's predominantly-dipole magnetic field.The magnetic field impedes and deflects particles with lesser energies at lower latitudes.As a consequence cosmic ray intensity and therefore nuclide production is higher at the poles than at the equator.At sea level production rates are about half at the equator compared to what they are at the poles.The altitude and latitude dependence of cosmogenic nuclide production rates reflects this modulation of the cosmic ray flux by the earth's magnetic field and the atmosphere (Fig. 2) (GOSSE & PHILLIPS 2001, MASARIK et al. 2001).Several physical models have been presented for the scaling of production rates from their sea level and high latitude values to the altitude and latitude of the sampling site (LAL 1991;DUNAI 2000;2001;STONE 2000;DESILETS & ZREDA 2001;PIGATI & LIFTON 2004;MUZIKAR 2005;DESILETS et al. 2006a) (see detailed discussion in BALCO et al. 2008).Key differences between the systems include the method of modelling the variation of the neutron flux with altitude and how past changes in the magnetic field are incorporated (BALCO et al. 2008).At present most studies use the production rates and scaling system of STONE (2000) for 10 Be and 26 Al.This allows a certain degree of intercomparison.Although the suitability of this protocol for samples at high altitude is under discussion (see BALCO et al. 2008).In any case it is crucial when Fig. 2: Production of rate of 10 Be in quartz as a function of geomagnetic latitude and altitude (based on STONE 2000).The production rates have been normalized to sea level and high latitude.At low latitude, production rates are lower than at high latitude.Production rates increase exponentially with increasing altitude.

SUSAN IVY-OCHS & FLORIAN KOBER
calculating an exposure age to use the same scaling formalities as were used for the original production rate calculations.The University of Washington/CRONUS-Earth website (http: //hess.ess.washington.edu/math/)allows the consistent comparison of data from different sites as well as presents detailed discussion of the differences and similarities of the various scaling systems (BALCO et al. 2008).Another website where these calculations can be done is cosmocalc.googlepages.com(VERMEESCH 2007).Cosmogenic nuclides are produced within minerals by several reactions (Table 2).These include spallation, muon-induced reactions and low-energy (epithermal and thermal) neutron capture (LAL & PETERS 1967).During spallation a secondary cosmic ray neutron with sufficient energy hits the target element and one or more particles are ejected from the nucleus leaving the cosmgenic nuclide in the target element's site in the mineral lattice.Cosmogenic nuclides are also produced through interactions of muons with the target element (capture of slow muons and stopping of fast muons).Because they are reacting with target elements in rocks the flux of secondary cosmic ray particles decreases (is attenuated) with depth into rock or sediment.Production due to spallation decreases exponentially with depth (LAL 1991) (Fig. 3).Muons are less apt to react than neutrons thus they penetrate deeper into the Earth's surface and production due to muons becomes increasingly important below depths of about 2 m (in a rock of density 2.7 g cm -3 ).The decrease of production depth profile for muons is described by equations with several exponentionals (SCHALLER et al. 2001;2002 Cl and to a lesser extent on 39 K (Table 2); in addition to spallation and muon-related reactions.Low-energy neutrons can diffuse back out of a rock surface (FABRYKA-MARTIN 1988).As a result of this neutron leakage, production of 36 Cl by neutron capture peaks about 20 cm down into the rock (Fig. 4).The shape of this curve depends on rock composition (proportion of target elements) and density (FAB- RYKA-MARTIN 1988;LIU et al. 1994;PHILLIPS et al. 2001) & PHILLIPS 2001).Calculation of these correction factors is also possible through the internet (http://hess.ess.washington.edu/math/).Cosmogenic nuclides [atoms g -1 ] build-up in an exposed rock surface according to the following equation: Eq. 1 where P (0) [atoms g -1 a -1 ] is the production rate at the sampling site, t [a] is the exposure age of the surface, λ [a -1 ] is the decay constant, ρ [g cm -3 ] is the density of the irradiated material, ε [cm a -1 ] is the erosion rate, and Λ [g cm -2 ] is the attenuation length.C in [atoms g -1 ] is the nuclide concentration already present at the beginning of exposure and is called inheritance.The presence of inherited nuclide concentrations will yield ages older than the true age (see also below).
Radionuclides build-up in exposed minerals until secular equilibrium is reached (saturation) which happens after about three to four halflives.At secular equilibrium the number of nuclides produced per unit time is equivalent to the number that decays; the concentration is at steady state.This is seen in Figure 5 where the the upper concentration growth curve flattens out.When the rock surface is eroding (weathering), both erosion and decay lead to loss of the nuclide.Saturation is reached earlier (after a shorter exposure period) (Fig. 6).
In the first instance, exposure ages are calculated assuming zero inheritance and without erosion such that Eq. 1 simplifies to: Eq. 2 Eq. 2 is solved for t.The production rate is scaled for site latitude (and when past magnetic field changes are considered longitude), altitude (Fig. 2), as well as, sample thickness, and topographic shielding.The attenuation length is 157 g cm -2 and the denstiy of crystalline rock is 2.7 g cm -3 (GOSSE & PHILLIPS 2001).As shown in Figure 7 the measured nuclide concentration is a direct measure of the length of the exposure period.Where field evidence indicates that rock surface weathering (erosion) has been significant, an age is calculated by using an assumed or measured erosion rate with the following equation (Eq. 1 without the inheritance term): Eq. 3

184
Cosmogenic nuclides [atoms g -1 ] build-up in an exposed rock surface according to the following equation: Eq. 1 where P (0) [atoms g -1 a -1 ] is the production rate at the sampling site, t [a] is the exposure age of the surface, ��[a -1 ] is the decay constant, ��[g cm -3 ] is the density of the irradiated material, � 6 following equation: where P (0) [atoms g -1 a -1 ] is the production rate at the sampl the surface, ��[a -1 ] is the decay constant, ��[g cm -3 ] is the de [cm a -1 ] is the erosion rate, and � [g cm -2 ] is the attenuation nuclide concentration already present at the beginning of ex The presence of inherited nuclide concentrations will yield also below).
Radionuclides build-up in exposed minerals until secular eq which happens after about three to four half-lives.At secul nuclides produced per unit time is equivalent to the number at steady state.This is seen in Figure 5 where the the upper flattens out.When the rock surface is eroding (weathering) loss of the nuclide.Saturation is reached earlier (after a sho In the first instance, exposure ages are calculated assuming erosion such that Eq. 1 simplifies to: Eq. 2 is solved for t.The production rate is scaled for site la field changes are considered longitude), altitude (Fig. 2), as topographic shielding.The attenuation length is 157 g cm -2 is 2.7 g cm -3 (GOSSE & PHILLIPS 2001).As shown in Figure concentration is a direct measure of the length of the expos Where field evidence indicates that rock surface weathering age is calculated by using an assumed or measured erosion (Eq. 1 without the inheritance term): Crystalline rock surface erosion rates are typically less than Crystalline rock surface erosion rates are typically less than 10 mm ka -1 (COCKBURN & SUMMERFIELD 2004).For spallation-dominated nuclides rock surface erosion has the effect of lowering the concentrations so that the measured age is younger than the true period of exposure.For 36 Cl exposure dating this is not always the case depending on the relative contribution of production due to low-energy neutron capture and the size of the hump in the depth profile (Fig. 4).The upper and lower age limits for each nuclide are set by a combination of geological and methodological factors.The lower age limit for 10 Be, 26 Al and 36 Cl is limited predominantly by measurement capabilities.A certain amount of carrier must be added and a certain ratio of radionuclide to stable nuclide must be attained to be above the background levels.Under appropriate conditions exposure ages in the range of several hundred years can be determined (DAVIS et al. 1999).Upper age limits are constrained by the nuclide half-life and the weathering rate of the rock surface (Fig. 6).In areas where rock weathering is slow exposure ages up to several million years can be determined with 10 Be or 26 Al; to tens of millions with 21 Ne.On the other hand, when a granitic rock surface is weathering at a rate of 5 mm ka -1 no ages older than several hundred thousand years can be calculated.But the calculated ages may be minimum ages for that landform.Although the noble gases are stable, they too are affected by erosion and will reach pseudo-saturation just like the radionuclides (Fig. 6).Often additional information can be gleaned about the history of a rock surface by measuring more than one nuclide (LAL 1991;NISHIIZUMI et al. 1993;BIERMAN et al. 1999;GOSSE & PHILLIPS 2001).As Eq. 3 contains two unknowns, the age and the erosion rate, using two nuclides can in some settings allow for the determination of both.One way to view two nuclide data (typically Al/Be, Ne/Be) is with the erosion island (or banana) plot (KLEIN et al. 1986;LAL 1991) (Fig. 8, 9).An important use of the erosion island plot is to readily distinguish data points that represent surfaces that have undergone continuous simple or single stage exposure (plot on or near the erosion island) versus those that reflect complex exposure involving burial (plot below the Al/Be erosion island, above the Ne/Be erosion  (1994), STONE et al. (1996, 1998), and PHILLIPS et al. (2001)).
Surface exposure dating with cosmogenic nuclides  SCHLÜCHTER 2008).Spalling of slabs of rock of tens of centimenters thick may lead to data that plots inside the banana or in the complex field below the banana (BIERMAN et al. 1999;KOBER et al. 2007).The minimum exposure time represented by data plotting below the erosion island is made up of continuous exposure along the outer line (bold black line) followed by burial with complete shielding (zero production).But an infinite number of periods of exposure and burial are possible which require more time (cf.BIERMAN et al. 1999).Indeed data points for rocks or sediment that were exposed then buried can arrive back inside the erosion island after a long enough period of (re)exposure.The erosion island plot can be used to estimate both exposure age and erosion rate, especially at sites where erosion rates are very low and where rocks have been exposed longer than 100 ka, for example in desert environments.For slowly eroding arid and hyperarid environments the use of 21 Ne (stable) measured in quartz in concert with 10 Be provides increased sensitivity for older, more slowly eroding surfaces (Fig. 9).The basic assumptions that are implicit in the use of in situ-produced cosmogenic nuclides to address problems of landscape evolution are: • the half-life of the radionuclide is known, • production pathways and production rates including their variation in space (including with depth into the rock or sediment) are known, • the initial nuclide concentration (inheritance) is zero or can be determined or estimated, • the mineral has remained a closed system, i.e. there has been no gain or loss of the nuclide except due to production or decay (or through erosion).
3 Specific nuclide characteristics

Radionuclides
AMS (accelerator mass spectrometry) is used to determine concentrations of long-lived radionuclides 10 Be, 14 C, 26 Al, 36 Cl by measuring ratios relative to a standard material.In AMS interfering isobars or like-mass molecules are separated at the ion source (when they do not produce negative ions), through mass discrimination with magnetic or electrostatic analyzers (as in traditional mass spectrometry), during stripping and/or during detection (FINKEL & SUTER 1993).Quartz is used in nearly all 10 Be studies.This is because it is a ubiquitous, resistant mineral that can be consistently cleaned of meteoric 10 Be.Meteoric 10 Be is produced in the atmosphere.Its presence in the analyzed mineral separate Fig. 5: Increase in concentration of the radionuclides 10 Be, 26 Al, 36 Cl and the stable nuclides 3 He and 21 Ne with time.Secular equilibrium, where production of radionuclides equals radioactive decay, is approached after 3-4 half-lives.The secular equilibrium concentraton sets the limit of the maximum exposure age that can be determined with a given radionuclide.

SUSAN IVY-OCHS & FLORIAN KOBER
Fig. 6: Increase of the concentrations of 10 Be, 26 Al, and 21 Ne with time taking into account different erosion rates of the exposed surface.Secular equilibrium, where radionuclide gain due to production equals loss due to radioactive decay and erosion, is approached earlier for more rapid erosion rates.A blow-up for the region 0 to 1 Ma is shown.
Surface exposure dating with cosmogenic nuclides 187 would lead to spurious age results.Minerals other than quartz have been tried; for example olivine and pyroxene (NISHIIZUMI et al. 1990;BLARD et al. 2008).It is important to note that in contrast to quartz, these minerals chemically weather to clay minerals, consequently problems removing meteoric 10 Be have been reported (SEIDL et al. 1997;IVY-OCHS et al. 1998a).Several groups have tried to measure 10 Be in carbonate rocks (BRAUCHER et al. 2005); the affinity of 10 Be for clay minerals poses a significant obstacle (MERCHEL et al. 2008).The use of 26 Al is restricted to minerals with low 27 Al content.Too much 27 Al would yield a 26 Al/ 27 Al ratio too low to be measured with AMS.Conveniently, quartz satisfies the requirements for both 10 Be and 26   36 Cl methodology is the implementation of isotope dilution.By adding a spike of known isotopic composition but different from the natural ratio of about 3:1 ( 35 Cl: 37 Cl), both the total rock Cl concentration and 36 Cl can be determined in a single target using an AMS set-up, through the measurement of 37 Cl/ 35 Cl as well as 36 Cl/Cl (ELMORE et al. 1997;IVY-OCHS et al. 2004).This has led to marked improvements in both precision and accuracy in 36 Cl results (DESILETS et al. 2006b).Because 36 Cl is produced from spallation of Ca and K and low-energy neutron capture on 35 Cl and 39 K and because each rock has a different chemistry, production rates for 36 Cl must be calculated individually.In addition to determination of major element oxides, concentrations of B, Gd, and Sm must be determined.These elements are strong neutron absorbers and influence the proportion of low-energy neutrons that are available for neutron capture reactions on 35 Cl and 39 K. U and Th concentrations are needed to correct for background (subsurface non-cosmogenic) neutron-capture 36 Cl production (FABRYKA-MARTIN 1988). 14C is produced in quartz by spallation of 16 O  (JULL et al. 1992; LIFTON et al. 2001; YOKOYAMA  et al. 2004).Atmospheric 14 C contamination is removed from the surfaces and crevices of the quartz grains with acid etching (similar to that used for 10 Be studies; KOHL & NISHIIZUMI 1992) and preheating.About five grams of quartz are flux melted in a flow of oxygen (LIFTON et al. 2001) to produce CO 2 .which is converted to graphite using standard procedures.Carbon ratios are measured with AMS.The possibility to use a gas ion source for direct analysis of the CO 2 with AMS is an exciting development.Because of difficulties in sample preparation and extraction, 14 C is currently used infrequently but has great potential.

Noble gases
For noble gas studies, mineral separates (tens to hundreds of milligrams) are pre-concentrated using heavy liquids and/or magnetic separation.Quartz is separated using the method of KOHL & NISHIIZUMI (1992).Because of the small amounts necessary mineral separates can also be hand-picked under a binocular microscope.Noble gases are extracted and measured with high sensitivity static noble gas mass spectrometry (NIEDERMANN 2002).Mineral separates Step-wise heating or crushing in vacuo are performed to discriminate non-cosmogenic components (nucleogenic, radiogenic, or trapped).Gas purification (separation from e.g., CO 2 , water vapour, and heavy noble gases) is accomplished by a combination of cryogenic traps and hot getters. 3He is produced by spallation reactions of nearly all elements. 3He is not measured in quartz as it diffuses out (TRULL et al. 1991;BROOK & KURZ 1993).Olivine and pyroxene phenocrysts and microphenocrysts are retentive for 3 He (KURZ et al. 1990;CERLING & CRAIG 1994b;LICCIARDI 1999;FENTON et al. 2001;MARCHETTI et al. 2005).The presence of radiogenic 3 He may limit studies to rocks with young crystallization ages (cf. NIEDERMANN 2002;WILLIAMS et al. 2005). 21Ne is produced by spallation of Si, Al, Mg and Na (Table 1) in quartz, olivine, pyroxene and sanidine.An important advantage of 21 Ne is that in contrast to 3 He it can be measured in quartz (GRAF et al. 1991;STAUDACHER & ALLÉGRE 1991).Thus three cosmogenic nuclides, 10 Be, 26 Al and 21 Ne can be determined on aliquots of a single quartz mineral separate (HETZEL et al. 2002a,b;KOBER et al. 2007).As 14 C is measured in quartz the potential to measure four nuclides in quartz exists. 21 1992;BRUNO et al. 1997;SCHÄFER et al. 1999).KOBER et al. (2005) showed that the volcanic potassium feldspar sanidine retains 21 Ne.By using sanidine one can measure 10 Be, 36 Cl and 21 Ne in aliquots of the same mineral separate (IVY-OCHS et al. 2007b).Interferences due to trapped and/or nucleogenic (both non-cosmogenic) neon isotopes (HETZEL et al. 2002a) can in most cases be deconvoluted using stepwise heating of mineral separates and plotting data on the three isotope ( 20 Ne, 21 Ne, 22 Ne) plot (GRAF et al. 1991;NIEDERMANN 2002).Surface exposure dating with cosmogenic nuclides

Sampling considerations
Cosmogenic nuclides can be used to address a variety of problems in the Earth Sciences (Fig. 1).Two factors are key in deciding which cosmogenic nuclide (Table 1) is best suited to the geological/geomorphological problem at hand: i) the half-life of the nuclide (or stable), and ii) the bedrock geology (mineralogy) of the study area.As described above certain nuclides can only be measured in certain minerals.Depositional landforms such as moraines (Fig. 10), fluvial terraces or alluvial fans can be dat-ed with cosmogenic nuclides.For the exposure date to represent as close as possible the true formation or abandonment age of the landform, the sampled object (boulder, clasts or bedrock) surface must have i) undergone single-stage exposure (no pre-exposure/inheritance), ii) been continuously exposed in the same position (not shifted), iii) never been covered, and iv) undergone only minimal surface weathering or erosion (not spalled Fig. 9: Plot of 21 Ne/ 10 Be ratios versus 10 Be concentration showing the evolution of the 21 Ne/ 10 Be ratios with time.Continuously exposed, non-eroding surfaces evolve along the bold black line.Continuously exposed surfaces eroding with steady-state erosion follow the trajectories (blue lines) that splay upward from the no-erosion line.The red line joins points of final 21 Ne/ 10 Be ratios with the given erosion rates but is not an evolution line.The prescribed area is called the «steady-state erosion island».Samples that plot below the steady-state erosion island experienced a more complex exposure that involves periods of burial.Samples may also plot above the erosion island if thick slabs have spalled off.

SUSAN IVY-OCHS & FLORIAN KOBER
and morphology of the landform (flatness of the upper surface steepness of the margins).‹›Too old ‹› ages arise when the initial nuclide concentration in the sampled rock surface was not zero (inheritance).In boulders, inheritance can be acquired in bedrock exposures before the boulder falls onto the glacier or before the landslide, or when the boulder is reworked from older depo-Surface exposure dating with cosmogenic nuclides 191 sits.For example boulders can be pushed into new moraines without having their orientation changed (see also IVY-OCHS et al. 2007a).
Where large boulders are not present amalgamated clast samples are analyzed (ANDERSON et al.1996;REPKA et al. 1997).This method is suitable for older landforms (> 100 ka) that probably never had boulders; for example fluvial terraces.The effects of unrepresentative concentrations (too high or too low) should be smoothed out by the amalgamation of >50 clasts of similar size (several centimeters in diameter) (ANDERSON et al. 1996;REPKA et al. 1997).Clasts are collected from the flat part of the landform surface well away from modifying channels.In many cases the nuclide concentration due to inheritance is revealed by measuring a depth profile.Amalgamated clast samples are taken every 10-20 centimeters down to 2 meters depth (Fig. 11) (ANDERSON et al. 1996;REPKA et al. 1997 NISHIIZUMI et al. 1989;KELLY et al. 2006).In general sampling of steep surface is avoided to circumvent the additional uncertainties associated with the dip correction.
5 Dating of Quaternary landforms

Glacial landscapes
Surface exposure dating has been used in a broad spectrum of settings in glacial landscapes.This  1996, REPKA et al. 1997)).Mischproben zeigen eine Abnahme der Nuklidkonzentration mit der Tiefe.
includes the dating of boulders on moraines (for a recent compilation see REUTHER et al. 2006a), boulders on glacial outwash fans (PHILLIPS et al. 1997), boulders on the former margins of icedammed lakes (DAVIS et al. 2006) and boulders deposited during catastrophic outburst of icedammed lakes (CERLING et al. 1994;REUTHER et al. 2006b).Glacially-polished bedrock is also analyzed for determining rates of ice retreat, and depth of subglacial erosion.The method of burial dating has great potential, especially in the dating of old, buried glacial deposits (see DEHNERT & SCHLÜCHTER 2008).BALCO et al. (2005) used 26 Al and 10 Be ratios in quartz to determine the age of deeply buried paleosols and underlying till units.

Dating of moraines
Moraines record the location of the margins of a glacier in the past.Changes in glacier volume and length themselves reflect changes in tem-Surface exposure dating with cosmogenic nuclides perature and precipitation patterns in a region with time (KERSCHNER 2005).Therefore if one can directly date moraines one can construct a chronological structure to past glacier fluctuations and therefore past climatic fluctuations (KERSCHNER & IVY-OCHS 2008).Depending on the detailed structure of the moraine complex the innermost moraine records the onset of glacier downwasting (GOSSE 2005).For example near synchrony of glacier downwasting at the end of the Last Glacial Maximum world-wide has been established with 10 Be dating (SCHAEFER et al. 2006).As described above large boulders in stable positions along the moraine crest are sampled preferentially.A boulder must be large and stable enough not to have toppled or shifted and high enough to have protruded above the matrix since moraine deposition (HALLET & PUTKONEN, 1994;PUTKONEN & SWANSON, 2003;PUTKONEN & O'NEAL, 2006).Key limitations include weathering of the boulder surface and degradation of the moraine itself (ZREDA et al. 1994;ZREDA & PHILLIPS, 1995;REUTHER et al. 2006a;IVY-OCHS et al. 2007a).Sampling several small clasts on a young (<20ka) moraine is unlikely to give the depositional age (IVY-OCHS et al. 2007a).Moraines ranging in age from hundreds of years to hundreds of thousands of years in both hemispheres have been dated (e.g.PHILLIPS et al. 1990;IVY-OCHS et al. 1999, 2006;BARROWS et al. 2002;KAPLAN et al. 2004;BRINER et al. 2005;BALCO & SCHAEFER 2006;Akçar et al., 2008b).The older the moraine the greater the spread in ages amongst the exposed boulders.This may make interpretation of the deposition age of moraines that are older than 100 ka difficult (KAPLAN et al. 2005;SMITH et al. 2005).In such cases the oldest age is assumed to be closest to the landform age.But this age may still be a minimum age for the landform.
In a suite of ages from a single moraine outliers that are too old reflect inheritance.Boulders that are deposited in moraines may have acquired inheritance in the bedrock setting or because they are reworked from older moraines.
For the case of a boulder suface exposed in the bedrock setting, the amount of inheritance is dependent on how long the bedrock surface was exposed (and the bedrock weathering rate) and how deep the boulder originated inside the bedrock surface (IVY-OCHS et al. 2007a).The effect of this will be greater in younger (Holocene) moraines.Based on a compilation of numerous published moraine boulder exposure ages, PUTKONEN & SWANSON (2003) found that pre-exposure was observed in moraine boulders in only a few percent of cases.

Glacially-modified bedrock surfaces and rates of sub-glacial erosion
Initial deglaciation of the valley bottom and rates of glacier downwasting can be determined by analyzing cosmogenic nuclides in glacially-scoured bedrock surfaces (e.g.roche moutonnées) (GOSSE et al. 1995;GUIDO et al. 2007).However, results from bedrock surfaces should be viewed with caution.If three meters or more of bedrock have not been removed by sub-glacial erosion during the last glaciation, then the rock surface may contain inherited nuclide concentrations.This has been noted where the rock is highly resistant (GUIDO et al. 2007).In the case where the timing of deglaciation is independently known (for example from exposure dating of erratics), the nuclide concentration measured in glacially-scoured bedrock can be used to determine sub-glacial erosion rates.Determined rates are on the order of 0.1 to 1 mm per year (BRINER & SWANSON 1998;COLGAN et al. 2002;FABEL et al. 2004).
Direct determination of such rates is only possible with cosmogenic nuclides.Cosmogenic nuclides have a unique characteristic in that they can be used to elucidate fundamental information about the thermal regime of past ice sheets.Lower 26 Al than 10 Be concentrations in bedrock reveals areas where ice was frozen to its bed and unerosive (BIERMAN et al. 1999;FABEL et al. 2002;MARQUETTE et al. 2004;STAIGER et al. 2005;SUGDEN et al. 2005;LINGE et al. 2006;PHILLIPS et al. 2006).Production ceased when the surfaces were covered by ice, 26 Al decays faster than 10 Be thus 26 Al/ 10 Be ratios plot below the erosion island (see also DEHNERT & SCHLÜCHTER 2008).

Alluvial, lacustrine and marine systems
Determining the age of alluvial landforms (for example fluvial terraces and alluvial or debrisflow fans) provides fundamental information about timing and rates of depositional processes.Constructing an age sequence also allows estimation of the contribution of the various external forcing mechanisms, such as tectonically induced base-level or regional slope changes, climatically (discharge) induced or sediment supply dependent variations.Fluvial or marine terraces; or alluvial fans can be dated using depth profiling and measuring one or more nuclides (samples from the upper 2 meters) (PERG et al. 2001;WARD et al. 2005;RYERSON et al. 2006;FRANKEL et al. 2007) or using burial dating (WOLKOWINSKY & GRANGER 2004) (see also DEHNERT & SCHLÜCHTER 2008).

Wave-cut bedrock platforms and paleoshorelines
Lake and marine paleoshorelines delineate higher water levels and thus record past changes in the balance between precipitation vs. evaporation plus lake drainage and inflow or fluctuations in sea-level, respectively.In order to integrate these periods of lake/sea highstands into existing regional chronologies a time frame is required.

Alluvial fans
Patterns of alluvial fan deposition record variations of erosional processes in the fan catchment and changes in incision of the fan head over time (DÜHNFORTH et al. 2008).When individual lobes Surface exposure dating with cosmogenic nuclides can be clearly mapped then surface exposure dating is a useful tool for dating of abandoned fan lobes (DÜHNFORTH et al. 2007).Original depositional forms such as flow snouts and levees are often still present on debris-flow fans that are less than 100 ka old.To date the different fan lobes large boulders in clear position on snout or levee are sampled (BIERMAN et al. 1995;ZEHFUSS et al. 2001;DÜHNFORTH et al. 2007).On older alluvial fans degradation of fan surfaces and boulder grusification leads to smoothing out of the original bar and swale morphology.After a long enough period of time (50 ka?) desert pavement of interlocking clasts develops (RYERSON et al. 2006) with associated desert varnish.In these cases amalgamated clast samples are analyzed.An important question is whether or not the dates represent the interval of lobe construction or whether they point to timing of fan abandonment.Another serious concern for the dating of alluvial and debris-flow fans is inheritance.Similar to the case of the strandline clasts inheritance can have been acquired in the catchment, in intermediate storage, or as clasts are moved from older to younger lobes.Reflecting the bedrock rezeroing effect of glaciers, inheritance is often greater in fans originating in unglaciated catchments.

Fluvial incision rates
Fluvial incision rates and their variation with time are calculated by using the exposure age and height of fluvial (WARD et al. 2005) or strath terraces (BURBANK et al. 1996;LELAND et al. 1998;PRATT-SITAULA et al. 2004;REUSSER et al. 2004).Boulders or amalgamated clast samples on debris-flow deposits on ancient bedrock straths have been dated to determine incision rates (FENTON et al. 2004;MARCHETTI & CERLING 2005).Based on 10 Be in samples from three strath terraces REUSSER et al. ( 2004) reported that incision rates of the Susquehanna River (eastern U.S.A.) more than doubled to 0.5 mm a -1 during the last glaciation (32 to 16 ka).In Taiwan 36 Cl measured in samples from fluvially-sculpted limestone channel walls in Taroko Gorge indicate incision rates of 26 mm a -1 (SCHALLER et al.

2005
). Steep gorge walls may provide a more continuous record of incision than strath terraces as the latter is a step-like sequence.In any case it is often difficult to verify that strath terraces were never covered by sediment during their exposure histories.Past incision rates can also be calculated by combining burial ages of cave sediments with the height of the cave above the active river channel (GRANGER et al. 1997;STOCK et al. 2004) (see DEHNERT & SCHLÜCHTER 2008).

Tectonic and mass movement studies
Cosmogenic nuclides can be used to determine rates of tectonic activity in two ways i) by dating landforms that have been offset by movement along faults and ii) by dating bedrock fault surfaces directly.

Dating of offset landforms
Slip rates on strike-slip faults or rates of uplift on normal faults have been determined by taking the age and offset distances of moraines (LASSERRE et al. 2002;BROWN et al. 2002), fluvial terraces (HETZEL et al. 2002a) and alluvial fans (BIERMAN et al. 1995;SIAME et al. 1997;BROWN et al. 1998; VAN DER WOERD et al. 1998;ZEHFUSS et al. 2001;RITZ et al. 2003;MÉRIAUX et al. 2005;RYERSON et al. 2006;FRANKEL et al. 2007).Cautious interpretation of field evidence for offset distances (piercing point, offset terrace risers vs. treads; offset stream channels) is important (MÉRIAUX et al. 2004).

Direct dating of fault surfaces
Limestone bedrock fault surfaces are dated directly with 36 Cl (ZREDA & NOLLER 1998; MIT-CHELL et al. 2001;BENEDETTI et al. 2002;2003;PALUMBO et al. 2004).Suitable fault surfaces are several meters high several kilometers long, fresh, and uneroded.Samples are taken at centimeter intervals along the exposed fault surface often with a circular saw.In principle, the recurrence interval of earthquakes along the fault segment can be reconstructed based on step changes in 36 Cl concentrations determi-

SUSAN IVY-OCHS & FLORIAN KOBER
ned in tens of samples from a single fault face.
Earthquakes that occurred as recently as only a thousand years ago are dated.In principle, such a study is also possible with 10 Be in quartz, but may be limited by sample preparation time for the large number of samples required.The age of the fault surface that can be determined is controlled by the karst weathering rate of the limestone and is generally restricted to those less than tens of thousands of years old.

Landslides
Landsliding is an important process for the modification of valley slopes and cross pro-

Volcanic sequences
Based on detailed mapping relative eruption sequences are constructed.Lava flows can often be 14 C dated with entrained charcoal or by analyzing underlying burnt soil, when such material is found.Time of eruption of lava flows and domes are dated with K/Ar and 40 Ar/ 39 Ar (KELLEY 2002).Not all flows are suitable for Ar/Ar dating: i) high-K minerals may be lacking, ii) excess Ar or iii) Ar loss may be a problem.As erupted volcanic rocks are newly formed surfaces with no pre-exposure, their initial nuclide concentration is zero, they are well-suited for surface exposure dating.The most challenging aspect is to unequivocally identify the original eruptive surface to sample.This may be established by the presence of primary eruptive features such as spatter (Fig. 10), frothy glassy texture of the cooling rinds on flows or bombs, vesicles near the tops of flows, or the ropey texture of pahoehoe flows (CER-LING & CRAIG 1994a;FENTON et al. 2001;PHIL-LIPS 2003).Mafic lava flows with olivine and/or Surface exposure dating with cosmogenic nuclides pyroxene phenocrysts or microphenicrysts are dated with 3 He and/or 21 Ne (KURZ et al. 1990;ANTHONY & POTHS 1992;CERLING & CRAIG 1994b;LICCIARDI et al. 1999;2006;FENTON et al. 2004;WILLIAMS et al. 2005;DUFFIELD et al. 2006). 21Ne and 10 Be in quartz and/or sanidine phenocrysts have been used to date ignimbrites (welded tuffs) (LIBARKIN et al. 2002;KOBER et al. 2005).Volcanic rocks of intermediate composition lacking quartz or olivine or pyroxene or flows lacking phenocrysts completely can be exposure dated with 36 Cl (PHILLIPS 2003; ZREDA  et al. 1993).This also affords the opportunity for intercomparison between the eruptive (K-Ar or 40 Ar/ 39 Ar age) and the exposure age (FENTON et al. 2001;PHILLIPS 2003).For the dating of landforms (moraines, flood deposits, etc.) comprised of volcanic rocks 3 He and 36 Cl are well suited.

Ancient Landscapes
It is only since the advent of cosmogenic nuclides that rock surfaces in the arid and hyperarid deserts of southern Africa (COCKBURN et al. 1999;2000;FLEMING et al. 1999 BELTON et al. 2004), western South America (NISHIIZUMI et al. 2005;KOBER et al. 2007) and Antarctica (SUMMERFIELD et al. 1999;SUGDEN et al. 2005) can be directly dated.The long-lived radionuclides ( 10 Be and 26 Al) and especially the noble gases ( 3 He and 21 Ne) are well suited to dating landforms that have been exposed for millions of years.Combining nuclides, for example 10 Be and 26 Al or 10 Be and 21 Ne, is critical to check for continuous exposure and rule out intermittent coverage.Studies using two or more nuclides indicate that many of the ancient bedrock surfaces in the deserts have been exposed continuously, but are never-theless weathering, albeit remarkably slowly.In most cases only minimum ages can be determined.Maximum erosion rates are calculated by assuming that nuclide concentrations are in steady state.At several of these sites erosion rates range down to less than 1 mm kyr -1 (COCKBURN & SUMMERFIELD 2004).With cosmogenic nuclide methods, fundamental information about the age and mode of formation of ancient landscapes has been gained.High nuclide concentrations measured in Australia led to the conclusion that inselbergs are direct descendents of early Cenozoic or even Mesozoic landforms (BIERMAN & CAFFEE 2002).Combined fission-track and low cosmogenic nuclide-derived erosion rates across the Namibian escarpment indicate that great escarpments do not form by rapid (and ongoing) escarpment retreat, but must have formed early on after continental break-up (COCKBURN et al. 2000). 10Be in combination with 26 Al data showed that many surfaces in Antarctica have experienced single-stage, continuous exposure with remarkably low erosion rates for at least the last several million years (NISHIIZUMI et al. 1991;BROOK et al. 1995;IVY-OCHS et al. 1995;SUMMERFIELD et al. 1999;MATSUOKA et al. 2006).The high nuclide concentrations found in rocks in the Dry Valleys Antarctica provide irrefutable support for the premise that the East Antarctic Ice Sheet has been a stable feature since its inception (SCHÄFER et al. 1999;ACKERT & KURZ 2004).Measured pre-Pleistocene apparent exposure ages and corresponding low denudation rates characterize the deserts of northern Chile, despite active uplift (DUNAI et al. 2005;KOBER et al. 2007;NISHIIZUMI et al. 2005).

Summary and outlook
The ability to use cosmogenic nuclides to determine how long minerals have been exposed at the surface of the earth provides an unrivaled tool for determining ages of landforms and rates of geomorphic processes.Depending or rock and landform weathering rates, landforms ranging in age from a few hundred years to tens of millions of years can be dated.Because of this unique capability, the variety of applications of cosmogenic nuclides will continue to grow.Concern about methodological uncertainties, such as those associated with the production rates, the site latitude and SUSAN IVY-OCHS & FLORIAN KOBER altitude scaling factors, as well as the effect of past changes in the Earth's magnetic field, has led to the establishment of an international consortium made up of CRONUS-Earth (www.physics.purdue.edu/cronus)and CRO-NUS-EU (www.cronus-eu.net).Analysis of artificial targets and samples from natural sites with independent age control are underway to refine production rates.Scaling factors are being evaluated with neutron monitors and analysis of same age natural samples taken along altitudinal transects (for example lava flows).Numerical modeling is being used to constrain production rates and scaling factors both now and in the past.The half-lives of radioactive nuclides must be accurately known.In the case of 10 Be, two different half-lives have been published, 1.51 and 1.34 Ma (GRANGER 2006;NISHIIZUMI et al. 2007).When these factors are better constrained the errors of the final ages will be closer to the range of the AMS and noble gas mass spectrometry measurement uncertainties (of the order of 1-4 %).With improved knowledge of production rates and their scaling to the site, the precision of obtained ages will improve.But the accuracy of the ages remains a question of geological uncertainties.The degradation of both rock surfaces and the landforms with time imposes clear limitations on the time range and accuracy of dating.Similarly, the natural variability of samples depends on landform morphology and its age.Obtained exposure ages must be evaluated individually for conformity with field relationships, including local terrace or moraine stratigraphy and regional morphostratigraphic relationships; as well as with independent age constraints for the same or correlative features.For older landforms (more than a hundred thousand years) measurement of mulitple cosmogenic nuclides can reveal fundamental information, such as non-continuous exposure, which must be factored into interpretations (ALVAREZ-MARRÓN et al. 2007;KOBER et al. 2007).Cosmogenic nuclides provide a powerful and multifaceted tool whose potential has yet to be fully realized.But this power is tempered with the need for careful sampling based on detailed field mapping.

Fig. 1 :
Fig.1: Schematic diagram showing the various landforms that can be dated and approaches for using cosmogenic nuclides to address questions of timing and rates of landscape change (see alsoBIERMAN & NICHOLS 2004).

Fig. 12 :
Fig. 12: 10 Be and 21 Ne concentrations with depth into a rock surface.The nuclide concentration in a sample is a composite of the inherited and the post-depositional nuclide concentration.The inherited concentration can be approximated by measuring several samples from different depths, tens of centimeters apart down to about 2 m. A. 10 Be and 21 Ne concentrations in a depth profile after exposure for 100 ka with a constant erosion rate of 0.20 mm/ka.B. The upper 50 cm have been mixed by bioturbation resulting in constant nuclide concentrations.C. Curve showing the effect of inheritance on both the 10 Be and 21 Ne concentrations.Abb.12: Entwicklung der Nuklidkonzentration ( 10 Be und 21 Ne) in Abhängigkeit von der Tiefe.Die Gesamtnuklidkonzentration ist die Summe aus ererbten Konzentrationen und der Konzentration aus der Produktion nach Ablagerung der Sequenz oder eines vormalig exponierten Tiefenprofils im Festgestein.Die ererbte Nuklidkonzentration kann durch die Beprobung in unterschiedlichen Tiefen (mehrere Proben über eine Tiefe von mindestens 2 m) ermittelt werden.A. 10 Be und 21 Ne Nuklidkonzentrationen in einem Tiefenprofil nach einer Expositionszeit von 100000 Jahren unter einer konstanten Erosionsrate von 0,20 mm/1000 Jahre.B. Der obere Bereich des Profils (50 cm) ist durch Bioturbation homogenisiert, was zu einer konstanten Nuklidkonzentrationen über dieses Intervall führt.C. Darstellung des Einflusses einer ererbten Nuklidkonzentration auf die 10 Be und 21 Ne Nuklidkonzentrationen.

Table 2 : Main reactions to produce cosmogenic nuclides on the Earth. Tab. 2: Hauptreaktionen kosmogener Nuklide an der Erdoberfläche. Target element Spallation Negative-muon capture Low-energy neutron capture
Al, 36Cl und Edelgase 3 He und 21 Ne in Al so that both can be extracted by dissolving a single quartz mineral separate.Pure quartz is obtained by selective chemical dissolution in a hot ultrasonic bath and/or on a shaker table (KOHL & NISHIIZUMI 1992).Most BIERMAN et al. 2002) (see also http: //depts.washington.edu/cosmolab/chem).Although it has a higher production rate (Table 1), one disadvantage of 26 Al is that a separate accurate measurement (with its own uncertainties) is required to determine the 27 Al content of the quartz.Ratios of 10 Be/ 9 Be or 26 Al/ 27 Al are measured with AMS.Because Cl is hydrophyllic contaminating meteoric 36 Cl (or 37 Cl or 35 Cl) can be removed with rinsing procedures.Under most conditions any rock type or mineral separate can be used for 36 Cl exposure dating.On the other hand if secondary minerals which include meteoric 36 Cl precipitate (for example secondary calcite) then the system is no longer closed.Sample preparation procedures for 36 Cl are given in ZREDA (1994), IVY-OCHS (1996), STONE et al. (1996, 1998), IVY-OCHS et al. (2004) and DESILETS et al. (2006b) (see also http://depts.washington.edu/cosmolab/chem/). Crushed rock samples are first leached several times to release any non-in situ produced Cl.Several milligrams of carrier (in solution) of known isotopic composition (pure 35 Cl, 37 Cl, or a mixture of both) are added.Carbonate rocks are dissolved with HNO 3 and silicate rocks with HF.Sulfur is removed by precipitation of BaSO 4 ( 36 S interferes with AMS measurement of 36 Cl).A crucial improvement in Ne is also measured in pyroxene and olivine (MARTI & CRAIG 1987; STAU- DACHER & ALLEGRE 1991; POREDA & CERLING, DEHNERT & SCHLÜCHTER 2008)versus 10 Be concentration showing the evolution of the26Al/ 10 Be ratios with time.Continuously exposed, non-eroding surfaces evolve along the bold black line.Continuously exposed surfaces eroding with steady-state erosion follow the trajectories (blue lines) that splay downward from the no-erosion line.The red line joins points of final26Al/ 10 Be ratios with the given erosion rates but is not an evolution line.The prescribed area is called the «steady-state erosion island».Samples that plot below the steady-state erosion island experienced a more complex exposure that involves periods of burial (seeDEHNERT & SCHLÜCHTER 2008).Samples may also plot below the erosion island if thick slabs have spalled off.Abb.8:Darstellung des Verhältnisses von 10 Be/26Al zur 10 Be-Konzentration, welches die Entwicklung des Verhältnisses von 10 Be/26Al mit der Zeit aufzeigt (als Funktion der Konzentration).Im Falle einer kontinuierlichen Exposition und Null-Erosion würde eine Probe auf der schwarzen, dicken Linie liegen. Kontinuirliche Exposition mit einer konstanten Erosionsrate DEHNERT & SCHLÜCHTER 2008)ktoren (blaue Linien) bewegen, die von der Null-Erosionslinie abzweigen.Die rote Linie verbindet die Endpunkte dieser Trajektoren (Gleichgewichtsendpunkt, Verhältnis von 10 Be/26Al ist konstant) und ist artifiziell.Das Feld zwischen Null-Erosionslinie und der artifiziellen Linie wird als "Gleichgewichts-Erosionsinsel" bezeichnet.Probenpunkte unterhalb dieser Insel haben eine komplexe Expositionsgeschichte, einschliesslich Perioden mit Abschirmung zur kosmischen Strahlung (siehe auchDEHNERT & SCHLÜCHTER 2008).Probenpunkte unterhalb der "Gleichgewichts-Erosionsinsel" können auch durch eine schalige Abspaltung vom Festgestein verursacht werden.