Explicit treatment of inheritance in dating depositional surfaces using in situ 10Be and 26Al

Robert S. Anderson Department of Earth Sciences and Institute for Tectonics, University of California, Santa Cruz, James L. Repka California 95064 Gregory S. Dick

ABSTRACT (Bierman, 1994; Hallet and Putkonen, We describe a new strategy for dating depositional landscape surfaces using in situ– 1994), inheritance during exhumation and produced cosmogenic radionuclides (CRNs) that removes the complication of nuclide in- transport places an additional limitation on heritance by clasts prior to deposition. Two amalgamated samples, each consisting of 30 the application of in situ–produced CRNs to clasts, one from the surface and one from a fixed depth in the subsurface, constrain this the dating of depositional surfaces. The CRN inheritance and date the surface. The inheritance may be used to estimate minimum clasts within most deposits have undergone exhumation rates and maximum transport times within the geomorphic system. We test the not only CRN production at the final site of technique using 10Be and 26Al to date the third (FR3) of five terraces along the Fremont deposition, but also potentially significant River, Utah, and the third (WR3) of 15 along the Wind River, Wyoming. Whereas effective and highly variable production throughout ages based solely upon the surface samples yield 118–138 ka (FR3) and 93 ka (WR3), the their path to this site (Fig. 1). Dating of a subsurface samples reveal that inheritance accounts for 22% (WR3) to 43% (FR3) of the single surface clast by CRN methods reveals total CRN concentration. Taking this into account yields terrace ages of ϳ61–81 ka for FR3 at best only a maximum age of the surface, and ϳ67–76 ka for WR3. We explore the dependence of age estimates on the accumulation leading Cerling and Craig (1994, p. 310) to history of the terrace silt caps. state that ‘‘in the end such studies will re- quire numerous samples to overcome the INTRODUCTION often contain good target material for the stochastic nature of sediment transport and Dating of depositional surfaces, such as production of CRNs, and several commonly surface residence history.’’ fluvial and beach terraces, alluvial fans, and used CRNs have half-lives that allow use Measurement of many clasts, taking the moraines, is necessary for constraining rates through the Quaternary. Whereas in situ minimum effective age to be that of the sur- of tectonic uplift, river incision, and pedo- CRN studies using 10Be, 26Al, and 36Cl have face, is expensive in both time and funds, genic evolution, and for establishing links focused largely on dating of erosional bed- and may still result in the wrong answer. As between oceanic and terrestrial records of rock surfaces (Lal, 1991; Nishiizumi et al., a more robust and economical alternative, climate change. However, dating of these 1993; Bierman, 1994; Cerling and Craig, we present a method in which many indi- landforms currently relies on methods with 1994), Phillips et al. (1990) and Chadwick et vidual clasts are amalgamated to construct significant temporal, geographic, or litho- al. (1994) have dated moraines and outwash two samples, one from the surface and one logic limitations. We explore a method that terraces, and Trull et al. (1995) have at- from the subsurface, that explicitly ad- employs in situ–produced cosmogenic ra- tempted to date shorelines. dresses the issue of CRN inheritance and dionuclide (CRN) concentrations to esti- Although many potential problems have allows strong constraints to be placed on mate the residence time of clastic material been addressed by previous workers, espe- both the surface age and the transport his- on the depositional surface. These surfaces cially in dealing with dynamic landforms tory of clasts within the geomorphic system.

Figure 1. Cosmogenic ra- dionuclide concentration history of clast through its transport history in hill- slope and fluvial systems, and in its subsequent res- idence on depositional site. Production rate is dictated by depth beneath local surface, z, which falls off exponentially with depth (inset). Exhumation on hillslope results in monotonic increase in concentration. Produc- tion history is stochastic within fluvial system, as clast travels between point bars, within which it is buried to differing depths. Evolution of con- centration on final terrace site is shown for two pos- sible burial depths, one on surface, other in subsurface. One scenario for surface clast (see Fig. 3 for others) is that it remains on silt surface, in which case it always sees surface production rate Po, and attains sampled concentration of Ns. Subsurface clast within underlying gravels (stippled) will undergo lower production rate, and is sampled with concentration Nss. Clasts on terraces much older than nuclide half lives attain secular equilibrium, with labeled concentrations. Amalgamated samples consisting of numerous clasts allow back-calculation of terrace age, ␶, and of mean inherited radionuclide concentration, Nin.

Geology; January 1996; v. 24; no. 1; p. 47–51; 3 figures; 1 table. 47 METHOD OF CONSTRAINING during transport, to be a random function INHERITANCE over the range [0, H], where 0 represents the We focus on the radionuclides 10Be and surface, and H the maximum possible burial 26 Al, with half-lives (t1/2) of 0.7 and 1.5 Ma, depth within the transport system. Integrat- respectively, although the results are appli- ing the accumulation history using the sim- cable to other stable and radionuclide sys- plest case of uniform probability between tems. It is sufficient to know (1) that the in z ϭ 0 and H yields situ production rate of the nuclide at the surface, P , is dependent upon the altitude 1 Ϫ eϪ͑H/z*͒ o N ϭ P T , (3) and latitude in a way that has been well con- trans o ͭ ͑H/z*͒ ͮ strained (Lal, 1991), and (2) that the pro- duction rate, P, falls off exponentially with where T is total time spent in the fluvial sys- depth, z, with a characteristic length scale z* tem (both in transport and in storage). For (ϭ⌳/␳, where ⌳ is the mass attenuation co- H greater than a few times z*, this reduces efficient, and ␳ is the bulk density of the to the yet simpler expression N ϭ trans Figure 2. Schematic illustration near-surface material; Lal, 1991; Nishi- (Po Tz*)/H. The combination Tz*/H repre- of postdepositional accumula- izumi, 1994); i.e., P ϭ Po exp(Ϫz/z*). Be- sents the effective time spent by the clast in tion of cosmogenic radionu- cause z* Ϸ 0.6 and 0.8 m in rock and allu- the production zone. For large H, the clast clides. (CRNs) at various depths. vium, respectively, it is the clast residence spends less time close to the surface and ac- Terrace gravels mantling strath time and depth trajectory within roughly 1 m cumulates fewer CRNs. Given that varia- surface cut into bedrock arrive on site with varying cosmogenic of the surface during exhumation and trans- tions in z* are expected to be small within inheritance (wide fuzzy line), but port that results in nuclide inheritance. such fluvial deposits (Ͻ10%), the clast-to- are assumed to have been em- placed with uniform distribution A clast may accumulate CRNs that con- clast variation in Ntrans arises principally tribute to the inherited signal throughout its from variations in total time in the fluvial of CRN inheritance with depth, characterized by mean, Nin, and history. The final CRN concentration, N,in system, T. spread represented by fuzzy line a clast from a depositional surface is Following transport, a clast comes to rest and shape of probability density, within the deposit we wish to date. Because p(N). Subsequent accumulation N ϭ N ϩ N ϩ N ϩ N ϭ N pre exh trans dep in of the stochastic nature of exhumation and of CRNs results in exponentially transport, any two clasts within the deposit decaying component, altering ϩ Ndep, (1) mean of distribution, but not are likely to have different CRN concentra- spread (curved fuzzy line). Mean where Npre is the concentration retained tions, resulting in unknown error when dat- concentrations derived from from an ancient transport-depositional cy- ing the surface using single clasts. However, amalgamating samples (black and (0 ؍ boxes) from surface (z cle; Nexh is the concentration produced dur- measurement of a large collection of clasts z ) will differ ؍ subsurface (z ing exhumation; Ntrans is that produced dur- ss should constrain the mean inheritance, Nin. ing transport to the final depositional from one another only by this The average CRN concentration of many postdepositional component. surface; and Ndep is that produced during clasts immediately after deposition repre- Both initial uniform inheritance, residence on the final depositional surface. sents a combination of the mean hillslope Nin, and elapsed time since The first three terms correspond to inheri- exhumation rate ͗⑀˙͘ and transport time, ͗T͘, abandonment of surface can be calculated from measured CRN tance, Nin. All terms must include radioac- undergone by the population of clasts. If Nin concentrations of amalgamated tive decay. does not vary with position on or depth samples. Consider the history of an individual clast. within the deposit, and if the clasts were de- During exhumation, the clast spends time posited over a period short relative to the within the zone of cosmic ray penetration; deposit age, then we may use the following the resulting production of CRNs yields strategy to account for inheritance and con- gamated samples. However, the postdepo- Nexh upon arrival at the ground surface strain the deposit age. sitional production rate in the subsurface (Fig. 1), which depends upon the time it Subsequent to final deposition, accumu- sample should be reduced below that of the takes to pass through the zone of cosmo- lation of radionuclides depends only upon surface sample (Fig. 2). We therefore use genic production. Ignoring small changes in the depth of the clast and the bulk density of the CRN production rates for each sample altitude that take place during exhumation the overlying deposit (Fig. 2). We first mea- depth to back calculate both the time at through the CRN production zone, assum- sure the cosmogenic radionuclide concen- which the samples had similar CRN concen- ing that erosion is steady, yields (Lal, 1991): tration, Ns, of an ‘‘amalgamated sample’’ trations (the surface age, ␶), and the mag- created by combining equal masses from nitude of this concentration (the CRN in- Po N ϭ , (2) many surface clasts (depth z ϭ 0; production heritance, N ). Graphically, this is exh ␭ ϩ ˙⑀/z* in rate Ps ϭ Po). This concentration is the sum analogous to tracing back the CRN trajec- where ␭ϭln 2/t1/2 is the nuclide decay of the mean of the stochastic inherited com- tories from their measured values, Ns and constant. ponent, Nin, and that acquired during resi- Nss, along lines inclined at Po for the surface Once the clast is free to move, processes dence on the depositional surface, Ndep. Sec- sample and at Po exp(Ϫzss/z*) for the sub- ranging from glacial and fluvial transport to ond, we measure the CRN concentration, surface sample until they intersect (star in longshore drift move it through the geomor- Nss, from a similarly amalgamated sample Fig. 1). For the case in which we may safely phic system. The clast undergoes periods of derived from many clasts collected from a assume no radioactive decay (stable nu- storage at varying depths within the geomor- known depth in the subsurface, zss. If there clides, or surface age young relative to the phic system, introducing a stochastic com- is no vertical variation in Nin, the contribu- nuclide half-life), this intersection occurs at ponent to the production history (Fig. 1). tion from inheritance should be similar ␶ϭ⌬N/⌬P, where ⌬P ϭ Ps Ϫ Pss, and ⌬N ϭ Consider z(t), the depth history of the clast between the surface and subsurface amal- Ns Ϫ Nss. Taking into account decay,

48 GEOLOGY, January 1996 1 ⌬P and sieved individually to a uniform size of ably even greater. The terrace surfaces are ␶ ϭ ln . (4) ␭ ͩ⌬P Ϫ ␭⌬Nͪ 200–500 ␮m, and an equal mass is removed at least as young as the youngest surface from each to create an amalgamated sam- clasts. If inheritance were not taken into ac- The inheritance can then be calculated: ple. Organic coatings, Fe and Mg oxides, count, the amalgamated surface sample

NinϭNssϪPss␶ϭNsϪPs␶. and carbonate coatings are eliminated with from FR3 would suggest that this terrace Application of this strategy requires esti- weak acids, after which the samples are sub- was abandoned at 118 ka (Be) or 138 ka (Al) mation of the number of clasts needed to jected to a series of leaches to remove non- (Table 1). Using the CRN concentration of establish Nin in the amalgamated samples. quartz silicates (Kohl and Nishiizumi, 1992) our amalgamated subsurface sample, and Numerical modeling using plausible exhu- and to eliminate ‘‘garden variety’’ 10Be. Sub- employing equation 4, we both revise this mation and transport scenarios suggests that sequent total dissolution, ion chromatogra- age estimate downward to 61 Ϯ 3 ka (Be) or

30 clasts are sufficient to constrain the mean phy, and accelerator mass spectrometry (El- 81 Ϯ 3 ka (Al) (Tadj in Table 1) and estimate to within 5% (Repka et al., 1994). more and Phillips, 1987; Davis et al., 1990) the inheritance, Nin. Roughly 43% of the are standard. CRN concentration of the surface sample is APPLICATIONS IN FLUVIAL SYSTEMS apparently due to inheritance. At present We have tested this technique on fluvial FREMONT RIVER we have no explanation for the discrepancy strath terraces in the western United States. Five well-expressed terraces of the Fre- between 10Be and 26Al results on the amal- Along the Fremont River in Utah, terrace mont River (FR1–FR5, from lowest to high- gamated FR3 samples, and we will use the ages have been estimated (Howard, 1970, est) are constructed of outwash from gla- mean of the adjusted ages (ϳ71 ka) in our 1986) through correlation with glacial mo- ciers that bounded the Aquarius and Fish discussion. raines upstream thought to correlate with Lake plateaus to the west. Each terrace is a the Pinedale and Bull Lake moraine chro- strath surface cut into the Mancos shale and WIND RIVER nology (Flint and Denny, 1956); no absolute sandstone bedrock, mantled by 3–5 m of The Wind River drains the northeast ages exist. On the Wind River terraces in coarse gravel and sand. All terrace surfaces flank of the Wind River Mountains and the Wyoming, the date of one terrace (WR3) display a single layer of varnished clasts cap- southern flank of the Absaroka Mountains. was determined by application of 36Cl ping a few centimetres of nearly clast-free Fifteen well-preserved terrace surfaces, (Chadwick et al., 1994), although inheri- silt. numbered WR1 through WR15 from lowest tance was not considered. The largest terrace-cover clasts (up to 1 (youngest) to highest (oldest), probably At both sites, we use quartzite clasts for m) are basalts; locally derived sandstone and span the entire Pleistocene (Chadwick et al., 10 26 Be and Al because: (1) SiO2 is an ideal more far-traveled quartzite and chalcedony 1994). The terraces are strath surfaces man- target for production of both nuclides clasts have maximum diameters of 20 cm. tled with several metres of coarse alluvial (Nishiizumi et al., 1993), and quartzites are The quartzites are derived from the Triassic gravel and capped by tens of centimetres of close to being pure SiO2; (2) quartzite is very Shinarump conglomerate and the Jurassic silt. resistant to weathering, and thus interpre- Brushy Basin member of the Morrison For- Using cosmogenic 36Cl, Phillips and tation is not complicated by postdeposi- mation (Billingsly et al., 1987). Zreda dated the WR3 surface (Chadwick et tional erosion of the clast (Phillips et al., al., 1994). They sampled the crests of four 1990; Chadwick et al., 1994); and (3) the Results large (several metres diameter) Precam- quartzites are derived from formations 26Al and 10Be results from the laterally brian metamorphic boulders on the terrace whose ages are several orders of magnitude extensive terrace FR3, 90 m above the surface, which minimize postdepositional greater than the half-lives of either nuclide, floodplain, are in very good agreement for erosion by ventifaction and fire spall (e.g., so that Npre is safely neglected. the single surface clast samples (Table 1). Bierman and Gillespie, 1991; Zimmerman In the field, we document the latitude and The spread of effective exposure ages (Teff ϭ et al., 1994). Three of the boulders had ex- elevation of the sample, from which the flat- 1/␭ ⅐ ln[Po/PoϪ␭Ns] taking into account de- posure ages of ϳ115 ka, the other had an surface production rate is calculated (e.g., cay) estimated from single surface clasts on age of 290 ka. They concluded that the age Lal, 1991), and measure the angle to the ho- this and other terraces (Repka et al., 1994) of the surface is 115 ka, interpreting the 290 rizon, ␪, in eight directions in order to cal- is very large (e.g., 130 ka [n ϭ 7] for FR2 and ka age to reflect a significant prior exposure culate the topographic shielding factor (Lal, 25 ka [n ϭ 3] for FR3), indicating that in- history for that particular boulder. This 1991; Cerling and Craig, 1994). Subsurface heritance is indeed significant. Given the places the terrace and associated moraine in sampling depth and thickness of the silt cap small number of single clasts analyzed, the isotope stage 5. are noted. In the lab, ϳ30 clasts are crushed true age spread of individual clasts is prob- We collected quartzite cobbles (several to

GEOLOGY, January 1996 49 20 cm in diameter) from the WR3 terrace from the top of the silt layer, and subsurface idence of bioturbation within the gravels; roughly 65 km downstream of the site sam- cobbles were collected from gravel exposed depositional imbrication is still intact, and pled by Phillips and Zreda (Chadwick et al., in soil pits (Table 1). Surface clasts did not carbonate coatings are confined to clast 1994). The surface clasts were collected display carbonate coatings. There is no ev- bases.

Results 26Al and 10Be analyses of our surface amalgamated sample (Table 1) yield effec- tive exposure ages of 97 Ϯ 6 ka (Al) and 95 Ϯ 5 ka (Be). These estimates are close to three of the four boulder ages reported in Chadwick et al. (1994). However, using the subsurface sample to correct the surface age for inheritance (equation 4) results in age estimates of 76 Ϯ 8 ka (Al) and 67 Ϯ 7ka (Be). These estimates overlap at ϳ72 ka.

DISCUSSION The downward revision of the age esti- mate required by the measured inheritance is dramatic. The FR3 age is revised down- ward to ϳ71 ka, potentially associating this terrace with isotope stage 4, and doubling the implied incision rate of the Fremont River. The WR3 age becomes ϳ72 ka, strongly supporting the inference, based upon morainal soil development, of a Wind River Mountains glaciation between the classic Pinedale and Bull Lake glaciations (Hall and Shroba, 1995). These calculated terrace ages are depend- ent upon the accretion history of the silt cap, and the postdepositional vertical motion of clasts within it. In our initial calculation, we assumed that the silt was deposited very near the time of abandonment of the sur- face, and that the sampled surface clasts reached the surface (by whatever means) shortly thereafter and have remained on the surface. Several other scenarios must be en- tertained (Fig. 3). The rise of clasts through the silt may have taken substantially more Figure 3. A: Cosmogenic radionuclide (CRN) histories for two possible postdepositional sce- time, perhaps through upfreezing under narios, resulting in identical measured CRN concentrations of surface and subsurface samples, periglacial conditions (Anderson, 1988). On Ns and Nss. Time scales here are assumed to be short relative to decay times for these radio- nuclides. In first ‘‘eolian inflation’’ case (light lines, solid for surface sample, dashed for sub- the other hand, the silt could be eolian, and surface; circle depicts inherited concentration and time of surface abandonment), clasts sam- the monolayer of clasts remains on the sur- pled at surface have remained on surface throughout, riding upward on growing silt layer face as a desert pavement during slow eolian through desert pavement processes. Their production rate is always P , as denoted by slope o accumulation (Wells et al., 1995). Both of line. Eolian inflation progressively buries subsurface sample, reducing rate of production cases revise the postdepositional CRN pro- through time, reaching final value appropriate for its sampled depth, zss (see inset). In second ‘‘upfreezing’’ case (heavy lines; box depicts inherited concentration and time of surface aban- duction histories of clasts. Numerical model donment), mantling silt is taken to be fluvial, and does not further inflate through time. CRN results (Repka et al., 1994) show that for the accumulation in subsurface sample is therefore steady, at rate P exp( z /z*). Clasts sampled -o ؊ ss silt caps on these terraces the necessary cor from surface are assumed to rise through silt layer, emerging at surface soon before sampling. rection due to the increasing depth of the They undergo increasing rates of production, eventually reaching rate Po. Back projection of production histories using steady production rates appropriate for sampled depths intersect at subsurface sample is not large (order 5%), is star; in both cases, surface age is underestimated. B: Four possible postdepositional scenarios not sensitive to the details of the inflation in which silt is rapidly accumulated and remains of steady thickness, all of which yield same history, and always increases the estimated measured surface and subsurface CRN concentrations, N and N . CRN accumulation in sub- s ss surface age (Fig. 3A). Our estimate of the surface sample is slow but steady (dashed line), at rate dictated by sample depth, zss (see inset). If gravels constituting surface sample remain on surface essentially since deposition (early WR3 age would rise at most from ϳ72 ka to upfreezing, or desert pavement), then CRN history of such a sample is depicted by line 1, whose ϳ77 ka. slope is Po, and estimated age (star) is youngest. Case in which gravels remained at gravel-silt Uncertainties in the vertical motion his- interface until just prior to sampling (very late upfreezing), undergoing slowest possible pro- tories of clasts in the surface sample can duction rate (line 2), yields oldest age estimate (circle). Steady upfreezing yields monotonically increasing production rate (line 3), while bioturbation and cryoturbation result in rates that significantly affect age estimates, however (Fig. 3B). Whereas subsurface gravels have .(fluctuate erratically from Po to Po exp(؊zg/z*), zg being depth of gravel-silt interface (line 4 These result in intermediate surface age estimates (square and oval). not undergone relative movement (due to

50 GEOLOGY, January 1996 cryoturbation, bioturbation, upfreezing), long-term exhumation rates and transport Hallet, B., and Putkonen, J., 1994, Surface dating this is not certain for the surface clasts. The times that are otherwise very difficult to of dynamic landforms: Young boulders on greatest increase in the estimated age of the constrain. aging moraines: Science, v. 265, p. 937–940. Howard, A. D., 1970, Study of process and history surface occurs if (1) the silt is deposited rap- in desert landforms near the Henry Moun- idly (either fluvially or by rapid postaban- ACKNOWLEDGMENTS tains, Utah [Ph.D. thesis]: Baltimore, Mary- donment deposition of eolian silt), and (2) Supported by the Petroleum Research Fund of land, Johns Hopkins University, 198 p. the American Chemical Society, a grant from the the clasts resided at the base of the silt until Howard, A. D., 1986, Quaternary landform evo- Center for Accelerator Mass Spectrometry at lution of the Dirty Devil River system, Utah: very recently (Fig. 3B). Sampled surface Lawrence Livermore National Lab, and a Geo- Geological Society of America Abstracts clasts then undergo the lowest possible logical Society of America Cole Award to Ander- with Programs, v. 18, p. 641. mean postdepositional production rate. son. We thank O. Chadwick for discussing unpub- Kohl, C. P., and Nishiizumi, K., 1992, Chemical lished Wind River results and for suggesting This increases our estimated terrace age for isolation of quartz for measurement of in- sampling sites on WR3; F. Phillips for sharing his situ-produced cosmogenic nuclides: Geo- the WR3 terrace from 72 ka to 133 ka. In- 36 unpublished Cl results and for candid discussion chimica et Cosmochimica Acta, v. 56, termediate cases arise if the clasts are cycled of them; and C. Brauderick, A. Densmore, R. p. 3583–3587. through the silt via cryoturbation and bio- Finkel, J. Koning, C. Lundstrom, E. Small, M. Lal, D., 1991, Cosmic ray labeling of erosion sur- turbation (Fig. 3B). On WR3, this would al- Swartz, N. Humphrey, J. Harper, and J. Stock for faces: In situ nuclide production and erosion their help in the field, in the lab, or in reviewing ter the age estimate to roughly 105 ka. Com- models: Earth and Planetary Science Letters, this paper. We also thank S. Wells for a thorough v. 104, p. 424–439. binations of eolian silt accumulation and review. Nishiizumi, K., 1994, Cosmogenic production of upfreezing that keep clasts near the surface 10Be and 26Al on the surface of the earth and reduce the correction considerably. For ex- REFERENCES CITED underground, in Eighth International Con- Anderson, S. P., 1988, The upfreezing process: ference on Geochemistry, Cosmochronology ample, using a steady silt accumulation sce- Experiments with a single clast: Geological So- nario, the oldest the WR3 terrace could be and Isotope Geology, abstracts: U.S. Geo- ciety of America Bulletin, v. 100, p. 609–621. logical Survey Circular 1107, p. 234. is 84 ka. Bierman, P., 1994, Using in situ produced cosmo- Nishiizumi, K., Kohl, C. P., Arnold, J. R., Dorn, Although the lack of carbonate coatings genic isotopes to estimate rates of landscape R. I., Klein, J., Fink, D., Middleton, R., and on surface clasts argues against long resi- evolution: A review from the geomorphic Lal, D., 1993, Role of in situ cosmogenic nu- perspective: Journal of Geophysical Re- 10 26 dence in the subsurface, the wide range of clides Be and Al in the study of diverse search, v. 99, p. 13,885–13,896. geomorphic processes: Earth Surface Pro- possible surface clast production histories Bierman, P., and Gillespie, A. R., 1991, Range cesses and Landforms, v. 18, p. 407–425. lowers our confidence in the calculated sur- fires—A significant factor in exposure-age Phillips, F. M., Zreda, M. G., Smith, S. S., Elmore, face ages. Work in progress, in which nu- determination and geomorphic surface evo- D., Kubik, P. W., and Sharma, P., 1990, Cos- lution: Geology, v. 19, p. 641–644. mogenic chlorine-36 chronology for glacial merous amalgamated samples are taken at Billingsly, G. H., Huntoon, P. W., and Breed, depth within the gravels, is designed to dis- deposits at Bloody Canyon, eastern Sierra W. J., 1987, Geologic map of Capitol Reef Nevada: Science, v. 248, p. 1529–1532. tinguish among these silt scenarios, and to National Park and vicinity, Emery, Garfield, Repka, J. L., Anderson, R. S., and Dick, G. S., test the assumption that the gravels were Millard and Wayne counties, Utah: Salt Lake 1994, Rate constraints of river incision using rapidly deposited. City, Utah Geological and Mineral Survey, cosmogenic radionuclides: Examples from scale 1:62 500. The measured mean CRN inheritance, fluvial terraces along the Fremont River, Cerling, T. E., and Craig, H., 1994, Geomorphol- Utah, in Eighth International Conference on Nin, can constrain the transport times and ogy and in situ cosmogenic isotopes: Review Geochemistry, Cosmochronology and Iso- hillslope exhumation rates (Table 1). If the of Earth and Planetary Sciences, v. 22, tope Geology, abstracts: U.S. Geological measured inheritance in the FR3 (WR3) p. 273–317. Survey Circular 1107, p. 265. Chadwick, O., Hall, R., Conel, J., Phillips, F., samples was acquired entirely during trans- Trull, T. W., Brown, E. T., Marty, B., Raisbeck, Zreda, M., and Gosse, J., 1994, Glacial de- 10 port (N ϭ N ), inversion of equation 3, G. M., and Yiou, F., 1995, Cosmogenic Be in trans posits and river terraces in Wind River Basin, and 3He accumulation in Pleistocene beach assuming H/z* ϭ 10, yields ͗T͘ϭ548 ka in Chadwick, O., et al., eds., Quaternary ge- terraces in Death Valley, California, USA: (260 ka) (Table 1). If instead the entire in- ology of the Wind River Basin, Wyoming: Implications for cosmic-ray exposure dating heritance was acquired during exhumation Friends of the Pleistocene, Rocky Mountain of young surfaces in hot climates: Chemical Cell field guide: Pasadena, California Insti- Geology, v. 119, p. 191–207. (Nin ϭ Nexh), inversion of equation 2 implies tute of Technology, Jet Propulsion Labora- ˙ Wells, S. G., McFadden, L. D., Poths, J., and ⑀ ϭ 11 ␮m/yr (23 ␮m/yr). Because inheri- tory, p. 1–15. Olinger, C. T., 1995, Cosmogenic He-3 sur- tance is the sum of that obtained during Davis, J. C., and 10 others, 1990, Lawrence Liv- face-exposure dating of stone pavements— each of these processes, we can only state ermore National Laboratory–University of Implications for landscape evolution in California Center for Accelerator Mass deserts: Geology, v. 23, p. 613–616. that the calculated mean hillslope exhuma- Spectrometry facility and research program: tion rates are minima, and the mean trans- Zimmerman, S. G., Evenson, E. B., Gosse, J. G., Nuclear Instruments and Methods in Physics and Erskine, C. P., 1994, Extensive boulder port times are maxima. It is clear, however, Research, v. B52, p. 269–272. erosion resulting from a range fire on the that the two geomorphic delivery systems Elmore, D., and Phillips, F., 1987, Accelerator type-Pinedale moraines, Fremont Lake, are distinctly different. mass spectrometry for measurement of Wyoming: Quaternary Research, v. 42, long-lived radioisotopes: Science, v. 236, p. 255–265. Our technique provides an important ad- p. 543–550. vance in the dating of depositional surfaces Flint, R. F., and Denny, C. S., 1956, Quaternary Manuscript received April 17, 1995 while raising a cautionary flag. The inheri- geology of Boulder Mountain, Aquarius Pla- Revised manuscript received September 18, 1995 tance of CRNs is potentially significant and teau, Utah: U.S. Geological Survey Bulletin, Manuscript accepted September 25, 1995 v. 1061-D, p. 103–164. must be considered in estimating surface Hall, R. D., and Shroba, R. R., 1995, Soil evidence ages. Taking the youngest effective age of a for a glaciation intermediate between the few individual clasts to represent the age of Bull Lake and Pinedale glaciations at Fre- a surface is risky; all clasts are likely to have mont Lake, Wind River Range, Wyoming, inherited nuclides within the geomorphic U.S.A.: Arctic and Alpine Research, v. 27, p. 89–98. system. The inheritance is not simply noise to be removed; it can be used to quantify

GEOLOGY, January 1996 Printed in U.S.A. 51