The Null Hypothesis: Globally Steady Rates of Erosion, Weathering Fluxes and Shelf Sediment Accumulation During Late Cenozoic Mountain Uplift and Glaciation

Home , Erosion and tectonics, Sadler effect, Tectonic uplift

doi: 10.1111/ter.12185 Debate Article The null hypothesis: globally steady rates of erosion, weathering fluxes and shelf sediment accumulation during Late Cenozoic mountain uplift and glaciation

Jane K. Willenbring and Douglas J. Jerolmack Department of Earth and Environmental Science, University of Pennsylvania, 240 South 33rd Street, Philadelphia, PA 19104-6316, USA

ABSTRACT At the largest time and space scales, the pace of erosion and over the last 10 Ma and support the null hypothesis – that glo- chemical weathering is determined by tectonic uplift rates. bal rates of landscape change have remained constant over this Deviations from equilibrium arise from the transient response time period, despite global climate change and mountain build- of landscape denudation to climatic and tectonic perturbations. ing events. This work undermines the hypothesis that increased We posit that the constraint of mass balance, however, makes weathering due to mountain building or climate change was it unlikely that such disequilibrium persists at the global scale the primary agent for a decrease in global temperatures. over millions of years, as has been proposed for late Cenozoic erosion. We synthesize weathering fluxes, global sedimentation Terra Nova, 28: 11–18, 2016 rates, sediment yields and tectonic motions to show a remarkable constancy in the pace of Earth-surface evolution

FORUM: The Editors of Terra Nova invite readers to contribute to a discussion of this topic on our forum at https://terranova- debates.wordpress.com/

(Berner, 1991; Berner and Caldeira, rates (Herman et al., 2013). Introduction 1997; Maher and Chamberlain, 2014) Researchers have attempted to cau- The evolution of the Earth’s surface and physical erosion (which causes sally link these trends to global cool- occurs through the erosion of rock unloading that can induce a tectonic ing with two related hypotheses. In and redistribution of mass. This is response; Montgomery, 1994; Whip- the first, rapid erosion associated ultimately driven by energy from two ple, 2009). In order to unravel these with Himalayan uplift acted to drive sources: tectonics, which contributes feedbacks and linkages, many Cenozoic cooling by reducing CO2 to erosion indirectly through uplift; researchers have turned to global levels through increased silicate and climate, which influences erosion datasets of the relevant parameters. weathering (Raymo and Ruddiman, and weathering via temperature and Of particular interest has been the 1992). The second reverses the arrow precipitation. Global changes in tec- late Cenozoic period, when a global of causality: in this case it is argued tonic motions or climate might be cooling trend, of unknown origin, climatic cooling and glacial-intergla- expected to produce global changes began at 50 Ma and culminated in cial swings acted to increase denuda- in the rate of landscape evolution; continental glaciations during the last tion rates, perhaps even inducing however, feedbacks among climate, 2–3 Ma in Eurasia and North and enhanced mountain uplift (Molnar tectonics and geomorphology may South America (Ruddiman, 2010). and England, 1990; Zhang et al., enhance or obscure the relation Global rates of tectonic motion have 2001; Molnar, 2004). between ‘signal’ and ‘response’. For been either constant over this inter- The climatic cooling trend of the example, although the engine for val, as determined from ages of crust late Cenozoic is not in dispute. We plate tectonics is mantle convection, (Fig. 1; Rowley, 2002, 2013), or challenge the premise that global researchers have proposed that rates decreasing, as determined from slab rates of landscape denudation have of tectonic uplift may depend and flux models that incorporate the changed, however, on both concep- feedback on global climate through geometry of the subduction zone tual and empirical grounds. There silicate weathering (which draws (Van Der Meer et al., 2014). Obser- are numerous possibilities for feed- down CO2 and leads to cooling) vations also point to an apparent backs and linkages among climate, increase (c.f. Sadler, 1981) in land- tectonics, erosion and weathering. scape denudation over the same time However, mass balance places a con- Correspondence: Dr. Jane Kathryn Wil- period, as indicated by global sedi- straint on the magnitude and persis- lenbring, Department of Earth and Envi- ment accumulation curves from the tence of denudation rates. Enhanced ronmental Science, University of world’s oceans and sedimentary weathering and erosion reduces Pennsylvania, 240 South 33rd Street, basins (Hay et al., 1988; Molnar, slopes and grows valleys, which act Philadelphia, PA 19104-6316, USA. Tel.: 2004) and, more recently, a global to decrease denudation back towards +001 215 746-8197; e-mail: erosion@sas. compilation of mountain erosion being in balance with tectonic uplift upenn.edu

Crust production these results for understanding feed- production (km

mean rate of crust backs between climate and geomor-

) phology are discussed.

–1 27 3.2

26 3.0 Evidence from geochemical 2 25 2.8 year proxies –1 rate (mm year rate mean half-spreading mean half-spreading 0 ) Recent proxies suggest stabilization 0246810 of atmospheric CO2 at near-modern Age (My) levels since the late Oligocene, despite continued global cooling Fig. 1 Rates of tectonic forcing over the last 10 Ma are steady. Figure adapted (Beerling and Royer, 2011; Fig. 2A). from Rowley (2002, 2013) and data digitized from Rowley (2002; Fig. 4). This stationary greenhouse gas composition implies nearly constant chemical weathering over the last (Whipple et al., 1999). As with many ple question of how climate impacts 10–20 Ma, unless there were signifi- problems, it is an issue of scale. landscape probably has as many cant changes in the organic carbon Large regions of the globe may sus- answers as there are landscapes. cycle (Pagani et al., 2009) that some- tain denudation rates that are far The simplest, that is, null, hypoth- how compensated for changes in from being in equilibrium with tec- esis is that global rates of erosion, weathering. Several recent hypotheses tonics, but for only limited periods weathering and sediment accumula- attempting to relate global cooling to of time. Alternatively, at-a-point tion have been constant over the the Northern Hemisphere glaciation rates of denudation may be much time interval in question. This con- have therefore abandoned increased larger than tectonic rates over long stancy is expected from the mass weathering fluxes as a cause (Knies periods of time, but are compensated conservation constraint, given that et al., 2014; Woodard et al., 2014). by lower rates elsewhere. To make global rates of tectonic motion have Other proxies brought to bear on this the case for sustained and enhanced not varied over the late Cenozoic issue have ambiguities that limit our denudation rates globally over sev- (Rowley, 2002, 2013). Ultimately, the interpretive ability. In the past, the eral millions of years, one needs to long time-scales needed to affect glo- rising late Cenozoic 87Sr/86Sr record demonstrate that: (1) erosion and bal heat production, and the correla- was thought to be an indicator of weathering rates are somehow decou- tion of mantle viscosity to mantle increased continental weathering pled from tectonics at the largest temperature, preclude any global (Raymo et al., 1988; Capo and scales; and (2) there is a convincing changes in tectonic rates. We do not DePaolo, 1990). However, this record and causal relation between climate have sufficient knowledge of the has been shown to be sensitive to the and denudation rates at the global mechanisms of weathering, erosion, isotopic ratio of the sedimentary scale. Regarding (1), while numerous geodynamics and climate to address source area (Richter et al., 1992) and studies have reported disequilibrium competing hypotheses from a theo- to the proportion of physical to between erosion and tectonics over retical/modelling perspective. Rather, chemical weathering (Derry and shorter timescales, models and data we must rely on empirical findings, France-Lanord, 1996). Similarly, Nd show a convergence of the two over which fall into three main categories. and Pb isotopes indicate only a timescales greater than 106 years The first approach is finding a geo- change in weathering style or source, (Koppes and Montgomery, 2009; chemical proxy that indirectly not a total weathering flux (Willen- Whipple, 2009), although transience records processes such as the flux of bring and von Blanckenburg, 2010a). may also persist for millions of years erosion or weathering products to Two proxies that are thought to (Willenbring et al., 2013). Further, the oceans and integrates over large describe the total weathering flux are for studies that infer denudation areas (continents, ocean basins or the 10Be/9Be ratio and the d7Li from basin-integrating (i.e. spatially global). The second approach is mea- record. 10Be/9Be systematics are rela- averaging) sediment yields, global suring erosion or weathering rates tively simple, thus may finally con- rates do not show any significant (or total denudation rate) in as many strain the flux of weathering trend over timescales longer than places as possible and trying to products to the ocean. This ratio is a 101 years (Koppes and Montgomery, extrapolate to unmeasured areas. combination of 9Be that is released 2009; Sadler and Jerolmack, 2014). The third approach is to use the sedi- from weathering and cosmogenic As for (2), global compilations using mentary record as a proxy for ero- 10Be, which acts as a constant flux cosmogenic radionuclides fail to sion and weathering loads. Below, tracer over time periods long enough reveal any relation between denuda- we show how converging lines of evi- to average several geomagnetic field tion rates and temperature or precip- dence from these three approaches cycles. Over the last 10 Ma, this itation (Portenga and Bierman, provide strong support for steady ratio is invariant from ocean basin 2011). Moreover, a landscape evolu- rates of global landscape evolution (Atlantic) (Fig. 2B) to ocean basin tion model that can realistically per- over the Cenozoic, and find that (Pacific) (Fig. 2C; Willenbring and turb a landscape via climate has not observed rates are remarkably con- von Blanckenburg, 2010a), which yet been produced. This topic is an stant over the last 10 million years. precludes the ~2–4-fold increase in active research area, and the not-sim- The significance and implications of weathering flux inferred from records

(A) Atmospheric CO2 for a modest (20%) increase in the weathering ﬂux. Moreover, since 9Be does not directly measure erosion, it 600 is indeterminate regarding an erosion rate increase. The 9Be weathering 400 Phytoplankton (ppmv) proxy is also not sensitive to basalt 2 Stomata or carbonate weathering, and the CO 200 Boron 10Be ﬂux is assumed to be quasi- 0 constant but also depends on the 0246810 ocean basin where the samples are Age (My) taken because of latitudinal variations in the delivery of 10Be from the (B) Arctic and Atlantic Ocean - 10Be/9Be atmosphere (Willenbring and von 7 ) d 2 Blanckenburg, 2010b). The Li val- –7 ues over the last 10 Ma are invariant 10 x 1.5 within the range of variability and

Be ( Be uncertainty in the measurements 9 1 (Fig. 2D). Interpreting how the d7Li

Be/ Modern record relates to chemical weathering 10 0.5 Atlantic Ocean ﬂuxes is more complex (Misra and Froelich, 2012; Torres et al., 2014).

Initial Initial 0 Nonetheless, the 10Be/9Be ratio (Wil- 02468 10 lenbring and von Blanckenburg, Age (My) 2010a) and the original interpretation d7 (C) - 10Be/9Be of the Li record (Misra and Froe- lich, 2012) support the view from ) 2 –7 palaeo-CO2 proxies that there has

10 been no large change in the silicate x 1.5 weathering ﬂux from the continents Modern Be ( Be

9 over the last 10 Ma. 1 Ocean Be/ 10 0.5 Measuring erosion, weathering and denudation

Initial Initial 0 0246810 The apparent increase in global Age (My) denudation rates is coincident with the onset of northern hemispheric (D) Whole Ocean δ7Li glaciation. Enhanced glacial erosion 34 in response to global cooling has thus been invoked as the major process

) 32

oo responsible, fuelled by a much-cited / o 30 view that glaciers can be more effec- Li ( Whole tive erosive agents than rivers (Hallet

7 Ocean δ 28 Range et al., 1996), even limiting the height of mountains regardless of tectonics 26 (the glacial buzzsaw hypothesis; Bro- 0246810 zovic et al., 1997; Mitchell and Mont- Age (My) gomery, 2006). This climate-driven erosion hypothesis has the potential to Fig. 2 (A) Palaeo-CO2 proxy reconstructions from Beerling and Royer (2011) 10 9 feed back to long-term climate: (excluding paleosol proxies) are essentially steady over the last 10 Ma. (B) Be/ Be increased denudation and glacial from ferromanganese crusts and dated sediments from the Arctic and Atlantic 10 9 grinding of debris creates fresh min- Oceans (adapted from Willenbring and von Blanckenburg, 2010a). (C) Be/ Be eral surfaces, accelerating weathering from ferromanganese crusts and dated sediments from the Pacific Ocean (adapted rates and thus causing the sequestra- from Willenbring and von Blanckenburg, 2010a). Records in (b and c) were cor- tion of atmospheric CO , which fur- rected for radionuclide decay of 10Be over time and show no systematic variation 2 over the last 10 Ma. (D) Whole ocean d7Li records for the last 10 Ma are essen- ther decreases temperatures. Data tially steady. Figure adapted from Misra and Froelich (2012). show, however, that while glacial cover may induce a rapid pulse of erosion, it is not sustained (Koppes and of sediment accumulation (Zhang ties in the behaviour of beryllium at Montgomery, 2009). Further, only et al., 2001; Molnar, 2004). However, the neutral pH of the oceans (Boschi some climates produce high rates of the scatter in the plot and uncertain- and Willenbring, 2013) would allow glacial erosion (Yanites and Ehlers,

2012); rates may be diminished at high measured rates, results from their sta- was that tectonics, and not climate, is elevations due to intense cold and lack tistical model are biased by not the ultimate driver of landscape ero- of moisture, which lead to ice-shed- accounting for the influence of im- sion rates over long timescales. ding ‘Teflon peaks’ (Anderson, 2005; measurably slow rates. Naylor et al. Ward et al., 2012). Finally, recent (2015) have demonstrated the effect Sediment generation and research indicates that pre-Quatern- of this bias for detrital ther- deposition ary topography – and not glacial ero- mochronologic datasets. The essence sion – sets the elevation of high of the argument is that slower rates The primary piece of evidence that plateaus (Van Der Beek et al., 2009), of exhumation are progressively past researchers have employed to and that correlations between the clipped from the measured age distri- infer a global late Cenozoic increase occurrence of glaciers and mountain bution of rocks as one approaches in erosion and weathering is the piles hypsometries are merely coincidental the modern because of the precision of sediment accumulated in the (Hall and Kleman, 2013). of the technique. The reason is that oceans (Hay et al., 1988; Metivier Cosmogenic nuclide data support thermochronometers measure the et al., 1999; Zhang et al., 2001; Mol- the idea of fleeting fluxes of sediment time since achieving an associated nar, 2004) (and, to a lesser extent, on from glacial erosion (Charreau et al., closure depth; rate is simply the mea- land (Kuhlemann, 2000)). As pointed 2011); thermochronologic data sured time divided by this depth. If out by Schumer and Jerolmack demonstrate that over the long peri- erosion rates have not been high (2009), however, the apparent accu- ods of time recorded by exhumation enough to bring rocks to the surface, mulation rates determined from sedi- rates, glaciers do protect some land- the rates associated with those buried ment piles do not reflect true scape surfaces (Thomson et al. 2010). rocks cannot be measured (Fig. 3). sediment accumulation rates and can- Only approximately half of palaeo- While the full distribution of not be used as direct evidence for denudation records show an exhumation rates can be measured increasing global denudation rates increased rate during the last 2 Ma for rocks older than, say, 20 Ma, over the late Cenozoic. The confusion compared to pre-Quaternary time slower rates are progressively clipped in interpreting sediment accumulation (Granger and Schaller, 2014). Data from the measurable distribution for arises from the now notorious ‘Sadler also show that in the Antarctic, Arc- younger and younger ages. Based on effect’: due to the unsteady nature of tic and sub-Arctic, glaciation may be the Naylor et al. (2015) model, Sin- sediment deposition, rates of accumu- geomorphically protective on high clair et al. (in review) found that, in lation appear to decrease with peaks and flat plateaus (Staiger the western Alps, the effect becomes increasing measurement interval et al., 2005; Briner et al., 2006), and pronounced for rocks in the range of (Sadler, 1981, 1994). This effect arises glacial erosion may recycle past gla- 3–5 Ma. If the mean of measured from a bias that is not unlike the clip- cial sediment rather than erode rock exhumation rates is assumed to rep- ping effect described above for ther- (Roy et al., 2004; Ebert et al., 2012). resent the true mean erosion rate, mochronologic data. Sediment In summary, enhanced erosion in then any thermochronologic dataset accumulation rates in the modern are some places would be at least par- could (mistakenly) suggest an accel- only measured where sediment is tially compensated for by inhibited eration in erosion rate that is actu- accumulating, and therefore preclude erosion in others. ally an artefact of the technique. In negative or zero values from being Herman et al. (2013) recently com- Fig. 3, we plot a variation of an idea incorporated; as one averages over piled a global thermochronologic of Anders et al. (1987) that describes longer and longer time intervals, mea- dataset to explore the idea of glacial two clipped distributions of sedimen- sured rates incorporate intervals of vs. non-glacial erosion, by comparing tation rate data both in the maxi- non-deposition and erosion. Because cooling rates in the 0–2 Ma time mum depth of a population sampled measurement interval and age are slice to those in the 4–6 Ma time and in the precision possible for the inseparable (Sadler, 1981), the upshot slice and inverse modelling erosion youngest ages. We view this near- is an apparent increase in accumula- rates based on a set of assumptions. reproduction of the Anders et al. tion rates on approach to the present. These authors found that the most (1987) figure (intended for sedimen- For every type of accumulation or recent 2-Ma interval often had tation rates) applied to exhumation erosion one could measure, one sees exhumation rates >4 times higher rates as a thought experiment. the same kind of apparent increase than the 6–4 Ma interval, when com- Although that was not the original (Gardner et al., 1987; Sadler, 1994; paring areas covered by glaciers. intent of the figure, the conceptual Schumer and Jerolmack, 2009; Fin- They acknowledge a bias in the data- analogy is useful. The reader is negan et al., 2014). Theory shows us set where only sufficiently fast referred to Naylor et al. (2015) for a that these apparent trends are an exhumation rates would achieve the rigorous discussion of bias in ther- inevitable consequence of stochastic precision necessary to answer this mochronologic systems. variations in erosion and deposition, question. Another source of bias Finally, the data compilation from and likely do not reflect any real acknowledged is that the samples Koppes and Montgomery (2009) change in the mean rate through time were not randomly selected, but shows that overall rates of glacial ero- (Strauss and Sadler, 1989; Pelletier occur most often in orogenic belts. sion are comparable to fluvial values and Turcotte, 1997; Jerolmack and While the authors employed a over long timescales, and that both Sadler, 2007; Schumer and Jerol- sophisticated inversion technique to converge with tectonic uplift rates over mack, 2009; Schumer et al., 2011; account for errors and variability in millennial timescales. Their conclusion Finnegan et al., 2014).

Thermochronometer bias eroded in a given time interval, no measurement bias would exist. This body of work is often ignored in Thickness limit from closure depth 12 favour of the more ‘straightforward’ interpretation of a positive trend in curves of measured rates. 10 Recent work and summaries of ‘ the literature (Romans, in press) 2 8 2 state there is little expectation that sediment simply records changes in humat 2 ‘ forcing. Even if the climatic or tec- 6 t exhumation2 rate tonic regime has changed, the sedimentary record might remain

Exhumed thickness (km) Constan 4 constant. Diffusive processes on hillslopes and in other sedimentary bodies are very efﬁcient at ﬁltering 2 Limit of precision climatic oscillations because of slow signal propagation (Metivier et al., 1999; Furbish and Fagherazzi, 2001; 246 810 Armitage et al., 2013; Godard Age (My) et al., 2013). Also, the signal 4 response to climate or tectonic forcing is likely nonlinear – at least 3.5 over some timescales (Jerolmack

) and Paola, 2010). –1 3 A remarkable finding from global Observed 2.5 data compilations has recently exhumation emerged; however, that provides the rate 2 most convincing physical evidence Thickness limit Constant yet that global erosion rates have 1.5 2 ‘ exhumation been constant over the Cenozoic. 2 rate Sadler and Jerolmack (2014) show 1 Exumation rate (km My 2 that a newly compiled curve of 2 progradation (lateral migration) rates 0.5 ‘ Limit in continental shelf settings mirrors of precision 0 the compilation of aggradation (ver- 0246810 tical deposition) rates. The data Age (My) shown in Fig. 4 are derived from means of thousands of individual Fig. 3 Figures (thickness – top; rate – bottom) adapted from Anders et al. (1987). measurements from all over the Illustration of a potential bias in thermochronometers by filtering of exhumation his- globe. Based on mass balance, the tories (idea by Sinclair et al., in review; Naylor et al., 2015) by the thickness limit product of these two curves repre- from the maximum closure depth and the limit of precision of the method. For the sents the global sediment flux depos- precision limit, in order to resolve an erosion/exhumation rate using a given ther- ited on shelves, and overcomes many mochronometer, there must have been sufficient exhumation since the passage of the of the problems of 1D measure- rock through the depth of effective mineral closure. Therefore, the younger the time ments. The data show that this flux period of interest, the higher the exhumation rate necessary to exhume the rock to is constant over timescales from the surface to be measured. Due to the vagaries of thermochronometry, the only years to tens of millions of years rates that are recordable during the last few millions of years are those with higher (Fig. 4). This study also presented values. Slow rates with insufficient time to reach the surface are clipped for recent the most extensive compilation yet of times but not clipped for long time periods. The ‘limit of precision’ denotes a zone upland denudation rates determined of unmeasurable rates. The variance in the dataset is either zero as in the case of – ‘constant exhumation rate’ or increased to 2r and 2rʹ are the variability around the from sediment yields which, impor- measurements (see top panel). The running median value is the ‘observed exhuma- tantly, are spatially integrated mea- – tion rate.’ All rates should be constant, but because of the selective inclusion of only surements and these data also fast rates for recent times and a limit on the thickness possibly exhumed ‘thickness indicate constant rates over a similar limit’, the observed exhumation rate increases (black dots) in the lower figure. If time span. In sum, when denudation these clipped values are assumed to represent the rates for the range as a whole, then and accumulation rate measurements it would lead to an observed acceleration in rates for the last 2 Ma. incorporate appropriate spatial averaging, they do not support any sig- The ‘Sadler effect’ bias arises due erosion measurements (Sadler and nificant increase in the pace of to the primarily one-dimensional Jerolmack, 2014). If we knew the landscape evolution over the late (1D) nature of accumulation and volumes of sediment deposited and Cenozoic. Instead, global denudation

Sedimentation catchment response. Earth Planet. Sci. Lett., 369, 200–210. Beerling, D.J. and Royer, D.L., 2011. Convergent Cenozoic CO2 history. –1 1 Nature Geosci., 4, 418–420. Berner, R.A., 1991. A model for

2 0 atmospheric CO2 over Phanerozoic time. Am. J. Sci., 291, 339–376. –1 Berner, R.A. and Caldeira, K., 1997. The Rate) (m year ) Rate) (m

log (Areal growth need for mass balance and feedback in the geochemical carbon cycle. Geology, 0246810 25, 955–956. log (Averaging time) (year) Boschi, V. and Willenbring, J.K., 2013. The behavior of beryllium in soils and Fig. 4 Sediment flux, expressed as volume discharge per unit width of the transport acquatic environments. Mineral. Mag., À system (i.e. cross-sectional growth rate, m2 a 1), does not vary systematically from 77(5), 744. timescales of years to one hundred million years (Sadler and Jerolmack, 2014). Briner, J.P., Miller, G.H., Davis, P.T. and Large rectangles plot growth rates for shelf systems; small rectangles plot growth Finkel, R.C., 2006. Cosmogenic rates for the shore zone and floodplain systems. These growth rates are products of radionuclides from fiord landscapes aggradation and progradation components determined separately from the mean of support differential erosion by thousands of empirical values from a global database (described in more detail and overriding ice sheets. Geol. Soc. Am. 118 – adapted from fig. 10 in Sadler and Jerolmack, 2014). Bull., , 406 420. Brozovic, N., Burbank, D.W. and Meigs, rates have been remarkably constant strongly to tectonics. The constancy A.J., 1997. Climatic limits on landscape for tens of millions of years. of denudation rates does not diminish development in the northwestern 276 – the importance of feedbacks among Himalaya. Science, , 571 574. Capo, R. C. and DePaolo, D. J., 1990. climate, tectonics, erosion and weath- Conclusions Seawater strontium isotopic variations ering; indeed, these feedbacks are from 2.5 million years ago to the A suite of geochemical proxies indi- likely responsible for returning present. Science, 249(4964), 51–55. cate that global weathering rates denudation rates to be in balance Charreau, J., Blard, P.H., Puchol, N., have not changed significantly over with tectonic forcing following a dis- Avouac, J.P., Lallier-Verges, E., the last 10 Ma (Willenbring and von turbance. In the future, careful atten- Bourles, D. and Roy, P., 2011. Paleo- Blanckenburg, 2010a; Beerling and tion to how measured rates converge erosion rates in Central Asia since Royer, 2011; Misra and Froelich, to constant values as a function of 9 Ma: a transient increase at the onset 2012). Compilations of global spatial and/or temporal averaging of Quaternary glaciations? Earth 304 – denudation rates determined from may reveal new information Planet. Sci. Lett., (1), 85 92. Derry, L.A. and France-Lanord, C., global sediment accumulation rates regarding the nature of these feed- 1996. Neogene Himalayan weathering determined from 2D data also indi- backs, and provide important con- history and river 87Sr/86Sr: impact on cate no significant changes over this straints for geodynamic models. the marine Sr record. Earth Planet. Sci. time interval. In fact, given the Lett., 142(1), 59–74. uncertainties and noise in the mea- Ebert, K., Willenbring, J., Norton, K., surements, the data ‘curves’ are Acknowledgement Hattestrand,€ C. and Hall, A., 2012. 10Be remarkably flat. In addition, global We acknowledge the NSF grant EAR- inventories from northern Sweden: compilations of cosmogenic radionu- 1331841 and thank Peter Sadler, Bruce implications for dating till and saprolite. Quatern. Geochron. 12 – clide data fail to show any clear rela- Wilkinson, Friedhelm von Blanckenburg, , ,11 22. Finnegan, N.J., Schumer, R. and tion between climate and denudation Shanan Peters, and Hugh Sinclair for Finnegan, S., 2014. A signature of ideas and discussions. We thank two (Portenga and Bierman, 2011). transience in bedrock river incision anonymous reviewers and Simon Brockle- Importantly, these results are robust rates over timescales of 104-107 years. hurst for their constructive reviews. empirical trends and are not inferred Nature, 505(7483), 391–394. from any model, and are consistent Furbish, D.J. and Fagherazzi, S., 2001. with expectations from mass balance. References Stability of creeping soil and Application of Occam’s razor sug- implications for hillslope evolution. gests that the null hypothesis is the Anderson, R. S., 2005. Teflon Peaks: The Water Resour. Res., 37, 2607–2618. evolution of high local relief in Gardner, T.W., Jorgensen, D.W., most viable. glaciated mountain ranges. AGU Fall A steady rate of global denudation Shuman, C. and Lemieux, C.R., 1987. Meeting Abstracts 1, 04. Geomorphic and tectonic process rates: provides an important baseline for Anders, M.H., Krueger, S.W. and Sadler, effects of measured time interval. regional studies, and implies that P.M., 1987. A new look at Geology, 15(3), 259–261. changes in global climate are accom- sedimentation rates and the Godard, V., Tucker, G.E., Burch Fisher, modated at the largest scales by completeness of the stratigraphic G., Burbank, D.W. and Bookhagen, B., changes in landscape form rather record. J. Geol.,1–14. 2013. Frequency-dependent landscape than the rate of landscape change. At Armitage, J.J., Dunkley Jones, T., Duller, response to climatic forcing. Geophys. appropriately averaged time and R.A., Whittaker, A.C. and Allen, P.A., Res. Lett., 40(5), 859–863. 2013. Temporal buffering of climate- space scales, it appears that rates of Granger, D.E. and Schaller, M., 2014. driven sediment flux cycles by transient landscape change are tied most Cosmogenic nuclides and erosion at the

watershed scale. Elements, 10(5), 369– Molnar, P., 2004. Late Cenozoic increase denudation and progradation rates: the 373. in accumulation rates of terrestrial case for time-scale invariance at Hall, A.M. and Kleman, J., 2013. Glacial sediment: how might climate change sediment sources and sinks. Geol. Soc. and periglacial buzzsaws: fitting have affected erosion rates? Annu. Rev. London Spec. Publ., 404, SP404. metaphors and mechanisms? Quatern. Earth Planet. Sci., 32,67–89. Schumer, R. and Jerolmack, D.J., 2009. Res., 81, 189–192. Molnar, P. and England, P., 1990. Late Real and apparent changes in sediment Hallet, B., Hunter, L. and Bogen, J., Cenozoic uplift of mountain ranges and deposition rates through time (2003– 1996. Rates of erosion and sediment global climate change: chicken or egg? 2012). J. Geophys. Res.-Earth Surface, evacuation by glaciers: a review of field Nature, 346,29–34. 114, F3. data and their implications. Global Montgomery, D.R., 1994. Valley Incision Schumer, R., Jerolmack, D. and Planet. Change, 12(1), 213–235. and the Uplift of Mountain Peaks. J. McElroy, B., 2011. The stratigraphic Hay, W.W., Sloan, J.L. and Wold, C.N., Geophys. Res., 99, 13913–13921. filter and bias in measurement of 1988. Mass Age Distribution and Naylor, M., Sinclair, H.D., Bernet, M., geologic rates. Geophysical Research Composition of Sediments on the van der Beek, P. and Kirstein, L., 2015. Letters, 38, 11. Ocean-Floor and the Global Rate of Bias in detrital fission track grain-age Sinclair, H. D., Naylor, M., Kirstein, Sediment Subduction. J. Geophys. Res.- populations: implications for L.A., Carter, A., Stanton, S., Solid Earth, 93(B12), 14933–14940. reconstructing changing erosion rates. Glotzbach, C., Bernet, M. and van der Herman, F., Seward, D., Valla, P.G., Earth Planet. Sci. Lett., 422,94–104. Beek, P., in review. Decreasing late Carter, A. , Kohn, B., Willett, S.D. and Pagani, M., Caldeira, K., Berner, R. and Cenozoic erosion rates in the Western Ehlers, T.A., 2013. Worldwide Beerling, D.J., 2009. The role of Alps. acceleration of mountain erosion terrestrial plants in limiting atmospheric Staiger, J.K.W., Gosse, J.C., Johnson, under a cooling climate. Nature, 504, CO2 decline over the past 24 million J.V., Fastook, J., Gray, J.T., Stockli, 423–426. years. Nature, 460(7251), 85–94. D.F., Stockli, L. and Finkel, R., 2005. Jerolmack, D.J. and Paola, C., 2010. Pelletier, J.D. and Turcotte, D.L., 1997. Quaternary relief generation by Shredding of environmental signals by Synthetic stratigraphy with a stochastic polythermal glacier ice. Earth Surf. sediment transport. Geophys. Res. Lett., diffusion model of fluvial Process. Landf., 30, 1145–1159. 37, L19401. sedimentation. J. Sediment. Res., 67(6), Strauss, D. and Sadler, P.M., 1989. Jerolmack, D. and Sadler, P., 2007. 1060–1067. Stochastic models for the completeness Transience and persistence in the Portenga, E.W. and Bierman, P.R., 2011. of stratigraphic sections. Math. Geol., depositional record of continental Understanding Earth’s eroding surface 21(1), 37–59. margins. J. Geophys. Res., 112, F03S13. with 10-Be. GSA Today, 21(8), 4–10. Thomson, S.N., Brandon, M.T., Tomkin, Knies, J., Mattingsdal, R., Fabian, K., Raymo, M.E. and Ruddiman, W.F., J.H., Reiners, P.W., Vasquez, C. and Grøsfjeld, K., Baranwal, S., Husum, 1992. Tectonic forcing of late Cenozoic Wilson, N.J., 2010. Glaciation as a K., De Schepper, S., Vogt, C., climate. Nature, 359, 117–122. destructive and constructive control on Andersen, N., Matthiessen, J., Raymo, M.E., Ruddiman, W.F. and mountain building. Nature, 467(7313), Andreassen, K., Jokat, W., Nam, S.-I. Froelich, P.N., 1988. Influence of late 313–317. and Gaina, C., 2014. Effect of early Cenozoic mountain building on ocean Torres, M.A., West, A.J. and Li, G., Pliocene uplift on late Pliocene cooling geochemical cycles. Geology, 16(7), 2014. Sulphide oxidation and carbonate in the Arctic-Atlantic gateway. Earth 649–653. dissolution as a source of CO2 over Planet. Sci. Lett., 387, 132–144. Richter, F., Rowley, D.B. and DePaulo, geological timescales. Nature, 507 Koppes, M.N. and Montgomery, D.R., D.J., 1992. Sr isotope evolution of (7492), 346–349. 2009. The relative efficacy of fluvial and seawater: the role of tectonics. Earth Van Der Beek, P., Van Melle, J., Guillot, glacial erosion over modern to orogenic Planet. Sci. Lett., 109,11–23. S., Pêcher, A., Reiners, P.W., timescales. Nature Geosci., 2(9), 644– Rowley, D.B., 2002. Rate of plate Nicolescu, S. and Latif, M., 2009. 647. creation and destruction: 180 Ma to Eocene Tibetan plateau remnants Kuhlemann, J., 2000. Post-collisional present. Geol. Soc. Am. Bull., 114(8), preserved in the northwest Himalaya. sediment budget of circum-Alpine 927–933. Nat. Geosci., 2, 364–368. basins (Central Europe). Mem. Sci. Rowley, D.B., 2013. History of Global Van Der Meer, D.G., Zeebec, R.E., van Geol. Padova, 52,1–91. Mean Spreading Rate: 83 Ma to Hinsbergena, D.J.J., Sluijsa, A., Maher, K. and Chamberlain, C.P., Present. Geol. Soc. of Amer. Abstr. Spakmana, W. and Torsvik, T.H., 2014. Hydrologic regulation of Prog., 45(7), 87. 2014. Plate tectonic controls on chemical weathering and the geologic Roy, M., Clark, P.U., Raisbeck, G.M. and atmospheric CO2 levels since the carbon cycle. Science, 343(6178), Yiou, F., 2004. Geochemical constraints Triassic. PNAS, 111, 4380–4385. 1502–1504. on the regolith hypothesis for the middle Ward, D. J., Anderson, R. S. and Metivier, F., Gaudemer, Y., Tapponier, P. Pleistocene transition. Earth Planet. Sci. Haeussler, P.J., 2012. Scaling the and Klein, M., 1999. Mass accumulation Lett., 227(3), 281–296. Teflon Peaks: rock type and the rates in Asia during the Cenozoic. Ruddiman, W.F., 2010. A paleoclimatic generation of extreme relief in the Geophys. J. Int., 137, 280–318. enigma. Science, 328(5980), 838–839. glaciated western Alaska Range. J. Misra, S. and Froelich, P.N., 2012. Sadler, P.M., 1981. Sediment Geophys. Res., 117, F01031, Lithium Isotope History of Cenozoic accumulation rates and the doi:10.1029/2011JF002068. Seawater: changes in Silicate completeness of stratigraphic sections. Whipple, K.X., 2009. The influence of Weathering and Reverse Weathering. J. Geol., 89, 569–584. climate on the tectonic evolution of Science, 335, 818–823. Sadler, P.M., 1994. The expected duration mountain belts. Nat. Geosci., 2(2), 97– Mitchell, S.G. and Montgomery, D.R., of upward-shallowing carbonate cycles 104. 2006. Influence of a glacial buzzsaw on and their terminal hiatuses. Geol. Soc. Whipple, K., Kirby, E. and Brocklehurst, the height and morphology of the Am. Bull., 106, 791–802. S., 1999. Geomorphic limits to Cascade Range in central Washington Sadler, P. and Jerolmack, D.J., 2014. climatically induced increases in State, USA. Quatern. Res., 65,96–107. Scaling laws for aggradation, topographic relief. Nature, 401,39–43.

Willenbring, J.K. and von Blanckenburg, Willenbring, J.K., Gasparini, N., Crosby, Yanites, B.J. and Ehlers, T.A., 2012. F., 2010a. Long-term stability of global B. and Brocard, G., 2013. What Does a Global climate and tectonic controls on erosion rates and weathering during Mean Mean? The temporal evolution the denudation of glaciated mountains. late-Cenozoic cooling. Nature, 465 of detrital cosmogenic denudation rates Earth Planet. Sci. Lett., 325–326,63–75. (7295), 211–214. in a transient landscape Geology, 41 Zhang, P., Molnar, P. and Downs, W.R., Willenbring, J.K. and von Blanckenburg, (12), 1215–1218. http://dx.doi.org/ 2001. Increased sedimentation rates and F., 2010b. Meteoric cosmogenic 10.1130/G34746.1 grain sizes 2–4 Myr ago due to the Beryllium-10 adsorbed to river Woodard, S.C., Rosenthal, Y., Miller, inﬂuence of climate change on erosion sediment and soil: applications for K.G., Wright, J.D., Chiu, B.K. and rates. Nature, 410, 891–897. Earth-surface dynamics. Earth-Sci. Lawrence, K.T., 2014. Antarctic role in Rev., 98(1), 105–122. doi: 10.1016/ Northern Hemisphere glaciation. Received 26 December 2014; revised j.earscirev.2009.10.008. Science, 346(6211), 847–851. version accepted 29 September 2015