Late Quaternary climatic controls on erosion rates and geomorphic processes in western Oregon, USA

Climate regulation of erosion in unglaciated landscapes remains difficult to decipher. While climate may disrupt process feedbacks that would otherwise steer landscapes toward steady erosion, sediment transport processes tend to erase past climate landforms and thus bias landscape evolution interpretations. Here, we couple a 50 k.y. paleoenvironmental record with 24 10Be-derived paleo-erosion rates from a 63-m-thick sediment archive in the unglaciated soil-mantled Oregon Coast Range. Our results span the forested marine oxygen isotope stage (MIS) 3 (50–29 ka), the subalpine MIS 2 (29–14 ka), and the forested MIS 1 (14 ka to present). From 46 ka through 28.5 ka, erosion rates increased from 0.06 mm yr–1 to 0.23 mm yr–1, coincident with declining temperatures. Mean MIS 2 erosion rates remained at 0.21 mm yr–1 and declined with increasing MIS 1 temperatures to the modern mean rate of 0.08 mm yr–1. Paleoclimate reconstructions and a frost-weathering model suggest periglacial processes were vigorous between 35 and 17 ka. While steady erosion is often assumed, our results suggest that climate strongly modulates soil production and transport on glacialinterglacial time scales. By applying a cosmogenic paleo-erosion model to evaluate 10Be concentrations in our sedimentary archive, we demonstrate that the depth of soil mixing (which is climate-dependent) controls the lag time required for cosmogenic erosion rates to track actual values. Our results challenge the widely held assumption that climate has minimal impact on erosion rates in unglaciated midlatitude terrain, which invites reconsideration of the extent to which past climate regimes manifest in modern landscapes. INTRODUCTION In unglaciated settings ranging from temperate to tropical, we are currently unable to accurately predict if climate change over glacialinterglacial intervals will increase or decrease erosion. As a result, the extent to which climate modulates landscape dynamics, such as river incision or aggradation, and invokes process thresholds that may promote new equilibrium states remains unresolved (Chorley et al., 1984; Tucker and Slingerland, 1997). Under steady uplift, geomorphic process feedbacks steer landscapes toward a dynamic equilibrium, such that erosion balances uplift over long time scales (Ahnert, 1994; Hack, 1975). However, it remains unclear how variations in precipitation or temperature patterns in unglaciated terrain disrupt landscape adjustment (Chorley et al., 1984) and produce transient signatures. The onset of increased climatic variability during the Pleistocene (Molnar, 2004) may have increased the frequency of landscape process perturbations such that repeated departures from steady-state adjustment led to accelerated sedimentation rates starting 4–2 Ma (Zhang et al., 2001). However, a preservation bias in the sediment accumulation record skewed toward younger deposits may negate this interpretation (Sadler, 1981; Schumer and Jerolmack, 2009; Willenbring and Jerolmack, 2015). While a global analysis of 18,000 bedrock thermochrono metric ages from mountainous regions suggests erosion rates have increased rapidly since ca. 2 Ma, even in unglaciated terrain (Herman et al., 2013), the extent to which this analysis is biased due to precision limits inherent to the thermochronologic method remains unresolved (Willenbring and Jerolmack, 2015). Further, these global-scale analyses have limited ability to infer process controls on climateerosion linkages, which is paramount for establishing robust and testable models of landscape dynamics. Without a mechanistic framework for the way in which climate controls processes and rates, it is difficult to parse how climate-driven changes in hillslope or fluvial processes influence landscape evolution. The onset of glaciation is viewed as an “abrupt and radical change” in a landscape’s erosional environment and history (Church and Ryder, 1972, p. 3059), with morphologic, modeling, and cosmogenic studies quantifying landscape response in terms of wider and deeper valleys, greater relief, and rapid denudation (Brocklehurst and Whipple, 2002; Herman and Braun, 2008; Montgomery, 2002). Despite recent advances, the legacy of past climates in unglaciated, soil-mantled settings is still difficult to discern, as topographic evidence such as solifluction lobes are likely turbated by biota over millennia and shielded from view in forested settings. Additionally, in tectonically active areas, soil residence times are short; modern processes quickly erase past signals, and sediment archives are rare. Our knowledge gap regarding the legacy of past climates in unglaciated terrain hampers progress in a broad array of problems in geomorphology and critical zone science, such as: quantifying fluxes of sediments and solutes, modeling landscape response to past and present climate change, and estimating the regulation of global CO2 by silicate weathering (Anderson et al., 2013; Dietrich and Perron, 2006; Dietrich et al., 2003; National Research Council, 2010). Cosmogenic nuclides, with applicability over time scales of 10 to 10 yr, overlap with the time scales over which rocks weather, soils form, climates cycle between glacial and interglacial, and rivers incise or aggrade (Granger and Schaller, 2014). To better understand how variations in temperature or precipitation may control landscape response, one approach is to quantify changes in sediment production or erosion rates across a suite of study sites with similar lithology and tectonics but differences in precipitation or temperature regimes. In a set GSA Bulletin; Month/Month 2016; v. 128; no. X/X; p. 1–17; doi: 10.1130/B31509.1; 9 figures; 3 tables; Data Repository item 2016362.; published online XX Month 2016. Present address: Department of Earth and Planetary Science, University of California–Berkeley, 307 McCone Hall, Berkeley, California 94720, USA; jmarshall@ berkeley .edu. For permission to copy, contact [email protected] © 2016 Geological Society of America