Are Landscapes Buffered to High-Frequency
Total Page:16
File Type:pdf, Size:1020Kb
Watkins et al. Are landscapes buffered to high-frequency climate change? A comparison of sediment fluxes and depositional volumes in the Corinth Rift, central Greece, over the past 130 k.y. Stephen E. Watkins1,†, Alexander C. Whittaker1, Rebecca E. Bell1, Lisa C. McNeill2, Robert L. Gawthorpe3, Sam A.S. Brooke1, and Casey W. Nixon3 1Department of Earth Science and Engineering, Imperial College London, Prince Consort Road, London, SW7 2AZ, UK 2Ocean and Earth Science, National Oceanography Centre Southampton, University of Southampton, Southampton, SO14 3ZH, UK 3Department of Earth Science, University of Bergen, Allégaten 41, 5007 Bergen, Norway ABSTRACT vary spatially around the gulf, and we used istics of sediment supply to depocenters (e.g., them to derive minimum catchment-aver- Humphrey and Heller, 1995; Allen and Dens- Sediment supply is a fundamental control aged denudation rates of 0.18–0.55 mm/yr. more, 2000; Gawthorpe and Leeder, 2000; on the stratigraphic record. However, a key Significantly, our time series of basin sedi- Densmore et al., 2003; Cowie et al., 2006; question is the extent to which climate affects mentary volumes demonstrate a clear reduc- Backert et al., 2010). However, the importance sediment fluxes in time and space. To address tion in sediment accumulation rates during of climate in controlling sediment fluxes from this question, estimates of sediment fluxes the last glacial period compared to the cur- catchments to depositional basins over a range can be compared with measured sediment rent interglacial. This implies that Holocene of timescales remains contentious (e.g., Collier volumes within a closed basin that has well- sediment fluxes must have increased rela- et al., 2000; Jerolmack and Paola, 2010; Simp- constrained tectonic boundary conditions tive to Late Pleistocene times. Furthermore, son and Castelltort, 2012; Armitage et al., 2013; and well-documented climate variabil- BQART-derived sediment flux predictions Godard et al., 2013; Foreman, 2014; Hidy et al., ity. The Corinth rift, central Greece, is one of indicate a 28% reduction in supply during 2014). Field-based and seismic stratigraphic the most actively extending basins on Earth, the last glacial period compared to the Holo- studies have demonstrated that sediment fluxes with modern-day GPS extension rates of up to cene; likewise, seismic sediment accumula- can change from glacials to interglacials (Col- 15 mm/yr. The Gulf of Corinth forms a closed tion rate estimates indicate a similar magni- lier et al., 2000; Fuller et al., 2009; Sømme system, and since ca. 600 ka, the gulf has tude of reduction (32%). At the Last Glacial et al., 2011; Hidy et al., 2014); however, the fluctuated between marine and lacustrine. Maximum, mean annual temperatures in the magnitudes, trends, and causes are still debated. We esti mated suspended sediment fluxes for region were lower by 5 °C, but mean annual Moreover, some authors have argued that sedi- rivers draining into the Gulf of Corinth using precipitation rates were broadly similar. We ment routing systems may be buffered to high- the empirically derived BQART method over hypothesize that although weathering rates frequency climate variation (e.g., Castelltort the last interglacial-glacial-interglacial cycle might be greater under glacial conditions, and Van Den Driessche, 2003; Jerolmack and (0–130 k.y.). Modern temperature and pre- warmer interglacial temperatures may be Paola, 2010). Additionally, disagreement be- cipitation data sets, Last Glacial Maximum more conducive to generating larger storms, tween numerical models with different setups reconstructions, and paleoclimate proxy in- which do more geomorphic work, driving and assumptions about whether landscapes are sights were used to constrain model inputs. greater sediment fluxes. Our results dem- insensitive (Jerolmack and Paola, 2010; Armi- Simultaneously, we exploited high-resolution onstrate that sediment export to the basin is tage et al., 2013) or sensitive (Simpson and two-dimensional seismic surveys to interpret sensitive to glacial-interglacial cycles, and we Castell tort, 2012; Godard et al., 2013) to glacial- three seismic units from 130 ka to present, explore the potential mechanisms behind this interglacial climate change highlights the need and we used these data to derive an inde- sensitivity. for model validation by further field studies. pendent time series of basin sedimentary In this study, we address these fundamental is- volumes to compare with our sediment input INTRODUCTION sues by examining the effect of high-frequency flux estimates. Our results predict total Holo- climate change (104–105 yrs) on a complete cene sediment fluxes into the Gulf of Corinth Sediment supply from source regions is a source-to-sink system: the active Gulf of Corinth of between 19.2 km3 and 23.4 km3, with a pre- fundamental control on the stratigraphic record rift, central Greece, which is a tectonically and ferred estimate of 21.3 km3. This value is a (Gawthorpe and Leeder, 2000; Allen, 2008a; climatically well-understood study area. This factor of 1.6 less than the measured Holocene Blum and Hattier-Womack, 2009; Armitage sedimentary basin is underfilled, meaning any sediment volume in the central depocenters, et al., 2011; Covault et al., 2011; Allen et al., changes in volumes are due to sedi ment sup- even without taking lithological factors into 2013; Hampson et al., 2014; Michael et al., ply and not accommodation changes. In addi- account, suggesting that the BQART method 2014; Romans et al., 2016). In areas of active tion, the basin is closed at glacial lowstands and provides plausible estimates. Sediment fluxes extension, landscape and rift evolution models effectively closed at highstands (Peris soratis have all shown the importance of tectonics in et al., 2000), enabling mass balance compari- †stephen .watkins10@ imperial .ac.uk driving the volume, grain size, and character- sons. Therefore, this setting allows us to make GSA Bulletin; March/April 2019; v. 131; no. 3/4; p. 372–388; https://doi.org/10.1130/B31953.1; 9 figures; 4 tables; published online 21 September 2018. 372 Geological Society of America Bulletin, v. 131, no. 3/4 © 2018 The Authors. Gold Open Access: This paper is published under the terms of the CC-BY license. Downloaded from https://pubs.geoscienceworld.org/gsa/gsabulletin/article-pdf/131/3-4/372/4651308/372.pdf by University of Southampton/National Oceanography Centre user on 02 May 2019 Are landscapes buffered to high-frequency climate change? a simpler comparison of source to sink dynam- therefore buffered. They concluded that initial tonic time scales. Similarly, for a series of small ics than many other settings. A rift naturally step changes in climate might be recorded in catchments in California, Covault et al. (2011) produces a series of drainages that wholly drain stratigraphy, but ultimately large sediment rout- compared cosmogenic radionuclide– derived into the basin, unlike many other types of basins, ing systems (300–1000 km length) are insensi- catchment-averaged denudation rates with seis- which are characterized by more complex ex- tive to high-frequency precipitation variations mically derived sedimentation rates offshore ternal and internal drainages. Rift drainages are (cf. Castelltort and Van Den Driessche, 2003; and showed that, during falling and lowstand also relatively small, unless a regional-scale river Allen, 2008b). However, Simpson and Castell- sea level (40–13 ka), sediment budgets balanced has been captured by the rift system at one of the tort (2012) argued that catchment systems are well with denudation rates, whereas during ris- ends, which is not the case here. In this contribu- sensitive to high-frequency changes in water ing and high sea levels, sediment budgets were tion, we: (1) predict Holocene sediment fluxes discharge, which can be related to precipitation greater than denudation rates. to the basin using the BQART methodology changes. Additionally, their model suggested Strath terraces and paleo-erosion-rate esti- (Syvitski and Milliman, 2007); (2) evaluate the that time-averaged sediment fluxes misrepresent mates along the Eel River, California, suggest extent to which these fluxes match with depo- the actual variability of sediment fluxes. The that high erosion rates are correlated to the 70– si tional volumes calculated from seismic reflec- model of Godard et al. (2013) also found similar 18 ka glacial period (Fuller et al., 2009), with tion data; and (3) use our seismic constraints on results to Simpson and Castelltort (2012), where higher inferred precipitation rates for that region sediment volumes and rates to investigate the sediment fluxes were demonstrably sensitive to (Adam and West, 1983). However, spatially response of the Gulf of Corinth system to cli- high-frequency changes in precipitation. Finally, averaged catchment-wide cosmogenic denuda- mate change since 130 ka. The results provide a number of modeling studies have identified tion rates from river catchments in the southern new insights into the sensitivity and response of autogenic behavior in sediment yields, mean- United States implied that interglacial denuda- sediment routing systems on a regional scale to ing fluxes vary even when there is no change in tion rates were higher than glacial rates (Hidy climate over a full glacial-interglacial cycle. the system boundary conditions (Jerolmack and et al., 2014); these results agreed well with Paola, 2010; Van De Wiel and Coulthard, 2010).