Paleodischarge reconstruction using oxbow lake sediments complicated by shifting hydrological connectivity

Highlights

• Local morphology has the first-order control on flood deposits in oxbow lakes.

• Flood magnitude controls lake-sediment coarseness during stable morphologic control.

• Shifting hydrological connectivity may compromise paleodischarge reconstruction.

Abstract

Alluvial sedimentary records attract great attention for quantitative paleoflood studies because they may preserve continuous high-resolution archives in lowland floodplains that are most vulnerable to inland flooding. The coarse-grained flood deposit in oxbow lakes emerges as a promising paleoflood proxy due to its sensitivity to flood magnitude and high preservation potential. Previous work has established a coarseness-discharge correlation for the instrumental period that was subsequently used to retrodict paleodischarge of older flood deposits. However, it has not been tested whether this correlation applies to older deposits that may represent different lake hydrological connectivity due to plug-bar growth and river migration. Furthermore, overbank sedimentation is known to be episodic and may not scale with flood magnitude. Here we investigate this issue by comparing peak annual discharge and the coarseness of a two-stage infill sequence of a ~60 yr old oxbow lake in South Carolina, USA. We found that local morphology had the first-order control and was responsible for the sedimentation shift in the lake. There is a strong linear coarseness-discharge correlation in the upper section of the lake infill. However, this correlation overestimated the flood magnitude for flood deposits in the lower section. This finding suggests that oxbow lake sediments can be greatly useful for paleoflood reconstruction, but this may be complicated by shifting hydrological connectivity. Therefore, the stratigraphic context of an oxbow lake needs to be well understood to derive paleodischarge from the sediments properly.

Introduction

Riverine flooding is among the most costly and deadly natural hazards globally that are expected to get worse with climate change (e.g., Davenport et al., 2021; Dottori et al., 2018; Jongman et al., 2012; Rojas et al., 2013; Wallemacq et al., 2018). The intensity of heavy precipitation has increased and is projected to continue due to anthropogenic global warming (e.g., Allen and Ingram, 2002; Easterling et al., 2017; Min et al., 2011). However, the correlation between riverine flooding and anthropogenic climate change is unclear (e.g., Dickinson et al., 2019; Mallakpour and Villarini, 2016), particularly at the river-basin scale (Dankers et al., 2014), adding significant uncertainty to flood hazard projection. There are several challenges in correlating changes of instrumental riverine flooding with anthropogenic climate change. First, instrumental records can be affected by land-use changes and river engineering (e.g., Munoz et al., 2018; Villarini and Smith, 2010). Second, most instrumental records are relatively short and not useful for separating the effects of anthropogenic climate change and long-term natural climate variability. Third, instrumental records may not capture the most extreme riverine flooding. Therefore, paleoflood hydrology emerged to use geomorphological (e.g., Baker, 1977, 2008; Knox, 1985; Kochel and Baker, 1982; Shi et al., 1987) and alluvial sedimentary proxies (Fuller et al., 2018; Ishii et al., 2017; Jones et al., 2010; Korponai et al., 2016; Leigh, 2018; Munoz et al., 2015, 2018; Peng et al., 2019, 2020, 2020; Toonen et al., 2017, 2020, 2020; Wolfe et al., 2006) to reconstruct long-term pre-instrument records to investigate climatic and anthropogenic controls on riverine flooding.

Alluvial sedimentation is episodic (Aalto et al., 2003; Shen et al., 2015) and may not scale with flood magnitude (Gomez et al., 1995; Lecce et al., 2004; Magilligan et al., 1998; Sambrook Smith et al., 2010). However, recent studies demonstrate that continuous sedimentary records in some alluvial niches may be used for quantitative high-resolution paleoflood reconstruction on centennial to millennial timescales (Toonen et al., 2020). The coarseness of historic infill sediments in oxbow lakes was shown to correlate with instrumental discharge (e.g., Munoz et al., 2018; Toonen et al., 2015), forming the basis to retrodict paleodischarge beyond instrumental records. Natural oxbow lakes are relict channels cut off from rivers due to flooding and channel migration (Constantine et al., 2010; Fisk, 1947; Hooke, 1995; Rowland et al., 2009; Toonen et al., 2012). These lakes are naturally dammed by plug bars at the cutoff location and receive coarser sediments from higher-stage floodwaters that entrain and transport more coarse sediments in suspension overtopping the plug-bar barrier. Therefore, extreme floods are indicated by silt to sand laminae in lake infill sequences (Munoz et al., 2018; Toonen et al., 2015). Plug-bar growth and river migration can change the hydrological connectivity between oxbow lakes and rivers, which may cause distinct shifts in the sedimentary architecture of the infill (Toonen et al., 2012). Furthermore, climate (Knox, 1993) and land-use changes (Syvitski et al., 2005) can change the sediment load of rivers, which may affect sedimentation in an oxbow lake. Therefore, the reliability of paleodischarge reconstruction depends on whether the sediment coarseness-discharge correlation established using the most recent infill sediments can be applied to older sediments of different infilling stages (Toonen et al., 2015, 2020).

To investigate this issue, we compared sediment coarseness of a ~60 yr old oxbow lake of the Waccamaw River in South Carolina, USA (Fig. 1) with peak annual discharge at a nearby United States Geological Survey (USGS) river gauge station (USGS 02110500). The lake infill shows distinct lithologic variations and can be divided into two sections. The instrumental discharge corresponding to the upper section was used to establish criteria to identify flood-event deposits and define the sediment coarseness-discharge correlation. The correlation was then used to retrodict discharges of event deposits in the lower section. The retrodicted discharges were compared with the instrumental data to test the paleodischarge reconstruction.