Feldspar single grain luminescence of modern fluvial sediments as a new tool to study fluvial transport

Luminescence has been developed as a dating tool for Quaternary deposits. One approach is the single-grain post-infrared luminescence (SG-pIRIR) protocol that provides high-resolution equivalent dose (D_e) distributions. This protocol is well-suited for fluvial deposits that often present large scatter in D_e distribution because of heterogeneous bleaching (zeroing) of the grains by sunlight exposure during transport.

Here we present a SG-pIRIR analysis of 14 samples of modern sediments from the Rakaia (RK) and the Waimakariri (WK) rivers in the South Island of New Zealand. Those rivers are output channels for tonnes of sediment eroded annually from the Southern Alps, they are braided in the Canterbury Plains on about 70 km, downstream of short sections of 10-20km where they are running into incised gorges.

The aim was to test and develop SG-pIRIR as a tool to document and quantify transport as proposed in some recent publications (McGuire and Rhodes, 2015; Gray et al.,2018; Sawakuchi et al., 2018). We focused on the fractions of saturated and well-bleached grains from D_e distributions, and on the mean D_e calculated with the central age model (CAM), as proxies for bleaching rates, transport and transient storage of particles in floodplains. In the Canterbury Plains, we found for both rivers that the percentage of saturated grains follow an exponential decay expressed as y= y₀.e^(-x/Lsat) with a characteristic length L_sat = 24 km, whereas on the opposite the quantity of well-bleached grains increase towards downstream at a rate of +4 to 7%/km. Similarly to the saturation, we observed an exponential decay of the CAM doses (characteristic length L_cam = 42 km). Those results reveal a strong alongstream bleaching of the grains.

We complement our natural-system analysis by building a numerical model that simulate the successive displacement and D_e evolution of a set of individual grains along a river of length l. The code includes three main processes that repeat until grains reach the river outlet: (1) displacement of a distance L_T set with an exponential probability density function (PDF); (2) temporary storage in the floodplain between two displacements, for a period R_t set with a PDF that follows a Pareto law (alpha=2). During R_t, D_e can increase by 3 Gy/kyr; (3) bleaching of grains during displacement (fluvial transport) or storage (if exposed at the surface of the floodplain during Rt) according to a probability P_Bl (tested from 0.01 to 0.3). We consider l=200 km and the input of 400 grains, 50%  with an initial D_e=50 Gy, which is the mean D_e measured upstream both rivers, and 50% at 1000 Gy (saturated). To first order, this simple model simulates well natural observations (L_{SAT and} L_CAM) along WK and RK for L_T on the order of a kilometer, Rt values of several decades and bleaching probability of ~0.05. This very simple transport model allows to better decipher SG-pIRIR data and to estimate transport length and resting times of sand-sized fluvial particles.

Future works should consider testing these tools on other contexts, either in other tectono-climatc context or on different flow styles.