Research
Research
My research uses recent granular physics approach to investigate different fundamental questions of the particulate sediment transport and associated landscapes dynamics. What is the vertical structure of the sediment transport? How it relates to the other granular transport processes, creep and suspension? Can we treat it like a soil, like an avalanche? (1)
Sedimentary systems dynamics, like rivers, dunes and avalanches often exhibit beautiful patterns. One of the main source of this complexity is the diversity of object sizes composing them. This complexity on one hand puzzles our understanding of the sediment transport processes but on the other hand offers new interpretation of the data to understand past and present evolution of landscapes. Nonetheless, if sedimentologists have long interpreted deposits grain sizes fluctuations, certain basic aspects of the mechanisms of transport of a polydisperse sediment remain mysterious. I have developed a new framework to explore the bimodal size distribution case. (2)
A fundamental consequence of the granular aspect of rivers transport dynamics is the shape of their channel. Indeed, if a granular pile of sand has a equilibrium slope due to grain-grain contact friction balancing gravity, it appears that rivers also develop an equilibrium shape able to balance friction, gravity, and hydrodynamic forces. Recent experiments in my Paris group demonstrated recently that a single, fully mobile, river channel indeed converges toward a predictable shape (Seizilles et al, PoF, 2013). To complete these experimental results I participated in a field trip in India on a very dynamic braided river system, the Kosi river. Even carrying a lot of sediment and so being far from equilibrium, our carefully analyzed dataset was able to show that the equilibrium-channel theory continues to work. (3)
In parallel to my research on rivers dynamics, I started a side project these last years about giant dunes sizes, based on fruitful discussions with Doug Jerolmack, Raleigh Martin and Federico Falcini. (4)
Vertical structure of bed load dynamics and river hysteretic behavior
Most of gravel bed rivers exhibit sediment transport in a regime very close to the onset of particles motion. This singularity is related to the river channel specificity of adjusting its shape : the width increases as the flow discharge increases (see section 4 below). To improve our understanding of the bed load transport, we image the inside of the bed over a long time doing a specific experiment in an annular flume where we match the refractive indexes of sedimenting plastic particle and a viscous oil.
The experiment is schematized below, and on the right.
The B&W image is an example of images we capture of of inside of the bed. The 3 circles represent 3 different grains whose we had tracked the displacement over time.
We track all the particles by image analysis and observe an important vertical gradient of particle velocity. (see example of tracks below)
River transport of a bimodal sediment
Almost all sedimentary systems present different grain sizes. Sediment transport is very sensitive of sizes differences, thus, especially when the different grain sizes are of different colors, it draws beautiful patterns.
To study this problem physically, during my PhD I designed a very simple experiment, where a steady turbulent flow shear a homogeneous bimodal bed. The figure below represent a) a scheme of the experiment, b) a top view of a bimodal bed I prepared for each experiment, where a fraction of big particles are dyed in black. For each experiment a record black particles displacements with an high-speed camera (250images/s). Analyzing images, I computed the averaged particles velocity Vi and number of moving grains per unit area ni (size i = 1,2) for different shear stresses and bed compositions.
For different compositions (represented by different symbols) and the two different grain sizes (small in white, large in black), we found universal trends with the shear stress using meaningful normalisations of ni and Vi., and carefully taking account the evolution of the critical shear stresses with the composition.
River channel and river networks of channels
Bed load transport rivers always present channels close to their equilibrium shape : where particles laying at their surface are close to their threshold of motion. With a laminar flow river experiment, Seizilles et al., (2014) demonstrated recently that this equilibrium shape can be simply derived from the condition of bed slopes being at the threshold of motion.
In order to better understand this extraordinary system, but also to verify laboratory results, I participated to a campaign of measurements of river flow and geometry on the Kosi river fan. This system presents the double interest of being very dynamic and simple (composed of sand only, with very little vegetation and almost no soil).
As represented on this map, we documented the transition from the fully braided state to the single channel state of the river., which allows us to test the Width-Discharge relationship over several orders of magnitudes.
Ref : Gaurav et al, (2014).
Side project : What controls giant dunes size ?
Desert dunes grow with time by coarsening as small dunes move faster than big ones. However, the limiting mechanisms responsible of the maximum dunes size of a given desert remain unclear.
As illustrated in these figures of Ewing et al., Geomorphology (2009) eolian dunes stature increases with time. Experiments and models in flume show that the maximum size dunes can reach over time scales with the the flume depth. On an other hand, numerous field datasets exhibit trend without saturation.
Recently, a study proposed that hydrodynamic feedback between the bed elevation and the atmospheric boundary layer (ABL, non-stratified first atmospheric layer, of zero to few km) surface was limiting dunes growing (Andreotti et al., Nature, 2009). We suspect that the correlation one can observe with ABL thickness is still an egg-chicken problem. Indeed, depending on the flow conditions, the boundary layer surface can conversely adapt to the topography. An attempt to analyse new data to document the atmosphere structure dynamics close to the surface of deserts is conducted now to bring new constrains on this problem.
Photo of the longitudinal dunes of the Namibia Desert (extracted from Andreotti et al,. 2009 ).