Abstract
Investigating experimentally the interactions between tectonics and surface processes is a challenging task imposing the development of an analogue material capable to simulate jointly geomorphic processes and geological deformation mechanisms. Under external forcings, typically surface water run-off induced by a rainfall system and deviatoric stresses imposed by a mechanical device, the analogue material must be able to develop a wide variety of morphological and tectonic markers such as watersheds, channels networks, alluvial fans, fluvial terraces, folds, faults scarps, etc. The physical and mechanical processes, behind the formation and evolution of tectonically controlled morphologies, act over a wide range of time and space scales and affect both the model surface and its internal structure. Rigorous model scaling can become, then, a serious issue. Laboratory constraints and the time and space characteristics of natural geological processes lead to typical length and temporal scalings in the ranges of 1 cm = 10-1000 meters, and 1 s = 1-1000 years, respectively. Based on these imposed parameters, more or less well-established scaling rules are used to determine the suitable properties of the analog material. In practice, they are probably not sufficient (see Paola et al., 2009 for a review).In the framework of the ANR Topo-Extreme project, we recently put some efforts to improve the morphological properties of the MATIV analog material used by our group to investigate Tectonics-Erosion-Sedimentation couplings in different geological contexts (see Graveleau et al., 2015 for a review). We completed the first step by correcting some limitations of the MATIV material such as its relatively low erodibility and drainage density. Adding pumice powder, capable to store water at a microscopic scale while preserving the very low permeability of the material, clearly boosted model surface erosion. By decreasing the granulometry of the coarser MATV components, we also significantly improved the level of detail exhibited by the experimental landforms.Results (figure 1) show that if we succeed in improving qualitatively and quantitatively model topography without degrading the simulation of tectonic processes, some of its morphological characteristics are still unsatisfactory. For instance, as shown by the length/width ratios, channel vertical incision appears not vigorous enough compare to channel lateral erosion. This could be partly corrected by increasing the material cohesion but it would strongly decrease material erodibility and generate also out-of-scale deformation processes. This recurrent issue can be limited by performing rainfall cycles, alternating dry and wet periods of a few seconds. By doing this, it becomes possible to modify the balance between channel erosion and slope diffusion processes. Indeed, as soon as the wet stage ends, slope processes are inhibited while water and sediments continue to be transported in the drainage network for a few more seconds until the complete cessation of water flows. During this period, vertical channel incision is more efficient because channel transport capacity is enhanced. However, using rainfall cycles does not solve totally the problem. What we suspect is that the use of dense silica particles, in both the MATIV and MATV mixtures causes two undesirable effects. First, it decreases the transport distance of the eroded particles, and, as shown by the erosion/sedimentation maps, part of these eroded materials are found stored in the lower part of the watersheds instead of being evacuated into the foreland basin. Second, the critical slope angle from which the analog material starts to erode is up to a few degrees while in nature this value is much lower (< 1°). To improve model scaling procedure and address the issues described above, we initiate a second step by including a numerical modeling approach. Our objective is to clone numerically the analog model using its morphometric characteristics to calibrate the numerical model. The latter is then used to analyze quantitatively model morphology and identify potential scaling discrepancies. First results are encouraging and provide new investigational pathways to further improve the MATV.