SLOW, UNCONFINED SPREADING OF A MUDFLOW

Citation
P. Coussot et S. Proust, SLOW, UNCONFINED SPREADING OF A MUDFLOW, J GEO R-SOL, 101(B11), 1996, pp. 25217-25229
Citations number
71
Categorie Soggetti
Geochemitry & Geophysics
Journal title
JOURNAL OF GEOPHYSICAL RESEARCH-SOLID EARTH
ISSN journal
21699313 → ACNP
Volume
101
Issue
B11
Year of publication
1996
Pages
25217 - 25229
Database
ISI
SICI code
2169-9313(1996)101:B11<25217:SUSOAM>2.0.ZU;2-2
Abstract
Mudflows are natural, highly concentrated water-clay-grain mixtures th at flow in mountain streams after long or intense rainy periods and ma y cause considerable damage if they overflow on the alluvial fan. The possibility of predicting the extent of these flows on the basis of ma terial and flow parameters is examined. The simplest realistic case of a yield stress mudflow moving through a narrow open channel followed by a wide, long plane is considered. It is demonstrated that the uncon fined flow of a yield stress fluid over an inclined plane cannot be un iform; even in steady state the flow width should increase continuousl y from the channel exit. A complete treatment of the flow equation on the basis of the long-wave approximation, including an appropriate thr ee-dimensional expression for the constitutive equation, makes it poss ible to establish a system of equations from which flow characteristic s at any point (longitudinal and lateral mean velocities and fluid dep th) can be deduced. In particular, for a Herschel-Bulkley fluid with a flow index of 1/3 it is found that the lateral extent will increase p roportionally to the distance from the channel exit to the power 9/20 and that, in the sheared part, the fluid depth in a cross section will have a parabolic distribution. Experiments have been carried out with fine mud suspensions (at different solid fractions) whose rheological behavior is similar to that of natural mudflows. The theory is in fai r agreement with experimental data concerning fluid depth distribution but systematically overestimates lateral extent (by 30%). This is cer tainly due to the fact that the assumption of lateral extent much smal ler than flow length is not respected in our tests.