A series of laboratory experiments has been conducted in order to elucidate the sediment‐induced mixing processes accompanying riverine outflows; specifically, the discharge of a warm, fresh, particle‐laden fluid over a relatively dense, cool brine. In a parameter regime analogous to recently acquired field measurements, hypopycnal (surface) plumes were subject to a convective instability driven by some combination of heat diffusing out of the warm, fresh, sediment‐laden plume and particle settling within it. Convection was robust in the presence or absence of intense turbulence, at sediment concentrations as low as 1 kg m−3, and took the form of millimetre‐scale, sediment‐laden fingers descending from the base of the surface plume. A consequence of the convective instability of the original hypopycnal plume is the generation of a hyperpycnal (bottom‐riding) flow. The experiments presented here indicate that natural river outflows may thus generate hyperpycnal plumes when sediment concentrations are 40 times less than those required to render the outflow heavy relative to the oceanic ambient. The resulting hyperpycnal plumes may play an important role in transporting substantial quantities of sediment to the continental slope and beyond.

JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about JSTOR, please contact support@jstor.org.. The University of Chicago Press is collaborating with JSTOR to digitize, preserve and extend access to The Journal of Geology. A B S T R A C TWe have performed a series of laboratory experiments that clarify the nature of the transition between fluid-mud and grain-flow behavior. The surface velocity structure and the speed of the nose of debris flows in channels with semicircular cross sections were measured with several cameras and visual tracers, while the mass flow rate was recorded using a load cell at the exit chamber. Other rheological tests were used to calculate independently the yield strength and matrix viscosity of the debris-flow mixture. Shear rates were varied by nearly an order of magnitude for each mixture by changing the channel radius and slope. Shear rates were significantly higher than expected (6-55 s Ϫ1 ), given the modest slopes examined (10.7Њ-15.2Њ). The large values were primarily a result of the concentration of shear into narrow bands between a central nondeforming plug and the sidewall. As a result, the shear rate of interest was calculated by using the width of the shear band and the plug velocity, as opposed to the flow depth and front velocity. The slurries exhibited predominantly fluid-mud behavior with finite yield strength and shear-thinning rheologies in the debris-flow body, while frictional behavior was often observed at the front, or snout. The addition of sand or small amounts of clay tended to make the body of the flows behave in a more Bingham-like fashion (i.e., closer to a linear viscous flow for shear stresses exceeding the yield stress). The addition of sand also tended to accentuate the frictional behavior at the snout. Transition to frictional grain-flow behavior occurred first at the front, for body friction numbers on the order of 100. Similar behavior has been observed in an allied field site in the Italian Alps. In the experiments, it was hypothesized that the snout-grain-flow transition was a result of concentration of the coarsest material at the flow front, reduced shear near the snout, and loss of matrix from the snout to the bed. Regardless of the frictional effects at the snout, flow resistance in the body was nearly always regulated by yield-stress and shearthinning properties, with no discernible boundary slip, despite volumetric sand contents in excess of 50%.

[1] As part of the STRATAFORM program, a bottom-boundary layer (BBL) tripod was deployed at 120 m depth in the northern thalweg of the Eel Canyon during winter 2000. Increases of the near-bottom suspended-sediment concentrations (SSC) recorded at the canyon head were not directly related to the Eel River discharge, but were clearly linked to the occurrence of storms. BBL measurements revealed that during intensifications of the wave orbital velocity, sediment transport at the head of the canyon occurred as sediment gravity flows directed down-canyon. Observational evidence for near-bed sediment gravity-flow transport included an increase toward the bed of the down-canyon component of wave-averaged velocity and high estimated SSC. At higher sampling frequencies (1 Hz), the current components during these events fluctuated at the same periodicity as the pressure, reflecting a clear influence of the surface-wave activity on the generation and maintenance of the sediment gravity flows. The origin of such flows is not related to the formation of fluid muds on the shelf or to intense wave-current sediment resuspension around the canyon head region. Rather, liquefaction of sediment deposited at the head of the canyon (induced by wave-load excess pore water pressures during storms) combined with elevated slopes around the canyon head appear to be the mechanisms initiating sediment transport. The resulting fluidized-sediment layer can easily be eroded, entrained into the water column, and transported down-canyon as a sediment gravity flow. Results from this study reveal that storm-induced sediment gravity flows occur periodically in the Eel Canyon head, and suggest that this kind of sediment transport process can occur in other submarine canyons more frequently than previously expected.

We performed a series of laboratory experiments to investigate the interactions of a turbulent wave boundary layer with a predominantly silt-size sediment bed. Quasi-steady, turbulent, highdensity suspensions (HDS) formed over a wide range of wave conditions and had near-bed (ϳ1 mm above bed) concentrations ranging from 17 to 81 g/l scaling roughly with the wave orbital velocity. HDS were defined by the presence of a lutocline, an abrupt change in vertical concentration gradient. Despite the initial bed being 70% silt and 20% sand, HDS had significant near-bed sand fractions ranging from 27 to 78%. Winnowing of the bed caused more concentrated HDS to be coarser grained, which in turn caused the suspensions to be thinner because of the greater settling velocity of the sediment. Our experiments are consistent with a dynamic feedback model where suspended sediment is limited through sediment-induced stratification expressed with a bulk Richardson number. However, our computed values of the bulk Richardson number converge to a value that is an order of magnitude less than the critical value of 0.25 that is typically assumed. The experimental wave orbital velocities (15-60 cm/s) and periods (3-8 s), as well as the characteristics of the HDS and the bed in our experiments, were comparable to observations made on the Eel shelf, California, during storm conditions when fluid mud has been observed.

We performed a multibeam survey of Eel Canyon, offshore northern California. The survey revealed a signifi cant bend in the canyon that appears to be due to the oblique compressional tectonics of the region. A series of steps within a linear depression, ~280 m above the canyon fl oor, extends from the canyon rim at this bend to the subduction zone and a distinct fan-like topographic rise. We hypothesize that the linear depression is a distributary channel and the steps are cyclicstep bedforms created by turbidity currents. Our interpretation indicates that turbidity currents are able to run up and overspill the 280-m-high canyon wall, resulting in a partial avulsion of the canyon and the construction of a fan lobe that is offset from the canyon mouth. Simple hydraulic calculations show that turbidity currents generated in the canyon head from failure of 2-3 m of material would be capable of partially overfl owing the canyon at this bend, assuming steadyuniform fl ow, full conversion of the failed mass into a turbidity current, and a range of friction coeffi cients. These estimates are consistent with analyses of sediment cores collected in the head of Eel Canyon, which suggest that 2-3 m of material fails on decadal time scales. Our calculations show that the overfl owing parts of the currents would have large shear velocities (>10 cm/s) and supercritical Froude numbers, consistent with erosion of the distributary channel and formation of cyclic steps by turbidity currents.