Morphodynamics of Levees Built by Turbidity Currents: Observations and Models OS23B-1662 Levees are the primary element of self-formed channels, are faithful recorders of channel history, and connect channels to their overbank surface yet little is known about their morphodynamics. Using an industry-grade 3D seismic survey we have studied a submarine network of channels located offshore Brunei Darussalam. We have mapped the seafloor and a shallow regional surface beneath the network of interest. The subsurface horizon defines the geometry of a scarp and slide plane associated with a mass-failure event that reset the margin to an unchannelized state. A map of deposit thickness created by differencing the seafloor and subsurface horizons was used to create plots of deposit thickness as a function of distance from a channel thalweg for channels of varying relief. Levee steepness increased from 0.01 m/m to 0.05 m/m as channel depth increased from 5 to 50 m, but this trend rolled over to a near constant steepness value of 0.05 m/m for channels greater than 50m in depth. A similar trend of levee steepness vs. local channel depth was observed in a reduced scale laboratory experiment. We model levee growth using a simple advection settling model for currents with multiple grain sizes and a vertical sediment concentration profile defined by the Rouse equation. This model reproduces the field and laboratory observations of levee growth and suggests that the most important parameters controlling levee deposition rates and steepness are the degree of channel confinement and the vertical structure of the suspended-sediment concentration profile. Straub, K.M., EAPS, MIT, 77 Massachusetts Avenue, Cambridge, MA Mohrig, D., Department of Geological Sciences, The University of Texas at Austin Support for our research was provided by Brunei Shell Petroleum and Shell International Exploration and Production Inc. Additional funding provided by the National Center for Earth- Surface Dynamics, an NSF Science and Technology Center A’ A B B’ z1z1 y z2z2 R ~100 m 4 Current Properties Determine Change in Levee Taper as a function of deposition Channel Relief = R Depositional taper = ∆z/y y ≈ 1 – 2 channel widths km 05 Seismic Dip Line Seismic Strike Line A A’ B B’ 0300 Deposit (m) Deposit Thickness Detachment Surface Slope Map ~ 50 m 2 km ~ 200 m 5 km Detachment Surface 200m water depth 1200m water depth 30 km Borneo S. China Sea Study Region Brunei Regional Overbank Deposition How does Levee Morphology Change as Channel Grows?? Each Deposit has constant Taper. Cumulative Taper increase with T. Individual deposit taper decreases with T. Cumulative Taper increase with T. Levees grow with self-similar form Importance of Levees to Seascape Models - Primary element of self-formed channels - Faithful recorders of channel history (channel bed subject to strong erosion & deposition). - Record vertical structures of currents - Levees connect channel to the overbank. Submarine channels offshore Brunei are bounded by prominent levees. We use an industry-grade seismic cube to unravel their growth history Horizontal Boundary of Subsurface Horizon Since last mass-failure event Brunei slope has been site of net deposition by turbidity currents Last mass-failure event reset margin to unchannelized state, allowing levee growth to be studied Observations 1st order control on deposition is distance from shelf-edge: points to highly progradational system Abrupt changes in slope induce changes in relative deposition, but key point is that system is net depositional Increasing Water Depth Failure Scarp Failure Scarp Horizontal Data Resolution = 25 m by 25 m Vertical Resolution ~ 5 m 0 0 N, E 0 0 N, E 14 0 N, E Background Deposition Extent of Channel Overbank Deposition Deposit thickness & distance to closest channel were calculated for every node point on deposit thickness map This data was binned by 25 m increments of distance to closest channel Channel center has ~ same deposit thickness as far-field deposit thickness Deposition from overbanking channelized flows extends ~ 2.5 km from channel. This distance is less than distance separating most channels We use a laboratory study and an advection- settling model to explore the influence of these 4 parameters Levee taper was defined from linear regression best-fit lines through plots of average cumulative deposit thickness vs. distance from channel. What causes roll-over in trend with increasing relief????? How does bulk levee taper and taper of individual flow event deposits change as channel relief increases?? Brunei Channels Preserved Mass-failure scarps

Laboratory Study of Submarine Levee Growth Parameters & Scaling for Channelized Turbidity Current Geometric Scaling (L) Model = (L) Prototype = 1/1000 Approximate Dynamic Similarity (Fr) M = (Fr) P, (p) M = (p) P Re ≥ 6400, ensuring turbulent flow conditions Model Width chnl = 77.0cm Depth chnl = 10.0cm Length chnl = 300cm U = 6.5cm/s H = 10.0cm T = 576 sec D50 = 2.9×10 -3 cm, m silt Prototype Width chnl = 770m Depth chnl = 50m Length chnl = 3000m U = 2.2m/s H = 100m T = 2.8hr D50 = 1.0×10 -2 cm, vf sand Governing Dimensionless Parameters Experimental Setup The experiment included 9 flow events Currents had 2.4% excess density from suspended crushed silica sediment. Currents were purely depositional with some reworking of suspension fallout as bedload Inlet current height was constant for all runs at 10 cm. Reduced scale laboratory experiments provide data needed to describe DYNAMICS that are missing in field scale studies of channel morphology. Experimental Results Comparison of Brunei and laboratory levees Summary Levee Growth Model We couple a suspended sediment concentration profile defined by a Rouse equation to an advection-settling scheme. As levee deposition occurs, channel relief increases causing progressive confinement of current Rouse Equation : ; Model Results Deposition rate is influenced by the near bed concentration and settling velocity of each (i) particle class ; Near bed concentration at a given location, x, on the levee is determined by concentration at a height in current that advects at a rate, U y, and settles at a rate, W s. A A’ Bathymetry (T = 0) Deposit (T8 – T0) Bathymetry (mm) Deposit (mm) C. I. = 2.5 mm C. I. = 6.5 mm Best-fit Parameters to Brunei Data Model is not indented to invert stratigraphy for exact levee forming C and U x Model does suggests degree of concentration profile stratification & amount of current confinement are most important parameters in setting levee taper Roll-over in trend of levee taper as a function of channel relief occurs once heavily stratified portion of concentration profile is confined in channel Observations from Brunei and lab experiments suggests taper of individual beds comprising levee is influenced by degree of current confinement m 01 Levee Deposit Properties Turbidity Current Properties Spatial change in levee deposit taper is greater than change in gradient of deposit grain-size with distance from levee-crest Stratification of concentration profile exceeds stratification of settling velocity profile Low mixing across low-shear zone for depositional turbidity currents Rate of levee taper increase is greatest at low values of Relief/Relief max for both channels that increase in relief through time (Brunei) and channels that decrease in relief through time (lab experiments). 1) Growth of tributary network of submarine channels offshore Brunei is the result of net depositional turbidity currents which construct prominent levees 2) Deposition rate and taper of levees bounding straight submarine channels are primarily controlled by structure of suspended sediment concentration profile and degree of current confinement 3) In the absence of direct measurements, the morphology and stratigraphy of levees can be used to constrain current properties, specifically current thickness 4) Further work is needed to quantify the effect of channel bends on levee morphodynamics Elevation of heavily stratified and high sediment concentration, lower portion of the turbidity current above levee crest results in rapid growth of levee thickness and taper Jim Buttles (University of Texas at Austin) provided additional help in conducting experiments