The fluid dynamics of natural turbidity currents, Lillooet Lake, British Columbia, Canada

[Note: click on any image for a larger version.]

Jim Best's current research page.


In August 2001 and 2002, we conducted two field campaigns examining the dynamics of density currents generated by river inflow into Lillooet Lake, British Columbia, Canada. This work has been funded by the UK Natural Environment Research Council and had four principal aims:

  1. To quantify the vertical flow & suspended sediment structure of these turbidity underflows, and compare these with existing experimental and numerical approximations,
  2. To document the temporal variability of the flow structure,
  3. To investigate the interaction of turbidity currents with topography (slumps/channel margins), &
  4. To examine the nature of possible interflows in this stratified lake.

This study has provided the first dataset of this kind from natural sediment-laden turbidity currents & bridges the significant gap between natural currents & both laboratory & numerical simulations. Significantly, this will allow us to assess the validity of these previous approaches when applied to the natural environment.

This website presents the background to the study, details of the methodology and will be used to place accessible results/data as they are processed and published.

Aerial view of the Lillooet delta.

Figure 1: Aerial view of the Lillooet delta with the 2001 measurement stations at 25 and 50m depth marked, together with one of the long profile lines. Note the very distinct plunge line of the sediment-laden flow and the scalloped shape to the plunge line. Continuous underflows were generated to the lakeside of this plunge line.

The Research Team

The research team at Lillooet Lake consisted of Jim Best, Jeff Peakall and Mark Franklin(Leeds), Ray Kostaschuk (Guelph, Canada), Paul Villard (ex UBC and now Parish geomorphic) and also Arjoon Ramnarine ('Ram') from UBC, Vancouver who piloted the UBC launch in all weathers and with supreme accuracy!

Ray and Ram.
Figure 2: Ray and Ram

Ram on a rainy day at Lillooet.
Figure 3: Ram on a rainy day at Lillooet

Jim and Mark.
Figure 4: Jim and Mark

Jeff pulling up anchor.
Figure 5: Jeff pulling up anchor!

Paul keeping careful notes.
Figure 6: Paul keeping careful notes on all.

Mark loose on the lake.
Figure 7: Mark loose on the lake!

Ray and Mark at front of Lillooet delta.
Figure 8: Ray and Mark at front of Lillooet delta locating measurement stations

Jeff and Mark reach Lillooet River.
Figure 9: Jeff and Mark reach Lillooet River


Turbidity currents are the principal agent through which sediment is transferred to deep-sea environments & the deposits of these flows form volumetrically the largest sediment accumulations in the oceanic basins. Recent years have seen a great resurgence in research on density currents that has been driven by their significance within environmental management & the importance of turbidity current deposits in hydrocarbon exploration. Study of the dynamics & deposits of turbidity currents has progressed largely through both numerical & physical modelling, allied to the study of the deposits of these flows, & this has yielded more complete models for the fluid dynamics of these currents, their interaction with topography & depositional characteristics. However, this substantial progress has not been accompanied by comparable quantitative studies of larger-scale turbidity currents within natural environments, principally because these currents have been inherently difficult, if not impossible, to monitor. This has led to the current situation where our knowledge of these vital flows is dominated by results derived from small-scale laboratory experiments, numerical simulations or deductions made from the ancient sedimentary record. However, recent developments in field monitoring technology, principally acoustic Doppler profiling (ADP, with which we have extensive experience in fluvial/estuarine environments), now permit, for the very first time, the opportunity to document & quantify the fluid dynamics of larger-scale natural turbidity currents. This project aimed to examine the dynamics of quasi-continuous turbidity underflows generated by snowmelt in Lillooet Lake, British Columbia and, using ADP, echo-sounding & suspended sediment sampling, achieve the objectives listed above.

Field Area & Techniques

The project has quantified the fluid dynamics of natural, hyperpycnal density currents generated by seasonal snowmelt flooding in Lillooet Lake, British Columbia, Canada.

Aerial photo of Lillooet delta.

Figure 10: Aerial photo of Lillooet delta: note the distinct plunge line and main distributary channels of Lillooet River. The delta is approximately 1.5 km wide.

This site is ideal for the aims of this project, since it has previously been studied by Gilbert (1973, 1975) & shown to: i) produce sustained hyperpycnal underflows which peak in the period of early August-early September, ii) deposit a range of turbidite sediments, iii) have a varied delta slope morphology consisting of channels & slide/slump blocks (which therefore provides a range of topographies over which the currents flow), iv) possess suspended sediment concentrations similar to those in which we have successfully used ADP, & v) can provide stratified conditions which may yield interflows.

The field methodology was based around use of several principal techniques:

1) Acoustic Doppler profiling (ADP) was used to quantify the three-dimensional flow structure throughout the flow depth from a survey boat: we used SONTEK 1500 kHz and 500 kHz ADP's, that enabled measurement water depths up to 70m.

Acoustic Doppler profiler.

Figure 11: Sontek 500 MHz three-beam acoustic Doppler profiler.

2) Analogue & digital echo sounders were used to both quantify bed morphology & visualise density interfaces within the flow. In the 2002 field season we used a new Navisound echo sounder that produced both digital records and spectacular analogue chart traces (see results), depicting the form of the underflows and even allowing imaging of some large particles (probably large organic matter/wood).

Transducer of dual frequency echo sounder.

Figure 12: Transducer of dual frequency echo sounder

Navisound 215 digital echo sounder with analogue trace.

Figure 13: Navisound 215 digital echo sounder with analogue trace. Note record of underflow on chart paper.

3) Sediment sampling and thermistor records.

4) Real-time kinematic DGPS: the position of the boat and surveying instruments were fixed using a NERC GEP Leica RTK DGPS, with base station being set-up near the lake-head and rover receiver on the boat.

Leica RTK DGPS base station

Figure 14: Leica RTK DGPS base station.

Field Sampling Strategy

All of this equipment was mounted on an aluminium work launch hired from UBC and deployed in the lake.

Aluminium work launch.

Figure 15: Aluminium work launch used in Lillooet surveys. Note ADP mount and two DGPS aerials (for Leica and back-up Trimble systems). The hoist was used to lower water sampling equipment, bed grab sampler and thermistor probes.

Two field seasons were undertaken in August 2001 and 2002. The hydrographs of Lillooet river recorded at Pemberton (discharge records courtesy of Environment Canada) show clear diurnal variations due to snowmelt, as well as some longer term fluctuations due to longer periods of fine weather that increase discharge (see August 10-15th 2002). However, on August 22-23rd 2001, a large storm passed through the area and generated a large peak in the hydrograph on August 22nd. We were fortunate to be able to record underflows during this period in which peak discharge reached 540 m3 s-1 in comparison to the base level of approximately 240 m3 s-1 for the rest of the month.

Hydrograph 2001.Hydrograph 2002.

Figures 16a and b: Hydrographs of the Lillooet River at Pemberton for August 2001 and 2002. Note the diurnal snowmelt-related fluctuations and the large storm event on August 22nd 2001.


Delta Morphology

Echo sounder records allowed maps of the delta front to be constructed. Figure 17 depicts a map of the delta front in August 2001 and depicts three main features: 1) a steep delta front down to ~ 10-15 m; 2) a smooth, low-angle (1-5° ) delta front that extends down to ~120 m; 3) the delta slope shows a lack of well defined channels and displays little morphological evidence for large scale slump scars.

More detailed maps are under preparation for the 2002 field season.

Morphological map of the Lillooet delta.

Figure 17: Morphological map of the Lillooet delta, August 2001.

Qualitative Flow Visualisation

The echo sounders allowed an excellent qualitative picture to be obtained of density underflows as they progressed into the lake.

Analogue echo sounder trace of underflows.

Figure 18: Analogue echo sounder trace of underflows passing beneath the launch moored in ~15 m water depth at the front of Lillooet delta. Note the clear front to the underflows and the very distinct pulsing to the current.

Click here to view a larger image of the front of the current.

This Figure shows the passage of an underflow underneath the launch as it was moored in 15m depth. Several remarkable features are apparent:

  1. The clear head of the underflow with a bulbous nose that passes laterally into a zone with evidence of large-scale eddies (Kelvin-Helmholtz instabilities) on the top of the current
  2. Dark strong reflectors probably represent pieces of water-saturated wood that are floating within the water column (Lillooet lake possesses a lot of woody debris ranging in size from small twigs/logs to entire trees!). Note how these darker reflectors indicate the passage of the flow that displaces some of the neutrally buoyant wood up, over and then down behind the current.
  3. Most remarkable is the fact that at this one point, the flow is seen to pulse - notice how the first underflow passes by the measurement point but is then followed, a few minutes later, by another pulse in the underflow. We have thus recorded, for the first time, pulsing in underflows that is generated from a continuous input and our work (in review) shows the velocity signature and periodicity of these flows.

The plunge line was very distinct in both field seasons (see aerial photographs) and always possessed a scalloped pattern, with distinct lobes of sediment-laden fluid (Figure 19) progressing out into the lake before decelerating and retreating.

Sediment laden plume.

Figure 19: Sediment-laden plume progressing out towards the launch at the front of Lillooet delta. The advance and retreat of these lobes possessed a periodicity that was matched to the pulsing within the underflows.

ADCP Quantification

A paper detailing the first results from this research is currently under review: details will be posted here in the near future.

Data Repository:

Summary data plots, processed ADP data and sediment concentration profiles from the research project will be posted on this site as the data is published. Please contact Jim Best if you require any more details on the data format.


Gilbert, R. (1973) Observations of lacustrine sedimentation at Lillooet Lake, British Columbia. Unpubl. PhD. Thesis, University of British Columbia, 193pp.

Gilbert, R. (1975) Sedimentation in Lillooet Lake, British Columbia. Can. J. Earth Sci. 12, 1697-1711.

Jim Best's current research page.