A series of submarine canyons occurs on the coastline off Kohala that appear to be down slope continuations of the spectacular and intensively eroded subaerial canyons that cut the windward flanks of the Kohala Mountains. These subaerial canyons are concentrated in an area of the coast that receives very high rainfalls and contains many springs, waterfalls and streams. A series of dives were made in the heads of two of these submarine canyons (Polola and Wainamu canyons) to investigate their origin and present-day activity. Our objective was to determine if the sediments that are shed from the land pass directly through these canyons and if fresh water sapping may play a role in canyon erosion. The high water table head on the adjacent land increases the potential for submarine springs near the heads of these canyons. 

The headward parts of these canyons are fairly straight but have large numbers of landslides along their walls and in the intercanyon areas. These landslides suggest that failure may be related to: fluid seeps, the deposition of unstable sediments, or the headward (upslope) advancement of progressive slumps.

On Dive 304 we collected 8 rock samples, 4 push cores and 2 Goflo water samples. Two rock samples and 2 push cores were obtained on Dive 305. Four rock samples, 3 push cores, 1 scoop bag, and 2 Goflo bottle samples were collected on Dive 306.  Continuous water samples were collected on all 3 dives for radium analysis.

Two dives (304 and 305) were made into the head of Polola Canyon. The first dive (304) was designed to investigate the basal part of the canyon’s walls for indications of erosive processes. The second dive (305) was planned for the examination of a landslide scar.

Dive 304
The start of the Dive 304 was in 257 m of water. The canyon floor is sandy and contains extensive low amplitude ripples that indicate both up and down canyon bottom currents. However, locally a stronger and confused ripple pattern was observed and at one locality a deep scour associated with a boulder contained both longitudinal and cuspate-like ripples forming a scallop-like pattern (at right). These types of ripples are generally associated with strong bottom currents.

The base of the canyon’s northern wall is extensively undercut forming many caves and overhangs where it was observed at a depth of 250 m.  A few fish occupied some of the deeply incised caves. On several occasions, while studying the conditions of the wall, a shimmering in the water was observed. However, Salinity-Temperature plots indicate that rapid mixing between the surface and deeper waters within the canyon probably produced the shimmering water.

The traverse up the northern side of Polola Canyon during Dive 304 encountered a 23-meter-high erosional scarp composed of well-layered mud overlain by a gentle mud covered slope. At a depth of 191 m, just below the ~10 m high landslide head scar, we found that mud chunks and carbonate cemented boulders with delicate white but apparently dead "finger" corals attached were scattered across the slope. The last 10 m of the canyon’s sides  (191-181 m) exhibits differential erosion that forms overhangs and other irregular surfaces that are developed in friable volcano-clastic sandstones. Upon traversing the scar we found a flat terrace-like sea floor onto which volcano-clastic boulders are scattered.  The tops of the boulders were extensively colonized with small delicate white “finger” corals. Although most all of these corals were in the living positions, the ones we collected did not appear to have been alive. Many beautiful tropical fish were found in and around the boulders, such as the angelfish at left.

Dive 305 was located near the head of Pololu Canyon where a well-defined landslide scar was obvious in the existing bathymetry, shown below. The dive commenced in 190 m of water and traversed the full length of the landslide from toe to head scar, a distance of about 330 m. The floor of the canyon at the toe is rippled sand with occasional depressions and scattered rubble. The dark sands that were sampled in the push cores were remarkably similar to the slightly lithified pieces of the sandstone talus that were sampled.  The highest concentration of rubble is located at the neck of the slide at a depth of 164 m where the slope is about 15° to 20°.  At 162 m we encountered a 10 m high scarp. The dive ended at 148 m on flat rippled seafloor, outside the canyon. Below are tiny cup corals on rubble on the Pololu Canyon floor.

Dive 306 was made in Waimanui Canyon to investigate a slump. This dive started in the canyon axis at a depth of 417 m and traversed a gentle sediment covered and bioturbated sea floor to the head scar of the slump. The base of the slump head scar lies at a depth of 393 m and is composed of mudstones with a distinctly more competent surface coating the vertical faces. This surface may be a vein along which the slope failure occurred. The surface coating is spalling from the outcrop and produces rubble in the form of slabs that are concentrated at the base of the scar. Overall, the basal part of the scar is smooth and clean looking with little biological growth.  

Two benthic ctenophores attached to a piece of wood, left,  reminded us of our own ctenophore specialist, George Matsumoto.

A less pronounced upper part of the scar is also composed of mudstones and forms a vertical cliff. This cliff is capped with slabs of extensive bioturbated well lithified mudstones. Sea fans grow on the vertical face of the mudstone wall, at right.

The canyons that were investigated during this series of dives appear to be active sand transport conduits.  Whether the sand is locally derived or transported from the streams that drain from the Kohala area is unclear. However, the rounded sides of the canyon walls suggest that erosional undercutting associated with the moving sand continues to destabilize the canyon walls.  No evidence of fresh water sapping or other fluid seeping processes were found.