A deep borehole drilled at Warrenton, Oregon in the ancestral valley of the Columbia River represents a geologic record that extends from 2 ka to 16 ka. Prior to the onset of the Holocene marine transgression at 16 ka, the incised Columbia River valley was cut to 112 m below present sea level at this location. The onset of estuarine circulation in the Columbia River estuary occurred at 11.5 ka as determined by the first appearance of brackish water diatoms in sediments from borehole cores at a depth of 70 m. Grain size and heavy mineral analyses indicate that the Columbia River tidal basin served initially (16-11 ka) as a bedload-bypassing conduit to the continental shelf, and/or the Astoria Canyon. With the ongoing Holocene transgression, the Columbia River tidal basin became more efficient as a river sediment sink between 11.5 and 9.0 ka. After 9.0 ka, the filling tidal basin again served as a sediment-bypassing conduit (source) of sand to the coastline and continental shelf. Heavy mineral analyses indicate that between 16 and 11 ka the Columbia River tidal basin was dominated by sediment from the metamorphic interior basins of the Columbia drainage basin. The dominant sediment source changed from the metamorphic interior basins of the Columbia drainage basin to the Cascade volcanic arc between 11.5 and 9.0 ka. After 9.0 ka, the tidal basin was dominated by Cascade volcanic arc derived sediments. The total volume of Holocene sediment, primarily bedload, which has accumulated in the lower Columbia River valley, is 73 km3. This compares to only 13-km3 accumulation during the last 5 ka. The tidal basin accumulation rate gradually increased from 0.6 million m3/yr to slightly over 18 million m3/yr between the depths of 112 m and 30 m (16 to 9 ka). Above the 30 m depth (corresponding to 8.2 ka), the rate of basin volume fill dramatically decreased to just over 4 million m3/yr. Such a large decrease in sediment accumulation rate suggests that after 9 ka sediments were bypassing the nearly full tidal basin to the beaches and inner shelf. The rates of bedload bypassing the lower Columbia River valley, substantially greater than 2.4 million m3/yr, supported the shoreface progradation in the littoral cell during the late Holocene.

The Columbia River littoral cell (CRLC) consists of four subcells (totaling 160 km in length) that are unique in the West Coast of the United States, in that they contain prograded barriers and beach plains, reaching 0.5�3 km in width (Fig. 1). The prograded beach deposits (1�5 ka in age) overlie shoreface deposits (1�8 ka in age), as identified in 18 ground penetrating radar profiles, and sampled from 24 boreholes. Two competing hypotheses were initially proposed to account for the origins of these unique, progradative shorelines: (1) cross-shore feeding by onshore wave transport of pre-Holocene sand from the submerged shelf, and (2) longshore dispersal of nearshore sand that was supplied to the littoral system by bedload sediment discharge from the Columbia River during the Holocene. The CRLC sand forming the shoreface deposits is fine (diameter 0.2 ± 0.02 mm) and rich in lithic fragments (20�40% by volume). Gravel and shell lag layers are uncommon in most of the CRLC shoreface deposits, but they show greater abundance locally near ravinement surfaces, tidal inlets, and in the Clatsop subcell, located south of the Columbia River mouth. Gravel and granule layers increase upsection in barriers south of the Columbia River and downsection in barriers at the northern end of the littoral system. These trends suggest different mechanisms of shoreface sediment feeding within the four subcells. However, borehole samples from all four subcells show the same sand provenance, i.e., post-glacial Columbia River sand, which is identified by high ratios of hypersthene:augite in heavy-mineral fractions. Selected shoreface sections were dated (0.5�8 ka) by AMS radiocarbon analysis of articulated-shell and wood fragments recovered from auger flights (3�22 m depth subsurface). Relatively young shoreface deposition (2.5 ka at − 6.5 m elevation NGVD88) in the Clatsop subcell south of the Columbia River shows a net-southward beach transport that fed shoreface and beachface progradation into deeper water. Older and deeper shoreface deposition (4.4 ka at − 7.1 m elevation) in the Long Beach subcell north of the Columbia River was a result of the filling of innermost-shelf accommodation space prior to beachface progradation. The total volume of shoreface sand deposited under the barrier spits and beach plains of the CRLC is estimated to be 6�7 km3 deposited since 6�8 ka. There was a net-northward transport of littoral sand ( 1 × 106 m3 year− 1) along the nearshore and inner-shelf; subsequently some of this sand was transported onshore to feed beaches of the northernmost subcells. Columbia River sand was also the source for the formation of the offshore shelf wedge above the transgressive ravinement surface, and for the filling of major tidal basins located north of the Columbia River. In summary, the unique progradational history of the CRLC barriers and beach plains derives from the combination of (1) longshore dispersal of fine sand discharged from the Columbia River during Holocene time, and (2) across-shore feeding of beaches at the northern end of the littoral system from fine sand carried north along the nearshore and the inner-shelf.

Sediment derived from the Columbia River has been deposited on the continental shelf, along the barriers and beaches, and in the bays of the Oregon and Washington coast during the Holocene. The barrier and beach deposits of this 150-km section of coast comprise approximately 6 km3 of these Holocene sediments (Peterson et al., 2010-this issue) while the fluvial and bay deposits comprise about 104 km3 (Baker et al., 2010-this issue), and the shelf deposit is approximately 79 km3. Seismic-reflection, sidescan sonar, and surface sediment data show that the shelf deposit is not uniform in distribution or composition. The shelf deposit is 15�50 m thick off the beaches of the southern part of the study area but is less than 3 m thick, and, in places, absent from the inner shelf in the northern third of the study area. Surface sediment texture of the shelf deposit varies as well. Pleistocene-age gravel covers parts of the inner shelf in the northern third of the area. To the south, the surface of the Holocene shelf deposit is composed of fine sand near shore that grades offshore to dominantly very fine sand in 25�30 m water depth and muddy sand on the middle and outer shelf (> 50 m depth). Although a huge volume of sediment covers the shelf, its uneven distribution indicates that in places only small amounts are available as a potential offshore source to the adjacent beaches, and in other places the finer-grained nature of the shelf deposit indicates that significant winnowing of fine sediment would be necessary to make it compositionally equivalent to sediment on adjacent beaches.