top of page

Sill Injectites w/ Shattered Fills - Indian Creek, WA


At Indian Creek, WA I discovered 3 breccia-filled sills injected into varved Glacial Lake Columbia sediments, each extending for more than 10m across the exposure. The sills are one of several flavors of flood injectites found along the path of the Ice Age megafloods.


Sedimentary sills extend for several meters across exposures along Indian Creek Road. Indian Creek is a tributary to the upper Columbia River north of Davenport, WA. They kind of have flat tops....hmmm. The enormous smooth-bladed exposures in which I found them are now covered by erosion control matting and have been hydroseeded. I was fortunate to arrive just in time to inspect them before crews covered everything up.


Notes from my field notebook...


Upper Columbia River at Lower Indian Creek Road / Hawk Creek area, WA (Site #17-01)

UTM 11T-042642, 5295464, 1475’ elevation

October 2-3, 2017


October 2 - Newly cut, continuous outcrops exposing as much as 25m of vertical section over a lateral distance of nearly a kilometer. County contractors (M&L Construction of Spokane; Luke is the foreman) are actively covering the cut faces (smooth, ~20 degrees slope back) with erosion control mat in long rolls. All will be covered within a few days.


I measured 2 clastic sills (injectites) before sundown tonight. Both cut horizontally across tilted bedding. Fill was homogeneous, granular, but not sheeted. Rather, the fill appeared shattered and “grains” similar to peds, not sand-silt as in the typical Touchet-type dike. Rip-up clasts of silt or fine sand, silt, and clay. No silt skins observed. “Shattered intrusions” may be an appropriate term.


My first impression of the sills is that they are injectites in the sense of Cartwright, Hurst and others in the North Sea (and Parize and Fries). These are not sheeted dikes, but something directly related to flood loading and/or local slumping/landsliding, though the latter idea seems less likely. Slumped packages of the section were coherent (block slides) with abrupt, sharp slide surfaces visible and cutting at angles across the exposed cut faces. A few small faults offset bedding, but in all, the section is intact even if slumped in places. Perhaps as much as 25% of the length of the cut face (~1 km long) is contained in one tilted block or another. A full, untilted section could be readily measured near the middle of the exposure. See map (p. 53, Book M) for suggested locations.


October 3 - Measured 3rd sill today. Write Brian Atwater and Pat Spencer of large exposures I observed yesterday. Perhaps something will come of a described section there, if Lincoln County is willing to allow access and trenching of their new slope.


Gray sill cuts brown Glacial Lake Columbia deposits at a low angle. The intruded sediments are varved silty muds (lacustrine) repeatedly interrupted by gray sand layers (outburst flood deposits). Muds below the sands are typically deformed.


Brecciated mudstone, stratified rip-up clasts, and sand comprise the fill in the sill injectites. Would you call that diamict? I'm not sure it would be my first choice. 3' lens cap.


Maybe they should be called dikes. The crosscutting angle here is pretty large.



This sill's rubbly fill and tapered shape are clear here.


Section is located just upstream from the confluence of Indian Creek and Hawk Creek. Hawk Creek Campground on Lake Roosevelt is less than a mile away. The sills occur at the base of the exposed section in the contorted layer noted "meter-scale SSD, breccia, injectites". Twenty-four sand beds (floods) counted.


Contorted beds at the base of the exposure contain the sills. Note the meter-scale folds are truncated and overlain by a pile of mostly flat-lying lake beds and sands.



Field sketch of the outcrop at road level. The sills are associated with mud-clast breccias, rip-ups, and meter-scale folds.


Stair-stepping lower boundary of the lighter-colored sill is super cool. I interpret these dike-sill features as synsedimentary structures that formed rapidly. They intruded soupy sediments on the floor of Glacial Lake Columbia during inundations by Missoula floods. The mechanism is fluid-driven fracture (hydrofracture) triggered by the rapid overloading and facilitated by silt, which seals the fractures. Shattered sedimentary material was driven downward, laterally, and upward through the substrate. I think these are smaller-scale versions of "sand injectites" described in petroleum basins worldwide. Sand injectites are large scale intrusive features found in marine environments (i.e., North Sea Paleogene strata), specifically seafloor fans and the channels that feed them (turbidites). Formation involves intrusion of fluidized sand into fractures either during fan deposition or remobilization of buried, fan-channel sands. The oil industry targets sand injectites for drilling specifically because they conduct hydrocarbon fluids out of source rocks and are sealed in mudstone. See work by Dixon et al. (1995), Surlyk and Noe-Nygaard (2001), Parize and Fries (2003), GEO ExPro (October 2005), Hurst and Cartwright (2007), Braccini et al. in Oil Field Review (Summer 2008), Hurst et al. (2011), Cobain et al. (2016), and many others. Similarly-scaled features to those at Indian Creek, WA and Steamboat Rock may also form in glacial deposits overridden by glacial ice (Broster, 1991; Roshoff and Costgrove, 2002; Phillips et al., 2013; Crossen, 2014; Ravier et al., 2014; Shanmugam, 2015) and volcaniclastic sediments overridden by lahars (Herriott et al., 2014).


While I was at Indian Creek, I was given permission to scramble around and take notes on a section of Glacial Lake Columbia/Missoula flood deposits. The M&L Construction foreman, whose crew was installing erosion control matting along a nearly 1km long cut face was very accommodating. They had bladed the slope smooth and back to an angle that was just barely walkable. I had only 40 minutes to complete my description of the approximately 30m high section. I did not have time to make measurements, so the thicknesses shown are relative. Only three units were used: "Sand" (gray, outburst flood beds), "Sand-Silt" (tan-brown, waning flood deposition), and "Clay" (green-gray-brown, lacustrine deposition). Rip-ups (RU) and soft sediment deformation features (SSD) were also noted.

Twenty-four sandy outburst flood beds interrupt more or less continuous lacustrine sedimentation during the phase of Glacial Lake Columbia represented in the exposure. Soft sediment deformation comes in two forms here: syn-flood deformation (contorted bedding, injectites, small scale soft sediment deformation features) and later deformation related to landsliding (coherent, rotational slide blocks).


I'm calling them flood injectites. I coined the term, so I'd better use it. But maybe they formed in response to overriding glacial ice, or were remobilized by a subaqueous landslide, or are somehow related to isostatic rebound following the retreat of glaciers and the draining of proglacial lakes.

Deformed lower portion of section.

Once terrific outcrops are now gone. Imagine what we would see if these were left to weather.

I was lucky to arrive when I did. Crews were actively covering large portions of the exposure with erosion control matting.

The 24 flood beds preserved in the section at Indian Creek validates Eugene Kiver's hunch that "at least 27 floods" entered Glacial Lake Columbia (Kiver et al. in Baker et al. 1991, p. 241). My hasty stratigraphic column represents the only "continuous section of alternating flood and normal glacial lake sedimentation...exposed in the low-level Lake Columbia terrace" (p. 241) recorded to date. Wish I'd had more time to do detailed work.


The friendly locals stopped by a couple times to chat.




Last 50 Posts