Predicting liquefaction extent in the Yakima Fold Belt

Geologists have long debated how large of an earthquake Yakima Fold Belt faults can produce. Some claim these shallow (<10km), intraplate structures are capable of generating quakes up to 7.0 magnitude. Others estimate their capability no higher than about 6.0 magnitude. The difference between 6.0 and 7.0 is huge. I've personally experienced two Alaska quakes larger than 7.0 M. Believe me, crazy sh*t happens above magnitude 7. I've also been in several 6.0 M quakes. Those, by contrast, were little more than brief amusements. Geologists should draw a bright line at magnitude 6.8 (VII) and ask that paleoseismologist how many M 7 quakes has he/she personally experienced.

Here's an excerpt from the the Walla Walla Union on July 17, 1938 describing the 1872 Chelan/North Cascades Earthquake,

And the same article's description of the 1936 Umapine/Stateline Earthquake,

Interesting to find this note about the 1872 Chelan/North Cascades Earthquake in Chleborad and Schuster (1990),

"As was noted by Noson and others (1988), 14 earthquakes, occurring from 1872 to1980, are known to have triggered landslides in Washington. However, information on the distribution and nature of landslides and other ground failures related to historic western Washington earthquakes prior to the 1949 Olympia earthquake is meager. For example, credible accounts of landslides generated by the 1872 North Cascades earthquake, the largest known seismic event in Washington or Oregon (M 7.0-7.5), are limited to the few reports of landslides along the Columbia River east of the Cascades, along the shores of Lake Chelan north of Wenatchee, and in southern British Columbia in the vicinity of Fort Shepard and at Lake Okanagan (Coombs and others,1977). An assertion, based on contemporaneous and later accounts, that the 13-million-m3 Ribbon Cliffs rockslide along the Columbia River, north of Wenatchee, was triggered by the 1872 event (Coombs and others, 1977), is presently disputed. Kienle and others (1978) concluded, on the basis of tree and tree-stump dating, that the slide did not occur during the 1872 earthquake and any significant movement of the landslide debris took place more than 215 years ago."

New seismic hazard database and probability maps. Oregon just completed their latest earthquake hazard assessment, the Oregon Seismic Hazard Database 1.0 (Madin et al., 2021). The map above is an excerpt from Plate 3 showing the "probability over the next 50 years of experiencing shaking strong enough to damage weak buildings". Medium blue to light blue colors, which correspond to Low (6-10%) to Very Low Probability (<5%), appear in the southern Walla Walla Valley, Umatilla Basin, and Columbia Gorge between Wallula Gap and Hood River.

I throw my lot in with the lower-estimate camp, based on my review of the historical record and 20+ years of experience working in the unconsolidated sediments of the greater Columbia Basin.

The evidence of liquefaction in Pliocene, Pleistocene, and Holocene sediments from Kettle Falls, WA to The Dalles, OR and from Lewiston, ID to White Swan, WA is almost non-existent. In fact, the region's most obvious liquefaction features (meter-scale t-shaped sand blows) are located way up in Priest River, ID in clay-rich proglacial lake sediments obviously involved in landsliding.

Local deformation along local faults, yes. Widespread liquefaction, no. Small earthquakes, yes. Large earthquakes, no.

The Seattle Office of the U.S. Geological Survey is currently assessing liquefaction evidence in the Yakima Fold Belt by examining and dating deformed strata exposed in the walls of shallow trenches they excavate across scarps found in lidar images. Its not yet clear what all they have found, but early reports suggest liquefaction features are present in southern Pasco Basin and perhaps elsewhere. The group documented no liquefaction in trenches in the western Boylston Mountains (Barnett et al., 2013) or at Wenas Creek (Sherrod et al., 2013). The jury is out on whether YFB structures are capable of "strong shaking" (>7.0 M).

Trenching and age dating of sediments is today the way Geomorphologists and Paleoseismologists assess fault activity and the threat faults pose to local infrastructure and human safety. One limitation of fault trenching is that trenches are shallow, typically just a few meters deep. Shallow trenches provide targeted views of the subsurface, which may be better than natural outcrops. But trenches are a compromise. Cost, safety issues, environmental impacts, and landowner restrictions all factor into the decision of where and how deep to dig. Faults and the strata they cut reach much deeper than the view a trench provides. The good stuff might be 5m, 50m, or 500m down and unreachable by the excavator's bucket. Also, scarps with clear surface expression (lidar imagery, aerial photos) may not actually correspond to the main fault. Trenches sometimes miss the mark. Nevertheless, trenches are the closest thing we've got to drilling and high wall mining - and they're a whole lot cheaper than even the simplest drill rig or open pit operation.

Tight corridors. Light blue ovals are my prediction of where liquefaction evidence in the Yakima Fold Belt will be reported by USGS in the future. Liquefaction will be limited to tight corridors at low elevations in unconsolidated sediments along mapped faults, consistent with modest shaking (MCE <M 6.8, <VIII).

New avenues for exploration. Faults in the Yakima Fold Belt have been growing over the past 10+ million years, presumably generating large earthquakes all the while. So why have geologists not found widespread, unambiguous evidence of soft sediment deformation ("seismites") in the a.) Ringold, Ellensburg, Thorp Formations, b.) borehole logs and physical cores stored at Hanford, c.) Ocean Drilling Program cores collected offshore of the Columbia River mouth, d.) fine grained non-flood deposits in the Channeled Scabland, e.) fine grained Pliocene and Pleistocene deposits just beyond the floodway margins (i.e., Kittitas and Naches-Tieton Valleys), f.) cores from Cascade Range alpine lakes and other lakes in Eastern Washington, g.) Okanogan/Okanagan Valley glaciolacustrine sediments, h.) Glacial Lake Missoula beds, and i.) thick bodies of Holocene alluvium in Eastern Washington. Disturbed beds in Lindsey's detailed stratigraphic sections at White Bluffs are not extraordinary (mostly confined to flooding surfaces). Newcomb's hardly mention any deformation (Newcomb, 1972). Are a few trench logs through surface faults enough or does inconclusive trench logs create controversy (i.e., West and Shaffer/GEI, 1988 work)?

Attention graduate students: Start looking here. The Kittitas Valley and Naches-Tieton Valley are two basins that do not contain flood deposits, but do contain significant volumes of young, semi-consolidated, fine grained sediment deposited during growth of the Yakima Folds. Both lie on the east slope of the Cascade Mountains in close proximity to active faults. A field-based survey for features attributable to strong shaking along the Yakima River corridor (Cle Elum to Richland) would be cheap and relatively quick to do - a perfect project for a couple of field-savvy graduate students. Simply plot the types of deformation features found along a longitudinal profile of the river.