Graduation date: 1997

The analysis of marsh cores from the tidal zones of the Siuslaw, Umpqua, and Coos River systems on the south-central Oregon coast provides supporting evidence of coseismic subsidence resulting from megathrust earthquakes and reveals the landward extent of the zero-isobase. The analysis is based on lithostratigraphy, paleotidal indicators, microfossil paleotidal indicators, and radiocarbon age. Coseismic activity is further supported by the presence of anomalous thin sand layers present in certain cores. The analysis of diatom assemblages provides evidence of relative sea-level displacement on the order of 1 to 2 m. The historic quiescence of local synclinal structures in the Coos Bay area together with the evidence of prehistoric episodic burial of wetland sequences suggests that the activity of these structures is linked to megathrust releases. The distribution of cores containing non-episodically buried marshes and cores that show episodically buried wetlands within this area suggests that the landward extent of the zero-isobase is between 100 km and 120 km from the trench. The zero-isobase has a minimum width of 10 to 15 km. Radiocarbon dating of selected buried peat sequences yields an estimated recurrence interval on the order of 400 years. The apparent overlapping of the landward margin of both the upperplate deformation zone (fold and/or thrust fault belt) and the landward extent of the zero-isobase is interpreted to represent the landward limit of the locked zone. The earthquake magnitude is estimated to be 8.5 based on an arbitrary rupture length of 200 km and a locked zone width of 105 km. The identification of the zero-isobase on the southcentral Oregon coast is crucial to the prediction of regional coseismic subsidence and tsunami hazards, the testing of megathrust dislocation models, and the estimation of megathrust rupture areas and corresponding earthquake magnitudes in the Cascadia Margin.

Nine west-northwest-trending faults on the continental margin of Oregon and Washington, between 43° 05'N and 470 20'N latitude, have been mapped using seismic reflection, sidescan sonar, submersibles, and swath bathymetry. Five of these oblique faults are found on both the Juan de Fuca and North American plates, and offset abyssal plain sedimentary units left-laterally from 2.0 to 5.5 km. These five faults extend 8-18 km northwestward from the deformation front. The remaining four faults, found only on the North American plate, are also inferred to have a left-lateral slip sense. The age of the Wecoma fault on the abyssal plain is 600±50 ka, and has an average slip rate of 7-1 0 mm/year. Slip rates of the other four abyssal plain faults are 5.5 ± 2 - 6. 7 ± 3 mm/yr. These faults are active, as indicated by offset of the youngest sedimentary units, surficial fault scarps, offsets of surficial channels, and deep fluid venting. All nine faults have been surveyed on the continental slope using SeaMARC 1A sidescan sonar, and three of them were surveyed with a high-resolution AMS 150 sidescan sonar on the continental shelf off central Oregon. On the continental slope, the faults are expressed as linear, high-angle WNW trending scarps, and WNW trending fault-parallel folds that we interpret as flower structures. Active structures on the shelf include folds trending from NNE to WNW and associated flexural slip thrust faulting; NNW to N trending right-lateral strike-slip faults; and WNW trending left-lateral strike-slip faults. Some of these structures intersect the coast and can be correlated with onshore Quaternary faults and folds, and others are suspected to be deforming the coastal region. These structures may be contributing to the coastal marsh stratigraphic record of co-seismic subsidence events in the Holocene. We postulate that the set of nine WNW trending left-lateral strike-slip faults extend and rotate the forearc clockwise, absorbing most or all of the arc parallel component of plate convergence. The high rate of forearc deformation implies that the Cascadia forearc may lack the rigidity to generate M > 8.2 earthquakes. From a comparison of Cascadia seismogenic zone geometry to data from circum-Pacific great earthquakes of this century, the maximum Cascadia rupture is estimated to be 500 to 600 km in length, with a 150-400 km rupture length in best agreement with historical data. / Graduation date: 1994

The Clarno Formation is a series of volcanic, volcaniclastic, and related intrusive rocks located in central Oregon. It is the westernmost extent of a broader Eocene magmatic belt that covers much the western United States. The magmatic belt stretches eastward from Oregon to western South Dakota, and from the Canadian Yukon to northern Nevada. While once attributed to subduction of the Farallon Plate under North America, more recent work suggests that a more complex tectonic regime involving extension was in place during the early Cenozoic. In the vicinity of Mitchell, Oregon, the Clarno Formation is well represented along with Mesozoic metamorphic and sedimentary units, and younger Tertiary volcanic and volcaniclastic units. In this area, Clarno volcanic activity occurred from ~52-42 Ma, producing mostly andesites and related volcaniclastic rocks. The Mitchell area is also underlain by related intrusive bodies ranging from basalt to rhyolite in composition. The Clarno was most active at ~49 Ma, and is dominantly calcalkaline. In addition, there are several coeval alkaline and peraluminous intrusives also scattered throughout the Clamo Formation. While these suites are less voluminous than the calc-alkaline magmatism, they offer insight into the tectonic and magmatic processes at work in this area during the Eocene. Whereas silicic intrusions are common in the Clarno, the high-silica rhyolite dike on the south face of Scott Butte is unusual due to its large garnet phenocrysts. The existence of primary garnet in rhyolitic magmas precludes middle to upper crustal genesis, a common source for silicic magmas. ⁴⁰Ar/³⁹Ar age determinations of the biotite indicate an age of ~51 Ma. This is after andesitic volcanism had commenced, but prior to the most active period of extrusion. The presence of the almandine garnet indicates that the dike represents partial melting of lower crustal (18-25 km) material. The presence of a high field strength element (HFSE) depletion commonly associated with subduction are magmatism indicates that either the source material had previously been metasomatised, or that some subduction melts/fluids (heat source) mixed with the crustal melt. Two alkaline suites, a high-K calc-alkaline basanite (Marshall and Corporate Buttes) and alkaline minette/kersantite lamprophyres (near Black Butte and Mud Creek), were emplaced ~49 Ma, during the height of calc-alkaline activity. The basanite lacks the HFSE depletion common in the other Clarno rocks. Instead it has a HIMU-type (eg. St Helena) ocean island basalt affinity, resulting from partial melting of enriched asthenospheric mantle. In contrast, the lamprophyres represent hydrous partial melts of metasomatized litho spheric mantle veins and bodies. Alkaline magmatism was not limited to the most active periods of calc-alkaline activity. The emplacement of an alkali basalt (Hudspeth Mill intrusion) at ~45 Ma occurred four million years after the largest pulse of volcanism, but still during calcalkaline activity. This alkali basalt represents partial melting of metasomatized lithospheric mantle. The occurrence of these alkaline suites coeval with the calc-alkaline activity is significant in that it disputes prior subduction theories for the broader Eocene magmatism that are based on spatial and temporal variations from calc-alkaline to alkaline magmatism. These suites also give further insight into the complex tectonic regime that existed in Oregon during the Eocene. The occurrence of asthenospheric melts not caused by fluid fluxing, along with lower lithospheric alkaline melts, are normally associated with extension. Extension provides these magmas with both the mechanism for melting, and the ability to reach shallow crust with little or no contamination. Extension is in agreement with both White and Robinson's (1992) interpretation that most Clarno Formation deposition occurred in extensional basins, and with other provinces in the broader Eocene magmatic belt. / Graduation date: 2002

Beach shoreline retreat induced by coseismic subsidence in the Cascadia subduction zone is an important post-earthquake hazard. Sand on a beach acts as a buffer to wave attack, protecting dunes, bluffs and terraces. The loss of sand from a beach could promote critical erosion of the shoreline. This study was initiated in order to estimate the potential amount of post subsidence shoreline retreat on a regional scale in the Central Cascadia Margin. The study area is a 331 km stretch of coastline from Copalis, Washington to Florence, Oregon. Several erosion models were evaluated, and the Bruun model was selected as the most useful to model shoreline retreat on a regional scale in the Central Cascadia Margin. There are some factors that this model does not address, such as longshore transport of sediment and offshore bottom shape, but for this preliminary study it is useful for estimating regional retreat. The range of parameter input values for the Bruun model include: the depth of closure (h) range from 15 m to 20 m water depth; the cross-shore distance (L) range from 846 m to 5975 m; and the estimated subsidence amount (S) range from O m to 1.5 m. The minimum to maximum range of post-subsidence shoreline retreat is 142 to 531 m in the Columbia River cell, 56 to 128 m in the Cannon Beach cell, 38 to 149 m in the Tillamook cell, 25 to 91 m in the Pacific City cell, 11 to 126 m in the Lincoln City cell, 30 to 147 m in the Otter Rock cell, 0 to 165 m in the Newport cell, 0 to 76 m in the Waldport cell, and 0 m in the Winchester cell. Results of the study suggest that many of the beaches in the study area are at risk of beach and personal property loss. Beach communities could limit the amount of potential damage in these areas through coastal zone planning.