IMS-9
Relative Earthquake Hazard Maps
for Selected Urban Areas in Western Oregon

Ashland, Cottage Grove, Grants Pass, Roseburg, Sutherlin-Oakland

By Ian P. Madin and Zhenming Wang, Oregon Department of Geology and Mineral Industries

This is one of four companion publications presenting earthquake hazard maps for small to intermediate-sized communities in western Oregon. Each publication includes a geographic grouping of urban areas.

NOTICE
The results and conclusions of this report are necessarily based on limited geologic and geophysical data. The hazards and data are described in this report. At any given site in any map area, site-specific data could give results that differ from those shown on this map. This report cannot replace site-specific investigations. Some appropriate uses are discussed in the report. The hazards of an individual site should be assessed through geotechnical or engineering geology investigation by qualified practitioners.

INTRODUCTION
Since the late 1980s, the understanding of earthquake hazards in the Pacific Northwest has significantly increased. It is now known that Oregon may experience damaging earthquakes much larger than any that have been recorded in the past (Atwater, 1987; Heaton and Hartzell, 1987; Weaver and Shedlock, 1989; Yelin and others, 1994). Planning the response to earthquake disasters and strengthening homes, buildings, and lifelines for power, water, communication, and transportation can greatly reduce the impact of an earthquake. These measures should be based on the best possible forecast of the amount and distribution of future earthquake damage. Earthquake hazard maps such as those in this publication provide a basis for such a forecast.
The amount of damage sustained by a building during a strong earthquake is difficult to predict and depends on the size, type, and location of the earthquake, the characteristics of the soils at the building site, and the characteristics of the building itself. At present, it is not possible to accurately forecast the location or size of future earthquakes. It is possible, however, to predict the behavior of the soil at any particular site In fact, in many major earthquakes around the world, a large amount of the damage has been due to the behavior of the soil. In this report, "soil" means the relatively loose and soft geologic material that typically overlies solid bedrock in western Oregon.
The maps in this report identify those areas in selected Oregon communities that will be at higher risk, relative to other areas, during a damaging earthquake. The analysis is based on the behavior of the soils and does not depict the absolute earthquake hazard at any particular site.

It is quite possible that, for any given earthquake, damage in even the highest hazard areas will be light. On the other hand, during an earthquake that is stronger or much closer than our design parameters, even the lowest hazard categories could experience severe damage.

This report includes a nontechnical description of how the maps were made and how they might be used. More technical information on the mapmaking methods is contained in the Appendix.
The printed report includes paper-copy Relative Earthquake Hazard Maps for each urban area, overlaid on U.S. Geological Survey topographic base maps at the scale of 1:24,000. In addition, for each area, three individual hazard component maps are included as digital data files on CD-ROM. The digital data are in two formats: (1) high-resolution -.JPG files (bitmap images) that can be viewed with many image viewers or word processors and (2) MapInfo and ArcView GIS vector files.
These maps were produced by the Oregon Department of Geology and Mineral Industries and were funded by the State of Oregon and the U.S. Geological Survey (USGS), Department of the Interior, under USGS award #1434-97-GR-03118. The views and conclusions contained in this document are those of the authors and should not be interpreted as necessarily representing the official policies, either expressed or implied, of the U.S. Government.

EARTHQUAKE HAZARD
Earthquakes from three different sources threaten communities in western Oregon (Figure 1).

[See JPEG file "Figure 1.jpg" - Figure 1. Plate-tectonic map of the Pacific Northwest. Oregon is cut in half to show where earthquakes originate below the surface (asterisks).]

These sources are crustal, intraplate, and subduction-zone earthquakes. The most common are crustal earthquakes, which typically occur in the North American plate above the subduction zone at relatively shallow depths of 6-12 mi (10-20 km) below the surface. The March 1993 earthquake at Scotts Mills (magnitude [M] 5.6) (Madin and others, 1993) and the September 1993 Klamath Falls main shocks (M 5.9 and M 6.0) (Wiley and others, 1993) were such crustal earthquakes.
Deeper intraplate earthquakes occur within the remains of the ocean floor (the Juan de Fuca plate) that has been subducted beneath North America. Intraplate earthquakes caused damage in the Puget Sound region in 1949 and again in 1965. This type of earthquake could occur beneath much of western Oregon at depths of 25-37 mi (40-60 km).
Great subduction-zone earthquakes occur around the world where the plates that make up the surface of the Earth collide. When the plates collide, one plate slides (subducts) beneath the other, where it is reabsorbed into the mantle of the planet. The dipping interface between the two plates is the site of some of the most powerful earthquakes ever recorded, often having magnitudes of M 8 to M 9 on the moment magnitude scale. The 1960 Chilean (M 9.5) and the 1964 Great Alaska (M 9.2) earthquakes were subduction-zone earthquakes (Kanamori, 1977). The Cascadia subduction zone, which lies off the Oregon and Washington coasts, has been recognized for many years. No earthquakes have occurred on the Cascadia subduction zone during our short 200-year historical record. However, in the past several years, a variety of studies have found widespread evidence that very large earthquakes have occurred repeatedly in the past, most recently about 300 years ago, in January 1700 (Atwater, 1987; Yamaguchi and others, 1997). The best available evidence indicates that these earthquakes occur, on average, every 500 to 540 years, with an interval between individual events that ranges from 100-300 years to about 1,000 years (Atwater and Hemphill-Haley, 1997). We have every reason to believe that they will continue to occur in the future.
Together, these three types of earthquakes could cause strong shaking through most of western Oregon. Maps are available that forecast the likely strength of shaking for all of Oregon (Geomatrix Consultants, 1995; Frankel and others, 1996; Madin and Mabey, 1996). However, these maps show the expected strength of shaking at a firm site on bedrock and do not include the significant influence of soil on the strength of shaking. They forecast a uniform level of shaking and damage in most communities, and as such they do not provide a useful tool for planning earthquake hazard mitigation measures.

EARTHQUAKE EFFECTS
Damaging earthquakes will occur in the cities and towns of western Oregon. This fact was demonstrated by the Scotts Mills earthquake (M 5.6) in 1993 (Madin and others, 1993). Although we cannot predict when the next damaging earthquake will strike, where it will occur, or how large it will be, we can evaluate the influence of site geology on potential earthquake damage. This evaluation can occur reliably even though the exact sources of earthquake shaking are uncertain.
The most severe damage done by an earthquake is commonly localized. One or more of the following phenomena generally will cause the damage in these areas:
1. Amplification of ground shaking by a "soft" soil column.
2. Liquefaction of water-saturated sand, silt, or gravel creating areas of "quicksand."
3. Landslides triggered by shaking, even on relatively gentle slopes.
These effects can be evaluated before the earthquake occurs, if data are available on the thickness and nature of the geologic materials and soils at the site (Bolt, 1993). Knowing the exact nature and magnitude of these effects is useful to technical professionals, and such data (in digital format) are included in this publication. For others, what is more significant is that these effects increase the damage caused by an earthquake and localize the most severe damage.

HAZARD MAP METHODOLOGY
Selection of map areas
Urban areas were mapped if they had a population greater than 4,000, were in Uniform Building Code (UBC) Seismic Zone 3 or 4, and were not likely to be the subject of a more detailed future hazard mapping program. The goal of this project was to provide an inexpensive general hazard assessment for small communities that could not afford their own mapping program but were not large enough to justify a major state-funded mapping effort. Such major, full-scale projects have been undertaken for the Portland, Salem, Eugene-Springfield, and Klamath Falls urban areas; they typically take several years and cost several hundred thousand dollars. In contrast, this project involved about two weeks of work and a few thousand dollars for each urban area mapped.
For each urban area selected, the hazard map area (inside the thick black line) was defined by the urban growth boundary plus a 3,300-ft (1-km)-wide buffer.

Geologic model
The most important element of any earthquake hazard evaluation is the development of a three-dimensional geologic model. For analysis of the amplification and liquefaction hazards, the most important feature is the thickness of the loose sand, silt, and gravel deposits that usually overlie firm bedrock. For an analysis of the landslide hazard, the steepness of the slopes and presence of existing landslides is important. For each urban area, the geologic model was developed as follows:
The best available geologic mapping was used to determine what geologic materials were present and where they occurred. Air photos were used to help make these decisions where the mapping was poor or of low resolution. All data were plotted digitally on USGS Digital Raster Graphics (DRG) maps (the digital equivalent of USGS 1:24,000-scale topographic maps).
Drillers' logs of water wells were examined to determine the geology beneath the surface and map the thickness of the loose surficial deposits and the depth to firm bedrock. Water wells were located according to the location information provided on the logs, which often is accurate only to within about 1,000 ft. Field location of the individual logs would have been prohibitively expensive.
The water well data were combined with the surface data to produce a three-dimensional geologic model, describing the thickness of the various geologic materials in the top 100 ft (30 m) throughout each urban area. For this procedure, MapInfo and Vertical Mapper Geographic Information System (GIS) software programs were used. The models take the form of a grid of thickness values spaced every 165 ft (50 m).
The resultant models were reviewed by geologists knowledgeable about each area, who judged whether the models were reasonable and consistent with the data.
Existing landslides were mapped where depicted on existing geologic maps or where air photos showed clear signs of landslide topography.
Slope data were derived from USGS Digital Elevations Models (DEMs) with elevation data every 100 ft (30 m). They were then used in MapInfo and Vertical Mapper to map the steepness of slopes.
The details of the local geology and data sources for each urban area are described in the "Urban Area Summaries" section of this report.

Hazard analysis
Ground shaking amplification
The soils and soft sedimentary rocks near the surface can modify bedrock ground shaking caused by an earthquake. This modification can increase (or decrease) the strength of shaking or change the frequency of the shaking. The nature of the modifications is determined by the thickness of the geologic materials and their physical properties, such as stiffness.
This amplification study used a method first developed for the National Earthquake Hazard Reduction Program (NEHRP) and published by the Federal Emergency Management Agency (FEMA, 1995). This method was adopted in the 1997 version of the Uniform Building Code (ICBO [International Conference of Building Officials], 1997) and will henceforth be referred to as the UBC-97 methodology. The UBC-97 methodology defines six soil categories that are based on average shear-wave velocity in the upper 100 ft (30 m) of the soil column. The shear-wave velocity is the speed with which a particular type of ground vibration travels through a material, and can be measured directly by several techniques. The six soil categories are Hard Rock (A), Rock (B), Very Dense Soil and Soft Rock (C), Stiff Soil (D), Soft Soil (E), and Special Soils (F). Category F soils are very soft soils requiring site-specific evaluation and are not mapped in this study, because limited funding precluded any site visits.
For the amplification hazard component maps, we collected shear-wave velocity data (see Appendix for data and methods) at one or more sites in each urban area and used our geologic model to calculate the average shear-wave velocity of each 165-ft (50-m) grid cell in the model. We then assigned a soil category, using the relationships in Table 1.

[See MS-Excel file "Tables 1-4.xls" - Table 1. UBC-97 soil profile types. From ICBO, 1997]

According to the UBC-97 methodology, none of the urban areas in this study had Type A soils. UBC-97 soil category maps for each urban area are presented in the accompanying digital map set.

Liquefaction
Liquefaction is a phenomenon in which shaking of a saturated soil causes its material properties to change so that it behaves as a liquid. In qualitative terms, the cause of liquefaction was described very well by Seed and Idriss (1982): "If a saturated sand is subjected to ground vibrations, it tends to compact and decrease in volume; if drainage is unable to occur, the tendency to decrease in volume results in an increase in pore water pressure, and if the pore water pressure builds up to the point at which it is equal to the overburden pressure, the effective stress becomes zero, the sand loses its strength completely, and it develops a liquefied state."
Soils that liquefy tend to be young, loose, granular soils that are saturated with water (National Research Council, 1985). Unsaturated soils will not liquefy, but they may settle. If an earthquake induces liquefaction, several things can happen: The liquefied layer and everything lying on top of it may move downslope. Alternatively, it may oscillate with displacements large enough to rupture pipelines, move bridge abutments, or rupture building foundations. Light objects, such as underground storage tanks, can float toward the surface, and heavy objects, such as buildings, can sink. Typical displacements can range from centimeters to meters. Thus, if the soil at a site liquefies, the damage resulting from an earthquake can be dramatically increased over what shaking alone might have caused.
The liquefaction hazard analysis is based on the age and grain size of the geologic unit, the thickness of the unit, and the shear-wave velocity. Use of the shear-wave velocity to characterize the liquefaction potential follows Andrus and Stokoe (1997). Liquefaction hazard categories were assigned according to Table 2. 

[See MS-Excel file "Tables 1-4.xls" - Table 2. Liquefaction hazard categories]

In all communities we assumed that the susceptible units were saturated. This is reasonable and conservative, since most of the susceptible units are either alluvial deposits in floodplains, coastal deposits, or silt deposits in areas of low relief and high rainfall in the Willamette Valley.

Earthquake-induced landslides
The hazard due to earthquake-induced landsliding was assessed with slope data derived from USGS DEMs with 100-ft (30-m) data spacing and from mapping of existing slides, either from air photo interpretation or published geologic maps. The analysis was based on methods used by Wang, Y.,  and others (1998) and Wang, Z., and others (1999) but was greatly simplified because no field data were available. Earthquake-induced landslide hazard categories were assigned according to Table 3.

[See MS-Excel file "Tables 1-4.xls" - Table 3. Earthquake-induced landslide hazard zones]

RELATIVE EARTHQUAKE HAZARD MAPS
The Relative Earthquake Hazard Map is a composite hazard map depicting the relative hazard at any site due to the combination of the effects mentioned above. It delineates those areas that are most likely to experience the most severe effects during a damaging earthquake. Areas of highest risk are those with high ground amplification, high likelihood of liquefaction, existing landslides, or slopes steeper than 25°. Planners, lenders, insurers, and emergency responders can use these simple composite hazard maps for general hazard mitigation and response planning.
It is very important to note that the relative hazard map predicts the tendency of a site to have greater or lesser damage than other sites in the area. These zones, however, should not be used as the sole basis for any type of restrictive or exclusionary development policy.
The Relative Earthquake Hazard Maps were created to show which areas will have the greatest tendency to experience damage due to any combination of the three hazards described above. For the purpose of creating the final relative hazard map for each urban area, the zones in each of the three component maps were assigned numerical values according to Table 4.

[See MS-Excel file "Tables 1-4.xls" - Table 4. Hazard zone values assigned to the individual relative earthquake hazard map zones]

For every point (in a 165-ft [30-m] grid spacing) on the map, the zone rating for each individual hazard type was squared, and the resulting numbers were added together. Then the square root of this sum was taken and rounded to the nearest whole number. A result of 4 or more was assigned to Zone A, 3 to Zone B, 2 to Zone C, and 1 to Zone D.
While the production of the individual hazard maps is different from previous DOGAMI relative earthquake studies (Wang and Priest, 1995; Wang and Leonard, 1996; Mabey and others, 1997), the method of production of the final relative hazard map is very similar. Thus, these relative hazard maps are directly comparable to DOGAMI studies in Eugene-Springfield, Portland, Salem, and Siletz Bay.
The GIS techniques used to develop these maps involved several changes between vector data and raster data, with a data grid cell size of 165 ft (50 m) for the raster data. As a result, the relative hazard maps often had numerous zones that were very small, and probably not significant. The final maps were hand-edited to remove all hazard zones that covered less than 1 acre.

USE OF RELATIVE EARTHQUAKE HAZARD MAPS
The Relative Earthquake Hazard Maps delineate those areas most likely to experience damage in a given earthquake. This information can be used to develop a variety of hazard mitigation strategies. The information should, however, be carefully considered and understood, so that inappropriate use can be avoided.

Emergency response and hazard mitigation
One of the key uses of these maps is to develop emergency response plans. The areas indicated as having a higher hazard would be the areas where the greatest and most abundant damage will tend to occur. Planning for disaster response will be enhanced by the use of these maps to identify which resources and transportation routes are likely to be damaged.

Land use planning and seismic retrofit
Efforts and funds for both urban renewal and strengthening or replacing older and weaker buildings can be focused on the areas where the effects of earthquakes will be the greatest. The location of future urban expansion or intensified development should also consider earthquake hazards.
Requirements placed on development could be based on the hazard zone in which the development is located. For example, the type of site-specific earthquake hazard investigation that is required could be based on the hazard.

Lifelines
Lifelines include road and access systems including railroads, airports, and runways, bridges, and over- and underpasses, as well as utilities and distribution systems. The Relative Earthquake Hazard Map and its component single-hazard maps are especially useful for expected-damage estimation and mitigation for lifelines. Lifelines are usually distributed widely and often require regional as opposed to site-specific hazard assessments. The hazard maps presented here allow quantitative estimates of the hazard throughout a lifeline system. This information can be used for assessing vulnerability as well as deciding on priorities and approaches for mitigation.

Engineering
The hazard zones shown on the Relative Earthquake Hazard Maps cannot serve as a substitute for site-specific evaluations based on subsurface information gathered at a site. The calculated values of the individual component maps used to make the Relative Hazard Maps may, however, be used to good purpose in the absence of such site-specific information, for instance, at the feasibility-study or preliminary-design stage. In most cases, the quantitative values calculated for these maps would be superior to a qualitative estimate based solely on lithology or non-site-specific information. Any significant deviation of observed site geology from the geologic model used in the analyses indicates the need for additional analyses at the site.

Relative hazard
It is important to recognize the limitations of a Relative Earthquake Hazard Map, which in no way includes information with regard to the probability of damage to occur. Rather, it shows that when shaking occurs, the damage is more likely to occur, or be more severe, in the higher hazard areas. The exact probability of such shaking to occur is yet to be determined.
Neither should the higher hazard areas be viewed as unsafe. Except for landslides, the earthquake effects that are factored into the Relative Earthquake Hazard Map are not life threatening in and of themselves. What is life threatening is the way that structures such as buildings and bridges respond to these effects.
The map depicts trends and tendencies. In all cases, the actual threat at a given location can be assessed only by some degree of site-specific assessment. This is similar to being able to say demographically that a zip code zone contains an economic middle class, but within that zone there easily could be individuals or neighborhoods significantly richer or poorer.
Because the maps exist as "layers" of digital GIS data, they can easily be combined with earthquake source information to produce earthquake damage scenarios. They can also be combined with probabilistic or scenario bedrock ground shaking maps to provide an assessment of the absolute level of hazard and an estimate of how often that level will occur. Finally, the maps can also be easily used in conjunction with GIS data for land use or emergency management planning.
This study does not address the hazard of tsunamis that exists in areas close to the Oregon coast and is also earthquake induced. The Oregon Department of Geology and Mineral Industries has published separate tsunami hazard maps on this subject (Priest, 1995; Priest and Baptista, 1995).

URBAN AREA SUMMARIES

Ashland Urban Area
The Ashland geologic model was developed using recently completed 1:24,00-scale geologic mapping by Dr. Jad D'Alllura of Southern Oregon University.  Subsurface geology was interpreted from 77 approximately located water wells.
The geology consists of a thin, irregular body of varied Quaternary alluvial clay, sand, and gravel (Qac) deposited on an irregular surface of Jurassic through Eocene bedrock (KJg).
The geologic model consists of a single body of Quaternary alluvium.
Shear-wave velocity is assigned as follows:
Qaf	Two direct measurements, 194 and 327 m/sec, average 260 m/sec.
KJg	Two direct measurements; weathered zone (grus), 7-13 m thick, is 640-720 m/sec (average 680 m/sec); unweathered rock is 1,015-1,220 m/sec, average 1,117 m/sec.
The relatively thin and high-velocity alluvium has low amplification hazard; the bedrock areas have none.
The Quaternary alluvium is typically clayey or gravelly and is likely to have very low liquefaction susceptibility.
The steeper slopes in the western part of the area are underlain by Jurassic granite and are unlikely to be very susceptible to landslides, except perhaps where there is thick grus (a layer of decomposed granite). The more moderate slopes along the eastern edge of the area consist of Eocene rocks overlying Cretaceous metamorphic rocks. One large slide occurs along this contact in the northeast edge of the area.
The majority of the area is in relative hazard Zone D, reflecting low amplification, no liquefaction, and low landslide hazard. Small areas of higher hazard are associated largely with steep slopes and existing landslides.

Cottage Grove Urban Area
The geologic model for the Cottage Grove area was developed using published surface geologic mapping (Walker and McLeod, 1991), air photo interpretation, and subsurface data from 69 approximately located water wells.
The geology consists of Pleistocene and Holocene alluvial sand and gravel (Qac) deposited on Eocene-Miocene sedimentary and basalt sedimentary bedrock (Tbs). The model consists of a single body of Qac over bedrock.
Shear-wave velocities were assigned as follows:
Qac	Two direct measurements, 187 and 219 m/sec, average 203 m/sec.
Tbs	Two direct measurements, 973 and 1,270 m/sec, average 1,121 m/sec.
Amplification is low to moderate in the bedrock slopes adjacent to the floodplains of the Coast Fork Willamette and Row Rivers and moderate in the floodplains, where the alluvium is fairly thick.
Liquefaction hazard is nil.
Earthquake-induced landslide hazard is generally low on the floodplains of the Coast Fork Willamette and Row Rivers and moderate on the adjacent slopes. High landslide hazard is restricted largely to areas of existing landslides.
Most of the area is in relative hazard Zone D, with a large area of Zone C along the valley floor, and some areas of Zone B in the hills associated with existing landslides.

Grants Pass Urban Area
The Grants Pass geologic model was developed using surface geologic data from Walker and McLeod (1991) and Ramp and Peterson (1979) and air photo interpretation. Subsurface data were interpreted from 224 approximately located water wells.
The geology consists of Quaternary alluvial gravel and sand (Qac) filling the valley of the Rogue River. The gravel overlies granite and metamorphic bedrock; the granite is typically covered by a blanket of decomposed material (grus) up to tens of meters thick. The geologic model consists of a body of gravel, a body of grus, and bedrock (Kjg).
Shear-wave velocity was assigned as follows:
Qac	Three direct measurements, from 257 to 554 m/sec, average 394 m/sec.
Grus	Two direct measurements, 868 and 925 m/sec, average 896 m/sec.
Amplification is low to nil, because the Qac alluvium is relatively high-velocity.
Liquefaction is nil, because the Qac is dense and gravelly.
Earthquake-induced landslide hazard is mostly low, with some moderate hazard in the hills adjacent to the Rogue River valley floor and a few small areas of high hazards associated with the steepest slopes in the hills.
Most of the area is in relative hazard Zone D, with a few spots of higher hazard associated with the steepest slopes.

Roseburg Urban Area
The Roseburg geologic model was derived using unpublished digital 1:24,000-scale geologic mapping provided by Dr. Ray Wells of the U.S. Geological Survey and subsurface data from 40 approximately located water wells. Quaternary alluvial terraces and alluvium (Qac) overlie complex, largely Eocene and older bedrock (Tbv). The geologic model consists of a body of Qac over bedrock.
Shear-wave velocity was assigned as follows:
Qac	One direct measurement, 181 m/sec.
Tbv	One direct measurement, 944m/sec.
Amplification hazard is low to nil, since most of the area is either bedrock, or fairly thin alluvium.
Liquefaction hazard is nil, because the alluvium is gravelly and thin.
Earthquake-induced landslide hazard is generally low to moderate with a few small areas of high hazard, mostly associated with existing slides.
Most of the area is in relative hazard Zone D, with some small areas of higher hazard associated with existing landslides and steep slopes.

Sutherlin-Oakland Urban Area
The Sutherlin-Oakland geologic model was derived using digital 1:24,000-scale geologic mapping provided by Dr. Ray Wells of the U.S. Geological Survey and subsurface data from 35 approximately located water wells. Quaternary alluvial terraces and alluvium (Qaf) overlie complex, largely Eocene and older bedrock  (Tbv). The thickness model consists of a body of Qaf over bedrock.
Shear-wave velocities were assigned as follows:
Qaf	Two direct measurements, 198 and 426 m/sec, average 312 m/sec.
Tbv	Two direct measurements, 842 and 1,079 m/sec, average 960 m/sec.
 Amplification hazard is nil to low, because the area is either bedrock or very thin alluvium.
Liquefaction hazard is nil, because the area is either bedrock or very thin, relatively dense alluvium.
Earthquake-induced landslide hazard is generally low to moderate with a few areas of high hazard associated with existing landslides and very steep slopes.
Most of the area is in relative hazard Zone D, with some small areas of higher hazard associated with existing landslides and steep slopes.

ACKNOWLEDGMENTS
Geological models were reviewed by Marshall Gannet and Jim O'Connor of the USGS Water Resources Division, Ken Lite of the Oregon Water Resources Division, Dr. Ray Wells of the US Geological Survey, Dr. Curt Peterson of Portland State University, Dr. Jad D'Allura of Southern Oregon University, Dr John Beaulieu, Gerald Black and Dr. George Priest of the Oregon Department of Geology and Mineral Industries. The reports were reviewed by Gerald Black and Mei Mei Wang. Marshall Gannet and Jim O'Connor provided unpublished digital geologic data which were helpful in building the geologic models. Dr. Marvin Beeson provided unpublished geologic mapping. We are very grateful to all of these individuals for their generous assistance.

REFERENCES CITED
Andrus, R.D., and Stokoe, K.H., 1997, Liquefaction resistance based on shear-wave velocity (9/18/97 version), in Youd, T.L., and Idriss, I.M., eds., Proceedings of the NCEER Workshop on Evaluation of Liquefaction Resistance of Soils, Jan. 4-5, Salt Lake City, Utah: Buffalo, N.Y., National Center for Earthquake Engineering Research Technical Report NCEER-97-0022, p. 89-128.
Atwater, B.F., 1987, Evidence for great Holocene earthquakes along the outer coast of Washington State: Science, v. 236, p. 942-944.
Atwater, B.F., and Hemphill-Haley, 1997, Recurrence intervals for great earthquakes of the past 3,500 years at northeastern Willapa Bay, Washington: U.S. Geological Survey Professional Paper 1576, 108 p.
Baldwin, E.M., 1964, Geology of the Dallas and Valsetz quadrangles, rev. ed.: Oregon Department of Geology and Mineral Industries Bulletin 35, 56 p., 1 map 1:62,500.
Beaulieu, J.D., 1977, Geologic hazards of parts of northern Hood River, Wasco, and Sherman Counties, Oregon: Oregon Department of Geology and Mineral Industries Bulletin 91, 95 p., 10 maps.
Beaulieu, J.D., and Hughes, P.W., 1975, Environmental geology of western Coos and Douglas Counties, Oregon: Oregon Department of Geology and Mineral Industries Bulletin 87, 148 p., 16 maps.
---1976, Land use geology of western Curry County, Oregon: Oregon Department of Geology and Mineral Industries Bulletin 90, 148 p., 12 maps.
Bela, J.L, 1981, Geology of the Rickreall, Salem West, Monmouth, and Sidney 71/2' quadrangles, Marion, Polk, and Linn Counties, Oregon: Oregon Department of Geology and Mineral Industries Geological Map Series GMS-18, 2 pls., 1:24,000.
Bolt, B.A., 1993, Earthquakes: New York, W.H. Freeman and Co., 331 p.
Bretz, J.H., Smith, H.T.U., and Neff, G.E., 1956, Channeled Scabland of Washington: New data and interpretations: Geological Society of America Bulletin, v. 67, no. 8, p. 957-1049.
Brownfield, M.E., 1982, Geologic map of the Sheridan quadrangle, Polk and Yamhill Counties, Oregon: Oregon Department of Geology and Mineral Industries Geological Map Series GMS-23, 1:24,000.
Brownfield, M.E., and Schlicker, H.G., 1981, Preliminary geologic map of the McMinnville and Dayton quadrangles, Oregon: Oregon Department of Geology and Mineral Industries Open-File Report O-81-6, 1:24,000. 
FEMA (Federal Emergency Management Agency), 1995, NEHRP recommended provisions for seismic regulations for new buildings, 1994 edition, Part 1: Provisions: Washington, D.C., Building Seismic Safety Council, FEMA Publication 222A / May 1995, 290 p.
Frankel, A., Mueller, C., Barnhard, T., Perkins, D., Leyendecker, E.V., Dickman, N., Hanson, S., and Hopper, M., 1996, National seismic hazard maps, June 1996 documentation: U.S. Geological Survey Open-File Report 96-532, 110 p.
Gannet, M.W., and Caldwell, R.R., 1998, Geologic framework of the Willamette Lowland aquifer system: U.S. Geological Survey Professional Paper1424-A, 32 p., 8 pls.
Geomatrix Consultants, Inc., 1995, Seismic design mapping, State of Oregon: Final Report to Oregon Department of Transportation, Project no. 2442, var. pag.
Heaton, T.H., and Hartzell, S.H., 1987, Earthquake hazards on the Cascadia subduction zone: Science, v. 236, no. 4798, p. 162-168.
ICBO (International Conference of Building Officials), 1997, 1997 Uniform building code, v. 2, Structural engineering design provisions: Whittier, Calif., International Conference of Building Officials, 492 p.
Kanamori, H., 1977, The energy release in great earthquakes: Journal of Geophysical Research, v. 82, p. 2981-2987.
Mabey, M.A., Black, G.L., Madin, I.P., Meier, D.B., Youd, T.L., Jones, C.F., and Rice, J.B., 1997, Relative earthquake hazard map of the Portland metro region, Clackamas, Multnomah, and Washington Counties, Oregon: Oregon Department of Geology and Mineral Industries Interpretive Map Series IMS-1, 1:62,500.
Madin, I.P., and Mabey, M.A., 1996, Earthquake hazard maps for Oregon: Oregon Department of Geology and Mineral Industries Geological Map Series GMS-100.
Madin, I.P., Priest, G.R., Mabey, M.A., Malone, S., Yelin, T.S., and Meier, D., 1993, March 25, 1993, Scotts Mills earthquake-western Oregon's wake-up call: Oregon Geology, v. 55, no. 3, p. 51-57.
National Research Council, Commission on Engineering and Technical Systems, Committee on Earthquake Engineering, 1985, Liquefaction of soils during earthquakes: Washington, D.C., National Academy Press, 240 p.
O'Connor, J.E., Sarna-Wojcicki, A., Wozniak, K.C., Polette, D.J., and Fleck, R.J., in press, Origin, extent, and thickness of Quaternary geologic units in the Willamette Valley, Oregon: U.S. Geological Survey Professional Paper 1620.
Priest, G.R., 1995, Explanation of mapping methods and use of the tsunami hazard maps of the Oregon coast: Oregon Department of Geology and Mineral Industries Open-File Report O-95-67, 95 p.
Priest, G.R., and Baptista, A.M., 1995, Tsunami hazard maps of coastal quadrangles, Oregon: Oregon Department of Geology and Mineral Industries Open-File Report O-95-09 through O-95-66 (amended 1997 by O-97-31 and O-97-32), 56 quadrangle maps (as amended).
Ramp, L., and Peterson, N.V., 1979, Geology and mineral resources of Josephine County, Oregon: Oregon Department of Geology and Mineral Industries Bulletin 100, 45 p., 3 geologic maps.
Schlicker, H.G., Deacon, R.J., Beaulieu, J.D., and Olcott, G.W., 1972, Environmental geology of the coastal region of Tillamook and Clatsop Counties; Oregon Department of Geology and Mineral Industries Bulletin 74, 164 p., 18 pls.
Schlicker, H.G., Deacon, R.J., Newcomb, R.C., and Jackson, R.L., 1974, Environmental geology of coastal Lane County, Oregon: Oregon Department of Geology and Mineral Industries Bulletin 85, 116 p., 3 maps.
Seed, H.B., and Idriss, I.M., 1982, Ground motions and soil liquefaction during earthquakes: Earthquake Engineering Institute Monograph, 134 p. 
Snavely, P.D., Jr., MacLeod, N.S., Wagner, H.C., and Rau, W.W., 1976, Geologic map of the Cape Foulweather and Euchre Mountain quadrangles, Lincoln County, Oregon: U.S. Geological Survey Miscellaneous Investigations Series Map I-868, 1:62,500.
Ticknor, R., 1993, Late Quaternary crustal deformation on the central Oregon coast as deduced from uplifted wave-cut platforms: Bellingham, Wash., Western Washington University master's thesis, 70 p.
Trimble, D.E., 1963, Geology of Portland, Oregon, and adjacent areas: U.S. Geological Survey Bulletin 1119, 119 p.
Waitt, R.B., 1985, Case for periodic, colossal jökulhlaups from Pleistocene glacial Lake Missoula: Geological Society of America Bulletin, v. 96, no. 10, p. 1271-1286. 
Walker, G.W., and McLeod, N.S., 1991, Geologic map of Oregon: U.S. Geological Survey Special Geologic Map, 1:500,000.
Wang, Y., Keefer, D.K., and Wang, Z., 1998, Seismic hazard mapping in Eugene-Springfield, Oregon: Oregon Geology, v. 60, no. 2, p. 31-41.
Wang, Y., and Leonard, W.J., 1996, Relative earthquake hazard maps of the Salem East and Salem West quadrangles, Marion and Polk Counties, Oregon: Oregon Department of Geology and Mineral Industries Geological Map Series GMS-105, 1:24,000.
Wang, Y., and Priest, G.R., 1995, Relative earthquake hazard maps of the Siletz Bay area, coastal Lincoln County, Oregon: Oregon Department of Geology and Mineral Industries Geological Map Series GMS-93, 1:12,000 and 1:24,000.
Wang, Z., Wang, Y., and Keefer, D.K., 1999, Earthquake-induced rockfall and slide hazard along U.S. Highway 97 and Oregon Highway 140 near Klamath Falls, Oregon, in Elliott, W.M., and McDonough, P., eds., Optimizing post-earthquake lifeline system reliability. Proceedings of the 5th U.S. Conference on Lifeline Earthquake Engineering, Seattle, Wash., August 12-14, 1999: Reston, Va., American Society of Civil Engineers, Technical Council on Lifeline Earthquake Engineering Monograph 16, p. 61-70.
Weaver, C.S., and Shedlock, K.M., 1989, Potential subduction, probable intraplate, and known crustal earthquake source areas in the Cascadia subduction zone, in Hayes, W.W., ed., Third annual workshop on earthquake hazards in the Puget Sound/Portland area, proceedings of Conference XLVIII: U.S. Geological Survey Open-File Report 89-465, p. 11-26.
Wiley, T.J., Sherrod, D.R., Keefer, D.K., Qamar, A., Schuster, R.L., Dewey, J.W., Mabey, M.A., Black, G.L., and Wells, R.E., 1993, Klamath Falls earthquakes, September 20, 1993-including the strongest quake ever measured in Oregon: Oregon Geology, v. 55, no. 6, p. 127-134.
Wilkinson, W.D., Lowry, W.D., and Baldwin, E.M., 1946, Geology of the St. Helens quadrangle, Oregon: Oregon Department of Geology and Mineral Industries Bulletin 31, 39 p., 1 map, 1:62,500.
Yamaguchi, D.K., Atwater, B.F., Bunker, D.E., Benson, B.E., and Reid, M.S., 1997, Tree-ring dating the 1700 Cascadia earthquake: Nature, v. 389, p. 922.
Yeats, R.S., Graven, E.P., Werner, K.S., Goldfinger, C., and Popowski, T., 1991, Tectonics of the Willamette Valley, Oregon: U.S. Geological Survey Open-File Report 91-441-P, 47 p.
Yelin, T.S., Tarr, A.C., Michael, J.A., and Weaver, C.S., 1994, Washington and Oregon earthquake history and hazards: U.S. Geological Survey Open-File Report 94-226-B, 11 p.


APPENDIX

1. GEOLOGIC UNITS USED IN TABLE A-1
Qaf	Fine-grained Quaternary alluvium; river and stream deposits of sand, silt, and clay
Qac	Coarse-grained Quaternary alluvium; river and stream deposits of sand and gravel
Qmf	Fine-grained Quaternary Missoula flood deposits; sand and silt left by catastrophic glacial floods
Qmc	Coarse-grained Quaternary Missoula flood deposits; sand and gravel left by catastrophic glacial floods
Qmf1	Fine-grained Quaternary Missoula flood deposits; upper, oxidized low-velocity layer
Qmf2	Fine-grained Quaternary Missoula flood deposits; lower, reduced high-velocity layer
Qe	Quaternary estuarine sediments; silt, sand, and mud deposited in bays and tidewater of major rivers
Qs	Quaternary sands; beach and dune deposits along the coast
Qmt	Quaternary marine terrace deposits; sand and silt deposited during previous interglacial periods
QPe	Pleistocene estuarine sediments; older sand and mud deposited in bays and tidewater reaches of rivers
QTac	Older coarse-grained alluvium; sand and gravel deposited by ancient rivers and streams
QTaf	Older fine-grained alluvium; sand and silt deposited by ancient rivers and streams
Grus	Decomposed granite
Tbs	Sedimentary bedrock
Tbv	Volcanic bedrock
KJg	Granite bedrock
KJm	Metamorphic bedrock

2. TABLE A-1, MEASURED SHEAR-WAVE VELOCITIES

[See MS-Excel file "Table A-1.xls" - Table A-1, Measured shear-wave velocities]

3. COLLECTION AND USE OF SHEAR-WAVE VELOCITY DATA 
This section describes our technique for collecting and applying the shear-wave velocity data shown in the preceding table (Table A-1). The table is also available on the accompanying CD-ROM disk as a Microsoft ExcelTM spreadsheet.
SH-wave data were collected by means of a 12-channel Bison 5000 seismograph with 8-bit instantaneous floating point and 2048 samples per channel. The data were recorded at a sampling rate between 0.025 and 0.5 ms, depending upon site conditions. The energy source for SH-wave generation is a 1.5 m section of steel I-beam struck horizontally by a 4.5-kg sledgehammer. The geophones used for recording SH-wave data were 30-Hz horizontal component Mark Product geophones. Spacing between the geophones is 3.05 m (10 ft). We used the walkaway method (Hunter and others, 1984), in which a group of 12 in-line geophones remained fixed and the energy source was "stepped out" through a set of predefined offsets. Depending upon site-geological conditions, the offsets of 3.05 m (10 ft), 30.5 m (100 ft), 61.0 m (200 ft), 91.5 m (300 ft), 122 m (400 ft), and 152.4 m (500 ft) were used. In order to enhance the SH-wave and reduce other phases, 5-20 hammer strikes on each site of the steel I-beam were stacked and recorded for each offset.
The SH-wave data were processed on a PC computer using the commercial software SIP by Rimrock Geophysics, Inc. (version 4.1, 1995). The key step for data processing is to identify the refractions from different horizons. Figure A-1 shows the composited SH-wave refraction profile generated from the individual offset records, at site McMin03 (Table A-1) near Dayton, Oregon. 

[See TIFF file "FigureA1.tif" - Figure A-1. Composited SH-wave refraction profile at site McMin03.]

Four refractions, R1, R2, R3, and R4 are identified in the profile. 
Arrival times of the refractions were picked interactively on the PC using the BSIPIK module in SIP. The arrival time data picked from each offset record were edited and combined in the SIPIN module to generate a data file for velocity-model deduction.
Figure A-2 shows the arrival times for the refractions identified in the profile (Figure A-1). The shear-wave velocity model is generated automatically using the SIPT2 module. 

[See TIFF file "FigureA2.tif" - Figure A-2. Arrival time curves of the refractions at site McMin03.]

Figure A-3 shows the shear-wave velocity model derived from the refraction data at site McMin03 (Figure A-1). The model is used to calculate an average shear-wave velocity.

[See TIFF file "FigureA3.tif" - Figure A-3. Shear-wave velocity model interpreted from refraction data at site McMin03.]

The average shear-wave velocity (Vs) over the upper 30 m of the soil profile is calculated with the formula of the Uniform Building Code (International Conference of Building Officials, 1997):
Vs = 30m/Sum{di/Vsi} 
Where: di = thickness of layer i in meters and
Vsi = shear-wave velocity of layer i in m/s. 

Based on the average shear-wave velocity and the UBC-97 soil profile categories as shown in Table 1 above (page 4), the UBC-97 soil classification map is generated with MapInfo(r) and Vertical Mapper(r). Soil types SE and SF can not be differentiated from the average shear-wave velocity. SE and SF are differentiated based on geologic and geotechnical data, and engineering judgement.

ß