SEARCH · USGS Science
Search indexed USGS publications on groundwater, aquifers, geologic maps, mineral resources and earthquakes. Explore source records by subject and place.
Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.
Glacial Lake Lind developed in the pre–late Wisconsinan St. Croix River valley, Minnesota and Wisconsin, and lasted more than 1000 yr during the retreat of the Superior lobe at the end of the Wisconsinan glaciation. Lake Lind sediment consists primarily of red varved silt and clay, but also includes mud-flow deposits, nearshore silt (penecontemporaneously deformed in places), nearshore rippled sand, and deltaic sand.
Tacaná volcano, located on the border between Mexico and Guatemala, marks the northern extent of the Central American volcanic chain. Composed of three volcanic structures, it is a volcanic complex that has had periodic explosive eruptions for at least the past 40 k.y. The most recent major eruption occurred at the San Antonio volcano, the youngest volcanic edifice forming the complex, about 1950 yr ago. The Peléan style eruption, issued from the southwest part of the dome, and swept a 30° sector with a hot block and ash flow that traveled about 14 km along the Cahoacán ravine. Deposits from this event are well exposed around the town of Mixcun and were therefore given the name of that town, the Mixcun flow deposit. The Mixcun flow deposit is, in the channel facies, a light gray, massive, thick (>10 m), matrix-supported unit with dispersed lithic clasts of gravel to boulder size, divisible in some sections into a variable number of flow units. The overbank facies is represented by a thin (<1 m), massive, matrix-supported unit. In both of these facies the deposit has disseminated charcoal, fumarolic pipes, and juvenile lithics with cooling joints. The Mixcun flow deposit contains clasts of (1) light gray, dense andesite, (2) dark gray, glassy and banded andesite, and (3) minor altered red andesite from the edifice, set in a matrix of sand and silt. The Mixcun flow deposit covers an area of at least 25 km 2 and has a minimum estimated volume of 0.12 km 3 . Basaltic-andesite inclusions (54% SiO 2 ) and various signs of disequilibrium in the mineral assemblage of the two-pyroxene andesitic products (60%–63% SiO 2 ) suggest that magma mixing may have triggered the eruption. Following deposition of the Mixcun flow deposit andesitic to dacitic (62%–64% SiO 2 ) lava flows were extruded and a dacitic dome (64.4% SiO 2 ) at the San Antonio summit formed. Syn-eruptive and posteruptive lahars flooded the main drainages of the Cahoacán and Izapa-Mixcun valleys in the area of the present city of Tapachula (population 250000) and the pre-Hispanic center of Izapa. Three radiocarbon ages date this event between A.D. 25 and 72 (range ±1σ, 38 B.C.–A.D. 216), which correlates with a halt in construction at Izapa (Hato phase of ca. 50 B.C.–A.D. 100), probably due to temporary abandonment of the city caused by lahars. Another similar event would produce extensive damage to the towns (population of about 68,000 people) now built upon the Mixcun flow deposit. The main summit of Tacaná volcano continues to show signs of fumarolic activity; the most recent period of activity in 1985–1986 culminated in a minor phreatic explosion.
The Cook Inlet basin is a northeast-trending forearc basin above the Aleutian subduction zone in southern Alaska. Folds in Cook Inlet are complex, discontinuous structures with variable shape and vergence that probably developed by right-transpressional deformation on oblique-slip faults extending downward into Mesozoic basement beneath the Tertiary basin. The most recent episode of deformation may have began as early as late Miocene time, but most of the deformation occurred after deposition of much of the Pliocene Sterling Formation. Deformation continued into Quaternary time, and many structures are probably still active. One structure, the Castle Mountain fault, has Holocene fault scarps, an adjacent anticline with flower structure, and historical seismicity. If other structures in Cook Inlet are active, blind faults coring fault-propagation folds may generate M w 6–7+ earthquakes. Dextral transpression of Cook Inlet appears to have been driven by coupling between the North American and Pacific plates along the Alaska-Aleutian subduction zone, and by lateral escape of the forearc to the southwest, due to collision and indentation of the Yakutat terrane 300 km to the east of the basin.
The Sea of Galilee, located in the northern part of the Dead Sea rift, is currently an intermediate fresh-water lake. It is postulated that during a short highstand phase of former Lake Lisan in the late Pleistocene, saline water percolated into the subsurface. Since its recession from the Kinarot basin and the instantaneous formation of the fresh-water lake (the Sea of Galilee), the previously intruded brine has been flushed backward toward the lake. Numerical simulations solving the coupled equations of fluid flow and of solute and heat transport are applied to examine the feasibility of this hypothesis. A sensitivity analysis shows that the major parameters controlling basin hydrodynamics are lake-water salinity, aquifer permeability, and aquifer anisotropy. Results show that a highstand period of 3000 yr in Lake Lisan was sufficient for saline water to percolate deep into the subsurface. Because of different aquifer permeabilities on both sides of the rift, brine percolated into a aquifers on the western margin, whereas percolation was negligible on the eastern side. In the simulation, after the occupation of the basin by the Sea of Galilee, the invading saline water was leached backward by a topography-driven flow. It is suggested that the percolating brine on the western side reacted with limestone at depth to form epigenetic dolomite at elevated temperatures. Therefore, groundwater discharging along the western shores of the Sea of Galilee has a higher calcium to magnesium ratio than groundwater on the eastern side.
In central British Columbia, ice flow during the late Wisconsinan Fraser glaciation (ca. 25–10 ka) occurred in three phases. The ice expansion phase occurred during an extended period when glaciers flowed westward to the Pacific Ocean and east-southeastward onto the Nechako Plateau from ice centers in the Skeena, Hazelton, Coast, and Omineca Mountains. Initially, glacier flow was confined by topography along major valleys, but eventually piedmont and montane glaciers coalesced to form an integrated glacier system, the Cordilleran ice sheet. In the maximum phase, a Cordilleran ice divide developed over the Nechako Plateau to 300 km inland from the Pacific coast. At this time, the surface of the ice sheet extended well above 2500 m above sea level, and flowed westward over the Skeena, Hazelton, and Coast Mountains onto the continental shelf, and eastward across the Rocky Mountains into Alberta. In the late glacial phase, a rapid rise of the equilibrium line caused ice lobes to stagnate in valleys, and restricted accumulation centers to high mountains. Discordant directions in ice flow are attributed to fluctuations of the ice divide representing changes in the location of accumulation centers and ice thickness. Ice centers probably shifted in response to climate, irregular growth in the ice sheet, rapid calving, ice streaming, and drainage of proglacial and subglacial water bodies. Crosscutting ice-flow indicators and preservation of early (valley parallel) flow features in areas exposed to later (cross-valley) glacier erosion indicate that the ice expansion phase was the most erosive and protracted event.
Significant sea-cliff erosion and storm damage occurred along the central coast of California during the 1982–1983 and 1997–1998 El Niño winters. This generated interest among scientists and land-use planners in how historic El Niño–Southern Oscillation (ENSO) winters have affected the coastal climate of central California. A relative ENSO intensity index based on oceanographic and meteorologic data defines the timing and magnitude of ENSO events over the past century. The index suggests that five higher intensity (relative values 4–6) and 17 lower intensity (relative values 1–3) ENSO events took place between 1910 and 1995. The ENSO intensity index correlates with fluctuations in the time series of cyclone activity, precipitation, detrended sea level, wave height, sea-surface temperature, and sea-level barometric pressure. Wave height, sea level, and precipitation, which are the primary external forcing parameters in sea-cliff erosion, increase nonlinearly with increasing relative ENSO event intensity. The number of storms that caused coastal erosion or storm damage and the historic occurrence of large-scale sea-cliff erosion along the central coast also increase nonlinearly with increasing relative event intensity. These correlations and the frequency distribution of relative ENSO event intensities indicate that moderate- to high-intensity ENSO events cause the most sea-cliff erosion and shoreline recession over the course of a century.
Elongate clasts in subglacial till and in fault gouge align during shearing, but the relation between clast-fabric strength and cumulative shear strain for such materials is effectively unknown. This relation was explored in experiments with a large ring-shear device in which a till and a viscous putty that contained isolated clasts were sheared to high strains. As expected, rotation of clasts in the putty is closely approximated by the theory of G.B. Jeffrey, who derived the orbits of rigid ellipsoids in a slowly shearing fluid. Clast rotation in the till, however, is strikingly different. Rather than orbiting through the shear plane as predicted by Jeffery, most clasts rotate into the shear plane and remain there, resulting in strong fabrics regardless of the aspect ratios and initial orientations of clasts. This divergent behavior is likely due to slip of the till matrix along the surfaces of clasts, which is a natural expectation in a granular material but violates the no-slip condition of Jeffery's model. These results do not support the widespread belief that subglacial till deformation results in weak clast fabrics. Thus, many tills with weak fabrics thought to have been sheared subglacially to high strains, like many basal tills of the Laurentide Ice Sheet, may have been sheared only slightly with little effect on either ice-sheet dynamics or sediment transport. In addition, these results indicate that in simple shear the rotation of clasts in till and in fault gouge is best analyzed with the model of A. March, who treated inclusions as passive markers.
The corrugated form of the Harcuvar, South Mountains, and Catalina metamorphic core complexes in Arizona reflects the shape of the middle Tertiary extensional detachment fault that projects over each complex. Corrugation axes are approximately parallel to the fault-displacement direction and to the footwall mylonitic lineation. The core complexes are locally incised by enigmatic, linear drainages that parallel corrugation axes and the inferred extension direction and are especially conspicuous on the crests of antiformal corrugations. These drainages have been attributed to erosional incision on a freshly denuded, planar, inclined fault ramp followed by folding that elevated and preserved some drainages on the crests of rising antiforms. According to this hypothesis, corrugations were produced by folding after subacrial exposure of detachment-fault foot-walls. An alternative hypothesis, proposed here, is as follows. In a setting where preexisting drainages cross an active normal fault, each fault-slip event will cut each drainage into two segments separated by a freshly denuded fault ramp. The upper and lower drainage segments will remain hydraulically linked after each fault-slip event if the drainage in the hanging-wall block is incised, even if the stream is on the flank of an antiformal corrugation and there is a large component of strike-slip fault movement. Maintenance of hydraulic linkage during sequential fault-slip events will guide the lengthening stream down the fault ramp as the ramp is uncovered, and stream incision will form a progressively lengthening, extension-parallel, linear drainage segment. This mechanism for linear drainage genesis is compatible with corrugations as original irregularities of the detachment fault, and does not require folding after early to middle Miocene footwall exhumations. This is desirable because many drainages are incised into nonmylonitic crystalline footwall rocks that were probably not folded under low-temperature, surface conditions. An alternative hypothesis, that drainages were localized by small fault grooves as footwalls were uncovered, is not supported by analysis of a down-plunge fault projection for the southern Rincon Mountains that shows a linear drainage aligned with the crest of a small antiformal groove on the detachment fault, but this process could have been effective elsewhere. Lineation-parallel drainages now plunge gently southwestward on the southwest ends of antiformal corrugations in the South and Buckskin Mountains, but these drainages must have originally plunged northeastward if they formed by either of the two alternative processes proposed here. Footwall exhumation and incision by northeast-flowing streams was apparently followed by core-complex arching and drainage reversal.
Major controls on the retention, distribution, and discharge of surface water in the historic (precanal) Kissimmee drainage basin and river were investigated to determine reference conditions for ecosystem restoration. Precanal Kissimmee drainage-basin hydrology was largely controlled by landforms derived from relict, coastal ridge, lagoon, and shallow-shelf features; widespread carbonate solution depressions; and a poorly developed fluvial drainage network. Prior to channelization for flood control, the Kissimmee River was a very low gradient, moderately meandering river that flowed from Lake Kissimmee to Lake Okeechobee through the lower drainage basin. We infer that during normal wet seasons, river discharge rapidly exceeded Lake Okeechobee outflow capacity, and excess surface water backed up into the low-gradient Kissimmee River. This backwater effect induced bankfull and peak discharge early in the flood cycle and transformed the flood plain into a shallow aquatic system with both lacustrine and riverine characteristics. The large volumes of surface water retained in the lakes and wetlands of the upper basin maintained overbank flow conditions for several months after peak discharge. Analysis indicates that most of the geomorphic work on the channel and flood plain occurred during the frequently recurring extended periods of overbank discharge and that discharge volume may have been significant in determining channel dimensions. Comparison of hydrogeomorphic relationships with other river systems identified links between geomorphology and hydrology of the precanal Kissimmee River. However, drainage-basin and hydraulic geometry models derived solely from general populations of river systems may produce spurious reference conditions for restoration design criteria.
Mostly stabilized late Holocene eolian sands on the Southern High Plains of the United States were studied to determine their origins and to assess whether present dune stability depends more strongly on sediment supply, sediment availability, or transport limitations. Geomorphic, sedimentological, and geochemical trends indicate that late Holocene dunes formed under westerly paleowinds, broadly similar to those of today. Mineralogical and geochemical data indicate that the most likely source for the sands is not the Pecos River valley, but the Pleistocene Blackwater Draw Formation, an older, extensive eolian deposit in the region. These observations suggest that new sand is supplied whenever vegetation cover is diminished to the extent that the Blackwater Draw Formation can be eroded, in agreement with modern observations of wind erosion in the region. We conclude, therefore, that Southern High Plains dunes are stabilized primarily due to a vegetation cover. The dunes are thus sediment-availability limited. This conclusion is consistent with the observation that, in the warmest, driest part of the region (where vegetation cover is minimal), dunes are currently active over a large area. Geochemical data indicate that Southern High Plains dunes are the most mineralogically mature (quartz rich) sands yet studied in the Great Plains, which suggests a long history of eolian activity, either in the dune fields or during deposition of the Blackwater Draw Formation.
Hundreds of circular features on Venus known as coronae are characterized by annular fractures and commonly associated radial fractures and lava flows. Coronae are thought to have been produced by buoyant mantle diapirs that flatten and spread at the base of the lithosphere and cause fracturing, uplift, and magmatism. The interior of Artemis Corona, by far the largest corona at 2100 km diameter, is divided in half by a northeast-trending deformation belt that contains numerous rounded ridges resembling antiforms. The largest of these ridges, located at the center of Artemis Corona, is ∼5 km high on its steep northwest flank where it is adjacent to a flat-bottomed, 10-km-wide trough interpreted as a rift valley. The 280-km-long antiformal ridge is marked by perpendicular grooves that cross the ∼50-km-wide ridge and extend southeastward as far as 120 km across adjacent plains. The grooves abruptly terminate northwestward at the rift trough. The large antiformal ridge terminates southwestward at a transform shear zone that parallels the grooves. These features—rift valley, antiformal uplift, grooves, and transform shear zone—are morphologically and geometrically similar to grooved, elevated, submarine metamorphic core complexes on the inside corners of ridge-transform intersections of slow-spreading ridges on Earth. As with submarine core complexes, the grooved surface on Venus is interpreted as the footwall of a large-displacement normal fault, and the grooves are inferred to be the product of plastic molding of the footwall to irregularities on the underside of the hanging wall followed by tectonic exhumation of the molded grooves and conveyer-belt–like transport up and over the large antiform and across the southeastern plains. According to this interpretation, the trend of the grooves records the direction of extension, which is perpendicular to the thrusts at the leading edge of the annular thrust belt 1000 km to the southeast. Both may have formed at the same time as a result of uniform southeastward displacement of the southeastern half of the interior of Artemis Corona. The location of this grooved core complex at the center of Artemis Corona may reflect genesis above the buoyant, ascending tail of the corona-producing mantle diapir.
Oriented samples of Jurassic radiolarian chert were collected from the Tuscan domain (continental margin) and the Ligurid domain (oceanic) of the northern Apennines for paleomagnetic study to determine the paleogeographic origins of these rocks. The oceanic rocks are all thermochemically overprinted by a mostly reversed-polarity component of magnetization (B) that was likely acquired during late Miocene regional uplift of the northern Apennines. This component also dominates the lower brittle chert of the Tuscan Cherts, but disappears upsection in the more clay-rich and ductile siliceous marlstones. In addition, the Tuscan Cherts retain an inferred primary magnetization (C), isolated at temperatures between 560 and 660 °C, which passes a fold test and shows a polarity stratigraphy. This component indicates a paleolatitude of 11° ± 4°N, and a counterclockwise vertical-axis rotation of 29° ± 9° with respect to the southern Alps of Italy, of 49° ± 8° with respect to Africa, and of 91° ± 8° with respect to Eurasia. Our results suggest that the Tuscan domain was farther south than other deep-water continental margin sections of Adria, and that transcurrent faulting might have played a significant role in the orogenic evolution of the northern Apennines.
Exposed channel banks along distributaries of the lower Snohomish delta in the Puget Lowland of Washington reveal evidence of at least three episodes of liquefaction, at least one event of abrupt subsidence, and at least one tsunami since ca. A.D. 800. The 45 measured stratigraphic sections consist mostly of 2–4 m of olive- gray, intertidal mud containing abundant marsh plant rhizomes. The most distinctive stratigraphic unit is a couplet comprising a 0.5−3-cm-thick, laminated, fining-upward, tsunami-laid sand bed overlain by 2−10 cm of gray clay. We correlated the couplet, which is generally ∼2 m below the modern marsh surface, across an ∼20 km 2 area. Sand dikes and sand-filled cracks to 1 m wide, which terminate upward at the couplet, and sand volcanoes preserved at the level of the sand bed record liquefaction at the same time as couplet deposition. Differences in the type and abundance of marsh plant rhizomes across the couplet horizon, as well as the gray clay layer, suggest that compaction during this liquefaction led to abrupt, local lowering of the marsh surface by as much as 50–75 cm. Radiocarbon ages show that the tsunami and liquefaction date from ca. A.D. 800 to 980, similar to the age of a large earthquake on the Seattle fault, 50 km to the south. We have found evidence for at least two, and possibly as many as five, other earthquakes in the measured sections. At two or more stratigraphic levels above the couplet, sand dikes locally feed sand volcanoes. Radiocarbon ages and stratigraphic position suggest that one set of these dikes formed ca. A.D. 910–990; radiocarbon ages on a younger set indicate a limiting maximum age of A.D. 1400–1640. We also interpret a sharp lithologic change, from olive-gray, rhizome-rich mud to grayer, rhizome-poor mud, ∼1 m above the couplet, to indicate a second abrupt lowering of the marsh surface during an earthquake ca. A.D. 1040– 1400, but no conclusive liquefaction structures have been identified at this horizon. Two distinctive coarse-sand laminae, 30–80 cm below the couplet, may record tsunamis older than A.D. 800. Thus, study shows that in the past ∼1200 yr, this part of Washington's Puget Lowland has been subjected to stronger ground shaking than in historic times, since ca. 1870.
The Sebree Trough is a relatively narrow, shale-filled sedimentary feature extending for several hundred kilometers across the Middle and Late Ordovician carbonate platform of the Midcontinent United States. The dark graptolitic shales within the trough stand in contrast to the coeval bryozoan-brachiopod-echinoderm– rich limestones on the flanking platforms. We infer from regional stratal patterns, thickness and facies trends, and temporal relations established by biostratigraphy and K-bentonite stratigraphy that the Sebree Trough initially began to develop during late Turinian to early Chatfieldian time (Mohawkian Series) as a linear bathymetric depression situated over the failed late Precambrian–Early Cambrian Reelfoot Rift. Rising sea level and positioning of a subtropical convergence zone along the southern margin of Laurentia caused the rift depression to descend into cool, oxygen-poor, phosphate-rich oceanic waters that entered the southern reaches of the rift from the Iapetus Ocean. The trough apparently formed in a system of epicontinental estuarine circulation marked by a density- stratified water column. Trough formation was accompanied by cessation of carbonate sedimentation, deposition of graptolitic shales, development of hardground omission surfaces, substrate erosion, and local phosphogenesis. The carbonate platforms on either side of the trough are dominated by bryozoan-brachiopod- echinoderm grainstones and packstones that were deposited in zones of mixing where cool, nutrient-rich waters encountered warmer shelf waters. Concurrently, lime mudstone and wackestone were deposited shoreward (northern Illinois, Wisconsin, Iowa, Minnesota, Michigan) in warmer, more tropical shallow seas. Coeval upward growth of the flanking carbonate platforms sustained and enhanced development of the trough shale facies. Five widespread diachronous late Mohawkian and Cincinnatian omission surfaces are present in the carbonate facies of the Midcontinent. These surfaces include sub-Deicke K-bentonite, DS1; top of Black River Limestone, DS2; base and top of the Guttenberg Limestone Member of the Decorah Formation, DS3 and DS4; and top of the Trenton Limestone, DS5. Some of the surfaces correspond to previously described depositional sequence boundaries. All five surfaces, which embody deepening phases on top of highstand-systems tracts, converge in the Sebree Trough, indicating that the trough was a long-lived feature and was the source of eutrophic waters that episodically spread across the adjacent platforms, terminating carbonate production. Late Turinian and early Chatfieldian incipient drowning episodes were followed by a final drowning event that began in the Sebree Trough during the late Chatfieldian ( Climacograptus spiniferus Zone) and reached southernmost Minnesota and other regions far within the platform interior by Richmondian time ( Amorphognathus ordovicicus Zone).
Detailed study of core samples of silicic tuff recovered from three geothermal wells along the strike-slip Great Sumatran fault zone near Silangkitang, North Sumatra, supports a model for enhanced hydrothermal circulation adjacent to this major plate-boundary fault. Two wells (A and C) were drilled nearly vertically ∼1 km southwest of the eastern (i.e., the principal) fault trace, and the third, directional well (B) was drilled eastward from the site of well A to within ∼100 m of the principal fault trace. The examined core samples come from depths of 1650–2120 m at measured well temperatures of 180–320 °C. The samples collected near the principal fault trace have the highest temperatures, the largest amount of secondary pore space that correlates with high secondary permeability, and the most extensive hydrothermal mineral development. Secondary permeability and the degree of hydrothermal alteration decrease toward the southwestern margin of the fault zone. These features indicate episodic, localized flow of hot, possibly CO 2 -rich fluids within the fault zone. The microstructure populations identified in the core samples correlate to the subsidiary fault patterns typical of strike-slip faults. The geothermal reservoir appears to be centered on the fault zone, with the principal fault strands and adjoining, highly fractured and hydrothermally altered rock serving as the main conduits for vertical fluid flow and advective heat transport from deeper magmatic sources.
U-Pb isotopic dating of detrital zircon from supracrustal Proterozoic and Cambrian arenites from the western United States and northern Mexico reveal three main age groups, 1.90 to 1.62 Ga, 1.45 to 1.40 Ga, and 1.2 to 1.0 Ga. Small amounts of zircons with ages of 3.1 to 2.5 Ga, 1.57 Ga, 1.32 Ga, 1.26 Ga, 0.7 Ga, and 0.5 Ga are also present. Detrital zircons ranging in age from 1.90 to 1.62 Ga and from 1.45 to 1.40 Ga are considered to have been derived from Proterozoic crystalline basement rocks of these known ages, and probably in part from reworked Proterozoic supracrustal sedimentary rocks, of the western United States. The 1.2 to 1.0 Ga detrital zircon ages from California, Arizona, and Sonora are characterized by distinct spikes (1.11 Ga, in particular) in the age-probability plots. These spikes are interpreted to indicate the influx of zircon from major silicic volcanic fields. Igneous rocks such as the Pikes Peak Granite (1.093 Ga) of Colorado, and the Aibo Granite (1.110 Ga) of Sonora, Mexico, may represent the deeply eroded roots of such volcanic fields. Samples from farther north along the Cordilleran margin that contain abundant 1.2–1.0 Ga detrital zircons do not show spikes in the age distribution, but rather ages spread out across the entire 1.2–1.0 Ga range. These age spectra resemble those for detrital zircons from the Grenville province, which is considered their source. Less common detrital zircons had a variety of sources. Zircons ranging in age from 3.36 to 2.31 Ga were apparently derived from inland parts of the North American continent from Wyoming to Canada. Zircons of about 1.577 Ga are highly unusual and may have had an exotic source; they may have come from Australia and been deposited in North America when Australia and North America were juxtaposed as part of the hypothetical Rodinian supercontinent. Detrital zircon of ∼1.320 Ga apparently had the same source as that for tuff (1.320 Ga) in the Pioneer Shale of the Apache Group in Arizona. Detrital zircons of about 1.26 Ga in the Apache Group and Troy Quartzite appear to be related to local, approximately coeval volcanic fields. Zircons of about 0.7 Ga may have had a source in igneous rocks related to rifting of the Proterozoic supercontinent of Rodinia, and 0.5 Ga zircons a source in relatively small areas of granitic rocks of this known, or inferred, age in Oklahoma, Texas, New Mexico, and Colorado.
A high-resolution aeromagnetic survey of the Puget Lowland shows details of the Seattle fault zone, an active but largely concealed east-trending zone of reverse faulting at the southern margin of the Seattle basin. Three elongate, east-trending magnetic anomalies are associated with north-dipping Tertiary strata exposed in the hanging wall; the magnetic anomalies indicate where these strata continue beneath glacial deposits. The northernmost anomaly, a narrow, elongate magnetic high, precisely correlates with magnetic Miocene volcanic conglomerate. The middle anomaly, a broad magnetic low, correlates with thick, nonmagnetic Eocene and Oligocene marine and fluvial strata. The southern anomaly, a broad, complex magnetic high, correlates with Eocene volcanic and sedimentary rocks. This tripartite package of anomalies is especially clear over Bainbridge Island west of Seattle and over the region east of Lake Washington. Although attenuated in the intervening region, the pattern can be correlated with the mapped strike of beds following a northwest-striking anticline beneath Seattle. The aeromagnetic and geologic data define three main strands of the Seattle fault zone identified in marine seismic-reflection profiles to be subparallel to mapped bedrock trends over a distance of >50 km. The locus of faulting coincides with a diffuse zone of shallow crustal seismicity and the region of uplift produced by the M 7 Seattle earthquake of A.D. 900-930.