Active Tectonics and Seismic Potential of Alaska E-Book
124,99 €
Mehr erfahren.
- Herausgeber: John Wiley & Sons
- Kategorie: Wissenschaft und neue Technologien
- Serie: Geophysical Monograph Series
- Sprache: Englisch
Published by the American Geophysical Union as part of the Geophysical Monograph Series, Volume 179.
This multidisciplinary monograph provides the first modern integrative summary focused on the most spectacular active tectonic systems in North America.
Encompassing seismology, tectonics, geology, and geodesy, it includes papers that summarize the state of knowledge, including background material for those unfamiliar with the region; address global hypotheses using data from Alaska; and test important global hypotheses using data from this region.
It is organized around four major themes:
- subduction and great earthquakes at the Aleutian Arc,
- the transition from strike slip to accretion and subduction of the Yakutat microplate,
- the Denali fault and related structures and their role in accommodating permanent deformation of the overriding plate, and
- regional integration and large-scale models and the use of data from Alaska to address important global questions and hypotheses.
The book's publication near the beginning of the National Science Foundation's EarthScope project makes it especially timely because Alaska is perhaps the least understood area within the EarthScope footprint, and interest in the region can be expected to rise with time as more EarthScope data become available.
Sie lesen das E-Book in den Legimi-Apps auf:
Seitenzahl: 1155
Veröffentlichungsjahr: 2013
Ähnliche
CONTENTS
PREFACE
Section I: Overviews
Active Deformation Processes in Alaska, Based on 15 Years of GPS Measurements
1. INTRODUCTION
2. DATA AND GPS DATA ANALYSIS
3. OBSERVED SITE VELOCITIES
4. SEISMIC CYCLE AND TECTONIC MODELS
5. ALASKA-SCALE DEFORMATION MODELS
6. BLOCK MODELS
7. CONCLUSIONS
Paleoseismicity and Neotectonics of The Aleutian Subduction Zone—An Overview
1. INTRODUCTION
2. NEOTECTONIC SETTING OF THE ALEUTIAN SUBDUCTION ZONE
3. SUBDUCTION ZONE SEGMENTATION AND GREAT EARTHQUAKES OF THE RECENT DISLOCATION CYCLE
4. PALEOSEISMOLOGY
5. CHRONOLOGY OF GREAT ALEUTIAN SUBDUCTION ZONE PALEOEARTHQUAKES
6. SUMMARY AND CONCLUSIONS
Neotectonics of the Yakutat Collision: Changes in Deformation Driven by Mass Redistribution
1. INTRODUCTION
2. REGIONAL TECTONICS
3. DEFORMATION FRONTS
4. DISCUSSION
5. CONCLUSIONS
An Overview of the Neotectonics of Interior Alaska: Far-Field Deformation From the Yakutat Microplate Collision
1. INTRODUCTION
2. THE DENALI FAULT AND THE 2002 EARTHQUAKE
3. RISE OF THE ALASKA RANGE
4. TOPOGRAPHY OF DENALI AND THE ALASKA RANGE
5. FLANKS OF THE ALASKA RANGE
6. EXHUMATION OF THE CHUGACH — KENAI MOUNTAINS
7. COOK INLET REGION
8. DISCUSSION—HOW IT WORKS
9. QUESTIONS FOR FUTURE STUDY
Active Tectonics of Interior Alaska: Seismicity, GPS Geodesy, and Local Geomorphology
1. INTRODUCTION
2. OVERVIEW OF SEISMICITY AND TECTONIC MODELS
3. ANALYSIS OF RECENT SEISMICITY
4. GEOMORPHIC DATA
5. GPS CONSTRAINTS ON DEFORMATION
6. DISCUSSION
7. CONCLUSIONS
Section II: The Alaska-Aleutian Subduction Zone
New Paleomagnetic Data From the Central Aleutian Arc: Evidence and Implications for Block Rotations
1. INTRODUCTION
2. PREVIOUS WORK
3. PALEOMAGNETIC STUDIES
4. CONCLUSIONS
Exhumation in the Chugach-Kenai Mountain Belt Above the Aleutian Subduction Zone, Southern Alaska
1. INTRODUCTION
2. TECTONICS OF THE CHUGACH—KENAI MOUNTAINS
3. TOPOGRAPHY OF THE CHUGACH—KENAI MOUNTAINS
4. THERMOCHRONOMETRIC APPROACH
5. RESULTS AND INTERPRETATIONS
6. DISCUSSION AND CONCLUSIONS
Active Faults on Northeastern Kodiak Island, Alaska
1. INTRODUCTION AND SEISMOTECTONIC SETTING OF THE KODIAK ISLANDS
2. PALEOSEISMICITY
3. SUMMARY
Paleoseismological Records of Multiple Great Earthquakes in Southcentral Alaska: A 4000—Year Record at Girdwood
1. CONTEXT AND AIMS
2. ANALYSIS OF GIRDWOOD MARSH
3. DISCUSSION
4. CONCLUSIONS
Seismicity of the Prince William Sound Region and Its Relation to Plate Structure and the 1964 Great Alaska Earthquake
1. INTRODUCTION
2. PREVIOUS STUDIES OF SEISMICITY
3. ANALYSIS TECHNIQUES
4. RESULTS
5. DISCUSSION
6. CONCLUSIONS
Section III: The Yakutat Collision Between the Subduction and Transform Boundaries
Geological and Geophysical Evaluation of the Mechanisms of the Great 1899 Yakutat Bay Earthquakes
1. INTRODUCTION
2. GEOLOGIC AND TECTONIC SETTING
3. 1899 DEFORMATION
4. A REVISED DEFORMATION MODEL
5. PALEOSEISMOLOGY
6. EARTHQUAKE SEISMOGRAMS AND SEISMIC MOMENT
7. CRUSTAL DEFORMATION AND FAULTING MODELS
8. EARTHQUAKE HAZARDS AND SEISMIC GAPS
9. SUMMARY AND CONCLUSIONS
Yakataga Fold-and-Thrust Belt: Structural Geometry and Tectonic Implications of a Small Continental Collision Zone
1. INTRODUCTION
2. GEOLOGIC AND TECTONIC SETTING
3. STRUCTURAL TRANSECT ACROSS THE CENTRAL YAKATA GA FOLD-AN D-THRUST BELT
4. AGE AND RATE OF DEFORMATION
5. TECTONIC IMPLICATIONS
5. SUMMARY AN D CONCLUSIONS
Identifying Active Structures in the Kayak Island and Pamplona Zones: Implications for Offshore Tectonics of the Yakutat Microplate, Gulf of Alaska
1. INTRODUCTION
2. DATA
3. OBSERVATIONS AND INTERPRETATIONS
4. DISCUSSION
5. CONCLUSIONS
Section IV: Deformation Inboard of the Plate Boundary
Neogene Exhumation of the Tordrillo Mountains, Alaska, and Correlations With Denali (Mount McKinley)
1. INTRODUCTION
2. GEOLOGIC SETTING
3. METHODS
4. FISSION TRACK THERMOCHRONOLOGY
5. AHe THERMOCHRONOLOGY
6. EXHUMATION
7. DISCUSSION
Does a Boundary of the Wrangell Block Extend Through Southern Cook Inlet and Shelikof Strait, Alaska?
1. INTRODUCTION
2. GEOLOGIC SETTING
3. SEISMIC REFLECTION DATA
4. DISCUSSION
5. CONCLUSION
Section V: Integrative Models, Stress Transfer, and Seismic Hazards
Tectonics, Dynamics, and Seismic Hazard in the Canada—Alaska Cordillera
1. INTRODUCTION
2. CURRENT TECTONICS
3. DYNAMICS OF THE CORDILLERA
4. IMPLICATI ONS FOR SEISMIC HAZARD MODELS
5. CONCLUSION
APPENDIX A: GRAVITATIONAL POTENTIAL ENERGY
Contemporary Fault Mechanics in Southern Alaska
1. INTRODUCTION
2. TECTONIC QUESTIONS
3. OBSERVATIONAL CONSTRAINTS
4. ANALYSIS APPROACH
5. RESULTS
6. DISCUSSION
7. CONCLUSIONS
Orogenesis From Subducting Thick Crust and Evidence From Alaska
1. INTRODUCTION
2. ALASKA TECTONIC SETTING
3. MECHANICAL MODEL FOR PLATE COUPLING
4. DISCUSSION
5. CONCLUSIONS
APPENDIX A: DIPPING VISCOUS SHEET
Stress Map for Alaska From Earthquake Focal Mechanisms
1. INTRODUCTION
2. PREVIOUS STUDIES
3. FOCAL MECHANISM DATA
4. STRESS TENSOR INVERSION METHOD
5. RESULTS AND DISCUSSION
6. CONCLUSIONS
Rapid Ice Mass Loss: Does It Have an Influence on Earthquake Occurrence in Southern Alaska?
1. INTRODUCTION
2. GLACIO-SEISM OTECT ONICS OF SOUTHER N ALASKA
3. STRESS CHANGES DUE TO ICE MASS FLUCTUATIONS AND CALCULATION OF FSM
4. SEISMICITY CATALOG
5. SIGNIFICANCE OF SEISMICITY RATE CHANGES
6. DISCUSSION AND SUMMARY
Challenges in Making a Seismic Hazard Map for Alaska and the Aleutians
1. INTRODUCTION
2. EARTH QUAKE HISTORY AND PLATE TECTONIC SETTING
3. METHODOLOGY
4. REGIONAL ANALYSIS OF SEISMIC SOURCES
5. DESCRIPTION OF MAP
6. PROBLEMS REQUIRING FUTURE WORK
7. CONCLUSIONS
Toward a Time-Dependent Probabilistic Seismic Hazard Analysis for Alaska
1. INTRODUCTION
2. OVERVIEW OF METHODOLOGY
3. COMPONENTS OF THE ANALYSIS
4. STRESS CHANGES
5. EARTHQUAKE PROBABILITIES
6. RESULTING EARTHQUAKE HAZARD
7. DISCUSSION
8. CONCLUSIONS
Fault Interaction in Alaska: Static Coulomb Stress Transfer
1. INTRODUCTION
2. STRESS MODELING
3. COULOMB STRESS CHANGE
4. SOURCES AND TARGETS
5. RESULTS
6. CUMULATIVE COSEISMIC COULOMB STRESS CHANGES BEFORE SEGMENT FAILURE
7. POSTRUPTURE STATIC STRESS TRANSFER: IMPLICATI ONS FOR SEISMIC HAZARD
8. CONCLUSIONS
Index
Download CD/DVD content
Geophysical Monograph Series
Published under the aegis of the AGU Books Board
Kenneth R. Minschwaner, Chair; Gray E. Bebout, Joseph E. Borovsky, Kenneth H. Brink, Ralf R. Haese, Robert B. Jackson, W. Berry Lyons, Thomas Nicholson, Andrew Nyblade, Nancy N. Rabalais, A. Surjalal Sharma, Darrell Strobel, Chunzai Wang, and Paul David Williams, members.
Library of Congress Cataloging-in-Publication Data
Active tectonics and seismic potential of Alaska / Jeffrey T. Freymueller ... [et al.], editors.
p. cm. — (Geophysical monograph, ISSN 0065-8448 ; 179)
Includes bibliographical references and index.
ISBN 978-0-87590-444-3
1. Geology, Structural—Alaska. 2. Earthquake zones—Alaska. 3. Geodynamics—Alaska. I. Freymueller, Jeffrey T.
QE627.5.A4A28 2008
551.809798—dc22
2008046658
ISBN: 978-0-87590-444-3
ISSN: 0065-8448
Cover Photo: An arial view of the Denali fault in the central Alaska Range, looking west toward Denali (Mt. McKinley). (Courtesy of Jeffrey T. Freymueller)
Copyright 2008 by the American Geophysical Union
2000 Florida Avenue, N.W.
Washington, DC 20009
Figures, tables and short excerpts may be reprinted in scientific books and journals if the source is properly cited.
Authorization to photocopy items for internal or personal use, or the internal or personal use of specific clients, is granted by the American Geophysical Union for libraries and other users registered with the Copyright Clearance Center (CCC) Transactional Reporting Service, provided that the base fee of $1.50 per copy plus $0.35 per page is paid directly to CCC, 222 Rosewood Dr., Danvers, MA 01923. 0065-8448/08/$01.50+0.35.
This consent does not extend to other kinds of copying, such as copying for creating new collective works or for resale. The reproduction of multiple copies and the use of full articles or the use of extracts, including figures and tables, for commercial purposes requires permission from the American Geophysical Union.
PREFACE
Alaska’s tectonic environment, one of the most dramatic and active on Earth, results in frequent great earthquakes, spectacular topography, crustal deformation, and high rates of volcanism. The Pacific–North America plate boundary system in Alaska and Canada is composed of four main types of faulting environments: a strike-slip boundary near the continental margin in the southeast (Fairweather, Queen Charlotte, and related faults), subduction of oceanic crust and block rotation in the west (Alaska-Aleutian megathrust), continental accretion and shortening in the north (Eastern Chugach–St. Elias region), and strike-slip and transpressional faulting well within the continent (Denali and other interior Alaska faults). These faults gave rise to the largest earthquakes in North America in the last hundred years, including the 1964 magnitude 9.2 Great Alaska earthquake and the 2002 magnitude 7.9 Denali fault earthquake. From the middle of the twentieth century to the present, virtually the entire Pacific–North America plate boundary from the Queen Charlotte Islands of northern British Columbia to the western end of the Aleutian Islands, has ruptured in about a dozen large to great earthquakes. The relatively small areas, segments only a few hundred kilometers long, along this plate boundary that have not ruptured during this time are of great interest. Some may be candidates for future gap-filling events, or if they are not, the differing character of these zones may hold clues to the fundamental nature of earthquake generation.
Perhaps the most visible consequence of the intensity and rate of such active tectonism are three of the four highest mountain ranges of North America: the Alaska Range, Chugach–St. Elias Mountains, and Wrangell Mountains. Subduction of the Pacific Plate has also led to formation of the Aleutian arc, one of the most active volcanic regions on the planet, and, in a less well-understood way, to formation of the voluminous volcanoes of the Wrangell Mountains. Active tectonics is not only a matter of what is happening at present, but also how active processes over time produce major geologic features. New techniques in thermochronology have helped reveal the geologically recent exhumation history of some of these mountain ranges. Traditional geological studies combined with the tools and perspectives of neotectonics, seismology, and geodesy, allow us to develop integrative models for the recent tectonic history and compare those to observations of present tectonism in order to understand how plate boundaries evolve. Studies in Alaska are aided in some cases by the very high rates of deformation that also cause the high levels of present seismicity.
Given this interesting tectonic setting and the wealth of new research and tools for analysis of the plate boundary in Alaska, we feel that the time is ripe for a synthesis of the current state-of-knowledge about the active tectonics and seismic potential of Alaska. The previous comprehensive summary of Alaska tectonics was that of the Decade of North American Geology (DNAG) volume, published in the early 1990s but written in the 1980s and based on previous work. Since then, space-based geodesy (GPS and InSAR) and new seismic instrumentation and analysis methods have provided additional information on the contemporary rates of deformation and the character of earthquakes. New techniques in Quaternary geology, paleoseismology, and thermochronology have also been developed and applied to Alaska problems, and regional geophysical studies have identified large-scale heterogeneities in physical properties and structure of the crust. The Denali fault earthquake of 2002 provided the impetus for considerable new work, and underscored the relevance and importance of ongoing studies, including many new insights into the role of thrust faults in initiating large strike-slip earthquakes and the process of mountain building in a plate boundary setting.
Subsequent renewed scientific interest and effort prompted by the Denali earthquake led to a 2006 AGU Chapman Conference titled “Active Tectonics and Seismic Hazards in Alaska,” which was attended by more than 60 participants affiliated with 31 research institutions from 5 countries. This monograph grew out of the presentations at that conference and summarizes both the state of the science and our understanding of the active tectonics and seismic potential of Alaska and the adjacent parts of Canada. Some areas of Alaska are poorly covered in this book because they are minimally understood or there has been little or no recent work. For example, the reader will find very little in this book about the tectonics of western Alaska, or about much of the northern tier of the state. These regions are tectonically active at a low level, and there is very little new data with which to evaluate the previously proposed hypotheses. We hope this lack of attention will be rectified in the near future so that various tectonic models can be clarified, seismic hazards characterized, and the nature of continental deformation illuminated. At this time, the NSF EarthScope program is finishing the construction of a new and much denser continuous GPS network across Alaska (part of the Plate Boundary Observatory), and in a few years EarthScope is scheduled to install seismic instrumentation of unprecedented density and regional scope (part of USArray). We hope the papers in this volume will spur new research in many disciplines of earth science that bear on the processes of plate boundary tectonics. The Chapman Conference and the papers in this volume have also served to inform the next revision of the seismic hazard maps issued by the U.S. Geological Survey. These maps are an important application of basic research results into societally relevant science products.
The papers in this book are organized into five groups. The first group consists of five invited overview papers, each summarizing a particular element of Alaska active tectonics studies: geodetic results and implications, paleoseismology of the Aleutian megathrust, the neotectonics of the Yakutat collision, and finally the active tectonics and seismotectonics of distributed deformation in the interior of the North American plate. The overview papers are intended to be understandable by all readers, whether they have a background in Alaska studies or not and while they are reviews, each also contains new results. The next four groups are topical papers, organized by four themes including the Alaska-Aleutian subduction zone, the Yakutat collision, and deformation inboard of the plate boundary. The final and largest group of papers combines regional models and syntheses with studies of stress transfer and seismic hazards. Not surprisingly, the boundaries between the groups can overlap, but the papers most closely related to each other in terms of tectonic problems rather than discipline, are placed near each other in the book.
Although this volume is focused on Alaska, it will be of interest to geophysicists and other earth scientists everywhere. Because Alaska contains ancient or modern examples of almost every tectonic environment on Earth (the notable exception being large-scale continental extension), the scientific questions highlighted here are broadly applicable to other regions. We also feel the approach to regional integration that we have taken in preparation of this book is also a good model for other regions. With the publication of this book and the promise of new Earth Scope data, we are now ready to encourage a broader array of scientists to apply the lessons learned from Alaska to other areas and vice versa.
Jeffrey T. FreymuellerUniversity of Alaska Fairbanks
Peter J. HaeusslerU.S. Geological Survey
Rob WessonU.S. Geological Survey (emeritus)
Göran EkströmColumbia University
Section I
Overviews
Active Deformation Processes in Alaska, Based on 15 Years of GPS Measurements
Jeffrey T. Freymueller,1 Hilary Woodard,2 Steven C. Cohen,3 Ryan Cross,4 Julie Elliott,1 Christopher F. Larsen,1 Sigrún Hreinsdóttir,5 and Chris Zweck6
1 Geophysical Institute, University of Alaska Fairbanks, Fairbanks, Alaska, USA.
2 Formerly at Geophysical Institute, University of Alaska Fairbanks, Fairbanks, Alaska, USA.
3 Formerly at NASA Goddard Space Flight Center, Greenbelt, Maryland, USA.
4 Now at Tetra Tech Inc., Seattle, Washington, USA.
5 Department of Geosciences, University of Arizona, Tucson, Arizona, USA.
6 Department of Hydrology and Water Resources, University of Arizona, Tucson, Arizona, USA.
We present a comprehensive average velocity field for Alaska, based on repeated GPS surveys covering the period 1992–2007, and review the major results of previously published papers that used subsets of this data. The spatially and temporally complex pattern of crustal deformation in Alaska results from the superposition of several processes, including postseismic deformation after the 1964 earthquake, spatial variations in plate coupling/slip deficit, translation and rotation of large crustal blocks or plates, and a large slow-slip event in Cook Inlet. Postseismic deformation from the 1964 earthquake continues today, mainly caused by viscoelastic relaxation, and causes trenchward motion. The behavior of the shallow seismogenic zone along the Alaska–Aleutian megathrust is characterized by dramatic along-strike variability. The width of the inferred seismogenic zone varies over along-strike distances that are short compared to the width. The along-strike distribution of locked and creeping regions along the megathrust is consistent with the persistent asperity hypothesis. A large slow-slip event occurred in upper Cook Inlet in 1998–2001, and a smaller event in the same area in 2005–2006. No sign of slow-slip events has been found in segments that are dominated by creep, which suggests that creep there occurs quasi-statically. The overriding plate in Alaska is subject to considerable internal deformation, and can be described in terms of the independent motions of at least four blocks: the Bering plate, the Southern Alaska block, the Yakutat block, and the Fairweather block.
1. INTRODUCTION
Figure 1. Location map showing major Quaternary active faults [Plafker et al., 1994], sites with GPS velocities, and the area of other figures. Faults shown with solid lines, small diamonds are sites with campaign surveys, and larger circles are continuous GPS sites. (a) All of Alaska, with outline area of other figures shown by dashed lines. Shaded regions show the rupture zones of great earthquakes, labeled by their year of occurrence. Outlined areas within the 1957 and 1965 earthquake rupture zones show the rupture areas of the M ~ 8 1986, 1996, and 2003 earthquakes. (b) Blowup view of region of densest data. The rupture areas of the 1964 Alaska earthquake and the 2002 Denali fault earthquake are outlined. YT, Yakutat terrane; PWS, Prince William Sound; KP, Kenai Peninsula; SGF, Susitna Glacier Fault.
The Chugach–St. Elias Range lies along the coast of southern Alaska in the gap between the subduction boundary to the west and the transform boundary to the east. These high and steep mountains lie at the northern boundary of the Yakutat terrane, an exotic terrane in the process of accreting to North America (Figure 1a). The Yakutat terrane moved north with the Pacific plate, and began to collide with North America at ~6 Ma [Plafker and Berg, 1994]. The Yakutat terrane consists either of a combination of continental and oceanic crust, or thickened oceanic crust [Pavlis et al., 2004]; in either case, it is buoyant and appears to resist subduction. The Yakutat terrane thrusts under the St. Elias Range (Figure 1b), and to the west it underthrusts Prince William Sound [Eberhart-Phillipset al., 2006; Plafker and Berg, 1994]. Subducted Yakutat crust may extend beneath North America as much as 500 km northwest from the trench [Eberhart-Phillips et al., 2006; Ferris et al., 2003]. A more complete description of the Yakutat terrane collision and subduction is given by Chapman et al. [this volume].
The southern Alaska margin features a wide zone of deformation inboard of the megathrust. The most intense inboard deformation occurs in a band that extends north and west of the St. Elias Range, where the impact of the Yakutat terrane has had the greatest effect. The right-lateral strike-slip Denali fault curves northwestward, well inboard of the St. Elias, and bounds a block of southern Alaska crust that rotates about a pole near Prince William Sound [Fletcher, 2002; Lahr and Plafker, 1980; St. Amand, 1957; Stout and Chase, 1980]. The central part of the Denali fault shows abundant Holocene offset features consistent with a slip rate of several mm/yr or faster over much of its length [Matmon et al., 2006; see also Haeussler, this volume]. The Denali fault system also includes a southeastern splay called the Totschunda fault, which may be part of an active connection between the Fairweather fault system in SE Alaska and the central Denali fault [Kalbas et al., this volume; Richter and Matson, 1971]. The 2002 Denali Fault Earthquake ruptured almost 300 km of the Denali and Totschunda faults, after initiating on the Susitna Glacier thrust [Crone et al., 2004; Eberhart-Phillips et al., 2003; Haeussler et al., 2004; Hreinsdóttir et al., 2006].
1.1. Terrestrial Geodetic Measurements
The USGS established several terrestrial geodetic networks in Alaska in the early 1980s. They repeated surveys of triangulation points [Lisowski et al., 1987; Savage et al., 1981], and carried out repeated electronic distance measurement (EDM) line-length surveys of networks crossing the Denali, Totschunda, and Fairweather faults, and networks to study the strain associated with subduction in the Yakataga and Shumagin segments (Figure 1b) [Lisowski et al., 1988; Savage and Lisowski, 1986, 1988, 1991].
The results of this work for the strike-slip faults were mixed. The EDM data from the Fairweather fault showed that the slip rate must be very high, although a precise rate could not be determined without independently constraining the locking depth [Lisowski et al., 1987]. A repeat survey of a triangulation network crossing the Totschunda fault also revealed right-lateral strain consistent with a slip rate of 10 ± 5 mm/yr [Lisowski et al., 1987]. Work on the Denali fault was less successful. During the 1980s, it was commonly thought that the Denali fault might have a slip rate almost comparable to the San Andreas fault, 10–20 mm/yr or more [Plafker et al., 1977]. However, two EDM networks across the fault showed nearly negligible shear strain, but significant fault-normal extension [Savage and Lisowski, 1991]. Still, given the error bars of the EDM measurements, the low right-lateral shear strain rate is not in conflict with later GPS measurements, because the EDM measurements were made very close to the fault. It is possible that some of the fault-normal extension was due to postseismic deformation after the 1964 earthquake (H. Suito and J. T. Freymueller, A viscoelastic and afterslip postseismic deformation model for the 1964 Alaska earthquake, submitted to J. Geophys. Res., 2008, hereinafter referred to as Suito and Freymueller, submitted manuscript, 2008), although geodetic monument instability or systematic errors are also possible explanations.
The network in the Yakataga segment (Figure 1b) showed rapid contractional strain with only small deviations from uniform uniaxial contraction. A simple subduction-type dislocation model could explain the observed deformation reasonably well, as long as the direction of plate convergence was taken to be N36°W instead of the N15°W direction of Pacific–North America relative motion [Savage and Lisowski, 1988]. Although this was not recognized at the time, later GPS measurements showed that the motion of the Yakutat terrane relative to North America was oriented very close to N36°W, and the strain in this region probably reflects Yakutat–North America motion [Fletcher and Freymueller, 1999, 2003].
The EDM measurements in the Shumagin Islands segment (Figure 1b) of the subduction zone showed only a small amount of strain in the direction of relative plate motion, much lower than anticipated from locked subduction zone models [Lisowski et al., 1988; Savage and Lisowski, 1986]. The Yakataga and Shumagin Islands segments were the only segments of the entire Alaska–Aleutian megathrust system that did not rupture in a great earthquake between 1938 and 1965, and there was a vigorous debate in the 1980s over whether the Shumagins segment represented a seismic gap that was overdue for rupture, or a segment with low strain and unlikely to rupture in great earthquakes [Boyd et al., 1988]. Vertical measurements in the Shumagins seemed to support the presence of subduction-related strain in that segment [Beavan et al., 1984], although later and more precise measurements showed smaller or negligible vertical motions [Beavan et al., 1986; Hurst and Beavan, 1987]. A later combination of EDM and GPS data showed that a small amount of horizontal contraction did occur across the Shumagin Islands, but only about 25% of what was expected from simple subduction zone models [Larson and Lisowski 1994; Zheng et al., 1996].
1.2. Mobile Very Long Baseline Interferometry Measurements
A permanent very long baseline interferometry (VLBI) station was established in Fairbanks in the early 1980s, which operated until the end of 2005. In addition, mobile VLBI measurements were made at five sites in Alaska and one in Canada from 1984 to 1989, in order to study the Pacific–North America plate boundary. The most comprehensive summary of the velocities was presented by Ma et al. [1990]. The mobile VLBI data were also used to estimate displacements from the 1987–1988 Gulf of Alaska earthquake sequence, a series of earthquakes that occurred on a north–south trending feature on the Pacific plate south of the megathrust [Argus and Lyzenga, 1994; Sauber et al., 1993].
The mobile VLBI data suffered from two significant drawbacks. The first was that most of the Alaska mobile VLBI campaigns contained very little data from sites outside of Alaska. As a result, the rotation of the network around the permanent Fairbanks point was poorly constrained, and as a result, the velocities of the sites differed considerably in different VLBI solutions (by as much as 3–4 mm/yr). However, the Ma et al. [1990] velocities compare quite well overall with the GPS velocity field, and orientation errors in that solution appear to be minimal. The main problem with the VLBI data was the small number of sites, which made interpretation of their velocities difficult. Ma et al. [1990] showed that the velocities of Kodiak and Cape Yakataga were generally consistent with the predictions of subduction zone models with a wide locked zone, whereas the velocity of Sand Point was intermediate between that expected for aseismic subduction and a locked subduction zone model. Sauber et al. [1993] compared the VLBI data from Cape Yakataga with the EDM line-length network, and found that these two data sets were compatible. The velocities for the interior sites were not easily explained. The mobile VLBI site at Whitehorse may have been unstable, because the VLBI velocity for this site is several mm/yr different from the velocity of the later GPS site WHIT.
1.3. GPS Measurements in Alaska
The first GPS measurements in Alaska date back to 1984, made by the National Geodetic Survey (NGS). Because the terrestrial geodetic network in Alaska was sparse and outdated, NGS carried out a series of statewide measurement campaigns from 1988 to 1991, aimed at providing a statewide High Accuracy Reference Network. Unfortunately, most of those campaigns came too early for the data to be useful for high-precision deformation measurements. Since then, NGS continued to carry out occasional surveys at airports and tide gauges, including some repeat measurements. The USGS carried out high-precision measurements in 1991 in the Shumagin Islands, and in 1992 in the ice fields near Yakutat, as final surveys of the old EDM networks. Much more extensive campaign GPS measurements were begun in 1993 by the USGS, Goddard Space Flight Center, and several universities, with three NASA-supported projects to study subduction-related processes in southern Alaska. Another important data set comes from the work of Avé Lallemant and Oldow [2000] in the western Aleutians, where data collection began in 1996.
We began to collect GPS data in many locations across Alaska in 1995, in several cases building on previous work done by other groups [e.g., Cohen et al., 1995; Sauber et al., 1997; Savage et al., 1998]. We carried out numerous GPS campaign surveys throughout Alaska, which provided the bulk of the data used here. Although the continuous GPS site FAIR has operated since late 1991, no additional continuous sites were established until 1996, when the U.S. Coast Guard set up seven sites to support real-time navigation. Another 10 continuous sites were established from 1997 to 2002, and 16 more sites were set up shortly after the 2002 Denali Fault Earthquake to measure postseismic deformation. For the most part, GPS work in Alaska has been carried out through repeated campaign surveys. Today there are many more continuous sites, as the Plate Boundary Observatory has set up sites across all of Alaska, but these data were not used in the present compilation. This new network of continuous sites will usher in a new generation of GPS studies.
This paper summarizes research into Alaska active tectonics carried out over a roughly a decade and a half, using GPS geodesy. These observations (Table 1) provide a wealth of information about the active tectonic processes affecting the region. The earliest data used in this paper were collected in 1992, because the earlier data are difficult to place into the same reference frame as the modern data due to the weak global tracking network at the time. This paper presents and documents the contemporary deformation field, as determined from GPS site velocities at 78 sites using data spanning 1992–2002, and from 497 additional sites spanning 1992–2007 (575 total sites). Only sites located far enough from the rupture zone of the 2002 Denali Fault Earthquake [Eberhart-Phillips et al., 2003] to have minimal effects from postseismic deformation [e.g., Freedet al., 2006a, 2006b] are included in the second set, which is based on the solution used by Cross and Freymueller [2008]. Spatial and temporal subsets of this velocity field have been published previously [Cohen and Freymueller, 1997, 2004; Cross and Freymueller, 2007, 2008; Fletcher, 2002; Fletcher and Freymueller, 1999, 2003; Fletcheret al., 2001; Fournier and Freymueller, 2007; Freymueller and Beavan, 1999; Freymueller et al., 2000; Larsenet al., 2004, 2005; Mann and Freymueller, 2003; Ohtaet al., 2006; Sauberet al., 1997, 2006; Savageet al., 1998, 1999; Zweck et al., 2002], but the version presented here is the most complete and comprehensive, spanning the entire region and includes all usable data. In addition to data we collected, we incorporate data from GPS surveys collected by other academic groups, state and federal agencies, and by private land surveyors (Table 1). Surveys conducted for land surveying purposes often use much shorter observation times than those conducted for measurements of crustal deformation, so we normally used only surveys that lasted a minimum of ~5 h per day. Despite the lower precision, these surveys provide critical data in some areas. Compared to previous published results, we improved the individual solution quality by a systematic reanalysis of all data using consistent software and models, used a better and more accurate definition of the global and North America-fixed reference frames, and carefully rechecked field notes and logs to correct errors in antenna heights and antenna types. The result is the most precise and accurate three-dimensional (3-D) velocity field available for Alaska to date.
Table 1. Sources of Data Used in This Study
Table 2. Summary of Solutions Used in This Study
2. DATA AND GPS DATA ANALYSIS
We have compiled the most complete GPS velocity field possible for both the pre-Denali earthquake period and the full study period. All data have been analyzed in a consistent manner using the same software and models, and combined into a single velocity solution to estimate site velocities. We divide the data analysis steps into three stages—GPS data analysis, reference frame, and velocity estimation—and describe each in order.
2.1. GPS Data Analysis
We analyzed all data presented in this paper using the GIPSY/OASIS GOA4 software developed at the Jet Propulsion Laboratory (JPL) [Gregorius, 1996; Zumberge et al., 1997]. Beginning 1 January 1996, we analyzed and used every day of data from permanent GPS stations in and around Alaska and included all campaign stations. We did not analyze every day of data prior to 1996, but included all days with some campaign data. We then combined all of these daily GPS solutions into a single velocity solution to estimate linear velocities. In total, 2801 individual daily GPS solutions were used in the pre-earthquake velocity solution, and 4623 in the full solution (Table 2).
For data collected before 1995, we combined the data from Alaska with data from the global (IGS) network and estimated satellite orbits. In these solutions, we fit a satellite trajectory from existing orbit estimates (broadcast orbits, or a precise orbit), and integrated the equations of motion to generate an a priori orbit and associated partial derivatives for each satellite. We estimated the initial conditions (position and velocity) for each day’s orbit, plus time-varying solar radiation pressure, along with station positions and other parameters. All site coordinates are estimated with loose constraints in our global solutions. For 1992 and 1993, we used every available GPS site around the world, no matter where it was located. By 1994, dense arrays of continuous GPS sites began to be deployed in Southern California and Japan, and the density of continuous sites in western Europe became higher than needed to determine orbit parameters, so for data from 1994 we use a well-distributed set of ~60 global stations in each daily solution. The exact stations used varied day-to-day based on data availability.
For data beginning in 1995, we combined data from Alaska with data from continuous GPS sites spanning North America, the Arctic regions of Eurasia, and a few additional sites in the central Pacific. The exact stations used varied day-to-day based on data availability. For these solutions, we fixed JPL’s fiducial-free orbit (estimated without significant a priori site position constraints). This orbit is generated based on data from a global network of 50–85 stations, and because no station positions are constrained it is in no particular reference frame. However, the orbits are internally self-consistent, and the resulting solutions can be transformed into any reference frame by use of a seven-parameter similarity transformation.
We used both phase and pseudorange data from all sites (some receivers provided only phase data), and analysis models similar to those described by Larson et al. [1997] and Freymueller et al. [1999], although many models internal to the software have been updated since that time. We use network solutions rather than point positioning, and we normally used the site ALGO (Algonquin Park, Ontario, Canada) as a reference clock. When ALGO was unavailable or had clock or data problems, we used the site AMC2 (Colorado Springs, CO) as a reference clock. We applied elevation-dependent antenna phase center corrections based on the IGS_01 model, and applied the NOAA antenna calibrations (http://www.ngs.noaa.gov/ANTCAL/) for antennas not included in the IGS_01 model. We used all data from satellites 10° or more above the horizon, and estimated zenith wet tropospheric path delays using the Niell mapping function [Niell, 1996]. The a priori tropospheric path delay was set to an elevation-dependent value for the dry (hydrostatic) component, and 10 cm for the wet component, and we estimated a residual wet delay with azimuthal gradient. Finally, we applied an ocean tidal loading model calculated for each site based on the TPXO.2 ocean tidal model, using the SPOTL software [Agnew, 1997]. Ocean tidal loading amplitudes are very large for sites in southern Alaska, often in excess of 40 mm, due to the very large tidal range in the Gulf of Alaska. The even larger tidal range in Cook Inlet is not represented in any existing tidal model, and we see residual periodic errors (a strong fortnightly period) in our solutions consistent with aliasing of unmodeled ocean loading displacements [Penna et al., 2007], especially for sites in that region.
2.2. Daily Reference Frame
We transformed each day’s loosely constrained solution into the International Terrestrial Reference Frame 2000 (ITRF2000) reference frame, using the IGSb00 realization (reference URL given above), using ~20 reference frame sites to define the transformation. For each day, we compute the predicted position of each reference frame site based on the IGSb00 realization of ITRF2000 (using SINEX file ftp://igscb.jpl.nasa.gov/igscb/station/coord/IGS03P33_RS106. SNX), and then compute the seven-parameter similarity transformation between our solution and the ITRF2000 prediction that minimizes the residuals at all reference frame sites. In computing the transformation, all sites present in both the solution and ITRF2000 are weighted based on the joint uncertainties in the ITRF and our solution. The typical 3-D weighted root-mean-square (WRMS) misfit after transformation is 4–6 mm for recent solutions, and the posttransformation 3-D WRMS residuals are usually ~15 mm for solutions in 1993 and ~20 mm for 1992. The WRMS values give a general idea of the precision with which the reference frame is defined on a daily basis.
We did not add uncertainty to the daily positions to account for uncertainties in the ITRF, because random daily errors in the realization of ITRF are most likely small compared to other uncertainties. However, there are significant uncertainties in the definition of the terrestrial reference frame that we account for by adding uncertainty to our estimated velocities. The most significant uncertainty in the definition of the ITRF is probably in the definition of the geocenter. It is well known that ITRF2000 and ITRF2005 have a geocenter Z rate difference of 1.8 mm/yr, and there are differences of similar magnitude between earlier versions of ITRF [Argus, 2007]. These differences make absolute plate rotations computed in ITRF specific to that version of ITRF, and for sites in Alaska they also affect the vertical velocities because a change in the definition of the geocenter affects the Z component of all site velocities, which maps more into the vertical than the horizontal for sites at high latitude.
2.3. Velocity Estimation and North American Frame
We used all daily GPS solutions in a linear network velocity estimation, to estimate positions at an epoch time (2000.0) plus velocities for the GPS sites (Figure 1). We present two solutions here, a pre-earthquake solution (1992–2002) and a 1992–2007 solution for sites far from the 2002 Denali fault earthquake, and we use the pre-earthquake solution only for sites not in the longer solution. The 1992–2007 solution is based on that used by Cross and Freymueller [2008]. In a few cases, colocated sites were assumed to have the same velocity, and only one velocity is presented here. Less than 10% of the sites in the solution are continuous GPS sites. We estimated only linear velocities with time, without periodic seasonal terms, since most sites are not continuous sites. The continuous GPS sites all have at least 5.5 years of data, so the velocities are unlikely to be biased by neglecting the seasonal terms, and all campaign surveys were done at nearly the same time of year (summer). All daily solutions were weighted according to the inverses of their covariance matrices resulting from the GIPSY analysis. We scaled the covariances of all solutions by a constant factor, 6.1 for the pre-earthquake solution and 8.9 for the 1992–2007 solution, based on the misfit to the velocity solution so that the χ2 per degree of freedom of the velocity solution was equal to 1.0. This results in an increase in the uncertainties of all observations by a factor of 2.5–3. We would most likely have a slightly smaller scaling factor for the uncertainties if seasonal terms were estimated, because the continuous sites show significant seasonal variations (mostly in height) [Freymueller, in press]. However, almost all campaign surveys were carried out in the same 3- to 4-month period each year, so seasonal variations should have little effect on their site velocities. Master tables of all site velocities are given in Tables ES1 and ES2 (found on the CDROM accompanying this volume). All figures are based on this master site velocity table.
Velocities were then referred to the North American plate by subtracting the rotation of North America in ITRF from the ITRF2000 velocities. We use the recent determination of the motion of North America of Sella et al. [2007], which also used the IGSb00 realization of ITRF2000 and is thus the most consistent with our velocity reference frame. Using an older definition of North America, such as that of the REVEL 2000 model [Sella et al., 2002], would change the horizontal site velocities relative to North America by 2–3 mm/yr. Experiments with different reference frame realizations show that this difference is almost entirely due to differences between ITRF97 (used for REVEL) and ITRF2000/IGSb00 (used by Sella et al. [2007]). The most significant difference between the frames is in the geocenter rate. The recent SNARF 1.0 realization of a North America-fixed reference frame (http://www.unavco.org/research_science/workinggroups_projects/snarf/snarf.html) predicts motions that differ by about 1 mm/yr from those of Sella et al. [2007]. This difference between the two recent studies results from the use of a different site distribution to define stable North America and a different strategy to account for Glacial Isostatic Adjustment across northern North America. Differences between predicted velocities are much smaller in the lower 48 states of the United States, but are amplified in Alaska by the large distance to the stable part of North America. This difference may be indicative of the level of uncertainty in the definition of the North American frame.
The uncertainties in ITRF and the North American frame discussed above are much larger than the random errors in site velocities for many sites, including most of the campaign sites with long measurement histories. We thus augment the covariance matrix of the velocities to account for both elements of the reference frame uncertainty. The first component of uncertainty is in the Z component, where we add 1.8 mm/yr uncertainty to all sites based on the difference between ITRF2000 and ITRF2005. This uncertainty is perfectly correlated from site to site, so that it affects only absolute velocities and not relative velocities. At 60°N, this maps into an additional uncertainty of 1.6 mm/yr in the local vertical, and 0.9 mm/yr in the north component, added in quadrature to the uncertainty from random error. We also add an uncertainty in the North American plate rotation equal to the difference between the angular velocities of Sella et al. [2007] and SNARF 1.0. The difference between these two is equivalent to a rotation about a pole located near Mobile, AL. For sites in Alaska, this amounts to an additional uncertainty of ~1 mm/yr, with an elongate error ellipse oriented NW–SE. Because we propagate these uncertainties using the full covariance matrix, this uncertainty is also highly correlated for nearby sites and has a limited effect on relative site velocities.
3. OBSERVED SITE VELOCITIES
The estimated site velocities (Figures 2–12) show substantial spatial variation, resulting from a variety of active tectonic and volcanic processes. These include elastic strain resulting from the slip deficit of the locked subduction zone, which varies along strike, postseismic deformation following the 1964 earthquake, glacial isostatic adjustment from the post-Little Ice Age (LIA) ice melting, the relative motion of large tectonic blocks making up part of the crust of Alaska, and inflation of active volcanoes.
Each of the subsections below discusses the observed velocities for a region of Alaska. We highlight the most important features of the data, based on the published papers that used it. More extensive discussion of models to explain the data in some of these regions will be presented in a later section; in this work we present the observations and summarize the main contributions to the velocities. Because of the large variations in the density of sites and magnitude of velocities, map scale and velocity scales differ for each figure. As the discussion for each of these regions necessarily uses local place names, we encourage the reader to refer to the regional maps.
3.1. Southeast Alaska
Southeast Alaska undergoes extraordinary uplift due to glacial isostatic adjustment from the rapid melting of glaciers and ice fields that followed the LIA. The LIA glaciation in southern Alaska reached its peak ~1900 A.D., although deglaciation in some regions began up to a century earlier [Calkin et al., 2001; Motyka, 2003; Wiles et al., 1999]. Peak uplift rates in the region exceed 30 mm/yr, and most sites along the coast and coastal mountain belt uplift at rates exceeding 10 mm/yr (detailed later in section 3.7). Larsen et al. [2004, 2005] showed that this rapid uplift can be explained using the known ice load history with an elastic crust about 60 km thick overlying a low-viscosity asthenosphere. Horizontal motions from glacial isostatic adjustment cause radial extension centered on Glacier Bay, and may cause horizontal motions larger than 5 mm/yr in some cases (Elliott et al., in preparation), although this effect is not considered in this paper.
The Fairweather fault is the most active part of the Pacific–North America plate boundary in Southeast Alaska (Figure 2; Plafker et al. [1978]). The Fairweather fault is the onshore extension of the offshore Queen Charlotte transform fault to the south, and it terminates in a complex transition with the Chugach–St. Elias fault in the north. Pacific–North America plate motion is roughly parallel to the Queen Charlotte fault, but the strike of the Fairweather fault (N34°W) is rotated clockwise by 23° relative to the Pacific–North America convergence direction (N11°W); this requires significant deformation to occur on other faults because the motion on the Fairweather fault is strike slip. A splay of the Fairweather fault may extend further to the northwest, to connect with the Totschunda and Denali faults, as originally proposed by Richter and Matson [1971], although no such fault has yet been mapped. The Denali fault lies ~100 km inboard of the Fairweather fault in this region, and Quaternary offsets of the Denali fault have been documented at several locations [Matmon et al., 2006; Plafker et al., 1994], although evidence for activity along its southern extension, the Chatham Strait fault, is lacking.
Right-lateral shear on the Fairweather fault dominates the GPS velocity field for this region. The site in Yakutat (YKTT) on the Pacific coast moves 45 mm/yr toward N30°W relative to Whitehorse (WHIT), located ~300 km inland. This relative motion vector is within 4° of being parallel to the strike of the Fairweather fault (N34°W). Velocities of all other sites along the Pacific coast are also very close to parallel to the Fairweather fault, indicating that right-lateral strike-slip motion parallel to the Fairweather fault is the dominant tectonic motion. Fletcher and Freymueller [1999, 2003] interpreted this direction of motion to indicate that the Yakutat terrane moves (relative to North America) parallel to the Fairweather fault, rather than in the direction of the Pacific plate as had been suggested earlier [e.g., Lundgren et al., 1995].
Lisowski et al. [1987] estimated the slip rate of the Fair-weather fault to be 41–51 mm/yr based on repeated line-length measurements of a dense network across the fault near Yakutat. The wide range of possible slip rates resulted from a strong tradeoff between the fault slip rate and locking depth; any locking depth ≥4 km could fit the data equally well, with lower slip rates associated with shallower locking depths. This is a defect of the network, because all sites were located within 10 km the fault. Several sites from that network have been surveyed repeatedly using GPS, and Fletcher and Freymueller [2003] modeled these GPS velocities and EDM line length changes and estimated the best-fitting slip rates on the Fairweather and eastern Denali faults to be 45.6 ± 2.0 and 3.8 ± 1.4 mm/yr, respectively, with a locking depth of 9.0 ± 0.8 km for the Fairweather fault.
Figure 2. Horizontal velocities from southeast Alaska. The sites Whitehorse (WHIT) and Yakutat (YKTT) are labeled. Major faults are shown in black and marked with letters inside boxes: D, eastern Denali Fault; DR, Duke River Fault; C, Connector Fault; CSE, Chugach–St. Elias Fault; F, Fairweather Fault (almost entirely covered by green arrows); TZ, Transition Zone; QC, Queen Charlotte Fault; LI, Lisianski Inlet–Peril Strait Fault; CS, Chatham Strait Fault.
Similar Fairweather-parallel motion persists to the west of this region, for coastal sites on the south side of the St. Elias Range, almost as far west as Prince William Sound (Figure 3). However, the coastal sites on the south side of the St. Elias Range move at a slower rate than the sites from Yakutat to the south, which may indicate that active convergent structures in the St. Elias Range extend nearly as far southeast as Yakutat Bay. This is consistent with the EDM data from the 1980s, which showed the strain tensor in the region to be dominated by contraction in the N32°W ± 2.4° direction [Savage and Lisowski, 1988], which is parallel to the Fairweather fault and the GPS velocities. Sauber et al. [1997] modeled one component (Pacific-parallel direction) of the velocities across the St. Elias Range using a subduction model with deformation caused by the subducting Pacific plate. Although their model fit the one component modeled well, it failed to explain the orientation of the observed velocity vectors. However, a very similar model in which the subducting plate is the Yakutat terrane, moving in a Fairweather-parallel direction, provides a good first-order fit to the data [Elliott et al., 2006].
Figure 3. Horizontal velocities from southeast Alaska and the St. Elias region. Major faults are shown in black and marked with letters inside boxes: D, Denali Fault; T, Totschunda Fault; C, Connector Fault; CSE, Chugach–St. Elias Fault; TZ, Transition Zone. Inset shows the Pacific–North America and Yakutat–North America relative motion directions. The site YKTT is labeled.
Inland sites show an east to northeast-directed motion relative to North America (Figures 2 and 3). Much of this can be explained by glacial isostatic adjustment (Elliott et al., in preparation), but the remainder may be caused either by an error in the definition of stable North America or by NE-directed motion of the northern Canadian Cordillera, as suggested by Mazzotti and Hyndman [2002]. The increasing velocity to the north observed for the inland sites is consistent with the latter explanation.
3.2. South-Central Alaska
Site velocities across South-central Alaska are spatially complex, showing the influence of several significant contributions to the deformation field (Figure 4). From east to west, velocities of coastal sites rotate systematically from a Fairweather-parallel orientation in the St. Elias Range east of Prince William Sound (see also Figure 3) to a Pacific-parallel direction in the eastern Kenai Peninsula (Figure 4). This rotation of velocities must result from a transition in the subducting plate from the Yakutat terrane in the east to the Pacific plate in the west. Beneath Prince William Sound, the North American plate, Yakutat terrane, and Pacific plate form a three-plate “sandwich,” with the Yakutat terrane in the middle. The Pacific plate underthrusts the Yakutat terrane at the eastern end of the Alaska–Aleutian trench, and the Yakutat terrane underthrusts North America [Brocher et al., 1994; Fuis et al., 2008]. This three-plate system makes a transition westward to normal subduction of the Pacific plate as the Yakutat terrane pinches out. The southwestern limit of the Yakutat terrane in the subsurface has generally been inferred to be the Slope Magnetic Anomaly, which extends from the Transition Zone to the SW edge of Montague Island [Griscom and Sauer, 1990]. Von Huene et al. [1999] proposed that the edge of the Yakutat terrane in the subsurface and the edge of the Prince William Sound asperity of the 1964 Alaska earthquake coincide. Brocher et al. [1994] suggested that the Prince William Sound asperity lies between the Yakutat terrane and North America, not between the Pacific slab and North America. This question will be revisited in section 4.4.
Figure 4. Horizontal velocities from Prince William Sound, the Kenai Peninsula, and Upper Cook Inlet. The shaded area is the rupture zone of the 1964 earthquake, and the star is its epicenter. Important geographic features are labeled: KP, Kenai Peninsula; PWS, Prince William Sound, KI, Kayak Island. Sites referred to in the text are labeled in italics. The inset shows the Pacific–North America and Yakutat–North America relative motion directions. The light dotted line offshore south of Montague Island shows the location of the Slope Magnetic Anomaly (SMA), considered to represent the southwest edge of the subducted Yakutat terrane crust. TZ, transition zone.
Across the Kenai Peninsula, the magnitude of coastal site velocities drops dramatically with distance west from Prince William Sound, from 56 mm/yr at Montague Island (MOTG) to 35 mm/yr at Seward (UAMF), to 10 mm/yr at Nuka Bay (2201). Further west, the velocities reverse orientation and become trench-normal. Inland, a similar pattern is found, although sites with trenchward velocities are found throughout the Cook Inlet region. The regions of landward (Pacific-parallel) and trenchward velocities roughly correspond to the mountainous and lowland regions of the Kenai Peninsula, but at the SW end of the Peninsula even sites in the Kenai Mountains move toward the trench. The average orientations of the two sets of velocities are nearly opposite to each other. Between the two groups lie several sites with near-zero velocities, and most of these site velocities are deflected to the west.
Savage et al. [1998] presented the first quantitative deformation model for this region, based on relative velocities within a profile at the western edge of Prince William Sound, observed from 1993 to 1997. This profile did not show the trenchward motion, and they explained the velocities using a 2-D (uniform in the along-strike direction) dislocation model with the plate interface dipping 3° to the NNW. Although they obtained a reasonable fit to the data, the deformation across their network could not be matched unless they assumed the plate convergence rate to be ~20% faster than the long-term plate convergence rate. They inferred that the additional strain resulted from continuing postseismic relaxation.
Freymueller et al. [2000] explained the complexity of the deformation field as resulting mainly from the superposition of two major components: postseismic deformation after the 1964 earthquake (trenchward motion), and elastic deformation from the shallow locked part of the megathrust. The latter component shows significant along-strike variations, caused by a contrast between the wide locked zone beneath and south of Prince William Sound and the eastern Kenai Peninsula with a narrow or nonexistent locked zone beneath and south of the western Kenai Peninsula. Where the elastic signal is smaller or missing, the trenchward velocities extend closer to the trench. Where the elastic signal is large, such as in the Prince William Sound area, velocities in the direction of the Pacific–North America relative motion extend much further inland. Where these two components cancel, the remaining westward motion of sites may indicate a westward tectonic escape of material away from the colliding Yakutat terrane.
The pattern of vertical velocities (Plate 1) agrees with the general predictions of the classic subduction zone model [Savage, 1983], with subsidence found near the coast and offshore, and uplift found inland. A possible second region of subsidence is observed NW of Upper Cook Inlet. The amplitudes of vertical velocities are substantial, with the highest subsidence rates being observed on the Pacific coast of the Kenai Peninsula, seaward of Seward, and across Prince William Sound. There are two nearly separate regions of subsidence along the coast, one in Prince William Sound and the other on the coastal Kenai Peninsula, separated by a zone of essentially zero vertical motion that passes through the SW end of Montague Island. Middleton Island, well offshore, shows extremely rapid uplift, as noted by Savage et al. [1998]. Middleton Island shows geological evidence for long-term uplift, including a series of uplifted marine terraces [Plafker et al., 1992].
3.3. Interior Alaska and the Denali Fault
There is abundant seismicity north of the Denali fault, with dense clusters of microseismicity in the Fairbanks area (Figure 5), and a diffuse band that extends north to the Arctic Ocean [Ruppert et al., this volume]. Some of these earthquakes occur just north of the Denali fault on faults that parallel the Denali, and others may occur on faults that make up a foreland fold-and-thrust belt at the northern edge of the Alaska Range [Bemis, 2004; Bemis and Wallace, 2007; Lesh and Ridgway, 2007]. Most, however, occur on a series of NNE-trending strike-slip faults that extend well to the north of the Alaska Range [Page et al., 1995]. Focal mechanisms for the larger earthquakes on these faults show left-lateral strike-slip motion [Ruppert et al., this volume]. The Kantishna cluster, the region of the most abundant microseismicity in Alaska, is located where one of the NNE-trending zones intersects the fold belt. Seismicity in the Kantishna cluster includes earthquakes on structures parallel to the Denali fault as well as NNE-trending structures [Burris, 2007].
Velocities of sites in the Alaska Range and Interior Alaska to the north (Figure 5
