Paleoseismic Analysis of the South Flank of Kilauea Volcano, Hawaii

BY

Eric Christopher Cannon

B.S. (University of California, Davis) 1997

THESIS

Submitted in partial satisfaction of the requirements for the degree of

MASTER OF SCIENCE

in

Geology

in the

OFFICE OF GRADUATE STUDIES

of the

UNIVERSITY OF CALIFORNIA

DAVIS

Approved:

_____________________________________

_____________________________________

_____________________________________

Committee in Charge

1999

-i-

Eric Christopher Cannon

December 1999

Geology


Abstract

I analyze fault offsets from the 1975 Kalapana earthquake and fault offsets in prehistoric lava flows to characterize the past faulting behavior of the Hilina fault system, on the south flank of Kilauea Volcano, Hawaii. Historical accounts of earthquakes on Hawaii only date back to 1823, whereas evidence of prior fault offset exists in the prehistoric lava flows. The main assumptions for the fault offset analysis are that the fault offsets of the Kalapana earthquake are typical of past large (M>7) earthquakes, and that the Hilina faults only rupture due to large (M>7) earthquakes. Conclusions reached suggest that the Hilina fault system is probably a shallow normal fault system (1-5 km depth) that slips only due to large (M>7) earthquakes originating on the basal detachment. Coastal displacements associated with the Kalapana earthquake occurred from slip both on the shallow Hilina fault system and along the ~8-10-km-deep basal detachment. For the time period back to 200-750 yr B.P. (years before present relative to 1950), horizontal and vertical offset rates across the Hilina fault system are 0.4 to 1.2 cm/yr and -0.2 to -2.0 cm/yr respectively. For this time period, 1.2-17.5 prehistoric faulting events, with a recurrence interval of 650-20 years, could have produced the total fault offset measured in the prehistoric lava flows. The wide range for estimates on the number of prehistoric faulting events suggests these events (1) do not necessarily produce fault offsets equal to fault offsets associated with the Kalapana earthquake, (2) do not produce repeated fault offsets of equal magnitude and orientation at the same location along the faults, and (3) do not have uniform recurrence intervals.


-ii-

Table of Contents

ii.......Abstract

iv......List of Figures

vi......List of Tables

vii.....Acknowledgements

1......Chapter 1. Introduction


4......Chapter 2. Shallow Normal Faulting and Block Rotation Associated with the 1975 Kalapana Earthquake, Kilauea Volcano, Hawaii

4......Abstract

5......Introduction

7......Tectonic Setting

11....Kinematic Models of the South Flank

12....Measurements of Fault Offsets Associated with the Kalapana Earthquake

15....Comparison of Observed and Model Displacements with Fault Offsets

19....Discussion of Proposed Model of Shallow Normal and Deep Detachment Slip Associated with the Kalapana Earthquake

22....Conclusions

23....Acknowledgements

24....References Cited


27....Chapter 3. Prehistoric Fault Offsets on Hilina Fault System, South Flank of Kilauea Volcano, Hawaii

27....Abstract

28....Introduction

30....South Flank Tectonic Setting

33....The Kalapana and Great Kau Earthquakes

34....Methods of Determining Fault Offset and Time-Averaged Offset Rates

38....Kalapana Earthquake Ground Fracture and Fault Offset Data

39....Fault Offsets in Prehistoric Lava Flows

50....Discussion

55....Conclusions

56....Acknowledgements

57....References Cited


60....Chapter 4. Conclusion


67....Appendix A. Fracture maps and fault offset measurements

-iii-

List of Figures

Chapter 2

6......Figure 2.1 - Location map of south flank of Kilauea Volcano

8......Figure 2.2 - South flank seismicity (1960-1991)

10....Figure 2.3 - Map of horizontal and vertical fault offsets for Kalapana earthquake

16....Figure 2.4 - Model and observed displacements, and fault offsets for Kalapana earthquake

21....Figure 2.5 - Northwest-southeast cross-section C-C' of south flank of Kilauea Volcano


Chapter 3

29....Figure 3.1 - Location map for Hilina fault system

31....Figure 3.2 - Shaded relief map of topography and bathymetry for south flank

36....Figure 3.3 - Location map for Kalapana earthquake ground fractures

37....Figure 3.4 - Geologic map of south flank of Kilauea Volcano, Hawaii

40....Figure 3.5 - Horizontal and vertical fault offsets along Hilina faults

42....Figure 3.6 - Location map and fault offsets at Kealakomo Overlook site

45....Figure 3.7 - Diagram of horizontal and vertical fault offset rates from prehistoric lava flows

46....Figure 3.8 - Graphs of horizontal and vertical fault offsets verses lava flow age in Hilina fault system

48....Figure 3.9 - Maps of horizontal and vertical fault offsets across Hilina fault system

49....Figure 3.10 - Graphs of horizontal and vertical fault offsets verses lava flow age across Hilina fault system

-iv-

Chapter 4

62....Figure 4.1 - Map of proposed future work within Hilina fault system

64....Figure 4.2 - Map of fractures between Puueo Pali and Puu Kapukapu

65....Figure 4.3 - Map of western end of Holei Pali


Appendix A

68....Figure A-1 - Location map for fracture, width, and fault offset maps of 1975 Kalapana earthquake ground deformation

69....Figure A-2 - Map of fractures in Mauna Ulu lava flows on Poliokeawe Pali

70....Figure A-3 - Map of fracture width in Mauna Ulu lava flows on Poliokeawe Pali

71....Figure A-4 - Map of horizontal fault offsets in Mauna Ulu lava flows on Poliokeawe Pali

72....Figure A-5 - Map of vertical fault offsets in Mauna Ulu lava flows on Poliokeawe Pali

73....Figure A-6 - Map of fractures in Mauna Ulu lava flows on Holei Pali

74....Figure A-7 - Map of fracture width in Mauna Ulu lava flows on Holei Pali

75....Figure A-8 - Map of horizontal fault offsets in Mauna Ulu lava flows on Holei Pali

76....Figure A-9 - Map of vertical fault offsets in Mauna Ulu lava flows on Holei Pali

77....Figure A-10 - Map of fractures in Mauna Ulu and prehistoric lava flows on Apua Pali

78....Figure A-11 - Map of fracture width in Mauna Ulu and prehistoric lava flows on Apua Pali

79....Figure A-12 - Map of horizontal fault offsets in Mauna Ulu and prehistoric lava flows on Apua Pali

-v-

80....Figure A-13 - Map of vertical fault offsets in Mauna Ulu and prehistoric lava flows on Apua Pali

81....Figure A-14 - Map of fractures in Mauna Ulu and prehistoric lava flows on Poliokeawe and Holei Pali


List of Tables

Chapter 2

14....Table 2.1 - Kalapana earthquake fracture characteristics for Hilina faults

Chapter 3

44....Table 3.1 - Summary of south flank double-fracture locations


-vi-

Acknowledgements

This work was supported by a United States Geological Survey National Earthquake Hazards Reduction Program grant 14534-HQ-98-GR-1024, a Geological Society of America Student Research grant, and U. C. Davis Department of Geology Cordell Durrell funds. I appreciate the comments and support from my thesis committee members Roland Bürgmann, Jim McClain, and Rob Twiss. Roland Bürgmann and Mike Poland conducted initial GPS surveys in the Hilina fault system during the summer of 1995. James Kellogg and William Chadwick provided preliminary fault offset measurements and field notes from their work in 1987. The following folks offered invaluable assistance at Hawaiian Volcano Observatory: Don Swanson, Mike Lisowski, Asta Miklus, Kristine Larson, Taeko Jane Takahashi, Arnold Okamura, Richard Fiske, Paul Okubo. I appreciate the help provided by numerous staff at Hawaii Volcanoes National Park, including the park rangers and visitor center staff. Tim Tunison and Bobby Camara from the Resource Management division of the National Park Service assisted in permitting and questions related to the natural history of the park. Sue Owen and Peter Cervelli from Stanford University, and David Schmidt and Michelle Wibler from U. C. Berkeley, provided help with my research. With the completion of yet another geologic adventure, I thank my parents for their support of my goals and help with fieldwork logistics.


I acknowledge many people in the Department of Geology here at Davis for their assistance. Louise Kellogg, Doug Neuhauser, Paul Waterstraat, Kate Hill, Donna Hunt, Gerald Bawden, Jai Sukhatme, and Larry Guenther provided computer assistance, while David Manaker, Kaylene Keller, and Jim Burke provided Global Positioning System (GPS) equipment training. I completed two

-vii-

successful field seasons in summer 1998 and 1999 with volunteer field assistants Gwen Pikkarainen, Chad Fleschner, Jason Bariel, and Jim Weigel. Numerous staff have helped me over the years: Norm Winter, Joseph Abril, Mary Graziose, Janice Fong, Judy Hendrickson, Helen Rogers, Susan Olson, Marilyn DeMoss. Fellow graduate students Glenn Jaecks, Jason Mayfield, and David Tinker provided helpful comments on early versions of manuscripts.

-viii

1

Chapter 1.


Introduction

"Pele struck the ground heavily with her feet. Again and again she stamped in wrath. Earthquakes swept the lands of Kahuku. Then the awful fiery flood broke from the underworld, and swept down over Kahuku. On the crest of the falling torrent of fire rode Pele, flashing the fires of her anger in great explosions above the flood."

From Westervelt (1916) describing Pele's violent actions over Kahuku, the land covered by lava flows.

Given her fiery demeanor, Pele, Goddess of the Hawaiian Volcanoes, certainly makes studying the volcanic and tectonic processes of the south flank of Kilauea Volcano an exciting and demanding research experience. This research focuses on evaluating south flank ground deformation from past large (M>7) earthquakes. My field crew and I mapped fractures and measured fault offsets associated with the 1975 Kalapana earthquake (hereafter referred to as the Kalapana earthquake). We also measured fault offsets in prehistoric lava flows to characterize the past faulting behavior of the Hilina fault system, on the south flank of Kilauea Volcano, Hawaii. Historical accounts of earthquakes on Hawaii only date back to 1823, requiring geologic fieldwork to identify ground deformation from prehistoric earthquakes. Evaluating past faulting behavior can improve kinematic models of south flank displacement by providing a longer-term history of fault offset than can modern geodetic measurements. These kinematic models may help improve assessment of volcanic, seismic, tsunami, and mass movement hazards of the south flank region.


2


This thesis contains two individual papers to be submitted to earth science journals. Chapter 2 contains a paper titled Shallow Normal Faulting and Block Rotation Associated with the 1975 Kalapana Earthquake, Kilauea Volcano, Hawaii, to be submitted as a brief paper to Short Notes of the Bulletin of the Seismological Society of America. Analysis of fault offset in the Hilina fault system, observed displacements from geodetic measurements, and model displacements of basal detachment slip suggest that the coastal regions experienced slip on the Hilina fault system in addition to ~8-10-km-deep basal detachment slip during the Kalapana earthquake. The Hilina faults are probably shallow normal faults, descending possibly to 1-5 km depth, rather than deep normal faults splaying off the basal detachment. Slip on these faults probably only occurs during large (M>7) south flank earthquakes.


Chapter 3 presents a paper to be submitted to Journal of Geophysical Research titled Prehistoric Fault Offsets of the Hilina Fault System, South Flank of Kilauea Volcano, Hawaii. Prehistoric faulting behavior of the south flank can only be determined by identifying fault offsets of prehistoric flows within the Hilina fault system. This study begins by evaluating fault offset due to the Kalapana earthquake. The Kalapana earthquake is unique in two aspects. First, the Kalapana earthquake is the only historical large (M>7) south flank earthquake for which records of geodetic measurements of displacements, seismicity, and fault offset measurements are all available. Second, observations suggest the Hilina fault system has only slipped twice in historical times, during the 1868 M=7.9 Great Kau and 1975 M=7.2 Kalapana earthquakes. Slip on the Hilina faults from the Kalapana earthquake is used to define fault offsets for a "typical large (M>7) south flank earthquake", a "Kalapana-type" event. By comparing fault offsets in prehistoric lava flows of known age with neighboring


3

fault offsets from the Kalapana earthquake (expressed as fractures in the 1969-1974 Mauna Ulu lava flows), initial estimates are presented of the number of Kalapana-type earthquake events to have occurred since the cooling of the lava flow.


Chapter 4 concludes this thesis, summarizing the findings of this research. The discussion includes interesting questions and unanswered problems that will require continued fieldwork. Perhaps these regions will be visited in future research projects.


References Cited

Westervelt, W. D. (1916). Hawaiian Legends of Volcanoes (Mythology), Ellis Press, Boston, 204 p.



4

Chapter 2.

Shallow Normal Faulting and Block Rotation Associated with the 1975 Kalapana Earthquake, Kilauea Volcano, Hawaii

Eric C. Cannon and Roland Bürgmann


Abstract


We analyze fault offset measurements and geodetic data to improve kinematic models of fault slip along the Hilina fault system due to the 1975 M=7.2 Kalapana earthquake. The Hilina fault system is located on the south flank of Kilauea volcano. M>6 earthquakes and aseismic displacement transport the mobile south flank southeast along a ~8-10-km deep basal detachment. Only the Kalapana earthquake and the 1868 M=7.9 Great Kau have produced slip on Hilina faults, in addition to basal detachment slip. The relationship between basal detachment slip at depth and surface slip is unclear. Observed displacements from geodetic measurements at the coast significantly exceed those expected from a dislocation model of slip on a ~8-10-km-deep detachment. Slip on basal detachment dislocation alone cannot explain surface displacements. We suggest slip on the Hilina fault system and slip on the basal detachment can explain coastal displacements. To explain observed displacements and fault offsets associated with the Kalapana earthquake and other large (M>7) earthquakes, we envisage a kinematic model of the Hilina fault system with shallow normal faults descending to 1-5 km depth, only slipping due to large (M>7) earthquakes, in addition to slip on the deep basal detachment. Horizontal offsets greater than vertical offsets, marine seismic surveys, leveling surveys, paleomagnetic analysis of rotated lava flows, a hyaloclastic layer at ~1-3 km depth, and surface expressions of lava flows all suggest hangingwall blocks of the Hilina fault system may have slipped and rotated on a set of shallow normal fault surfaces during the Kalapana earthquake.


5

Introduction

Major earthquakes (M > 6) have occurred on the south flank of Kilauea Volcano in historic times, for example in 1823, 1868, 1975, 1954, and 1989. These earthquakes involved slip along an ~8-10-km-deep, subhorizontal basal detachment, and the larger events caused extensive surface rupture along the Hilina fault system (Lipman et al., 1985; Wyss, 1988; Bryan, 1992). Hazards associated with major Kilauea earthquakes include significant ground subsidence, tsunamis, and potential catastrophic mass movements (Tilling et al., 1975; Lipman et al., 1985, Ma et al., 1999). The Ml=7.2 November 29, 1975 Kalapana earthquake (hereafter referred to as the Kalapana earthquake), the largest south flank earthquake this century, generated a local tsunami 14 m high and produced up to 8 m horizontal displacement seaward and 3.5 m subsidence in coastal regions (Lipman et al., 1985). Similar coastal subsidence was observed following the 1868 M=7.9 Great Kau earthquake (Swanson et al., 1976; Wyss, 1988). Major south flank earthquakes are thought to involve slip on the basal detachment driven by rift intrusions and gravitational spreading (Swanson et al., 1976; Dieterich, 1988; Delaney and Denlinger, 1999).


We plan to study the poorly understood relationship between slip on the Hilina fault system (Figure 2.1) and slip on the basal detachment by comparing observed displacements, model displacements of basal detachment slip, and fault offset measurements. The kinematic model that best explains current aseismic displacements assumes slip on the basal detachment only (Owen et al., 1995). However, our surface ground deformation studies of the Kalapana earthquake, complimenting geodetic measurements of displacements across the Hilina


6



7

fault system suggest this kinematic model significantly underestimates coastal displacements. Our revised kinematic model of normal fault slip on the Hilina fault system, in addition to basal detachment slip, better explains coastal displacements associated with the Kalapana earthquake. Surface and subsurface evidence suggests that the Hilina fault system may be a shallow normal fault system, rather than a deep normal fault system with splays off of the basal detachment. We suggest that large (M>7) earthquakes displace coastal regions by shallow normal faulting of the Hilina fault system as well as by slip on the basal detachment.


Tectonic Setting

The Kalapana earthquake mainshock and aftershocks, in conjunction with geologic, geodetic observations, and tsunami studies, provide an improved understanding of the south flank structure of Kilauea Volcano. Mainshock focal mechanism studies by Ando (1979) and Furumoto and Kovach (1979) and observed displacements from geodetic measurements (Lipman et al., 1985; Bürgmann et al., 1999) indicate southeast transport of the mobile wedge-shaped south flank along a ~8-10-km-deep subhorizontal fault striking approximately northeast. Kalapana earthquake aftershock activity was concentrated along a ~8-10-km-deep, shallowly north-dipping band of microseismicity (Ando, 1979; Got et al., 1994; Gillard et al., 1996). Microseismicity (Figure 2.2) delineates the 2-3-km-thick zone, indicating the region of the weak boundary of ocean sediment separating the mobile south flank block and Cretaceous Pacific oceanic lithosphere (Hill, 1969; Thurber and Gripp, 1988). Multiplet relocation (Got et. al, 1994) of earthquakes collapsed this 2-3-km thick zone into a 100-200-m thick band of seismicity at 8.5±1.5 km depth, dipping 6°±-4° northward. Bryan (1992) identified a similar band of aftershock concentration at 10 km depth for the 1989 earthquake, again interpreted as microseismicity on the basal detachment


8



9

following the rupture (Árnadóttir et al., 1991). Even though the majority of Kalapana earthquake coseismic moment release occurred as subhorizontal slip on the basal detachment, extensive normal faulting was documented on the surface along the Hilina fault system.


Over 25 km of surface rupture occurred along the Hilina fault system from the Kalapana earthquake (Lipman et al., 1985). Kellogg and Chadwick (1987) collected the first detailed fault offset measurements from the Kalapana earthquake (Figure 2.3). Lipman et al. (1976) suggested that ~1.5 m of vertical offset along much of the Hilina faults could account for approximately two-thirds of the coastal subsidence, implying the Hilina fault system played a significant role in south flank deformation associated with the Kalapana earthquake. Fault morphology of the Hilina fault system suggests prior fault offset from major earthquakes (Tilling et al., 1985; Cannon and Bürgmann, 1999).


The Kalapana earthquake produced a marked change in the style of south flank deformation. From 1896 to the Kalapana earthquake, as much as 1 m northwest-southeast shortening occurred across much of the coastal south flank region (Swanson et al., 1976). After the Kalapana earthquake, geodetic baselines experienced periods of extension and contraction from 1976 until the dike intrusion event initiating Puu Oo eruption occurred in January 1983 (Delaney et al., 1998). The south flank has been displacing to the southeast since 1983 (Delaney et al., 1998; Owen et al., 1995, Owen et al., 1999) with the displacement field modeled as aseismic slip on a 9-km-deep horizontal basal detachment with concurrent rift zone opening at 2-9-km-depth (Owen et al., 1995). Seismicity rates have not regained levels observed prior to the Kalapana earthquake (Delaney et al., 1998), suggesting the south flank is still recovering from the


10



11

event. Assuming the Hilina faults have remained inactive since the Kalapana earthquake (Delaney et al., 1998), the fractures we observe almost 25 years after the event represent the original surface rupture.


Kinematic Models of the South Flank

The Hilina fault system is located in the coastal region of the mobile south flank block. The Hilina faults display arcuate, south-facing, normal fault scarps trending east to northeast, with a maximum scarp height of ~500 m. Two kinematic models exist for the south flank, a "shallow" and "deep" model. An ~8-10-km-deep basal detachment and surface expression of the Hilina normal faults are common to both models. However, the "shallow" model treats slip on the Hilina fault system as independent of the basal detachment slip. In the "deep" model, the Hilina faults descend to the basal detachment as normal fault splays. Both models allow for normal faulting in the Hilina fault system and basal detachment slip associated with the Kalapana earthquake. Triggering mechanisms for slip on the Hilina faults may be different for both models. The "shallow" model could produce a greater risk for catastrophic slumps and mass movement due to slip on shallow fault structures than with the "deep" model of the Hilina fault system.


The Hilina fault system may be a set of shallow (Ando, 1979; Swanson et al., 1976, Gillard et al., 1996). Several lines of evidence support this shallow fault interpretation. The arcuate surface traces of the Hilina faults resemble spoon-shaped listric normal faults. Riley (1996) conducted a paleomagnetic study of the rotation of lava flows in the Puu Kapukapu block relative to the Hilina Pali. She determined 12°±6° of landward rotation of the Puu Kapukapu block occurred on a listric normal fault with 5.2 km depth. Ponding of lava flows against the fault


12

surface suggests landward block rotation of hangingwall (Swanson et al., 1976). A hyaloclastic layer at 1-3 km depth (Swanson et al., 1976) could act as a shallow basal detachment for Hilina faults. Marine seismic reflection surveys offshore of Halape (Figure 2.1) (Morgan et al., 1998; Morgan et al., 1999) indicate the presence of slump structures, possibly related to the Hilina normal faults. An offshore submarine bench at 1 km depth (Chadwick et al., 1993) may represent the submarine trace of the shallow basal detachment. Studies of compressive wedges (Yin, 1993) support the existence of normal faults within the mobile wedge. Gillard et al. (1996) state that shallow slip on predominantly aseismic Hilina faults could have produced Hilina vertical fault offset associated with the Kalapana earthquake.


A second model considers the Hilina fault system as deeply-rooted normal faults splaying off the ~8-10-km-deep basal detachment (Lipman et al., 1985; Okubo et al., 1997). Microseismicity is detected at 8-10 km depth beneath the upper south flank (Figure 2.2), possibly representing the splay junction of the deep Hilina Pali descending to the basal detachment (Okubo et al., 1996). P-wave tomographic studies show a significant lateral velocity gradient steeply dipping to the southeast beneath the Hilina fault system (Okubo et al, 1996). This velocity gradient, separating low velocity seaward rocks from high velocity inland rocks, may indicate the presence of the Hilina Pali descending to the basal detachment.


Measurements of Fault Offsets Associated with the Kalapana Earthquake

We analyze ground fractures and geodetic data associated with the Kalapana earthquake to try to understand the mechanisms of displacement for coastal regions. To document Kalapana earthquake ground fractures, we


13

measure fault offset using piercing point features in fractured 1969-1974 Mauna Ulu lava flows that drape the central Hilina fault system. Sawtooth-shaped fractures in Mauna Ulu pahoehoe best preserve fault offset as piercing points. Fault offset is measured as plunge, azimuth, and magnitude of the fault offset. Fractures and fault offsets are not concentrated along a single fault strand, but rather distributed over a fracture zone. To measure fault offset across a fracture zone, we sum multiple individual fault offsets along traverses orientated perpendicular to the general trend of the fault scarp. As an example, horizontal offset vector T50 in Figure 2.3 is calculated by summing 13 individual measurements of horizontal fault offset (labeled "a" through "m". 73 traverses fault offset traverses contain over 700 individual fault offset measurements during, an average of 10 individual fault offset measurements per traverse.


We summarize the trend of over 18,000 m of fracture and over 200 fault offset measurements on the Poliokeawe Pali, Holei Pali, and Apua Pali in Table 2.1. Fracture trends and fault offset azimuths are perpendicular to each other within uncertainties for all three faults, indicating extension without a horizontal shear component. Average fault offset azimuths trend generally southeast, parallel to the south flank displacement associated with the Kalapana earthquake (Figure 2.1). The largest Kalapana fault offset for any Hilina fault located on the Holei Pali, with 3.3 m of total offset summed along a traverse of 16 individual measurements of fault offset.


We assume fractures in the Mauna Ulu lava flows are tectonic in origin, resulting from fault slip along the Hilina fault system associated with the Kalapana earthquake. Compressional fold and fault structures are identified at the base of fault scarps, suggesting mass-movement did not produce the surface


14



15

fractures. Where possible, we descend into fractures to observe the fracture surface. Slip surfaces parallel or sub-parallel to the bases of lava flows do not occur, indicating that the upper few meters of Mauna Ulu pahoehoe had not detached from the subsurface to produce surface fractures. Even though fracture azimuth trends (Table 2.1) are generally parallel to the direction of south flank aseismic displacement and Kalapana earthquake displacement, fractures seem to trend parallel with local fault traces. This suggests local fault geometry influences fracture geometry.


Comparison of Observed and Model Displacements with Fault Offsets

We propose slip of the Hilina fault system and ~8-10-km deep basal detachment slip contributed to coastal displacement associated with the Kalapana earthquake. Our initial kinematic model is presented in Figure 2.4a. A single dislocation represents the basal detachment and rift zone opening is permitted, no dislocation is included for the Hilina fault system. The initial kinematic model best explains horizontal aseismic displacement of the south flank measured with GPS equipment (Owen et al., 1995). For a basal detachment slip event, geodetic stations landward of the Hilina fault system (i.e. the footwall of the Hilina fault system) displace toward the southeast due to basal detachment slip only. Model displacements (open vectors in Figure 2.4a) should agree with observed displacements from geodetic measurements (solid vectors in Figure 2.4a). If in addition to basal detachment slip, secondary slip occurs on the Hilina fault system, geodetic stations on the hangingwall of the Hilina fault system will experience observed displacement due to basal detachment slip and slip on the Hilina normal faults. Residuals s1 and s2 (Figure 2.4a) represent difference between model and observed displacement. We suggest these residuals represent slip associated with slip on the Hilina fault system.


16



17

Secondary slip on the Hilina fault system may result from a shallow or deep fault geometry. If the Hilina faults are shallow normal faults not attached to the basal detachment, the Hilina slip will be independent of basal detachment slip. If the Hilina faults descend to the basal detachment, hangingwall displacement will result from normal slip on the deep fault splay transferred to slip on the basal detachment. We cannot determine the geometry of the Hilina fault system from model displacements due to basal detachment slip. However, we can determine if coastal displacements are not well explained by basal slip alone, and propose slip on the Hilina fault system as a solution to produce significant coastal displacements.


Figure 2.4b displays observed horizontal displacements (solid vectors, computed from trilateration data) and vertical displacements (open vector, from tide gage and leveling data) calculated relative to geodetic station HVO162 (Bürgmann et al., 1999). By calculating observed displacements relative to HVO162, which is located on the footwall of the Hilina fault system, coastal geodetic stations on the hangingwall of the Hilina fault system reveal displacement relative to the footwall block. We note that the extension and elevation changes for the hanigingwall coastal stations relative to footwall inland stations are maximum in the region up to 5 km inland from the coast. The Hilina footwall block itself moved significantly to the southeast with respect to stations north of Kilauea caldera and the rift zones, as noted by northwest-trending solid vectors of 1-2 m magnitude.


To separate and evaluate the contributions of the Hilina fault system and basal detachment on the surface displacements, we calculate model displacements as a function of basal detachment slip only (open vectors in


18

Figures 2.4c and 2.4d). Bürgmann et al. (1999) present a dislocation model of the Kalapana earthquake derived from an inversion of available geodetic data (reanalysis of trilateration, leveling, and tilt data collected by Hawaiian Volcano Observatory, Lipman et al., 1985) but excluding those geodetic data from sites in the hangingwall of the Hilina faults. The dislocation model is constrained to be consistent with Kalapana earthquake seismicity and geodetic measurements of south flank displacements. Initial constraints include an 8-10-km-depth and 20° landward to 10° seaward dip for the dislocation. The basal detachment is represented by a dislocation plane 47 km by 36 km in size (midpoint depth 8.2 km), 060° NW strike and 12° dip, 9.5 m of slip to the southeast. The model displacements represent displacements expected at south flank locations due to basal detachment slip only.


Comparison of observed and model displacements indicates observed displacements significantly exceeded model displacements for coastal geodetic stations (Figures 2.4c and 2.4d). Footwall stations (Goat, Goat 2, Pilau-3, Panau) remain stationary relative to HVO162, while Apua Pt2, Kaena Pt, and Laeapuki displace horizontally to the southeast 0.8 to 3.4 m. Horizontal displacements for geodetic measurements (solid vectors) exceed horizontal model displacement (open vectors) by as much as ~3 m. Geodetic measurements of observed subsidence also exceed model displacements by up to ~1 m.


In order to compare fault offsets across the Hilina fault system with observed and model displacements, we identify traverses located near three geodetic baselines. Fault offset traverses (Figure 2.3) are summed along GOAT-APUA and H162-KAEN baselines to calculate horizontal and vertical fault offsets across the Hilina fault system (Figures 2.4e and 2.4f). No Mauna Ulu lava flows


19

exist near the PANU-LAEP baseline, so no fault offset can be calculated. Solid vectors represent fault offset across the Hilina fault, and open vectors are the residual vectors determined by subtracting the model displacements from observed displacements (Figures 2.4c and 2.4d). We note that horizontal residual vectors show a transition from southeast to east azimuths, for stations farther to the east. The H162-KAEN and PAUN-LAEP vertical residual vectors are 11 cm and 9 cm respectively, almost not detectable given the displacement scale in Figure 2.4f. Horizontal and vertical fault offset measurements are greater than residual vectors.


Discussion of Proposed Model of Shallow Normal and Deep Detachment Slip Associated with the Kalapana Earthquake

Faulting in the south flank region due to major earthquakes, such as the Kalapana earthquake and probably the Great Kau earthquake, involves slip on Hilina normal faults, in addition to slip on the ~8-10-km-deep basal detachment. Analysis of the Hilina fault offsets, and observed and model displacements, show that fault offsets greatly exceed model displacements of basal detachment slip (Figures 2.4c and 2.4d), and that fault offsets are greater than residual displacements (Figures 2.4e and 2.4f). Fault offsets may exceed model displacements calculated from basal detachment slip if significant slip occurred within the Hilina fault system. Revised dislocation models need to include dislocations for the Hilina fault system. Comparison of fault offset with observed and model displacements suggests that faulting of the Hilina fault system produced significant displacement of coastal regions.


We speculate that the geometry of the central Hilina fault system is a set of shallow normal faults rather than deep normal faults splaying off the basal


20

detachment. We define the offset ratio as horizontal fault offset divided by vertical fault offset. For a simple model of normal dip-slip on a fault surface dipping less than 45°, the value of horizontal displacement exceeds vertical displacement producing an offset ratio greater than one. As the offset ratio increases, the angle of the fault dip decreases. 39 of 43 traverses have horizontal fault offsets greater than vertical fault offsets for the Kalapana earthquake (Figure 2.3). Kalapana earthquake offset ratios of slip range from 0.2 to 31.5, suggesting that most slip on the Hilina faults occurred on low-angle normal faults. Deep normal faults splaying off the basal detachment would have offset ratios less than one, showing vertical fault offset greater than horizontal fault offset. Leveling along the Chain of Craters Road (line of section B-B' in Figure 2.5 from data in Figure 2.4b) indicates subsidence of leveling stations located in the Holei Pali and Apua Pali fault zones. Subsidence in these zones is interpreted as landward block rotation of the hangingwall blocks. Low-angle normal faults implied from offset ratios, and landward block rotation of the hangingwall interpreted from leveling along the Chain of Craters Road, adds evidence to support to the shallow normal fault interpretation of the Hilina fault system.


Shallow normal faulting on the south flank coast has several implications regarding kinematic models of wedge-shaped mobile volcanic flanks, geodetic measurements of displacement on volcanic flanks, and hazard assessment. Observed displacements along the south flank coast resulted from slip on the Hilina faults and basal detachment. Data sets containing coastal observed displacements should not be used to evaluate basal detachment slip unless the effects of the Hilina fault slip are removed from coastal observed displacements. Tsunami generation from the Kalapana earthquake probably did not result from slip on Hilina fault system, but rather submarine slumping or basal detachment


21



22

slip (Ma et al., 1999). However, future large (M>7) earthquakes could conceivably produce a catastrophic mass movement of the Hilina fault system and generate a significant tsunami.


Slip on the Hilina fault system plays an undoubtedly important and poorly understood role in the kinematics of the south flank. Inland seismic reflection and refraction surveys, borehole logging, and high spatial-resolution geodetic measurements from synthetic-aperture radar, and densely-spaced GPS networks might provide more evidence to evaluate the structure and displacement of Hilina faults. Future work should focus on identifying depths and structures of faults in the Hilina fault system, removing the effects of Hilina fault slip on coastal geodetic stations when evaluating only slip on the basal detachment, and evaluating earthquake and tsunami hazard hazards using a shallow fault geometry for the Hilina fault system.


CONCLUSIONS

We have accomplished detailed mapping of the 1975 Kalapana earthquake surface rupture along the central Hilina faults. Comparison of fault offsets, observed displacements from geodetic measurements, and model displacements of basal detachment slip, supports a kinematic model for the Hilina faults as a set of shallow normal faults. Our dislocation model for basal detachment slip does not include coastal geodetic displacements in the inversion, nor dislocations for Hilina faults. Thus coastal displacements calculated from this model should represent the effects of displacement on the basal detachment alone. We show that observed displacements greatly exceed model displacements for coastal geodetic stations, by as much as ~3 m horizontal and ~1 m vertical. We propose that slip on the Hilina fault system, in addition to basal detachment slip, accounts


23

for coastal displacements associated with the Kalapana earthquake. Coastal displacements associated with the Kalapana earthquake were accommodated both by slip on the Hilina fault system and basal detachment slip.


We interpret the Hilina fault system as a shallow normal fault system rather than a deep fault system splaying off the basal detachment. Offset ratios of Kalapana earthquake and subsidence in fault zones along the Chain of Craters Road suggests the Hilina fault system is a set of shallow normal faults. We suggest slip on Hilina faults only occurs when triggered by a large (M>7) south flank earthquake. Further work is needed to provide geometric and slip constraints on the Hilina fault system. Accurate kinematic models are fundamental for evaluating significant seismic and tsunami hazards associated with mass movement of the Hilina fault system and mobile south flank block.


ACKNOWLEDGEMENTS

This work is supported by U. S. Geological Survey NEHRP grant 14534-HQ-98-GR-1024 and a Geological Society of America Student Research grant (E.C.C.). Many thanks to folks at Hawaiian Volcano Observatory for assistance, especially Mike Lisowski, Asta Miklus, Arnold Okamura, Don Swanson, and Taeko Jane Takahashi. Peter Cervelli, Sigurjon Jonsson and Kristine Larson provided assistance during fieldwork. James Kellogg and William Chadwick provided fault offset data from their Hilina fieldwork. We thank our field assistants Jason Bariel, Chad Fleschner, Gwen Pikkarainen, and Jim Weigel for their excellent efforts. The Generic Mapping Tools (GMT) program has been invaluable (Wessel and Smith, 1995).


24

REFERENCES CITED

Árnadóttir, T., P. Segall, and P. Delaney (1991). A fault model for the 1989 Kilauea south flank earthquake from leveling and seismic data, Geophys. Res. Lett. 18, 2217-2220.

Ando, M. (1979). The Hawaii Earthquake of November 29, 1975: Low Dip Angle Faulting Due to Forceful Injection of Magma, J. Geophys. Res. 94, no. B13, 7616-7626.

Bürgmann R., S. E. Owen, and P. T. Delaney (1999). Mechanics of the 1975 Kalapana, Hawaii, earthquake, manuscript in preparation for submission to J. Geophys. Res.

Cannon, E. C. and R. Bürgmann (1999). Hilina fault offset rates on the south flank of Kilauea Volcano, Hawaii, manuscript in preparation for submission to J. Geophys. Res.

Bryan, C. J. (1992). A Possible Triggering Mechanism For Large Hawaiian Earthquakes Derived From Analysis Of The 26 June 1989 Kilauea South Flank Sequence, Bull. Seism. Soc. Am. 82, no. 6, 2368-2390.

Chadwick, W. W. Jr., Smith, J. R. Jr., Moore, J. G., Clague, D. A., Garcia, M. O., and C. G. Fox (1993). Bathymetry of South Flank of Kilauea Volcano, U.S. Geol. Surv Misc Field Studies, map MF-2231.

Delaney, P. T. and R. P. Denlinger (1999). Stabilization of Volcanic Flanks by Dike Intrusion, an Example from Kilauea, in press Bull. Volcan.

Delaney, P. T., R. Denlinger, M. Lisowski, A. Miklius, P. Okubo, A. Okamura, and M. K. Sako (1998). Volcanic spreading at Kilauea, 1976-1996, J. Geophys. Res. 103, no. B8, 18003-18023.

Denlinger, R. P. and P. Okubo (1995). Structure of the mobile south flank of Kilauea Volcano, Hawaii, J. Geophys. Res. 100, no. B12, 24499-24507.

Dieterich, J. H. (1988). Growth and persistence of Hawaiian volcanic rift zones, J. Geophys. Res., 93, 4258-4270.

Furumoto, A. S., and R. L. Kovach (1979). The Kalapana Earthquake of November 28, 1975: An Intra-plate Earthquake and its Relation to Geothermal Processes, Phys. Ear. Plan. Int. 18, 197-208.

Gillard, D., M. Wyss, and P. Okubo (1996). Type of faulting and orientation of stress and strain as a function of space and time in Kilauea's south flank, Hawaii, J. Geophys. Res. 101, no. B7, 16025-16042.


25

Got, J.-L., F. W. Fréchet, and W. Klein (1994). Deep fault plane geometry inferred from multiplet relative relocation beneath the south flank of Kilauea, J. Geophys. Res. 99, no. B8, 15375-15386.

Hill, D. P. (1969). Crustal structure of the island of Hawaii from seismic-refraction measurements, Bull. Seism.Soc. Am. 59, no. 1, 101-130.

Hill, D. P. and J. J. Zucca (1987). Geophysical constraints on the structure of Kilauea and Mauna Loa volcanoes and some implications for seismomagmatic processes, U. S. Geol. Surv. Prof. Pap. 1350, 903-917.

Kellogg, J. N. and W. Chadwick (1987). Neotectonic Study of the Hilina Fault System, Kilauea, Hawaii (abstract), Geol. Soc. Am. Abstracts with Programs 19, no. 6, 394.

Lipman, P. W., J. P. Lockwood, R .T. Okamura, D. A. Swanson, and K. M. Yamashita (1985). Ground Deformation Associated with the 1975 Magnitude-7.2 Earthquake and Resulting Changes in Activity of Kilauea Volcano, Hawaii, U. S. Geol. Surv. Profess. Pap. 1276, 45 pp.

Ma, K-F., H. Kanamori, and K. Satake (1999). Mechanism of the 1975 Kalapana, Hawaii, earthquake inferred from tsunsmi data, J. Geophys. Res., 104, no. B6, 13153-13167.

Morgan, J. K., G. F., and Leslie, S. (1998). Seismic reflection lines across Papa'u Seamount, south flank of Kilauea: The submarine expression of the Hilina Slump? EOS Trans. AGU, Fall Meeting Suppl., v. 79, p. 1008.

Morgan, J. K., G. F. Moore, D. J. Hill, and S. Leslie (1999). Confirmation of Volcanic Spreading Models for Kilauea's Mobile South Flank, Hawaii, from Marine Seismic Reflection Data, manuscript submitted to Geology.

Okubo, P. G., H. M. Benz, and B. A. Chouet (1997). Imaging the crustal magma sources beneath Mauna Loa and Kilauea volcanoes, Hawaii, Geology 25, no. 10, 867-870.

Owen, S., P. Segall, J. Freymueller, A. Miklius, R. Denlinger, T. Árnadóttir, M. Sako, and R. Bürgmann (1995). Rapid Deformation of the South Flank of Kilauea Volcano, Hawaii, Science 267, 1328-1332.

Owen, S., P. Segall, M. Lisowski, A. Miklius, R. Denlinger, J. Freymueller, T. Árnadóttir, T., and M. Sako (in press). The Rapid Deformation of Kilauea Volcano: GPS measurements between 1990 and 1996, J. Geophys. Res.

Riley, C.. (1996). A Paleomagnetic Study of Movement in the Hilina Fault System, South Flank of Kilauea Volcano, Hawaii (M.S. thesis), Houghton, Michigan Technological University, 61 pp.


26

Swanson, D. A., W. A. Duffield, and R.S. Fiske (1974). Displacement of the South Flank of Kilauea Volcano: The Result of Forceful Intrusion of Magma Into the Rift Zones, U. S. Geol. Surv. Profess. Pap. 963, 39 pp.

Thurber, C. H. and A. E. Gripp (1988). Flexure and Seismicity Beneath the South Flank of Kilauea Volcano and Tectonic Implications, J. Geophys. Res., 93, no. B5, 4271-4278.

Tilling, R. I.., R. Y. Koyanagi, P. W. Lipman, J. P. Lockwood, J. G. Moore, and D. A. Swanson (1976). Earthquake and Related Catastrophic Events Island of Hawaii, November 29, 1975: A preliminary report, U. S. Geol. Surv. Circular C 0740, 33 pp.

Wessel, P. and W. H. F. Smith (1995). New version of the Generic Mapping Tools released, EOS 76, 329.

Wyss, M. (1988). A Proposed Source Model For The Great Kau, Hawaii, Earthquake of 1868, Bull. Seismo. Soc. Am. 78, no. 4, 1450-1462.

Yin, A. (1993). Mechanics of Wedge-Shaped Fault Blocks 1. An Elastic Solution for Compressional Wedges, J. Geophys. Res., 98, no. B8, 14245-14256.


27

Chapter 3.

Prehistoric Fault Offsets of the Hilina Fault System, South Flank of Kilauea Volcano, Hawaii

Eric C. Cannon and Roland Bürgmann

Abstract

We evaluate fault offsets in prehistoric dated lava flows to characterize the past faulting behavior of the Hilina fault system. Historical accounts of earthquakes on Hawaii only date back to 1823, whereas lava flows of 750-1500 yr B.P. age still contain piercing points of fault offsets (yr B.P. - radiocarbon years before present relative to 1950). We define a "Kalapana-type" earthquake as a large (M>7) earthquake with fault offset values equal to fault offsets of the Hilina faults for the Kalapana earthquake. We compare fault offsets in prehistoric lava flows with neighboring Kalapana earthquake fault offsets to determine the number of Kalapana-type earthquakes that could have produced the observed prehistoric fault offset. We calculate the horizontal and vertical fault offset rate across the Hilina fault system for the time interval back to 200-750 yr B.P. as 0.4 to 1.2 cm/yr and -0.2 to -2.0 cm/yr respectively. Our best estimate of the number of Kalapana-type earthquakes to occur in lava flows 400-750 yr B.P. age at the Kealakomo Overlook site is 3-5 events, with a recurrence interval of 260-80 years. Evaluating three double-fracture outcrops, the number of Kalapana-type earthquakes to occur in lava flows of 200-750 yr B.P. age is 1.2-17.5 events, with recurrence intervals of 650-20 years. These wide ranges of past earthquake events and recurrence intervals indicate our assumption of a "Kalapana-type" earthquake is an over-simplification of past faulting events. Past large (M>7) earthquakes probably did not displace the Hilina faults with "Kalapana-type" fault offsets for each prehistoric faulting event. For a given location in the Hilina fault system, multiple faulting events probably did not produce uniform fault offset, but rather varied fault offset with each event. Large (M>7) earthquakes most likely have non-uniform recurrence intervals.


28

Introduction

The south flank of Kilauea Volcano, Hawaii has experienced two large (M>7) earthquakes that ruptured a large sub-horizontal detachment fault at the base of the volcanic edifice, the 1975 M=7.2 Kalapana earthquake and 1868 M=7.9 Great Kau earthquake (Figure 3.1) [Ando, 1979; Furumoto and Kovach, 1979; Tilling et al., 1975; Lipman et al., 1985; Wyss, 1988]. Estimates of recurrence intervals for these large (M>7) earthquakes remain disputed [Wyss and Koyanagi, 1992]. Hazard potential for the south flank includes catastrophic coastal mass movements and submarine slumping offshore [Lipman et al., 1985; Moore et al., 1995; Ma et al., 1999]. Tsunamis from the 1975 Kalapana and 1868 Great Kau earthquake reached maximum heights of ~15 m and ~14 m respectively on the south flank coast. Large (M>7) south flank earthquakes pose a significant hazard potential to the immediate Hawaiian Islands. Tsunamis generated in catastrophic south flank failure events might threaten cities around the Pacific Rim.


This research focuses on evaluating past faulting behavior of the Hilina fault system associated with large (M>7) south flank earthquakes. Records of historical seismicity on Hawaii date back to 1823 [Wyss and Koyanagi, 1992]. Evaluating the faulting behavior prior to historic accounts requires geologic field observations of surface deformation caused by large (M>7) earthquakes, specifically fault offset measurements of prehistoric lava flows. Triggered slip along the normal faults of the Hilina fault system, recorded in prehistoric lava flows, is the only accessible source of geologic information about prehistoric large (M>7) earthquakes. First, we document ground fractures from the 1975 Kalapana earthquake (hereafter referred to as the Kalapana earthquake), the only large (M>7) south flank earthquake with available geodetic measurements of displacement, seismicity, and fault offset measurement of the Hilina fault system.


29



30

Second, we determine fault offsets accumulated over several hundreds of years to calculate time-averaged slip rates across the Hilina fault system. Comparison of the Hilina fault offsets associated with the Kalapana earthquake to the Hilina fault offsets in prehistoric dated lava flows may reveal the number of large (M>7) earthquakes to affect the south flank. South flank seismic hazard assessment and kinematic models will benefit from our improved understanding of the past faulting behavior of the Hilina fault system.


South Flank Tectonic Setting

Seismic and geodetic data of the Kalapana earthquake, distribution of microseismicity, and geological features help define the structural geometry of the south flank. The wedge-shaped mobile south flank block is bounded at ~8-10-km-depth by a shallowly landward-dipping basal detachment fault, illuminated by microseismicity beneath the upper south flank [Ando, 1979; Got et al., 1994; Gillard et al., 1996], and imaged offshore at shallower depths in seismic reflection data [Morgan et al., 1999]. This weak boundary layer probably represents ocean sediments deposited on the Cretaceous ocean floor [Hill, 1969; Thurber and Gripp, 1988]. Focal mechanism studies from the Kalapana earthquake mainshock and aftershock sequence indicate south flank seaward transport along the basal detachment fault [Ando, 1979; Furumoto and Kovach, 1979; Gillard et al., 1996]. Bounded on the north by Kilauea caldera and Southwest and East rift zones, the south flank experiences seaward motion to the southeast. South flank lateral boundaries are not as well defined as the northern rift zone boundary. To the west, the Papa'u Seamount submarine ridge (Figure 3.2) exists but the lineation does not continue onto land in the form of a right-lateral shear zone as would be expected for the western boundary [Chadwick et al., 1993; Denlinger and Okubo, 1995; Morgan et al., 1999]. In the east, low electrical


31



32

resistivity lineations interpreted as conductive zones above dikes may indicate lateral boundaries of large-scale slump structures [Flanigan and Long, 1987]. The south flank southern boundary exists 40-50 km offshore, at the distal toe of chaotic debris from submarine landslides [Lipman et al., 1985; Moore et al., 1995; Morgan et al., 1999].


A combination of rift zone dike intrusions [Fiske and Jackson, 1972; Swanson et al., 1976; Denlinger and Okubo, 1995] and gravitational spreading [Dieterich, 1988; Delaney et al., 1998] probably displaces the south flank to the southeast. A variety of geodetic methods, including triangulation, trilateration, leveling, and Global Positioning System (GPS), have detected overall seaward south flank motion since 1896 [Swanson et al., 1976; Lipman et al. 1985; Owen et al., 1995; Delaney et al., 1998]. Geodetic station Panau ("P" in Figure 3.1) has moved seaward at least 10 m horizontally since 1896, with approximately 3.4 m horizontal displacement from the Kalapana earthquake and the remaining amount by aseismic basal detachment slip [Swanson et al., 1976]. However the Panau-Laeapuki baseline shortened ~1 m from 1898 to 1970, indicating the coastal south flank was under compression prior to the Kalapana earthquake [Swanson et al., 1976]. After the 1975 Kalapana earthquake [Swanson et al., 1976; Lipman et al., 1985], the south flank experienced phases of contraction and extension, with continued extension since the January 1983 eruption of Puu Oo [Delaney et al., 1998]. Surface displacement velocities from 1990 onward indicate a maximum of 10 cm/yr motion for the south flank [Owen et al., 1995; Owen et al., 1999]. The observed displacement rates can be modeled as 15-25 cm/yr slip on a 9-km-deep horizontal detachment [Owen et al., 1995]. Geodetic data suggest the Hilina faults have not significantly displaced except in the Kalapana earthquake [Delaney et al., 1998].


33

The wedge-shaped mobile south flank is crosscut by east-west trending Koae and Hilina normal fault systems (Figure 3.1). The Koae faults dip to the north, and show 1-2 m opening with less than 20 m vertical offset [Duffield, 1975]. The Hilina morphological fault scarps dip seaward, with up to 500 m vertical offset. Hilina faults may be deep-rooted, splaying off the basal detachment [Lipman et al., 1985]. Seismic tomography studies from Okubo et al. [1997] show a steeply southeast-dipping lateral gradient in P-wave velocity down to 10 km depth beneath the Hilina fault system. This gradient is interpreted as the Hilina fault boundary between high velocity rocks of the volcanic edifice and low velocity rocks of the Hilina fault system.


Alternatively, the Hilina faults are interpreted as shallow normal faults [Swanson et al., 1976; Ando, 1979; Hill and Zucca, 1987; Gillard et al., 1996]. Riley [1996] used paleomagnetic measurements from landward-rotated lava flows within the Puu Kapukapu fault block to model fault slip on a cylindrical fault surface with depth of 5.2 km. Fault offset measurements and geodetic data suggest the Hilina faults are a set of shallow normal [Cannon and Bürgmann, 1999]. Morgan et al. [1999] interpret offshore seismic reflection data near Halape (Figure 3.4) as related to a shallow fault or slump structure.


The Kalapana and Great Kau Earthquakes

The M=7.2 Kalapana earthquake is the largest earthquake to strike the south flank in the twentieth century [Tilling et al., 1975; Stover and Coffman, 1993]. Mainshock focal mechanisms indicate seaward transport of the south flank [Ando, 1979; Furumoto and Kovach, 1979] along the basal detachment. Over 25 km of surface rupture occurred on the Hilina faults, with a maximum of 1.5 m vertical offset (Figure 3.1) [Tilling et al., 1975]. An estimated 1000 square


34

kilometers of fault surface area ruptured on the basal detachment [Lipman et al., 1985]. Coastal geodetic stations displayed up to 8 m horizontal motion and 3.5 m vertical subsidence [Tilling et al., 1975; Lipman et al., 1985]. Approximately 60 km of coastline from Punaluu to Kaimu (Figure 3.1) subsided due to the Kalapana earthquake [Lipman et al., 1985].


The 1868 M=7.9 Great Kau earthquake (hereafter referred to as the Great Kau earthquake) is one of the fifteen largest earthquakes recorded in the United States [Stover and Coffman, 1993], and the largest historical earthquake to occur in Hawaii (Figure 3.1). The earthquake sequence included a M~6.7 foreshock with on 28 March 1868, the M=7.9 mainshock on 2 April 1868, and possibly a decade of aftershock activity [Wyss, 1988]. A ~14 m high tsunami inundated coastal villages along the south flank coast [Tilling et al., 1975; Wyss, 1988; Wyss and Koyanagi, 1992]. Brigham [1909] reported 1.2 - 2.1 m of permanent coastal subsidence along ~80 km of coastline from Punaluu to Kapoho (Figure 3.1). Wyss [1988] estimated 8m of horizontal slip over a basal detachment area of 4000 square kilometers bounded by Mauna Loa rift zones.


Methods of Determining Fault Offsets and Time-Averaged Offset Rates

To improve our understanding of faulting behavior of the Hilina fault system prior to the Kalapana and Great Kau earthquakes, we measure Hilina fault offsets in prehistoric lava flows. Fault offsets are measured from piercing points that indicate plunge, azimuth, and total length of offset. Total offset measurements are decomposed into horizontal and vertical components. We utilize GPS-RTK (Real-Time Kinematic) and measuring tape and compass techniques to measure fault offset. With GPS-RTK equipment, we measure the absolute locations of each end of the piercing point and then calculate the


35

piercing point attitude. The GPS-RTK equipment yields 1-2 cm horizontal accuracy and 2-4 cm vertical accuracy for absolute positions. We use the measuring tape and compass technique for fault offsets in rugged terrain or remote field locations where GPS-RTK equipment is difficult to operate. A compass is used to obtain plunge and azimuth of the piercing point, and total length is measured with a tape measure. A handheld GPS receiver records the location of the piercing point. Post-processed differentially-corrected GPS fault offset positions from the handheld GPS equipment have 2-5 m horizontal and 4-10 m vertical absolute position accuracy.


Fault offset measurements across a fault zone are collected in traverses conducted perpendicular to main fault traces of the Hilina faults (Figure 3.3d). For each traverse, individual fault offsets are summed to calculate total fault offset for each traverse. For example, individual horizontal measurements of fault offset "a" through "k" (11 measurements) are summed to produce horizontal fault offset T53. 73 traverses contain over 700 measurements, an average of 10 measurements per traverse. Using the "measuring tape and compass" technique, measurement error for fault offset components is estimated at ±1 cm. GPS-RTK equipment yields an estimated ±1 cm horizontal error, and ±2 cm vertical error for each individual measurement of fault offset. To obtain error bounds for fault offset traverses, we utilized the square root (sum of the (measurement errors)^2) convention.


Time-averaged calculations for fault offset rates require offset measurements recorded in dated lava flows. We use ages for south flank lava flows given in radiocarbon year ages (Figure 3.4) Wolfe and Morris [1996] utilized several methods to date lava flows, including observed stratigraphic


36



37



38

relationships, alteration of lava flow surfaces from an initially glassy surface, radiocarbon ages from charcoal samples [Lipman and Lockwood, 1980], and paleomagnetic secular variation [Holcomb, 1987]. Radiocarbon years are defined as "years before present relative to 1950", represented by the symbol "yr B.P.". Add 25 years to "calibrate" the lava flow age relative to the Kalapana earthquake; add 50 years to "calibrate" the lava flow age relative to the year 2000. Ages of south flank lava flows include 200-750 yr B.P., 400-750 yr B.P., 750-1500 yr B.P., and 1500-3000 yr B.P. from Wolfe and Morris [1996]. A flow of age 330-470 yr B.P. was determined using radiocarbon dating methods [D. A. Swanson, 1999, personal communication]. Given a lava flow of known age, a minimum time-averaged offset rate is calculated by dividing the fault offset for a given traverse by known age of the lava flow containing the traverse. This time-averaged calculation is a theoretical minimum fault offset rate because fault offsets probably occurred sometime after the lava flow cooled. By studying fault offsets in lava flows of up to 1500-3000 yr B.P. age, we can evaluate past faulting behavior of the Hilina fault system.


Kalapana Earthquake Ground Fracture and Fault Offset Data

The first step in understanding past faulting behavior of the Hilina fault system requires study of the ground fracture and fault offsets from the Kalapana earthquake. Utilizing GPS-RTK equipment, we present the first ground fracture map for the Kalapana earthquake (Figure 3.3b-e). Mauna Ulu lava flows (1969-1974) covered the central south flank region in pahoehoe and aa flows. These flows were consequently fractured along the Hilina faults during the Kalapana earthquake [Cannon and Bürgmann, 1999]. Fracture of Mauna Ulu lava flows display a complex pattern of fractures on Hilina faults. Single fractures are up to 170 m long (Figure 3.3b) and fracture zones extend up to 200 m in width.


39

Fault ramps and steps produce fractures distributed throughout the fracture zone, not necessarily concentrated at the northern or southern boundaries of the fracture zone (Figure 3.3b-e).


To determine fault offset of the Hilina fault system associated with the Kalapana earthquake, we present 43 fault traverses in Mauna Ulu lava flows (solid vectors in Figure 3.5). Maximum horizontal and vertical fault offset occur on the Holei Pali, with 2.8 m of horizontal offset and 1.6 m of vertical offset determined from summing 16 individual fault offset measurements. We define an "offset ratio" as horizontal fault offset divided by vertical fault offset for an "offset pair". An "offset pair" is defined as a location where both horizontal and vertical fault offsets are measured. 39 of 43 measurement locations have horizontal fault offset exceeding vertical fault offset, with offset ratios of horizontal to vertical fault offsets ranging from 0.2 to 31.5 (Figure 3.5).


Fault Offsets in Prehistoric Lava Flows

Limited by historic earthquake records, we measure the Hilina fault offsets in prehistoric south flank lava flows (open vectors in Figure 3.5) to extend our knowledge about south flank earthquakes as far back as 1500-3000 yr B.P. Quality of fault offset preservation decreases with increased flow age. Lava flows with ages greater than 200-750 yr B.P. have degraded fault fracture surfaces that make fault offset measurements difficult or impossible to obtain. Often with flows greater than 200-750 yr B.P. age, only vertical fault offset measurements can be estimated with large offset errors.


Fault offsets in prehistoric lava flows offer crucial information toward understanding past slip on the Hilina fault system. 45 locations with fault offset


40



41

in prehistoric lava flows are summarized in Figure 3.5 (open vectors). Horizontal fault offsets range up to 5 m and vertical fault offsets up to 7-8 m. The maximum vertical fault offset of ~10±2 m occurs in 750-1500 yr B.P. age lava flows, a rough quantitative estimate due to the poor preservation of fault offsets in older prehistoric lava flows. We only evaluate locations with offset for lava flows of less than 750-1500 yr B.P. age, as these lava flows have the best fault offset preservation. 31 of 45 locations are located in prehistoric lava flows of less than 750-1500 yr B.P. age, and 23 of these 31 locations contain offset pairs. 2 offset pairs are from lava flows of 750-1500 yr B.P. age. For lava flows of 200-750 yr B.P. age, 14 of the 21 offset pairs have horizontal fault offsets greater than vertical fault offsets (Figure 3.5). Offset ratios of horizontal to vertical fault offset range from 0.2 to 680 for prehistoric flows.


To estimate the number of large (M>7) earthquakes that have occurred on Hilina faults, we collect offset data at locations with two important requirements. First, prehistoric lava flows and Mauna Ulu flows must be juxtaposed flows. Second, fault offsets in prehistoric lava flows must be well-preserved for measurement of piercing points. The Kealakomo Overlook region meets both of these requirements (Figure 3.6). The Mauna Ulu lava flows are bound to the east and west by 400-750 yr B.P. age lava flows. Traverses trend both the Poliokeawe Pali and Holei Pali, with traverses #1 and #4 in the prehistoric lava flows, and traverses #2 and #3 in the Mauna Ulu lava flows. Fault offsets in the Mauna Ulu lava flows display 1-2 m horizontal offset and 0.2-0.5 m vertical offset. Fault offsets in the prehistoric lava flows show 2-3 m horizontal fault offset and 3-4 m vertical fault offset.


At a few sites in the Hilina fault system, a prehistoric lava flow that drapes a fault has been offset numerous times by past earthquakes. Some prehistoric


42



43

fractures were filled by 1969-1974 Mauna Ulu lava flows. The Kalapana earthquake then produced fault offset of the prehistoric fractures and created new fractures in the Mauna Ulu lava flows. This scenario is referred to as a double-fracture outcrop (Table 3.1). Three south flank field sites reveal double-fracture outcrops (lettered locations in Figure 3.4), and present the opportunity to compare Kalapana earthquake and prior fault offsets at the outcrop scale. Fractures in the Mauna Ulu lava flows display the Kalapana earthquake fault offset only, and the prehistoric fractures contains total fault offset (Kalapana earthquake and previous earthquakes) since the cooling of the lava flow. The number of large (M>7) earthquakes to produce total fault offset of the prehistoric lavas is estimated by dividing the total fault offset in the prehistoric lava flows by the total fault offset in the Mauna Ulu lava flows. Total fault offsets in Mauna Ulu flows vary from 5-67 cm, while total fault offsets in prehistoric lava flows vary from 80-89 cm (Table 3.1).


Time-averaged fault slip rates for horizontal and vertical offsets in prehistoric flows are presented in Figure 3.7. Fault offset data from prehistoric lava flows (Figure 3.5) divided by age of lava flow produces a fault offset rate for each offset measurement location. Solid vectors indicate locations with offset pairs, whereas open vectors indicate locations lacking horizontal or vertical fault offset measurements. This figure simplifies south flank faults and lava flow geometry, preserving fault-lava flow cross-cutting relationships, fault-fault and lava flow-lava flow spatial relationships. Considering fault offsets in lava flows for ages back to 1500-3000 yr B.P., maximum horizontal and vertical fault offset rates equal ~1 cm/yr and ~2 cm/yr respectively.


44



45



46



47

Plotting fault offset rates locations with offset pairs only (Figure 3.8), 19 of 21 paired offsets occur in lava flow of less than 200-750 yr B.P. age (solid symbols). Two paired offsets occur in 750-1500 yr B.P. age lava flows, shown with by open squares. For horizontal fault offset rates, slip on individual Hilina faults varies from 0.03 to 1.1 cm/yr. For vertical fault offset rates, slip rates on individual Hilina faults ranges from -0.01 cm/yr to -1.4 cm/yr with negative values indicating hangingwall down fault offset over time with respect to the footwall. Fault offset rate values for 750-1500 yr B.P. age offsets fall within the range of the upper and lower bounds for horizontal and vertical fault offset rates in lava flows of 200-750 yr B.P age.


We also consider time-averaged fault offset rates across the entire Hilina fault system. Fault offsets are summed along prehistoric lava flows of the same age (Figure 3.9a) to produce traverses of total fault offset measurements across the Hilina fault system (Figure 3.9b). Solid vectors in Figure 3.9b indicate horizontal fault offset and open vectors represent vertical fault offset across the Hilina fault system. Traverses #5 and #6 occur in lava flows of 1500-3000 yr B.P. age and 750-1500 yr B.P. age respectively, and traverses #7-#11 occur in lava flows of 200-750 yr B.P. age. For traverses #7-#11, vertical fault offsets exceed horizontal fault offsets (#7, #8, #10), whereas for traverses #9 and #11 have horizontal fault offsets greater than vertical fault offsets. For south flank lava flows of less than 200-750 yr B.P. age, the maximum horizontal fault offset is 7±2 m and the maximum vertical fault offset is 10±2 m.


Prehistoric time-averaged total fault offset rates across the Hilina fault system are presented in Figure 3.10. These fault offset rates are calculated by dividing the total fault offset across the Hilina fault system (Figure 3.9b) by the


48



49



50

age of the lava flow. For lava flows of 200-750 yr B.P. age, fault offset rates across the Hilina fault system show ranges of 0.4 to1.2 cm/yr and -0.2 to -2.0 cm/yr for horizontal and vertical fault offset rates respectively. Horizontal fault offset rates for 200-750 yr B.P. age lava flows are greater than horizontal offset rates from fault offsets in 750-1500 yr B.P. age lava flows. Vertical fault offset rates for 750-1500 yr B.P. age lava flows fall within the range of -0.2 to -2.0 cm/yr values. With a completed fracture map and fault offset measurements of Mauna Ulu and prehistoric lava flows, we now evaluate the past faulting behavior of the Hilina fault system.


Discussion

We focus our efforts on quantifying fault offsets of the Hilina fault system for the Kalapana earthquake and prehistoric lava flows. With these data compiled and presented, we now characterize the past faulting behavior of the Hilina faults by: (1) interpreting ground fractures from the Kalapana earthquake; (2) comparing horizontal and vertical offsets associated with the Kalapana earthquake and prehistoric offsets; (3) establishing past offset rates of individual Hilina faults; (4) determining past total offset rate across the Hilina fault system; (5) establishing an initial estimate of the number of large (M>7) earthquakes to have occurred in the last ~3000 years; and (6) discussing our assumption of a characteristic Kalapana-type earthquake.


We interpret the formation of fractures in Mauna Ulu flows associated with the Kalapana earthquake as tectonic in origin. Compressional fold and fault structures are not observed at the base of fault scarps, indicating mass movements did not produce the fractures. Where fractures are wide enough to climb down, we examine fracture walls for piercing points and subsurface


51

detachment surfaces parallel to the base of flows. No subsurface detachment surfaces are observed on fracture walls, suggesting the upper few meters of lava flow did not detach from the subsurface to produce fractures observed at the surface.


Kalapana earthquake and prehistoric horizontal fault offsets display similar characteristics and relationships with respect to vertical offsets. Horizontal offsets in Mauna Ulu flows (solid vectors in Figure 3.5) and displacements from geodetic measurements across the south flank (solid vectors in Figure 3.1) show motion toward the southeast, in agreement with the overall southeast direction of south flank transport {Swanson et al., 1976; Tilling et al., 1975; Lipman et al., 1985]. Horizontal fault offsets for both Mauna Ulu and prehistoric lava flows (open vectors in Figure 3.1) generally trend southeast, but the local trend of the fault trace seems to constrain fault offsets to be in a perpendicular direction [Cannon and Bürgmann, 1999]. Considering offset pairs of fault offsets in Mauna Ulu lava flows, 39 of 43 offset pairs have horizontal measurements exceeding vertical measurements with ratio values of 0.2 to 31.5 (Figure 3.5). Offset ratios greater than one for a simple normal fault model indicate the fault surface dips less than 45°. 14 of 21 offset pairs in prehistoric lava flows of less than 200-700 yr B.P. age also show horizontal fault offset greater than vertical fault offset (offset ratios range from 0.2 to 680). Offset ratios for the Kalapana earthquake and fault offsets from prehistoric flows suggest the fault surfaces of the Hilina fault system have an overall dip of less than 45°.


For lava flows of the time period back to 200-750 yr B.P., we determine time-averaged fault offset rates for horizontal and vertical fault offsets. For individual fault offsets of this time period, horizontal fault offset rates range


52

from 0.03 to 1.1 cm/yr, and vertical fault offset rates span the values of -0.1 to -1.4 cm/yr. Considering fault offset across the entire Hilina fault system, we calculate a horizontal and vertical fault offset rate of 0.4 to 1.2 cm/yr and -0.2 to -2.0 cm/yr respectively. We expect horizontal fault offset rates to be greater than vertical fault offset rates, since prehistoric horizontal fault offsets are greater than vertical fault offsets for 14 of 21 pairs for the time period back to 200-750 yr B.P. Significant errors in lava flow ages might overprint any relationship between horizontal and vertical fault offsets. Individual faults do not show a characteristic fault offset rate. Since fault slip varies along the fault, calculated fault rates would also vary with the maximum fault offset rates for the maximum fault offset.


We estimate the number of large (M<7) earthquakes that might have occurred, and recurrence intervals, for the Poliokeawe Pali and Holei Pali at Kealakomo Overlook (Figure 3.6). Comparing the Kalapana earthquake fault offset traverse #2 to 400-750 yr B.P. age fault offset traverse #1 in the western region, 3-5 Kalapana-type earthquakes could have produced fault offsets observed in the 400-750 yr B.P. age lava flows. The time-averaged recurrence interval for Kalapana-type earthquakes on fault offset traverses #1 and #2 is 260-80 years. In the eastern region, horizontal offsets are about equal for traverses #3 and #4, fault offsets from the Kalapana earthquake and prehistoric offsets respectively. However, vertical fault offset in traverse #4 may be explained by as many as seven Kalapana-type fault offsets, measured in traverse #3. Large error bars on fault offset traverse #4 result from poor scarp preservation. 3-5 Kalapana-type earthquakes can explain offsets in prehistoric flows observed in traverse #2, but as few as one and as many as seven Kalapana-type earthquakes may have produced offsets in traverse #4. We place more confidence in


53

interpretations of the number of Kalapana-type earthquakes and recurrence intervals calculated from traverses #1 and #2, than traverses #3 and #4, due to poor offsets located along traverse #4.


Interpreting fault offsets at double-fracture outcrops also provides estimates for the number of past large (M>7) earthquakes and recurrence intervals (Table 3.1). Assuming multiple Kalapana-type earthquakes produced the total fault offset observed in the prehistoric lava flow, 1.2-17.5 Kalapana-type earthquakes with time-averaged recurrence intervals of 650 to 20 years may have occurred on the Hilina faults. The wide range of the number of estimated Kalapana-type earthquakes may be due to unique fault offsets on each fault for each event, rather than a uniform fault offset for each event. The large variance in time-averaged recurrence intervals suggests that large (M>7) south flank earthquakes may not produce a uniform fault offset on Hilina faults with each event.


Our recurrence interval estimates from double-fracture outcrops and Kealakomo Overlook agree with values determined from previous workers. Estimating a recurrence interval for large (M>7) south flank earthquakes remains difficult since only the Great Kau and Kalapana earthquakes have been documented. From the time interval between the 1868 Great Kau and 1975 Kalapana earthquakes, the recurrence interval for a M=7-8 earthquake in the Kalapana area is 108 years [Wyss and Koyanagi, 1992]. Lipman et al. [1985] concluded approximately 1000 Kalapana earthquake subsidence events may be responsible for Hilina fault morphology. Given 23 ky as the minimum age estimate for Hilina faults [Easton, 1978 - age of Pahala ash cap on Puu Kapukapu], a 230 year recurrence interval is calculated. Riley [1996] determined


54

from paleomagnetic rotations of Puu Kapukapu block a 200 year recurrence interval. Caution must be exercised with these recurrence interval estimates of large (M>7) earthquakes in a volcanic system. Time-averaged recurrence intervals are based on Hilina fault offset due to the Kalapana earthquake, not from the absolute ages of multiple earthquake events. Volcanic processes and changes through time, such as changes in magma transport, may play an important role in developing and triggering a large (M>7) earthquake, and most likely do not have uniform recurrence intervals.


Our estimates of the number of past large (M>7) south flank earthquake and recurrence interval calculations depend on the main assumption that all past large (M>7) earthquakes produced fault offsets comparable to fault offsets measured in the Hilina fault system for the Kalapana earthquake. The wide range of recurrence interval estimates and number of large (M>7) earthquakes suggest slip on Hilina fault systems is not uniform and periodic for past earthquakes. For the Poliokeawe Pali, fault offsets at Kealakomo Overlook and the Poliokeawe double-crack outcrop might be explained by as few as two earthquake events, the Kalapana earthquake, and the Great Kau earthquake. We speculate that the larger magnitude Great Kau earthquake may have produced more slip on the Hilina fault system than occurred with the Kalapana earthquake. Fault offsets of 400-750 yr B.P. age lava flows on the Poliokeawe Pali might have been produced by only the Great Kau and Kalapana earthquakes, again suggesting that large (M>7) south flank earthquakes do not have uniform repeat times. Geodetic monitoring of strain accumulation across the south flank is more appropriate method for evaluating large (M>7) earthquake hazards than wide variances in recurrence intervals based on Kalapana-type offsets of prehistoric lava flows.


55

Conclusions

Detailed mapping of the Kalapana earthquake fractures and the fault offsets in prehistoric flows allow us to characterize the past faulting behavior of the Hilina fault system. Accounts of historical earthquakes on Hawaii only date back to 1823, but we extend the earthquake record back as far back as 200-750 yr B.P. by comparing fault offsets in prehistoric lava flows with neighboring Kalapana earthquake fault offsets. Our first assumption is that the Hilina faults only slip due to large (M>7) earthquakes, since only the 1868 M=7.9 Great Kau and 1975 M=7.2 Kalapana earthquakes have produced surface ruptures of the Hilina fault system. Our second assumption is that each prehistoric faulting event has "Kalapana-type" fault offsets equal to measured Kalapana earthquake fault offsets on the Hilina faults. The Kalapana earthquake is the only earthquake to offset the Hilina fault system and have available geodetic measurements of displacement, seismicity records, and fault offset measurements.


We draw several conclusions about past faulting behavior of the Hilina fault system from our fault offset analysis. We focus our conclusions on prehistoric lava flows as far back as 200-750 yr B.P. age. Horizontal fault offset across the individual Hilina faults is greater than vertical fault offset for 39 of 43 offset pairs associated with the Kalapana earthquake, and 14 of 21 offsets pairs in prehistoric lava flows. Horizontal fault offset is an important mechanism for slip on the Hilina fault system. Fault offset rates for individual Hilina faults are 0.03 to 1.1 cm/yr and -0.01 to -1.4 cm/yr for horizontal and vertical rates as far back as 200-750 yr B.P. Horizontal and vertical fault offset rates across the entire


56

Hilina fault system are 0.4 to 1.2 cm/yr and -0.2 to -2.0 cm/yr. The number of large (M>7) earthquakes to occur as far back as 200-750 yr B.P. ranges from 1.2 to 17.6, with recurrence intervals of 650-20 years, and our best estimate suggests 3-5 events have occurred with recurrence intervals of 260-80 years in the time period back to 400-750 yr B.P. The wide range of estimates on the number of prehistoric faulting events suggests that these events do not necessarily produce Kalapana-type offsets. Fault offsets of prehistoric earthquakes probably are not uniform for each event, and these earthquakes do not have uniform recurrence intervals. Future seismic hazard assessment for the south flank must be based on geodetic observations of strain accumulation, not on the wide range of recurrence interval estimates determined from comparison of Kalapana earthquake and prehistoric fault offsets.


Acknowledgements

This work was supported by a USGS NEHRP grant, GSA student grant (E.C.C.) and Cordell Durrell grant (E.C.C.). Jim McClain and Rob Twiss provided comments to significantly improve this mamanuscript. We received help from many people at Hawaiian Volcano Observatory, especially Don Swanson, Mike Lisowski, Asta Miklus, Kristine Larson, Taeko Jane Takahashi, Arnold Okamura, Richard Fiske, and Paul Okubo. Five excellent field assistants included Mike Poland, Gwen Pikkarainen, Chad Fleschner, Jason Bariel, and Jim Weigel. Many figures were produced with the Generic Mapping Tools program [Wessel and Smith, 1995].


57

References Cited

Ando, M., The Hawaii Earthquake of November 29, 1975: Low Dip Angle Faulting Due to Forceful Injection of Magma, J. Geophys. Res., 94, no. B13, 7616-7626, 1979.

Brigham, W. T., The volcanoes of Kilauea and Mauna Loa on the island of Hawaii, Memories of the Bernice Pauahi Bishop Museum, 2, 478-497, 1909.

Cannon, E. C. and R. Bürgmann, Normal faulting and block rotation associated with the Kalapana earthquake, Kilauea Volcano, Hawaii, manuscript in preparation for submission to Bull. Seismol. Soc. Amer., 1999.

Delaney, P. T., R. Denlinger, M. Lisowski, A. Miklius, P. Okubo, A. Okamura, and M. K. Sako, Volcanic spreading at Kilauea, 1976-1996, J. Geophys. Res., 103, no. B8, 18003-18023, 1998.

Denlinger, R. P. and P. Okubo, Structure of the mobile south flank of Kilauea, Volcano, Hawaii, J. Geophys. Res., 100, no. B12, 24499-24507, 1995.

Dieterich, J. H., Growth and persistence of Hawaiian volcanic rift zones, J. Geophys. Res., 93, 4258-4270, 1988.

Duffield, W. A., Structure and origin of the Koae fault system, Kilauea Volcano, Hawaii, U.S. Geol. Surv. Prof. Pap. 856, 1975.

Easton, R. M., The stratigraphy and petrology of the Hilina Formation; the oldest exposed lavas of Kilauea Volcano, Hawaii, M.S. thesis, Univ. of Hawaii, 1978.

Fiske, R. S. and E. D. Jackson, 1972, Orientation and growth of Hawaiian volcanic rifts: the effect of regional structure and gravitational stresses, Royal Soc. London Proc. 329, 299-326, 1972.

Flanigan, V. J. and C. L. Long, Aeromagnetic and near-surface electrical expression of the Kilauea and Mauna Loa volcanic rift systems, U.S. Geol. Surv. Prof. Pap. 1350, 935-946, 1987.

Furumoto, A. S., and R. L. Kovach, The Kalapana Earthquake of November 28, 1975: An Intra-plate Earthquake and its Relation to Geothermal Processes, Phys. Ear. Plan. Int., 18, 197-208, 1979.

Gillard, D., M. Wyss, and J. S. Nakata, A seismotectonic model for western Hawaii based on stress tensor inversion from fault plane solutions, J. Geophys. Res., 97, no. B5, 6629-6641, 1992.

Gillard, D., M. Wyss, and P. Okubo, Type of faulting and orientation of stress and strain as a function of space and time in Kilauea's south flank, Hawaii, J. Geophys. Res., 101, no. B7, 16025-16042, 1996.


58

Got, J.-L., F. W. Fréchet, and W. Klein, Deep fault plane geometry inferred from multiplet relative relocation beneath the south flank of Kilauea, J. Geophys. Res., 99, no. B8, 15375-15386, 1994.

Hill, D. P, Crustal structure of the island of Hawaii from seismic-refraction measurements, Bull. Seismol. Soc. Amer., 59, no. 1, 101-130, 1969.

Hill, D. P. and J. J. Zucca, Geophysical constraints on the structure of Kilauea and Mauna Loa Volcanoes and some implications for seismomagnetic processes, U.S. Geol. Surv. Prof. Pap. 1350, 903-917, 1987.

Holcomb, R. T., Eruptive history and long-term behavior of Kilauea Volcano, U.S. Geol. Surv. Prof. Pap. 1350, 261-350, 1987.

Lipman, P. W., J. P. Lockwood, R. T. Okamura, D. A. Swanson, and K. M. Yamashita, Ground Deformation Associated with the 1975 Magnitude-7.2 Earthquake and Resulting Changes in Activity of Kilauea Volcano, Hawaii, U.S. Geol. Surv. Profess. Pap., 1276, 1985.

Lockwood, J. P. and P. W. Lipman, Recovery of Datable Charcoal beneath Young Lavas: Lessons from Hawaii, Bull. Volcanol. 43, no. 3, 609-615, 1980.

Ma, K-F., H. Kanamori, and K. Satake, Mechanism of the 1975 Kalapana, Hawaii, earthquake inferred from tsunami data, J. Geophys. Res, 104, no. B6, 13153-13167, 1999.

Moore, J. G., W. B. Bryan, M. H. Beeson, and W. R. Normark, Giant blocks in the South Kona landslide, Hawaii, Geology, 23, 125-128, 1995.

Morgan, J. K., G. F. Moore, D. J. Hill, and S. Leslie, Confirmation of Volcanic Spreading Models for Kilauea's Mobile South Flank, Hawaii, from Marine Seismic Reflection Data, Geology, in review 1999.

Okubo, P. G., H. M. Benz, and B. A. Chouet, Imaging the crustal magma sources beneath Mauna Loa and Kilauea volcanoes, Hawaii, Geology, 25, no. 10, 867-870, 1997.

Owen, S., P. Segall, J. Freymueller, A. Miklius, R. Denlinger, T. Árnadóttir, M. Sako, and R. Bürgmann, Rapid Deformation of the South Flank of Kilauea Volcano, Hawaii, Science, 267, 1328-1332, 1995.

Owen, S., P. Segall, M. Lisowski, A. Miklius, R. Denlinger, J. Freymueller, T. Árnadóttir, and M. Sako, The Rapid Deformation of Kilauea Volcano: GPS measurements between 1990 and 1996, J. Geophys. Res., in press 1999.

Riley, C., A Paleomagnetic Study of Movement in the Hilina Fault System, South Flank of Kilauea Volcano, Hawaii, M.S. thesis, Mich. Tech. Univ., 1996.


59

Stover, C. W. and J. L. Coffman, 1993, Seismicity of the United States, 1568-1989 (Revised), U.S. Geol. Sur. Prof. Pap. 1527, 1993.

Swanson, D. A., W. A. Duffield, and R. S. Fiske, Displacement of the South Flank of Kilauea Volcano: The Result of Forceful Intrusion of Magma Into the Rift Zones, U.S. Geol. Surv. Prof. Pap. 963, 1974.

Thurber, C. H. and A. E. Gripp, Flexure and seismicity beneath the south flank of Kilauea volcano and tectonic implications, J. Geophys, Res., 93, 4271-4281, 1988.

Tilling, R. I., R. Y. Koyanagi, P. W. Lipman, J. P. Lockwood, J. G. Moore, and D. A. Swanson, Earthquake and Related Catastrophic Events Island of Hawaii, November 29, 1975: A preliminary report, U. S. Geol. Surv. Circ. 740, 1976.

Wessel, P. and W. H. F. Smith, New version of the Generic Mapping Tools released, EOS Trans, AGU, 76, 329, 1995.

Wolfe, E. W. and J. Morris, Geologic Map Of The Island Of Hawaii, U.S. Geol. Surv. Misc Invest. Ser. Map I-2524-A, 1996.

Wyss, M., A Proposed Source Model For The Great Kau, Hawaii, Earthquake of 1868, Bull. Seismol. Soc. Amer, 78, no. 4, 1450-1462, 1988.

Wyss, M. and R. Y. Koyanagi, Seismic Gaps in Hawaii, Bull. Seismol. Soc. Amer., 82, no. 3, 1373-1387, 1992.


60

Chapter 4

Conclusion

Significant results of this study improve our understanding of the kinematics of the Hilina fault system. First, a detailed fracture map of the Kalapana earthquake now exists (Figure 2.2), containing over 18,000 m of fracture and 200 offset measurements. The fault offset data set includes an additional 700 offset measurements from fault offsets in Mauna Ulu and prehistoric lava flows throughout the Hilina fault system. Second, analysis of fault offset, and observed and model displacements suggests the Hilina normal faults are shallow normal faults, possibly 1-5-km deep, and displace only due to large (M>7) earthquakes (Figure 2.4). Third, time-averaged fault offset rates across the Hilina fault system for the time period back to 200-750 yr B.P. for horizontal and vertical offset rates are 0.4-1.2 cm/yr and -0.2 to -2.0 cm/yr respectively (Figure 3.10). Fourth, the estimated number of large (M>7) earthquakes to occur for the time period back to 200-750 yr B.P. varies from 1.2-17.5 events (Table 3.1). However, excellent fault offset preservation at the western Kealakomo Overlook traverses (Figure 3.6) indicate 3-5 events with a recurrence interval of 260-80 years for the time period back to 400-750 yr B.P. These variable recurrence interval estimates suggest past earthquakes that produced slip on the Hilina fault system did not necessarily produce magnitudes of slip associated with the Kalapana earthquake, nor did each event produce uniform offset. This research is a first step towards understanding the past faulting history of the Hilina fault system. Future improvements and research by others will undoubtedly provide more insight into the kinematics of the Hilina fault system.


61

Improvements in age dating of prehistoric lava flows will help constrain fault offset rates across the Hilina fault system and calculations of the number of large (M>7) earthquakes to rupture prehistoric lava flows of a particular age. Charcoal is extremely difficult to locate in the field, and conditions for charcoal preservation do not always persist (Lockwood and Lipman, 1980). Improvements in paleomagnetic dating of lava flows using secular variation curves (Hagstrum and Champion, 1995) offer refinements in lava flow age estimates. Large errors on prehistoric fault offset rates will be reduced with smaller errors on lava flow ages.


Future research should be focused on detecting subsurface fault surfaces and understanding the kinematics of the Hilina fault system. Potentially catastrophic mass movements of hangingwall blocks could occur due to large (M>7) south flank earthquakes. Seismic reflection and refraction experiments on land might be able to detect the fault surface of shallow normal faults or landward-dipping lava flow layers produced by landward block rotation. Thickening of lava flows proximal to the fault scarp can be interpreted as landward block rotation of hangingwall blocks, similar to sediment deposition in an asymmetric normal fault graben. Vibroseis equipment could be deployed along the Chain of Craters Road (Figure 4.1), resulting in a seismic survey strike-perpendicular to the Hilina fault system. Small-scale seismic surveys could be accomplished with a minimal setup employing a 24-channel seismometer, and an explosive or weight drop energy source.


Monitoring of fault slip within the Hilina fault system should be continued to evaluate slip produced by the next large (M>7) south flank earthquake Permanent monuments could be installed in linear geodetic arrays


62



63

strike-perpendicular to Hilina faults (Figure 4.1). 1 cm diameter x 10 cm length stainless steel slugs, permanently glued into lava flows with epoxy, spaced every 10-20 m, could be installed across Hilina faults. Present GPS-RTK technology and equipment would require an estimated one to two days to survey 5-10 fault traverses. Existing nail arrays along Chain of Craters could be resurveyed, and nails added to increase the monument density of the nail arrays. Densely-spaced geodetic arrays, combined with south flank geodetic and seismic networks, and synthetic-aperture radar analysis would help monitor the south flank for future large (M>7) earthquakes. Annual surveys of geodetic arrays in the Hilina fault system would be extremely useful in determining slip of the Hilina fault system from the next large (M>7) south flank earthquake.


Two locations in the western portion of the Hilina fault system offer challenging opportunities to study fault-fault interactions, and fault termination. A relay ramp probably separates Puueo Pali and Puu Kapukapu (Figure 4.2). Fractures in the upper part of the relay ramp generally trend northeast. Detailed offset measurements would be required to analyze the interaction of offset on the relay ramp with Puueo Pali and Puu Kapukapu. Crider and Pollard (1998) developed a numerical model to analyze relay ramp fractures in extensional systems. They show that en echelon fractures forming on the relay ramp are produced by dip-slip extension along the overall trend of the fault system, rather than strike-slip motion. The primitive Halape campground could be used as a base camp for fieldwork.


Another area of potential study is the western end of Holei Pali (Figure 4.3). The trace of the western Holei Pali fault intersects the coastline between Halape and Keahou Landing (Figure 4.3). The Holei Pali ground fractures trend


64



65



66

southwest, change direction to the west to form a semi-circular trace of about one kilometer diameter, then return to a southwest trend at the coast. 750-1500 yr B.P. age flows cover this region, with poor fault offset piercing points to the west. Fractures are well-displayed on 9" x 9" NASA color aerial photographs [Nasa/JPL Roll 03541 5000' scale, images 026 and 027; Nasa/JPL Roll 03541 13000' scale, images 0074 and 0076] and the Kau Desert 7.5-minute topographic map. Tent camping at Keahou Landing and Halape offer excellent base camps for fieldwork.


References Cited

Crider, J. G. and D. D. Pollard (1998). Fault linkage: Three-dimensional mechanical interaction between echelon normal faults, J. Geophys. Res., 103, no. B10, 24373-24391.

Hagstrum, J. T. and D. E. Champion (1995). Late Quaternary geomagnetic secular variation from historical and 14-C-dated lava flows on Hawaii, J. Geophys. Res., 100, no. B12, 24393-24403.

Lockwood, J. P. and P. W. Lipman (1980). Recovery of Datable Charcoal beneath Young Lavas: Lessons from Hawaii, Bull. Volcanol. 43, no. 3, 609-615.


67

Appendix A.

Fracture Maps and Fault Offset Measurements

Appendix A contains plots of fracture, width and piercing points collected from fault offsets on Poliokeawe Pali, Holei Pali, and Apua Pali. Over 18 000 m of fracture and 200 piercing points were measured from fractures in 1969-1974 Mauna Ulu pahoehoe lava flows. Figure A-1 shows field locations for Figures A-2 to A-14. Lava flow ages in this appendix are from Figure 3.4.

Figure A-1. Location map for fracture, width, and fault offset maps of 1975 Kalapana earthquake ground deformation

Figure A-2. Map of fractures in Mauna Ulu lava flows on Poliokeawe Pali

Figure A-3. Map of fracture width in Mauna Ulu lava flows on Poliokeawe Pali

Figure A-4. Map of horizontal fault offsets in Mauna Ulu lava flows on Poliokeawe Pali

Figure A-5. Map of vertical fault offsets in Mauna Ulu lava flows on Poliokeawe Pali

Figure A-6. Map of fractures in Mauna Ulu lava flows on Holei Pali

Figure A-7. Map of fracture width in Mauna Ulu lava flows on Holei Pali

Figure A-8. Map of horizontal fault offsets in Mauna Ulu lava flows on Holei Pali

Figure A-9. Map of vertical fault offsets in Mauna Ulu lava flows on Poliokeawe Pali

Figure A-10. Map of fractures in Mauna Ulu lava flows on Apua Pali

Figure A-11. Map of fracture width in Mauna Ulu lava flows on Apua Pali

Figure A-12. Map of horizontal fault offsets in Mauna Ulu lava flows on Apua Pali

Figure A-13. Map of vertical fault offsets in Mauna Ulu lava flows on Apua Pali

Figure A-14 Map of fractures in Mauna Ulu flows on Poliokeawe and Holei Pali


68



69



70



71



72



73



74



75



76



77



78



79



80



81