Tullis, Terry

Professor Emeritus of Geological Sciences, Professor of Earth, Environmental, and Planetary Sciences (Research)

Overview

I received my BA from Carleton College in 1964, then went on to get both my MS (1967) and PhD (1971) from UCLA. I came to Brown University in 1970 and was promoted to full Professor in the Department of Geological Sciences in 1989. Though officially retired as of 2005, I am still extremely active within the department.

Brown Affiliations

Earth, Environmental, and Planetary Sciences

Research Areas

deformation of rocks | earthquake prediction | nature of fault surfaces | tectonophysics

Research

Research Overview

I focus on understanding the mechanics of earthquakes. This involves a combination of lab experiments on the frictional behavior of rock and computer modeling of earthquakes that employs the constitutive relations that arise from the lab experiments. Computer modeling is used to determine if there are any signals that might be used for short - or intermediate - term earthquake prediction. We are trying to understand the processes responsible for the observed frictional behavior.

Research Statement

The experimental work by my group in the past several years has concerned: 1) testing whether previously proposed constitutive relations adequately describe rock friction in laboratory experiments, 2) experimentally determining the values of the constitutive parameters in these relations, and 3) testing whether theories of instability that employ these constitutive relations agree with observations of stability behavior in the experimental setting. We have found that the stability theories work well and we are able to understand quantitatively the transition from stick-slip to stable sliding in the laboratory. However, the constitutive descriptions only describe the behavior over more limited ranges of velocity than we have studied experimentally or are important in nature, and thus they represent only a partial description of constitutive behavior. Consequently we are attempting to measure the frictional response at intermediate slip velocities with and without pore fluid and are building equipment to slide at higher rates in the future. With pore fluid shear heating can cause the pore pressure to rise and thus the shear resistance to fall. Without pore fluid, shear melting can occur this is expected during earthquakes, but it does not occur at the rates we can now attain. Both processes could be important in allowing dynamic stress drops during earthquakes to be larger than static stress drops and can be important in affecting the magnitude of accelerations at the earthquake source and consequently the magnitude of the damage-causing strong ground motions.

To do the laboratory friction measurements I have designed and built a unique high-pressure rotary shear gas apparatus for the past 30 years I have been using and improving this machine. I am now modifying it so that we can slide at seismic slip rates of 1 meter per second. To do this at 100 MPa normal stress requires 80 horsepower, and the current electric motor and hydraulic power supply can only deliver 10 HP this is what presently limits us to only intermediate slip rates. Consequently, I have purchased a 100 HP motorcycle engine and I am partway through designing a way to adapt this to the rotary shear apparatus so we can slide at seismic slip rates and investigate the new processes that are expected to occur at these more rapid rates. I have also nearly finished building the sample assembly parts that incorporate an internal furnace so we can do experiments in rotary shear at both high pressure and elevated ambient temperature the most complex of the parts have already been constructed by our machinist.

The constitutive relations we presently use to describe the frictional response are primarily derived empirically from experimental data and are not based upon an understanding of the micromechanical processes that operate during frictional sliding. This is an unsatisfactory situation, both from the point of view of gaining a true understanding of the observed phenomena and that of applying laboratory observations to natural faults. One of the principal difficulties is that the processes that cause frictional resistance in rocks are not really known. Various possible processes have been proposed, in part by analogy with friction in metals, but the relative importance of these for rocks is unknown. One of the major continuing efforts I am engaged in is to try to understand the processes that operate and the implications these have for the constitutive behavior, both in the laboratory and on natural faults. This is an important problem, but a difficult one. The difficulties arise in part because it is hard to make observations of the samples on the scale at which the frictional resistance occurs. One of the important processes in frictional resistance is probably adhesion acting across very small areas of contact. In order to understand the role played by adhesion, I have been investigating the effect of chemical environment on frictional behavior, since recent advances in understanding intermolecular and surface forces suggest that different chemical environments should show different tendencies for attraction and repulsion between surfaces. I am also investigating the quantitative role played by small changes in the thickness of the fault zone and whether these may demonstrate that time-dependent indentation creep is responsible for the changes in friction that occur with time and velocity. This part of my planned research involves fundamental materials science and also has important practical applications for understanding the earth.

I also initiated a series of indentation experiments on quartz in a variety of chemical environments in order to determine whether the time dependencies observed in rock friction could be due to plastic flow of contact points. More work on this has been conducted at Oak Ridge National Laboratory with Dr. David Goldsby, an Associate Professor, Research at Brown.

In the last few years I have been conducting research on numerical modeling of fault behavior using constitutive descriptions of fault zone materials obtained from laboratory experiments. This is not only interesting in helping to understand the behavior of faults, it can be of value in determining the extent to which earthquake prediction can be aided by accurate constitutive descriptions. With the availability of increased computer power it is becoming more feasible to study realistic three dimensional models. I have made three dimensional instability models of the series of characteristic earthquakes that occur at Parkfield, California, and have made videos of these models to aid in understanding their complex behavior. This was intended to evaluate whether the model of Parkfield earthquakes could be predicted by existing arrays of field instruments in the presence of earth noise, something that does not seem possible. This numerical modeling of earthquakes suggests that although accelerating slip probably occurs at the earthquake hypocenter prior to the main earthquake, its magnitude may be too small to detect from the ground surface by most measurement methods. However, foreshocks are likely to occur as part of the accelerating slip process and, if they can be identified, they could form the basis for predicting the earthquake. I am interested in working with NASA to determine whether the space-based geodetic methods of GPS and INSAR might be able to detect the small displacements of the Earth's surface that we expect will precede some earthquakes.

I have become involved with a group of other scientists who are interested in trying to understand earthquakes through a combination of simulations of earthquakes and analysis of a wide variety of observations made of the earth that are relevant to earthquake occurrence and prediction. This group was intitally called GEM (General Earthquake Models) and I was the cochairman of the GEM Committee on Simulations. As part of this GEM effort I have attended eight workshops and have taken the lead in applying some advanced computation methods (Fast Multipoles) to this earthquake modeling effort so that we should be able to make much more realistic models using parallel processing on large supercomputers. My results so far on using this method suggest that it is possible to make models with enough elements that we can properly represent a continuum and can simulate microearthquakes as well as major ones in the same model. As part of this the GEM group received a 3-year, $2.2 million grant from NASA that allowed us to make some substantial progress solving these problems. As my part of this I have completed the programming and the code improvement that involves converting the earthquake code to run in parallel using MPI. Documentation on this is publicly available at http://www.servogrid.org/slide/GEM/PARK/ as one of the computer codes on the web site set up by our group. A subsequent evolution of this group into QuakeSim (http://quakesim.jpl.nasa.gov/) has provided a framework for continued efforts and obtaining grants.

Funded Research

Current Funding:
Laboratory Experiments on Fault Shear Resistance Relevant to Coseismic Earthquake Slip (with David Goldsby)

Quasi-Dynamic Parallel Numerical Modeling of Earthquake Interactions Over a Wide Magnitude Range Using Rate and State Friction and Fast Multipoles

Implications of Hydrothermal Laboratory Faulting Experiments for the Depth of the Stable-Unstable Transition, Slip-Length Scaling, and Seismic Hazard

A Collaborative Project: Comparison and Validation of Earthquake Simulators

Workshop on Earthquake Simulators

Laboratory Experiments on Fault Shear Resistance Relevant to Coseismic Earthquake Slip (with David Goldsby)

Quasi-Dynamic Parallel Numerical Modeling of Earthquake Interactions Over a Wide Magnitude Range Using Rate and State Friction and Fast Multipoles

Implications of Hydrothermal Laboratory Faulting Experiments for the Depth of the Stable-Unstable Transition, Slip-Length Scaling, and Seismic Hazard

A Collaborative Project: Comparison, Verification, and Validation of Earthquake Simulators

Workshop on Earthquake Simulators

Scholarly Work

Beeler, N. M., T.E. Tullis, D. L. Goldsby (2008) Constitutive relationships and physical basis of fault strength due to flash-heating, J. Geophysical Res., 113, B01401, doi:10.1029/2007JB004988.

Beeler, N.M., T.E. Tullis, L.A. Reinen, and A.K. Kronenberg (2007) The instantaneous rate dependence in low temperature laboratory rock friction and rock deformation experiments, Journal of Geophysical Research, 112, B07310, doi:10.1029/2005JB003772.

Tullis, T. E., et al. (2007), Rheology of fault rocks and their surroundings, in Tectonic Faults – Agents of Change on a Dynamic Earth, edited by M. Handy, et al., pp. 183-204, MIT Press, Cambridge, MA.

Trotta, J. and T. E. Tullis, An independent assessment of the Load/Unload Response Ratio (LURR) proposed method of earthquake prediction, Pure and Applied Geophysics, 163, 1-13, doi:10.1007/s00024-006-0128-9, 2007.

Tullis, T. E. (2007), Friction of rock at earthquake slip rates, in Treatise on Geophysics, G. Schubert (ed.), v. 4, Earthquake Seismology, H. Kanamori, (ed.), Chapter 5, p. 131-152, Elsevier Ldt., Oxford.

Background

Education and Training

Year	Degree	Institution
1971	PhD	University of California, Los Angeles
1967	MS	University of California, Los Angeles
1964	AB	Carleton College

Honors and Awards

Phi Beta Kappa, 1964

Sigma Xi, 1964

National Science Foundation Graduate Fellowship, 1964-68

Alfred P. Sloan Research Fellowship, 1973-75

U. S. National Committee for Rock Mechanics Annual Award for 1990 for Outstanding Basic Research in Rock Mechanics for the paper "Roughness and Wear During Brittle Faulting," J. Geophys. Res., 93, 15268-15278, by W. L. Power, T. E. Tullis, and J. D. Weeks.

Editors' Citation for Excellence in Refereeing, American Geophysical Union, 1998.

Fellow, American Geophyisical Union, 2002

Researchers@Brown