Skip to main content
SearchLoginLogin or Signup

John F. Hawley (1958–2021)

Hawley made fundamental contributions to the field of numerical astrophysics and the understanding of gas accretion around black holes.

Published onAug 22, 2022
John F. Hawley (1958–2021)

Photo of Hawley taken in 2006 by Robert Rood, University of Virginia Department of Astronomy.

John Frederick Hawley, the John Dowman Hamilton Professor of Astronomy and Senior Associate Dean for Academic Affairs in the College and Graduate School of Arts and Sciences at the University of Virginia, died of cancer on 12 December 2021. He was 63.

John was born in Annapolis, Maryland, on August 23, 1958. When he was in the first grade his parents moved to Salina, Kansas. He graduated from Salina Central High School in 1976 and attended Haverford College in Haverford, Pennsylvania, graduating in 1980 magna cum laude with a degree in physics and astronomy. For his senior thesis he computed a nuclear reaction network appropriate for a helium-burning shell in a red giant star to determine the production of various neon isotopes; this project hooked him on numerical simulations.

When he was considering graduate studies he visited the University of Illinois at Urbana-Champaign, where Larry Smarr showed him simulations carried out with Jim Wilson of Lawrence Livermore National Laboratory. John went to UIUC, with Smarr as his advisor. His Ph.D. thesis focused on the numerical simulation of the infall of gas with angular momentum and the subsequent creation of an orbiting gas torus around the black hole in a fixed Kerr metric. The equations assumed axisymmetry, since three-dimensional simulations were not practical with the computers of the era. John and Larry decided to hold off on magnetohydrodynamics (MHD) until the hydrodynamic case was better understood and the MHD algorithm could be improved.

His planned thesis simulations could not be performed with the computers available at Illinois. Smarr had established a collaboration with Karl-Heinz Winkler and Michael Norman at the Max Planck Institut für Astrophysik in Garching bei München, in what was then West Germany. Winkler and Norman were carrying out the first detailed 2D simulations of astrophysical jets at resolutions never before achieved. This was possible because the MPI had a Cray-1 computer. With one million words of central memory and over 100 million floating point operations per second, it was the most advanced computer available at the time. John spent a month at the Institute in the summer of 1983, followed by a longer period in the spring of 1984. He completed his Ph.D. in the summer of 1984 and in September moved to the California Institute of Technology as a Bantrell Fellow in Theoretical Astrophysics, working in the group led by Roger Blandford and Peter Goldreich. Also while at Caltech, he worked with fellow postdoc Charles Evans on the problem of adding magnetohydrodynamics (MHD) to the general-relativistic simulation code.

In October of 1985 he attended a conference on numerical general relativity in Philadelphia where he met a graduate student named Katherine Holcomb, who was presenting her work on numerical simulations of non-isotropic cosmologies. She became his life partner and wife.

John started as an assistant professor with the Astronomy Department at the University of Virginia in the fall of 1987. He soon began to collaborate with Steven Balbus, who had joined the faculty two years before him. In 1990 he and Steve had started to consider the nature of waves in a magnetized disk with an eye toward understanding wave transport of angular momentum. Applying this analysis to differential rotation produced a surprise. Rather than a wave, one MHD disk mode emerged as a pure linear instability, now generally called the magneto-rotational instability or MRI. John had already set up a local simulation for the cross section of a shearing flow with orbital dynamics which, combined with recent code developments in MHD, allowed him to immediately run a short simulation of a vertical magnetic field in a disk. A plot from that simulation showed the previously smooth field lines developing kinks. A subsequent, more systematic study permitted them to match the theoretical linear growth rates with those observed in the simulation, and led to a simple intuitive understanding of the nature of the instability.

The next step was to determine that the MRI led to genuine turbulence and significant angular momentum transport. Long-time simulations of the shearing sheet demonstrated that turbulence was the inevitable end result. An analysis of the MRI for a variety of radial and vertical field orientations, and another study using toroidal fields along the orbit, showed the complete generality of the instability. These results were presented in a series of papers published in 1992.

For his work in developing numerical techniques for computational MHD and the significant discoveries made with those tools, John was awarded the 1993 Helen B. Warner Prize of the AAS.

John’s numerical work was always marked by the most meticulous care. He was justifiably proud of his MRI work, but he also was very proud of his work with Steven Balbus on investigating the absolute stability of Keplerian flow without a magnetic field. For decades it was simply taken for granted that, when viscosity is small, any shear flow would be nonlinearly unstable. It apparently did not occur to anyone to actually investigate the problem of the nonlinear stability of Keplerian flow numerically. John led the first systematic calculations to do so, and despite the fact that simple Cartesian shear was readily shown to be nonlinearly unstable, in agreement with observations, there was not a hint of any such nonlinear instability for Keplerian shear. This finding was very controversial and proponents of nonlinear instability insisted that it was really there, just unfortunately beyond the domain of all modern computation. The consensus now is that Keplerian flow is in fact nonlinearly stable, though a mathematical proof has yet to be found.

Around the turn of the millennium, John began a collaboration with Julian Krolik of Johns Hopkins University on global simulations of disks in the region close to the black hole. In 2003 he and Jean-Pierre De Villiers finally completed the task begun while he was at Caltech of constructing a general-relativistic MHD code. Along with Krolik and Shigenobu Hirose, they produced a series of papers reporting on simulations of disks around black holes with different spins. One result that emerged from those simulations was somewhat unexpected: a collimated jet along the black hole axis. In these simulations accretion, driven by the MRI, carries the magnetic field found in the initial condition and deposits it on the black hole. Material drains off the field lines leaving behind a stable overall magnetic field; the spin of the hole accounts for the energy carried off in the resulting jet.

In 2006 his career path shifted when he took on the role of Chair of the Department of Astronomy. He was reappointed twice before being asked in 2012 to join the Dean’s office of the College of Arts and Sciences at UVA as the Associate Dean for the Sciences. In 2020 he took on his final administrative role as Senior Associate Dean for Academic Affairs in the College.

In 2013 John and Steven Balbus won the Shaw Prize in Astronomy for their work on the MRI. John was elected a Fellow of the American Astronomical Society in 2021.

John’s lasting monument is the field of numerical accretion disk physics itself, of which he was a founding father. But he was also a great mentor, training many of today’s leading numericists, and a talented classroom teacher. He was famous for his sharp wit and dry sense of humor; a colleague in the UK called him “the funniest man in astrophysics.” His Astronomy 101 and Cosmology courses became legendary for his clarity of presentation and his extemporaneous jokes. Students came from around the university. One anecdote is representative of his wit. When John was trying to convey to an Astronomy 101 class what the nature of exponential growth is, he showed that if the human population growth continued at its current pace for another 10,000 years or so, then every atom in the universe would be in a human being. A student raised her hand. “But,” she said, “aren’t new atoms made every time a baby is born?” The class tittered. John slowly replied: “I’ll have to tell you how babies are made after class is over.” The class roared.

He enjoyed hiking and, in his younger years, camping. Music was a particular passion; he was especially fond of the works of Gustav Mahler. He was survived by his wife, Katherine Holcomb of Earlysville, Virginia; brothers Steven (Eileen), of Lawrence, Kansas and James (Amy) of Salina, Kansas; sister Diane (Bernie Robe) of Eugene, Oregon; niece Jamie and nephew Aaron; and his mother Jeanne, also of Lawrence. He was predeceased by his father Bernard Hawley.

See also Hawley’s AstroGen information.


No comments here