David Jackson (1943 - 2023)

Professor David Jackson passed away from a heart attack on 30 March 2023 at his home in Pacific Palisades, California. He is survived by his wife Kathy and children Kelly and Morgan.

Professor Jackson graduated in Physics at Caltech in 1965, followed by a PhD in geophysics at MIT in 1969. He was appointed Professor in Residence at UCLA in 1969 and Researcher in Leon Knopoff’s group. After three years he was appointed Assistant Professor in 1972. He received tenure in 1975, was appointed to Full Professor in 1981 and Distinguished Professor in 2000. He retired in 2011. He was Chair of his Department 2004-2008 and held a joint appointment in the Institute of the Environment and Sustainability, 2008 – 2011.

Professor Jackson was an iconoclast who challenged conventional wisdom and sought rigorous methods to recognize the truth, refuting and replacing widely held paradigms. As the validity of his work became apparent, his expertise was sought after by numerous government and professional bodies. He was appointed Director of SCEC, the Southern California Earthquake Center (1996 – 1999). He served as Secretary (2003 – 2007) of the U.S. National Committee of the International Union of Geodesy and Geophysics (IUGG) and on its Finance committee (2003 – 2007) and set up a program for the IUGG General Assembly. He wrote the document used by member nations to request support from their governments. He was elected the U.S. Delegate to the IUGG and Chair of the U.S. National committee for IUGG (1999 – 2003) and continued as a member of the governing body of IUGG (2007 – 2011). He was elected Secretary (1989), then President (1991) of the Seismology Section of the American Geophysical Union (AGU) and AGU Fellow (1993).

Below are listed several of the topics that Professor Jackson addressed.

Inverse Theory. In the 1970s standard least squares methods were used to infer model parameters and their uncertainties from the data. Jackson developed methods to give realistic error bounds based on defining the boundary of the multi-parameter volume with the data called the 'Edgehog’ method in contrast to the ‘Hedgehog method’ that had been developed in Russia by V. Valus. In the latter, all parameters were tested within the volume that fit the data. Jackson’s method, extended to a paper titled ‘Most Squares’ explored the outer edge, and so was much more computer efficient. These methods take into account covariance in the data and model parameters to define more extreme error estimates than standard least squares. This development was followed by his extension of standard least squares to include prior information in a formal manner by augmenting the normal equations with prior data equations normalized by their uncertainties and data equations normalized by their uncertainties. His method has been shown to be the basis of modern simulated annealing methods.

Palmdale Bulge. In the mid-1970s repeated surveying suggested a large bulge in the ground had developed around the section of the San Andreas fault near Palmdale, California, similar to bulges that had been reported before giant earthquakes in China. There was considerable worry that Los Angeles was about to experience a huge earthquake. He applied his analysis to reveal that the data supporting the bulge could be explained by cumulative, hitherto-hidden, surveying errors. In 1975, the USGS report of the development of the Palmdale Bulge caught the attention of then-Vice President Nelson Rockefeller who directed his science advisor to convene a panel to recommend expanding earthquake studies. This led to funds allocated to NSF and the USGS for earthquake studies, a precursor to the National Earthquake Hazards Reduction Program or NEHRP. Eventually, after this start, and further attempts by others, in 1979 the NEHRP was established. NEHRP funds earthquake research by FEMA, NIST, USGS and NSF. Jackson’s paper in Science: Aseismic uplift in southern California: An alternative interpretation, coming 2 years after the formation of NEHRP, was seen as a serious challenge. The USGS spokesperson chosen to challenge Jackson was a young, articulate scientist, Ross Stein, who engaged Jackson in debate.

Now, four decades later, recalling his slow reversal of opinion, Ross writes in Temblor “David had been right, the Bulge was gone. I immediately switched sides.”

But added “So, while he approached popular ideas with skepticism, he approached their advocates with equanimity.” (https://temblor.net/temblor/essay-memorium- remembering-david-jackson-13088/).

Gap theory. Many thought and still do, that earthquakes would occur in seismic gaps, that is regions along plate boundaries that had experienced no historic earthquakes and so needed to catch up. David Jackson and Yan Kagan analyzed 10 years of data and showed the reverse is true. They found that earthquakes are more likely to occur where previous earthquakes happened. To some this suggests that plate motions in gaps might be happening by aseismic slip, to others, 10 years may be too short a time for a process that happens over millions of years. In either case, on the human timeframe the gap model does not work.

Characteristic earthquakes and periodicity. The simplest model for earthquake hazard held that they occur on characteristic faults of a given length as determined by geological fault scarps, and therefore had a given size, and that earthquakes repeat quasi- periodically. National hazard maps based on seismic records and paleontological trenching used this model. The California Landers (1992) earthquake showed that motions on different faults could combine to generate a larger event than each alone. The seismic record had shown that perceived periodicity can be an artefact of examining small records in an otherwise chaotic sequence. Jackson led a group of multi-disciplinary collaborators to develop the statistical formalism to take into account trenching results and geological faults, and included non-periodicity, earthquake clustering, and fault cascading. It set up the basis of modern hazard maps.

Pre-earthquake strain increase. An early model of potential precursors to earthquakes, proposed that the development of strain before an earthquake would accelerate over the time leading up to the event, allowing prediction. Jackson showed the opposite is true. After the Landers earthquake, a group led by Professor Jackson installed the newly developed portable GPS instruments to measure the post-earthquake strain in the vicinity of the fault trace. They were the first to continuously measure this strain signal that decayed over months. Other instruments had shown there was no accelerated strain change prior to Landers, or before other events. Professor Jackson and his students then combined historic trilateration records and current GPS measurements to map southern Californian strain as a function of time. They discovered that indeed many of the regions of increased stain rate in southern California are located not where earthquakes are about to occur but where they had occurred in the past. This pattern is now explicable as viscoelastic relaxation in the substrate beneath the fault.

Professor Jackson was one of the leaders of the group that wrote the original proposal to form the Southern California Earthquake Center. His experience of using GPS at Landers made him a leading figure in establishing the southern California GPS network, over a hundred sites that track the relative plate motion between the Pacific and North American plates, showing where strain is accumulating. Eventually Jackson along with Kagan, led what is regarded as the foremost earthquake-statistics team, guided by a principle espoused in the title of his early article ‘Interpretation of Inaccurate, Insufficient, and Inconsistent Data.’ His first major demonstration of this philosophy was the afore- mentioned deflation of the Palmdale bulge.

Earthquake forecasting. Earthquake prediction has been a subject rife with anecdotal evidence, optimistic interpretation, and reinterpretation of pre-earthquake events. This set the stage for the statistical likelihood testing methods that were used by Jackson and continue to be used by colleagues today. He encouraged the use of the word ‘forecasting’ rather than ‘prediction’ to reflect the uncertainties involved. He was a coauthor on a paper published in Science titled ‘Earthquakes Cannot be Predicted.’ Based on his rejection of the gap model, he quantified the clustering model in which a forecast is based on the proximity in space and time of recent earthquakes.

He issued a challenge to earthquake forecasters to make prospective rather than retrospective forecasts. Largely, as a result of his input, SCEC set up an earthquake forecasting contest called RELM (Regional Earthquake Likelihood Models) in which different ideas are translated into earthquake probability as a function of space, magnitude, and time, against which null hypotheses can be compared in a formal manner. Different groups across the country, including Professor Jackson’s, signed up for RELM and contributed a 5-year forecast. After the 5-years had elapsed, the results were published by the National Academy of Sciences. The HJK forecast, authored by Helmstetter, Jackson and Kagan, is still (Bayona et al., 2022) regarded as one of the most successful methods. The formalism established by Jackson and Kagan was the basis for expanding RELM to an international forecasting effort, CSEP, the Collaboratory for the Study of Earthquake Predictability and GEAR, Global activity rate.

Professor David Jackson rose to a position of national and international distinction that implicitly recognized the need for his careful statistical approach in assessing the likelihood of natural hazards such as earthquakes. Keiti Aki (1930-2005), first SCEC Director, wrote: “One of my wishes I had when I left Japan for MIT 33 years ago was to witness in person how the western civilization, finally confronting earthquake hazards in California, would deal with the problem. …. (Remember the Palmdale Bulge) David has abundant courage to fight an entire government organization and a large segment of the geological community in these endeavors. Not only that, when it became time to integrate earth science information for the benefit of the public, he served unselfishly and patiently for building consensus among scientists of various disciplines and produced a product (SCEC II Phase Report) which became truly useful for the public. I see a genuine spirit in western civilization in what David has accomplished in the past 30 years and feel that my wish was fulfilled by him.”

Professor Jackson’s career could be used as a case study of the application of careful objectivity in science. He challenged advocates of various positions, but it never became personal. As a person, Professor Jackson was a kind man, with a wonderful sense of humor and was ever ready to lend a helping hand to all. He is greatly missed.

Paul Davis, UCLA