Data and curiosity: Working with mentor Donald V.Helmberger

Edward Garnero

School of Earth and Space Exploration Arizona State University Tempe, AZ 85287, USA

I arrived at Caltech in June 1986, with interests in both mineral physics and seismology. I was excited to either study mineral physics with Professor Tom Ahrens, or investigate great earthquakes with Hiroo Kanamori. Upon arriving, I decided seismology would be my path. I was particularly interested in Hiroo’s work on earthquake size,from some papers I read as an undergraduate (Kanamori,1978, 1983). In my first week, I met with Hiroo, and he asked me what my plan would be for studying great earthquakes. Well, I had no plan, no clue! He wisely noted that we are in need of more and better seismic data –which of course came during the decade that followed with the digital data explosion. Hiroo suggested I introduce myself to seismologist Don Helmberger next door – maybe he would have some interesting research projects for me to consider.

I knocked on Don’s open door, and he welcomed me in with an enthusiastic smile. His office was big with an open feeling. Tall windows lined one wall, the long side of his rectangular office. A considerably large ficus tree sat off to one side, and a very large table space occupied most of the center of his office (it probably measured 2 by 4 meters). I mentioned I was a new graduate student and hoping to do seismology for my PhD. Don immediately invited me to sit down at his huge table. There were stacks of papers and seismograms on it. With no small talk, he grabbed a pile of printed seismograms, plopped them down in front of us,and started going through them, one page at a time, page after page. Each page was filled with “wiggles”. They were printouts of microfiche film copies of World-Wide Standardized Seismograph Network (WWSSN) seismograms (Murphy, 1961). The sub-basement of Caltech’s Seismological Laboratory had a room filled with file cabinets of WWSSN data which could be scanned and printed with a very big enlargement printing machine at the end of the room. Don had piles upon piles of file folders filled with these printouts. On each page, he pointed to wiggles he called “SKS”, and talked about its wave shape, page after page after page. With each new page there was an excited, “…and look atthisone!”

I had no idea what Don was talking about, and he gave me no indication that he was aware of my cluelessness about the seismogram after seismogram he was showing me. But his excitement and enthusiasm were so incredibly palpable. In that very first meeting I thought to myself that something this exciting to a Caltech seismology professor would probably be an excellent path forward for me! Little did I know that SKS would become my religion over the next many years. I quickly descended to the sub-basement,spending hours upon hours scanning and printing WWSSN records that contained SKS waves, then digitizing them on a table digitizer (Figure 1). This was a nearly daily routine until digital data became available (which was many years later). So, on day one I learned about Don’s unbridled excitement and enthusiasm for seismic data. Don had such a wonderful demeanor of sharing things with you,regardless of your knowledge of the topic. He was like a kid sharing a new favorite toy,every timehe showed you seismic data. His curiosity and joy were infectious.

After a whole summer of collecting data in the subbasement, the Fall semester began and I found myself sitting in Don’s “Advanced Seismology” course, with about four other students. I quickly learned of another side of my data loving advisor: theoretical and computational seismology. Don was not only an ultimate observer, he also actively pioneered ways to reproduce and explain seismic observations. It didn’t take long to be buried in Cagniard de Hoop theory in class and generalized ray theory codes in research, which enabled us to explore the contribution to seismic data from individual layers in a velocity-depth model. Several of Don’s early papers covered these, such as Helmberger (1968) and Helmberger et al. (1985), as well as his nice summary paper,Helmberger (1983), all which became our daily scriptures.Many of us worked with Don’s codesaseries.forbounce1or2.f, and other flavors of the old Fortran codes in our research. It was common to arrive at my desk to see a yellow sticky note with the words, “See me, OK”. When I went to Don’s office, he usually had a printout of synthetic seismograms from these codes to explain some interesting aspect of SKS waves.

Our work quickly evolved to study an interesting phenomenon that happens to SKS when it encounters the core-mantle boundary (CMB). At the critical angle of incidence of an SV wave converting to a P wave at the CMB, diffracted P energy is created (i.e., an ScPdiff wave). This wave transmits P energy to the core to become SPdKS (and from reciprocity, on the upgoing K-leg of SKS, SKPdS is also made). Our first paper on the phenomena (Garnero et al., 1993) established how SPdKS,which first appears and then grows out of the shoulder of SKS, shows up earlier in distance for very strong P velocity reductions at the CMB (Figure 2). This would come to be called an ultra-low velocity zone, or ULVZ,and modeled as a thin layer (e.g., 5–40 km thick) at the CMB in later papers (Mori and Helmberger, 1995; Garnero and Helmberger, 1996). Restricting the low velocities to a thin layer at the CMB was Don’s idea.

Figure 1. (a) Scan of an original WWSSN long period recording of a Fiji earthquake recorded at station ALQ in Albuquerque New Mexico, with the digitized version to the right (image originally from Garnero (1994)). (b) An example of a digitizing table similar to the one used to digitize paper records (image from https://www.alibaba.com/product-detail/Jindx-Clothing-Plate-Making-Scanner-Gerber_62555983285.html).

As another example of Don continuing to think about the problem happened after I moved to my first postdoc job at the University of California, Santa Cruz , to work with Don’s former student Thorne Lay. In that first year I would receive faxes with pages upon pages of step function seismograms (output from Don’s generalized ray theory codes that just needed a derivative and convolution with a seismic instrument to be compared to data). An example is presented in Figure 3, which shows that Don had identified a new and important feature in SKS data used to image ULVZ structure. Here, Don was showing that a multiple internal reflection in a low velocity boundary layer at the CMB also contributed significantly to the wavefield (Helmberger et al., 1996). This is just one example of Don’s uncanny ability to dissect, model, and understand the seismic wavefield. This additional unique seismic arrival combines with SPdKS (and SKPdS) to make anomalous SKS waveforms.

Figure 2. (a) Seismic ray paths of the phases SKS, SPdKS, and SKPdS. (b) Paths from two source regions, the Fiji-Tonga area, which have the Pd parts of SPdKS (thicker line segments) traveling in the Pacific LLSVP, and South America (mostly Bolivia events). (c) Digitized long-period WWSSN SV data with arrows pointing to SKS and SPdKS arrivals (which is a combination of SPdKS and SKPdS). Seismographic station names and epicentral distances are to the right of each SV recording. The Tonga event has SPdKS developing at a sooner distance (e.g., see station AAM) than for the Bolivia data path geometry (e.g., see station JER at nearly the identical distance — there is no SPdKS yet). ULVZ structure causes SPdKS to arrive at earlier distances. These figure panels are reproduced and slightly modified from Garnero and Helmberger (1996).

Figure 3. A 1994 Fax message from Don Helmberger to me, several months into my first postdoc job, showing step(Green’s) function responses for SKS and an additional arrival due to a multiple in a low velocity layer at the CMB (a 40 km thick ULVZ), SppKS (and SKppS). Epicentral distance is from 106 (top trace) to 120 (bottom trace). The asterisk depicts the arrival of interest for which Don sketched the ray paths in hand drawn cartoons on the right. This extra arrival contributes to creating anomalous SKS data that are used to map ULVZs.

Figure 4. Photos of Don Helmberger during the “Gala Dinner” of the 2016 Study of Earth’s Deep Interior Conference, at the Chateau de la Poterie, which is a Louis-XVI style castle built at the end of the 18th century. On the left, Don curiously takes in beautiful trees on the property. On the right, Don being Don, full of enthusiasm and joy.

Don’s insight into data analyses and methods always seemed to be far ahead of his time. I recently looked at a binder I have called “DVH conversations” (‘DVH’ for Don’s full name, Donald Vincent Helmberger). I learned to always write down anything Don said about data. I was very excited to find a page from nearly 30 years ago that is titled “DVH conversation summary” filled with bullets that summarized things Don told me to investigate over the previous year or two. The 15 or so bullets on that summary page have a number of things that, to this day, still have not been done and are incredible ideas!

I was incredibly fortunate to have Don Helmberger as an advisor. Of all the things he helped me with, the most profound for me was something he never explicitly verbalized. Don consistently showed me how the data, if you patiently study it without an agenda or a model you might want to prove, willalwayslead you to something.The data will always teach us and reveal Earth’s secrets.Sure, we have our ideas, our intuition, our speculations and hypotheses. But turning these preconceptions off to openly look at all the data is where true discoveries lie. This was so compelling for me because Don’s love of data always struck me as innocent, humble, and pure. No agenda. No ego. Rather, just pure curiosity. If you spent enough time with Don, you would see his wonderful innate curiosity towards a number of other things, for example, like human behavior, nature, politics, sports (basketball!), and more.Figure 4 shows two pictures from the SEDI 2016 meeting in France. Don, with great curiosity, looks at the beautiful tall trees. Standing there behind him, I marveled at him really pausing to take it all in.

Finally, to celebrate Don and his infectious curiosity, I purchased my own office ficus tree. It has grown quickly and almost reaches the ceiling. It is my own little daily reminder tolook at data, as often as possible, with an open mind, without agenda or a model in mind, to innocently and curiously observe. Thank you Don!

Acknowledgements

I thank Xiaodong Song, Thorne Lay, and David Wald for providing edits and comments that helped to clarify this paper.