Interview TLE 2003

7/25/2019 Interview TLE 2003

1/4

Arthur B. Weglein received his PhD inphysics from the City University of NewYork in 1975 and then spent two years as aRobert Welsh Postdoctoral Fellow at theUniversity of Texas at Dallas. He enteredseismic petroleum research in 1978, first atCities Service Oil Company ResearchLaboratory in Tulsa (1978-81) and SohioPetroleum Company Research Laboratoryin Dallas (1981-85). Weglein spent the next15 years as a member of ARCOs researchstaff. He spent a sabbatical year (1989-90)as visiting professor at the FederalUniversity of Bahia, in Brazil, and threeyears (1990-94) as scientific advisor atSchlumberger Cambridge Research in Cambridge, England.In 2000, Weglein joined the University of Houston as theMargaret S. and Robert E. Sheriff Endowed Faculty Chairin Applied Seismology.

Weglein started the Mission-Oriented Seismic ResearchProgram and industry consortium in January 2001. The goal

is to address highest prioritized problems whose solutionswould produce the biggest positive step-change in the abil-ity to locate and produce hydrocarbons. This program isdesigned to meet the highest standards and aligned inter-ests of university education and research, seismology, andthe petroleum industry. In 2002, Weglein was promoted toa university-wide chair, the Hugh Roy and Lillie CranzCullen Distinguished Professorship in Physics, with a jointprofessorship in the Department of Physics and theDepartment of Geosciences.

The following interview was conducted by Bill Goodway,CSEG president, and Recorder editors Satinder Chopra and

Jason Noble on 28 April 2003, on the occasion of WegleinsDistinguished Lecture in Calgary. The DL, A perspectiveon the evolution of processing seismic primaries and mul-tiples for a complex multidimensional earth, was ultimatelypresented in 25 SEG Sections in six countries. The follow-ing article (edited for TLE and published with kind per-mission of the Canadian Society of ExplorationGeophysicists) first appeared in the September 2003 Recorder.

Your work has been unique, beyond even the best minds in ourbusiness. Its hard to look back at this point and see who woulddoubt the conclusions youve come up with, but I can see at cer-tain points you must have been seen as something of a heretic.

Many thanks for those kind words. Oh yes, a heretic.Early on I was nave and surprised by how much opposi-tion to new thinking there would be from both academicand industry geoscientists, mathematicians, and physicists.

Human nature, inertia, egos, fear of being superseded, andcollisions of ambition account for some of the resistance andnegativism. Now, Ive come to believe that new ideas willhave a concomitant and expected resistance, and to under-stand, predict, and anticipate it. If you want most people tothink well of your efforts, go into mild variation of currentthinking research, which by the way is a useful and neces-sary enterprise, but different from developing fundamen-tal new concepts and step change capability.

In hindsight, Im even more impressed when I think

about the amount of support we did (anddo) receive, in particular the depth, extentand duration of support we received frommanagementnot necessarily universalsupport, but all you need is enough to carryforward the program. Among those we aregrateful to are Hamid Al-Hakeem, JackGolden, Jamie Robertson, Jim OConnell,Dodd DeCamp, Mike Wiley, Phil Christie,David Campbell, Marlan Downey, BillClement, James Martin, Ken Tubman, ReidSmith, Bee Bednar, Craig Cooper, AndreRomanelli Rosa, Lincoln Guardado,Vandemir Oliveira, Paulo Siston, EdoardoFaria, and Jurandyr Schmidt for providing

the support and environment necessary to carry out long-range high impact prioritized research.

In our experiences these enlightened and courageousmanagers shared a certain scale of inner character, imagi-nation, confidence, a fascination (rather than fear) of newvisions of what might be possible, and those traits often pro-

pelled their own careers. They were often geologists or geo-logically oriented. They just had a certain personal opennessand orientation because they had this kind of courage andflexibility and willingness to support those ready to be thefirst to walk into a dark room.

Many mathematicians and physicists are not necessar-ily better suited than others to do creative things. It surprisedme to come to this way of thinking. They are often attractedto their discipline because things seem reasonable andordered. One can move away from the trouble of the realworld to a world where if you add one to both sides of theequation and integrate it all works, its all fair. The realworld is often not fair, people get hurt. Math can be anescape to a world of fairness. However, when you do reallycreative things, it doesnt come from math and physicsalone. Most mathematicians and physicists, like most peo-ple in general, want comfort in the current format, in thecurrent framework. To do creative things you have to jumpoutside that framework, where logic cannot take you, andtrust your intuition, to attempt to reach a new, broadersuperseding vision, and predictive capability. Its yourhumanness that separates you from merely manipulatingformulas and its intuition and feeling that are both thesource of inertia and creativity. In my view, mathphysicscapability and knowledge of geosciences is a necessary butfar from sufficient condition for significant creative contri-

bution.With the inverse scattering multiple attenuation research,

we didnt start with rigor; it largely began as a loose set of

guesses and then we sought a mathematical expression forthat thinking, and refinement of concepts by testing and eval-uation. If this is sort of how the internal multiple mightstart to be created in a forward scattering series maybe thisis how it starts to be removed in the inverse now can weexpress that thought as a guessed formula and evaluate itsvalidity, and refine and rethink the concept if necessary,thats how it started. It became clear that the inverse scat-tering series had the potential to remove multiples andimage/invert primaries at depth without the traditionalneed for a velocity model. Mining that potential into stableand practical algorithms is far from trivial. The entire seriespromises to input all your data at once and output the earth unfortunately the entire series doesnt have favorable con-

976 T HE LEADING EDGE OCTOBER 2003 SEPTEMBER 2003 T HE LEADING E DGE 00

CSEG interviews Art Weglein

Art Weglein

Editors note: In CSEG interviews Olivier Dubrule (TLE,September 2003) the name of one of the interviewers, Oliver Kuhn,assistant technical editor of the Recorder, was inadvertently omittedin the credits. TLE regrets this oversight.

Downloaded 01 Sep 2011 to 129.7.53.226. Redistribution subject to SEG license or copyright; see Terms of Use at http://segdl.org/


2/4

vergence properties, without a good starting model that isnot available in practice how to cajole and finesse usefulinformation from that overall series gave rise to a stagedand task specific subseries strategy: Locate where free sur-face multiple removal, internal multiple removal, imagingprimaries, inverting primaries are located within the series that strategy has been useful and provides significantpractical benefit and added-value.

When you take a step where there is no framework, youtrust in your humanness. Its your humanness that gives you

your edge. There are computer programs now that can domathematics better than we can, it can arrange equations,and it can solve equations. Our edge is first of all under-standing what it means, interpreting it, and going where thatmanipulation cant go because there is no logic step yet. Ithink we need to educate and require that our students havestrong capabilities in the tools, but they should also trusttheir intuition. Gut feelthats what gives you theedge. Roger Penrose, a pro-fessor at Oxford, wrote TheEmperors New Mind, a bookwhich demonstrates that it is impossible for a computer tomatch the human mind now because the current state of

physics is fundamentally inadequate to describe the humanmind. Theres a part of creativity that has to do with takinga step that you cant explain. In our field there is a big com-ponent of nonrigor at the beginning. Then others come andfancy it up and make themselves happy; that often takesyears. Theres an obvious important place for rigorous math-ematics to clearly understand the assumptions and condi-tions behind your algorithm, but if youre trying to take astep that will provide a significant new concept, and poten-tial capability, I know of no step that first came from onlyderiving things.

Art, I was looking at your research interests, and it says, Researchand development of new seismic technology that enables explorationand production of hydrocarbons, a prioritized list of problems is iden-

tified which is felt will have the highest impact. So what is yourprioritized list and what kind of problems are you working on?

In general, imaging at depth and delineating large con-trast targets, with complex geometry, beneath a complexoverburden such as basalt, salt, karsted sediments, and vol-canics are outstanding and highest priority issues for E&Pand hence, are at the top of our list. For example, we hearfrom oil company operations people that imaging beneathsalt was, and is, their biggest obstacle to current effective-ness in deep water Gulf of Mexico. Thats something we did-nt have particular knowledge of; we certainly didnt haveknowledge of velocity analysis. We were involved in devel-opment of migration-inversion theory, which along with allcurrent depth-accurate imaging techniques assumes an ade-

quate velocity is achievable. We were spending a lot of timeon multiples, 10 or 12 years, we could have continued, thatwould be an easy road. We chose not to do that. That would

be boring. Its not dangerous anymore. Theres a bit of excite-ment in hearing today at the Distinguished Lecture, Youregoing to get the accurate depth image when you havent gotthe velocity, before or after, give me a break! Those samepeople thought it impossible to attenuate all multiples with-out knowing the subsurface now its an industry standard.It was hard to fathom that that could be possible back then.Now we are pursuing imaging and inversion at the correctdepth beneath complex media, without the detailed veloc-ity. Its intriguing, it is relevant, and most important it is veryhighest priority for E&P and it deserves our attention. In

exploration and production you want more effective capa-bility, not solving insignificant problems that publish use-less papers, with low priority issues/parameters. If you aresolving the most significant science, stepping out, then youare having a step improvement in prediction and a stepreduction in risk. There is an alignment between the inter-ests of science and the petroleum industry; because they are

both best served by solving prioritized problems. A drill isempirical, its experiment. From a scientific perspective,mathematics and models dont mean anything until you

experiment; it isnt science until then. To paraphrase FrancisBacon and Albert Einstein, all the chicken scratching on the

blackboard is philosophy until you experiment and, for us,the experiment is when we drill into the earth. Theres a veryhigh bar for exploration and productionit has to be effec-tive. It has to move the drill from a less to a more reliablelocation.

Publishing papers is awalk in the park. Producingtruly new and impactfulmethods and algorithms isanother thing altogether.

You spent a lot of time with Bob Stolt, whose f-k algorithm came

at a very opportune time to help people to migrate data quickly.What is Stolt working on now?After hisf-kpaper, Stolt made pioneering contributions

to migration-inversion and inverse scattering multiple atten-uation. He published a paper just last year on data recon-struction, extrapolating, and interpolating data with a bettermodel. Stolt is very deep, very capable, enormously pro-ductive and creative, and very quiet. I often have a need toexplain to people what we do. When someone doesntunderstand I feel it necessary to try to make it clearer. Stolthas a confidence that, he understands, and if you dontunderstand perhaps thats your problem. Its refreshing tosee. Working together with Stolt is a special privilege; it alsoprovides a good balance to that issue for me.

I have followed your interaction with Guus Berkhout and the con-troversy over the need for the background velocity field for atten-uating internal multiples. I was impressed with your reply whenhe acknowledged that you were correct.

The statements that Berkhout made that you are refer-ring to speak to his personal and professional integrity, andI stated that at the time in reply to his statement.

In our personal history, Guus and I had a bit of an adver-sarial relationship early on. People found it amusing andentertaining, but then we sat down and decided to collab-orate. He was a very gracious and warm host when anopportunity was arranged for me to be a visiting professorat Delft University. What impressed me most working oneon one with Guus was his physical intuition and creativity.

He has made numerous significant theoretical contribu-tions, and he makes sure they are applied and have impact.We learned things from Berkhout and his group, and vice

versa, about the nature of multiples. Each group got stronger.We cooperated but we also were competing. And we bothdrove harder to the internal multiple on field data becausewe knew the other one was after it.

The main controversy, as you indicate, Bill, was aboutthe need for an accurate background velocity field for atten-uating internal multiples, which Guus defended and I saidwe didnt need. His agreement on that point at the SEG andEAGE meetings some years back speaks to his technical acu-men and strength and character. So now we are at a simi-lar juncture on that point, but there are still substantive


If you study whatever you are studyingin great depth, that will pay a dividend.



3/4

differences between our actual methods/approaches forfree surface and internal multiples. When you can isolatethe reflector in your data where an internal multiple has itsdownward reflection, there are computational advantagesto his approach but when you are interested in attenuat-ing all internal multiples without any interpretive inter-vention, or picking reflectors or anything else, the inversescattering internal demultiple algorithm is the method ofchoice. Both methods belong in your toolbox.

In our current research, were saying you dont need the

background velocity detail to get a detailed image in depth.As you could see there were quite a few people in the DLaudience today who were not comfortable with that asser-tion. At the least, we know its new when there is that degreeof discomfort.

I think the discomfort today at the lecture was in the lack of back-ground field for the depth image.

Right. Because thats new. Twelve years ago if you saidyou were going to predict the multiples without any infor-mation about the earthin other words, you were going tohave a section with the data and a section with the ringingfrom the salt predicted without knowing top or bottom ofsaltthey would have thought you were insane. Fifteen

years ago processing primaries were conceptually moreadvanced than processing multiples; now that situation isreversed. Previously multiples were treated as deriving froma 1D earth with a known 1D velocity, whereas primaries wereviewed as deriving from a multi-D earth with a knownmulti-D velocity. Free surface and internal multiples can beattenuated today with absolutely no subsurface informationwhatsoever about the heterogeneous and/or anisotropicmulti-D earth by arranging certain distinct additive andmultiplicative communications between the entire recordedwavefield of primaries and multiples. Multiplicative com-munication is the key difference that allowed multiple atten-uation to leap-frog processing primaries in concept andefficacy.

Current leading-edge imaging methods, once you have

settled on your velocity model, only allow additive com-munication between events that have experienced the samereflector. For this reason , an inadequate velocity model willresult in an inadequate image in depth. Sums of the surfacedata produce an image in depth, and weighted sums of theimaged data over angle produce the AVO response. Our cur-rent research seeks to extend the vision and capability toprocessing primaries that we earlier provided to removingmultiples. Allowing other specific types of distinct additiveand multiplicative communication between all primaries atonce, contains the conceptual possibility of imaging andinverting all primaries at their correct spatial locations, with-out knowing or determining an adequate velocity model.This provides a totally new vision for primaries aimed at

directly addressing our most significant and intractable E&Pproblems. That some might find this unbelievable today andnot deserving their attention and support is not surprising;what is truly amazing to me is the breadth and depth ofindustry support worldwide and the direct industry col-laboration we have received. Thats an extremely optimisticand positive note about the petroleum industry.

In our early history, migration-inversiona form ofmigration before AVOwas received with the biggest neg-ative blowback, both from people in migration and peoplein AVO. They made it clear in the strongest terms possiblehow crazy we were. Today its just another industry com-mon practice.

Youre saying the migration people didnt really believe the ampli-tudes they came up with?

Many migration people stated they were not interestedin amplitudes and the AVO mantra was, We can ignoremulti-D effects on propagation and reflection, because our 1Dmodel is closer to the real 3D world (since both are odd num-

bers) than 2D is to 3D. The guys who are really doing newthings more often than not can accept other new things. Thatsa big test of the mettle of a scientist. You first demonstrate yourrespect of science by trying to improve upon it. The toughest

test is to know that you yourself will be superseded. Of coursewe dont try to make that easy, we try to keep moving, butits going to happen. Everything we do has assumptionsincluding all the things we said at the Distinguished Lecturetoday, and we make them, acknowledge them, and argue fortheir reasonableness. Todays reasonable assumption is invari-ably tomorrows high obstacle to effectiveness.

I dont think anyone would accuse you of being an obstacle, but oftesting their ability to think beyond the box.

Thank you, Bill. In real research you have to be preparedto fail. Stolt and I worked on something in the late 1970s andearly 1980s that we eventually decided was intractable. Ifyoure really in the new research business, thats something

for managers to understand, and also in universities. If every-one were showing progress, I would say, Where are the fail-ures? If you are never failing youre not taking chances withnew thinking. Youve got to have some rate of failing. Stoltand I, separatelyhe at Stanford, I at Cities Service Oilwerelooking for a closed form solution to the simplest multi-D earth:Velocity varying in x and z. Just find an algorithm for that typeof model, without small earth property changes or other lin-ear assumptions, which didnt require a series. We workedvery hard, looked all over the world, and tried to solve it our-selves, and couldnt find a solution. We shut it down; youvegot to know when its enough. When you do that it isnt awaste of time. We learned a lot of things that guided otherapproaches to these prioritized problems.

How do you in a tenure system, or an industry judging

research, allow people who have potential big steps withinthem to have the flexibility to fail? If youre not allowing thatflexibility for some portion of your research staff, all youregoing to get are small, insignificant cosmetic forms of change.Counting the number of published papers is easy and quan-titative but rarely a metric of real contribution. Youve got tohave the right people. You give some people too much free-dom, theyll go sit on the beach and drinkpia coladas. I hada great manager at ARCO, Al-Hakeem who said, Why shouldwe send you to Brazil? If I think of whats best for ARCO fornext year, I shouldnt send you. But if I think whats best forARCO for the next 10 years, it makes sense. It will give youa place to think. It was that, and I am indebted to him. Allof the research on multiples had its origin there.

Is it your perspective that companies have all changed for the worsenow?

I think the oil companies are moving in the wrong direc-tion in generaltoo much short-term thinking on research andtechnology development, too much outsourcing. There are sev-eral hopeful exceptions, however, and reason for measuredoptimism. You dont outsource what you consider critical toyour business success and competitive advantage. On theother hand, I have also been extremely encouraged, and franklypositively surprised, by the broad and extensive industry sup-port of our efforts within the Mission-Oriented SeismicResearch Program at University of Houston. It speaks volumesabout the interest that exists within the industry for support-




4/4

ing fundamental research when it focuses on relevant issuesthat have the potential of providing significant increased effec-tiveness and reduced risk.

I also think that there isnt an overabundance of new think-ing from researchers overall, rather more complaining aboutlack of support. People who complain about industry notsupporting research generally have it backwards: if they werein fact actually doing real research they would have support.In my own experience, I have never seen a single creativeapproach that could provide, if successful, an increased capa-

bility that would not find support within the industry. I havealso found that industry by and large wants universities to beuniversitiesto concentrate on the core educational andresearch responsibility, on fundamentals, to be aware of indus-try issues and to be responsivebut not too responsivetoindustry needs.

I dont think there is much time allowed. The reliance is upon peo-ple like yourself and your students.

There is also a certain bit of responsibility to the turningoff of research, thats the responsibility of the researchers.Even if researchers are actually working on relevant problems,they are still looked upon suspiciously by operations.Operations have pressure; they need something done now, and

for them a great thing a week from now is useless. It is alwayseasier to show an increased profit by cutting costs than by find-ing and producing more hydrocarbons; and, research can bean easy target. On the other hand, researchers that either over-sell, hide assumptions and prerequisites, solve irrelevant andlow priority or convenient problems, and publish useless butrsum expanding papers, or try to masquerade technical ser-vice as research, also give support to short-sighted near-termthinking that damages the development of relevant and dif-ferential capability.

Do you interact with the majors in terms of how you might changeconcepts of sampling in acquisition?

First we look at a new concept for intrinsic capability givenperfect data conditions. Then we start to take away perfec-

tion to bring realism in. If it remains robust, then we ask whatdifferent kinds of acquisition would enhance effectiveness? Isit achievable, that is, does it have an added cost that makes itunrealistic, or is the added cost more than offset by increasedeffectiveness with increased reliability and reduced risk? Weget involved in all aspects. We are dealing with all levels ofthe seismic experiment because these new methods put ahigher bar on our expectations.

When you speak about multinationals, the majors, do they comedirectly to you with their problems?

We dont see their data. How we figure out what to workon is, for example, a company might say, We have a prob-lem with deghosting ocean bottom pressure measurements,

so we check around to see if this is that companys problemor a universal problem. For us to look at it, it has to be a globalhigh priority matter for the industry. The company might

show us the data as an example of the problem, but they dontgive us the data. We provide code, but its research prototypecode. We dont provide production strength code; its researchcode that has been tested on field data. We dont do tech ser-vice. Many universities are doing tech service; thats gener-ally a maltreatment of graduate students. If you have a codeand youre a professor, and some company wants to get theirdata processed, the faculty will use the graduate students asessentially processors. Processing data once is an education,

but processing it routinely is not an education and has little

or no research value. The company saves money becausegraduate students are not as well paid as contractor proces-sors. Further, the company can write it off as an educationalexpense, a tax write-off. The company saves money, the pro-fessor is a hero because he brings money in, and the only vic-tims are the students, because they are not getting an education,nor performing research. Fortunately most companies, espe-cially larger companies, avoid that practice. Universities losetheir way when they only measure success by the number ofpapers published and how much money theyre bringing in.We need to be vigilant to the educational mandate and roleand responsibility of a university in our society.

How different is it lecturing as an SEG distinguished lecturer ver-

sus the normal lectures that you give?The objective of the distinguished lecturer tour is to speakto the average SEG member, and thats a challenge when theactivity is very technical. If you really know what youre doingyou can explain it to an intelligent farmer. If you understandthe machinery, there is a way of explaining what it is tryingto do, without the math and the physics, just give some senseof why this new possibility is there. Its a wonderful part ofthe tour; you get to see people you havent seen for years, meetnew people, see wonderful places, and see places you haventseen at all. The difficult part is you are away from your fam-ily a long time, weeks at a time. What Ive done at most uni-versities and would have done here in Calgary if there weremore time is to follow the DL with a separate talk which ismore technical.

What advice would you give young people who are just entering thegeophysical profession?

Thats a tough one. When I go and speak at universitieson this DL tour, I tell them I dont really have a crystal ballabout what their livelihoodyou know, oil, environment,etc.will be. The only thing Im pretty sure of is that if theystudy whatever they are studying in great depth, that thatwill pay a dividend. They can transfer that to other things.We mentor students toward PhDs in seismic physics, whiledemonstrating and guiding them on how to choose andsolve problems that havent been solved. Thats a transfer-able thing. Dont assume that they will be in the narrow placethat we provide them, but make sure they understand that

what we are doing exemplifies a process, rather than beingthe process. TLE


Downloaded 01 Sep 2011 to 129 7 53 226 Redistribution subject to SEG license or copyright; see Terms of Use at http://segdl org/

Documents

Interview TLE 2003