Invited Paper Submitted on Direct Inversion for Parameter Estimation and Comparison with Indirect Methods Like FWI and AVO
Posted on November 17, 2016
Dear Sponsors, Colleagues and Friends,
I hope that this note finds you very well. All is well on this end: very busy, productive and happy.
In the link below, please find : (1) a recently submitted invited paper entitled ’ A direct inversion method for subsurface properties: the conceptual and practical benefits and added-value in comparison with all indirect methods, for example, AVO and FWI” for an upcoming special section on ‘Seismic Inversion’ for the SEG journal ’ Interpretation’ , and (2) please find within the link several key background papers that are referenced within this recently submitted paper. A related video on this subject can be found in Key–note address, Abu Dhabi, March 31st , 2015 presented at the SEG FWI, Workshop Filling the gaps in Abu-Dhabi
Among the key contributions and messages of the new invited paper submitted to SEG “Interpretation” are:
What is the precise difference between modeling, and direct and indirect inversion?
What are the bottom-line E&P strategic and drill decision differences, the added-value and benefit of a direct inverse method compared to indirect methods?
Providing a detailed analysis and examples that demonstrate that solving a forward problem in an inverse sense is an indirect method, and not in any way equivalent to a direct inverse solution.
We show that methods like AVO and FWI are solving a forward problem in an inverse sense, and hence are indirect inverse methods.
A direct inverse method provides several important benefits- among benefits is knowing that you have actually solved the problem you had an interest in, and set about to solve.
In addition, ( as we explain below) only a direct solution can inform you of whether your problem of interest is in fact the problem you need to be interested in.
If the direct inverse solution doesn’t improve your drilling success rate you cannot assign blame to the method of solution, and you have identified that the issue and problem is the starting point and the problem you decided to invert and was considered the relevant problem of interest.
On the other hand , with an indirect solution you don’t know if the breakdown or lack of improved drilling decisions is due to the problem you are seeking to solve or whether the problem resides within the plethora of indirect so-called ‘solutions’ or resides within both.
The latter indirect methodology is therefore much less clear and valuable for problem isolation and identification- a critical part of addressing challenges- but that lack of clarity is great for generating workshops, for make-work up-dating search engines and huge HPC purchases and commitments and special technical session opportunities.
We provide a direct method for elastic parameter estimation for all amplitude analysis objectives- and explain how all current AVO and FWI approaches are fundamentally ad hoc and clearly violate the fundamental math-physics identity between changes in medium properties and the concomitant changes in the wave-field
We show that all direct acoustic or elastic inversion methods only require primaries for all structural determination and amplitude analysis applications, and all medium characterization objectives.
Direct methods only input primaries and specify precisely how those primaries are to be used to achieve direct migration for structure and direct parameter inversion objectives.
Ad hoc methods can be useful. Indirect ad hoc methods like FWI are seismic tools that can have a range of usefulness and can be made more useful when multiples are included in the indirect model matching schemes.
But with or without multiples the FWI methods are, and remain completely ad hoc, without any firm math-physics basis or foundation of knowing when, in principle, a solution will be forthcoming.
And beyond technical issues, a main problem with indirect methods like FWI is the groupthink, the intense and frequent over-selling, over -marketing and hype , that ( if history is any judge and guide) will have an inevitable blow-back and bubble bursting, that actually will discount and undermine a measured view that finds the actual range of potential usefulness of these ad hoc FWI methods within the seismic toolbox.
We show that indirect methods are never equivalent to a direct inverse method, even under the simplest and most pristine circumstances. We show exactly why indirect inversion has a much greater interest, requirement and need for low frequency data than the corresponding direct method.
We summarize all the conceptual and practical differences between direct and indirect inversion, and describe our view of the proper and necessary role of indirect methods within an effective and comprehensive direct and indirect inversion strategy.
We advocate direct inversion for the part of the real world captured by your physics, and indirect approaches to accommodate the part of reality that’s outside our assumed physical model.