In the second episode of Beneath the Subsurface we pick back up with a deep dive into onshore seismic technology in unconventional plays. Wayne Millice, Mike Perz, and Jason Kegel dig through seismic technologies, pre-stack seismic attributes, acquisition developments, and our predictions for the future of seismic and the unconventional realm. Erica Conedera, your host, new to the onshore seismic world, explores the challenges and sometimes over-hyped solutions with onshore acquisition and processing with our guests.
In the second episode of Beneath the Subsurface we pick back up with a deep dive into onshore seismic technology in unconventional plays. Wayne Millice, Mike Perz, and Jason Kegel dig through seismic technologies, pre-stack seismic attributes, acquisition developments, and our predictions for the future of seismic and the unconventional realm. Erica Conedera, your host, new to the onshore seismic world, explores the challenges and sometimes over-hyped solutions with onshore acquisition and processing with our guests.
TABLE OF CONTENTS
0:00 - Intro
1:51 - Onshore TGS History
2:35 - Acquiring Onshore Data
5:00 - The Migrated Stack
7:28 - Resolution: The Bug Bear of Processing
8:38 - Pre-stack Migration
9:55 - Pre-stack Attributes; The Good and the Bad
12:05 - Pre-stack: The Secret Sauce
13:48 - Noise, Noise, Noise
15:38 - The Future of Unconventionals; ARLAS, AI, and ML
18:35 - Joint Study with FracGeo: Pre-stack Depth Migration
20:39 - Analytic Ready LAS (ARLAS) and velocity Models
24:33 - Acquisition Technology; Surface and Subsurface
27:10 - Azimuthal Sampling - AVO and Velocity Inversion
28:22 - The Q Problem (Anelastic Attenuation)
30:08 - Frequency Problems
35:21 - Interaction with Acquisition and Processing
37:42 - The Future of Seismic in Unconventionals
41:24 - Conclusion
EXPLORE MORE FROM THE EPISODE:
EPISODE TRANSCRIPT
Erica Conedera: 00:12 Hello and welcome to Beneath the Subsurface a podcast that investigates the intersection of geoscience and technology. In our second episode, we'll deep dive into seismic technologies, pre-stack seismic attributes, acquisition developments, and our predictions for the future of seismic and the unconventional realm. From the software development department here at TGS. I'm Erica Conedera, your host and complete newcomer to the world of onshore seismic. I hope you'll find our discussion today as informative and enjoyable as I did.
Erica:
00:45
Um, so let's start with introductions to my left.
Jason Kegel
00:49
Yeah. My name is Jason Kegel. I've been with TGS for six years. I'm a geologist. I've worked on almost every one of the onshore US seismic programs that we have.
Erica:
00:59
Awesome.
Wayne Millice:
01:00
I'm Wayne Millice. I'm the gray beard of the group. I've been with TGS only about 11 years, but are, sorry, eight years. But I've been in the business about 35 years I'm the VP of onshore multiclient. And I'm here to hopefully teach some people about the value of seismic in our business.
Mike Perz:
01:19
I'm Mike Perz. I am the director of technology and the onshore group. So I'm responsible for looking after all matters technical in support that group. And I'm not quite as gray bearded as the gentleman sitting to my right, but I have been in the industry for about 25 years. So I'm kind of blondish with whisps of gray, I guess you'd say. (Laughter) No spring chicken.
Erica:
01:42
Awesome. So let's kick off the discussion for today. If you will Wayne by giving us a brief description of TGS' involvement in onshore.
Wayne:
01:51
Sure. TGS was primarily an onshore-offshore company. Up until about 2011 and 2011, we started the onshore business, January I believe, if I remember correctly. And that's how long I've been here, since January, 2011. In 2012, we acquired a company called Arcis in Canada that gave us an instant library of about 15,000 square kilometers in the western Canadian sedimentary basin. And in 2012 we started our first project in the US. And, we have you a since grown the library from the initial 15,000 square kilometers or so until about a 34,000 square kilometer based our database based in the US and Canada. So it's been a, it's been a fun run and it's going well.
Erica:
02:35
Awesome. So Mike, can you take it over for seismic technology? What do we do with the data once we get it?
Mike:
02:44
Sure. So the first thing that happens is that data has to be processed and I always like to call a seismic processing the Rodney Dangerfield of the E&P chain. And the reason I say that is as you might predict, it gets very little respect, certainly in terms of the almighty buck and the price, the price point's
Wayne:
03:04
Very little budget.
Mike:
03:05
Yeah, very, very little budget. And it's kind of ironic because as Wayne and I have discussed a lot, it's the seismic processing step where we have maximal client engagement usually during the course of a multi client project and reputations are won and lost on the processing. But again, very little dollar value flows with it. I don't fully understand why the valuation isn't higher, but it's a problem that I certainly can't fix. So we kind of, in a way, we try to almost leverage that fact that it's a fairly, fairly cheap technology and we take it very seriously at TGS. So with that preamble about why it isn't the most highly valued element of the, of the chain, let's talk about some of the key outputs from processing. So the thing called the migrated stack is probably the single most important processed attribute in an unconventional play in say, offshore environments like the Gulf of Mexico seismic technology is no one buys CEO's of a big oil companies as an important de-risking tool for say sub salt plays the, in the case of unconventionals, I would not say that seismic has that same kind of universal traction whereby everybody in the c suites on down know about seismic. Nevertheless, it is gaining a lot of momentum.
Erica:
04:34
And when you say unconventionals, can you elaborate on that?
Mike:
04:38
Yeah, I'm talking actually we're all going to be restricting the scope of this discussion to the shale plays onshore shale plays. In a, well North America primarily, primarily
Wayne:
04:50
Our primary focus on probably the Permian and the scoop and stack too. But there are several, several basins in the, in the US market that you could consider unconventional.
Erica:
04:58
Okay.
Mike:
05:00
Right? Yeah. So back to this business of the migrated stack, it is well accepted that it's a very useful thing in unconventional, development. And the primary reason for that is it helps in a delineating landing zones for the lateral wells and also geosteering and hazard avoidance. And I don't know, Jason, if you wanted to expand on a geological perspective of why those things are so important in the, in the depth domain. With seismic, you can start really understanding how to land your wells and doing geosteering in the unconventional world. That's one of the most important things that people are doing right now with their seismic.
Jason:
05:41
Geosteering in particular and finding these landing zones has been important because these reservoirs are, we're looking for is the conventional reservoirs can be anywhere from 10 to 50 feet, which is a lot of times right around the [Clears throat]. The area of seismic resolution, what we found to be more difficult is sort of calibrating everything together. So when we have the data, so calibrating the well logs, the tops, some of the understanding the differences in the different tool parameters your measured while drilling tool parameters versus your after drilling parameters and how that relates back to a depth calibration has been very important in the seismic industry. bringing all those things together to geosteer real-time to actually find these landing zones has been something that a lot of different softwares have attempted to do. And bring this into a multi-client aspect where the operator can instantly get a depth to calibrate and volume that they can geosteer on or look at their regional area of interest onshore has been very different than offshore seismic, which has traditionally had that depth migrated volume to begin with.
Wayne:
06:53
I can expand on one thing that Jason said too when we're talking about regional views on the petroleum systems. So our TGS has a strategy to date has been to get assets that are contiguous within these, with these within these basins so you can understand the regional view of it or of an oil producing basin or hydrocarbon producing basin. So it's important in our opinion that we get a large regional view. That's why you'll see you somewhere databases online. When you look at our, when you look at our projects, they're very contiguous and very focused on one area.
Mike:
07:28
Yeah. Jason gave a nice description of of why we might want to use migrant stacks, for geosteering. And he touched on something important. You brought up resolution and you talked about thin beds on the order of 10 feet to 50 feet. And one of the real bug bears are an unfortunate reality in the seismic processing world is the fact that we really cannot dive down to smaller resolutions than, than those beds. In fact, we're probably operating in, in the order of like, wavelengths of hundreds of feet. So resolving those beds is pretty tricky. We can detect them sometimes but not resolve them and we're always being pushed on the processing side to do a better job. And it's disappointing because all, sometimes all the acquisition equipment in the world isn't gonna help you through that. Mother Nature is cruel in a way and she chews up the high frequencies and there really hasn't been a breakthrough in seismic processing technology to allow us to bash through that, that limitation. So resolution is an ongoing issue and we're always squeezed by it in the unconventional context in the, especially for this geosteering. So that's worth noting. And one other quick thing, Jason mentioned pre-stack depth migration and that's an important new technology in unconventionals. Technology has been around forever for 20-25 years in the Gulf of Mexico, but it's really gaining ground in unconventionals and in in fact, TGS, shameless plug for a talk. TGS is going to be hosting a talk in early June, June 6th. Mariana Roche Davies is going to talk about pre stack depth migration and why it's valuable in unconventional plays.
Wayne:
09:07
We should be plugging a lots of things here, shouldn't we all sorts of-all sorts of shameless
Mike:
09:11
shamelessly plug. (Laughter)
Mike:
09:13
So, so if, if I could move away from the migrated stack, I just want to talk about the second big thing that seismic data is used for on the and the processing side. And that's the, the pre-stack data are used for generating attributes and we sometimes call this AVO analysis or Pre-stack and conversion. And the interesting thing here is that while the migrated stack has quite a lot of acceptance as a, as a really good de-risking tool for the reasons we mentioned, there is less universal acceptance o- the, these pre stack derived seismic attributes.
Mike:
09:55
Some I can think of one really technically astute interpreter from a Permian player who's very successful and they don't touch the pre-stack attributes because there are too contaminated by noise. On the other hand, you go to the SEG or URTeC and all that, there's tons of talks on using these pre-stack attributes. So it depends on who you talk to. Some people use them, some people don't. My hope is that they're going to be used more and more down the road. We're kind of pinning a lot of our own technical direction on that, on that premise.
Jason:
10:22
No pre-stack attributes have always sort of been the holy grail for, for people to find their, find their sweet spots. Right. I mean, looking at AVO in context, I mean that's the, the number one thing, right? And people are always absolutely to define their bright spot, right? And there's been tons of wells drilled just on that. But then to bring in rock mechanics and what they're doing with, with more pre-stack attributes in rock Brittleness and actually trying to look at Poisson's ratio and Young's modulus. When we start to look at those, we start to actually correlate the actual rock properties to what we're getting from our are sound frequencies. The more we can, we can do that and the better we can actually accomplish that is in the academic world has always been the, the, the driver. Right? And you can't talk to hardly any anybody that's teaching geophysics or rock mechanics or geology nowadays that doesn't want to talk about how to correlate your, your wells to your seismic. And it all comes down to understanding densities and shear wave and you're, you're compressed wave wireline tools and bringing that back to the, to the seismic world. unfortunately Mike is correct in saying that a lot of operators in these unconventional zones don't necessarily don't necessarily use it. They'll use it on their, on their own. They'll use a proprietorially, they'll use their own individual softwares to do that. But in a multi client aspect, it hasn't really caught as much traction as is, I think it will. And I think one of the big things that might push that is, regional is that, that's something you guys think the idea to have more regional studies of pre-stack attributes in pre stack, volumes.
Mike:
12:05
Yeah, I think, I think that's a good idea. I mean what one of the nice things with our huge well database at TGS as we can, we can leverage that massive information source into these regional studies. And one thing I forgot to mention was that this pre-stack conversion or attribute business, it does very well to have a lot of well control and we've got lots of that here. So that would, that would certainly help garner interest. One of the big problems, I think that that detracts from acceptance is just that there are not kind of generic workflows for what to do with the pre-stack attributes. Once you, once you have them, it's quite easy to, stare down a migrated stack and figure out, I steer here, I land here.
Mike:
12:49
That's it. You know that that protocol is easy to understand. What do you do with all these attributes? And different companies have their own secret sauce for that and sometimes they're quite tightly guarded about what they, what they do. So I think that may change in the future. We hope it does.
Erica:
13:02
Why do you think it might change?
Mike:
13:04
I just, I just think it will behoove everybody to leverage the seismic more everybody would win from, from that
Erica:
13:12
To be more transparent with their methodologies or?
Mike:
13:16
Possibly, I mean I think as as technologies emerge that-
Wayne:
13:19
Or we push or we push the methodology, for instance, we have the data points internally that we need to start pushing those to those new solutions so to speak or so push them out and then our customers will create their own secret sauce from hopefully some of our solutions that we're aware of or a team.
Mike:
13:34
And even as they push their secret sauce as the years tick away, typically people give up, they cough up their secret sauce to make a bad extended, a lousy metaphor. But they tend to divulge it and public domain and we all benefit from it.
Wayne:
13:46
It's another paper at URTeC.
Mike:
13:48
Exactly. So yeah, I guess this seismic technology thing is my bailiwick. That's why I'm doing a lot of the talking here that I was going to move on now to future future looking at data processing first of all and take a stab at what what I think are important technologies of the future. One is an old thing, it's noise, noise, noise, getting rid of noise, especially in places like the Permian. The Permian is so nasty as regards seismic soundings. You've got these horrible near surface layers, of anhydrites and salts interspersed and then you get these, these fills zones where the salt collapses and it, it kind of bedevil's all your seismic tools in many ways. And so that's why that one operator I was telling you about is reluctant to look at at their pre stack data for fear of the noise, screwing up their analysis. So we've got to do a better job at noise. We've got to do a better job at eliminating multiple energy. Full wave form inversion is a fairly well established technology offshore. We need to leverage that knowledge and get it going. Working better onshore for us, gets a nice velocity models among other things. Those are good for feeding this pre-stack depth migration technology.
Erica:
15:02
What are the challenges of leveraging that?
Mike:
15:04
Good question. The data are noisier on land typically. And so that isn't totally compatible with the full waveform inversion model to
Erica:
15:13
So you have to adapt the model.
Mike:
15:14
Adapted, got adapted to handle topography, things like that. And there are people are, people are doing that. We were certainly very active in that, in that space at TGS. Some of our competitors are as well. But again, I don't think there was this sort of routine commercial use at this point. I mean I know there's not just yet, but we're getting there. So yeah, those, those are kind of the big, the big things.
Mike:
15:36
Now the last thing I was going to ramble on about a bit was taking a future look at interpretation. So where would interpretation be going for for unconventionals? Cause I mean, Jason, check me if I'm wrong, it's really a different beast than conventional plays where interpreters have, there there special ways to stare down data and pick sweet spots and bright spots. This is not that, that same thing. I and I could be off base here. I'm just prognosticating. I think that, one important thing in the future we'll be using machine learning and at TGS we could leverage our data and analytics group for this stuff and basically use machine learning to tease out complicated relationships between seismic attributes and production and completion data points with the view towards being able to predict from the attributes alone where the next landing zone should be the next well.
Erica:
16:32
It's shameless plug. Our first episode was all about machine learning and AI. So please check it out if you haven't already.
Wayne:
16:38
on there. So there're interesting conversations that our AI summit to sort of speak about who would be picking the next location. Would it be AI being confirmed by a human or human confirming AI. So there was a, that was pretty interesting discussion of that, that ti's a good point to bring up.
Mike:
16:57
Yeah, for sure.
Jason:
16:57
And when it comes to interpretation in particular with seismic and how machine learning can help having all of that data readily available in the cloud is, or the first step, right? So when it comes to machine learning, it's just a matter of the more data you have the in the, in the machine, the better you're going to have it coming out. But that's everything that TGS does have, right? The well data start including tops, production completion techniques, different attributes for seismic. Then you actually get the machines starting to actually tell you where your reservoirs are going to have sort of different permeabilities, right? If you could start understanding where these different permeabilities come in and these shales, very slight variations can lead to huge benefits in production. So that's a, that's a very big thing that we would love to be able to do, but it's not quite there yet.
Mike:
17:48
Yeah, I mean I think you've raised a good point. We feel like we have all the ducks in a row here at TGS and it's, it's interesting because there- others before us have played around with multivariate analysis too to try to fit these attributes to things like production. They don't have the breadth of data that we have at TGS and they don't have as ready access to a lot of these things. So we're, we're poised to do some, some pretty cool stuff. So watch this space as they say. The only other thing I was going to say on on future looking interpretation wise, and I again I - disclaimers cause I could be wrong, but I believe that that combining seismic with geomechanical modeling software, may be an important thing to that end. And again, what is this our third shameless plug?
Wayne:
18:32
Well we keep doing it because that's what we're here for. (Laughter)
Mike:
18:35
So we're undertaking a joint study with FracGeo, a Geo mechanical modeling software and Services Company in the Permian Basin on our west Kermit Dataset in the Delaware. And we're going to be reporting back on that soon. But basically we're just, we're taking our seismic data and post-stack attributes like curvature to predict fault locations and that becomes feedstock for their Geo mechanical modeling stuff. And also the stuff you brought up, Jason Poisson's ratio and all the things we glean from inversions, those will go into their geomechanical modeling process as well. So that you know, hopefully that's a new sector in which seismic can be used.
Erica:
19:11
We realized that we missed something, We need to circle back around to the topic related to pre-stack depth migration gentlemen.
Mike:
19:20
Yeah. Pre-Stack definite migration in unconventionals. We kind of give it short shrift. I just wanted to add a few more more things. I had mentioned that it's a very established technology pre-stacked depth migration in offshore plays, Gulf of Mexico and such, and it's only been over the last couple of years that operators are using pre-stack depth migration a lot for unconventionals.
Mike:
19:40
It's interesting to note you don't get the jaw dropping improvements on the migrated stacks that you do in the Gulf of Mexico because the data are not nearly as structured. Right, Jason?
Jason:
19:50
Right, in most areas when people say railroad tracks, they're not kidding.
Mike:
19:54
Yeah, yeah. So, so you don't get these amazing glossy brochure image improvements on the stocks, but the, the benefits come in subtler but still important ways. For example, you get natural output and in depth is one, one really important thing and another thing you get better fault definition after pre-stack depth migration. Sometimes I think the real prize can be the actual velocity model itself. One really important difference in velocity model building for pre-stack depth migration in the unconventional onshore case compared to offshore is that in the former case, in the onshore case, we've got so much more well data to constrain or lock down our velocity models, especially at TGS with our massive well database.
Mike:
20:39
And so that's, that's a really, really good thing. So that's why I feel quite confident at the end of the day the velocity models are so responsibly constructed that you really can trust those depths and you get this natural depth conversion after depth migration that's as good or better than what an interpreter would do using his favorite or her favorite method for for depth converting time process data and on that well topic are TGS so-called ARLAS synthetic, well construction using machine learning. That's really gonna help our depth model building. We've yet to exploit it, but we're going to basically be able to get way more sonic wells through this ARLAS process to constrain interval velocities
Jason:
21:24
And that's, that's a big benefit in the shallow, we start looking at the, the shallower area for drilling hazards and drilling risk. we also start looking at that for water, for water. So in the Delaware, it's a big issue, not only just produced water and injected water and saltwater disposal, but making sure that the, the drinking water in the aquifer water that's usually in the shallower intervals is safe. So it's an environmental concern that we look into having that velocity model better structured in the upper sections that we normally don't look too much into and we're looking at exploration per se onshore, helps quite a bit with that, both environmental and with, with hazard mitigation.
Mike:
22:05
And the ARLAS construction will help that process, right?
Jason:
22:10
Oh, absolutely. The ARLAS dataset- any type of velocity model that can improve on the, the prior velocity model is of big concern. So you can get back to geosteering. Anything that helps that velocity model. A lot of times when they are geosteering, they'll have realtime velocity model building as the mud loggers are providing new information. They cross different faults, they notice different things that can instantly update the velocity model they're using to help steer that well. So it just goes back to the fact that having the best velocity model up front is going to help the, the final piece of the puzzle, which is landing that well on the, the zone where you can get the most oil or gas out of it.
Jason:
22:53
And that's been shown there. There's been a bunch of studies that have shown this, but there was one in the Balkan a few years ago that showed that using 3D depth seismic helped reduce their costs with 75% just by having their geosteerers use seismic. So that's you know, it's a known value for, for the, the seismic industry and the oil and gas industry to, to geosteer with depth migrated volumes. And it's nice to see that and the multiclient aspect that starting to really catch hold.
Mike:
23:26
Absolutely. And let's just push it onto those pre-stack attributes.
Jason:
23:29
No, I know, we just need it in the attributes.
Erica:
23:33
Okay.
Jason:
23:34
Particularly with faults. All right, so you're talking about some of the coherence studies with the post-stack, but when we can take some of that pre-stack ideas about Brittleness and Poisson's and Young's modelists and looking at those pre-stacks, bring it to the post-stack to where we can start identifying the fault structures and how those faults work. If you're interpreting those faults on your seismic before you go into your completion plan, then you have a much better idea of how you can track that well horizontally. So these wells nowadays, are a mile two miles long, some cases, I mean there, there they go for quite a ways going over some of these faults that have 20 feet to 50 feet to throw can greatly throw off where you're steering that well. So any type of better velocity model, will help you guide that. And a lot of times these faults, they're under seismic resolution. Again, so any type of fault or any kind of deviation that you can see in the seismic or with that velocity model is going to help you with your, your drilling plan and your completion plan.
Erica:
24:33
Okay, so to pivot a little bit; acquisition technology?
Wayne:
24:36
Well, I can chat a little bit about that. So I was in the contractor community for many, many years and back in the day we are pretty happy with, if you take it up from a spatial sampling standpoint, we were pretty happy at the end of the day when we were getting 100,000, 200,000 traces per square mile.
Mike:
24:56
How long ago was it? How long have you been? 55
Wayne:
24:58
Long time, yeah
Mike:
24:59
when did you enter the industry? 65 years?
Wayne:
24:59
At least 65 years. Yeah, (Laughter)
Wayne:
25:04
I was still microfilming, right? (Laughter)
Erica:
25:04
Sick burn
Wayne:
25:04
I've been getting- yeah, I get that usually from him, so that's okay. But now, the contractor community has made significant investments in equipment and we're actually acquiring datasets that are, millions have millions of traces per square mile, not just 1 million, but millions of traces per square mile. Now they've been doing this quite a bit in the, Middle Eastern markets because of the terrain. The train's fairly simplistic over there. So the ability to put several thousand source points in one square mile or one square kilometer or whichever you choose to measure by Canadian or US, has- is quite simple. Whereas in the US, or the North American market per se, there is a lot more, what do we call, obstructions and they come from several people from several things. Mostly people I didn't slip there. That was a purposeful-
Mike:
25:58
Freudian slip.
Wayne:
25:58
Freudian slip yeah, But, so now that technology that high trace density wide azithmuth fully azimuthly sampled, that technology or that product is now available in the North American market. So, and it's getting more prevalent. We're starting to see a new acquisition techniques mostly with surface source because you're still limited in what you can do. Subsurface source, for instance, a dynamite, right. But with a vibroseis or any or other surface sources, you're able to acquire data probably for about the same amount of money. It was, like I said, I was getting 250,000 per square mile in 1996 and I'm getting millions for the same number today. Right. So it's a, they've seen significantly increased their their traits count, unfortunately haven't increased their profitability so that that's still a problem in the industry for the most part. But they're working on that. Hopefully at some point we can hopefully at some point we can, (Laughter) we can, get to a 10 million traces per squad or mildly because, go ahead.
Mike:
27:10
I was going to say, you brought up the azimuthal sampling and that, that reminds me, I, I've been conspicuous by my silence on azimuthal AVO and velocity inversion techniques and these techniques are, are in use today using surface seismic to help characterize horizontal stress anisotropy and the presence of fractures and I kind of on purpose didn't get into it too much. I'm bringing it up now because I know that some, some of the, some of the listeners are probably wondering why we're not talking about it, that these things can be, can be useful and unconventional plays. But I'm avoiding too much mentioned because there's somewhat controversial and they have a, in my opinion, limited realm of applicability when they work, they work very well, but they have been oversold in over-hyped. So like I could, I felt I had to, I had to go there cause you brought up azimuthal. I'm going to turn you back to your, to your, your comments though.
Wayne:
28:01
So as Mike, as Mike mentioned earlier, denser is better, but, as we've seen and we've tested and we've done all kinds of things in the field that mother nature has different ideas no matter how dense, we shoot these things. Once we drive that sound signal into of the ground, we don't know what's going to happen to it at the end of the day. So,
Mike:
28:22
Yeah, for, for example, Q, I like to say Q can rear its ugly head Q mean is my proxy for anelastic attenuation. And I don't care how, how many sources and receivers you deploy, you can deploy them every, every fraction of an inch and you're not, you're not gonna change the fact that you lose your high temporal frequencies. And so that you know that that's a real problem. And then certain brands of noise are really well suited to being crushed or eradicated through dense spatial sampling. So that's wonderful. But some things like random noise, sorry, like, like really, really tricky linear noise. that's heavily aliased. If it's complicated enough, then you might need really, really fine sampling to deal with it. And that's still kind of a research topic. Random noises, easier, random noise. The denser, the denser it is, the more you'll, you'll beat down the random noise. No quibbles about it..
Erica:
29:12
Maybe this is overly simplistic, but what causes Q, where does that come from?
Mike:
29:18
Oh no, that's, that's a good question. It basically, every time the earth vibrates because a seismic wave is passing through it, the vibration has some loss to heat. And so it's not a pure elastic phenomenon. There's an energy bleed off and that, that basically that, that, that effect winds up, it's been, it's fairly, fairly straight forward and demonstrate that that kills the high frequencies of your seismic waves.
Erica:
29:45
Okay.
*Mike:
29:46
So yeah
Erica:
29:48
If it's straight forward, then what-
Wayne:
29:50
It's straight forward for Mike (Laughter)
Mike:
29:53
It's straight forward from the viewpoint of the textbooks. I not going to derive that in real time, are you kidding me? No. My mind is mush over the years as I become more managerial and sales focused. So, but it's, it's well appreciated. It's well established in the community.
Jason:
30:08
So how can new acquisition technologies help to mitigate some of those issues? Like are there other things on the horizon that there we're doing or you think that might, that might be out there to increase the frequency spectrum both low and high?
Mike:
30:20
I, well maybe, let me return to the, the noise thing that first before I forget to reiterate, some of the spatial sampling might help to, to kill coherent noise that's alias. If you get a sample, fine enough to remove the alias. So that's, that's a good thing. But back back now to acquisition and the spectrum, the temporal frequency spectrum. Well on the high end with this Q effect or anelastic attenuation, honestly I don't think all the acquisition in the world is going to help you. If we need, we need to break through in other ways. Then there are some ideas about sparse spike deconvolution that had been around for a while. Maybe those will, those will improve over the years. On the low frequency side we are doing tangible things in the field. I don't know, Wayne, if you wanted to speak to them on the source and receiver side or,
Wayne:
31:11
Sure. We're starting to do some, some experimenting, I think it's actually become more than experiment. We're actually acquiring projects with what we call either low frequency or low dwell sweeps, so we're starting in a real low frequencies and moving, moving slowly through the lower frequencies and then ramping up through the high frequency. So we're driving that spectrum a little bit wider so to speak. Right. So there's a lot of analytics going on on whether that works or not right now. Like you can comment from the processing side, but-
Mike:
31:40
well it's interesting. Yeah.
Wayne:
31:42
The equipment's there to do it as always. There's always been the equipment to do all this neat stuff, but stuff we create the data. Three C's a good example. We create three component data, but a lot of times we only use the p wave and not the transverse and the inver- and the, the, the, the three. So we don't use the three, we just use two and we create these volumes, but we got other stuff that sits on the shelf. But now we're starting to utilize some of these, low frequency start points, so to speak with a vibrators.
Mike:
32:09
Yeah. Right. And same ditto on the receiver side, right?
Wayne:
32:11
Yep. Yeah. Oh yeah. We're trying to, trying to go with the five hertz damp and phones instead of 10 hertz. We're trying all these things, but have we gotten there and put it into production mode yet? I think we're on the cusp.
Mike:
32:23
Well, it's, it's, it's, it's interesting because a lot of clients are very interested in these technologies and there's definitely theoretical promise and we've demonstrated on synthetics that, you know, you can get good results by, by caring a lot about the low end. And we ran it a fascinating test that hopefully we're going to publish at an upcoming SEG workshop. Shameless plug number five, right?
Wayne:
32:42
Four or five?
Mike:
32:44
Five, six, I can't remember. So, so I'm a co organizer. Christof Stork is, is the chief organizer and along with Bruce Hootman and Rodney Johnston and myself work organizing this SEG workshop on land processing and acquisition. And we're gonna, we're gonna dive into some of these, some of these, some of these topics. And one of the things we're talking about is, are we actually really enjoying the benefits of this low frequency attention that we're, you know, that we're foisting on the soundings in the field. Are those low frequencies coming out at the end of the day after all our inversion products? And Are we really reaping the benefits? It's not clear. We ran an interesting internal tests where we, we acquired data with the low low hertz or low frequency phones and I think we had low dwell sweeps. We certainly had have lots of energy on the source side, on the low end and after preliminary processing the result, cause we had a control experiment where we didn't do all this low frequency attention and the preliminary processing showed that that when you were really attentive in the field to these low frequencies, you got a better answer. But guess what? After we got to final processing and we're able to use a second pass of something called deconvolution to really widen the spectrum, we found very little difference between the conventional acquisition mode and the and the the low frequency effort. This is at odds with some of the, some of the literature, and I'm not disputing other people's findings, but there might be a subtle effect with an area dependency to it. We'll see.
Wayne:
34:13
But is a subtle effect enough to justify asking one of our contractors to go spend x number of dollars on equipment to upgrade their crews, right? Or it's,
Mike:
34:24
I know it's a, it's a tough, it's a tough question. Tough question. You know, I guess if price points on the cruise side drop enough, sure it's Gravy, why not? But if not it might not be worth it. You might spend your money on other other things. I'm not sure.
Jason:
34:35
Was it not the low frequencies that help you differentiate liquids in, in some of the inversions that you do further down the road? Is that the, that's the the biggest benefit, right?
Mike:
34:47
That's I believe, I believe it's very helpful. The low frequencies certainly helped to, to lock down the low frequency model for the inversion they give you support. Where are you, at low frequencies, where you don't typically have such support with conventional surface seismic and, and I'm not an expert in inversion, but my understanding is some of the fluid effects do tend to show themselves better when you've got the right answer for the low frequency model. And that's facilitated by having some of these low frequency acquisition techniques in play.
Jason:
35:21
You had mentioned earlier how the seismic technology and processing is the sort of the, the biggest area where we get interaction with our clients. Right. And it seems to be undervalued in that sense with acquisition. Is that a way we can of push that to, to fill that gap so we have that interaction and on both sides?
Mike:
35:44
So interaction on the acquisition side?
Jason:
35:46
Yeah.
Mike:
35:46
Well it's a good question. I mean, my understanding is there's typically not a ton of engagement at the field acquisition stage yet. There's obviously some,
Wayne:
35:54
Actually I would say yeah, there certainly is our one, our pre funders, write a check, they want to have some, implement some, some say so to speak what's going on. But mostly once we've made an agreement, on parameters, all that stuff is pretty much on us to deliver what we said we'd deliver. So, but we do where we really interact with our customers, we help them, we take problems off their plate so to speak, by taking on the acquisition piece, the acquisition piece is the most labor intensive, right. And, but where we really start to get in with our customers and when we, after we get the data, we've done the field acquisition, we interact with our customers from the processing side a lot. So it's important to us that like we said processing's a small piece of our AFE, but it's the most important because that's what we deliver, and that's what they see. Right. So, the, the nobody, no, I always say this to my guys to say nobody remembers the farmer that shot at you. Nobody remembers the vibrator they got stuck in the field, but they always remember if you're AVO volume was crap when they delivered it. Right. So they always remember that. But none of us other than other stuff that went on the field ever matters when they're looking at and looking at data on that workstation. Right? Yeah.
Mike:
37:07
So this, the poor sister in the E&P chain is the processing somehow is, it seems to continually be this, this critical, critical engagement point for, for the client. I mean, I guess the client, they don't, they don't like having to deal with permitting and stuff.
Wayne & Mike:
37:29
No, they only pay - like you guys - take the load off.
Wayne:
37:31
We're taking that load off them. That's a big load. Trust me.
Erica:
37:34
So jumping ahead, what do you predict for the future of seismic in the unconventional space?
Mike:
37:42
Well, I think I state this without proof of course, but I believe that there's going to be an increased use of seismic, including outside-
Wayne:
37:51
Well, the, the data that there's a lot of, there's a lot of data that's been acquired in the US and Canada for that matter. But a lot of it's getting dated, right? So when we're talking about, just like denser is better. We mentioned that earlier, right? Denser is better. So we're finding that a lot of these processing techniques that, Mike has been mentioning earlier, don't apply very well to older data data sets that don't have high resolution and aren't sampled very well. So we're finding, probably a lot of these older servers, you're going to get over it or getting acquired again, right? So that's, that's one marketplace. But as the unconventional space goes on, I think you're going to find, find it. A lot of these, like I said, a lot of these older datasets and a lot of the, are you going to make some discoveries within these data as the processing techniques get better and as we use the attributes better and all those things.
Mike:
38:42
Yeah, 100% yeah. And I was going to say, I believe from my conviction that there'll be an increased use of seismic for that to reach for that to actually come into play. I think that we need to, as an industry use these pre-stack attributes that Wayne just mentioned more and more. And we also, I believe need to start using 3C converted wave data more. We didn't get into converted wave data at all on this Chit Chat.
Wayne:
39:06
That's another, maybe that's another podcast.
Mike:
39:08
It's - it could, in of its own, but you know, there, there's some great promise with that technology, like so many technologies, it's been oversold and over hyped to some degree. But there's some really interesting case studies in western Canada that show that it's got great potential. We had awesome converted wave soundings.
Wayne:
39:24
Yeah.
Mike:
39:24
On the loyal survey. Yeah. And that's so, so that might help to propel the increased use of seismic as well as increased use of these attributes. So that's, that's what I think is going to, it's going to happen.
Jason:
39:35
One other thing, I really think that seismic is going to help in completion engineering. I'm going, I think that's sort of where it's going to now and where it's sort of, we've seen that happen with some of the pre-stack attributes and just to use seismic first off and understanding exactly where to perf and exactly where to make your completion intervals and where you're going to get the best production, on top of all the regional work you do to, to start out.
Wayne:
39:58
And that'll impact the funding cost per barrel for our customers. So that's going to, we hope that that's the, again, the value of seismic, right? So how's that going to drive our business? How it's going to drive our customer's business at the end of the day.
Mike:
40:11
Yeah, absolutely. And I mean one fundamental thing I forgot to mention, and Jason, you check me if I'm wrong, but I think what's happening in the unconventional spaces that there's a a slowly growing recognition that's actually probably accelerating right now. That to the tune that hey, we can't just go factory production style with completing all of our acreages there's enough geological heterogeneity that the production in this set of laterals here from this pad is kind of different than over here or even among the laterals in a pad. Why is this one so different? Parent Child Interactions, let's understand them better and all these burning questions, they're demanding some sort of better gaze into the subsurface and that is seismic.
Jason:
40:53
That is seismic and that's where I think that's where you're absolutely right. That's where the future is driving it. If you can understand the parent child relationships between your multi well pads and pads next to you and how you're going to complete the entire basin on a stacked play basis, using seismic is going to be your, one of your only real tools to help out. And the better you have the air velocity models hammered down, the better you have your pre-stack attributes that can be involved in that study, the better off we are and I think we're well on our way.
Erica:
41:24
Awesome. Well, thank you gentlemen for being here for our second episode. This was a really, educational discussion for me as someone who is not from a seismic background. And I'm sure I've heard listeners as well.
Mike:
41:37
Been our pleasure, Erica. Yeah, yeah, yeah.
Jason:
41:39
Thanks Erica.
Wayne:
41:40
Yup. Good for-Thanks for dragging us all in here.