VideoPoints

• ••• • •• • ••••• • •
GEOL 6390 3D Seismic Exploration - Day2 10-28-23 AM
Help Tour
© Distribution of this video is restricted by its owner
Transcript ×
Auto highlight
Font-size
00:00 Um Oh, ok, thanks. the uh other thing is um when
00:08 working with a company, they will standards that they'll expect you to
00:14 So they'll expect you to use Seismic . It might be for positive yellow
00:21 red and then to white and then blue and Cyan, that's a,
00:26 a common one if you're doing amplitude , like for a bo um but
00:32 wanna keep it standard. So when talking to other workers, uh that
00:41 you know what you're talking about you know, this is a fairly
00:46 class, but there are a bunch I, I showed a couple of
00:52 um in color yesterday. Uh like with the interpolation because you're gonna,
01:00 know many, many people said, , I never use that attitude because
01:03 looks like garbage. And the reason looks like garbage is because it's the
01:09 the default in landmark Kingdom suite the way it interpolates things like ASAM
01:16 and base or flustering. So I into old rock doc commercial package and
01:28 ask it to come up with five faces and I'm gonna have a yellow
01:32 , a green, a blue and faces fine. Gonna save that as
01:36 segue loaded into patrol, take a through it. It's got to
01:43 So around every little, you red blob, it's gonna have a
01:49 of rainbow colors circling that is and to the next one. It really
01:55 . Ok. But a lot of has to do with color. The
01:59 pitfall of color is not displaying your bar. So I'll ask you in
02:04 uh when you give me a always capture a color bar, you
02:08 capture it as a separate file. usually cleanest. OK? That's usually
02:14 . Um instead of like putting it the, in the volume itself and
02:20 reason is um velocity, what color you like to use for high
02:29 And what color for a low OK. Red for which?
02:38 Red is fast, blue is OK? Like temperature, high temperature
02:44 , low temperature blue obvious. What we gonna use for uh shallow and
02:58 ? All if you like scream? good. OK. OK. Typically
03:03 gonna use red for shallow because red your eye. That's why there's red
03:11 in the green forest. You animals can see it at least monkeys
03:16 see it. OK? And so red grabs your eye. So on
03:20 time structure map, the convention is red is a structural high blue and
03:28 tend to be structural lows, green be in between. So we typically
03:32 a rainbow color bar. But red high blue being well, now then
03:37 have somebody who's just a full waveform and all this beautiful geophysical stuff.
03:45 they're gonna plot a time structure map they're gonna say, well, red
03:50 gonna be deep and blue is gonna shadow and then everybody is like they're
03:55 at it, but they're looking at inside out upside down. They're
04:01 your eye is telling you the red shallow because that's what you're used
04:05 So you have to be aware of for your audience. Now,
04:13 we, I don't, she's from , so I don't know where she's
04:16 . But one of the guys Michael , he was a student here in
04:19 science. He was from Taiwan and showed me the Taiwan stock market.
04:25 , red in China is good, . Ok. You know, in
04:29 front, front, front of your , you might have a red light
04:32 at the water feature, things like . Very inviting. Good luck.
04:37 , when the stock market goes they use red. When the stock
04:41 goes down, that use green Dow stock market goes up, green goes
04:48 red. Ok. So, you need to be aware of
04:54 Good morning. You found us. you go to the other place
04:59 Oh, you saw the Note? . Good. Um, so those
05:03 , those are kind of pitfalls you to be aware of. So there's
05:06 things you always want to have on display in a map meeting and,
05:10 in a report or uh a uh technical paper or dissertation, whatever,
05:19 always want to have a collar you always want to have a scale
05:23 . Now, um if you're working that soon to be Chevron with or
05:33 you, that's the way it OK? Um Then sure, they
05:39 allow you to show a presentation at geological Society Houston meeting or at the
05:46 meeting and but they do not want numbers cross line numbers, culture
05:59 So culture data would be the names a town, ok? A
06:04 the old lake, you gotta clean of that up. They will allow
06:09 to show the survey. Oh, may have you crop it, they
06:13 have you rotated, you know, kind of thing with a scale bar
06:18 says five kilometers or two miles. fine. So you always have to
06:22 a scale bar for a publication, all the little stuff that you're gonna
06:27 in portray all that information that never as well in a document. It's
06:33 clutter and then the companies don't like . OK? So um so you
06:40 want to keep your figures as clean possible, simple as possible with a
06:46 bar, scale bar. And if a map, a North arrow and
06:50 anyhow. So um we're gonna start with the next um lecture and Utah
07:00 making coffee. I'm sorry, Jessica Alicia. No coffee for you.
07:08 Are you, you're not giving them ? Are you, you're giving
07:11 What's the little uh you could email a little thing and have coffee
07:16 right? OK, Venmo or something that? O OK, I gotta
07:23 my presentation here. We are gonna it in that. So we're gonna
07:35 uh a very simple concept but one that several of you are very
07:42 with, but everybody needs to know . OK. So we're starting with
07:48 basics and then going on to So I, yeah, well,
07:53 am sharing my screen. Thank you much. Um And I'm gonna minimize
08:01 , put that up there and then have to go up to uh slide
08:10 from the beginning and then I have go to display settings, duplicate.
08:21 right. So we're on a lecture three which is um spectral content limits
08:28 vertical resolution. OK. So after section, I want you to be
08:34 to visualize time and space variant signals terms of their spectral components.
08:40 So like is it low frequency, it high frequency, is it long
08:45 , is it short wavelength uh be to describe the impact of a low
08:51 filter? That means I'm gonna allow low frequencies to go into the result
08:59 high pass voter means I'm gonna allow high frequencies to go into the
09:05 And a band pass filter, which the frequencies between a lower and an
09:13 limit go through the results. And this is true for filtering of
09:18 kinds on your stereo system that grandpa . Um And uh probably your,
09:27 telephone, your television, uh you , sound bar and things, it
09:31 all those kind of filters on Then we want to evaluate strong constructive
09:37 from thin beds in terms of something call thin bed tuning thickness.
09:44 So here's spatial resolution. I'm gonna to, I can move that up
09:55 . Oh Hang on. He showed how to get rid of that I
10:03 channel. Oh OK. OK. here's fine resolution, medium resolution horse
10:14 . OK. Our friend Mona Lisa to the Puente Lima, the,
10:25 T PC K, any of you artistic. No. OK. So
10:37 uh France, they have um in , we got a great right here
10:40 town, there's beautiful impressionist paintings here the, at, at the Houston
10:46 , at the, at the Fine Museum. OK. And uh what
10:53 manet Monet and Serra and several of other artists uh of the late 18
11:00 through the early 19 hundreds found out that you could have a particularly bright
11:07 picture with high contrast if you didn't green, but you put dots of
11:17 and blue next to each other. then if you stood 20 ft
11:25 your eye would kind of 40 a if it goes through the lens of
11:30 eye, and you're gonna lose some and would mix them into a very
11:35 green. Instead of you seeing the . Then when you walk up,
11:39 see the dots. So let's look this um color in the middle,
11:44 is, I think magenta, but forgotten now. And depending on the
11:51 , you see how it looks it looks like the color is changing
11:56 on the background. So that's this of uh differentiation. And here's a
12:03 of uh by Signac Report de Saint . And you can, if you
12:08 at it carefully, it's a bunch dots. But even back, if
12:12 look, move back, if you up close at your screen because most
12:15 you are looking at a screen, see the dots and you move back
12:19 5 ft and you say, that's really bright and stunning. And
12:23 happening? Your eye is mixing the . It doesn't have that resolution.
12:28 you're seeing the mixed color. Here's the Hubble space, space uh
12:35 . This is the, I think is the web on the right.
12:39 the new one, right? So I should say web versus
12:43 In fact, well, I'll do next time. Oh, I'll do
12:47 now because I just hit this uh horses with satellites. So the way
13:03 um the way they've improved the image by having um a larger camera.
13:23 . Right. Should come up. . So you're gonna have AAA larger
13:29 on the camera and then they did things like to measure very, very
13:35 amplitude events. Um They actually refrigerate to keep the uh electronic noise of
13:50 of the circuitry from affecting the So it, it's really high
13:54 But anyhow, it's a great example improved resolution, how you get a
13:58 picture. All right. So in time domain here is a sinusoidal
14:11 So one frequency OK. And we're describe the period as the peak to
14:17 distance could be trough to trough. ? It could be zero crossing to
14:26 crossing. I'll spend two sentences on because as you're going to interpret and
14:36 uh on seismic data, there's a crossing that goes from a positive to
14:43 . And if I were to sketch this, let's see if I'm clever
14:49 to sketch oer. Yeah. And , this guy looks like the letter
15:06 , right Z like in Zorro. then this one, this guy looks
15:21 the letter S, this is OK? It's a rotated Z but
15:27 a Z and the other one's a and in patrol, that's actually how
15:31 , they're gonna have a little icon a Z and a nest in
15:34 So the one of them is um 90 degrees and the other is minus
15:39 degrees. OK? So the frequency one over the period measured in Hertz
15:48 H because it's Mister Hertz Capital W Mister Watts Capital A is Mister Capital
15:59 is Madame Curie, I actually cu capital C for Celsius. OK.
16:06 get that I've been an editor so gotta get those capital letters for units
16:12 are named after people. OK. Hertz and then we have also a
16:17 frequency. Oops, sorry, I'm drawing. Um we have a radio
16:23 omega that is uh measured in radiance second. So instead of cycles per
16:32 or Hertz, it's measured in radiance second. Now, why do we
16:36 that? Well, it turns out you're a signal analysis person like a
16:40 processor, the arithmetic is a whole easier. You don't have to carry
16:45 P along every step of the So it's just for mathematical simplification.
16:51 I see something in chat. Let's . Oh, back in 10
16:55 Oh That's you ta saying he's back 10 minutes. OK. OK.
17:04 to that. I've got to get of this. That OK.
17:18 I should be able to go Now let's go horizontally and I have
17:23 bunch of um I can think of as geophones on the ground and the
17:29 coming into the geophones. But let just think of seismic data that you
17:34 at yesterday in the lab. So can look at how many cycles per
17:39 vertically and then everything is measured in . How many cycles per meter
17:45 Well, guess what if it's perfectly flat, it's zero cycles per
17:51 horizontal. If it's dipping at, say 45 degrees. And I measured
17:59 vertical depth. Uh the the vertical was in depth and the horizontal axis
18:05 in depth like from depth migrated data Howie Joe talked about. Well,
18:10 the cycles per meter would be the , both horizontally and vertically.
18:15 So here we're typically going to use word wavelength. If a step
18:20 we're gonna use wavelength vertically and that'll peak to peak dropped across zero
18:24 no zero crossing. And whereas we a radial frequency omega radiance per
18:32 we'll have a wave number for horizontal is what we'll use. And they're
18:37 be measured in radiance per meter or per kilometer or radiance per kilo
18:43 OK. Now, um the, gonna have an in line wave number
18:51 a cross line wave number because we two dimensions horizontally. All right.
18:57 this will give you a little comfort us spectral decomposition and four a theory
19:05 you haven't done any signal analysis. uh Bert Bauer is a structural geo
19:11 and he was uh this paper he's here is about curvature and he's talking
19:15 different wavelengths. So here is OK? I gonna a mountain in
19:27 little valley and a small hill and little hill and a little valley here
19:33 the 1200 m component. So it's a smooth version of that mountain.
19:40 is a 300 m component. Here the 100 m component. And if
19:47 add this one, was this was that one? I get the
19:53 topography. OK. So that's what doing with uh spectral components when we
20:01 the data into wave numbers in this , wavelengths or into frequencies and we
20:08 vertical that you're comfortable with that. the true topography. Yep. So
20:18 would be like taking a picture of room be better if we put the
20:25 on and uh saying OK, let's look at the green component of the
20:35 . Nobody is hot. I think is wearing green today. You're wearing
20:39 green sweater. Is it green? , her jacket's not green. I
20:47 look hang on my eyes are Oh It's blue. Oh OK.
20:55 we're gonna put a blue filter OK. And then hey, she's
21:00 be show up very, very well the blue filter which is so many
21:05 in wavelength. OK. It's a frequency red is gonna be a low
21:09 . So we see part of your jacket, just parts of it.
21:14 . And then uh those of us are boring wearing gray. Well,
21:19 gonna show up in all the colors a little bit. OK.
21:27 So now here we're back to causal when I talked about uh one
21:33 two potato, one alligator, two and thunder, thunder collapse and lightning
21:40 . So I'm gonna have even functions odd functions. So, and even
21:48 . Here's my zero. And I've a function going to the right and
21:52 it's just flipped to the left. the even function, the value at
21:58 value uh the amplitude value at minus is exactly equal to the value at
22:04 T. OK. Then an odd is to the left of zero.
22:12 is the amplitude is whipped upside OK? So the value of minus
22:21 is equal to minus the value with T. Bye. And then we
22:30 a mixed function which is causal and equal to the odd plus the even
22:37 . So what we're gonna do remember and cosines go from minus whatever to
22:42 whatever. So if I'm gonna correlate the data with minus uh from
22:48 to 4 seconds, I've got to something in the negative thing. So
22:52 gonna say, well, one part's the others uh odd. And I'm
22:58 cross correlate sines and cosines with OK. Now, in uh seismic
23:08 , we've got normal modes of And probably some of you play guitar
23:16 piano. OK? So definitely you violin. OK? Good,
23:25 Let's see if your daughter is gonna how to play violin. Wouldn't that
23:29 nice? You should play spooky music her this week. You know,
23:34 in Young Frankenstein. Yeah. My daughter plays that and,
23:40 so on the violin you have the string and then you put your finger
23:46 on the fret and you're gonna have fundamental mode is gonna be the length
23:53 , what's the bar you call the and where her finger is.
23:58 So if it's a 12 inch, a big violin, six inches,
24:06 inches. But it got AAA ZERO and a zero crossing and a cosine
24:13 , half of the cosine wave. if it's six inches, the length
24:17 the, of the wavelength of that is uh 12 inches, OK.
24:23 then you can have a harmonic instead a uh six inches. If you
24:29 a wavelength that was half of it would be a peak and then
24:33 trough and go to zero, that be a harmonic thing. And those
24:37 the normal modes of that violin string way she's held. OK.
24:44 um what 48 theory does? It the same thing. Think of that
24:51 seconds of seismic data that you're looking in the lab as a violin
24:57 How much of that can be represented the fundamental, how much of it
25:04 the first harmonic, the second the third harmonic, the fourth
25:08 that's all the sound that, that string can make, it can make
25:13 fundamental and all the harmonics, nothing . And in 48 theory, you
25:18 into it and uh the geophysicists here certainly at least a bit. Um
25:24 you make it repeat on and on negative and positive, you still get
25:30 same harmonic things. So that's how do four E analysis. So we
25:35 the recording, uh the seismic recording signs and cosigns, but they're a
25:42 number, the, the harmon, base, the, the fundamental and
25:48 it's harmonic, it's not an infinite . OK. The smallest one,
25:56 smallest frequency you can have for digital . And I might have a slide
26:00 chose this, you cannot have less two samples or wavelength or two samples
26:10 period. OK. If you you're going to have a O
26:14 I'm pretty sure I'll have a If not, I'll pull one
26:19 OK. So here's an example of little wavelet. It happens to be
26:26 causal wavel minimum base wavelength because this a seismic processing book. And then
26:33 uh Agios has done and he's broken into a six Hertz component. A
26:39 and 18, a 30 Hertz And if you look at this,
26:44 got the different spectral components. Um if you look at uh this peak
26:51 , see how everything lines up most the data lineup that gives you a
26:56 most of the data line up that you a peek, a lot of
27:00 data lines up that gives me a peak, nothing lines up. I
27:06 zero. So down here, I constructive interference of the component up
27:13 I have constructive interference of the And in this song, then I
27:19 constructive destructive destructive interference. Up I have destructive interference down here,
27:27 have destructive interference. And in this , I have various degrees of constructive
27:34 . OK. So that's how we the data is a component. So
27:40 have uh a fourier transform. And Doctor Joseph Fourier and we can
27:50 are the even part of our data cross correlated with a cosine wave.
27:57 that's gonna give us a cross correlation the fish. So those of you
28:01 are geologists, you're very comfortable with a, let's say a self potential
28:07 and cross correlating it with uh a ray. No, not gamma ray
28:13 be very good, too busy, maybe uh a resistivity log, you
28:18 cross correlate them and see how similar are. Well, I could also
28:23 a resistivity log and cross correlate it sines and cosines. I mean,
28:27 not? So all we're doing with fourier transform is cross correlating with sines
28:33 cosines in a deep voice, we'll them basis functions because they form the
28:39 of the transform. We're gonna do same thing with the sine wave and
28:45 give us, uh we're gonna cross with the odd coefficients and that'll give
28:50 the uh sine coefficients. And if remember from trigonometry in 11th grade,
28:59 you took trigonometry in 11th grade, that the exponential of IP is equal
29:09 cosine P plus I sine phi that's theorem. OK. So then the
29:16 of I Omega T is equal to omega T plus I sine of Omega
29:21 . So I can take these data come up with a complex coefficient.
29:27 right. And the complex coefficient is to the cosine transform was I the
29:33 number and unit imaginary number uh square of minus one times the uh sine
29:41 . It so those are a 48 . So it's gonna tell me how
29:45 is the lo uh component of light in this room and it's gonna home
29:55 on lately, but she's wearing a sweater today. OK. Now,
30:01 inverse 48 transform is uh just closely almost backwards. OK. And so
30:13 , the four a cosine transform, take my coefficients for each frequency.
30:20 Hertz 10 Hertz 1520 2500 and 25 correlated with the cosine 5,
30:27 2025. And I add them And I get the even part I
30:32 the same thing with the sine coefficient correlated with the sign. The only
30:37 is I use a minus sign in of it. And if I do
30:44 numbers, I can take the complex instead of E to the plus I
30:49 T like here, gonna use E the minus I omega T. That's
30:54 this minus sign came up here. gonna reconstruct the data. So I
30:59 into the 40 A domain, I the coefficients and then I uh do
31:10 I want with the data, then come back. If I do nothing
31:13 the 48 coefficient, I get the data back. Right. So why
31:18 we do this? Let's say I'm seismic data underneath the a big power
31:26 and the power line is generating or transmitting voltage at uh 60 Hertz.
31:36 that 60 Hertz magnetic, the voltage going back and forth. That's gonna
31:42 a magnetic field in circles around the power line. My geophones are laying
31:49 the ground and if they're connected by , which is in the traditional way
31:54 cable is a wire that will then electromotive force, the magnetic field from
32:03 power line would generate an electromotive force the wire. And this is how
32:07 steal electricity from power lines. By way, you take a big coil
32:13 and you put it underneath the power and then you run it to your
32:16 . It's against the law. So do it. But it's,
32:20 it's done. OK? Especially if a survivalist. OK? You're kind
32:26 off of the grid. Um so now I've got my seismic data
32:34 I've recorded and I have this 60 component. What do I do?
32:38 go into the 48 domain, I at all the components, let's say
32:44 0.025 Hertz. I look at 60 set it to zero and it is
32:53 . OK. Indeed, if you're seismic process, sure you can even
32:57 better. We happen to know the at any particular time of what that
33:04 . So we can actually take not the 60 Hertz but the 60 Hertz
33:09 whatever the phase is, is that of recording because the whole country is
33:15 the same phase. Otherwise the power uh from Texas couldn't communicate to
33:22 to New Mexico to Oklahoma. All right. So that's how we
33:28 the data and we'll, when we , we'll either get out a
33:32 which is what we'll do. That'll a notch filter to give you the
33:35 Hertz. It might be a manufacturing that is uh got a machine that's
33:41 to 25 Hertz. I'm gonna maybe to notch that out. OK.
33:50 You know the plugging, unplug OK. OK. Utah is very
34:14 , dedicated. Uh Can you guys remote? Can you hear the fan
34:18 well? I can walk closer to fan here. I'll put it up
34:28 . This is what old men hear hearing. AIDS. Oh,
34:34 you call it self conscious men with hair that they grow longer to cover
34:39 gray hearing aid. So you probably see them. Ok. That's what
34:45 hear in this room. I always , hear the loudest person, not
34:49 spouse that I'm trying to talk Ok. Wavelength. Here's the Oklahoma
34:55 . We're actually doing pretty good this . Ok. Last year we
35:00 All right. But anyhow, we're pretty good and a common wavelength for
35:07 a velocity for rock 10,000 ft per . It's pretty, pretty common.
35:11 velocity, water is 5500. So would be moderately consolidated sedimentary rocks,
35:19 say, uh sands and shales of gauge. OK. And I'm gonna
35:26 at a frequency of 50 cycles per , 50 Hertz. And the way
35:31 do, if you remember from high physics or, or maybe college
35:37 the way you always check unit conversion you always bring the units with
35:45 right? Did you guys learn it way? OK. So I got
35:48 per second over per second and then seconds cross out and I have 200
35:54 . So my wavelength is 200 OK. So let's go see where
36:01 seat was row 66 200 ft above stadium. So I was one wavelength
36:08 the stadium. And uh we're gonna a resolution criteria of a quarter wavelength
36:14 two way travel time and Um That I can see things on the scale
36:23 , of 50 ft. So I be able to, to resolve from
36:29 data. Uh this lower, lower but nothing finer than that.
36:38 So there's the 200 ft wavelength. . So I've got a data set
36:43 uh West Texas Central Basin platform. the original data that we got from
36:54 Resources. If I recall while I here at U A. OK.
36:59 I'm gonna apply a bunch of I'm gonna apply everything at the same
37:04 . So I've got a low pass . I'm just gonna look at 0
37:10 10 Hertz and then I'm gonna ramp down to 20 Hertz. OK?
37:14 I'm looking at the lowest frequencies not strong and then low pass 20 Hertz
37:24 40 50. You see how we're better resolution? OK. So low
37:30 ah now I'm starting to see little in between the others. OK?
37:36 60 70 80 90 100. There's almost everything. Now, I'm
37:46 do a high pass filter. So gonna get rid of the 10 Hertz
37:52 , get rid of the 20 Hertz , get rid of the 30 you
37:57 see it doesn't appear you're losing resolution you are what you're getting is more
38:06 . OK? So here, where is the reflection? Let's say
38:10 reflection were a white uh peak. , is this, it is
38:15 it is that it is that you see how it starts to look
38:20 so as you become narrower band and want to define it for purposes of
38:28 in terms of octaves, OK? more octaves you have the better resolution
38:35 have. So if I go from to 40 Hertz, I'm improving resolution
38:40 going 10 to 80 Hertz. I'm my resolution also. If I go
38:45 5 to 40 Hertz, I'm doubling resolution. Both of them are
38:49 really valuable. OK? And I now I'm just looking. So now
38:55 everything looks kind of ringing. So now we're gonna do a band
39:00 over and this tells us where the content is. So here is the
39:06 to 10 Hertz and I, I've kind of scaled them all.
39:12 to 1010 to 20 30 to 4040 5050 to 6060 to 70. Now
39:22 , let's say it's 70 to Now you can maybe see where I'm
39:27 . Now. Well, which one these is the actual reflector and which
39:31 are the side lobes? OK. eighties and nineties and then 90 to
39:38 . So we've got good usable data at 100 Hertz. All right,
39:43 go back cause I'll probably talk about on week three. I see good
39:52 here and then I see all of stuff and the stop is deeply
40:08 And when you look at it let's at this event coming down here.
40:16 I look at it perpendicular to you'll say, oh that's high
40:23 That's not low frequency. But what we're doing all of our analysis on
40:30 traces. So what I am looking is the apparent frequency of everything.
40:37 let's make this much more geological. as uh a geoscientist, you're gonna
40:43 a reservoir. All right, and wanna estimate what the capacity of the
40:48 for, let's say co2 sequestration. the reservoirs tilt it and my well
40:57 vertical and what I'm going to measure the top of the base of the
41:01 is an apparent thickness. Now, I know the depth of the reservoir
41:07 theta, then I can find the , the true thickness is going to
41:12 the apparent thickness times cosine of OK. So we have the same
41:17 here. I have events that are dipping at maybe 80 Hertz, but
41:25 showing up on the 10 Hertz Well, it's because that's their apparent
41:30 as well. They're steeply dipping. , I do know. OK.
41:35 those, I don't know Jessica, still with us? Uh Yes,
41:41 am. OK. So I'm gonna my hands or do you see my
41:45 ? I do. OK. This , this is weird doing this.
41:50 yeah, because I don't see her here anyhow. So what Howie Joe
41:55 in the class uh two or three ago I'm gonna take an event on
42:00 seismic trace. I don't know where came from, but I do know
42:06 it came. OK. So I when it came, I have an
42:11 of the velocity of the earth. talked about how you compute that as
42:16 . And I'm gonna take that event put it on the locus of all
42:20 events that it came from. So the earth were a constant velocity,
42:28 would put it on, on a . If it were a more variable
42:33 , it would be an ellipse that stepped on and expanded and crunched,
42:38 cetera. But anyhow, you're gonna it on in the lips when your
42:43 are close, measured, close let's say 10 m apart, then
42:51 ellipses will constructively interfere when they are to the reflector. That's what he
42:59 in the migration class. OK? they will destructively interfere. That means
43:04 peaks will align with troughs where there's reflector. Now, in terms of
43:12 , the problem is our spatial aliasing we put the geophones, where we
43:17 the sources cost money that cost money at one millisecond instead of two
43:24 that doesn't cost us anything. That's this space. But put more geophones
43:28 the ground means I have to have times as many locations on the surface
43:34 the earth, four times as many , four times as cost. So
43:38 goes up. All right. So if my geophones are spaced further
43:47 then two cycles, 22 geophones per horizontally. So I've got my
43:55 Here's my little wavel coming across. that little wavelet is under sampled,
44:01 not going to destructively interfere. So we're gonna have are these vertical edges
44:07 the migration ellipses. That's exactly what ugly things are here. OK.
44:15 it's a migration, oh, in deep voice, you can take migration
44:19 a, you see, because that's it is. OK? All
44:24 So there's the band pass filters and ask you a question like this on
44:29 test. So you need to know band test go three does in your
44:33 you know, when you have uh Petro has some nice tools for
44:37 Pasco three. First. Let me this. You see these pictures that
44:45 just showed. This is what people in a processing shop to decide what
44:53 do I keep? What data do throw away? And they look at
44:57 and they say jeez, even at something centered around uh five Hertz,
45:05 got good geology. I'm gonna keep . Yeah, it's got some noise
45:09 it but I'm gonna keep it. they're looking for geology. And then
45:14 at 100 or 95 Hertz is the I see geology. I'm gonna keep
45:22 . Then at 100 and 50 they may just see random noise at
45:26 Hertz. Or two Hertz they might just smears, you know, nothing
45:31 looks like geology. So that's how design their processing flow uh by unbeknownst
45:39 you guys or maybe counterintuitive because of the mathematics they use. They're gonna
45:46 I see geology there. I'm gonna it. Ok. So it's pretty
45:54 . So broader band data. And is a book by a guy called
45:58 out of uh the Netherlands here. got a wave with that's fairly narrow
46:05 centered around 20 Hertz. And he's a single sand and three closure
46:13 And you can see, well, can see the package but I can't
46:18 the individual sands. Now, what done here is he's broadened the spectrum
46:26 so he's adding more frequencies. So you can see the individual events by
46:34 more octaves. Now in spectral we'll talk about that in week
46:42 we're gonna have a reflectivity series. have a source wavelet for every spike
46:49 gonna copy paste a wavelet and scale upside down right side up, add
46:56 all together that's called convolution. And use a little asterisk. We do
47:02 convolution that gives us our seismic And then the frequency domain, we
47:10 say that the magnitude spectrum is kind sort of white oh how to get
47:20 smaller. Oh There it is. . So the spectrum is kind
47:24 it's got the same frequencies at one as a 200 Hertz kind of.
47:30 . Then my source wavelength, we like our wavelength to be flat,
47:36 say between 10 Hertz and 80 Hertz we can't. So in the frequency
47:43 , we convolve the data copy page the complex frequency domain. So this
47:51 the magnitude there's also phases. The complex number, we multiply it
47:57 the complex way, spectrum and we our band limited white spectrum.
48:04 So we can do these things in frequency domain or in the time domain
48:08 usually it's much faster in the frequency . All right. And we're gonna
48:14 spectral balancing. Petrel doesn't do a job of it. Don't tell
48:22 I said that but they do a bad job of it. Uh It's
48:26 interactive. Good way of doing this balancing is to generate the power
48:34 And some of you might use Kingdom . The kingdom suite does a good
48:39 , much cheaper package. So here's power spectrum, let's say for the
48:44 survey, maybe time are in and gonna apply a band pass filter to
48:50 data. You're gonna compute the power each frequency component. All right.
48:55 I've got the data something we call tilt transform 90 degree phase version of
49:00 data. OK. That's the power we're gonna smooth the power spectrum XY
49:05 T to compute an average what's the way to I calculate the power and
49:12 every trace and then I say, the average power at a given time
49:17 frequency for the entire survey? Then I'm gonna compute the peak of
49:22 . So here's the peak at a time. Once I do that,
49:30 is my power here is a noise and subjective. A human being is
49:38 look at this kind of spectrum and , OK, signal signal. Oh
49:47 is Jessica rushing on I 45 to here after John hearing that you'll forget
49:59 . And this is her rumbling down 45. OK. So a human
50:04 is gonna look at the spectrum and , oh it kind of flattens out
50:07 . That's noise, that's the background , all the traffic on I 45
50:11 the wave noise on the ocean, the wind noise in the trees.
50:17 ? That's a threshold and it's gonna a fraction of the peak. And
50:21 you apply a simple filter to increase peak moving around. We're gonna move
50:30 part that's above towards one and the that's below towards zero. And there's
50:36 formula I'm gonna take the peak, this fraction. This is my
50:42 I'm doing things and uh the data in uh this is on power.
50:48 I have to take the half power it. And then here's my original
50:53 pass filter version of data. Here's new one I do that for each
50:57 bank, add them up. Some of the results. I don't
51:03 why these keep moving around on OK? Should we add them
51:13 That's how you do spectra balancing. . So here's the original data I
51:20 you and then your respective rebalance Has it changed it radically?
51:27 but can I see? Oh, , I look right in here.
51:33 , I see an event in Now, I'll look in here.
51:37 I see one event. Oh, I see two events um down
51:43 Uh Maybe there's something in there. yeah. Now I see it pretty
51:48 . So that's what spectral balancing OK. Here's an example from a
51:55 of mine at EN I the Italian Company original data from the North
52:00 Then data after spectral balancing and he's some Dolomites in this song,
52:05 he's got more events to pick now he did up here. Here's his
52:10 spectrum in black, his balance spectrum red, the vertical axis is in
52:17 . So it's a logarithmic scale. why it doesn't look that different.
52:25 um here's time domain spectral balancing using sparse white technique. So here's my
52:35 data. And then the first step are the spikes that when involved with
52:44 seismic wavel? Give me the original . This is what we do with
52:50 Decom Pou and maybe did how we about the convolution a bit good.
52:56 . So we want to find out the event where the reflectors are.
53:01 we first asked the question if I what the seismic wavel it is,
53:06 may have to estimate it. But I know what the seismic wavel
53:09 what's the most likely location and strength the spikes and I wanna minimize.
53:15 so how we probably talked about these ? Did you talk about these
53:24 OK. Did he talk about L ? OK. So you guys do
53:37 , right? And now I've got bunch of data scattered X versus
53:44 I'm walking around the room here now we just waving in the arm in
53:48 air. I'm not really doing anything . So we got and Alicia.
53:52 I, I have an XY plot I got scattered data kind of following
53:58 line and then I got so I'm least square fit in Excel.
54:04 you just push the button that says the trend line. OK? Now
54:10 have one measurement that is sticking way in the middle of nowhere. If
54:16 use that event in the lee my trend line is now going to
54:23 heavily biased by the strength of the . So I take the distance of
54:28 point through the trend line, I it all the other data. It's
54:33 just errors of like 001002 and this got an error of like 10,
54:42 dominates it. So now that trend doesn't follow the trend of the
54:46 it's highly biased by that bad So instead of minimizing the distance from
54:54 straight line, instead of minimizing the distance, let's minimize the absolute
55:05 OK? And in mathematics, we'll one L two, that means square
55:11 . And absolute value would be L that would be it, that the
55:17 one norm is what they call OK. And then that is going
55:21 minimize that uh effect of that outlying . So we use that a lot
55:29 statistics. And in science, we away outliers. And so that's what
55:34 do with the sparse spike conversion. trying to, instead of giving a
55:40 bunch of little bitty points that represent data, we're gonna get a minimum
55:46 of bigger points that represent the Then I can replace my wavelet with
55:54 narrow broadband wavelet and come up with extension. So you'll see this technique
56:01 the commercial software as well. So are original data and then here's after
56:10 by de convolution and you can see it's really improve the spectrum mm original
56:19 bandwidth. And then another one from uh original and after *** and
56:28 So here's his, there's top of reservoir and his face of the reservoir
56:33 . OK. Here you can't see top and the bottom there, you
56:41 , right? So a best practice we're assuming spar spike, well,
56:47 the earth really isn't a bunch of reflectors. So, in a
56:53 that's not a bad thing to do Oklahoma where we got mostly paleozoic rocks
56:59 are highly urd. So we'll have a sandstone layer and a shale layer
57:04 a carbonate layer, a sandstone layer with near constant impedances. Uh Here
57:11 the coastal plain of the United States the Gulf coast, um You're gonna
57:17 much more continuous reflectors and the sparse assumption may not work. We actually
57:24 have a new reflection coefficient every 1 man. It's possible. So you
57:29 to validate your assumption and the way validate it is using a synthetic with
57:34 Well. So here is the original on the right and then the uh
57:40 holding this way as a microphone Um And then here is the
57:47 a good tie. I think he's a AP wave velocity and the density
57:53 then here's his reflectivity and then he's to do the same thing then after
57:59 spike de convolution. So here is data uh to this trace. It's
58:05 deviated. Well, so here are data, here is his synthetic,
58:11 had to have a different wavelength high wavelet to generate the synthetic. And
58:17 can see, oh, that looks good. So this is the way
58:20 say what I'm doing is valid. it didn't tie, if you weren't
58:25 to tie, then it means you the wrong assumption. And this farce
58:29 is not what you wanna do. ? And then you can go a
58:35 further and take your original data and impedance inversion uh and generate uh the
58:43 from you'll get uh uh basically density velocity. OK? And then after
58:54 , dark sparse spike de convolution Jersey's his reservoir area right there. And
59:03 you really couldn't see it at OK? Now, sometimes you have
59:09 throw away data and this is by fellow called uh Bob Hardin and he
59:18 at the Texas Geological Survey and he's data that goes from 80 to 80
59:27 . And he says, well, got a lot of shallow reverberations,
59:33 statics. Let me throw away everything 16 Hertz and above. So he's
59:41 away almost all of his data. now he starts to see some
59:48 Ok? So his data are so . He has to throw away the
59:51 frequencies and this is the kind of you might do in patrol. If
59:56 find all, I'm just contaminated with hyper noise. Let me cut it
60:01 to see what I have. And he starts picking his phone. There's
60:07 example, 8088 to 60 and he a bunch of faults. No seismic
60:21 might range from 10 to 100 We don't measure typically data down to
60:33 and five Hertz. Hm. So I want to do acoustic competence,
60:39 need to generate that impedance from other . So the typical measurement is I'm
60:47 have a density log and a velocity in a well. And then I
60:53 uh Ken Wells in my survey, going to pick horizons from the seismic
61:03 and then I'm going to assume, , I've got similar with allergies in
61:08 of those formations and I'm going to the impedances measured at the well structurally
61:20 along those formations. OK. And gonna do it uh kind of sort
61:29 inversely proportional to how far away I from a well. So the closer
61:34 am to a well, that's gonna weighted more and the further away I
61:38 from a well, that'll be weighted . We can go one step beyond
61:44 . And Krieg Krig named after a African professor named Creek and for
61:53 And then they'll actually look at the and say, well, how many
61:59 are within uh 1000 m from the ? How many wells are within
62:05 how many wells are within 3000 And then they'll generate statistics. So
62:09 does that mean? And the expectation the standard deviation according to the
62:16 And they'll use that to interpolate a better. That's a different course.
62:21 . So you get those background but patrol happens to use Krey to interpolate
62:28 . OK. And they call it . Now, if I have no
62:34 above 100 Hertz from the seismic. still have the well logs,
62:40 that might go every 1 ft, half a foot. But let's say
62:45 can use that well, log At least at the well,
62:49 I could find the impedances at every milliseconds in time. Ok. At
62:59 Hertz. Even though I don't measure , but only at the well,
63:03 what I'm going to do, I'm look at the statistics. What is
63:07 probability if I have? Let's make real simple, make it with our
63:12 instead of impedance. I've got shale, sandstone, different mixes.
63:22 pick Oklahoma. I've got a, , a 400 ft Merrimack sandstone.
63:28 got an 800 ft, uh, limestone. I've got a 200 ft
63:36 limestone. Ok. So I'm picking formations that I'm familiar with and now
63:40 going down here in the Merrimack. the probability of the next sample above
63:48 below being a merrimack? Oh, , it's 400 ft thick. Pretty
63:52 high. Close to like 99%. about 10 ft away? Oh,
63:59 high. Uh, 95%. 200 ft. Well, it's only
64:05 ft thing. Not close to Ok. And in between, it'll
64:09 in between. So what it's saying , what is the probability of having
64:14 formation at a distance from a known ? Ok. Now I go over
64:22 , I guess at a value at that represents my seismic data.
64:35 Yeah, that, that's one of that, yeah, that limestone,
64:38 kind of fits the seismic data that it fine. And then the next
64:43 , well, once I guessed then the ones next to it,
64:48 not independent, they have to fit probability distribution function that I have for
64:54 well data. So now everything is random anymore. You're kind of trying
65:00 fit the seismic data. But then profiles that you choose have to be
65:09 with the statistics, the GEOS statistics GEO because of geometry, not
65:15 but because of geometry of proximity to your first guess is. So really
65:20 you do and you're gonna find this huh That's crazy. You pick a
65:25 , a random value. It has be representative of your geology. So
65:30 can't pick 50,000 m per second because of your rocks are gonna be that
65:35 . OK? They're gonna be somewhere 1,507,000 m per second. But in
65:40 area, it's gonna be more You're gonna pick something random and then
65:46 gonna look at neighboring points, but have to fit that statistics. And
65:51 you have a candidate geologic profile. going to convolve that with my seismic
66:03 , generate a synthetic, compare that to my real seismic trace. The
66:09 is gonna be 10 to 100 It's not gonna be 10 to 250
66:14 would be 10 to 100 Hertz compare . Is that good? Uh,
66:19 , it's ok. I'll keep it guess. Oh, that really
66:23 Throw that one away. Oh, one's a little better and you can
66:26 your way down a tree to get and closer to a model that fits
66:35 data since seismically and obeys the statistics . Now I go to the next
66:43 . I pick the next trace It's a certain distance from that first
66:49 . Well, the statistic says when look at well, log to well
66:53 to well log at different distances. , if I'm 25 m next to
66:59 first, well, I'm gonna expect correlation. OK? For most
67:07 if I'm 1 500 m away, , for sale, I'm gonna say
67:14 likely to have the same. But it's a dolomite, well, maybe
67:18 50% because I get a lot you know, car stain and other
67:23 . So that gives you kind of flavor for what GEOS statistics does.
67:27 that expands both the uh the vertical the lateral resolution. So here's an
67:37 of a realization. OK. And realization means one model that fits the
67:46 and fits the statistics of the OK. Then here is the average
67:52 100 realizations. So this is the and this is what we get with
67:57 inversion. And here is the standard of all the possible realizations that fit
68:03 statistics and the data. And guess at the, well, the standard
68:10 at zero because it's constrained at well as you're further away the details
68:14 away. So now what do we ? We give these models to the
68:19 engineer and they might simulate the P P 25 p 75 situation. So
68:30 in a uh this is pretty intensive , But you can imagine if I'm
68:35 to put fluid, let's say the is porous. OK. Well,
68:39 everything's gonna work pretty well. And an actual realization. No, the
68:45 is only gonna go so far and the, I have uh a 5
68:52 thick baffle or permeability barrier that I resolve with seismic data could really hurt
69:00 production. Well, I can't tell if it's there, but I can
69:05 you how likely given the data we have that you'll have this problem.
69:10 that will help you predict your Then you can compare an asset in
69:16 field to an asset to another to a third field and to figure
69:19 well, where do we want to the money in terms of risk?
69:23 . All right. I'm not gonna you anything about that, but it's
69:27 you need to, to be familiar . OK. No, this,
69:33 do need to know. I've got geoscientist, geophysicists, love the wedge
69:38 , you're gonna see the wedge from and over and over again. So
69:42 got uh high impede and shale on of a low impede and sand,
69:48 impede and shale. Uh Gulf of Shallow Gulf of Mexico Geology. I'm
69:54 to have a negative reflection coefficient in top. A positive reflection coefficient from
69:59 bottom. So the bottom is gonna trough peak, trough, the top
70:04 gonna p peak trough peak when um milliseconds thick um resolved, I can
70:13 the distance between that trough MSP as go thinner and thinner. Well,
70:24 side low peak of the upper reflector interferes with the main lobe peak of
70:33 lower reflector. Uh to here where have maximum constructive interference and the side
70:41 trough of the bottom reflector constructively interfered with the main lobe trough of the
70:50 reflector. Yeah, there. So trough is biggest. So here I'm
70:59 , I can pick the top and bottom. Here's my tuning thickness where
71:07 construct of your interference is maximum. then on the left, I'm
71:12 And when I'm unresolved, my peak distance is is constant and just the
71:18 changes in the uh if you didn't it yesterday, you'll do it
71:25 One of the buttons you push in attribute world is in is to calculate
71:29 Hilbert transform. And what it does it rotates the data by 90
71:37 OK. And this example, if were, here's my reservoir right in
71:42 middle. If I go in the , I'm picking a zero crossing.
71:46 , that's not very interesting to If I look at the Hilbert
71:50 it rotates data by 90 degrees. my zero crossing becomes uh a peak
71:58 then I can map that amplitude very . OK. So there are some
72:04 of thin bed resolution I've got He's got two positive things. So
72:09 got a positive reflector, another positive , I can pick these two peaks
72:13 great confidence. Then here is our uh criteria that we use. So
72:21 depth, a half a wavelength thick time because we go down and
72:29 we measure twice a quarter wavelength thick time. All right, I
72:34 that's the limits of resolution. And there's this guy Ricker who tried to
72:39 it a little further but our us data usually don't allow you to do
72:43 . So the railway criteria is the one. And then here I'm
72:48 My reflectors are so close together, can't tell uh what's going on.
72:56 . So seismic data are band they limit therefore vertical and lateral
73:02 The loss of high frequencies limits our to resolve thin beds. The loss
73:08 low frequencies limits our ability to estimate actual value of impedance and hence lithology
73:16 ferocity. OK. So yeah, get around that by interpolating from well
73:22 , but that's not the same as a constructive and destructive interference give rise
73:29 tuning frequency phenomena which allow us to layer thicknesses below a quarter wavelength
73:36 And how are we gonna do Well, we're gonna go take a
73:41 . Sorry, we're gonna take a . If I were to map
73:46 I say, oh it's low This is high amplitude, I can
73:50 a map of relative thickness. So I have a horizon slice and I
73:56 some feature of different amplitude meandering I can't tell how thick that channel
74:03 , but I know it's a channel geomorphology and from the lateral change in
74:09 . OK. OK. So time a break. And uh what time
74:17 we gonna go to the lab? ty Howie Joe is gonna come all
74:28 way in? Ah OK. Good. All righty um discard
74:39 OK. Let's take a break we can turn them off. We
74:45 to drink coffee therapy. Is that now? Oh, that's true.
77:31 . Hey, wake up everybody. back. All right. So uh
77:40 is working on the web so we get in and uh we had 11
77:48 by um Javier that I thought was appropriate and it was on slide um
78:02 . And let me just um show one again, show share slide from
78:09 slide. So here is a slide and, and he was asking basically
78:26 was asking, yeah, this is . But hey, I'm down in
78:30 and we're doing exploration. We don't any wells with he wave sonic and
78:40 and density logs in it to generate synthetic to see whether spike in dcom
78:45 method X helped. So, um yeah, well, logs are
78:52 especially if you want to convince an because they see that as ground
78:57 But there are a lot of other you can do as well. And
79:01 uh quote a recent talk by a called Charles Pere who did his phd
79:08 with John Castagna, who's a professor . Uh And he and John
79:14 and several of their affiliated workers uh champions of bandwidth extension. We try
79:22 increase the frequencies beyond what you And they basically say that the process
79:29 done trace by trace independent. If of a sudden on the higher
79:35 if you see features that makes sense that you could not see before,
79:45 you're going to be confident that you're the data and not hurting it.
79:51 an example like on this one that not Javier um Carlos and I were
79:57 about like like this little event down . Well, it's got a little
80:01 on the top, it's a little on the bottom and maybe that's a
80:08 , an in size channel. So I'm gonna take a horizon slice along
80:14 red. Hey, we're, we're good. It only took us
80:21 minutes to figure it out. Uh just came back in. Um So
80:28 I look at a map and extract amplitude or extract some of the attributes
80:36 tuning frequency along that map, and see that this little anomaly here now
80:46 a dendritic channel with branches in it a meandering channel. Well, that's
80:54 seismic noise, that's geology. Then gonna feel OK. This is
80:59 I'm happy with this. I'm gonna this. OK. Now some other
81:06 and this sometimes processes help you get the day. So let's say you
81:15 something that is really hard to It's so noisy. None of your
81:20 pickers or auto trackers work. And is structure where they filtering. So
81:27 I'm gonna be real aggressed and over the data. I'm gonna help some
81:33 so that they're easier to pick in areas. I'm gonna hurt.
81:38 I might obliterate them. I'm gonna that easier to pick event. I've
81:45 my horizon picked. I take the , put it back on the original
81:52 and then the boss says, How did you pick that? That's
81:56 painful. It's oh, I'm just good. I'm just very good.
81:59 very careful. But if that intermediate where you posted the quality of the
82:06 to make something easier to pick, helps you to pick and then you
82:11 control it on the original data, OK. OK. In fact,
82:16 how these auto pickers work. They're cross correlating, seeing which part is
82:22 looking at peaks and troughs, smoothing little bit. They're doing all those
82:27 of things themselves. So that's all as well. You may not want
82:31 use that as your final product on data, but you can use the
82:35 so long as you quality control OK. So, uh Jessica
82:42 do you have any questions out I'll see their pictures. No,
82:50 at the moment. OK. Um . OK. Then the other guys
82:57 see. All right. So let's to um the next section.
83:11 Yes, that was three. Direct hydrocarbon indicators. So this is
83:21 be very basic. Fred Hiltermann is spend a whole class on this.
83:29 ? And go into great detail. right. This is section number four
83:43 uh we'll, we'll use the word I or direct hydrocarbon indicator. So
83:50 this section, I want you to able if I give you a compaction
83:55 for a shale, a water sand a gas sand and a basin,
84:00 I'll probably do that later this And again, next week, I
84:05 you to be able to predict the direct hydrocarbon indicator if any.
84:13 Then I want you to be able recognize bright spots, flat spots,
84:21 spots, they're real like this will your head around, go pick what
84:27 don't see who drill what you don't , like all kind of like a
84:35 of faith and then phase reversals on AM data and then correlate direct hybrid
84:43 indicators to structural control. Hey, a classic curve appropriate for tertiary basins
84:51 the world. Showed it to you little bit yesterday. So as I
84:55 older and deeper, I have more application, more diogenes is a little
85:01 of mechanical compaction. So my gas , so a gas, a sand
85:08 with gas, well, it impedes and notice when I get down to
85:17 level, my gas and shale have same impedes. I'm not going to
85:28 a reflection. I'm going to have dim spot. Ok. Then here
85:37 a water sand and the shale and cross at this location. So the
85:49 sand over a shale, it's refreshing , fishing is negative and a water
85:57 a shale in this area, the is higher in PS it will have
86:03 positive reflection coefficient. Ok. So gonna look at different, we're gonna
86:08 a reservoir and the reservoir has typically structure associated with it. The gas
86:14 part is always gonna be structurally So we're gonna look along a reservoir
86:20 see. Hm, how does the and the polarity of that reflection
86:27 And if it changes, it's indicative there being gas in the system versus
86:33 most of the time it's filled with , right? Uh So those are
86:38 two crossover points. We got dim , we got face changes between the
86:46 and the water filled part of the and we have bright spots, big
86:51 changes. OK. So here's an of an early Gulf of Mexico bright
86:57 . And this is one of those I recommended as a, as a
87:01 . The best thing is go borrow from somebody and maybe not return
87:05 you know, go find somebody my . I've given away all my books
87:08 like three. Uh So see if can find somebody like me and borrow
87:13 book for all eternity. But it's great book. A lot of good
87:18 . So, here is the You see the reservoir. That's the
87:22 . Well, maybe look at there's a nice, easy to pick
87:27 . It's a big dome. There's big gun like, what, what
87:31 heck is this thing? What's Flt Stephanie, what's an FLT technical
87:41 ? Funny working thing? That's what call. Ok. What's this funny
87:46 thing? Ok. So there's my . Here's my bright spot. Very
87:54 amplitude on top. There's a flat . Do we see this all the
88:00 ? No, but when we we're pretty happy. That's the gas
88:04 contact. So the gas has a impedance and the gas sand has a
88:11 being and the water sand has a impute. You might call the
88:18 I, yeah. Yeah, there's gas charged, there's water charged if
88:24 would just go back. Oops. , I wanted to go back so
88:29 you can see. Oh, the on top is significantly stronger than the
88:34 on the side. Ok. um, here is the moves.
88:54 , I stand on 1 ft and my hand on my head. And
88:58 what, what do I need to ? We get more, more?
89:08 . Oh, I can hide the , the floating control. I got
89:16 . Thank you. I'll never be to find them again, but that's
89:21 . All right. So here I'm shallow on this side, the relatively
89:30 , hear them deeper. And here's velocity differences from gas to, let's
89:37 brine, brine is just saltwater. here travel, I've got a big
89:44 like 0 10 to 50% difference in , sand to gas, sand.
89:53 highly sensitive, this is where we the bright spots. So within the
89:56 six or 7000 ft, but even to uh 10, 15,000, we
90:01 still see gas pretty well. And then to look at brine and
90:07 , well, that percent difference, couple percent differences, that's gonna be
90:14 very small reflection coefficient. It would hard to see oil as a direct
90:20 . Ok? You need to use more sophisticated techniques that Fred Hilter and
90:24 talk about. Pardon? Why? , because you've got a lot of
90:31 changes going on. So those reflection might be four and 5%. And
90:37 here I got an oil versus a sand and it's like 2%. It's
90:43 hard to see a 2% change when else is changing four and 5%.
90:48 as simple as that. Ok. here's a, here's a little cartoon
90:55 uh the kind of the notation that Brown uses, he's gonna call a
91:01 refreshing coefficient in red and the negative coefficient in in blue. Uh maybe
91:11 day, I'll take his picture and him look like troughs instead of blue
91:14 . But anyhow, so here in shallow section, shallow part of the
91:20 , the I'm going from high impedance low impedance gas charged. So I
91:28 a strong negative reflection coefficient on the , a strong positive refreshing coefficient on
91:35 bottom. Why? Because I'm going my gas hand back into my shall
91:40 shale. And then here I'm going high to me. So I'm having
91:46 moderately negative reflection coefficient, moderately positive here's my flat spot, moderately positive
91:55 sense. OK. Now let's go little uh quite deep. So the
92:02 one, the dim spot then now have a weak positive reflection over my
92:14 because I have a low impedance in shale, medium impedance in the gas
92:21 , high impedes in the water So here I have a strong
92:26 So I'm going, I see strong , strong reflector then the reflector kind
92:30 goes away. Ok. It becomes . So what do I pick?
92:37 picked a thing that looks less interesting maybe I don't see it all.
92:43 then on the base, same pattern opposite and polarity. So have negative
92:49 strong negative on the bottom, and then I'll have my gas and
92:57 sand, uh positive reflecting coal Then in the, in between death
93:02 the phase changes. So here I a weaker reflection of going from BD
93:11 to low MP D and then I'm from medium pinch to high. So
93:18 see, I go from trough, to peak trough. So here is
93:27 gas part of the reservoir. Here my water part of the reservoir.
93:32 the polarity has changed. OK? then the uh I have a strong
93:39 reflection again for the flat spot. right. So again, looking,
93:48 my shale impedance. Let's use that a constant. I showed you as
93:53 green curve earlier. It's increasing I'm showing you with respect to the
93:58 curve. OK? Then the bright , if I add uh gas,
94:06 gonna go from more gas, it's become brighter and brighter. OK?
94:12 the dim spot, more gas, gonna become weaker and weaker.
94:19 And then the phase change, uh gas I'm gonna change phase.
94:27 he's got a bunch of examples collected the years. He's got red is
94:33 blue is negative. Remember we have keep track. He has a whole
94:36 in his book on polarity. Oh . So you know a little bit
94:43 seismic acquisition. I got a P coming down from below hits the surface
94:52 the earth. OK. What's the of the earth going to do when
94:59 wave comes up gonna, is it move up and down or is it
95:04 be constant? Got a wave coming ? No, it's gonna, it's
95:12 move up and down. That's why not buildings down. But anyhow,
95:17 that wave is gonna come up. I measure the velocity, then
95:24 the boundary conditions, you may remember conditions from differential equations. You guys
95:31 to do differential equations. You you didn't. Oh OK.
95:38 I know we're gonna follow GD depth equation anyhow. So you have boundary
95:43 . So you have one that says , you can't move, it's rigid
95:46 you have another, it's free. the earth's surface in terms of velocity
95:51 velocity is free. So I have P wave of a certain amplitude coming
95:56 and then it goes down. And the amplitude of the surface is two
96:01 as much as the incident wave because have the amplitude of the upcoming and
96:05 down going in together in a marine , the pressure on the surface is
96:14 . So a P wave coming let's say positive reflection. It the
96:19 goes down, in order to have pressure, the pressure of the P
96:26 going down has to have negative OK. So if my earphones at
96:31 hydrophones at the top, I'm gonna nothing. So what I have to
96:35 is put the hydrophone down 5 m a streamer. And then I measure
96:40 up and down going wave, both them, the first wave and then
96:43 call the second one a ghost like old style television systems. All
96:50 So now in the United States, started a cornering seismic data in the
96:56 twenties in Texas and Oklahoma places like . So we use earphones,
97:05 He wave positive he wave coming gonna be positive reflection coefficient. We're
97:12 call her positive red and negative blue positive question. Now, shell
97:20 Well, at that time, there much oil in the Netherlands. They
97:25 working Indonesia, which was then uh East Indies or whatever. And so
97:34 were working marines. So they had use microphones and then they got a
97:39 wave coming up, negative wave going . They're measuring the pressure well,
97:44 to make everything look right for a refreshing coefficient. Their polarity on their
97:50 was just the opposite because they're showing . OK? Instead of particle
97:58 So we have different conventions. We like a European convention, an American
98:05 . Oh, and this guy Alistair , he's from Australia. So he
98:08 , well, we got the Australian by cranking. OK. So they
98:13 that. And then of course, BP buys Amoco and Arco and Exxon
98:21 a bunch of European companies and the buy everybody. Everything's all mixed up
98:27 . So nobody knows anything. So we'll have all our whales are
98:32 in feet. And then many of surveys in the Gulf of Mexico are
98:37 meters. There's lots of ways of your eye out. And the same
98:41 is true. When you trade data company to company to company, you
98:45 to know what the polarity is. . Sometimes they'll say North American
98:53 sometimes they'll say European polarity. Uh the best way is to look at
99:00 have a well log. That's the far the best in the absence of
99:04 well log. You gotta look for hard reflector. It's not the water
99:08 because the water bottle can be very soft. Uh If you got
99:13 carbonate someplace, the Gulf of go sail out to garden banks 100
99:18 10 miles off of uh Houston northern most northern carbonate reefs in the
99:27 . Run your geophones over there. . All my polarities are fine.
99:32 serious. This is what the service do. Uh make sure.
99:38 So here is a, here's my , I can see the flat
99:43 I got a bright spot. So my strong negative reflection. There's my
99:52 spot. My first one and then is the a second reservoir. So
100:00 I get side lobe interference there. right. And so I've got one
100:09 , two reservoirs, a flat spot , another flat spot there.
100:15 And then here's his map of the spot, another example of bright
100:21 So he's got kind of the brownish are negative. OK. So here's
100:26 one, number two, number number four. And you see how
100:32 are tilted up against the salt. really if I were, if let's
100:38 , look at this event, see this is kind of strong. And
100:42 over here it's weak. So here strong and over there it's weak.
100:48 yeah, you might look at the thing and say, oh those things
100:50 jumping out. But what you really do is look at the hypothesized
100:57 How does it change along the OK. Then um this is the
101:05 data. Yeah, same data. then he used the equipped color
101:12 Uh So he's using Cyan for the yellow for the negative. Uh And
101:18 says you gotta be careful when you your data. Now we loaded our
101:22 as eight bit that's dangerous. So using eight bit because it's gonna go
101:29 . If your survey is not too , you want to go to this
101:34 bit and then a safe thing to in between is voted at 16
101:41 So wait, wait, do you what I mean? By eight
101:47 16 bit, 32 bit 5, , all I hear is the
102:00 eight bit, what's eight bit Uh, yeah. What do we
102:07 when we store it as a big ? The? Mhm. Ok.
102:16 affects resolution but what's this, what we do in the computer, do
102:20 know? Oh, ok. That's . How about a knob? Do
102:24 know what? 858? All So all of our computers that we're
102:30 today are 32 bit data. Now separate supercomputers that use 64 bit.
102:36 . So eight bits, four bites a bit. Eight bitch. Um
102:47 as an integer, we could represent as big as two to the
102:52 Now, going back to high when you learn scientific notation, you
103:01 that? Ok. I'm gonna have mancha like I don't know, 4.306
103:11 10 to the 20th power. Maybe the US national debt today.
103:17 Something like that big number. So we have gonna have so many
103:21 are gonna use for the Mantis. I'd say about 24 and then,
103:26 , if I wanna go on my , I think I go to e
103:30 the cluster minus 32. That's five . So I have five bits for
103:35 exponent. OK? Either the either the zero up to 32.
103:41 ? 2 to 2 to 5 powers . 0, I can be plus
103:45 minus either to 32 either to minus . So I got six picks for
103:49 exponent that leaves how many bits? minus 626. 0, I need
103:56 bit for plus or minus sign of Mantissa. And so 25 bits represent
104:05 Mantissa of the data, the value the Mantissa. Ok. And it
104:09 me about six digits of that. see six digits isn't the same as
104:21 you could store the data as an . And uh if you use integer
104:26 instead. Ok. Now for a data, if I read in 32
104:33 data and I read in eight bit , it takes me four times as
104:37 to read in 32 bit data. takes four times as much this
104:42 Ok? So let me take my . I'm gonna pick am and the
104:47 I think I sat with Jessica yesterday maybe Stephanie and said pick minus 25,000
104:53 plus 25,000 is the range. What going to do is then divide that
105:00 by 255 and all the numbers that so 25,000 by 255. 0,
105:12 ok. They're gonna be in bins about 1000. So the values between
105:20 they're gonna be scale. We're not store the number 13,268 as 13,000,
105:30 we're gonna store as 13,000, we've it away. Ok. Now I
105:36 said, so there's a, there's trade off. If you're using eight
105:41 data, anything that's greater than 25,000 minus 25,000 is quipped. So if
105:49 have 40,000 as an amplitude, it's to 25,000 means if I'm looking at
105:55 spots that are greater than 25,000, might miss him or they might be
106:02 in with other strong things. So a problem with eight bit data.
106:07 I make my range too big, lose my resolution for the weaker
106:14 If I make my range too I quit my strong positive and negative
106:20 and make the bright spots funny So you gotta be careful with the
106:25 loading. If you can load it 32 bit, that's what you should
106:31 . Um But the date is not show up as fast. OK.
106:37 let's look uh here he's in He's got another bright spot here is
106:41 , it's blue. Notice he's always the color bar there. That's what
106:45 need to do. So I've got negative, positive and then here he's
106:51 some kind of salt film and here's bright spot right there. OK.
106:57 I see a flat spot. They have some yeah bright spot, black
107:02 . Yeah. Doing it. Water sand. See how that reservoir
107:10 . So, you know, if a normal person or a processing kind
107:14 person like me, yeah, you're see the bright spot, but you're
107:18 forget about the other part of the . All right, because it doesn't
107:22 out. But really you need to at that other part of the reservoir
107:26 that helps you define the model. doesn't mean it's an oyster bed.
107:34 here in the Gulf of Mexico, go down the, uh, you
107:41 the ferry from Galveston Island to Bolivar and then you walk along the beach
107:51 Bolivar Peninsula by the White House. the White House still there or did
107:55 get blown down? Either you guys have a life or the lighthouse is
108:02 longer there. Ok. It's a lighthouse. It's pretty cool.
108:08 Take your two year old on the to Bolivar Peninsula once you have a
108:15 . Oh, because it's three hours for you. All right. All
108:19 . Sorry. Another conversation anyhow. , you go to the Boulevard Peninsula
108:24 you walk along the beach there where no beach. It's just oyster
108:30 Ok. Now, think of burying . That's gonna have a very,
108:34 strong reflection postage. It'll happen to a positive reflection co instead of a
108:40 , that's one indicator. But it's really not a reservoir that's gas
108:46 . It's, it's basically a layer by itself. Ok? So being
108:50 see the map out the whole formation part of identifying bright spots and the
108:58 chimney. OK. I got you . OK. So the the question
109:12 is that are bright spots always negative, positive. OK.
109:20 you gotta worry about the polarity, let's say North American polarity. Um
109:24 it's a bright spot, let's put in context. If it's a ga
109:31 it's a hydro direct hydrocarbon indicator, is always positive, negative, positive
109:38 zero phase. OK? If it a, I might have an image
109:45 Brazil. If I have a igneous a very, very high impedance,
109:55 will be bright but it'll be positive, negative, it'll have opposite
110:03 . So I think it depends whether mean the bright spot as a buzzword
110:12 means, oh yeah, definitely. I or if you're saying a just
110:19 descriptive word like normal English or it's that's stronger than anything else. So
110:24 think I would be careful. So would call it a direct hydrocarbon
110:29 But uh I think most people will bright spot, that's what they
110:33 And the other thing that's bright. , you wanna be specific. I
110:44 pretty sure I have an example. not, I'll be able to find
110:47 when we take a break. So here's another example uh from the
110:54 Sea eco fe field. So that's chalk creases. Um They've got
111:02 a gas chimney in here. So um there's um falls down deeper and
111:10 happens if you have a, a , if your gas is leaking,
111:14 slows down the velocity and the velocity um when the velocity decreased.
111:21 uh Howie Joe talked about migration, probably showed, I hope he showed
111:26 bad migration. Did he show you ones or just good ones like a
111:32 with a long velocity? Did he you any of those? Ok.
111:38 in this case, hey, there good reflectors in here. In
111:41 you see how this guy kind of down? Well, that's strange.
111:46 is this dome going down? this gas is leaking from a reservoir
111:50 here. It's slowing the velocities down I get a pushdown effect.
111:56 The two way travel time is greater go through a slower velocity then because
112:04 doing my velocity analysis every one OK. I've got good velocity analysis
112:13 I nice have, have nice coherent and I simply didn't do a velocity
112:19 here. So I'm going to interpolate high velocities on the planks into the
112:27 because there was not either. I pick this because he didn't pay me
112:32 pick it. But they were a leader, right? Did they use
112:37 word? Cost leader? That means ? OK. Uh They're gonna pay
112:43 to do every one kilometer grid velocity . So they missed it or when
112:49 interpreter or the processor looked at it I can't pick anything there. It's
112:53 complicated and so they picked the wrong in this gas chimney. The energy
112:59 not really, it could, some it is absorbed. Uh, but
113:03 lot of it is because the velocities wrong. Ok. All right.
113:10 one's pretty cool. This is in North Sea as well. And it's
113:14 big flat spot. Anybody see the spot there, Hayden, you shoot
113:19 flat spot, right? It rains cutting right across Strat gray. That's
113:34 funny. Ok. Now you can Strat democratic effects like that due to
113:41 like in Wyoming and Montana, there be a digenetic change to anchorite,
113:54 type of a clay that forms a and you will have something they call
114:02 centered gas. Why? Because all a sudden I have this flat seal
114:07 the bottom part of the basin and gas gets stuck in there. So
114:11 I'm gonna drill the middle of the , like, come on. Now
114:14 gonna drill a sin co that's what do. So you'll see papers on
114:19 center gas and there they will have diet, genetic event cutting across.
114:25 this one's a flat desktop and nine long, pretty big cuts across the
114:31 and everything. There's another one from um bright spot on the top,
114:42 spot on the bottom kind of hard see where the reservoir goes beyond outside
114:50 . But you can kind of guess it is by using this event above
114:52 as a guide. Thanks. And uh here's the bright spot and a
115:02 spot. Now, hold on So, Carlos, why does this
115:08 spot stop and then come up Any idea? I got him.
115:37 . Yeah. So what, what saying? And I'll have to point
115:40 it on the, the screen Yes. Here is the base of
115:46 reservoir or I'm gonna ah darn Here is the, here is the
115:53 of the reservoir coming up and well, I don't have any reservoir
116:01 . OK. So my gas water contact is here and then I my
116:08 base of the reservoir. Here it here. This is probably the base
116:11 it's red. OK? And then the reservoir goes down again. So
116:17 simply missing the reservoir in the And that's why I don't have the
116:22 water content kind of obvious when you about it. But at first you
116:26 , why, why is my flat not going across then here he's got
116:31 map view on the side. Now, dim spots. So here
116:37 gotta pick the events that you don't well. And uh if you're gonna
116:44 at any of the papers, this the paper you wanna look. So
116:48 uploaded a bunch of the papers were the best paper in 2016. And
116:53 really clear definition of bright spots and spots. OK. So here's my
117:03 . I got a positive reflection coefficient up, going up. Oh,
117:09 look above the reservoir reflection going So I'm kind of doming across.
117:14 right. And then yeah, there's stuff going across. So what happened
117:20 ? I go from a good reflector nothing. OK. That's a dim
117:30 . This one's a little bit easier see. Here's my nice oh,
117:36 I've changed polarity. I got blue positive red for negative. Uh These
117:43 are from Shell. OK. So a European company even though the work
117:48 done here in Houston. Let's see things get complicated, right? Um
117:57 a strong positive reflection and then it gets weaker as you go to
118:05 high and then there's a fall. here I got a positive negative and
118:15 here is the negative, positive. . Oh God, a weak positive
118:24 on the one side, strong positive on the other with negative reflection,
118:30 positive. So these are I got dim spot on the lower one.
118:36 . Here's a phase reversal uh from Brown's book is using yeah positive positive
118:49 for shale over water sand than a reflection or jail over gas sand.
119:02 this is at an intermediate depth where I'm going positive, negative, positive
119:12 then here I'm sorry, negative, negative and now it's positive, negative
119:19 . Now how do you figure out event is, what? Well,
119:25 start picking events nearby that don't have of this in there and you can
119:30 on them. Use those as a . I can use this one down
119:35 as a reference. And then, know, oh, there's not a
119:38 here. There's something else going Ok. There was a wet sand
119:44 a guessing with a polarity reversed. one, maybe you can see this
119:51 reversal here, I got a polarity . And then here is my,
120:00 going blue and here is red. a gas charged uh volcano uh Caspian
120:16 . So probably Azerbaijan, OK. got positive and blue here. So
120:22 from what we were looking at so I got a a positive reflection
120:28 and there, here there's gas So we've got a negative reflection
120:35 OK. Then down here, positive , oh where the reversal normalcy for
120:45 with a positive regression coefficient. But there's gas at the top of this
120:50 volcano. There's a gas on a diaper. Yeah, Malaysia. This
120:59 be called a mass transport complex in , underwater landslide shale diap here coming
121:06 kind of like a salt dye but . And then we've seen very strong
121:12 here change of polarity flat spot here got and now in this one,
121:21 got negative as, as red, background, positive reflection polarity reversal.
121:31 water charge gas charged black box be if they all look this clean.
121:41 , they usually don't. But uh then on a horizon slices, uh
121:48 can map the tity. So here got a negative reflection coefficient with a
121:55 charged channel looks like a channel, the channel. And uh and it's
122:01 long horizon. And then here's a view I amplitude, negative clarity,
122:11 amplitude deeper in the section. And then here's what some of these
122:21 look like at the pinch outs. ? So here's a flat spot on
122:28 map of horizon. There's the pinch , there's kind of the top of
122:36 reservoir to the reservoir. OK. this is the question that I think
122:45 had not all strong reflection because So polarity is important and here is
122:54 that's gas that a negative event and that one probably an oyster band.
123:03 this is Gulf of Mexico data. positive reflection at the top of
123:09 that here's oyster beds in South Carolina the uh the both of our peninsula
123:18 just like that except it doesn't have promet. OK. Here's another one
123:24 um gas, so blue, blue, so negative event. And
123:34 this one is red, blue, , positive event. This was a
123:39 hole. This was a good, , you can see people make
123:42 right? I mean, they they drove them both thinking they were
123:46 good potential. OK. So here's picture again, depth of maximum burial
123:59 , younger, shallower rocks, you're see bright spots, intermediate depths in
124:04 , you're gonna see face. Can just and deep like the deep water
124:09 ? You're gonna see dim spot. . We gotta worry about gas
124:16 Um, for two reasons. Let's if Jessica is there, Jessica,
124:24 do we have to worry about gas ? Why do we care? See
124:31 she's still there? I'm here. . Why do, why do we
124:37 about gas hydrate? Um, I know, let's say you're Japanese.
124:48 do you care? You're not No. Ok. Why would the
125:00 care about gas hydrates? Oh, a gas hydrate? What's the gas
125:15 ? Any idea you can say? , that's fine. A knob.
125:21 know what a gas hydrate is? know what a gas hydrate is?
125:29 actually frozen. It's actually ice. down at the bottom of the
125:36 Ok. So now I understand why guys don't care. You need to
125:39 . Ok. Start caring. I'm at them, Jessica telling them to
125:45 . Ok. So it's gonna be the, at the base of deep
125:52 , the temperatures maybe four °C, pressure and the uh methane and water
126:04 to form and ice. So down deeper where, where um uh showing
126:12 mouse moving here here, the gas is very, very stable.
126:20 And then as it comes shallower, becomes unstable and can evaporate and what
126:25 does is it, it puts methane the atmosphere pure and simple.
126:32 So here's the permafrost area and then this permafrost has somehow dissolved a bit
126:41 melted and the gas gas hydrate comes and generates a big pock mark.
126:47 . So here is nice and stable . It's not. Now if the
126:52 warms or if the Arctic Northern Siberia, Northern Alaska Lapland, if
127:03 warm all of this permafrost contains a deal of gas hydrate and a great
127:08 of methane. And it will go the atmosphere. Methane is like 40
127:16 more uh impactful than carbon dioxide on atmosphere. I might be wrong with
127:22 number, might be a bigger Uh But, and it's gonna warm
127:28 the climate more. We'll have more hydrates come out and we run away
127:34 get out of control. Ok. that's why uh that's why um you
127:42 to be worried as an environmentalist now the Japanese, they don't have a
127:51 of oil and gas. So you actually mine the gas hydrate and have
127:58 and not have to import it from . So it's actually a resource that
128:04 be capped. Yeah. So here's example of, of seismic data for
128:11 gas hydrate trap which is kind of you all those, these right
128:18 Well, if you look at it detail, I think you'll see looks
128:22 a meandering channel system, right? charged with high amplitude reflection. Here's
128:28 meandering channel system. This little well, it's not filled with this
128:33 amplitude reflectivity. Here's what it looks on vertical seismic. There's a piece
128:42 gas hydrate and it's kind of it's ice but it, you can
128:47 it. Ok. And then the evaporates as you burn it. So
128:54 , um, here's my bottom reflection then in the gas hydrate community,
129:02 talk about the bottom stimulating reflection. BS R bottom simulating reflection, it
129:12 of looks like a multiple but it's, it's not because the multiple
129:17 be double the distance from the water to the water bottom. Rather it
129:26 a pressure, pressure, temperature stability . And below that, we have
129:36 above that, we have the hydrates below we have free gas. So
129:40 we have the free gas below, what? I got a flat
129:43 I got another flat spot. So here, I have free grass
129:47 I have a gas hydrate. When drilling through the gas hydrates, you
129:51 be careful. You don't mess up stability because well, your drilling rig
129:56 collapse. You might have some serious , but they gotta be very careful
130:00 that. Uh And here they're just 3D visualization, which you'll do in
130:05 web. OK. Another gas height this one from the Caribbean, I
130:11 my reflection coefficients on the side. my bottom simulating reflectors. See how
130:18 again, one of these things that across stratigraphy. In this case,
130:24 a digenetic change. OK. So rare to have events. Normally when
130:30 have an event cutting stratigraphy, it's multiple some migration artifacts, some kind
130:37 seismic noise, but you can have , genetic events that cut across photography
130:44 of course, the flat spots like showed earlier, there's some free gas
130:53 , maybe more free gas, maybe free gas, maybe more free
130:57 Ok. Gas chimneys, I guess this is the picture. I have
131:02 take the other one out and I about cash. So Alistair Brown has
131:07 checklist to validate the presence of It's a long long list. I'm
131:13 gonna read them, but you're gonna at it because whether you're working in
131:19 sequestration guess CO2 is uh oh, a gas. It's gonna look just
131:27 methane. So once you start injecting , everything I said about bright spots
131:33 for methane or Co2 storage. Uh you're gonna store hydrogen, everything I
131:39 about methane is gonna hold for hydrogen . A geothermal, a little different
131:47 . You're gonna have maybe water you might see some of that.
131:53 ok. So in summary, seismic can be quite sensitive to the presence
131:58 gas giving rise to direct hydrocarbon These uh DH is change as the
132:05 of burial changes due to different compaction sands and shales. That compaction di
132:12 know, gratification. OK. Bright analysis requires amplitude friendly data processing and
132:20 scaling prior to interpretation. I would today, all of your data are
132:28 because everybody is doing bright spots. a bo I would say all of
132:32 modern surveys of the last 20 the process trying as best they can
132:38 preserve the empathy. You go later earlier than 20 years. Well,
132:45 people would take every trade, try like they're easier to see and
132:51 Well, then you're gonna lose these effects, ehis are often correlated to
132:58 such as the flat spot, providing confidence in the interpretation. And I
133:03 you the one example of the channel to structure and uh and geology.
133:11 . So any questions comments on Yes, ma'am. I gotta walk
133:24 . Oh, the gas chimney. , the so and the, the
133:30 chimney, I think I talked about picture earlier. That's why I skipped
133:34 . And um, so what we is uh we have gassed.
133:43 there's two hypotheses about gas chimneys and Thompson who's gonna do maybe the next
133:50 , right? Or two classes from . Thanks doctor. Thank you.
134:00 , Fred Hill, come in. . Well, Leon will be next
134:03 then or we. Ok. So two theories of uh what happens with
134:10 . Why we see a gas The first theory is that the gas
134:18 coming up from below, very often a producing reservoir? OK. And
134:26 we have several ways. Well, , I'll make my, oh,
134:32 is terrible. But professors do this the time, right? You guys
134:35 used to this? All right. I'll take your, your an my
134:39 and make it into a question for . OK. How, how can
134:43 attenuate seismic data? How, what do you know to attenuate seismic
134:50 Yeah. How would you attenuate Absorption? How would you absorb
134:54 What's the, what's the physics? the mechanism? OK. Well,
135:00 where you say, hey Hayden, do we attenuate seismic data?
135:08 Let's pick on a, a second . How do we attenuate seismic
135:16 Pardon the energy, like I like kind of like, well,
135:23 it's elastic, you're not attenuating at , you're just reflecting some and
135:32 Bob Brilliant. Hard right now. . Well, you're gonna have geometric
135:41 . But how does, well, ? Yeah, it's gonna, it's
135:46 to decay in amplitude, but that's that's a geometric attenuation and that's easy
135:54 compensate for. So Zach's not looking me in the eyeball. So I'll
135:59 him, yo, Zach, I , I really, yeah, tell
136:06 everything you know about attenuation first. this uh you mess with the frequency
136:17 that up there? OK. So attenuation, you're typically gonna lose the
136:21 frequencies more than the lower frequency. right. So who's our far distant
136:30 ? Again, we got, you that? Yeah, hang on,
136:37 on. I'm just looking. I've forgotten everybody. I'm picking on
136:41 and I wanna, I wanna give Jessica. Yeah, Jessica, my
136:45 name is Jessica. That's why I . No, it can't be
136:49 How do we have attenuation? Any ? You, you've been looking at
136:54 web while everybody else is scratching their . I have not. Um,
137:00 I can. Ok. All So here, uh, I'm looking
137:05 the ceiling now in the in office , we've got ceiling tile. So
137:12 gonna stand up and put his hand the ceiling tile because he's really,
137:19 darn who's taller than Bob here? Guia. Oh My. OK.
137:27 it's rough. It's rough. It's GS, right? Do you see
137:35 this wall? Is it smooth? , very rough? OK.
137:42 what's the, so what's the wavelength sound that you hear? It's about
137:50 size of your ear? It's on scale of your ear. If you're
137:53 elephant, you can hear much longer . OK. But we tend to
137:58 wavelengths on the order of a, know, five centimeters to quarter
138:03 So when I have a reflection from smooth surface, I'm going to have
138:10 reflection angle of incidence equals angle of reflection coefficient wavelength change. Nothing changes
138:16 the wavel form exactly the same, a little different amplitude. When I
138:21 here on my, on this rough that I'm hitting, you hear me
138:26 the rough wall. OK. Hopefully does. And uh then the,
138:33 he hit this rough wall, now am going to reflect the different angles
138:39 on the local uh dip or slope the reflectors gonna go in all crazy
138:47 and the higher frequencies, the shorter are gonna scatter more than the lower
138:55 . So what this is doing, not mechanically absorbing the energy, it's
139:01 it incoherently. OK. So you're see this in seismic data. If
139:08 have a nice smooth limestone reflector, reflections are gonna be broadband high
139:16 If I have a rugose dolomite it's going to be the low frequencies
139:23 reflect fine because they'll constructively interfere as come up and the high frequencies will
139:30 interfere. OK. So that's a scattering. That's one way of
139:40 It's diffraction going into all these little . And whenever we go through very
139:47 uh boundaries, usually we lose frequencies we go deeper. Not because the
139:54 is gone. It's just because the is so incoherent and we don't have
139:59 detailed velocity model to put it back again. OK. Then we have
140:05 squirt mechanism. OK? You've never of that. OK. Fred will
140:12 talk about it but in the squirt , I gotta have it. So
140:16 concede my, my. So we , got a P wave. So
140:22 have cores and in the pores, of the pores are filled with gas
140:25 some are filled with water. So I'm gonna compress the pore with
140:31 water in it. And then that is gonna go through the throat into
140:37 pore filled with gas. Ok? then as the wave rarefies, the
140:44 wave rarefies goes negative. Then the in the one or going back through
140:52 throat and then back into the original . So it isn't going through the
141:00 . You have the friction and you're , converting mechanical energy of the P
141:06 into heat energy. OK. So the squirt mechanisms. That's the most
141:10 one. Now, we have another is deep in the earth like in
141:15 basement where we don't have much water there the attenuation is much less.
141:22 . Reflectors are less too but But so think of a P plagioclase
141:30 . So Hayden tell me everything you about plagioclase. Why could you tell
141:34 something that's ST distinct about it a time ago. It's longer for
141:43 Trust me. Do you remember anything twinning? OK. Say something profound
141:52 20. OK. All right. we've got a crystal structure, there's
142:00 in the crystal structure and they're called features or pals P er LS stress
142:10 . And you can actually in the , we because the earth is mostly
142:14 , you can actually take one That's what they call them in crystallography
142:19 move it. OK. So we move the dislocation along the crystal
142:25 So we would, for pla we would, if we had enough
142:30 , confining pressure and a strong enough , it would actually move those twin
142:35 locations by, you know, three four atoms. And when you do
142:41 convert the heat energy, so that's most common way of changing uh of
142:49 in crystalline rock. And then there other ones with phase changes. So
142:54 can actually maybe under proper temperature and . So, you know, a
142:59 of our gas is not, it's of in a super critical. So
143:06 , it's really a fluid down Well, if I change the pressure
143:10 putting a P wave across it, , I can go from dashes to
143:14 to fluid stage, dashes to fluid to lose energy. That way,
143:18 one's kind of a minor one, ones, the throat. OK.
143:22 I've got, got mostly water on sides. I got gas in
143:29 A P wave comes across doing all poor throat stuff losing energy.
143:37 So it is actually you could, will say it's absorbed and then sheer
143:44 are not sensitive to fluid. They feel fluid at all, they just
143:49 the rock matrix. So if I a P wave down and the sheer
143:53 up, I'm gonna get a great and Leon Thompson who talked to this
143:58 later later in the year. Uh an expert on that. You
144:02 one of his claims to fame is wave imaging. And then the other
144:07 I talked about earlier is, maybe I don't have diffused gas,
144:13 I have a lot of little sand , 1 ft 2 ft thick.
144:17 they're filled with gas and that's just my velocity so much that I have
144:21 bad image. But there's the gas question from le Yes sir.
144:30 Uh How can we like? Meaning example, for high? Ok.
144:41 . So Carlos is asking how can differentiate between a gas chimney like
144:46 which I'm saying is due to gas the system, a gas cloud and
144:51 that's so faulted that uh that we the same piece of garbage image.
145:01 . And um I'd say that's a . That's a fortunately it's carlos'
145:08 not my problems, but uh you know a bad image the
145:13 So that first example I said with the little gas charge channels. So
145:18 I was at Amaco, we had guy Svea Dahlberg. Uh uh
145:25 he's now chief, chief physicist at here in Houston. Um but he
145:32 had the well logs over eco field then he did a model and he
145:39 , OK, I'm gonna fill all little sands with gas that changes the
145:43 by maybe 10%. And then I'm generate a synthetic, then I'm gonna
145:49 uh a preset synthetic using a wave . Then I'm gonna migrate it using
145:55 velocities we normally pick and he got image just like that. OK.
146:00 bad velocity and this is your problem the faults. The faults aren't a
146:06 is that you don't have the right for the layers around that fault to
146:11 a good image. It's usually the or you may not have far enough
146:18 to get the diffraction. You need image those faults. So,
146:24 so that's ambiguous. That's hard. hard. OK. So gosh,
146:29 20 to 12. We're gonna take at noon, right? Is that
146:33 we've been doing? You guys have doing? Where do you go to
146:41 ? You go to where you Well, oh Moody, I've been
146:46 . Yeah. If I wear a shirt, do I get half
146:51 Oh damn. Last time I came a Friday, I was wearing my
146:55 ou shirt and I got half Ok. Uh So we'll go there
147:03 let me just do the next one uh for 20 minutes and then we'll
147:07 back to the lab. Is the gonna be open? Ok. So
147:12 see the next lecture then just keep along. I appreciate the questions,
147:21 . I'm gonna skip that at, skipped two B and this is easy
147:32 understand. I'm gonna go through So two D they're kind of the
147:39 you're gonna generate in the lab. White settings type of attribute display.
147:54 wish the book, here's a vertical through the seismic data happens to be
148:01 rendered amplitude and dip as mute uh actually reflect your convergence. OK.
148:09 a time slice. So at time , it just constant time through the
148:14 volume could be two data volumes could in this case, three data
148:18 OK? And most of you looked times slices yesterday. So time was
148:24 time depth slice, constant depth here a time structure horizon map kind of
148:31 product you would generate. OK? it's got structure on it and we
148:35 use shaded illumination to make it look . OK? Then we got attributes
148:42 from a picture horizon. So I my picture horizon. First thing I
148:46 is a time structure map like I you here and then when you pick
148:54 can make a difference, I'm not make it a big deal with you
148:59 . You're gonna pick what you think easiest. And this image I've got
149:06 in white noise and yellow and then you actually measure in magenta. Let's
149:14 I'm gonna pick a peek. Here's pee, here's the true peak.
149:22 ? Because when I have my peak , it's kind of flat. It's
149:25 changing rapidly. The noises go ahead changing like crazy. So it shifts
149:32 peak. Well, what about a ? Well, the signal is changing
149:37 at its trough. The noise is however, it wants to. So
149:43 shifted. How about a zero The signals changing like crazy. The
149:48 changing like at once, ah the crossings are gonna have less of an
149:53 then the peak in the trough. . Here's an example. My buddy
150:00 up in Calgary had picked, he a trough in Cyanne, uh zero
150:08 in yellow and a peak in And here's just looking at amplitude uh
150:15 , OK. Uh Just the amplitude it. And uh here you see
150:20 nice and smooth. This is the sorry, this is the time structure
150:25 . So nice and smooth, rough. OK. Here's a time
150:33 map. Notice the green arrow. means I made it in patrol.
150:40 let's do some sun shady. So I take a sun? A
150:45 I think in patrol, it's gonna like a flashlight and then I can
150:50 up and down, change the And here I'm kind of looking maybe
150:54 six o'clock in the afternoon where the going down this time of year.
150:59 . And I'm looking from kind of south. Here is the dip magnitude
151:05 that picture horizon. That's one of products you'll generate. Here's the dip
151:11 you of that picked horizon again, product and then you can call render
151:18 two and there you actually have to two separate horizons to co render it
151:24 in patrol. Here's the most positive of the Pict horizon and the most
151:33 curvature. Now I haven't defined the but red's gonna be anticlines and blue
151:37 gonna be sin coin. OK? I can co render them all
151:44 Then I can make a fault Now, we can't do this in
151:50 patrol software, but this gal in , so um Jessica can look her
151:57 , he's up in Aiken and here's fault surface. And then what's the
152:04 atomu to the fault surface? you can calculate the dip magnet to
152:08 fault surface and then she, oh the dip magnitude dip aute and this
152:13 the cylindrical. So here I'm holding uh water bottom. It's just kind
152:18 a cylinder. Here's my fault. ? It's my fault more like this
152:27 is my fault more like that. the asperity of your fault?
152:36 You can map the shapes of the . Then we calculate attributes parallel to
152:45 Pict horizon. OK. So we're extract them. Oh Did I do
152:51 ? I didn't do extract along a horizon. We're gonna call that a
152:54 strike. Let's do a phantom horizon . So I've got, I picked
152:59 A, I picked Horizon B. were easy horizons to pick in between
153:06 pain and a butt to pick nothing across there. So what I can
153:12 is I can pay Horizon B and patrol add 100 and 20 ft to
153:19 . So I shift it up. here is Horizon B 120 ft.
153:25 haven't changed the shape or anything. moved it. Let's look at the
153:31 . So I've got energy and amplitude , couple of attributes that I generated
153:36 here's along Horizon B that looks Go up Sheller. Oh, in
153:42 area that I couldn't pick the reason couldn't pick it. There's a big
153:45 fan with a little distribution channels. . Then we have attributes extracted proportionally
153:54 two pick slices. We'll call these slices or proportional slices. That'll be
153:59 lab exercise. OK? Gonna take and B and we're gonna go 10%
154:06 the way. 2030 40 50 60 80 90. And here's what I'm
154:12 do. OK. So you can my formula down here in the lower
154:16 . I'm taking 0.3 of horizon A of horizon B gives me a new
154:22 . I slice through the data. visualize it. I hit the little
154:25 I found I capture it. Then attributes extracted along horizons. That
154:35 for photography. We can't do this patrol. There are two packages do
154:41 , the horizon tube in the growth , detect and then uh pun in
154:51 paleo scan. OK. So here human interpreter picks three maybe four P
155:01 . In this case, uh Stan has tied them to a well.
155:06 then we're gonna pick every peak trough zero crossing both the Z and the
155:11 shaped ones for a given trace and to auto track it using the hand
155:16 as a guide. OK. And did all these intermediate picks. So
155:23 got, you know, 37 picks something in here. Now. And
155:28 from that, the geologist here will comfortable with. We were diagrams.
155:38 you wanna know what did the earth like at a particular time at the
155:46 ? Now, if you got Don here, Doctor Don, he actually
155:52 this at Amaco. We called him bug and swag uh a true aged
156:00 because he had the fossil data to it here. Usually seismic data,
156:05 don't have the fossil data. You're gonna hope that over a survey which
156:09 be 100 miles by 50 miles for big survey, you're gonna hope that
156:13 reflectors are of the same geologic It'll be approximately that way.
156:20 So like in this example, then you can see it, hey,
156:24 thicker over here, more accommodation space the west, less a common space
156:29 the east. There's a hiatus, deposition in this area, no deposition
156:34 , no deposition there, right? then you can generate attributes so he
156:41 generate the thickness maps at different levels show where the accommodation space was in
156:50 past. OK. Then we have computed between two horizons formation attributes.
156:59 simplest one is I just pick a slice and I go plus or minus
157:03 milliseconds take the R MS amplitude. could take a horizon go plus or
157:08 10 milliseconds around that R MS I could take a top reflector and
157:14 base reflex and calculate the R But now the thickness changes from point
157:19 point. OK. So here is separate horizons, our mi amplitude between
157:26 two horizons. Now we have GEO and GEO bodies. So here I
157:34 my seismic trace, let's say it's eight bit in VL GEO. The
157:41 used to be stored as actually 0 255. So your zero crossing was
157:46 127. They had a kind of zero crossing in your head. Thank
157:51 . And uh here are the very values he's colored in brown from 0
157:57 60. So the troughs you're gonna a seed point and then the software
158:04 go in line left to right, line in and out vertically up and
158:13 and say, well, is that 0 to 60? If so pick
158:17 , if not, don't pick it you go out, you try to
158:19 that GEO body with the constraint that wanna have potato shaped geo bodies or
158:26 shaped with geo bodies versus spaghetti shaped bodies. So there's a, you
158:32 a constraint on how connected it You want to be pretty well
158:36 So here's a GEO probe and uh lot of the words depend on the
158:44 you're using. So a GEO probe actually the word from Magic Earth and
158:53 landmark. So Halliburton company, so call it a GEO probe and then
158:58 vial geo, well, geo probe copyrighted so nobody else can use it
159:03 sell it. Well, let's call a fro fro. That's a good
159:09 . Ok. So, yeah, probe, vel probe, what are
159:14 gonna call it? You're working for ? What are you gonna call
159:18 The two good names have been What kind of lame name are you
159:22 come up with? Box, Good. That's exactly what they
159:32 Is it? I'm gonna call it Box Pro. Why? Because the
159:36 good geologic sounding or visualization names somebody used up and copyrighted. So they
159:41 to come up with a stupid name then the other companies they can't even
159:45 the name Box Pro. Ok. you'll have five and six names for
159:50 same darn thing. So you're partnering another company? Well, we're using
159:53 Box Probe and damn, I wish had a Box Pro. Boy,
159:57 got the Box probe. They just it something different. Ok. So
160:02 the box probe. Took one of curvature attributes. Oh, here's a
160:07 Pinnacle reef. And then I well, let's go look at all
160:10 ridges and domes in the volume. talk about this in week.
160:15 Ok. And here they are. these are all the little carbonate build
160:20 in the mid wind basin of central and you can pick one of them
160:25 say, oh, what's the how many acre feet is in there
160:31 , and drill it. So in time slices show unbiased views of the
160:38 if care, carefully pick horizon splicing show better fractures within a given litho
160:45 . Ok? If you're careful about it, horizon sizes are better for
160:49 gray, OK? Like channels carbonates horizon choice. That's the Alistair Brown
160:58 . It's a picture horizon that you up or down like a ghost.
161:03 didn't pick it because there's nothing to there. It's just stuff think of
161:08 that's got like uh the Johnno's Basin Colombia, there's like 10,000 channels in
161:14 subway. What are you gonna Pick the top and base of which
161:17 would you even do? No, gonna pick a nice maybe a volcanic
161:22 or some kind of un conformity and you're gonna slice through it using that
161:27 a as a guy. So they a means of visualizing, visualizing complex
161:33 that cannot be easily picked in size cars, mass transport complexes, et
161:40 . So stradle slices also called proportional between two piss. They approximate Strat
161:48 horizons if the rate of deposition does vary vertically. OK. So,
161:58 in other words, the rate of I should really say laterally that the
162:04 of deposition doesn't change. Jessica helped say that I, I wanna say
162:14 . OK. I'm filling 1/10 of , of the formation on the left
162:19 of the survey and 1/10 on the side and 1/10 of the time versus
162:25 gradation where I'm filling everything on the first, then I'm gonna fill in
162:30 middle, then I'm gonna fill on right. But what's the words I
162:35 use there? I'm back after Don't take it. OK. And
162:45 Chronos Strat democratic slices, which is higher tech stuff like paleo scan.
162:51 , they generate maps that approximate a time. So they're better than the
162:55 one except you got to find or software. And so you can map
163:01 between adjacent Chronos traffic graphic horizons. provide measures of accommodation, space of
163:08 , non deposition, et cetera. then finally, uh geo bodies you
163:23 geo bodies provide a mean to extract objects. That's what a computer scientist
163:29 them. Like a carbonate built Most commonly, these objects are zones
163:34 high porosity used in reservoir evaluation. the one most people do with
163:40 an interpreter can isolate carbonate reefs, cones, channels and other features of
163:46 . And of course, machine learning trying to do this on steel
163:49 OK. So any questions before lunch uh dinner for Jessica. OK.
164:01 then we'll come, we're gonna meet the lab at what? 115?
164:04 that good time? Is that enough to go to have lunch and come
164:11 ? Yeah. Ok. Ok. . But we gotta walk there
164:18 So. All right, so let's in the lab at 115.
164:26 And can you lock this room or we lock that a that?
164:39 we can come back here after the . But I'm, I'm thinking,
164:43 know, with all the computers, don't have to walk all that
164:46 Can we, is this room locked is it not? Ok? And
164:51 you can unlock it? Ok, . So we can leave our stuff
164:56 , come get it, go to lab in my case, I'll leave
165:00 suitcase here too. And, so we, we, we'll meet
165:05 115 and, and, uh, is gonna have some, uh,
165:13 and a
-
+