Statics corrections in reflection seismic processing
Statics corrections in reflection seismic processing
I have problems understanding statics correction for reflection seismic processing. In this Covid19 situation and home staying days, I thought it may be a good time to ask my questions and have a chance to have your explanations. Actually, I am going to write down what I realized and some questions and want to kindly ask you if these explanations are right or not. Thank you in advance.
I know that we need to correct for different elevations at source and receiver positions (to have a hyperbolic event and CDP gathers). For this, we need to get rid of two things, one is the topography and another is the low velocity layer (LVL). First, we need to deal with LVL and then topography. LWL is replaced with a constant velocity medium (or with the velocity of the underlying layer(let's call it V1)). So, we need to know the depth and velocity of the weathering layer (V0). Refraction can give us V1 using the FB picking. V0 can also be obtained using uphole surveys.
After these things, we need to define a flat horizontal datum and put sources and receivers on that to have the same elevation for source and receivers. Sum of these delay times (LVL and topography related) should be subtracted from traces times.
Are these explanations right?
how can we have the depth of LVL?
how to know V0 when we don't have uphole survey?
Somewhere I saw "datum velocity", is it the replacement velocity we use to substitute the LVL?
How about floating datum? what is that?!
What is the difference between "elevation statics" and "refraction statics"?
I know I have written too much for a post and asked lots of questions. I really appreciate your time and replies to finally solve my misunderstandings.
I know that we need to correct for different elevations at source and receiver positions (to have a hyperbolic event and CDP gathers). For this, we need to get rid of two things, one is the topography and another is the low velocity layer (LVL). First, we need to deal with LVL and then topography. LWL is replaced with a constant velocity medium (or with the velocity of the underlying layer(let's call it V1)). So, we need to know the depth and velocity of the weathering layer (V0). Refraction can give us V1 using the FB picking. V0 can also be obtained using uphole surveys.
After these things, we need to define a flat horizontal datum and put sources and receivers on that to have the same elevation for source and receivers. Sum of these delay times (LVL and topography related) should be subtracted from traces times.
Are these explanations right?
how can we have the depth of LVL?
how to know V0 when we don't have uphole survey?
Somewhere I saw "datum velocity", is it the replacement velocity we use to substitute the LVL?
How about floating datum? what is that?!
What is the difference between "elevation statics" and "refraction statics"?
I know I have written too much for a post and asked lots of questions. I really appreciate your time and replies to finally solve my misunderstandings.
Re: Statics corrections in reflection seismic processing
The way I think of this is that we want to simulate the seismic survey being collected on a flat, horizontal surface,so the reflections make pretty hyperbolas.
We do that by
a) stripping out the low velocity layer and
b) fill this in with some kind of magic concreterock with a constant velocity so that
c) it ends up flat
It's all based on the idea that the ray paths are nearlyvertical, so that a static shift will be fine. That's an okay approximation for "small shifts", which in general means you are dealing with lower velocities and small ranges of velocity. In hard rock, that doesn't work so well.
So
 yup your explanations look pretty good
 its common to to figure out the velocity/thickness of the LVL by looking at the refractions
 that means picking "first breaks" and doing some kind of inversion
 there's some specialised software packages for this
 they can include upholes or other a priori information
 they often resolve for elevations as well
Replacement velocity is the term we used in refraction statics; datum velocity is usually used with elevation corrections. These are normally the same value, but can be separate processes. You might apply none one or both; in some packages they are integrated so you have no choice  do refraction statics and the elevation statics are built in.
You use a floating datum if the topography is large. Its a smoothed version of the elevations. You do that to keep the corrections "small"  again that vertical approximation. Small usually means less that 1216ms; things can get a bit distorted after that.
And  yeah, staying at home too so ask away
We do that by
a) stripping out the low velocity layer and
b) fill this in with some kind of magic concreterock with a constant velocity so that
c) it ends up flat
It's all based on the idea that the ray paths are nearlyvertical, so that a static shift will be fine. That's an okay approximation for "small shifts", which in general means you are dealing with lower velocities and small ranges of velocity. In hard rock, that doesn't work so well.
So
 yup your explanations look pretty good
 its common to to figure out the velocity/thickness of the LVL by looking at the refractions
 that means picking "first breaks" and doing some kind of inversion
 there's some specialised software packages for this
 they can include upholes or other a priori information
 they often resolve for elevations as well
Replacement velocity is the term we used in refraction statics; datum velocity is usually used with elevation corrections. These are normally the same value, but can be separate processes. You might apply none one or both; in some packages they are integrated so you have no choice  do refraction statics and the elevation statics are built in.
You use a floating datum if the topography is large. Its a smoothed version of the elevations. You do that to keep the corrections "small"  again that vertical approximation. Small usually means less that 1216ms; things can get a bit distorted after that.
And  yeah, staying at home too so ask away
Re: Statics corrections in reflection seismic processing
a
Last edited by jefry123 on Sun Mar 22, 2020 8:31 am, edited 1 time in total.
Re: Statics corrections in reflection seismic processing
Thank you so much Guy, you always explains every thing in the easiest way which is great.
I still have a little problem in differentiating elevation correction and refraction statics.
I have written a note somewhere in my notebook that "elevation statics is the vertical alignment of different elevations of source and receiver" and "refraction statics is the correction for the weathered layer using first breaks", right?
However, as far as I understood, we have to deal with both elevation difference of source and receivers, and weathered layer. In this case, we have to use refractions and first breaks to figure out weathering layer velocity and thickness, use a magic concrete velocity replacement to fill it, and correct the elevation difference of sources and receivers using elevation statics, right!?
So, how can they be two separate methods?!
Re: Statics corrections in reflection seismic processing
"How can they be two different corrections?"
History, really. We could apply elevation corrections long before we had a refraction statics method, both to flat and floating datum planes.
We also had residual statics before we had refraction statics; residual statics allow for any "surfaceconsistent" parts of the solution.
So  elevation statics plus residual statics might get you to a solution in some situations, without actually picking all those first breaks.
I'd usually apply elevation statics and look at the refraction breaks  are they straight? Or close to it? Applying a LMO correction interactively will help.
You can do a lot of work on refraction statics and not make much difference in some cases.
Modern residual static are really very good  multiple terms and so on  and will always be able to deal with more detail in the nearsurface better than you can get from a refraction statics model, which is always relatively coarse.
It's all 80/20 rule stuff  what's the minimum work you need to do to get close enough to what you need?
History, really. We could apply elevation corrections long before we had a refraction statics method, both to flat and floating datum planes.
We also had residual statics before we had refraction statics; residual statics allow for any "surfaceconsistent" parts of the solution.
So  elevation statics plus residual statics might get you to a solution in some situations, without actually picking all those first breaks.
I'd usually apply elevation statics and look at the refraction breaks  are they straight? Or close to it? Applying a LMO correction interactively will help.
You can do a lot of work on refraction statics and not make much difference in some cases.
Modern residual static are really very good  multiple terms and so on  and will always be able to deal with more detail in the nearsurface better than you can get from a refraction statics model, which is always relatively coarse.
It's all 80/20 rule stuff  what's the minimum work you need to do to get close enough to what you need?
Re: Statics corrections in reflection seismic processing
So typically the way I've been working is
 get the geometry sorted out, look at the elevations and built a floating datum
 get a basic velocity field and look at a stack, shots, CDPs with and without a floating datum
 make a call then if I need to do refraction statics
 pick first breaks, run refraction statics
Now the kit I was using takes the error (ie modelled first break minus observed) and runs a surfaceconsistent decomposition to get a residual as well, which is a good trick.
 test apply refraction statics, look at shots, cdps, stacks
 compare that with my floating datum or fixed datum elevation statics
(test signal processing  decon, noise suppression, gain)
 pick velocities and do residual statics; two iterations
So  elevation, refraction and two passes of residual statics and velocities.
 get the geometry sorted out, look at the elevations and built a floating datum
 get a basic velocity field and look at a stack, shots, CDPs with and without a floating datum
 make a call then if I need to do refraction statics
 pick first breaks, run refraction statics
Now the kit I was using takes the error (ie modelled first break minus observed) and runs a surfaceconsistent decomposition to get a residual as well, which is a good trick.
 test apply refraction statics, look at shots, cdps, stacks
 compare that with my floating datum or fixed datum elevation statics
(test signal processing  decon, noise suppression, gain)
 pick velocities and do residual statics; two iterations
So  elevation, refraction and two passes of residual statics and velocities.
Re: Statics corrections in reflection seismic processing
1. Are these explanations right?
Yes, they are.
2. how can we have the depth of LVL?
Before the development of the refraction statics technology, the seismic crews used to shoot a WZ survey in parallel with the 2D reflection (production) survey. A 500m 24 geophonesymetrical refraction spread with one geophone per station, the geophones near the source being closely spaced, was laid on the ground and a seismic source was activated once at both ends of the spread. The first arrivals were recorded then picked manually on paper or screen in the office and interpreted to get Vo and the refraction velocities (sometimes the WZ had more than one layer). The depth of each WZ layer at each source point was then calculated using the standard intercept formulas, knowing Vo, V1,V2... and the intercept for each refractor. So, in the end, we get a nice LVL section along the 2D line from which the static corrections are calculated at each source or receiver station.
The refraction statics technology is based on the firstbreak picking of the production seismic records. There are several methods for the derivation of the static corrections from the firstbrek picks but for all of them you need to supply Vo from a separate source, generally an uphole survey. However, in the case of the Generalized Linear Inversion, which is one of the most widely used refraction statics package, you can start with a bestguess Vo to build the initial model and then the program will iteratively build a final LVL model which matches the firstbreak picks.
Yes, they are.
2. how can we have the depth of LVL?
Before the development of the refraction statics technology, the seismic crews used to shoot a WZ survey in parallel with the 2D reflection (production) survey. A 500m 24 geophonesymetrical refraction spread with one geophone per station, the geophones near the source being closely spaced, was laid on the ground and a seismic source was activated once at both ends of the spread. The first arrivals were recorded then picked manually on paper or screen in the office and interpreted to get Vo and the refraction velocities (sometimes the WZ had more than one layer). The depth of each WZ layer at each source point was then calculated using the standard intercept formulas, knowing Vo, V1,V2... and the intercept for each refractor. So, in the end, we get a nice LVL section along the 2D line from which the static corrections are calculated at each source or receiver station.
The refraction statics technology is based on the firstbreak picking of the production seismic records. There are several methods for the derivation of the static corrections from the firstbrek picks but for all of them you need to supply Vo from a separate source, generally an uphole survey. However, in the case of the Generalized Linear Inversion, which is one of the most widely used refraction statics package, you can start with a bestguess Vo to build the initial model and then the program will iteratively build a final LVL model which matches the firstbreak picks.
Re: Statics corrections in reflection seismic processing
3. how to know V0 when we don't have uphole survey?
See answer to question 2.
4. Somewhere I saw "datum velocity", is it the replacement velocity we use to substitute the LVL?
The most widely used term is "replacement" velocity. I think elevation statics are sometimes called datum statics, so in this case the replacement velocity is called "datum" velocity.
5. How about floating datum? what is that?!
The floating datum is generally a smoothed version of the elevation profile which is used as an intermediate datum during the velocity analysis step. The purpose is to avoid applying large static shifts (when we have very rough topography) which would have a distorting effect on the stacking velocities. However, we need to apply the final flatdatum statics before staking.
6. What is the difference between "elevation statics" and "refraction statics"?
In some areas, the weathering effect is not significant so the static shifts are mainly caused by the surface elevation variation from station to station, not by the velocity variation in the first layer. So all we need is the elevation profile and the replacement velocity. The standard practice is to test, at the beginning of the survey, elevation statics, refraction statics and sometimes uphole statics (a statics map built using only a dense upole grid), i.e. produce a brute stack for a small portion of data using the corresponding sets of statics, and then compare the stacks and decide if it is worth going for firstbreak picking, which is a heavy, timeconsuming, costly process. In some areas, the client's geophysicists in charge of the seismic survey know by experience from previous work in the area that simple elevation statics are good enough.
See answer to question 2.
4. Somewhere I saw "datum velocity", is it the replacement velocity we use to substitute the LVL?
The most widely used term is "replacement" velocity. I think elevation statics are sometimes called datum statics, so in this case the replacement velocity is called "datum" velocity.
5. How about floating datum? what is that?!
The floating datum is generally a smoothed version of the elevation profile which is used as an intermediate datum during the velocity analysis step. The purpose is to avoid applying large static shifts (when we have very rough topography) which would have a distorting effect on the stacking velocities. However, we need to apply the final flatdatum statics before staking.
6. What is the difference between "elevation statics" and "refraction statics"?
In some areas, the weathering effect is not significant so the static shifts are mainly caused by the surface elevation variation from station to station, not by the velocity variation in the first layer. So all we need is the elevation profile and the replacement velocity. The standard practice is to test, at the beginning of the survey, elevation statics, refraction statics and sometimes uphole statics (a statics map built using only a dense upole grid), i.e. produce a brute stack for a small portion of data using the corresponding sets of statics, and then compare the stacks and decide if it is worth going for firstbreak picking, which is a heavy, timeconsuming, costly process. In some areas, the client's geophysicists in charge of the seismic survey know by experience from previous work in the area that simple elevation statics are good enough.
Re: Statics corrections in reflection seismic processing
I didn't know seismic refraction surveys can be useful for seismic reflection data processing, very interesting. We use seismic refraction in engineering field very often. Was the seismic refraction method invented for helping seismic reflection data processing in the beginning?
Re: Statics corrections in reflection seismic processing
In the early times of oil and gas exploration, before the development of the seismic reflection technology, the refraction method was used to map the basement and find the major structures. Very long spreads and big explosive charges were needed in order to record the refracted first arrivals from the deep basement.
Re: Statics corrections in reflection seismic processing
Aha! It makes sense. The refraction method is easier in terms of data processing, especially when you don't have the computing power we have now. So it actually can be used first.
Re: Statics corrections in reflection seismic processing
"The refraction method is easier in terms of data processing, especially when you don't have the computing power we have now. So it actually can be used first."
Well, it was.
Most modern systems for determining the LVL are going to use a full tomographic inversion that is calibrated by any uphole data that you have; this allows for a vertically varying LVL. As the hunt for oil shifts to more challenging geographic areas (the easy stuff being found some time ago) so the sophistication of these packages has increased.
The big "reflection" software processing systems tend to provide a capability, but fully acknowledge there's specialist providers that make outstanding solutions. On top of that there's a certain amount of "art" involved in recognising the subtle variation in first breaks and the implications for the subsurface.
If you get a chance, it's well worth checking out Chuck Diggins work on this; he's been in the game for many years and helped to develop many products over that time.
Well, it was.
Most modern systems for determining the LVL are going to use a full tomographic inversion that is calibrated by any uphole data that you have; this allows for a vertically varying LVL. As the hunt for oil shifts to more challenging geographic areas (the easy stuff being found some time ago) so the sophistication of these packages has increased.
The big "reflection" software processing systems tend to provide a capability, but fully acknowledge there's specialist providers that make outstanding solutions. On top of that there's a certain amount of "art" involved in recognising the subtle variation in first breaks and the implications for the subsurface.
If you get a chance, it's well worth checking out Chuck Diggins work on this; he's been in the game for many years and helped to develop many products over that time.
Re: Statics corrections in reflection seismic processing
I always love colorcoded contour maps such as those provided by refraction tomography. The software we use probably are not as powerful as those used in oil and gas industry, but are convenient for us to use for smaller scale engineering surveys:
viewtopic.php?f=2&t=1817
viewtopic.php?f=2&t=1817

 Similar Topics
 Replies
 Views
 Last post