Mike Perz: 2015 CSEG Symposium Proposal for Presentation Title: “Washout Creek 3C: a multi attribute case study” Context: The proposed talk is centered around a detailed fracture analysis and reservoir characterization which was undertaken on Arcis’ Washout Creek Multi-client 3D-3C data set. The Washout Creek survey was acquired in March 2014, and it was designed to evaluate multiple zones from Cretaceous to Devonian, including the Cardium and Duvernay formations. A relatively high field effort was employed (SL spacing=280 m; RL spacing=240 m; SI=RI=40 m leading to natural bin size of 20 x 20 m) in order to maximize data quality for the prestack azimuthal analyses that were planned at the outset of acquisition. Most happily, the data quality has turned out to be very good, especially on the PS side. It is well-acknowledged that the very thin Cardium sands are below the resolution threshold for surface seismic experiments; however, the question of whether attributes characterizing lithology and/or fractures in the overlying and/or underlying formations (which are sufficiently thick to allow seismic detection) might provide indirect information about the Cardium proper remains unanswered. In an effort to address this important question, we decided to adopt an ‘everything but the kitchen sink’ approach to creating a broad range of seismic attributes. In particular, we generated all four of the surface-seismic-based fracture detection attributes which are commonly used in the industry today: shear-wave splitting (SWS) from the converted wave data, azimuthal AVO (AVAZ) and velocity analysis (VVAZ) from the prestack PP data, and curvature from the poststack PP data. In addition, we are generating Vp/Vs volumes from the PS processing; moreover, we are in the midst of generating various reservoir characterization attributes such as impedance inversion and LMR volumes as well as joint PP/PS prestack inversion. By the time the Symposium rolls around, we will have completed analysis of all these attributes. To date, we have examined the four fracture attributes (Figs 1—4) . Novelty: It is quite rare, if not unprecedented, that such a large number of new and diverse attributes be examined concurrently on a single data set. For example, I am only aware of a small handful of simultaneous comparisons of SWS and VVAZ in the public domain and all of them were focusing on the shallow overburden. This would be the first time, to my knowledge, that a simultaneous comparison between VVAZ and SWS at depth (that is, after employing layer stripping to remove effects of the shallow anisotropy) would be presented in the public domain. Weakness: As mentioned, we were expecting to have an interpreter work with us to render a detailed interpretation of our attributes, including a validation against existing production data, which exists in significant quantities over our survey area. It is unclear now whether or not that interpretation will be ready by the time the Symposium rolls around. Whether or not we have this interpretation has a profound bearing on how the talk will shape up. For now I will assume we will NOT have the interpretation ready.
4
Embed
Mike Perz: 2015 CSEG Symposium Proposal for Presentation · 2015. 11. 18. · Mike Perz: 2015 CSEG Symposium Proposal for Presentation Title: ^Washout Creek 3C: a multi attribute
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
Mike Perz: 2015 CSEG Symposium Proposal for Presentation
Title: “Washout Creek 3C: a multi attribute case study”
Context: The proposed talk is centered around a detailed fracture analysis and reservoir characterization which
was undertaken on Arcis’ Washout Creek Multi-client 3D-3C data set. The Washout Creek survey was
acquired in March 2014, and it was designed to evaluate multiple zones from Cretaceous to Devonian,
including the Cardium and Duvernay formations. A relatively high field effort was employed (SL
spacing=280 m; RL spacing=240 m; SI=RI=40 m leading to natural bin size of 20 x 20 m) in order to
maximize data quality for the prestack azimuthal analyses that were planned at the outset of
acquisition. Most happily, the data quality has turned out to be very good, especially on the PS side.
It is well-acknowledged that the very thin Cardium sands are below the resolution threshold for surface
seismic experiments; however, the question of whether attributes characterizing lithology and/or
fractures in the overlying and/or underlying formations (which are sufficiently thick to allow seismic
detection) might provide indirect information about the Cardium proper remains unanswered. In an
effort to address this important question, we decided to adopt an ‘everything but the kitchen sink’
approach to creating a broad range of seismic attributes. In particular, we generated all four of the
surface-seismic-based fracture detection attributes which are commonly used in the industry today:
shear-wave splitting (SWS) from the converted wave data, azimuthal AVO (AVAZ) and velocity analysis
(VVAZ) from the prestack PP data, and curvature from the poststack PP data. In addition, we are
generating Vp/Vs volumes from the PS processing; moreover, we are in the midst of generating various
reservoir characterization attributes such as impedance inversion and LMR volumes as well as joint
PP/PS prestack inversion. By the time the Symposium rolls around, we will have completed analysis of all
these attributes. To date, we have examined the four fracture attributes (Figs 1—4) .
Novelty: It is quite rare, if not unprecedented, that such a large number of new and diverse attributes be
examined concurrently on a single data set. For example, I am only aware of a small handful of
simultaneous comparisons of SWS and VVAZ in the public domain and all of them were focusing on the
shallow overburden. This would be the first time, to my knowledge, that a simultaneous comparison
between VVAZ and SWS at depth (that is, after employing layer stripping to remove effects of the
shallow anisotropy) would be presented in the public domain.
Weakness: As mentioned, we were expecting to have an interpreter work with us to render a detailed
interpretation of our attributes, including a validation against existing production data, which exists in
significant quantities over our survey area. It is unclear now whether or not that interpretation will be
ready by the time the Symposium rolls around. Whether or not we have this interpretation has a
profound bearing on how the talk will shape up. For now I will assume we will NOT have the
interpretation ready.
Themes explored in the talk: 1. I want to use this case study to convey my conviction that the industry is overly focused on
challenges on the rock physics side of unconventional seismic exploration while ignoring the
challenges faced in the imaging side—this despite the fact that imaging problems on noisy land
data will render absolutely useless the most elegant rock physics analysis. I’m still not sure how
I’ll weave this into the talk, but I’ll find a way.
2. I want to remind the audience of something that we now seem to be understanding as an
industry: namely that all these different fracture attributes aren’t expected to give consistent
results because they are sensing things at different measurement scales.
3. I could discuss pitfalls of the 4 fracture detection approaches, as you guys suggested in an email
last week…but my original intent was not to dwell on this simply because I’ve presented that
very thing before in a CSEG talk a couple years back. But maybe I should revisit this to some
degree. Please comment!
4. Obviously I want to spend a large amount of time showing the attributes themselves. Here are
some fussy specifics:
a. Compare Figs 1 and 2 (SWS and VVAZ in the shallow section). Note that the orientations
all trend from NE to SW. The fact that these are consistent with the orientation of the
max horizontal regional stress field inspires confidence in both results. What is quite
puzzling, however, is the lack of agreement of the intensity patterns. We have reviewed
this in detail and I still don’t have a solid explanation. The most obvious explanation
would be noise in the data; however, we have looked very carefully at the prestack PP
and PS gathers and they are clean. We are still investigating, but the most likely
explanation seems to be that the anisotropy is “lighting up” different elastic parameters
preferentially in different parts of the survey, some of which control the shear wave
splitting (Thomsen’s gamma) and others which control the VVAZ effect (Thomsen’s
delta_v). As far as I understand these are indeed independent parameters, so there’s no
guarantee they would have to track each other…but the whole thing is quite troubling,
or fascinating, depending on your perspective.
b. Compare Figs 3 and 4 (SWS and VVAZ in the deeper Viking to Banff interval). Clearly
both sections show that the anisotropy orientation has changed from the consistent NE
to SW observed in the shallows. This is consistent with the little bit of feedback we got
from the interpreter, as they figured that the sub-Viking anisotropy would be different
than that above owing to a certain Cretaceous thrust faulting episode. The “problem” is
that the two maps are not showing huge consistency in the orientations. Yet, there is
still enough similarity in orientation in places to make you believe that the two
attributes are picking up something meaningful.
c. I’ve given the AVAZ and curvature a cursory look over, and I don’t see any obvious
correlations between the two of them, nor with the SWS and VVAZ.
5. Obviously I would spend time explaining the processing steps leading up to the attribute
inversions. This is in keeping with theme 1 above.
6. I might comment on pros and cons of this ‘everything but the kitchen sink’ approach to fracture
characterization. It carries an identical tradeoff to what interpreters encounter when they
create a gazillion “conventional” attributes. On the one hand the more info the better, on the
other hand you could just spin your wheels being overwhelmed by too many maps.