Gravity and Magnetics in Southeast Asia - Modern Applications
Brian S. Anderson, John E. Bain, Harald van Hoeken, and
Mark Weber, Fugro-LCT, Inc.
Abstract
Recent advances in the acquisition, processing, and interpretation
of gravity and magnetics methods now enable the use of these
data for solving a wide range of seismic problems.
In the summer of 1995, LCT in conjunction with Robertson
Research International, Kevron Pty Ltd. and Oil Search Plc.
acquired a unique airborne gravity, magnetics, and seepfinder
survey over the Erlend Basin in the North Sea. A review
of the gravity data from this survey demonstrates an accuracy
approaching that of 2D marine surveys.
There is an increasing number of marine 3D seismic surveys
being acquired with concurrently recorded high resolution
gravity and magnetics data. Instrumentation, navigation,
and processing advances have led to significantly increased
gravity and magnetics data resolution. State of the art
workstation software tools provide a means for the integration
of seismic, gravity, and magnetic data. Integrated interpretations
have aided in the delineation of complex imaging problems
including the verification and enhancement of seismic velocity
models. A review is given of technical developments, economic
considerations, and case studies in support of integrated
3D marine surveys utilizing gravity, magnetics, and seismic
data.
Introduction
Airborne gravity is an important exploration tool in South
East Asia. It is often used in areas over which dense jungle,
transition zone geography, and inhospitable ground conditions
exist for seismic data acquisition operations. Although
airborne gravity has been in use in South East Asia for
over ten years, only recently has a system been developed
which takes advantage of the latest advances in DGPS positioning
technology and enhanced gravity processing. By virtue of
the advancements in DGPS and processing, a new approach
has been proven which allows for the acquisition of high
quality airborne gravity data under conditions which would
have previously been unachievable. By accurately recording
and computing the three dimensional velocities (and accelerations)
on the airborne gravity instrument, these effects can now
be quantitatively removed in processing. Other approaches
which do not utilize this sophtisticated processing technology
employ heavy filtering of the data, which are effective
not only in the suppression of noise, but also in distorting
the gravity signal.
In the marine environment, using new interpretation case
studies from the Gulf of Mexico and offshore Europe, it
is possible to gain insights into enhancing and constraining
the interpretation of 3D seismic data using gravity and
magnetics methods. Using digital horizons interpreted from
a 3D seismic volume and high resolution gravity and magnetics
data, an integrated and constrained 3D geologic model can
be quickly built and tested.
The acquisition of high resolution gravity and magnetic
data in conjunction with 3D seismic surveys is now an accepted
norm in Europe, and is quickly gaining acceptance in the
Gulf of Mexico. Over 300 OCS blocks of 3D-acquired gravity
and major areas of high resolution aeromagnetic data have
been and are now being recorded in the Gulf of Mexico. The
use of high resolution gravity in seismic velocity analysis
and the use of velocity grids for localized, focused density
input to gravity models is now possible. A detailed example
of a localized conversion of a velocity cube to a density
volume is provided from the Southern Additions, offshore
Louisiana.
A brief review of instrumentation, processing techniques,
costs, and integrated software applications is provided
to set the framework for the interpretation case studies.
Gravity and magnetic instrumentation has decreased in size
and increased in sampling and resolution power. Processing
of the data using high quality DGPS positioning data has
resulted in a dramatic increase in the resolution of shipborne
gravity. Workstation applications are now in use which facilitate
the direct transfer of data and models between seismic and
gravity/magnetic modelling software systems.
Data Acquisition and Processing: The State of the Art
With the development of digitally controlled marine gravity
systems as described by LaFehr et al (1,2) the restrictions
of hardware-defined filtering have been removed. This is
a major factor in recovering maximum signal in final processed
results, by allowing the data processor to quantitatively
determine optimum filter parameters for specific sea conditions
and induced noise levels. Marine surveys are now routinely
producing results of 0.1-0.5 mgal resolution over 500 to
1000 meter minimum wavelengths. This change reflects a)
the new digital gravity meter technologies, b) benefits
of DGPS positioning for the removal of vehicular motion
effects on the gravity meter, and c) the benefit of larger,
more stable multi-streamer seismic vessels. In addition,
the better spatial sampling of the data due to closely spaced
ship tracks on a 3D seismic survey greatly increase the
data resolution over 2D methods (Fig.
1).
Likewise in magnetics, increases in resolution have also
occurred, and are due to a combination of better equipment,
more frequent data sampling, and enhanced data processing
techniques.
Closely-sampled 3D-acquired potential fields data has presented
new challenges in data processing. High resolution results
(Figs. 2 and 3)
have required the development of new line leveling algorithms
and filtering techniques to address low amplitude random
noise in the data when profiles are combined to produce
grid results. Predictive gridding and narrow band Weiner
strike filtering are some of the techniques now employed.
Economics
With the typical 3D survey vessel pre-equipped with DGPS
navigation, power and space, little additional cost is incurred
in the addition of high resolution gravity and magnetics
to the 3D survey effort. Worldwide, the cost of acquiring
and processing this data is on the order of 1% - 3% of the
seismic data acquisition (before seismic processing) costs
(Fig. 4).
Interpreting High Resolution Data
As explorationists, we are all familiar with the important
industry trend of integration. This term has been used to
describe many things. It comes from the same root as the
word integrity - meaning the state of being truthful or
whole. In exploration it is used to describe the incorporation
of well data, geology, seismic, gravity, magnetics, cultural,
and other data to form a whole or integrated model of the
subsurface. 3D-acquired gravity and magnetic data is playing
a larger role than ever in finding oil in the Gulf of Mexico
and elsewhere in the world, through the use of truly integrated
subsurface models and interpretations.
As described by Saad (3) and Pawlowski (4), a breakthrough
in effective integration has been the emergence of workstation
applications for simultaneous modelling of seismic, gravity,
and magnetic data (Fig.5).
Team-oriented Exploration Tools. With the trend towards
highly focused exploration teams, the smooth interaction
and coupling of multiple geophysical disciplines is essential.
Explorationists are expected to employ and be familiar with
more disciplines on a continuing basis. The development
of workstation applications which enable the interpreter
to simultaneously refine the subsurface model using seismic,
gravity, and magnetic data has been a giant step forward.
Data Applications. The benefits of acquiring and
incorporating 3D-acquired gravity and magnetic data into
subsurface models is multifold:
 Increased
resolution on interpreting steeply dipping strata
Aids
in resolution of seismic "no data zones" (overthrust,subsalt,
etc.)
Base
of salt imaging
Determination
of salt vs. sediment for seismic velocity analysis
Seismic
velocity modelling
Even with high quality 3D seismic data, interpreters can
have problems in defining the salt/sediment boundary at
the flanks of a salt dome, salt sheet, or other complex
structure. For decades, gravity has been used to address
this problem. The most recent changes are: a) better acquisition
technology and processed data, and b) truly integrated workstation
software tools. By incorporating a co-recorded data set,
independently measuring a related property of the subsurface
(density from gravity and velocity from seismic), the interpreter
can place a much higher degree of confidence in the final
geologic interpretation. To quantify this observation, the
following case studies show that incorporating 3D seismic
with high resolution gravity and magnetics can alter the
base of salt interpretation by several thousand feet from
the 3D seismic interpretation alone. In some cases, results
from gravity modelling have provided excellent insights
into the geology below a salt body, enabling the seismologists
to refine their migration velocity model for the structure,
and as a result, refine the seismic image through reprocessing
the data using the new velocity model.
Velocity Modelling. At present gravity is commonly
incorporated into the subsurface model after the seismic
data has been: 1) fully processed, 2) specialty processed,
3) migrated, and 4) interpreted. Present work is underway
to incorporate the gravity and magnetic data into the seismic
data at a much earlier stage, ideally during the initial
velocity analysis process. The end result will be a velocity
model which respects the constraints of the gravity and
magnetic data, and a much more refined density model (from
seismic velocities) for use in the interpretation of the
gravity and magnetic data.
Case Study 1: Gulf of Mexico Velocity - Density Volume
& Deep Low Density Zone Mapping
As described by Bain et al (5) the primary determining
factor in gravity interpretation validity is the amount
and accuracy of density data. In the same way, magnetic
interpretation is limited by magnetic susceptibility control.
In this case study, data from 54 check shot velocity surveys
with co-located gamma-gamma density logs are analyzed to
determine a localized empirical relationship between near-surface
seismic velocity and density for the Southern Additions,
offshore Louisiana. When these data are plotted(Fig. 6) versus the more commonly used
Gardner's Equation for density/velocity conversion, it is
apparent that the locally-and empirically-derived LASA equation
provides a more suitable velocity to density conversion
for the data in this area. With the aid of this relationship,
and density logs from over 1,500 wells, a 32-layer (stacked
grids) density model is constructed for the Southern Additions,
for use in the regional and prospect level mapping of inter-salt
sediment thicknesses, and relief of the deep low density
zone (Fig. 7). The
LASA velocity density relationship is most useful in the
upper 5,000 feet of the subsurface model. In this area,
the impact of incorrect densities on the modelling results
is greatest. It is also the area where density logs are
most lacking. Effective use of velocity data enhances density
control for gravity modelling. Classical gravity and magnetics
modelling has often been performed using a single density
or susceptibility value for each geologic unit or "layer"
in the model. We compare a map view of an interval seismic
velocity grid (Fig. 8) for a 2,000 foot layer of an
area offshore Louisiana, with the corresponding density
grid as computed from the velocities using Gardner's Equation
(Fig. 9). The significant lateral variations
indicate that for the most accurate modelling, laterally
varying interval densities should be used as a better approximation
of geologic truth than a fixed value per layer.
Case Study 2: An Integrated Modelling Approach to Salt
Imaging
The tabular salt body shown in Fig. 10 is interpreted from a 3D seismic
survey. As with many salt features the top of the body is
easily interpreted (except where steeply dipping) but the
base of salt and the "Gumbo Zone" below the salt are difficult,
if not impossible, to interpret from the seismic data alone.
Furthermore, the time based seismic interpretation does
not provide depth information important to the development
of a successful interpretation. A real-time integrated modelling
technique using gravity, well log and seismic data is conducted
in order to:
Confirm
the seismic interpretation
Delineate
possible "Gumbo Zone" thickness
Determine
approximate depth information
Provide
constraints for depth migration velocity model
An initial 2D depth model is generated using the seismic
derived salt geometry and sediment velocities. Laterally
varying density data obtained from the 3D density volume
discussed in Case Study 1: Gulf of Mexico Velocity- Density
Volume & Deep Low Density Zone Mapping, and a salt density
of 2.08 g/cm3 are used to constrain the model (Fig.11). The calculated gravity of
the initial model has a similar shape to the observed gravity
data. However, the significant difference in the magnitudes
of the calculated and observed gravity fields indicates
that the initial interpretation is not entirely correct.
The fit of the calculated field more closely matches that
of the observed field after:
Conversion of the salt body from 2D to 2.5D. Seismic derived
top salt maps were used to determine approximate half-widths
for the
salt body to a depth of 8,000 feet. The salt body has significant
strike length below 8,000 feet allowing it to be treated
as a 2D
body.
Integration of density data derived from well logs with
sediment warping derived from the seismic image. The shapes
of the laterally
varying sediment density polygons derived from well and
seismic data were modified to include structural warping
contained
in the seismic data.
Modification of the base salt geometry. Little or no modification
was implemented where the base of salt was easily interpreted
rom
the seismic data. Large modifications were made to the base
of salt where the seismic data provided no clear indication
as
to the location of the salt base and where required to match
the gravity.
Incorporation of a low velocity, low density, "gumbo" zone
below the tabular salt. The geometry of the "gumbo" zone
was derived from
subtle amplitude indications in the seismic image and correlations
between the calculated and observed gravity fields.
The final depth model (Fig. 12) displays a high degree of
fit between the observed and calculated gravity field derived
from the integrated model. The shallowest steeply dipping
top of salt has been modified slightly from that of the
original seismic-derived interpretation in order to match
the high frequency component of the observed. This was deemed
to be acceptable by the explorationist as the seismic method
may indicate inaccurate dips near vertical structures.
The original base of salt, as interpreted from the seismic
data alone and displayed in Fig. 12, differs from the base
of salt derived from the integrated approach over portions
of the body. The vertical discrepancy between the original
and final models is as much as 3,500 feet. Also notable
is the large portion of the model over which the integrated
model confirms the initial seismic derived base salt interpretation.
Although high resolution magnetic data was not used in
this case study, its application would have provided an
additional inexpensive data set for the imaging of the complex
salt feature.
Case Study 3: North Sea Airborne Gravity and Magnetics-Regional
Structure Mapping
In the summer of 1995, an airborne gravity, magnetics,
and seepfinder survey was conducted over the Erlend Trough
area. 19,653 line km of airborne gravity data was recorded
in a one month time period as a part of the effort, using
a standard Cessna 404 aircraft under contract from Kevron
Pty. Ltd. under contract to SPT. The combined gravity and
GPS data set was processed using LCT-developed algorithms
capable of producing free-air gravity data accurate to better
than one mGals at four km half-wave lengths. Also shown
is a comparison between the airborne gravity and existing
regional marine gravity coverage. The resolution of the
airborne survey approaches that of the marine data.
In focusing on the gravity portion of this project, three
elements are important with respect to exploration in South
East Asia and other difficult or remote areas:
Speed
of acquisition and delivery of the data
Resolution
of the data with respect to previous airborne gravity systems
Application
of the data in optimizing seismic costs
Case Study 4: Reef or Volcano? - Offshore Lombok, Indonesia
BP Exploration conducted a seismic program covering their
Production Sharing Contract (PSC) area, offshore Lombok,
Indonesia in 1991 which included some 20,000 km of seismic,
gravity and magnetics data. Western Geophysical acquired
the seismic data, and LCT acquired and processed the gravity
and magnetics data. These data were then merged firstly
with a number of industry data sets, secondly with the existing
public domain gravity and magnetic data sets obtained from
the U.S. National Geophysical Data Center (NGDC), and finally
for the gravity, these results were merged together with
satellite-derived gravity data. The satellite-derived data
had been reprocessed from the original satellite tracks
using seismic stacking algorithms, which were adapted by
BP for this purpose (Lewis & Mitchell, 1991). The fully
merged data set included approximately 122,000 line km of
gravity and magnetics coverage.
The regional data sets proved to be highly valuable at
the regional scale,providing a birds-eye view of the tectonic
fabric of the area as well as an independent guide for linking
up the numerous (and oftentimes nebulous) fault systems
interpreted from the seismic data.
The portion of the work discussed herein involved a detailed
analysis of the seismic, gravity and magnetics data over
a subset of the total survey area, with the purpose being
to delineate several seismic "bumps" and attempt to determine
whether these were volcanic features, basement horsts, or
reefal buildups, all of which are known to occur in other
nearby areas. This multi-disciplinary modelling effort aided
the farm-in of a 50% partner, and assisted the selection
of drilling targets. One major benefit to the study discussed
herein was that the modelling and interpretation of the
seismic, gravity and magnetics were performed at the same
time, and by a closely coordinated group effort. In this
way, new ideas were driven by one geophysical method, quickly
tested using the other methods and, accordingly, the ideas
were either dismissed or strengthened in a truly real-time
integrated exploration sense.
The carbonate fairway area of investigation was approximately
240 km west to east and 130 km south to north. Four analyses
were undertaken simultaneously:
Seismic
mapping of the key horizons
Construction
of a depth to magnetic basement surface
Lineament
trend analysis using gravity and magnetics including delineation
of regional changes in basement composition
A
detailed analysis of the density and sonic logs for the
entire study area and a susceptibility analysis from outcrops
on surrounding
islands and limited borehole information.
The next phase in the study involved detailed modelling
over specific features of interest. The primary seismic
problem was to determine whether specific seismic bumps
were caused by: 1) thick carbonates, 2) buried volcanoes
or basement highs, 3) thin carbonates seeded on top of buried
volcanoes, or 4) non carbonate sequences. The seismic interpretation
was straight-forward down to the top of the "reef." However,
many differences of opinion were emerging for the interpretation
of the section below the "reef" top, and many spirited discussions
took place daily as the interpretation progressed. The variations
in seismic interpretation ranged (for individual features)
between a thin veneer of carbonates sitting on top of a
basement high (on the order of 250 ms reefal buildup) to
a reef with as much as 2500 ms of buildup.
Sensitivity models were constructed of the various lithological
units including recent sediments, deep water marine carbonates,
platform carbonates, volcanics and basement types. The density
and velocity study suggested that the gravity data would
be useful for corroborating the seismic interpretation of
the reef boundary, but indicated that there was insufficient
detail regarding density to utilize the gravity data as
a strong independent method for testing the various (deemed-to-be)
equally-viable seismic interpretations. The sensitivity
modelling indicated that the magnetics could provide an
independent assessment of the shallowest possible level
of volcanics, which could then be used to infer the maximum
thickness of the "reef."
Two of the initial seismic interpretations for one of the
seismic bumps are shown in (Figs. 13a and 13b).
Note that the variations in potential "reefal" thickness
ranged from two separated buildups with 250 ms maximum isochron
in (Figs.13a to 2200 ms in 13b). The first task was to model
the primary regional magnetic field components attributed
to changes in basement composition and structure. The magnetic
depth estimates were obtained using MAGPROBE, which allows
the user to apply multiple depth estimation techniques in
a real time session. The corroboration of depth estimates
using differing algorithms allows a higher confidence factor
to be assigned to a given depth estimate than if a depth
is supported with only a single method. The magnetic depth
and susceptibility estimates were found to be in very close
overall agreement, with the primary magnetic basement surface
at approximately 5,000 meters, with some suggestions of
slightly shallower picks just under the "basement uplift".
These depth estimates were, in general, somewhat deeper
than the interpreted seismic basement, suggesting that acoustic
basement is somewhat shallower than magnetic basement, which
is a common occurrence.
Fig.14 illustrates the basement model
using the depth and susceptibility estimates discussed above.
The primary magnetic field range and character across this
line is due to intrabasement susceptibility changes. Three
primary basement compartments are evident, with the central
core being "acidic", that is low susceptibility (assigned
here as 1200 µ CGS) relative to the surrounding basement
rocks, which appear to be more "basaltic" in character,
with average susceptibilities in the range of 3500 to 4000
µ CGS. The regional magnetic field, containing most of the
magnetic field relief, was easily matched with no structural
edits.
Note that the calculated and observed fields match well
over much of the model, but a localized anomaly is clearly
evident in the central portion of the profile, coincident
with the reef (or volcano?). This initial portion of the
regional modelling was performed using a 2D assumption (infinite
extent along the strike, which is perpendicular to the plane
of the cross section). The subsequent modelling of the localized
feature was performed firstly in a 2.5D mode and, after
all seismic units were mapped both above and surrounding
the individual features, 3D modelling was performed over
certain of the "reefs" to fully fold in all available geometrical
information.
With a reasonable model (i.e., cross corroborated with
all available control) of the regional basement in place,
and the imposition of seismic constraints for the top of
the reef and surrounding sedimentary sequences, the number
of possible interpretations was decreased from a non-unique
solution (with an infinite number of possible solutions)
to a small set of feasible solutions. The next phase involved
the selection of plausible lithological scenarios for the
"reefal" section between top of magnetic basement and the
top of the "reef". The scenarios deemed feasible (based
on knowledge of similar features in surrounding areas) included:
1) acidic volcanics filling all volume up to top of "reef",
2) basaltic volcanics filling all volume up to top of "reef,"
3) reefal buildup seeded on a basaltic basement horst block,
or 4) reefal buildup seeded on acidic basement horst block.
Modelling of the reef as acidic volcanics indicated a rapid
departure in the anomaly gradient and amplitude between
the calculated and observed fields. While one should recognize
the amplitude could be suspect owing to lack of susceptibility
control, the departure in gradients is quite diagnostic.
This model was deemed implausible. Similarly, modelling
of the "reef" as being basaltic volcanics was a further
exaggeration of the acidic volcanics in both amplitude and,
particularly, gradients. This model was deemed implausible.
The next alternative examined the "reef" modeled as a reefal
buildup seeded on a basaltic basement horst block. In this
model the central basement compartment was replaced with
"basaltic" susceptibilities, including the uplifted section.
The response and dramatic departure from the observed further
supports this central basement block being "acidic". This
model was also deemed implausible.
Fig. 15 illustrates the preferred
interpretation, wherein the reef was seeded on an acidic
basement horst block. With the assigned constraints in place
including the seismic control (with corroboration from gravity
modelling results), the magnetic source depth estimates,
and a limitation on the possible scenarios based on similar
features in the region and rock properties from borehole
and outcrop samples, this interpretation is deemed the most
reliable model which adheres to the data. This adherence
was both local, as well as regional, i.e., no "local" changes
for the sake of a nice "curve" match were made. These results
were used, in part, to high grade this feature's prospectivity,
and it was selected as a drilling target. Drilling results
proved that the structure was, indeed, a reef, with thickness
very close to that of the integrated model.
Conclusions
By incorporating high resolution gravity and magnetics
into 3D seismic surveys and interpretations, there is a
positive impact on the final interpreted results. The cost
effective high resolution data can be used to constrain
and verify seismic velocities, particularly in problem areas
near complex structures and seismic shadow zones. The mechanics
of integration have been simplified by the development of
real-time, integrated workstation applications.
In addition, airborne gravity surveys are producing higher
resolution results than ever before, providing large regional
data sets for the optimization of seismic exploration costs.
Acknowledgments
We thank Gunther Newcombe of BP for his work on the Lombok
interpretation, Jack Weyand of Sidney Schafer & Associates
for his work on the density models, TGS and Geco-Prakla
for their top/base salt interpretation and VDIP velocity
grid, IGC for their Gulf of Mexico models, GDC for well
data and compilation, and Elizabeth Johnson and Mark Odegard
of Unocal Corporation for their Gulf of Mexico data and
interpretation.
References
1. LaFehr, T.R., Valliant, H.D., and MacQueen, J.D.: "High-Resolution
Marine Gravity by Digital Control", paper presented at the
SEG Annual Conference and Exhibition, New Orleans, 1992.
2. Valliant, H.D., Halpenny, J. and Cooper, R.V., 1985:
"A microprocessor-based controller and data acquisition
system for LaCoste & Romberg air-sea gravity meters"
Geophysics, 50, 840-845.
3. Saad, A.H., 1993: "Interactive Integrated Interpretation
of Gravity, Magnetic, and Seismic Data: Tools and Examples."
Offshore Technology Conference Proceedings, Paper 7079,
35-44.
4. Pawlowski, R., 1994,: "Emerging Workstation-based Potential
Field Methodologies. " The Leading Edge (SEG), June, 1994,
687-689.
5. Bain, J.E., Weyand, J., Horscroft, T.R., Saad, A.H.,
and Bulling, D.N., 1993,: "Complex Salt Features Resolved
by Integrating Seismic,Gravity, and Magnetics." EAEG/EAPG
1993 Annual Conference and Exhibition, expanded abstracts.
6. Bain, J.E., and Newcombe, G., 1994,: "Reducing Exploration
Risk Using Multi-Disciplinary Geophysical Methods Case Studies
From Indonesia and the Gulf of Mexico." Venezuelan Geophysical
Congress 1994, expanded abstracts.
|
