APPLIED GEOPHYYSICS, Vol.4, No.4 (December 2007), P. 288 - 293, 5 Figures.

DOI:10.1007/s11770-007-0041-8

Evaluation and extraction of weak gravity

and magnetic anomalies*
Liu Yunxiang

Abstract: In this paper, I introduce what are called weak gravity and magnetic anomalies
and propose standards for estimating their reliability. I also introduce new techniques for
processing this kind of weak anomaly. These techniques consist of interference elimination
and weak signal extraction. Practical applications have proved their effectiveness. Weak
gravity and magnetic anomalies will get more attention with the development of targeted
exploration.
Keywords: gravity and magnetic exploration, weak anomaly, anomaly extraction, less
balance filtering, distortion correction.

Introduction Reliability of gravity and

magnetic anomalies
Anomalies smaller than 2.5 times the overall anomaly
accuracy (the leveled overall anomaly precision) are In gravity and magnetic surveys, ambiguous
traditionally regarded as unreliable or noise during anomalies with small, hard to distinguish, amplitudes are
gravity and magnetic data processing and interpretation. regarded as weak anomalies. There is no definition of
As a result, little attention is normally paid to them. “weak gravity and magnetic anomalies” in the literature.
In order to improve the reliability of anomalies and I take gravity and magnetic anomalies with amplitudes
remove interference, data is commonly smoothed or between 1.0 – 2.5 times the overall anomaly precision as
filtered, which results in reduced anomaly resolution. weak gravity and magnetic anomalies.
In traditional processing methods, weak gravity and In traditional gravity and magnetic data processing
magnetic anomalies are usually ignored or removed. and interpretation, anomalies larger than 2.5 times the
Wi t h t h e r a p i d d e v e l o p m e n t s i n o i l a n d g a s overall anomaly precision measured by two stations are
exploration, geologists have placed greater demands regarded as reliable anomalies. Anomalies less than that
on gravity and magnetic exploration precision and are regarded as unreliable, caused by interference or
resolution (Guan et al., 2002), while traditional concepts observation errors, and they are generally ignored.
restrain the development of new techniques for gravity Are gravity and magnetic anomalies with amplitudes
and magnetic data processing and interpretation. The less than 2.5 times the overall anomaly precision
study of the gravity and magnetic anomaly signal-to- reliable? What are the criteria for evaluating the
noise (S/N) ratio and their processing techniques is reliability of weak gravity and magnetic anomalies for
inevitably helpful for improving the ability to solve present high-precision gravity and magnetic surveys?
geological issues. We discuss these questions below.

Manuscript received by the Editor September 8, 2007; revised received October 4, 2007.
*The subject is sponsored by the National 863Project Fund (Project No. 2006AA06Z201).
1. Bureau of Geophysical Prospecting, Zhuozhou 072751, China.

288
 Weak gravity and magnetic anomalies

Previously, theodolites were used for positioning and resulting from gravity and magnetic anomaly calculation
mechanical magnetometers for observations. At the are also independent of each other. At the same time,
present time, the type G856/858 magnetometer is used observations of the total magnetic field with magnetometers
for single station observations and GPS receivers are used of type G-856/G-858 for single stations also make the error
for positioning. As a result, the independence between of adjacent stations independent of each other. Application
gravity and magnetic stations is dramatically improved. of electrical gravimeters of type EG and CG-3/CG-5 can
So it is necessary to define new criteria for evaluating the also reduce the relationships between system errors caused
reliability of weak gravity and magnetic anomalies. by operation and starting and ending basic station data
Based on error theory and the probability distribution records. So the causes for error transmission are dramatically
function for normal distributions (Teaching and reduced with present observations.
Research Group of Higher Mathematics of Mathematics If the observation probability is x for an anomaly
Department of Zhejiang University, 1984), the percent measured by a single station, its reliability is x and
of measurement probabilities are 68.26 %, 86.6 %, 95.44 its unreliability percentage is (1-x). For anomalies
%, 98.8 %, 99.74 %, and 99.9 % when the data standard delineated by n independent stations with an equal
deviations are ±σ, ±1.5σ, ±2σ, ±2.5σ, ±3σ, and ±3.3σ, observation probability of x, their unreliability
respectively. Measurement probabilities of single gravity percentage is (1-x) n according to probability theory
and magnetic observation stations abide by this rule. (Teaching and Research Group of Higher Mathematics
At present, GPS application makes it possible that of Mathematics Department, 1984). Based on this result,
adjacent stations are independent of each other in terms of we obtain probabilities of anomalies caused by 1 to 5
errors, so positioning error and errors of the correction terms stations (Table 1).

Table 1 Relations among distribution area (given as number of stations observing the anomaly), amplitude,
and reliability percentage of gravity and magnetic anomalies

Anomaly at Anomaly at Anomaly at Anomaly at Anomaly at

Anomaly amplitude
single station two stations three stations four stations five stations
±σ 68.3% 89.9% 96.8% 99.0% 99.7%
±1.5σ 86.6% 98.2% 99.7% 99.9% 99.99%
±2σ 95.4% 99.7% 99.99% 99.999%
±2.5σ 98.8% 99.98% 99.99%
±3σ 99.73% 99.99%

For relatively independent observations, we define to the overall anomaly precision, then the reliability of
observation values as reliable when the measurement this weak anomaly is 99.996%. The reliability of weak
reliability percentage reaches 99.5% (only one anomalies is dramatically improved with increasing
probability percent exceeding the mean square error number of stations recording the anomaly,, even if its
among 200 observations). From the table, we see that a amplitude ranges between 1 to 2.5 times the overall
single anomaly with amplitude no less than three times anomaly precision. So, when evaluating the reliability
the mean square error is regarded as a reliable anomaly. of weak anomalies, we will consider both anomaly
Similarly, anomalies observed at two stations with amplitude and number of stations recording it.
amplitudes greater than two times the mean square error
are taken as reliable. Anomalies covering three stations
with amplitudes greater than 1.5 times the mean square Techniques for processing weak gravity
error are taken as reliable. Anomalies measured over five
stations with amplitudes greater than the mean square
and magnetic anomalies
error are taken as reliable. These observations establish
the criteria for evaluating reliability. Techniques for processing weak gravity and magnetic
Given the observations pattern, a magnetic anomaly anomaly include eliminating near surface interference
delineated by nine stations (three adjacent stations on and weak anomaly extraction. Before weak anomaly
each of three adjacent lines) with an amplitude equal extraction, we need to analyze and correct for the terrain
289
 Liu Yunxiang

and culture affects of the near surface. After experiencing variable density changes, anomalies after terrain
the pre-processing, the data should reflect the subsurface correction with a constant density will have terrain-
geologic structure. related amplitude features. In order to remove this kind of
distortion, I perform the following procedures for variable
Distortion correction for weak gravity and density correction: 1) analyze the relations between the
gravity anomalies and the near surface density distribution
magnetic anomalies and terrain fluctuations; 2) determine equivalent density
Before conducting data processing and interpretation I based on the surface layer distribution, density, and
first study the interference and distortions and eliminate relation between the anomaly and layering and plot the
them. For gravity and magnetic surveys, the following derived equivalent density distribution; and 3) determine
interference and distortions should be considered: 1) the corrections due to the middle layer using the variable
culture interference; 2) complex terrain corrections; 3) density and residual elevation and calculate the terrain
variable density correction; and 4) curvature correction effect with variable density (Liu, 2004).
to a plane datum in mountainous areas. Figure 1 shows a geologic map and the gravity
The variable density correction is a relatively new anomaly map before and after the variable density
technique (Liu, 2004; Yan et al., 2005; Ji and Zheng, correction. I conclude that the gravity anomaly after
2007). When the ground surface is rugged with lateral correction is smooth and without local distortion.

8
N21

-36
N22

6
-370

-36
-370
-368

N21

N1

-368
N1 -378 -376-374-372 -370
mgal

Fig. 1 Geology map (left) and gravity anomaly map (right) before and after variable density correction (red contours are before
the correction and blue contours are after the correction).
(Q - Quaternary, N22 - Upper Youshashan Group of Upper Miocene, N21 - Lower Youshashan Group of Upper Miocene, N1 - Lower Miocene)

Techniques for extracting weak gravity and effective to some extent.

Relevant filtering should be performed to extract
magnetic anomalies weak gravity and magnetic anomalies more effectively.
Weak gravity and magnetic anomalies feature small In traditional filtering methods to extract residual
amplitudes, so special techniques are necessary to anomalies, such as moving average filtering, the sum of
extract them. As we know, the second vertical derivative n

of potential fields can strengthen local weak anomalies, positive and negative coefficients is 1.0: ¦ (1/ n) (where
i 1
but the values or gradient are completely different n is the number of stations), in other words, the ratio of
from the original gravity and magnetic anomaly values positive and negative coefficients is 1.0 or (-1.0), which
and the derivative strongly magnifies interference and is called a balance filter. We take the Elkin III formula as
noise at high frequencies (Luo and Guo, 1991; Wang an example of the second vertical derivative method. Its
et al., 1991). So, the second vertical derivative has filtering coefficient is 60 or (-60) and it is a balance filter
limitations for extracting local gravity or magnetic as well (Chen, 1986).
anomalies. Publications about methods for extracting In order to better extract weak gravity and magnetic
weak gravity and magnetic anomaly are few. Liu et al. anomalies, I propose less balance filtering for the first
(1996) proposed a method for analyzing weak gravity time, Less balance filtering is based on the coefficients
and magnetic signals based on statistical analysis. It is of local anomalies wich are different than those for
290
 Weak gravity and magnetic anomalies

regional anomalies. To be similar to the coefficients of spacing), the wavelength is equivalent to the radius
the second vertical derivative method or field average of the calculation circle; and 3) the regional field is
filtering and to improve local weak information, the 2D independent of the local field and the long wavelength
less-balance filter is designed for the spatial domain. weighted negative coefficients are large. This filter can
The weight coefficients (Chen, 1986) are a0, a1, …, a8 for strengthen local weak signals and remedy the amplitude
circles with grid space radii of 0, 1, 2, 3, 4, 5, 6, 7, and 8, loss of weak anomalies.
with respect to the survey station at the center. The ratios Figure 2 shows a reduced to pole anomaly of total
of their positive and negative coefficients are not equal magnetic intensity in a test area. Figure 3 shows residual
to each other. The expression is: anomalies obtained using an iterated sliding filter
gr = a0g0 +a1g1 + a2g2 +a3g3 + a4g4 (window size of 5 by 5 and 12 iterations) and using a
+ a5g5 + a6g6 + a7g7 + a8g8 (1) less-balance filter. From the comparison, we conclude
that the residual anomaly map obtained using the less-
where gr is the residual anomaly; g0 is the anomaly of the balance filter better reflects the weak anomaly detail with
calculated station; g1, g2, …, g8 are the average values of the two magnetic highs within a magnetic high belt in the
field at radii of 1, 2, …, and 8, respectively. The coefficient middle part of the area.
group can take different values based on the amplitude and
area of the weak gravity and magnetic anomalies.
Here are two groups of filter coefficients (obtained
by analogy): (1.04, 0, -0.02, -0.08, -0.12, -0.18, -0.22, 0
-0.24, -0.14) and (0.88, 0.18, 0, -0.06, -0.12, -0.18, -0.22,

50
-0.24, -0.18). We take the first group of coefficients as an
example. The coefficient group of the local field can be
regarded as (1.04, 0, 0, 0, 0, 0, 0, 0, 0), while that of the

25
regional field is (0, 0, -0.02, -0.08, -0.12, -0.18, -0.22,
-0.24, -0.14). So we can conclude this kind of filter is -75 -50 -25

0
characterized by: 1) the sum of positive coefficients of
the local field is slightly larger than the absolute sum nT 5 10 15km
of the negative coefficients of the regional field; 2) the
Fig. 2 Reduction to pole anomaly of total magnetic intensity
wavelength of the local field is short (less than a grid in a test area.

Fig. 3 Residual anomaly comparison (left: sliding filter, right: less-balance filter)

Application of the technique for extracting ms-2, while gas reservoirs can cause gravity anomalies
weak gravity and magnetic anomaly reaching 10 – 200 × 10-8 ms-2 (Li, 1997). In some areas,
such as the Sanhu area in the Qaidam basin, the gas bed
is shallowly buried (less than 2000 m) and large in scale
The first example is a study of a weak gravity anomaly and causes gravity anomalies with amplitudes reaching
responding to a gas reservoir. Gravity anomalies caused 200 – 1000 × 10-8 ms-2 (Fan and Li, 1997).
by oil reservoirs are generally weak, about 5 – 30 × 10-8 In an area of the Tarim basin, a high precision gravity
291
 Liu Yunxiang

survey was conducted with a survey grid of 250 m × high reliability. The locations of these weak gravity
250 m and an observation precision that reaches 20 anomalies correspond to Lower Tertiary to Cretaceous
× 10-8 ms-2. The work area features a flat surface with normal structures and the main body of these weak
no interference. The less balance filter was used to anomalies agrees well with the location of the gas
extract weak gravity anomalies of 30 – 40 × 10-8 ms-2. reservoir. So, the weak gravity anomalies indicate the
Although the gravity anomalies are weak, they have a presence of a highly enriched Tertiary gas reservoir at
large scale and are orderly distributed, so they possess 3500 m depth (Figure 4).

Weak Gravity Anomaly

-0.1
0

1 5
-0. -0.0 0 r Tertiaryervior
L owe as Res
u c t u re of G
Str

0 0.05
10-5ms-2 0 5 10 km

Fig. 4 Weak gravity anomaly extraction.

The second example is a study of a weak magnetic 5

anomaly corresponding to igneous rocks. High precision 5
0
magnetic surveys play an important role in studying
igneous rocks. Figure 5 shows the extracted weak
magnetic anomaly of an area in the Junggar basin, The 0
contours in Figure 5 come from right plot of Figure 0 D8 0
3. In this area, igneous rocks developed during the D5
Carboniferous are 3300 m deep and the reservoirs D14
are intermediate-acidic igneous rocks with moderate 0
magnetism. The survey grid is 250 m × 500 m and total
anomaly precision reaches 1.2 nT (see the right panel nT
0 5 10 15 km
of figure 2). Weak magnetic anomalies with amplitudes
of 3 – 4 nT are extracted using less balance filtering. Fig. 5 Extracted weak magnetic anomaly.
The surface area is flat with no interference and the Well D8 has a gas show, D5 is a low production gas well,
anomalies are large in scale, so the reliability of these and D14 is high-production commercial gas well. The red
contour is approximately the boundry of the igneous rocks.
anomalies is very high. The weak magnetic anomalies
reflect the igneous rock distribution. Theoretical
forward modeling shows the weak anomalies can be Conclusions
caused by intermediate-acidic igneous rocks with
magnetic intensity of 350 × 10-3 A/m (moderate to weak
magnetism) at depths of 3300 – 3800 m. Discovery 1. Criteria for evaluating weak gravity and magnetic
of these anomalies has significant meaning for further anomalies presented in this paper possess significant
hydrocarbon exploration in this area. practical meaning for hydrocarbon exploration since
In Figure 5, wells D5 and D8 are old exploration hydrocarbon-related anomalies are generally weak
wells. D5 is a low production gas well and D8 had a gas gravity and/or magnetic anomalies.
show. Well D14 is a new well based on 3D seismic and 2. The variable density gravity correction and less
magnetic data. High-production gas was encountered balance filter proposed in this paper provide matching
at depth where the Carboniferous igneous is developed. techniques for weak gravity and magnetic anomaly
The discovery of the D14 gas reservoir is a breakthrough analysis.
for gas exploration in this area. 3. The analysis of observed data shows that weak
292
 Weak gravity and magnetic anomalies

gravity and magnetic anomalies contain crucial north Tarim basin: Geophysical Prospecting for
information for studying hydrocarbon and igneous rocks. Petroleum (in Chinese), 35 (2), 81 – 88.
4. The techniques for extracting weak gravity and Luo, X. K., and Guo, S.Y., (Chief-Editors), 1991,
magnetic anomalies are effective. Applied geophysics tutorial-gravity, magnetic method
of exploration: Geological Publishing House, Beijing.
Teaching and Research Group of Higher Mathematics
References of Mathematics Department of Zhejiang University,
1984, Engineering math probability and math physical
statistics: High-Education Press, Hangzhou, China,
Chen, S., 1986, Gravity exploration: Geological 117 – 118.
Publishing House, Beijing, China. Wang, J. L., Wang, Y. X., and Wan, M. H. (Eds), 1991,
Fan, Z. P., and Li, G. Y., 1997, Application of high- Petro-gravity & magnetic interpretation: Petroleum
precision gravity survey in east area of Qaidam basin: Industry Press, Beijing.
Petroleum Geophysical Prospecting (in Chinese), 32 Yan, L. J., Wan, P., and Yao, C. L., 2005, Variable
(Supplement 2), 125 – 133, 139. density correction method and its application in
Guan, Z. N., Hao, T. Y., and Yao, C. L., 2002, Gravity calculating Bouguer gravity anomaly: Engineering
and magnetic exploration prospect for 21 century: Geophysical Transaction (in Chinese), 2 (3), 177 –
Progress in Geophysics (in Chinese), 17 (2), 237 – 180.
244.
Ji, L. S., and Zheng, L., 2007, High-precision gravity
correction methods for loess plateau areas, a case Liu Yunxiang, Senior Engineer, graduated from
study: Applied Geophysics, 4 (2), 89 – 93. Changchun Geological Institute
Li, Q. Z., 1997, Discussion of direct delineating in 1988 with a bachelor degree in
hydrocarbon with Aifei micro-gravity method and Geophysical Exploration. He received
GONG method: Petroleum Geophysical Prospecting a PhD in Geophysical Exploration
(in Chinese), 32 (2), 277 – 302. from the China University of
Liu, Y. X., 2004, Variable density gravity correction Geoscience in Beijing in 2007. He is
techniques in complex area: CPS/SEG 2004 Beijing currently Deputy-General Engineer of
Internat. Mtg., Expanded Abstracts, 467 – 469. Non-Seismic Surveys, BGP, CNPC.
Liu, T. Y., Cui, N., Qiu, S. D., and Hou, W. G., 1996, His research interests include gravity and magnetic
Weak signal analysis in high precision gravity and exploration. E-mail: lyx933@sina.com.
magnetic survey and its application in Yakela area,

293
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.