On the Structure of Weak Magnetic Fields of Vein-Type Gold Deposits Localized in Sedimentary Rocks

(Using the Lyubavinskoye Deposit of Transbaikalia as an Example)

B. G. Mordanov, Z. A. Myshko, L. Ya. Erofeev

(Submitted by the Department of Geophysical Exploration Methods)

A significant part of vein-type gold deposits is confined to sedimentary and sedimentary-metamorphic rocks, where the ore bodies are represented by low-thickness quartz veins with a poor sulfide content. At such deposits, geophysical methods, especially the widely used electrical-prospecting methods, do not yield positive results.

In recent years, the high effectiveness of detailed magnetic surveying has been established in the study of ore fields of a number of gold deposits localized in intrusive rocks.

In this connection, we began special magnetometric investigations at gold deposits whose ore fields are situated in sedimentary rocks. The aim of the work being carried out was to clarify the principal features of the structure of the magnetic fields, with a subsequent assessment of the possibilities of applying detailed magnetic surveying to the study of deposits of this type.

In the present article only the first part of the question will be discussed, i.e., the structure of the magnetic field. As an example, the Nikolaevsky sector was selected, which is the eastern flank of the Lyubavinskoye gold-ore deposit of Eastern Transbaikalia.

Within the sector, strongly metamorphosed sedimentary rocks are developed: sandstones–siltstones and shales of Permian age. Tectonically, the sector is confined to a zone of a deep-seated latitudinal fault and is part of the southern limb of a synclinal fold. The fault is fixed by a chain of granite stocks, a belt of dikes of quartz porphyries and granite-porphyries, and a concentration of ore mineralization within a belt about 1 km wide.

The sector is characterized by the absence of granite stocks cropping out at the surface, by the northward dip of the rocks, and by the confinement of ore bodies to gently dipping feathering fractures of latitudinal faults. The disturbances encountered are longitudinal and diagonal, more rarely transverse. The main ore zones have variable strike from 330° to 260°, an irregular thickness from several cm to 1–2 m, and flatten out with depth. Gold is distributed in the form of disseminations, and its confinement is noted to small longitudinal or oblique fractures, as well as to areas of sharp changes in vein thickness. The bedrock of the sector is overlain by loose deposits, the thickness of which varies from 0.5 to 2 m on watersheds and from 10 to 12 m in valleys.

The magnetic survey was carried out with an M-23 magnetometer over an area of 2 sq. km, on a 50×5 m and 100×5 m grid, with an accuracy of ±4γ. To determine

of the character of the change in the increment of the vertical component of the magnetic-field intensity in the Nikolayevsk sector, statistical methods of analysis were applied, making it possible to estimate:

1) the magnitude and complexity of the change in the amplitude of the increment of the vertical component;
2) the forms of change in the magnetic-field intensity;
3) the anisotropy of the magnetic field.

The estimate of the change in the magnitude of the amplitude of the increment of the vertical component—$\Delta Z$—was carried out as follows. The values of $\Delta Z$ were grouped into statistical series. The grouping interval for constructing the series was determined by Sturges’ formula:

\[ d=\frac{\Delta Z_{\max}-\Delta Z_{\min}}{1+3{,}322\lg N}, \]

(where $N$ is the number of $\Delta Z$ values in the statistical series) and proved to be equal to 5 m. From the statistical series, distribution polygons were constructed for six profiles and for the sector as a whole (Fig. 1, a).

The range of variation of the amplitudes of the increment of the vertical component of the magnetic-field intensity (without taking into account the anomaly over the granite-stock) is 60γ, the largest value being $+35\gamma$, the smallest $-25\gamma$. In 90% of all $\Delta Z$ values, the change in amplitudes occurs only in the interval from $-10$ to $10\gamma$, i.e., within 20γ.

The distribution polygons of the amplitudes $\Delta Z$ by profiles differ somewhat. For the western profiles the form of the distribution polygons is distinguished by slight positive asymmetry and is close to a flat-topped one. On the distribution polygons of the central part, two maxima are already clearly outlined, with weak negative (relative to the mean) asymmetry. For the eastern profiles a bimodal distribution with a sharply expressed right maximum is characteristic, as a result of which the form of the polygons becomes sharply peaked.

The graphs of cumulative frequencies, constructed on normal probability paper, according to the Kolmogorov criterion ($\lambda<1{,}35$), may be assigned to straight lines, which indicates that the empirical distribution is close to normal.

To estimate the complexity of the change in the amplitude $\Delta Z$, the normalized entropy was calculated by formula (1):

\[ H=\frac{-\sum p_i\cdot \ln p_i+\frac{N-1}{2n}}{\ln N}, \]

where $n$ is the number of measurement points on the profile for which the entropy was determined, $p$ is the probability of the grouping class, and $N$ is the number of classes in the distribution.

The normalized entropy—$H$, characterizing the complexity of the change in amplitudes $\Delta Z$, is 0.81–0.87 in the sector. The greatest entropy value, 0.85–0.87, is observed on the outermost eastern and western profiles, where the distribution polygons differ in character. However, the fluctuations of entropy are small, lying within 0.06, which indicates the uniformity of the change in amplitudes $\Delta Z$ in the sector as a whole.

The study of the form of change in the magnetic-field intensity was carried out as follows. Autocorrelation functions were calculated along the profiles

\[ R(\tau)= \frac{ \sum_{i=1}^{N-1}(x_i-x_{\mathrm{cp}})(x_{i+\tau}-x_{\mathrm{cp}}) }{ \sum_{i=1}^{N-1}(x_i-x_{\mathrm{cp}})^2 }, \]

where \(x'_{\mathrm{av}}\) is the mean value of \(\Delta Z'\) on the profile,
\(x_i\) is the value of \(\Delta Z'\) at the \(i\)-th point \((i=1\ 2\ 3\ldots N)\);
\(N\) is the number of points;
\(\tau\) is the correlation interval, \(\tau=\Delta x\cdot p\);
\(\Delta x\) is the discretization interval, equal to the measurement step;
\(p\) is the scale factor.

Fig. 1. a. Distribution polygons of the amplitudes of the observed magnetic field: I, II... VI — ordinal numbers of profiles, VII — total polygon, \(n\) — number of measurement points on the profile, \(n'\) — number of points in the grouping interval, \(H\) — entropy.
b. Autocorrelation functions: 1 — \(R(\tau)\) of the observed field \(\Delta Z\), 2 — \(R'(\tau)\) of the transformed field \(\Delta Z'\).

Since, in the magnetic field of the gold-ore deposits of the vein type studied earlier, there was a number of stable periodic components, the calculation of the autocorrelation function \(R(\tau)\) was carried out up to large comparison intervals \(\tau\) (0.8 of the profile length), which, when the spectral density is calculated from such \(R(\tau)\), is equivalent to constructing periodograms. We did not calculate the spectral density; the parameters of the periodic variations \(\Delta Z\) (amplitude and period) were determined directly from the graph of \(R(\tau)\), as is recommended in [2].

To obtain information on the spectral composition of the variations \(\Delta Z\), the periodic components must be subtracted from the total spectrum; this was done as follows: from the autocorrelation curve \(R(\tau)\) a periodic function was isolated, its mean amplitude—\(A\)—and period—\(T\)—were determined, and from these a cosine wave was constructed on tracing paper. The latter was superimposed on the graph of \(\Delta Z\), and by horizontal displacement the visually best coincidence of the curves was selected, i.e., the position for which the mutual-correlation coefficient was greatest.

By subtracting the cosine wave from the graph of \(\Delta Z\), a curve \(\Delta Z'\) was obtained, and for it the autocorrelation function \(R'(\tau)\) was calculated. If a periodic component was observed in its character, the isolation process was continued in the same way, i.e., \(A'\) and \(T'\) were determined, a new periodic component was constructed, it was subtracted from \(\Delta Z'\), and from the remaining values \(R''(\tau)\) was calculated, etc.

As a rule, after just one subtraction of a periodic component, the autocorrelation function assumed the form of a graph decreasing monotonically to zero, which, with further increase of \(\tau\), remained close to zero (Fig. 1, b).

From the functions \(R'(\tau)\), in an interval \(\tau\) equal to 15% of the realization length, the mean autocorrelation function was constructed; it was approximated by the exponential function \(y = le^{-d(\tau)}\) (Fig. 2, a), and from it the spectral-density graph was calculated (Fig. 2, b).

As a result of the spectral analysis, two periodic

Fig. 2. a—Autocorrelation functions: 1—mean over the segment, 2—theoretical; b—graph of the spectral density \(S(\omega)\).

components with frequencies of 0.0008 cycles/m and 0.0014 cycles/m, which corresponds to anomaly widths of 300–800 m.

The amplitude of the anomalies is 6–8 γ, i.e., 1/10 of the change in the amplitude of ΔZ over the plan.

In addition to the periodic variations distinguished here, isolated changes in intensity also occur; on the plans of the graphs they appear in a quite obvious form. We did not include the field areas where such anomalies appear in the statistical sample, since they constitute pronounced “inhomogeneities.” These are anomalies above the granite stock (width 160 m and intensity +340 γ) and above convergent dikes of basic composition (anomaly width 40 m, intensity +100 γ).

It is difficult to establish any features in the high-frequency part of the spectrum against the general background of ΔZ variations: the spectral density decreases smoothly with increasing frequency. This is how the field appears in the high-frequency part if it is studied in a single section, i.e., in two-dimensional space. Analyzing the field in three-dimensional space (which is practically impossible to carry out quantitatively), one can, purely qualitatively, detect a number of characteristic disturbances in the intensity variations in this part as well.

These are linearly elongated anomalous zones with a ratio of axis lengths of 1:1000 and more. The width of the anomalies is variable and ranges from 0.5 to 20 m. The amplitude varies within the limits of −10 and −25 γ. The predominant strike is northeastern; the length is from several meters to kilometers. The distribution of such anomalies over the area is relatively uniform, since the site lies entirely within the ore field.

To assess the anisotropy of the field, a micromagnetic survey was carried out at 10 sites measuring 20×20 m with a spacing of 2 m. The sites were surveyed in the area of development of the sandy-shale sequence, the granodiorite intrusion, and the dike belt. The isodynamic-line plans were processed with a protractor: every 0.5 cm the azimuths of directions along the isodynamic lines were measured with an accuracy of 2°. From these data, distributions were constructed in the form of a rose diagram of directions along the ΔZ isodynamic lines (Fig. 3).

Fig. 3. Rose diagram of directions along the ΔZ isodynamic lines above the dike belt.

For the sandy-shale sequence, the character of the distribution of directions is close to uniform. However, two small maxima are outlined, corresponding to azimuths of 15° and 345°, i.e., a submeridional direction predominates. According to the rose diagram, the dike belt is characterized by two maxima with azimuths of 0° and 50°. For the granite massif, the principal large maximum with an azimuth of 350° is clearly localized. On the whole, the observed magnetic field proved to be anisotropic.

Thus, according to the results of the magnetic survey in the Nikolaevskoye

In the area, insignificant changes are noted in the increments of the amplitudes of the vertical component of the magnetic-field intensity: 90% of all values of $\Delta Z$ lie within the limits from $-10$ to $+10$ gammas. However, definite regularities are apparent in the character of the changes in the magnetic field, consisting in the following:

1. The distribution law of the amplitudes of the observed magnetic field is close to normal. The complexity of the amplitude variation, estimated by entropy $(H = 0.81\text{--}0.87)$, is almost identical over the area; the spectrum is characterized by a deep minimum.

2. The magnetic field of the area is anisotropic, with a predominance of two rather clearly expressed directions of $\Delta Z$ isolines: submeridional and northeastern.

3. The magnetic field contains periodic components with a frequency of $0.0008\text{--}0.0014$ cycles/m, corresponding to anomalies $300\text{--}800$ m wide with an amplitude of $6\text{--}8\,\gamma$.

4. Against the background of a weak magnetic field, areal anomalies with an intensity of $+340\,\gamma$ and narrow, linearly elongated anomalies with an amplitude of $-10\text{--}25\,\gamma$ appear.

REFERENCES

1. B. V. Shilin. Quantitative assessment of the complexity of geophysical and geological maps. Exploration and Protection of the Subsurface, 1966, No. 5.

2. M. G. Serebrennikov, L. A. Pervozvanskii. Detection of hidden periodicities. Izd. “Nauka” FML, Moscow, 1965.