Quality of airborne geophysical surveys and reliability of geological interpretation

It is shown that improvements in the quality of airborne geophysical surveys (instrumentation and technologies) depend directly on the quality, reliability and completeness of the geological interpretation of the resulting data, primarily the high-frequency component of the geophysical field spectrum. A landscape-geological approach to data interpretation is briefly described. Examples and techniques for interpreting the high-frequency components of the gamma-ray and magnetic fields are presented. It is concluded that the approach described makes it possible not only to identify low-contrast anomalies at the level of “interference”, but also to obtain additional, highly important information from so-called “geological noise”.

Improving the accuracy of geophysical surveys increases the likelihood that low-contrast targets will be reflected in the measurements, but also sharply increases the number of recorded low-intensity local geophysical anomalies. Together, these form the natural background variations in the measured field, or so-called “geological noise”. Accordingly, the probability that any identified anomaly belongs to the specified exploration target is substantially reduced. In a number of cases, the distributions of anomaly and “noise” parameters coincide not only in signal amplitude but also in autocorrelation radius (1). This makes it impossible to separate useful anomalies adequately from noise at the interpretation stage, resulting either in substantial expenditure on investigating all anomalies using other methods (including excavation and drilling), or in widespread failure to detect exploration targets.

In any event, the circumstance described becomes an obstacle to further improvements in survey quality. It has a particularly strong effect on the effectiveness of airborne geophysical surveys, since these are usually conducted over extensive areas in which hundreds or thousands of local variations in the measured field may be identified that actually satisfy the criteria for anomalies. A similar situation was described previously (2) and was regarded as a condition rendering further improvements in geophysical survey quality impractical.

Our completed and ongoing studies (4) of the structure of background variations in various geophysical fields measured by high-accuracy methods suggest that, in most cases examined, the merits of improving airborne geophysical survey quality depend directly on the quality and reliability of geological interpretation specifically of the relatively low-contrast part of the spectrum of geophysical field variability. Here, the quality of geological interpretation means the ability to develop a satisfactory interpretation model, and reliability means its adequate implementation over the entire survey area.

Unlike traditional schemes for geological interpretation of data, which typically interpret each identified metrically reliable anomaly individually, the technology considered below may be termed landscape-geological interpretation of low-contrast background field variations, or, in abbreviated form, landscape-geological reduction. (4)

The technology under consideration rests on two basic propositions. The first is that a limited number of landscape-geological factors collectively determine metrically reliable background field variations in specific survey areas. The second is that most of these factors can be identified and modelled on the basis of limited additional data and knowledge.

Landscape-geological reduction of airborne gamma-ray spectrometric data is carried out by quantitatively removing from the observed fields their systematic components associated with known landscape-geological features of the area, or with features that can be mapped relatively easily. The algorithm is based on the following operations (4):

• dividing the area into provisionally homogeneous zones on the basis of a priori landscape-geological characteristics;
• determining background radiogeochemical parameters for each landscape-geological type of zone;
• reducing the observed values of the spectral components of the gamma-ray field using individual background parameters for each landscape-geological type of zone. The input data for reduction are the airborne gamma-ray spectrometric survey results and a landscape-geological map of the study area. This map is compiled by spatially superimposing the boundaries of a geological map, a map of genetic landscape types and a map of plant associations. The latter two maps can readily be obtained with the required accuracy through specialised interpretation of either accompanying aerial photography or acquired photographs. For this type of reduction, the three factors named above (rock type, landscape type and plant association) account for virtually all causes of background field variations. Possible inaccuracies in identifying geological complexes and delineating geological boundaries become a specific subject for investigation and correction through successive iterations of the landscape-geological reduction procedure.

Application of the technology described in areas prospective for kimberlites in western Yakutia and Arkhangelsk Region made it possible, for the first time, to identify secondary radiogeochemical halos associated with kimberlite bodies. Subsequent comparison of their configuration and intensity with the original data showed that these halos were reflected in the original data only in the highest-contrast zones.

This example confirms both the proposition that a limited number of landscape-geological factors determine metrically reliable background field variations in specific survey areas, and the proposition that most of these factors can be identified and modelled using limited additional data and knowledge. It also illustrates the effectiveness of landscape-geological interpretation of low-contrast background field variations, which led to the identification of a new predictive factor. Interestingly, after these results were obtained, it became necessary to improve the quality of airborne gamma-ray spectrometric surveys further by increasing the NaI crystal volume to 48 and 60 litres.

The search for kimberlite bodies in areas of trap-rock occurrence is one of the principal problems in diamond geology. This is due to the very frequent spatial association of kimberlites and trap rocks, the stronger magnetization of the latter, and the fact that the pipes generally occur beneath the trap rocks.

Aeromagnetic surveys conducted by various organisations in different years over areas of trap-rock occurrence showed that the intensity of the local magnetic-field component virtually precludes identification of local pipe-related anomalies against this background. The ranges of variation in autocorrelation radii and horizontal-gradient values were also found to be virtually identical for anomalies associated with pipes and trap rocks.

Our investigation of the causes of strong magnetic-field differentiation in areas where trap sills are widespread established that the principal factor controlling this variability is their block structure and abrupt changes in the elevation of the sill base between adjacent blocks. Individual fractures separating sill blocks were found collectively to form three distinct subparallel systems, corresponding to longitudinal, transverse and diagonal prototectonic fracturing of the sill. Most importantly, these fractures were found to be very clearly expressed in aerial photographs (3).

This last finding proved decisive in developing a rapid technique for reducing aeromagnetic survey data from areas of trap-rock occurrence. Specialised interpretation of aerial photographs was used to compile tectonic diagrams of the sills and estimate sill thickness in each block. A fundamentally new computer technology, TRAPP (ТРАПП), was developed for volumetric (3 D) modelling of the magnetic field of a geological model, calculating field values at specific points along the actual flight lines and at the actual flight altitudes. A computer technology, NAGYTS (НАГЫЦ), was developed for iterative solution of the problem in areas where model parameters were poorly defined.

Interestingly, analysis of the comparative plots presented here and elsewhere suggested that their partial discrepancies might be due to insufficient accuracy in positioning observation points along lines on the ground. This and other circumstances necessitated a sharp improvement in airborne geophysical survey quality in areas of trap-rock occurrence. Thus, in this case too, high-quality geological interpretation of the data led to increased requirements for survey quality.

It should be emphasised that in this case too, the two basic propositions formulated at the beginning of the article for landscape-geological interpretation of low-contrast background field variations proved valid.

Processing the results of high-accuracy aeromagnetic surveys over a number of areas showed that the corresponding digital data always contain a high-frequency part of the magnetic-field spectrum extremely rich in small anomalies, constituting its “fine” structure. These are magnetic anomalies several hundred metres wide, with intensities ranging from tenths of a nT to a few nT. Traditionally, this highest-frequency part of the recorded field spectrum is regarded as interference and disregarded.

The causes of this part of the spectrum can provisionally be divided into three groups. The first is apparently associated with random measurement errors and distortions arising when profile data are converted to a regular grid. An important characteristic of anomalous points caused by these factors is their randomness and lack of spatial correlation. The second group of causes produces anomalies and anomalous points through the effects of topography, heterogeneity in the facies composition of unconsolidated deposits, and anthropogenic factors. This group of anomalies often exhibits relatively elevated gradients where spatial correlation is present. The third part of the “fine” field structure comprises anomalies genetically associated, directly or indirectly, with local heterogeneities in low-contrast bedrock. In studies of platform areas with a substantial thickness of sedimentary cover, these anomalies are directly associated with heterogeneities in the sedimentary-cover rocks. Intensely deformed, strongly magnetized crystalline basement rocks do not affect this part of the magnetic-field spectrum at depths greater than one kilometre.

A dedicated software package, MILLIGAN (МИЛЛИГАН, MIcrolocal LInear Geophysical ANomalies), was developed to identify and analyse such anomalies. Its data-processing technology searches a digital field model for microlocal linear anomalies of various strikes and compiles maps of these anomalies and their spatial distribution density. This stage is fully formalised, performed computationally and requires no substantial labour. Digital data from any high-accuracy, large-scale magnetic survey can, if necessary, be reprocessed and analysed from the perspective outlined above. Typically, several hundred microlocal anomalies can be identified per thousand square kilometres of survey coverage.

Where basement depth is sufficiently great, band-pass filtering of the magnetic field makes it possible to separate microlocal anomalies into groups according to their horizontal dimensions and analyse them individually. This approach allows anomalies associated with near-surface heterogeneities to be isolated as a separate group, while the remaining anomalies are used to map zones of heterogeneity in the cover at various depths and examine them in three dimensions.

Without discussing in detail the full range of possible geological causes of such low-contrast heterogeneities in platform cover, we briefly consider how the technology identifies systems of tectonic disturbances in the cover. These disturbances are usually of very, very low contrast, and their detection requires substantial expenditure on seismic surveying and drilling.

The technology for identifying faults in platform cover is based on the following principal concepts:

• a fault forms a complex three-dimensional geological body—the zone of active fault influence—formed by an assemblage of shear fractures, tensile fractures and other fractures of various orders;
• fracture orientation relative to the overall fault direction is determined primarily by the directions of the principal stress vectors, as well as by rock mechanical properties, anisotropy and so forth, and therefore contains information on these factors;
• the specific mechanical and physicochemical conditions in fracture zones provide the prerequisites for variations in the petrophysical properties of cover rocks that are spatially associated with these zones
• the small dimensions of individual fractures and small changes in the total magnetization of the adjacent rock volumes determine the very weak intensity of the corresponding anomalies in aeromagnetic data. These particular features of faults (increased lineament density and a characteristic pattern) provide a basis for their identification and study using lineament maps produced with MILLIGAN. Linear areas of elevated microanomaly density are interpreted as zones of disruption in the sedimentary cover, and the geometric axes of these zones as fault axes. The characteristic lineament pattern within the zones enables a fault to be recognised and subjected to tectonophysical interpretation, with faults classified by the dominant orientations of individual disruption elements relative to the overall direction. Where displacement is strike-slip, the developmental stage of a given disruption zone can be estimated. If the entire range of possible relative displacements of rock blocks is considered, classification by the direction of principal shear stresses should be used. This makes it possible to determine the direction of the horizontal component of the true displacement and can be used to address a number of geodynamic problems.

The technology described allows the relative reservoir properties of both fault zones and areas (blocks) of sedimentary-cover rocks to be assessed by classification according to the distribution density of disruption elements. At the final interpretation stage, density maps for individual directions are superimposed, involving non-formalised interpretation and generalisation of the resulting structural pattern of the area.

The technology described was used to process aeromagnetic survey data for several areas and yielded important geological results in every case. In a study of the Serpukhov District of Moscow Region aimed at identifying tectonic structures in the area of a groundwater deposit, the tectonic disturbances identified in the cover were confirmed by drilling.

It appears evident that the geological information obtained with this technology is necessary both for general geological mapping and for exploration for mineral resources hosted in platform deposits (groundwater, hydrocarbons, diamonds, ores of sedimentary-infiltration origin, and others).

The word “geophysical” in the acronym MILLIGAN means that, in principle, this processing can be applied not only to magnetic survey results but also to airborne gamma-ray spectrometric survey data, digital terrain models and so on. To obtain information on fault fluid activity, MILLIGAN was used to analyse airborne gamma-ray spectrometric survey data, specifically the uranium channel after correction for atmospheric radon. Good correlation between the corresponding structural patterns was observed in a number of cases, indicating the presence of emanating faults.

The data presented in the third example again indicate that the number of natural factors affecting differentiation of the local high-frequency field component is limited and that these factors can be identified reasonably rapidly on the basis of a priori knowledge. In this example, that knowledge concerned the specific geological structure of faults.

These examples reaffirm the statement made at the beginning of the article: opportunities to improve airborne geophysical survey quality are directly linked to improvements in the quality and reliability of geological interpretation. To some extent, the dialectics of this relationship can be compared with the familiar relationship between projectile power and armour thickness. The main focus here is on geological interpretation of the high-frequency component of geophysical field spectra using landscape-geological interpretation of low-contrast background field variations (landscape-geological reduction). This enables not only the “identification of useful anomalies at the interference level”, but also the extraction of geological information of practical importance from the highest-frequency part of the field itself (geological noise). A distinctive feature of this approach is the incorporation into the workflow of additional data not normally used in airborne geophysics, such as landscape-geological maps and image-interpretation diagrams. Nevertheless, the nature of geophysical fields itself dictates this step, while the means of obtaining such information—remote and rapid—indicate that adding non-geophysical methods to an airborne geophysical survey entails extremely low additional costs.

References

1. Nikitin A.A. Theoretical foundations of geophysical information processing. Moscow: Nedra, 1986, 281 pp.

2. Tarkhov A.G., Bondarenko V.M., Nikitin A.A. Principles of method integration in exploration geophysics. Moscow: Nedra, 1977, 314 pp.

3. Tsyganov V.A., Mogilevsky V.E., Kulikov N.F. Capabilities of aeromagnetic surveying in the search for kimberlite bodies in areas of trap-rock occurrence. In: New developments in creating effective methods of ore geophysics. Proceedings of TsNIGRI, Issue 222, Moscow, 1988, pp. 92-97.

4. Tsyganov V.A. Reliability of geological exploration systems. Moscow: Nedra, 1994, 299 pp., illustrated.