Computational algorithm for downward continuation of gravity anomalies Preprint submitted to arXiv
D. K. Ivanov, L. N. Temirbekova, P. N. Vabishchevich · 2025
Read Computational algorithm for downward continuation of gravity anomalies Preprint submitted to arXiv by D. K. Ivanov, L. N. Temirbekova, P. N. Vabishchevich (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2507-08506.txt.
Introduction
Ivanov and co-authors present an algorithm that pushes gravity readings from the surface down toward their hidden sources. Geologists use such downward continuation to locate the bodies that cause gravity anomalies. The task is unstable by nature, since small errors in the data can blow up as the field is extended downward.
The method writes the deep field as the pull of a thin, simple layer, or of its vertical change. Keeping the layer's density to one sign keeps the anomalies' signs intact, as long as the layer wraps all the sources. A non-negative least squares solver enforces that sign rule in practice. Model examples show the algorithm recovering the buried sources well. The authors frame it as a steadier route into the subsurface.
Catalog record
| CREATOR | D. K. Ivanov, L. N. Temirbekova, P. N. Vabishchevich |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 25 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2507-08506
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Computational algorithm for downward continuation of gravity anomalies Preprint submitted to arXiv D. K. Ivanov, L. N. Temirbekova, P. N. Vabishchevich Abstract The downward continuation of potential fields from the Earth’s surface into the subsurface is a critical task in gravity exploration, as it helps to identify the sources of gravity anomalies. This problem is often addressed by solving a first-kind integral equation using regularization techniques to stabilize an inherently unstable process. A similar approach is used in our work, where the continued field is represented as the potential of a simple layer or its vertical derivative. The constancy of the density sign of this equivalent simple layer preserves the sign of anomalies, provided that the layer’s surface encloses all anomalous sources. This constraint is a key feature of our algorithm for the downward continuation of potential fields. To enforce, for instance, non-negativity in the simple layer density, we employ the NNLS (Non-Negative Least Squares) method. The efficiency of the proposed method is demonstrated on model examples. MSC Classification]35R30, 65R20, 86A22, 65F10, 45Q05 keywords potential field, gravity, method of integral equations, simple layer potential, method of least squares pacs [ 1 Introduction An essential class of geophysical problems involves gravimetry, which examines heterogeneities in the Earth’s crust and their influence on the gravitational field Gupta 2011 ; Mudrecova and Veselov 1990 . The primary measured quantity is gravitational acceleration, and the gravitational anomaly is defined as deviations from the normal gravity field after applying necessary corrections. Analysis of gravity survey data provides insights into mass distribution in the upper crust Parker 1994 ; Blokh 2009 . The direct gravimetric problem is associated with the calculation of the gravitational field generated by a mass with known physical and geometric parameters. In inverse gravimetric problems these parameters are deduced from observed field data Crossley et al. 2013 , typically collected the Earth’s surface, e.g. on earth’s surface, airborne surveys and satellite missions. Solutions of direct and inverse problems are used in data processing, qualitative interpretation of the observed data. Among these tasks, transformation, upward and downward continuation of the gravitational field, and characterization of the shape and density of anomalies are highlighted separately. When interpreting observed gravity fields, solving the forward problem makes it possible to predict and compare gravity anomalies for geological structures. Numerical methods for solving direct and inverse problems of geophysics specifically gravimetry and magnitometry are well-studied Vogel et al. 1992 ; Zhdanov 2002 ; Freeden and Nashed 2018 ; Yagola et al. 2014 ; Eppelbaum 2019 . The gravitational potential is defined as the Newtonian (volume) potential. The measured gravitational field is associated with the derivatives of the first- and higher-order potentials Blakely 1996 . The observations themselves are carried out on the Earth’s surface, taking into account the elevation Dolgal 2022 and the Earth’s sphericity for large-scale surveys. There are two main continuation tasks with respect to the observation surface and source locations: upward continuation and downward continuation Pilkington and Boulanger 2017 . The gravitational potential and its derivatives outside the anomalies satisfy the Laplace equation. Thus, we have problems of continuing solutions of elliptic equations beyond the boundary of the computational domain, which reduces to Cauchy problems for elliptic equations. Field downward continuation (towards sources) allows for clearer localization of the anomalies Pašteka et al. 2012 . The Cauchy problem for elliptic equations belongs to the class of ill-posed problems Lavrentiev and Saveliev 2006 , where small perturbations of the input data (measurements at the boundary) amplify solution errors. Approximate solutions employ regularization techniques Tikhonov and Arsenin 1977 ; Kabanikhin 2011 . For conditionally well-posed evolutionary problems, various approaches of approximate solution have been developed. For instance, we highlight Samarskii and Vabishchevich 2007 the Tikhonov regularization method, methods of equation perturbation (the quasi-inversion method), and separately note the algorithms with perturbation of initial conditions (non-local boundary value problems). In modern computing practice, iterative methods are most widely used Alifanov et al. 1995 ; Hansen 1998 ; Vasin and Ageev 2013 ; Bakushinsky et al. 2011 , where the number of iterations acts as a regularization parameter. For geopotential field continuation from one layer (irregular surface or level plane) to another, potentials are often approximated by equivalent sources, while the original generating sources are unknown. For brevity, we mention works Pilkington and Boulanger 2017 ; Dolgal et al. 2022 and references therein. For these problems, field-based and source-based methods can be distinguished. Field-based approaches are primarily based on simple and/or double-layer potentials derived from Green’s identities. For approximate solutions of Cauchy problems for elliptic equations in the case of unbounded domains, integral equations are more effective with well-developed numerical methods Hackbusch 1995 ; Kythe and Puri 2011 . Such an approach was applied to potential downward continuation problems by Tikhonov et al. 1968 . In the approximate solution of ill-posed problems, a fundamental increase in accuracy is achieved Tikhonov et al. 1995 by taking into account additional information about the solution. In a series of works Strakhov and Stepanova 2002 ; Stepanova et al. 2019 ; Stepanova et al. 2020 , the method of S-approximations and its modifications for separating geopotential fields by the sum of potentials of simple and double layers was proposed. When continuing potential fields from one plane to another, the problem can be more effectively solved by Fourier transforms in the spectral domain Pašteka et al. 2012 . Wavelet transformation techniques are used in multilevel source approximation in Li and Oldenburg 2010 ; Balk et al. 2016 ; in Balk et al. 2016 , this approach with Haar basis functions is compared with the quadtree algorithm. Wavelet analysis using the Legendre basis is applied in Temirbekov et al. 2022 for the solution of the first-kind integral equation. Source-based methods Dampney 1969 operate with elementary homogeneous objects such as points, prisms, etc., which generate parameterized potentials. Among recent works, in Leonov et al. 2024 , the authors pointed out the instability of source approximation, which can be effectively solved by a new approach based on improving the condition number of a matrix with the help of the minimal pseudoinverse matrix method. To continue the field into the lower half-space, the problem becomes unstable, requiring stable techniques based on iterative processes or regularization Zeng et al. 2013 ; Li et al. 2023 . In Vabishchevich 2024 , a two-dimensional problem of potential field continuation towards anomaly sources was considered using the integral equation method. The two-dimensional problem corresponds to masses extended along the horizontal direction, and gravity is calculated using the logarithmic potential. In this work, we extend the approach to the case of three-dimensional problems. The gravitational field is represented as a simple-layer potential or its derivative. It is shown that the equivalent layer density is positive (negative) for positive (negative) density anomalies if the surface encloses all anomalies. The continuation process is solved using the well-known method of non-negative least squares (NNLS) Lawson and Hanson 1995 , enforcing the required sign of the simple-layer densities. The article is structured as follows. The article begins by introducing the relevance of gravimetry and the challenges associated with downward continuation of gravity anomalies (Section 1). Section 2 formulates the mathematical framework, presenting integral equations for potential