- Research Article
- 10.3997/1365-2397.fb2026012
Predictive Geoscience: Machine Learning Integrating Globe, Gravity, Magnetics, and Remote Sensing Data
- Feb 01, 2026
- First Break
- Samuel Cheyney + 3 more +3
Publications from 2021 to 2026
Showing 10 of 18 papers
Predictive Geoscience: Machine Learning Integrating Globe, Gravity, Magnetics, and Remote Sensing Data
Arctic geophysical and geological data overview
Since early 1990s, the Arctic Ocean and surrounding seas have been witnessing a dramatic increase in geophysical and geological data acquisition related to the following major activities: (i) academia- and state-funded research; (ii) United Nations Convention on the Law of the Sea (UNCLOS) research; and (iii) industry-related hydrocarbon exploration programmes. The major advances have been related to the use of ice – capable research vessels assisted by icebreakers as research platforms, allowing use of a broad range of data-gathering tools to acquire a wealth of atmospheric, ice, ocean and sub-bottom data over relatively short periods of time. The data quality has increased greatly due to advances in the offboard towing equipment, GPS positioning, satellite communication and computing. In this chapter we provide a brief overview of geophysical data acquisition and geological sampling campaigns in the Arctic since 1990. Enclosure F is a part of this chapter, portraying locations of multichannel seismic reflection, wide-angle refraction profiles, sonobuoys, ocean-bottom seismometers and sampling sites including drilled wells. The corresponding metadata have been compiled in two tables in this chapter.
Read moreExploration Workflow for Natural Hydrogen Based on Potential Field Data
Summary This study presents an innovative exploration workflow for natural hydrogen, focusing on the HyAfrica Project's work in the Tendrara Basin, Morocco. Natural hydrogen, a sustainable energy resource, differs from industrial hydrogen due to its geological generation processes. The study introduces a specialized workflow integrating multiple data sets to explore hydrogen in three critical areas: Surface, Shallow, and Deep layers. Analysis of the Surface layer uses the Natural Hydrogen Seeps Detection (NHSD) algorithm, using multispectral satellite imagery to identify semi-circular structures indicative of hydrogen presence. The characterization of the Shallow layer involves geological/geophysical modelling, integrating well and seismic data to analyze geological structures and reservoir seals. The Deep layer is explored for hydrogen sources using the Quantification of Natural Hydrogen Generation Algorithm (QNHG), by means of gravity and magnetic data inversion in order to model serpentinization processes and estimate hydrogen production. The findings in this study highlight the potential of natural hydrogen as a significant energy resource, with implications for sustainable energy development.
Read morePaleogeographic and tectonic models of the Western Mediterranean
Recent advances in plate modelling and deep-earth visualisation software present excellent opportunities to integrate spatial and temporal data from several earth science disciplines. These tools can enhance our understanding of tectonic and palaeogeographic evolution, with clear applications to resource exploration.We present a case-study of our geodynamic visualisation package, in which we integrate ‘traditional’ rigid tectonic elements alongside dynamically evolving plate boundaries, tectonic and thermal events and paleogeographic mapping. We focus on the Western Mediterranean, where our developed tectonic model is derived from potential field data, structural analysis, and crustal architecture interpretations. Our model forms the basis for interpretations of palaeogeography at key reconstruction ages throughout the Cenozoic evolution of the area.Our crustal architecture interpretations illustrate the nature of the crust and its deformation since Variscan. Key components of the model include: Variscan age fold and thrust belts in Iberia, a Cenozoic oceanic domain (Algerian-Ligurian-Provencal Sea) bounded by a magma-poor continent-ocean transition domain, a relatively undeformed continental block (Corsica-Sardinia) and a large expanse of attenuated crust associated with the rapid roll-back of the Calabrian Arc. Our model is constrained by both structural interpretation of gravity and magnetic data and their derivatives and 2D gravity and magnetic modelling of crustal profiles to match the observed signature. The 2D crustal models confirm crustal thicknesses and density to aid the interpretation of crustal type, particularly in the continent-ocean transition where the geophysical signature can help to understand the role of magmatism, hyperextension, or mantle exhumation.Our tectonic model comprises a rigid, kinematic terrane reconstruction showing the evolution of geologically unique terranes through time, and a dynamic plate reconstruction constrained by geodynamically modelled plate boundaries. Rigid terranes in the model show the detailed tectonic evolution of the Western Mediterranean by reconstructing fault-bounded blocks using Euler rotations derived from geological and geophysical observations. Dynamic elements of the model show a bigger picture tectonic evolution including long and short-lived plates bounded by active tectonic boundaries. Together these elements illustrate the full tectonic evolution including rigid blocks, deforming margins and palaeo-oceanic domains.We also present an attributed catalogue of tectono-thermal events, which describe the location, age and duration of key events that have occurred through time; visualisation of these events aims to assist with understanding the thermo-tectonic evolution of potential resource exploration targets and geothermal prospectivity.   Paleogeographic and landscape evolution models are built alongside the tectonic model allowing iterative refinement of the models. Palaeogeographic interpretations include gross depositional environments, the cause and timing of uplifted and eroding areas, and likely lithologies in depositional areas. We build digital elevation models for each stage which are constrained by palaeogeographies and drainage networks. This palaeogeographic module therefore illustrates source-to-sink relationships, which is key for exploration across a range of energy sectors.
Read moreSlope-Weighted Eccentricity: Automatic Terrain Classification of Atlantic Ocean Crust
The shapes and directionality of bathymetry at slow-spreading ridges are key to understanding the magmatic or tectonic emplacement of the crust. Magmatic terrain is marked by linearly fault-bounded abyssal hills, while tectonic terrain is marked by long-lived detachment faults, forming Oceanic Core Complexes (OCCs). However, the quantitative description of these crustal regimes is still limited. We develop a novel automated terrain classification technique and test it at the 13-15° N section of the Mid-Atlantic Ridge. The algorithm uses the Slope-Weighted Eccentricity (SWE) of the horizontal eigenvalues to represent surface directionality and reveal crustal tectonic fabric. The application of this new technique yields results consistent with qualitative interpretation. Thus, it provides both new insights into the mid-oceanic ridge spreading and the potential to automate such mapping with different sets of grids, such as gravity and magnetic data in regions further away from the ridge where sediments mask sea-bed features.
Read moreRe-Formative Polymer Composites from Plastic Waste: Novel Infrastructural Product Application
A novel re-formative polymer composite manufacturing route has been developed by UK and Qatar-based Universities. This novel process recycles domestic-waste thermoplastic material, without the requirement for intensive filtering or washing operations. The produced polymer can be reinforced with recycled glass fibres, forming a structurally load-bearing composite, which may potentially be suitable for use in applications, including utility poles, railway sleepers, and fencing. Thus, infra-red (IR) analysis showed the presence of polypropylene (PP), polyethylene (PE), and polyethylene terephthalate (PET) in the commingled material. Differential scanning calorimetry (DSC) was used to determine glass transition temperatures and melting temperatures of each of the associated polymer types. Dynamic mechanical thermal analysis (DMTA) was used to determine the storage and loss modulus of the bulk commingled component. Lastly, flexural and tensile strengths of the re-formative polymer with differing proportions of glass fibre were assessed, giving a range of strengths at each glass fibre proportion for possible compositional variation in the polymer type. The recycled polymer is considered a viable structural material for replacing both wooden and concrete components, generating a polymer recycling route with concomitant environmental benefits. This plastic recycling route therefore offers a solution towards achieving climate change targets with a purposeful end-product component.
Read moreRegional Gravity Field Model of Egypt Based on Satellite and Terrestrial Data
This study presents a recent combined regional gravity field model over Egypt, developed by integrating satellite and terrestrial data via applying the remove-compute-restore (RCR) principle and the least-squares collocation (LSC) procedure. A high-resolution digital terrain model was exploited for the computation of the terrain and residual terrain corrections. Hereby, all the signals that can be modelled or deterministically computed are considered known and then removed in order to reduce the order of magnitude of the input gravity data prior to applying the LSC. Several GOCE-only and combined global geopotential models (GGMs) have been thoroughly investigated with respect to the EGM2008, in which the space-wise (SPW) solution, namely the SPW-R5 model, demonstrated the best performance. For the development of the combined model, the SPW-R5 GGM has been integrated with both the EGM2008 GGM and the terrestrial data retrieved from 56,250 gravity stations of the Getech data, acquired in the framework of the African Gravity Project. The combined regional gravity model was compared to the state-of-art XGM2016 global gravity model. The standard deviation of the differences is 18.0 mGal in terms of Bouguer anomalies. The combined regional model fits well with the terrestrial gravity data along the chosen North–South oriented profile through the Nile Delta region. The improvements of the developed combined regional model over the XGM2016 are due to the use of a more extensive terrestrial dataset. In conclusion, our model is more suitable than solely using the ground data or GGMs for regional density modelling over Egypt. As an example, the comparison of using a global or regionally defined gravity model with the forward gravity modelling based on Saleh (Acta Geodaetica et Geophysica Hungarica 47(4):402–429, 2012) density model is performed.
Read moreCurie Point Depth of the Iberian Peninsula and Surrounding Margins. A Thermal and Tectonic Perspective of its Evolution
Abstract In this work the thermal structure of the Iberian Peninsula is derived from magnetic data by calculating the bottom of the magnetization, assumed to be the Curie‐point depth (CPD) isotherm, which accounts for the depth at which magnetite becomes paramagnetic (580°C). Comparison of the CPD with crustal thickness maps along with a heat flow map derived from the CPD provides new insights on the lithospheric thermal regime. Within Iberia, the CPD isotherm has thickness in the range of 17 to 29 km. This isotherm is shallow (<18 km) offshore, where the lithosphere is thinner. In continental Iberia, the NW Variscan domain presents a magnetic response that is most probably linked to thickening and later extension processes during the late Variscan Orogeny, which resulted in widespread crustal melting and emplacement of granites (in the Central Iberian Arc). The signature of the CPD at the Gibraltar Arc reveals a geometry consistent with the slab roll‐back geodynamic model that shaped the western Mediterranean. In offshore areas, a broad extension of magnetized upper mantle is found. Serpentinization of the upper mantle, probably triggered in an extensional context, is proposed to account for the magnetic signal. The Atlantic margin presents up to 8 km of serpentinites, which, according to the identification of exhumed mantle, correlates with a hyperextended margin. The Mediterranean also presents generalized serpentinization up to 6 km in the Algerian Basin. Furthermore, a heat flow map and a Moho temperature map derived from the CPD are presented.
Read moreSouthern Natal Valley Basin: Integrating Gravity with Seismic for Basin Location and Prediction of Petroleum Systems
Summary One of the key attributes of gravity and magnetic data is to extend observations and interpretations from seismic profiles over a larger area. This data combination was used effectively in Getech’s study of the Southern Natal Valley Basin along the east coast of South Africa. Regional seismic data tied to on-shelf wells were integrated with Getech’s gravity and magnetic data and the resulting data set was used to unravel structures, identify sub-basins and develop a new plate model for the area. The data set provided us with the framework to predict the petroleum systems and prospectivity of the sub-basins of the Southern Natal Valley Basin.
Read moreApplying Magnetic Magnitude Transforms to Aid Structural Mapping in Areas Where the RTP Calculation Is Unreliable
A common use for magnetic data is to locate the edges of subsurface bodies. Numerous transformations have been developed in order to obtain more easily interpretable maps, however many initially require the data to be reduced to the pole (RTP) if they are to be used for edge detection. The RTP operation is unstable at low magnetic latitudes, which affects significant areas of equatorial regions currently of interest to the oil and gas industry. In addition, any remanent magnetisation will often lead to anomalous results. We demonstrate the advantages and disadvantages of using magnitude transformations, focusing on the scenarios where they are likely to be the most useful: at low latitudes and in the presence of remanent magnetisation. Based on synthetic modelling, magnitude transforms equivalent to 1st or 2nd order derivatives of the magnetic field appear most suitable, however when applied to real data the effects of enhancing the noise level limits the usefulness. The Total Amplitude, which is of the same order as the observed TMI data, has higher centricity than the TMI allowing lineament interpretation, without biasing the result to shorter-wavelength features and accentuating the noise level in the data.
Read more