Reference Map Research Articles

Avalanche size estimation and avalanche outline determination by experts: reliability and implications for practice

Abstract. Consistent estimates of avalanche size are crucial for communicating not only among avalanche practitioners but also between avalanche forecasters and the public, for instance in public avalanche forecasts. Moreover, applications such as risk management and numerical avalanche simulations rely on accurately mapped outlines of past avalanche events. Since there is not a widely applicable and objective way to measure avalanche size or to determine the outlines of an avalanche, we need to rely on human estimations. Therefore, knowing about the reliability of avalanche size estimates and avalanche outlines is essential as errors will impact applications relying on this kind of data. In the first of three user studies, we investigate the reliability in avalanche size estimates by comparing estimates for 10 avalanches made by 170 avalanche professionals working in Europe or North America. In the other two studies, both completed as pilot studies, we explore reliability in the mappings of six avalanches from oblique photographs from 10 participants and the mappings of avalanches visible on 2.9 km2 of remotely sensed imagery in four different spatial resolutions from 5 participants. We observed an average agreement of 66 % in the most frequently given avalanche size, while agreement with the avalanche size considered “correct” was 74 %. Moreover, European avalanche practitioners rated avalanches significantly larger for 8 out of 10 avalanches, compared to North Americans. Assuming that participants are equally competent in the estimation of avalanche size, we calculated a score describing the factor required to obtain the observed agreement rate between any two size estimates. This factor was 0.72 in our dataset. It can be regarded as the certainty related to a size estimate by an individual and thus provides an indication of the reliability of a label. For the outlines mapped from oblique photographs, we noted a mean overlapping proportion of 52 % for any two avalanche mappings and 60 % compared to a reference mapping. The outlines mapped from remotely sensed imagery had a mean overlapping proportion of 46 % (image resolution of 2 m) to 68 % (25 cm) between any two mappings and 64 % (2 m) to 80 % (25 cm) when compared to the reference. The presented findings demonstrate that the reliability of size estimates and of mapped avalanche outlines is limited. As these data are often used as reference data or even ground truth to validate further applications, the identified limitations and uncertainties may influence results and should be considered.

The phenotype-genotype reference map: Improving biobank data science through replication

Population-scale biobanks linked to electronic health record data provide vast opportunities to extend our knowledge of human genetics and discover new phenotype-genotype associations. Given their dense phenotype data, biobanks can also facilitate replication studies on a phenome-wide scale. Here, we introduce the phenotype-genotype reference map (PGRM), a set of 5,879 genetic associations from 523 GWAS publications that can be used for high-throughput replication experiments. PGRM phenotypes are standardized as phecodes, ensuring interoperability between biobanks. We applied the PGRM to five ancestry-specific cohorts from four independent biobanks and found evidence of robust replications across a wide array of phenotypes. We show how the PGRM can be used to detect data corruption and to empirically assess parameters for phenome-wide studies. Finally, we use the PGRM to explore factors associated with replicability of GWAS results.

An Integrated Map of Cell Type-Specific Gene Expression in Pancreatic Islets.

Pancreatic islets are comprised of multiple cell types that produce hormones required for glucose homeostasis, and islet dysfunction is a major factor in type 1 and type 2 diabetes. Numerous studies have assessed transcription across individual cell types using single cell assays; however, there is no canonical reference of gene expression in islet cell types that is also easily accessible for researchers to query and use in bioinformatics pipelines. Here we present an integrated map of islet cell type-specific gene expression from 192,203 cells from single cell RNA-sequencing of 65 non-diabetic, type 1 diabetes autoantibody positive, type 1 diabetic, and type 2 diabetic donors from the Human Pancreas Analysis Program. We identified 10 distinct cell types, annotated sub-populations of several cell types, and defined cell typespecific marker genes. We tested differential expression within each cell type across disease states, and identified 1,701 genes with significant changes in expression, with most changes observed in beta cells from donors with type 1 diabetes. To facilitate user interaction, we provide several single cell visualization and reference mapping tools, as well as open-access analytical pipelines used to create this reference. The results will serve as a valuable resource to investigators studying islet biology.

Statistical Analysis to Quantify the Impact of Map Type on Estimating Peak Discharge in Non-Instrumented Basins

The calculation of peak discharge in non-instrumented basins requires including morphometric parameters, which in turn depend on the map type used. This study analyses the impact of and variation in peak discharges of the Caño Ricaurte basin, Colombia, based on three types of maps at different resolution scales. The reference map used was the map made for the detailed designs of the channel analysed, which was extracted from the Master Plan of the City. Additionally, maps from a 90 × 90 m digital elevation model and contour lines extracted from Google Earth were used. The time of concentration was determined by different equations (Kirpich, Témez, Bureau, and TR-55) using the mapping methods described above, and the peak discharge was determined using rainfall-runoff models.

A Method for Intelligent Road Network Selection Based on Graph Neural Network

As an essential role in cartographic generalization, road network selection produces basic geographic information across map scales. However, the previous selection methods could not simultaneously consider both attribute characteristics and spatial structure. In light of this, an intelligent road network selection method based on a graph neural network (GNN) is proposed in this paper. Firstly, the selection case is designed to construct a sample library. Secondly, some neighbor sampling and aggregation rules are developed to update road features. Then, a GNN-based selection model is designed to calculate classification labels, thus completing road network selection. Finally, a few comparative analyses with different selection methods are conducted, verifying that most of the accuracy values of the GNN model are stable over 90%. The experiments indicate that this method could aggregate stroke nodes and their neighbors together to synchronously preserve semantic, geometric, and topological features of road strokes, and the selection result is closer to the reference map. Therefore, this paper could bridge the distance between deep learning and cartographic generalization, thus facilitating a more intelligent road network selection method.

Metagenomics and proteomics profiling of extracellular polymeric substances from municipal waste sludge and their application for soil and water bioremediation

This study assessed the components of anaerobically digested sludge, activated sludge, and microbial and extracellular polymeric substance (EPS) enzymes to identify the mechanisms underlying nitrogen removal and soil regeneration. 16S rRNA gene amplicon-based sequencing was used to determine the microbial community composition and the related National Center for Biotechnology Information (NCBI) protein database was used to construct a conventional library from the observed community. EPS components were identified using gel-free proteomic (Liquid Chromatography with tandem mass spectrometry-LC/MS/MS) methods. Alginate-like EPS from aerobically activated sludge have strong potential for soil aggregation and water-holding capacity, whereas total EPS from anaerobic sludge have significant potential for ammonia removal under salt stress. Fourier transform infrared spectroscopy (FTIR) revealed that both EPS may contain proteins, carbohydrates, humic compounds, uronic acid, and DNA and determined the presence of O–H, N–H, C–N, CO, and C–H functional groups. These results demonstrate that the overall enzyme activity may be inactivated at 30 g L-1 of salinity. An annotation found in Kyoto Encyclopedia of Genes and Genomes (KEGG)- KEGG Automatic Annotation Server (KAAS) revealed that the top two metabolic activities in the EPS generated from the anaerobic sludge were methane and nitrogen metabolism. Therefore, we focused on the nitrogen metabolism reference map 00910. EPS from the anaerobically digested sludge exhibited nitrate reductase, nitrite reductase, and dehydrogenase activities. Assimilatory nitrate reduction, denitrification, nitrification, and anammox removed ammonia biochemically. The influence of microbial extracellular metabolites on water-holding capacity and soil aggregation was also investigated. The KAAS-KEGG annotation server was used to identify the main enzymes in the activated sludge-derived alginate-like extracellular EPS (ALE-EPS) samples. These include hydrolases, oxidoreductases, lyases, ligases, and transporters, which contribute to soil fertility and stability. This study improves our understanding of the overall microbial community structure and the associated biochemical processes, which are related to distinct functional genes or enzymes involved in nitrogen removal and soil aggregation. In contrast to conventional methods, microbial association with proteomics can be used to investigate ecological relationships, establishments, key player species, and microbial responses to environmental changes. Linking the metagenome to off-gel proteomics and bioinformatics solves the problem of analyzing metabolic pathways in complex environmental samples in a cost-effective manner.

CoPR: Toward Accurate Visual Localization With Continuous Place-Descriptor Regression

Visual Place Recognition (VPR) is an image-based localization method that estimates the camera location of a query image by retrieving the most similar reference image from a map of geo-tagged reference images. In this work, we look into two fundamental bottlenecks for its localization accuracy: reference map sparseness and viewpoint invariance. Firstly, the reference images for VPR are only available at sparse poses in a map, which enforces an upper bound on the maximum achievable localization accuracy through VPR. We therefore propose Continuous Place-descriptor Regression (CoPR) to densify the map and improve localization accuracy. We study various interpolation and extrapolation models to regress additional VPR feature descriptors from only the existing references. Secondly, we compare different feature encoders and show that CoPR presents value for all of them. We evaluate our models on three existing public datasets and report on average around 30% improvement in VPR-based localization accuracy using CoPR, on top of the 15% increase by using a viewpoint-variant loss for the feature encoder. The complementary relation between CoPR and Relative Pose Estimation is also discussed.

Organization of the human intestine at single-cell resolution

The intestine is a complex organ that promotes digestion, extracts nutrients, participates in immune surveillance, maintains critical symbiotic relationships with microbiota and affects overall health1. The intesting has a length of over nine metres, along which there are differences in structure and function2. The localization of individual cell types, cell type development trajectories and detailed cell transcriptional programs probably drive these differences in function. Here, to better understand these differences, we evaluated the organization of single cells using multiplexed imaging and single-nucleus RNA and open chromatin assays across eight different intestinal sites from nine donors. Through systematic analyses, we find cell compositions that differ substantially across regions of the intestine and demonstrate the complexity of epithelial subtypes, and find that the same cell types are organized into distinct neighbourhoods and communities, highlighting distinct immunological niches that are present in the intestine. We also map gene regulatory differences in these cells that are suggestive of a regulatory differentiation cascade, and associate intestinal disease heritability with specific cell types. These results describe the complexity of the cell composition, regulation and organization for this organ, and serve as an important reference map for understanding human biology and disease.

How Sustainable Are Tourist Destinations Worldwide? An Environmental, Economic, and Social Analysis

Sustainable tourism is the idea of managing the negative impacts or potential for serious harm to the economic, environmental, or social elements of a destination, and thus reaching the goals of sustainable development. This research benchmarks the sustainability of 111 tourist destinations throughout the world, using an advanced integration of a two-stage network DEA under the meta-frontier concept, directional distance function, and network-based ranking methods. The empirical outcomes of these analyses clarify that the sources of sustainable inefficiencies are the self-production process and the technology gap among the tourism regions, rather than the capacity of utilizing the input resources or maintaining the production outcomes. The network-based ranking emphasizes the sufficiency and deficiency of each place, and provides a map of reference among destinations. Based on these, the article reports some managerial suggestions, theoretical implications, and future research directions.

A semantic vector map-based approach for aircraft positioning in GNSS/GPS denied large-scale environment

Accurate positioning is one of the essential requirements for numerous applications of remote sensing data, especially in the event of a noisy or unreliable satellite signal. Toward this end, we present a novel framework for aircraft geo-localization in a large range that only requires a downward-facing monocular camera, an altimeter, a compass, and an open-source Vector Map (VMAP). The algorithm combines the matching and particle filter methods. Shape vector and correlation between two building contour vectors are defined, and a coarse-to-fine building vector matching (CFBVM) method is proposed in the matching stage, for which the original matching results are described by the Gaussian mixture model (GMM). Subsequently, an improved resampling strategy is designed to reduce computing expenses with a huge number of initial particles, and a credibility indicator is designed to avoid location mistakes in the particle filter stage. An experimental evaluation of the approach based on flight data is provided. On a flight at a height of 0.2 km over a flight distance of 2 km, the aircraft is geo-localized in a reference map of 11,025 km2 using 0.09 km2 aerial images without any prior information. The absolute localization error is less than 10 m.

Creating a digital historical map of the German-inhabited areas in southern Transylvania (central Romania) in 1848

ABSTRACT The paper describes the creation of a historical reference map of the mid-southern region of Transylvania (nowadays in central Romania) in 1848. This area was shaped by the presence of the Transylvanian Saxons, the ethnic German descendants of the settlers invited here by the kings of Hungary during the High Middle Ages. Transylvania became part of the Habsburg Empire at the end of the seventeenth century and was elevated to the status of a Grand Principality in 1765. The map is primarily based on historical cartographic documents and written publications, but also integrates digital geo-spatial datasets. It was produced at a scale of 1:400 000 and includes the hydrography, the administrative boundaries, the main roads, the postal stations and all the settlements attested in the Transylvanian Saxon administrative subdivisions before the mid-nineteenth century.

No-reference shadow detection quality assessment via reference learning and multi-mode exploring

Shadow detection is a common computer vision task that identifies the location of shadows. It has been used as a key primitive for many high-level tasks such as object segmentation and video tracking. Due to the complexity and variety of shadows, it is challenging to design a one-size-fits-all shadow detection algorithm in real-world scenarios. Instead, it is often necessary in practice to pick the best shadow detection algorithm for a particular vision application. While shadow detection algorithms can be assessed by comparing the detected shadow with the ground truth, such ground truth is often unavailable in real-world vision systems consistently generating new data. In this paper, we propose a no-reference shadow detection quality assessment network (NSDQA-Net) to rank shadow detection algorithms without needing the ground truth. For a given image, the proposed assessment network first learns a reference map that compensates the false positives and false negatives of multiple input shadow detection algorithms. The reference map, as well as the original image and the input shadow detection results, are then used to regress the quality scores of the shadow detection methods. Our framework supports the assessment of one, two, or multiple algorithms in a single run, making it easier for deployment in practical scenarios. Experiments results show that the proposed network achieves optimal accuracy in no-reference shadow detection assessment. Relying on it for adaptive algorithm selection also improves the performance of shadow detection using a fixed algorithm on benchmark datasets.

Land degradability mapping using remote sensing data and soil chemical properties

Selection of the most influential soil properties to land degradability (LDy) mapping and their association is unclear. Accordingly, this study developed a geo-environmental modelling approach using multi-source datasets including remote sensing imagery, meteorological data, and soil properties (electrical conductivity (EC), soil organic matter (SOM), and calcium carbonate (CaCO3)) to LDy mapping in Jajrud and Karaj basin of Iran. LDy reference map obtained from soil properties data and Hedged Product overlay method was used to train and validate the random forest model. A sensitivity analysis was performed, which showed that normalized difference vegetation index (NDVI) is the most effective variable in LDy mapping. Under data constraints, the combination of biological soil crust index, leaf area index, and maximum wind speed is the best dataset for LDy mapping. Results revealed that the study area's most degraded lands were primarily barelands, followed by farmlands, grasslands, and fallow lands, respectively. Besides, considering their lower EC, the highlands and fallowlands did not witness land degradation.

Single-cell reference mapping to construct and extend cell-type hierarchies.

Single-cell genomics is now producing an ever-increasing amount of datasets that, when integrated, could provide large-scale reference atlases of tissue in health and disease. Such large-scale atlases increase the scale and generalizability of analyses and enable combining knowledge generated by individual studies. Specifically, individual studies often differ regarding cell annotation terminology and depth, with different groups specializing in different cell type compartments, often using distinct terminology. Understanding how these distinct sets of annotations are related and complement each other would mark a major step towards a consensus-based cell-type annotation reflecting the latest knowledge in the field. Whereas recent computational techniques, referred to as 'reference mapping' methods, facilitate the usage and expansion of existing reference atlases by mapping new datasets (i.e. queries) onto an atlas; a systematic approach towards harmonizing dataset-specific cell-type terminology and annotation depth is still lacking. Here, we present 'treeArches', a framework to automatically build and extend reference atlases while enriching them with an updatable hierarchy of cell-type annotations across different datasets. We demonstrate various use cases for treeArches, from automatically resolving relations between reference and query cell types to identifying unseen cell types absent in the reference, such as disease-associated cell states. We envision treeArches enabling data-driven construction of consensus atlas-level cell-type hierarchies and facilitating efficient usage of reference atlases.

Mapping the Posterior Ledge and Optic Foramen in Orbital Floor Blowout Fractures.

Background The posterior ledge (PL) is a vital structure that supports the implant posteriorly during orbital floor reconstruction. This study describes a technique for mapping the PL in relation to the infraorbital margin (IM) in patients with orbital floor blowout fractures. This study establishes the location of the optic foramen in relation to the PL. Methods Facial computed tomography (FCT) scans of 67 consecutive patients with isolated orbital floor blowout fractures were analyzed using Osirix. Planes of reference for orbital fractures, a standardized technique for performing measurements on FCT, was used. Viewed coronally, the orbit was divided into seven equal sagittal slices (L1 laterally to L7 medially) with reference to the midorbital plane. The distances of PL from IM and location of optic foramen were determined. Results The greatest distance to PL is found at L5 (median: 30.1 mm, range: 13.5-37.1 mm). The median and ranges for each slice are as follows: L1 (median: 0.0 mm, range: 0.0-19.9 mm), L2 (median: 0.0 mm, range: 0.0-21.5 mm), L3 (median: 15.8 mm, range: 0.0-31.7 mm), L4 (median: 26.1 mm, range: 0.0-34.0 mm), L5 (median: 30.1 mm, range: 13.5-37.1 mm), L6 (median: 29.0 mm, range: 0.0-36.3 mm), L7 (median: 20.8 mm, range: 0.0-39.2 mm). The median distance of the optic foramen from IM is 43.7 mm (range: 37.0- 49.1) at L7. Conclusion Distance to PL from IM increases medially until the L5 before decreasing. A reference map of the PL in relation to the IM and optic foramen is generated. The optic foramen is located in close proximity to the PL at the medial orbital floor. This aids in preoperative planning and intraoperative dissection.

Optimization of orthogonal separations for the analysis of oligonucleotides using 2D-LC

Oligonucleotides are commonly analysed using one dimensional chromatography (1D-LC) to resolve and characterise manufacturing impurities, structural isomers and (in respect to emerging oligonucleotide therapeutics) drug substance and drug product. Due to low selectivity and co-elution of closely related oligonucleotides using 1D-LC, analyte resolution is challenged. This leads to the requirement for improved analytical methods. Multidimensional chromatography has demonstrated utility in a range of applications as it increases peak capacity using orthogonal separations, however there are limited studies demonstrating the 2D-LC analysis of closely related oligonucleotides. In this study we optimised OGN size and sequence based separations using a variety of 1D-LC methods and coupled these orthogonal modes of chromatography within a 2D-LC workflow. Theoretical 2D-LC workflows were evaluated for optimal orthogonality using the minimum convex hull metric. The most orthogonal workflow identified in this study was ion-pair reversed phase using tributylammonium acetate (IP-RP-TBuAA) coupled with strong anion exchange in conjunction with sodium perchlorate (SAX-NaClO4) at high mobile phase pH. We developed a heart-cut (IP-RP-TBuAA)-(SAX-NaClO4) 2D-LC method for analysis of closely related size and sequence variant OGNs and OGN manufacturing impurities. The 2D-LC method resulted in an increased orthogonality and a reduction in co-elution (or close elution). Application of a UV based reference mapping strategy in conjunction with the 2D-LC method demonstrated a reduction in analytical complexity by reducing the reliance on mass based detection methods.

Exploring the use of multi-source high-resolution satellite data for snow water equivalent reconstruction over mountainous catchments

Abstract. The hydrological cycle is strongly influenced by the accumulation and melting of seasonal snow. For this reason, mountains are often claimed to be the “water towers” of the world. In this context, a key variable is the snow water equivalent (SWE). However, the complex processes of snow accumulation, redistribution, and ablation make its quantification and prediction very challenging. In this work, we explore the use of multi-source data to reconstruct SWE at a high spatial resolution (HR) of 25 m. To this purpose, we propose a novel approach based on (i) in situ snow depth or SWE observations, temperature data and synthetic aperture radar (SAR) images to determine the pixel state, i.e., whether it is undergoing an SWE increase (accumulation) or decrease (ablation), (ii) a daily HR time series of snow cover area (SCA) maps derived by high- and low-resolution multispectral optical satellite images to define the days of snow presence, and (iii) a degree-day model driven by in situ temperature to determine the potential melting. Given the typical high spatial heterogeneity of snow in mountainous areas, the use of HR images represents an important novelty that allows us to sample its distribution more adequately, thus resulting in highly detailed spatialized information. The proposed SWE reconstruction approach also foresees a novel SCA time series regularization technique that models impossible transitions based on the pixel state, i.e., the erroneous change in the pixel class from snow to snow-free when it is expected to be in accumulation or equilibrium and, vice versa, from snow-free to snow when it is expected to be in ablation or equilibrium. Furthermore, it reconstructs the SWE for the entire hydrological season, including late snowfall. The approach does not require spatialized precipitation information as input, which is usually affected by uncertainty. The method provided good results in two different test catchments: the South Fork of the San Joaquin River, California, and the Schnals catchment, Italy. It obtained good agreement when evaluated against HR spatialized reference maps (showing an average bias of −22 mm, a root mean square error – RMSE – of 212 mm, and a correlation of 0.74), against a daily dataset at coarser resolution (showing an average bias of −44 mm, an RMSE of 127 mm, and a correlation of 0.66), and against manual measurements (showing an average bias of −5 mm, an RMSE of 191 mm, and a correlation of 0.35). The main sources of error are discussed to provide insights into the main advantages and disadvantages of the method that may be of interest for several hydrological and ecological applications.

OPTIMUM WATER ALLOCATION FOR ABO-ZIRIQ MARSH ECOLOGICAL RESTORATION

Optimum allocation of water for restoration of Iraqi marshes is essential for different related authorities. Abo-Ziriq marsh area about 120 km2 is situated 40 km east of Al-Nassryia city. After comparing the measured annual water qualities with the Iraqi standards for surface water quality evaluation, Abo-Ziriq marsh water quality was in acceptable limit. Hydro balance computation were done for each month by using interface among the HEC-RAS, HEC-GeoRAS and ArcView GIS software and built a number of eco-hydro relationships to simulate the marsh ecosystem by using HEC-EFM program to estimate water allocation adequate for ecosystem requirement and constructs a GIS hydraulic reference map to show inundation area, depth grid and velocity distribution, the optimal flow result consists of three different scenarios (24, 30.3 and 33.6 m3 /s) for marsh operation during the year. A computer program in MATLAB 7 was developed to simulate the optimization model to determine the optimum flow value entry to lower zone. The priority of each parameter is represented by a weight associated with each of them (penalty factor). The model was used for different scheme of penalty factor value and examines three cases of flow (wet, moderate and dry years). The results obtained from the program run show that the optimum flow values are not affected significantly with changing the scheme of the penalty factors.Hence, any set of solutions can be use for operation the control structure of two inlets in the lower zone that best fits the objective of the system and increase flow release from Abo Jiry inlet to minimum deviation in water quality during the most time of the year

MDB-05. HUMAN FETAL CEREBELLAR CELL ATLAS INFORMS MEDULLOBLASTOMA ORIGIN AND ONCOGENESIS

Abstract Medulloblastoma (MB) is the most common malignant childhood brain tumor, yet the origin of the most aggressive subgroup-3 form remains elusive, impeding development of effective targeted treatments. Previous analyses of mouse cerebella or human counterparts from frozen tissue nuclei5 have not fully defined the compositional heterogeneity of MBs. Here, we undertook an unprecedented single-cell profiling of freshly-isolated human fetal cerebella to establish a reference map delineating hierarchical cellular states in MBs. We identified a unique transitional cerebellar progenitor connecting neural stem cells to neuronal lineages in developing fetal cerebella. Intersectional analysis revealed that the transitional progenitors were enriched in aggressive MB subgroups, including group-3 and metastatic tumors. Single-cell multi-omics revealed underlying regulatory networks in the transitional progenitor populations, including transcriptional determinants HNRNPH1 and SOX11, which are correlated with clinical prognosis in group-3 MBs. Genomic and Hi-C profiling identified de novo long-range chromatin-loops juxtaposing HNRNPH1/SOX11-targeted super-enhancers to cis-regulatory elements of MYC, an oncogenic driver for group-3 MBs. Targeting the transitional progenitor regulators inhibited MYC expression and MYC-driven group-3 MB growth. Our integrated single-cell atlases of human fetal cerebella and MBs reveal potential cell populations predisposed to transformation and regulatory circuitries underlying tumor cell states and oncogenesis, highlighting hitherto unrecognized transitional progenitor intermediates predictive of disease prognosis and potential therapeutic vulnerabilities.

Inferring geological structural features from geophysical and geological mapping data using machine learning algorithms

AbstractWe present an automated approach for inferring surface geological structures from geophysical survey data. Our method employs machine learning, using mapped geological structures as labels and filtered geophysical surveys as reference maps. We compared the performance of the eight main machine learning algorithms and their 32 branches. Applied to the Geological Survey of Victoria's database for the Bendigo Zone, following an appropriate choice of geological features, the 3‐class classification model using subspace K‐nearest neighbour methods achieves a stable and validated 92% accuracy in around 1 min. The fault‐only classification model achieves a stable and validated 97% accuracy in around 6 min. This shows that geological structural features on the surface may be inferred from between one and three of the following geophysical data types: gravity, airborne total magnetic intensity and first vertical derivative of total magnetic intensity. It shows the prospect of machine learning in geological research and suggests that geophysical data combined with machine learning may be useful and efficient in determining the existence of geological structural features.