Landform Recognition Research Articles

An Overview of Geomorphological Mapping: A Case Study of Rimbi Chhu River Basin, Sikkim, India

Geomorphology is the primary science which demonstrates the basic understanding and mapping of terrain features. A geomorphological map consists of scientific data and dynamic source of information on characteristics of landforms, their origin and evolution, which can be recorded in a map form. Basin hydrology also gets impacted by the terrain features of a basin as overland flow, stream flow and through flow is largely determined by the underneath surface. The configuration of the basin topography reflects the nature of the interaction between process and form. Profound understanding and in depth analysis of geomorphological features of a river basin can provide beneficial information for predicting, preventing and mitigating natural hazards; managing natural resources sustainably for future generation. Implementation of remote sensing and GIS techniques have fostered incredible improvements in landform recognition and accelerated the growth of geomorphological research. The river Rimbi Chhu has been selected as the study area in order to determine terrain configuration in relation with geomorphological features of the basin area. Visual representation of terrain (such as aspect, slope, channel networks, different river profiles, shaded relief and flow direction) and geomorphological (saddle, ridge line, convex-concave slope, rounded hills, fault for instance) derivatives is therefore illustrated, as well as a description of the challenges and crisis that can occur in this context.

Using music and music‐based analysis to model and to classify terrain in geomorphology

AbstractMusic has long served as a bridge between science and nature, allowing for the artistic expression of natural and human‐made phenomena. While music has been used in geomorphology as an engaging teaching strategy, its application in specific scientific inquiries within geomorphology remains relatively unexplored. Drawing on the morphological similarities between music and terrain relief, this work introduced a novel music‐based method for modelling and expressing terrain relief based on the drainage basin profile (DBP). It converts terrain relief into pitches and time values according to mapping rules and then describes terrain relief in an audible form. Based on 5 sample areas and 360 drainage basins on the Loess Plateau, we developed the application of the proposed method on four core geomorphic tasks, including landform interpretation, analysis, recognition and classification. Experimental results show that (1) terrain music can interpret the terrain relief and landform evolution processes through its musical structure and rhythmic variations; (2) music derivatives are related to different terrain features, such as terrain relief, terrain variation intensity, landform evolution degree and terrain complexity, and have a well‐functioning relationship with a series of conventional terrain derivatives; (3) leveraging machine learning techniques, the terrain music method is effective for landform recognition, achieving an overall accuracy of 88.85% and a mean accuracy of 88.85%; and (4) via a case study in Northern Shaanxi, music modelling successfully divided it into 12 distinct landform regions and 8 landform types. Different landform regions exhibit clear regional boundaries and gradual transition zones, while specific landform regions share prominent terrain, spatial clustering and landform processes. Our delineation provides reasonable and effective landform differentiation but captures additional bed‐rock mountain features compared with a traditional method. This study highlights that music is not only an artistic expression but also a valuable research paradigm for a wide range of geomorphic tasks, offering fresh perspectives and enhancing our understanding of loess landforms. The results show a promising effort to integrate music theory into practical geomorphic tasks, demonstrating the potential of using music as a medium for conveying and analysing spatial information in geomorphology.

Directed Positive Negative Terrain Structure Graph Attention Network for Genetic Landform Recognition

Directed Positive Negative Terrain Structure Graph Attention Network for Genetic Landform Recognition

Surface-subsurface connectivity in the morphological evolution of sandstone-capped tabular hills – How much analogy to karst?

This paper presents geomorphic evidence for the significant role of surface-subsurface connectivity in the geomorphological evolution of tabular hills (mesas) capped by moderately thick sandstone beds. The study area is located in Central Europe, in the Sudetes mountain range (Poland, Czechia), in an area built of a clastic succession of Late Cretaceous age. Research involved analysis of high-resolution LiDAR DTM and its various derivatives, providing insights into characteristics of drainage routes and patterns of ruiniform relief, as well as fieldwork. The latter included landform recognition and mapping, surveying representative cleft and cave systems. Specific surface landforms indicative of drainage diversion underground and subterranean pathways of sand removal include boulder-filled troughs and hollows, open clefts and plazas, pseudo-sinkholes and tilted blocks. Further evidence is provided by abundant, joint-aligned roofed slots and boulder caves, whose origin is attributed to selective sandstone weathering, sand evacuation by groundwater, and gradual subsidence. Sandy cones and sheets below exits of slots and fractures are the complementary sedimentary evidence. The percentage of caprock area drained via subterranean outflow varies from 16 % to 89 %. Reasons for the high efficacy of subterranean water and sediment transport include both structural factors (regular jointing pattern, high porosity) and, very likely, pre-conditioning role of dissolution of cement in the quartz arenites (arenization), which facilitates detachment of individual grains even by low flows under low hydraulic gradient. Thus, the sandstone-capped mesas show numerous analogies to the karst proper, especially silicate karst. They are manifested in both similar landform inventories, main preparatory processes, and considerable surface-subsurface connectivity.

Mountain railway alignment optimization based on landform recognition and presetting of dominating structures

AbstractMountain railway alignment optimization has always been a challenge for designers and researchers in this field. It is extremely difficult for existing methods that optimize alignments before major structures to generate a better alignment than the best one provided by human designers when the terrain is drastically undulating between the start and endpoints. To fill this gap, a “structures before alignments” design process is proposed in this paper. Primarily, a landform recognition method is devised for recognizing dominating landforms. Then, a bi‐level alignment optimization model is proposed, with the upper level dedicated to characterizing dominating structures and the lower level focusing on optimizing the entire alignments. To solve this bi‐level model, a three‐stage optimization method is designed. At the first stage, a scanning process and screening operators are devised for generating all the possible locations of dominating structures. At the second stage, a hierarchical multi‐criteria decision‐making procedure is applied for selecting the optimized dominating structure layouts. At the third stage, alignments are optimized based on the determined structure layouts using a bi‐objective optimization method, which minimizes construction cost and geo‐hazard risk simultaneously. The proposed model and solution method are applied to two real‐world cases whose results verify their capabilities in producing alignment alternatives with better combinations of construction cost and geo‐hazard risk than manually designed alternatives.

Extremely fast Holocene coastal landscape evolution in the Kachchh Upland (NW India): Clues from a multidisciplinary review

ABSTRACT The coastal landscape of the Kachchh Upland (KU) region (NW-India) changed over the last few thousand years from a shallow marine gulf to a salty desert (1-4 meters asl). In this area, bordered to the south by the Northern Hill Range (NHR), the tectonic-climatic interaction triggered the sea level fall from +2/4 m circa (6000-2000 BP) to zero. An ancient river pattern deposited a tidally regulated delta area during the sea level fall that stopped 2000-3000 years ago due to tectonic activity and a dry climate. Deltaic-alluvial fans (DAF) in front of the NHR suggest that the KU's tectonic activity led to fast landscape evolution. We explored such drastic changes by integrating scientific information from a multidisciplinary literature review, identifying terraces and DAFs, and inferring faults through landform recognition, quantitative morphometry, and field surveys. Our interpretation, summarized in a map, provides new information on active processes along the NHR.

Automatic Landform Recognition, Extraction, and Classification using Kernel Pattern Modeling

Automatic Landform Recognition, Extraction, and Classification using Kernel Pattern Modeling

Quantitative analysis and spatial distribution of landform spatial structure on Loess Plateau

The Loess Plateau is the largest gully geomorphic region in the world, characterized by the most intense soil erosion in a typical loess-covered area. Previous studies have focused on the terrain and texture of this region; however, there have been no systematic studies on the gully spatial structure of the Loess Plateau. Therefore, the present study investigated the characteristics and spatial distribution of landform spatial structure over the Loess Plateau. Specifically, gully weighted complex networks (GWCNs) were applied to simulate the gully spatial structure of loess landforms, and a series of quantitative indices were introduced to delineate these. Using 57 geomorphological units uniformly distributed across the Loess Plateau as test areas and six typical loess landforms as sample areas, GWCNs were constructed, and using these GWCNs, the spatial structure and internal mechanisms of typical loess landforms were explored. From a series of fresh insights, such as the regional scale-free distribution, homologous structure, tightness, community effect, connectivity, stability, and complexity, regular variations in quantitative indices delineate the spatial distribution of the characteristics of landform spatial structure over the plateau. Moreover, the spatial distribution of complex network indices exhibited strong spatial coupling with loess landforms. Overall, GWCNs could be effectively used for landform recognition and performed well. In conclusion, these experimental results suggest that introducing complex networks into landform studies can offer novel insights into landform quantitative analyses. The present work is of great significance, as it proposes a new methodology for describing the spatial structure and terrain features of landforms in quantitative analyses and furthers our understanding of landform genesis.

Enhancing performance of multi-temporal tropical river landform classification through downscaling approaches

ABSTRACT Multi-temporal remote sensing imagery has the potential to classify river landforms to reconstruct the evolutionary trajectory of river morphologies. Whilst open-access archives of high spatial resolution imagery are increasingly available from satellite sensors, such as Sentinel-2, there remains a fundamental challenge of maximising the utility of information in each band whilst maintaining a sufficiently fine resolution to identify landforms. Although image fusion and downscaling methods on Sentinel-2 imagery have been investigated for many years, there is a need to assess their performance for multi-temporal object-based river landform classification. This investigation first compared three downscaling methods: area to point regression kriging (ATPRK), super-resolution based on Sen2Res, and nearest neighbour resampling. We assessed performance of the three downscaling methods by accuracy, precision, recall and F1-score. ATPRK was the optimal downscaling approach, achieving an overall accuracy of 0.861. We successively engaged a set of experiments to determine an optimal training model, exploring single and multi-date scenarios. We find that not only does remote sensing imagery with better quality improve river landform classification performance, but multi-date datasets for establishing machine learning models should be considered for contributing higher classification accuracy. This paper presents a workflow for automated river landform recognition that could be applied to other tropical rivers with similar hydro-geomorphological characteristics.

Scattering Center Modeling of Typical Planetary Landforms for Planetary Geomorphologic Exploration

Radar imaging of typical planet landforms can be used for planetary exploration such as landform recognition and geometry measurement. This communication realizes high-resolution imaging simulations of typical planetary landforms, such as craters, hills, and volcanoes, by using parametric scattering center (SC) models. Based on the geometric models of these typical landforms, parametric SC models are constructed from scattered solutions of the physical optics (POs) method. Because the SC models take geometric parameters as model parameters, these models can simulate the scattered waves of landforms of different sizes. These SC models are validated by the PO results. In addition, the landform types can be recognized and their geometrical parameters are further extracted by SC model matching. The extracted geometric parameters show good agreement with real landforms, demonstrating that the SC models of typical planetary landforms have potential applications in planetary geomorphologic exploration.

A Fine-Grained Genetic Landform Classification Network Based on Multimodal Feature Extraction and Regional Geological Context

Deep learning networks have facilitated the automated scene recognition of landforms based on geomorphogenesis. However, current genetic landform classification methods do not consider regional geological context, which can more accurately reflect the formation and evolution mechanism of geomorphic landforms than local ones. Therefore, this study proposes a multimodal, deep learning landform recognition framework based on a joint contextual geological and channel attention module (GCMENET). First, the multibranch feature extraction network of DenseNet121 is used to extract the respective features from the target scene and the contextual geological scene. Second, the features similar to the landform features of the target scene are extracted from the contextual geological features based on the cosine method and then combined with geomorphic features of the target scene. Third, channel attention mechanism is used to reduce the interference caused by redundant contextual geological information after fusion of data. To measure the classification accuracy of GCMENET, we establish a fine geomorphogenic dataset consisting of remote sensing images of six landform types with a <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$64\times64$ </tex-math></inline-formula> -pixel size and 10-m resolution (JOS10m). During the training process of two geomorphogenic datasets, the feature extraction network without batchnorm2d (batch normalization) could preserve the distribution and spatial alignment of data from the components. Using different training-to-validation data ratios and combinations of input components, the results of the GCMENET supplemented with the joint contextual geological and channel attention module exhibited greater accuracy than those obtained without the module. This observation confirms the importance of contextual geological information in automated geomorphogenic landforms.

Automatic Landform Recognition from the Perspective of Watershed Spatial Structure Based on Digital Elevation Models

Landform recognition is one of the most significant aspects of geomorphology research, which is the essential tool for landform classification and understanding geomorphological processes. Watershed object-based landform recognition is a new spot in the field of landform recognition. However, in the relevant studies, the quantitative description of the watershed generally focused on the overall terrain features of the watershed, which ignored the spatial structure and topological relationship, and internal mechanism of the watershed. For the first time, we proposed an effective landform recognition method from the perspective of the watershed spatial structure, which is separated from the previous studies that invariably used terrain indices or texture derivatives. The slope spectrum method was used herein to solve the uncertainty issue of the determination on the watershed area. Complex network and P–N terrain, which are two effective methodologies to describe the spatial structure and topological relationship of the watershed, were adopted to simulate the spatial structure of the watershed. Then, 13 quantitative indices were, respectively, derived from two kinds of watershed spatial structures. With an advanced machine learning algorithm (LightGBM), experiment results showed that the proposed method showed good comprehensive performances. The overall accuracy achieved 91.67% and the Kappa coefficient achieved 0.90. By comparing with the landform recognition using terrain indices or texture derivatives, it showed better performance and robustness. It was noted that, in terms of loess ridge and loess hill, the proposed method can achieve higher accuracy, which may indicate that the proposed method is more effective than the previous methods in alleviating the confusion of the landforms whose morphologies are complex and similar. In addition, the LightGBM is more suitable for the proposed method, since the comprehensive manifestation of their combination is better than other machine learning methods by contrast. Overall, the proposed method is out of the previous landform recognition method and provided new insights for the field of landform recognition; experiments show the new method is an effective and valuable landform recognition method with great potential as well as being more suitable for watershed object-based landform recognition.

Remote sensing landform recognition algorithm based on improved attention mechanism

Remote sensing landform recognition algorithm based on improved attention mechanism

Bag of Geomorphological Words: A Framework for Integrating Terrain Features and Semantics to Support Landform Object Recognition from High-Resolution Digital Elevation Models

High-resolution digital elevation models (DEMs) and its derivatives (e.g., curvature, slope, aspect) offer a great possibility of representing the details of Earth’s surface in three-dimensional space. Previous research investigations concerning geomorphological variables and region-level features alone cannot precisely characterize the main structure of landforms. However, these geomorphological variables are not sufficient to represent a complex landform object’s whole structure from a high-resolution DEM. Moreover, the amount of the DEM dataset is limited, including the landform object. Considering the challenges above, this paper reports an integrated model called the bag of geomorphological words (BoGW), enabling automatic landform recognition via integrating point and linear geomorphological variables, region-based features (e.g., shape, texture), and high-level landform descriptions. First, BoGW semantically characterizes the composition of geomorphological variables and meaningful parcels of each type of landform. Based on a landform’s semantics, the proposed method then integrates geomorphological variables and region-level features (e.g., shape, texture) to create the feature vector for the landform. Finally, BoGW classifies a region derived from high-resolution DEM into a predefined type of landform by the feature vector. The experimental results on crater and cirque detection indicated that the proposed BoGW could support landform object recognition from high-resolution DEMs.

Geomorphology of the post-caldera monogenetic volcanoes at Deception Island, Antarctica — Implications for landform recognition and volcanic hazard assessment

Deception Island is one of the most active volcanoes in Antarctica, with more than 20 monogenetic eruptions during the Holocene. The latest episodes of 1967, 1969, and 1970 have shown that volcanic activity on Deception Island can become a concern for tourists, scientists, and military personnel working on or near the island.The objective of this work is to identify eruptive processes and the geomorphic evolution of post-caldera monogenetic volcanic edifices at Deception Island by morphometric analyses, supported by field observations. Morphometric analysis has been used since the 1970s to analyse scoria cones, but it has rarely been applied to other monogenetic volcanoes, such as tuff cones and tuff rings. Tuff cones and tuff rings represent the most common landforms during Deception Island's recent geological past, with over 70 scattered eruptive vents inside and along the caldera rim. These volcanic landforms have been studied based on field observations and later, Digital Elevation Model analyses. Their geometry ranges from 10 to 100 m in height, and diameters vary between 300 and 2500 m. The morphometric data suggest that tuff cones have a more circular crater than tuff rings, with a significant separation between the two landforms using the height (Hco) and basal width (Wco) parameters. Tuff cones have Hco/Wco ratios with values between 0.04 and 0.16 and outer slopes between 7° and 34°. Tuff rings have lower Hco/Wco ratios that are between 0.01 and 0.04 and lower outer slopes between 3° and 21°.This study shows that basic shape parameters in combination with slope angle analysis can be used to discriminate among different types of monogenetic volcanoes. The initial eruptive-related morphometric diversity, however prevents correlation of eruption ages with morphometric parameters, or constructing a relative chronology of the monogenetic eruptions. This work provides a better comprehension of the potential evolution of a future eruption and a broader understanding of volcanic hazards on Deception Island.

Mountain summit detection with Deep Learning: evaluation and comparison with heuristic methods

Landform detection and analysis from Digital Elevation Models (DEM) of the Earth has been boosted by the availability of high-quality public data sets. Current landform identification methods apply heuristic algorithms based on predefined landform features, fine tuned with parameters that may depend on the region of interest. In this paper, we investigate the use of Deep Learning (DL) models to identify mountain summits based on features learned from data examples. We train DL models with the coordinates of known summits found in public databases and apply the trained models to DEM data obtaining as output the coordinates of candidate summits. We introduce two formulations of summit recognition (as a classification or a segmentation task), describe the respective DL models, compare them with heuristic methods quantitatively, illustrate qualitatively their performances, and discuss the challenges of training DL methods for landform recognition with highly unbalanced and noisy data sets.

Multi-modal deep learning for landform recognition

Automatic landform recognition is considered to be one of the most important tools for landform classification and deepening our understanding of terrain morphology. This paper presents a multi-modal geomorphological data fusion framework which uses deep learning-based methods to improve the performance of landform recognition. It leverages a multi-channel geomorphological feature extraction network to generate different characteristics from multi-modal geomorphological data, such as shaded relief, DEM, and slope and then it harvests joint features via a multi-modal geomorphological feature fusion network in order to effectively represent landforms. A residual learning unit is used to mine deep correlations from visual and physical modality features to achieve the final landform representations. Finally, it employs three fully-connected layers and a softmax classifier to generate labels for each sample data. Experimental results indicate that this multi-modal data fusion-based algorithm obtains much better performance than conventional algorithms. The highest recognition rate was 90.28%, showing a great potential for landform recognition.

Geomorphology- and geophysics-based recognition of stages of deep-seated slope deformation (Sudetes, SW Poland)

Slopes of the Kamienne Mountains, SW Poland, have been affected by a variety of slope deformations, from DSGSDs through rock slope collapses, rotational and translational landslides sensu stricto, to valley-confined earth flows. Variable geomorphic expression suggests that major slope deformation events are of different ages and some slope deformation may be ongoing. In this paper we use the combination of landform mapping and Electrical Resistivity Tomography profiling to characterize two settings, considered to represent early stages of slope deformation and advanced stages of ridge collapse. Landform recognition involved field geomorphic mapping and interpretation of high-resolution LiDAR DTM. Diagnostic features of early stages of deformation are open cracks, minor and rather subdued scarps in the upper slope, and lower slope bulging. Irregular hummocky topography, ruiniform relief and closed depressions typify the late stages. ERT results are generally consistent landform interpretation and show weakening of bedrock, evolving failure planes, as well as subsidence of rigid caprock blocks into deformable sedimentary series for the late stage. The combined geomorphological-geophysical approach performed fairly satisfactorily and both methods can be used to cross-check interpretations, even in areas not subject to recent mass movement events, poorly accessible due to dense forest cover and lacking clear morphological evidence.

Automated mapping of relict patterned ground: An approach to evaluate morphologically subdued landforms using unmanned-aerial-vehicle and structure-from-motion technologies

Relict landforms provide a wealth of information on the evolution of the modern landscape and climate change in the past. To improve understanding of the origin and development of these landforms we need better spatial measurements across a variety of scales. This can be challenging using conventional surveying techniques due to difficulties in landform recognition on the ground (e.g. weak visual/topographic expression) and spatially variable areas of interest. Here we explore the appropriateness of existing remote sensing datasets (aerial LiDAR and aerial photography) and newly acquired unmanned aerial vehicle (UAV) imagery of a test site on the upland of Dartmoor in SW England (Leeden Tor) for the recognition and automated mapping of relict patterned ground composed of stripes and polygons. We find that the recognition of these landforms is greatly enhanced by automated mapping using spectral two-dimensional imagery. Image resolution is important, with the recognition of elements (boulders) of &lt;1 m maximised from the highest resolution imagery (UAV red-green-blue (RGB)) and recognition of landforms (10–100 m scale) maximised on coarser resolution aerial imagery. Topographic metrics of these low relief (0.5 m) landforms are best extracted from structure-from-motion (SfM) processed UAV true-colour imagery, and in this context the airborne LiDAR data proved less effective. Integrating automated mapping using spectral attributes and SfM-derived digital surface models from UAV RGB imagery provides a powerful tool for rapid reconnaissance of field sites to facilitate the extraction of meaningful topographic and spatial metrics that can inform on the origin of relict landform features. Care should be given to match the scale of features under consideration to the appropriate scale of datasets available.

Landform Recognition in Granite Mountains in East Asia (Seoraksan, Republic of Korea, and Huangshan and Sanqingshan, China) – A Contribution of Geomorphology to the UNESCO World Heritage

Abstract Applied research in geomorphology includes landform analysis and evaluation from a specific perspective of scientific significance and global relevance. In this paper, landform diversity of Seoraksan, Republic of Korea, a UNESCO World Heritage candidate, is compared with geomorphic characteristics of two World Heritage properties in China, Huangshan and Sanqingshan. Seoraksan represents an almost complete mountain geomorphic system of considerable contemporary dynamics, with outstanding scenery and spectacular landforms such as domes, fins, bedrock channels, waterfalls, and inherited block fields. It is argued that Seoraksan contains outstanding scientific and aesthetic values, not present at the Chinese properties, offering scope for successful nomination.