Large-scale Mapping Research Articles

Research on the A* Algorithm for Automatic Guided Vehicles in Large-Scale Maps

The traditional A* algorithm faces the challenges of low search efficiency and large node extension range in the field of path planning. These directly restrict the overall performance of the algorithm. In this study, we aimed to improve the search efficiency and path planning quality of the A* algorithm in complex and large-scale environments through a series of optimisation measures, including the innovation of weight design, flexible adjustment of the search neighbourhood, improvement of the heuristic function, and optimisation of the node selection strategy. Specifically, this study innovatively introduces the local obstacle rate as the core index of weight design, and it dynamically adjusts the weights according to the change of the obstacle rate during the node movement process, which effectively reduces the search space and significantly improves the search speed. At the same time, according to the real-time change of the local obstacle rate, this study dynamically adjusts the range of the search neighbourhood, so that the algorithm can choose the optimal search strategy according to different environmental information. In terms of the improvement of the heuristic function, this study adopted the diagonal distance as the benchmark for cost estimation, and it innovatively introduces the angle coefficient to reflect the complexity of path turning, thus providing the algorithm with a more accurate guidance for the search direction. In addition, this study optimises the node selection method by drawing on the idea of simulated annealing, which eliminates the need to calculate and compare all possible surrogate values during the node selection process, thus significantly reducing the running time of the algorithm. The results of the simulation experiments fully verify the effectiveness and practicality of the improved algorithm. Compared with the traditional A* algorithm, the improved algorithm achieved significant optimisation in terms of the average running time, the number of expansion nodes, and the path length, with the average running time shortened by 84%, the number of expansion nodes reduced by 94%, and the path length also shortened by 2.3%.

Rapid lithological mapping using multi-source remote sensing data fusion and automatic sample generation strategy

ABSTRACT The advancement of remote sensing technology aids geologists in obtaining lithological maps more quickly, comprehensively, and accurately. However, key challenges in lithological mapping include the limited spectral information from individual sensors and the difficulties in visually interpreting lithological samples. In this study, we integrated 241 scenes of optical data and 106 scenes of radar data on the Google Earth Engine (GEE) platform, proposing a rapid lithological identification framework that combines an automatic lithological sample data generation strategy with multi-source data. Using various machine learning algorithms, we evaluated the classification capabilities of heterogeneous predictive factors, feature optimization algorithms, and object-based algorithms. Results indicate that: (1) Combining optical and radar data improves prediction accuracy, with terrain data further enhancing mapping capabilities; (2) Terrain factors contribute most to classification, but SWIR and TIR bands of optical data are critical for lithological identification; (3) The feature optimization algorithm reduces feature redundancy and efficiency issues from multi-source data, achieving 96.51% accuracy with the optimal feature model, an improvement of 0.1%−2.02% over original features; (4) Object-based algorithms show significant potential in mapping areas with large rock outcrops. This study offers new insights for medium- to large-scale lithological maps and provides essential data support for geological work.

Optimum Combination of Spectral Variables for Crop Mapping in Heterogeneous Landscapes based on Sentinel-2 Time Series and Machine Learning

Abstract. This article aimed to determine a workflow for more efficient large-scale crop mapping using a time series of images from the Sentinel-2 Satellite, statistical methods of attribute selection, and machine learning. The proposed methodology explores the best possible combination of spectral variables related to vegetation (16 vegetation indices in the RGB, NIR, SWIR, and Red Edge regions) to characterize different spectro-temporal profiles of Land Use and Land Cover (LULC) in spatially heterogeneous landscapes. First, we applied a data dimensionality reduction analysis using the PCA (Principal Component Analysis) method. Subsequently, the variables that showed the highest statistical correlation between each other were used in the spectro-temporal classification process, using the Random Forest, TempCNN, and LightTAE algorithms, following three different strategies: C1 (ALL), C2 (BE + IV (Red Edge)) and C3 (BE + IV (without Red Edge)), where ALL – All variables; BE – Spectral Bands; IV – Vegetation Indices. Given the results found, the C2 classification scenario (with bands B3, B4, B5, B6, B7, B8, and B8A and the NDRE1, RESI, and MSR indexes) demonstrated the best LULC classification accuracy at the crop pattern level, compared to the other scenarios, with average values of 0.91, 0.88, 0.91, 0.89, and 0.89 (Global Accuracy, Producer Accuracy, User Accuracy, Kappa index, and F1-Score, respectively, for the TempCNN model), the which emphasized the importance of both qualitative and quantitative variability of sampling data and variables based on the Red Edge region for improving LULC classification processes in large-scale heterogeneous landscapes.

Forecasts and Statistical Insights for Line Intensity Mapping Cross-correlations: A Case Study with 21 cm × [C ii]

Intensity mapping—the large-scale mapping of selected spectral lines without resolving individual sources—is quickly emerging as an efficient way to conduct large cosmological surveys. Multiple surveys covering a variety of lines (such as the hydrogen 21 cm hyperfine line, carbon-monoxide rotational lines, and [C ii] fine-structure lines, among others) are either observing or will soon be online, promising a panchromatic view of our Universe over a broad redshift range. With multiple lines potentially covering the same volume, cross-correlations have become an attractive prospect, both for probing the underlying astrophysics and for mitigating observational systematics. For example, cross-correlating 21 cm and [C ii] intensity maps during reionization could reveal the characteristic scale of ionized bubbles around the first galaxies, while simultaneously providing a convenient way to reduce independent foreground contaminants between the two surveys. However, many of the desirable properties of cross-correlations in principle emerge only under ideal conditions, such as infinite ensemble averages. In this paper, we construct an end-to-end pipeline for analyzing intensity mapping cross-correlations, enabling instrumental effects, foreground residuals, and analysis choices to be propagated through Monte Carlo simulations to a set of rigorous error properties, including error covariances, window functions, and full probability distributions for power-spectrum estimates. We use this framework to critically examine the applicability of simplifying assumptions such as the independence and Gaussianity of power-spectrum errors. As worked examples, we forecast the sensitivity of near-term and futuristic 21 cm × [C ii] cross-correlation measurements, providing recommendations for survey design.

Neural Surfel Reconstruction: Addressing Loop Closure Challenges in Large-Scale 3D Neural Scene Mapping

Efficiently reconstructing complex and intricate surfaces at scale remains a significant challenge in 3D surface reconstruction. Recently, implicit neural representations have become a popular topic in 3D surface reconstruction. However, how to handle loop closure and bundle adjustment is a tricky problem for neural methods, because they learn the neural parameters globally. We present an algorithm that leverages the concept of surfels and expands relevant definitions to address such challenges. By integrating neural descriptors with surfels and framing surfel association as a deformation graph optimization problem, our method is able to effectively perform loop closure detection and loop correction in challenging scenarios. Furthermore, the surfel-level representation simplifies the complexity of 3D neural reconstruction. Meanwhile, the binding of neural descriptors to corresponding surfels produces a dense volumetric signed distance function (SDF), enabling the mesh reconstruction. Our approach demonstrates a significant improvement in reconstruction accuracy, reducing the average error by 16.9% compared to previous methods, while also generating modeling files that are up to 90% smaller than those produced by traditional methods.

CLOUDSPAM: Contrastive Learning On Unlabeled Data for Segmentation and Pre-Training Using Aggregated Point Clouds and MoCo

SegContrast TEST paved the way for contrastive learning on outdoor point clouds. Its original formulation targeted individual scans in applications like autonomous driving and object detection. However, mobile mapping purposes such as digital twin cities and urban planning require large-scale dense datasets to capture the full complexity and diversity present in outdoor environments. In this paper, the SegContrast method is revisited and adapted to overcome its limitations associated with mobile mapping datasets, namely the scarcity of contrastive pairs and memory constraints. To overcome the scarcity of contrastive pairs, we propose the merging of heterogeneous datasets. However, this merging is not a straightforward procedure due to the variety of size and number of points in the point clouds of these datasets. Therefore, a data augmentation approach is designed to create a vast number of segments while optimizing the size of the point cloud samples to the allocated memory. This methodology, called CLOUDSPAM, guarantees the performance of the self-supervised model for both small- and large-scale mobile mapping point clouds. Overall, the results demonstrate the benefits of utilizing datasets with a wide range of densities and class diversity. CLOUDSPAM matched the state of the art on the KITTI-360 dataset, with a 63.6% mIoU, and came in second place on the Toronto-3D dataset. Finally, CLOUDSPAM achieved competitive results against its fully supervised counterpart with only 10% of labeled data.

Android-based framework for condition assessment of existing RC hospital buildings

Rapid visual screening (RVS) is a practically viable tool implemented all over the world to rank structures based on their seismic vulnerability. Seismic assessment of hospital buildings is essential in seismic-prone regions to ensure their structural integrity and readiness to provide emergency care and medical treatment after significant seismic events. Therefore, in the current study, an approach is developed that is similar to the proven methodologies reported in the literature, for the seismic evaluation of structural and non-structural elements of hospital buildings in a metropolitan city like Karachi. Karachi is located in the southwest of Pakistan and is vulnerable to earthquake and tsunami hazards. To this end, an android-based RVS system has been developed for large-scale seismic risk mapping of hospital buildings. The proposed system acquires the physical status of the building features in real time and identifies the building performance obtained from the calculated seismic risk index. Moreover, three case study hospital buildings were employed and analysed through the proposed RVS system. Results illustrate the performance of case study hospital buildings at organisational, structural and non-structural levels based on seismic risk indices.

Precision crop mapping: within plant canopy discrimination of crop and soil using multi-sensor hyperspectral imagery

Leveraging diverse optomechanical and imaging technologies for precision agriculture applications is gaining attention in emerging economies. The precise spatial detection of plant objects in farms is crucial for optimizing plant-level nutrition and managing pests and diseases. High-resolution remote sensors mounted on drones have been increasingly deployed for large-scale crop mapping and field variability characterization. While field-level crop identification and crop-soil discrimination have been studied extensively, within-plant canopy discrimination of crop and soil has not been explored in real agricultural farms. The objectives of this study are: (i) adoption and assessment of spectral unmixing for discriminating crop and soil at within-plant canopy level, and (ii) generation of benchmark terrestrial and drone-based hyperspectral datasets for plant or sub-plant level discrimination using various spectral mixture modelling approaches and sources of endmembers. We acquired hyperspectral imagery of vegetable crops using a frame-based sensor mounted on a drone flying at different heights. Further, several linear, non-linear, and sparse-based spectral unmixing methods were used to discriminate plant and soil based on spectral signatures (endmembers) extracted from different spectral libraries prepared using in situ or field, ground-based, and drone-based hyperspectral imagery. The results, validated against pixel-to-pixel ground truth data, indicate an overall crop-soil discrimination accuracy of 99–100%, subject to a combination of endmember source and flying height. The influences of different endmember sources, spatial resolution as indicated by flying height, and inversion algorithms on the quality of estimated abundances are assessed from a verifiable and functionally relevant perspective. The generated hyperspectral datasets and ground truth data can be used for developing and testing new methods for sub-canopy level soil-crop discrimination in various agricultural applications of remote sensing.

Research on a Multi-Scale Clustering Method for Buildings Taking into Account Visual Cognition

Building clustering is a key problem that needs to be solved in the realization of the automatic synthesis of large-scale maps. The selection of different feature and spatial distance calculation methods has a great impact on the clustering results, and the need to manually select appropriate feature and distance metrics leads to the problem of not being able to fully consider the complexity and diversity of buildings. In this paper, we propose a multi-scale clustering method for buildings that takes visual perception into account using the Gestalt principle to simulate how humans classify buildings through visual perception. Moreover, by analyzing the spatial features and texture attributes of buildings, a visual distance is designed to be used as a condition for building classification to assess the similarity between buildings, solving the complexity of manually selecting feature vectors and spatial distances and realizing the adaptive selection of features. Through experimental validation at different scales (macro, meso and micro), the present method is able to achieve the accurate clustering of buildings, and a frequency threshold of 91% is found, which is able to determine the optimal clustering results. The experimental results show that the proposed method can not only fully consider the complexity and diversity of buildings but also effectively support the understanding and analysis of urban spatial structure and provide a scientific decision-making basis for urban planning and management.

A “Region-Specific Model Adaptation (RSMA)”-Based Training Data Method for Large-Scale Land Cover Mapping

An accurate and historical land cover monitoring dataset for Alaska could provide fundamental information for a range of studies, such as conservation habitats, biogeochemical cycles, and climate systems, in this distinctive region. This research addresses challenges associated with the extraction of training data for timely and accurate land cover classifications in Alaska over longer time periods (e.g., greater than 10 years). Specifically, we designed the “Region-Specific Model Adaptation (RSMA)” method for training data. The method integrates land cover information from the National Land Cover Database (NLCD), LANDFIRE’s Existing Vegetation Type (EVT), and the National Wetlands Inventory (NWI) and machine learning techniques to generate robust training samples based on the Anderson Level II classification legend. The assumption of the method is that spectral signatures vary across regions because of diverse land surface compositions; however, despite these variations, there are consistent, collective land cover characteristics that span the entire region. Building upon this assumption, this research utilized the classification power of deep learning algorithms and the generalization ability of RSMA to construct a model for the RSMA method. Additionally, we interpreted existing vegetation plot information for land cover labels as validation data to reduce inconsistency in the human interpretation. Our validation results indicate that the RSMA method improved the quality of the training data derived solely from the NLCD by approximately 30% for the overall accuracy. The validation assessment also demonstrates that the RSMA method can generate reliable training data on large scales in regions that lack sufficient reliable data.

Functional specialization and distributed processing across marmoset lateral prefrontal subregions.

A prominent aspect of primate lateral prefrontal cortex organization is its division into several cytoarchitecturally distinct subregions. Neurophysiological investigations in macaques have provided evidence for the functional specialization of these subregions, but an understanding of the relative representational topography of sensory, social, and cognitive processes within them remains elusive. One explanatory factor is that evidence for functional specialization has been compiled largely from a patchwork of findings across studies, in many animals, and with considerable variation in stimulus sets and tasks. Here, we addressed this by leveraging the common marmoset (Callithrix jacchus) to carry out large-scale neurophysiological mapping of the lateral prefrontal cortex using high-density microelectrode arrays, and a diverse suite of test stimuli including faces, marmoset calls, and spatial working memory task. Task-modulated units and units responsive to visual and auditory stimuli were distributed throughout the lateral prefrontal cortex, while those with saccade-related activity or face-selective responses were restricted to 8aV, 8aD, 10, 46V, and 47. Neurons with contralateral visual receptive fields were limited to areas 8aV and 8aD. These data reveal a mixed pattern of functional specialization in the lateral prefrontal cortex, in which responses to some stimuli and tasks are distributed broadly across lateral prefrontal cortex subregions, while others are more limited in their representation.

Surprisingly good fit of pressure-based cropland condition map and bird census data at the national scale

As the world is struggling to halt the rapid decline of biodiversity, the assessment and mapping of ecosystem condition is getting ever increasing attention. Croplands are artificial ecosystems, but as they occupy a large portion of land, they significantly influence biodiversity. Yet, there is a knowledge gap about their suitability to support and maintain wildlife on a national scale. Large-scale condition mapping is meant to address this gap; the good condition of croplands includes their ability to support biodiversity. However, the lack of suitable databases is often a challenge when creating such maps. As there is a strong causal relation between pressures on the ecosystem and its condition, pressure indicators can be used as proxies to approximate condition when more direct indicators are lacking. The validation of condition maps based on pressure proxies is a key but challenging step. In this study, we tested a previously designed pressure-based cropland condition map for Hungary using bird census data. Besides validating the composite condition indicator, we also tested some key elements of the mapping process, such as the choice of variables and thresholds.Using multiple comparisons of means by Tukey’s contrast and Random Forest modelling, we examined the relationship of (1) the continuous pressure-based cropland condition variables, (2) their rescaled, ordinal version (called sub-indicators), and (3) the composite cropland condition indicator (sum of the sub-indicators) with a biodiversity measure, the standardised relative richness of characteristic farmland bird species (rRRCS). To get a picture of the spatial patterns of the examined relationships across Hungary, individual Random Forest models were constructed for all the spatial units of the bird census database, using focal analysis with a 30 km radius moving window.We found significant differences in the mean rRRCS for nearly all sub-indicator categories, signifying that the literature-derived thresholds were mostly sound. Categories with higher (better) condition scores had higher mean rRRCS; the differences are significant in the mid-range but not in the extreme categories, indicating a need for a meaningful simplification of the categories. The goodness-of-fit (R2) of the Random Forest models was found to be high, but it is spatially heterogeneous (ranged from 0.69 to 0.89, with a median value of 0.81), similarly to the variable importance. The proportion of semi-natural areas proved to be the most important condition variable. The proportion of maize and alfalfa were more important than parcel size. Our results show that condition maps based on pressure proxies can reflect patterns of biodiversity surprisingly well. They also highlight the spatial context dependence of the uncertainty of condition maps.

Leveraging Open-Source Geographic Databases to Enhance the Representation of Landscape Heterogeneity in Ecological Models.

Wildlife abundance and movement are strongly impacted by landscape heterogeneity, especially in cities which are among the world's most heterogeneous landscapes. Nonetheless, current global land cover maps, which are used as a basis for large-scale spatial ecological modeling, represent urban areas as a single, homogeneous, class. This often requires urban ecologists to rely on geographic resources from local governments, which are not comparable between cities and are not available in underserved countries, limiting the spatial scale at which urban conservation issues can be tackled. The recent expansion of community-based geographic databases, for example, OpenStreetMap (OSM), represents an opportunity for ecologists to generate large-scale maps geared toward their specific research needs. However, computational differences in language and format, and the high diversity of information within, limit the access to these data. We provide a framework, using R, to extract geographic features from the OSM database, classify, and integrate them into global land cover maps. The framework includes an exhaustive list of OSM features describing urban and peri-urban landscapes and is validated by quantifying the completeness of the OSM features characterized, and the accuracy of its final output in 34 cities in North America. We portray its application as the basis for generating landscape variables for ecological analysis by using the OSM-enhanced map to generate an urbanization index, and subsequently analyze the spatial occupancy of six mammals throughout Chicago, Illinois, USA. The OSM features characterized had high completeness values for impervious land cover classes (50%-100%). The final output, the OSM-enhance map, provided an 89% accurate representation of the landscape at 30m resolution. The OSM-derived urbanization index outperformed other global spatial data layers in the spatial occupancy analysis and concurred with previously seen local response trends, whereby lagomorphs and squirrels responded positively to urbanization, while skunks, raccoons, opossums, and deer responded negatively. This study provides a roadmap for ecologists to leverage the fine resolution of open-source geographic databases and apply it to spatial modeling by generating research-specific landscape variables. As our occupancy results show, using context-specific maps can improve modeling outputs and reduce uncertainty, especially when trying to understand anthropogenic impacts on wildlife populations.

Highly Urbanized Mangrove Areas are Small in Size, Fragmented, and Missed by Large-scale Mapping Efforts

Highly Urbanized Mangrove Areas are Small in Size, Fragmented, and Missed by Large-scale Mapping Efforts

Assessing the Potential of UAV for Large-Scale Fractional Vegetation Cover Mapping with Satellite Data and Machine Learning

Fractional vegetation cover (FVC) is an essential metric for valuating ecosystem health and soil erosion. Traditional ground-measuring methods are inadequate for large-scale FVC monitoring, while remote sensing-based estimation approaches face issues such as spatial scale discrepancies between ground truth data and image pixels, as well as limited sample representativeness. This study proposes a method for FVC estimation integrating uncrewed aerial vehicle (UAV) and satellite imagery using machine learning (ML) models. First, we assess the vegetation extraction performance of three classification methods (OBIA-RF, threshold, and K-means) under UAV imagery. The optimal method is then selected for binary classification and aggregated to generate high-accuracy FVC reference data matching the spatial resolutions of different satellite images. Subsequently, we construct FVC estimation models using four ML algorithms (KNN, MLP, RF, and XGBoost) and utilize the SHapley Additive exPlanation (SHAP) method to assess the impact of spectral features and vegetation indices (VIs) on model predictions. Finally, the best model is used to map FVC in the study region. Our results indicate that the OBIA-RF method effectively extract vegetation information from UAV images, achieving an average precision and recall of 0.906 and 0.929, respectively. This method effectively generates high-accuracy FVC reference data. With the improvement in the spatial resolution of satellite images, the variability of FVC data decreases and spatial continuity increases. The RF model outperforms others in FVC estimation at 10 m and 20 m resolutions, with R2 values of 0.827 and 0.929, respectively. Conversely, the XGBoost model achieves the highest accuracy at a 30 m resolution, with an R2 of 0.847. This study also found that FVC was significantly related to a number of satellite image VIs (including red edge and near-infrared bands), and this correlation was enhanced in coarser resolution images. The method proposed in this study effectively addresses the shortcomings of conventional FVC estimation methods, improves the accuracy of FVC monitoring in soil erosion areas, and serves as a reference for large-scale ecological environment monitoring using UAV technology.

The Frasnian stage of the Upper Devonian in Nakhichevan Autonomous Republic (Southern Transcaucasia)

Marine carbonate-terrigenous deposits of the Lower (partially), Middle, and Upper Devonian, rich in coral-brachiopod and other fauna, are widespread across the territory of the Nakhichevan Autonomous Republic. Various ore and non-metallic minerals are associated with these deposits: the Gyumushlug and Yaydzhinsky polymetallic deposits, etc. Further ex-pansion of the mineral resources of the study area, as well as the challenges associated with the establishment of large-scale geological mapping in this region, make the development of a detailed stratigraphic scheme crucial. An indispensable condition for the creation of such a scheme is the study of the most characteristic groups of ancient organisms, among which brachiopods play a dominant role in the organic world. Devonian deposits in the Southern Transcaucasus are distributed over a limited area – in the western part of the Na- khchivan Autonomous Republic – but they are widespread here and are represented by all three well-characterized faunal divisions. For the first time, a diagram has been compiled, reflecting the sequence of appearance and distribution over time of Devonian brachiopods. On this basis, a scheme for the zonal division of the Devonian deposits in the region has been developed, identifying 20 biostratigraphic zones and 16 stratigraphic units (formations). This study summarizes the scattered information contained in various sources on the paleontology and stratigraphy of the Devonian period in the region. The identified brachiopod zones are linked to standard conodont zones. The great importance of brachiopods is determined by the fact that they are one of the most effective groups of fauna not only for dating sediments in natural outcrops, but they also provide good results for broad interregional correlations. Based on the macrofaunal complex, the Frasnian age of the sediments within the formation in question is firmly established. A detailed biostratigraphic scheme with zonal division of the Devonian deposits of Transcaucasia was developed, based on the most important stages in the development of the brachiopod fauna. Biostratigraphic zones, formations, and subformations identified in the Devonian deposits of the Nakhchivan Autonomous Republic and adjacent areas, based on the study of the entire complex of organic remains, mainly brachiopods and cono- donts, can be compared with corresponding deposits of typical sections of Western Europe (Ardennes and Eifel Mountains) and many regions of Eurasia. The obtained data are of practical importance for theoretical geologists, prospectors and developers.

Large-scale map of RNA-binding protein interactomes across the mRNA life cycle

mRNAs interact with RNA-binding proteins (RBPs) throughout their processing and maturation. While efforts have assigned RBPs to RNA substrates, less exploration has leveraged protein-protein interactions (PPIs) to study proteins in mRNA life-cycle stages. We generated an RNA-aware, RBP-centric PPI map across the mRNA life cycle in human cells by immunopurification-mass spectrometry (IP-MS) of ∼100 endogenous RBPs with and without RNase, augmented by size exclusion chromatography-mass spectrometry (SEC-MS). We identify 8,742 known and 20,802 unreported interactions between 1,125 proteins and determine that 73% of the IP-MS-identified interactions are RNA regulated. Our interactome links many proteins, some with unknown functions, to specific mRNA life-cycle stages, with nearly half associated with multiple stages. We demonstrate the value of this resource by characterizing the splicing and export functions of enhancer of rudimentary homolog (ERH), and by showing that small nuclear ribonucleoprotein U5 subunit 200 (SNRNP200) interacts with stress granule proteins and binds cytoplasmic RNA differently during stress.

Sparse large-scale high-order fuzzy cognitive maps guided by spearman correlation coefficient

Time series prediction is one of the most important applications of Fuzzy Cognitive Maps (FCMs). In general, the state of FCMs in forecasting depends only on the state of the previous moment, but in fact it is also affected by the past state. Hence Higher-Order Fuzzy Cognitive Maps (HFCMs) are proposed based on FCMs considering historical information and have been widely used for time series forecasting. However, using HFCMs to deal with sparse and large-scale multivariate time series are still a challenge, while large-scale data makes it difficult to determine the causal relationship between nodes because of the increased number of nodes, so it is necessary to explore the relationship between nodes to guide large-scale HFCMs learning. Therefore, a sparse large-scale HFCMs learning algorithm guided by Spearman correlation coefficient, called SG-HFCM, is proposed in the paper. The SG-HFCM model is specified as follows: First, the solving of HFCMs model is transform into a regression model and an adaptive loss function is utilized to enhance the robustness of the model. Second, l1-norm is used to improve the sparsity of the weight matrix. Third, in order to more accurately characterize the correlation relationship between variables, the Spearman correlation coefficients is added as a regular term to guide the learning of weight matrices. When calculating the Spearman correlation coefficient, through splitting domain interval method, we can better understand the characteristics of the data, and get better correlation in different small intervals, and more accurately characterize the relationship between the variables in order to guide the weight matrix. In addition, the Alternating Direction Multiplication Method and quadratic programming method are used to solve the algorithms to get better solutions for the SG-HFCM, where the quadratic programming can well ensure that the range of the weights and obtaining the optimal solution. Finally, by comparing with five algorithms, the SG-HFCM model showed an average improvement of 11.93% in prediction accuracy for GRNs, indicating that our proposed model has good predictive performance.

Prediction and spatial–temporal changes of soil organic matter in the Huanghuaihai Plain by combining legacy and recent data

Soil organic matter (SOM) is critical for soil fertility, crop growth, and plays an important role in the global carbon cycle and climate change. Therefore, spatial prediction of SOM is important to rational soil resource utilization, agricultural production, and ecological environment management. However, large-area SOM mapping research heavily relies on legacy soil data, and large-scale recent SOM mapping may not be possible or have lower accuracy due to limited or less recent data availability. In this study, we aimed to improve SOM prediction and mapping accuracy by combining legacy data with limited recent data. Three models, namely, partial least squares regression (PLSR), random forest (RF), and one-dimensional convolutional neural network (1D-CNN), were applied and compared. The results showed that combining legacy and recent data effectively improved SOM prediction accuracy compared to using only recent data. Among the three modeling methods, 1D-CNN exhibited superior performance, with an averaged determination coefficient of the prediction (R2) of 0.58, a root mean square error (RMSE) of 4.56 g/kg, and a ratio of performance to interquartile distance (RPIQ) of 2.05. The predicted SOM content for both legacy (1980 s) and recent (2010 s) periods showed similar spatial distribution patterns throughout the Huanghuaihai Plain. Generally, there was a noticeable trend of increasing SOM content from northwest to southeast, with higher values observed in Jiangsu and lower values concentrated in Henan, Hebei, and Shandong regions within the study area. Over time, SOM contents in the Huanghuaihai Plain showed an increasing trend, with an average increase of 5.90 g/kg from legacy to recent period. This study provides a promising approach for improving SOM prediction and mapping accuracy at large scales, particularly when recent data availability is limited.

Predictive GAM seabed maps can account for defined and fuzzy boundaries to improve accuracy in a Scottish sea loch seascape

Marine seabed mapping is an important element in marine spatial and conservation planning. Recent large scale mapping programmes have greatly increased our knowledge of the seafloor, yet at finer resolutions, large gaps remain. Loch Eriboll, Scotland, is an area of conservation interest with a diverse marine environment supporting habitats and species of conservation importance. Here we test and present strategies for a predictive seabed substrata map for Loch Eriboll using drop down Stereo Baited Remote Underwater Video (SBRUV) imagery collected as part of systematic underwater survey of the Loch. A total of 216 SBRUV deployments were made across the study site in depths of 3 m–117 m, with six seabed classes identified using an adaptation of the EUNIS (European Nature Information System) hierarchical habitat classification scheme. Four statistical learning approaches were tested, we found, Generalised Additive Models (GAMs) provided the optimal balance between over- and underfitted predictions. We demonstrate the creation of a predictive substratum habitat map covering 63 km2 of seabed which predicts the probability of presence and relative proportion of substratum types. Our method enables naturally occurring edges between habitat patches to be described well, increasing the accuracy of mapping habitat boundaries when compared to categorical approaches. The predictions allow for both defined boundaries such as those between sand and rock and fuzzy boundaries seen among fine mixed sediments to exist in the same model structure. We demonstrate that SBRUV imagery can be used to generate cost effective, fine scale predictive substrata maps that can inform marine planning. The modelling procedure presented has the potential for a wide adoption by marine stakeholders and could be used to establish baselines for long term monitoring of benthic habitats and further research such as animal distribution and movement work which require detailed benthic maps.