Article
Environmental Sciences
Chongzhi Chen, Zhangquan Shen, Yuhui Weng, Shixue You, Jingya Lin, Sinan Li, Ke Wang
Summary: In this study, models were developed to evaluate landslide susceptibility in forest-covered areas in Lin'an, southeastern China. Logistic regression, decision tree, and random forest techniques were used, and key predictors were identified as forest type, understory vegetation height, normalized differential vegetation index in summer, distance to road, and maximum daily rainfall. The results showed that forest cover information is essential for predicting landslides and conversion from natural forests to plantations could increase landslide risk. The study provides valuable information for understanding landslide occurrences and designing disaster mitigation.
Article
Geography, Physical
Hanne Hendrickx, Gaelle Le Roy, Agnes Helmstetter, Eric Pointner, Eric Larose, Luc Braillard, Jan Nyssen, Reynald Delaloye, Amaury Frankl
Summary: High mountain environments are increasingly affected by rockfall-related hazards, driven by climate change. Studying rockfall in these environments is challenging due to the inaccessibility of mountain ridges and complex controlling factors. This study presents a detailed investigation of a rock wall in the Swiss Alps using various data collection methods. The dataset is unique as it started before the destabilization of the rock wall, providing insights into precursory indicators of large-scale events.
EARTH SURFACE PROCESSES AND LANDFORMS
(2022)
Article
Geography, Physical
Alexandre Legay, Florence Magnin, Ludovic Ravanel
Summary: The study indicates that warm permafrost areas are more susceptible to rockfalls, with failures directly linked to high temperatures. Furthermore, surface temperatures and scar depths significantly increase in the days to weeks leading up to rockfall events.
PERMAFROST AND PERIGLACIAL PROCESSES
(2021)
Article
Multidisciplinary Sciences
Soongu Kwak, Hyun-Jung Lee, Seungyeon Kim, Jun-Bean Park, Seung-Pyo Lee, Hyung-Kwan Kim, Yong-Jin Kim
Summary: This study aimed to explore the differences in the association between cardiovascular risk factors and atherosclerotic cardiovascular disease (ASCVD) risk in men and women using machine learning. A random forest model was developed to predict the 10-year ASCVD probabilities in each sex. The results showed that there were significant sex-specific associations between cardiovascular risk factors and ASCVD events, with higher total cholesterol or LDL cholesterol levels being more strongly associated with the risk of ASCVD in men, while older age and increased waist circumference were more strongly associated with the risk of ASCVD in women.
SCIENTIFIC REPORTS
(2023)
Article
Computer Science, Artificial Intelligence
Nuttanan Wichitaksorn, Yingyue Kang, Faqiang Zhang
Summary: Random subspace logistic regression is a regression model that improves classification or prediction accuracy by randomly selecting features. It can be applied to both standard and lasso logistic regression models. The proposed method shows promising results in simulations and empirical data, indicating its potential as an alternative for traditional feature selection approaches.
EXPERT SYSTEMS WITH APPLICATIONS
(2023)
Article
Geography, Physical
Shibing Huang, Haowei Cai, Zekun Xin, Gang Liu
Summary: Global warming has caused increased rockfall activities in alpine mountains due to the thawing of ice-filled joints in rock masses. This study conducted compression-shear experiments to investigate the shear strength of ice-filled rock joints, considering various factors such as joint roughness, temperature, opening, shear rates, and normal stress. The research found that joint roughness and aggregation of rupture ice area significantly affect the shear strength of ice-filled joints. The opening of joints and increasing temperature reduce the effect of joint roughness, while shear rate and normal stress have opposing effects. The findings contribute to a better understanding of the degradation mechanism of ice-filled joints.
Article
Green & Sustainable Science & Technology
Arvind Chandra Pandey, Tirthankar Ghosh, Bikash Ranjan Parida, Chandra Shekhar Dwivedi, Reet Kamal Tiwari
Summary: The Indian Himalayan region has been experiencing frequent hazards and disasters associated with permafrost. However, there has been little or no research on permafrost in this region. It is important to have knowledge about the spatial distribution and state of permafrost in the Indian Himalayas. This study used modern remote sensing techniques and a geographic information system (GIS) to assess permafrost in the Alaknanda Valley of the Chamoli district in Uttarakhand state. The results showed that a logistic regression model based on mean annual air temperature produced the most accurate predictions of permafrost distribution.
Article
Computer Science, Artificial Intelligence
Ke Yuan, Yabing Huang, Jiabao Li, Chunfu Jia, Daoming Yu
Summary: This paper proposes a block cipher algorithm identification scheme based on hybrid random forest and logistic regression. The experimental results show that the proposed scheme achieves higher accuracy and stability compared to existing methods.
NEURAL PROCESSING LETTERS
(2023)
Article
Computer Science, Interdisciplinary Applications
Shangmin Zhao, Weiming Cheng, Yecheng Yuan, Zemeng Fan, Jin Zhang, Chenghu Zhou
Summary: This paper aims to simulate and predict global permafrost distribution, and analyze its change under different climate scenarios. The research found that the area of degraded permafrost is the largest under RCP85 scenarios, mainly distributed in East Asia, West Asia, North Europe and North America.
ENVIRONMENTAL MODELLING & SOFTWARE
(2022)
Article
Geography, Physical
Seamus Daly, Philip P. Bonnaventure, Will Kochtitzky
Summary: Access to permafrost distribution maps is limited for remote communities in the discontinuous permafrost zone. This study presents a time- and cost-efficient method for conducting community-scale permafrost mapping using binary logistic regression and various variables. Vegetation is found to be the strongest predictor of near-surface permafrost.
PERMAFROST AND PERIGLACIAL PROCESSES
(2022)
Article
Computer Science, Artificial Intelligence
Yuan Zhong, Hongyu Yang, Yanci Zhang, Ping Li
Summary: Continuous data streams mining poses challenges for machine learning. The ORB-RRF is an online rebuilding regression random forests model designed to adapt to dynamic data streams, showing significant improvements in adaptability and predictive accuracy through numerical experiments.
KNOWLEDGE-BASED SYSTEMS
(2021)
Article
Biochemical Research Methods
Cansu Alakus, Denis Larocque, Aurelie Labbe
Summary: This study proposes a new method called Covariance Regression with Random Forests (CovRegRF) for estimating the covariance matrix of a multivariate response given covariates. The method utilizes a random forest framework to build decision trees with a specially designed splitting rule to maximize differences between sample covariance matrix estimates of child nodes. A significance test for the partial effect of a subset of covariates is also introduced. Simulation studies and an application to thyroid disease data demonstrate the accuracy and control of Type-1 error of the proposed method. CovRegRF is available as a free R package on CRAN.
BMC BIOINFORMATICS
(2023)
Article
Computer Science, Information Systems
Chao-Yu Guo, Yi-Jyun Lin
Summary: A proper statistical approach is crucial in medical and health sciences to avoid erroneous conclusions. Different genders and drug interactions can have significant impacts on therapeutic effects and efficacy. This study proposes a new method called random interaction forest (RIF) based on random forest, which outperforms other models in considering interactions and making predictions under various scenarios.
Article
Forestry
T. J. Boettcher, Baburam Rijal, James Cook, Shuva Gautam
Summary: This study aimed to develop a predictive model for the occurrence and abundance of buckthorn in Wisconsin by establishing sample plots and constructing different types of regression models. The ZINB model was identified as the best model for estimating buckthorn presence and abundance, indicating that factors such as stem density, woody species diversity, and environmental variables were important for predicting buckthorn invasion.
FOREST ECOLOGY AND MANAGEMENT
(2022)
Article
Engineering, Multidisciplinary
Siru Wang, Guoqi Qian, John Hopper
Summary: In this study, a coherent procedure is developed to analyze the association between phenotype and genotype using genomic data with missing values and non-ignorable missingness mechanism, by integrating statistical learning methods and hypothesis testing.
APPLIED MATHEMATICAL MODELLING
(2023)
Article
Geography, Physical
Fabian Fleischer, Jan-Christoph Otto, Robert R. Junker, Daniel Hoelbling
Summary: Debris cover on glaciers is a key component influencing climate-glacier dynamics and glacier lifespan. The study in the Eastern Alps of Austria shows an increase of over 10% in debris cover on glaciers from 1996 to 2015, while glaciers retreated in response to climate warming. Different mountain ranges exhibit significant regional variability in debris cover on glaciers.
EARTH SURFACE PROCESSES AND LANDFORMS
(2021)
Article
Environmental Sciences
Eric Petermann, Hanna Meyer, Madlene Nussbaum, Peter Bossew
Summary: The study developed an improved spatial continuous GRP map based on a large amount of field measurements in Germany, utilizing three different machine learning algorithms and conducting prediction performance evaluation using spatial cross-validation.
SCIENCE OF THE TOTAL ENVIRONMENT
(2021)
Article
Geography, Physical
Stefan Haselberger, Lisa-Maria Ohler, Robert R. Junker, Jan-Christoph Otto, Thomas Glade, Sabine Kraushaar
Summary: This study investigated small-scale biogeomorphic interactions on proglacial slopes by setting up erosion plots along a plant cover gradient. The results showed that there were two significant declines in geomorphic activity when plant cover exceeded 30% and 75% respectively. Analysis of vegetation composition and environmental conditions revealed the impact of high-magnitude geomorphic events on the environment and species communities.
EARTH SURFACE PROCESSES AND LANDFORMS
(2021)
Article
Ecology
Hanna Meyer, Edzer Pebesma
Summary: Machine learning algorithms are popular for spatial mapping due to their ability to fit complex relationships, but their use is limited to data similar to the training set. The study proposes a method to assess the area where a prediction model can be reliably applied, using a Dissimilarity Index (DI) to define the Area of Applicability (AOA) and map estimated performance. Simulation studies show comparable prediction errors within the AOA to cross-validation errors, emphasizing the importance of considering the relationship between DI and cross-validation performance.
METHODS IN ECOLOGY AND EVOLUTION
(2021)
Article
Geography, Physical
Jan-Christoph Otto, Kay Helfricht, Guenther Prasicek, Daniel Binder, Markus Keuschnig
Summary: The study simulated the subglacial topography in the Austrian Alps and found that the current models overestimate the number and location of potential future glacial lakes, especially in valley type glaciers. The research suggests that the modeling approach performs better on large valley type glaciers and less accurately on mountain glaciers. The results indicate that up to 42 new lakes may form within 23 glaciers in Austria.
EARTH SURFACE PROCESSES AND LANDFORMS
(2022)
Article
Plant Sciences
Lisa-Maria Ohler, Sarah Seeleitner, Stefan Haselberger, Sabine Kraushaar, Jan-Christoph Otto, Birgit Mitter, Robert R. Junker
Summary: Bacterial communities in the phyllosphere are influenced by host genotype and phenotype, as well as spatio-temporal variation of the environment, with the potential to alter plant phenotype. Manipulation of leaf-associated microbial communities in different successional stages within a glacier foreland can significantly influence microbial composition and functional plant traits over the course of a growing season.
Article
Environmental Sciences
Alice Ziegler, Hanna Meyer, Insa Otte, Marcell K. Peters, Tim Appelhans, Christina Behler, Katrin Boehning-Gaese, Alice Classen, Florian Detsch, Juergen Deckert, Connal D. Eardley, Stefan W. Ferger, Markus Fischer, Friederike Gebert, Michael Haas, Maria Helbig-Bonitz, Andreas Hemp, Claudia Hemp, Victor Kakengi, Antonia V. Mayr, Christine Ngereza, Christoph Reudenbach, Juliane Roeder, Gemma Rutten, David Schellenberger Costa, Matthias Schleuning, Axel Ssymank, Ingolf Steffan-Dewenter, Joseph Tardanico, Marco Tschapka, Maximilian G. R. Vollstaedt, Stephan Woellauer, Jie Zhang, Roland Brandl, Thomas Nauss
Summary: The monitoring of species and functional diversity plays a crucial role in the development of biodiversity conservation and management strategies. This study explores the potential of using airborne LiDAR-derived variables and elevation models to predict animal species richness. The results demonstrate that elevation models are more effective in predicting species richness than LiDAR-based models, with little additional contribution from structural information.
Article
Environmental Sciences
Lilian-Maite Lezama Valdes, Marwan Katurji, Hanna Meyer
Summary: Land Surface Temperature (LST) plays a crucial role in monitoring environmental and biological processes, especially in the Antarctic Dry Valleys. Researchers successfully enhanced the spatial resolution of MODIS satellite LST product to 30 meters using machine learning models trained with Landsat 8 thermal infrared data, providing valuable data for various research applications.
Editorial Material
Multidisciplinary Sciences
Hanna Meyer, Edzer Pebesma
Summary: This article examines the data and methods commonly used in creating published global maps of ecological variables, and discusses the possibility of assessing the quality of predicted values on a global and local scale.
NATURE COMMUNICATIONS
(2022)
Article
Ecology
Carles Mila, Jorge Mateu, Edzer Pebesma, Hanna Meyer
Summary: This study proposes a new cross-validation strategy that takes into account the geographical prediction space and compares it with other established methods. The new method, called NNDM LOO CV, provides reliable estimates in all scenarios considered. The existing methods, LOO and bLOO CV, have limitations and only provide accurate estimates in certain situations. Therefore, considering the geographical prediction space is essential when designing map validation methods.
METHODS IN ECOLOGY AND EVOLUTION
(2022)
Article
Environmental Sciences
Luca Kleinewillinghoefer, Pontus Olofsson, Edzer Pebesma, Hanna Meyer, Oliver Buck, Carsten Haub, Beatrice Eiselt
Summary: Land cover area estimates can be derived using design-based or earth-observation-based mapping approaches. The former requires costly and effective sampling, while the latter may be biased. Combining reference samples and remote sensing products can provide a more efficient method. The study tests different methods to estimate imperviousness areas in selected European countries, highlighting the use and limitations of existing reference information and remote sensing products.
Article
Ecology
Marvin Ludwig, Alvaro Moreno-Martinez, Norbert Hoelzel, Edzer Pebesma, Hanna Meyer
Summary: Global-scale maps of the environment are crucial for researchers and decision makers, but the feasibility of making predictions beyond the training data location has been questioned. We propose a new workflow that combines recent developments in spatial predictive mapping to improve model transferability and performance assessment. However, the limited availability of reference data hampers the creation and evaluation of reliable gap-free global predictions.
GLOBAL ECOLOGY AND BIOGEOGRAPHY
(2023)
Article
Geography
Ikram Zangana, Jan-Christoph Otto, Roland Maeusbacher, Lothar Schrott
Summary: We present a detailed geomorphological map (1:5000-scale) of a middle mountainous area in Jena, Germany, using geographic information systems (GIS) and high-resolution digital data. The map features were extracted by manually interpreting the combination of different data sources using light detection and ranging (LiDAR) data. By incorporating the visual interpretation of multidirectional hillshade and land surface parameters (LSPs) composite maps, we were able to systematically delineate landforms and geomorphological process domains.
Article
Geochemistry & Geophysics
Maciej Dabski, Ireneusz Badura, Marlena Kycko, Anna Grabarczyk, Renata Matlakowska, Jan-Christoph Otto
Summary: This study focuses on the development of glacial forelands as a result of the contemporary retreat of glaciers, providing excellent areas for studying the initial stages of weathering. The research examines weathering rinds on glacially abraded Dachstein limestone surfaces in the Eastern Alps. Results show a time-dependent increase in micro-roughness, decrease in rock strength, and decrease in spectral reflectivity within visual light. Older sites reflect infrared radiation significantly better than younger ones. The study also suggests the potential role of microorganisms in limestone dissolution and the formation of secondary porous limestone layers.
Proceedings Paper
Geography, Physical
Marvin Ludwig, Jonathan Bahlmann, Edzer Pebesma, Hanna Meyer
Summary: This article presents a workflow that combines the enhanced processing capabilities of the openEO cloud computation platform with state-of-the-art machine learning model development in R. The workflow allows for standardized imagery acquisition and preprocessing, reduction of overfitting, and assessment of mapping accuracy to enhance and evaluate the spatial transferability of machine learning models.
XXIV ISPRS CONGRESS: IMAGING TODAY, FORESEEING TOMORROW, COMMISSION III
(2022)
Article
Geography, Physical
A. Securo, C. Del Gobbo, L. Rettig, S. Pillon, A. De Luca, D. Fontana, E. Benedetti Fasil, R. R. Colucci
Summary: Small glaciers in temperate mountain regions have experienced significant reduction and unprecedented melt rates in recent years. Some glaciers have transitioned from clean ice to debris-covered or even rock glaciers. This study examines the surface elevation change of the Popera Alto glacier in the Sesto Dolomites using LiDAR and Structure from Motion surveys, and analyzes its evolution in terms of surface cover and geomorphic processes. The glacier has lost an average of 0.35 m water equivalent per year over the past 16 years, with active modification of its surface cover by geomorphic processes. The role of debris and local topography feedback has allowed the resilience of the glacier, leading to a marked difference between the current environmental equilibrium line altitude (envELA) and the effective ELA (effELA) of the glacier.
Article
Geography, Physical
Zhenzhen Yan, Yaolin Shi, Lili Kang, Xiangtao Fan
Summary: This study proposes a quantitative regional deformation model based on global positioning system (GPS) data to quantitatively analyze the morphological evolution of rivers in the Three Rivers Region. It finds that tectonic deformation phases significantly control regional landscape development and drainage features.
Article
Geography, Physical
Said Mukhtar Ahmad, Nitheshnirmal Sadhasivam, Mona Lisa, Luigi Lombardo, Mustafa Kemal Emil, Amira Zaki, Cees J. Van Westen, Islam Fadel, Hakan Tanyas
Summary: In this study, we investigated a large slow-moving landslide in Northern Pakistan, using Interferometric Synthetic Aperture Radar (InSAR) analysis. Our results showed that the crown of the landslide is moving faster than the surrounding regions, while the footslope experienced high deformations. We discussed the possible roles of meteorologic and anthropogenic factors in causing these deformations.
Article
Geography, Physical
Shuang Bian, Xibin Tan, Yiduo Liu, Suoya Fan, Junfeng Gong, Chao Zhou, Feng Shi, Michael A. Murphy
Summary: The Yarlung River's drainage divide is primarily moving north due to variations in precipitation across the Himalayas. The Gangdese drainage divide shows predominantly northward and southward migration, controlled by base-level rise and downstream influences. The presence of north-trending rifts separates the drainage divides into five zones, each with a distinct migration pattern.
Article
Geography, Physical
Joon-Young Park, Seok Yoon, Deuk-Hwan Lee, Seung-Rae Lee, Hwan-Hui Lim
Summary: This study developed a multiple-regression model to estimate site-specific average growth rates of debris flow events. The proposed model was validated through a case study and showed reasonable predictions of debris flow velocities and heights.
Article
Geography, Physical
Nicholas Reilly Mccarroll, Arnaud Temme
Summary: New geochronological data from hillslope boulder armor in the Flint Hills reveal the rates and timing of lateral retreat in the landscape. Surfaces of limestone boulders dating back to the Pleistocene era were found, and the ages of the hillslope armor increased with distance from the limestone bench. The estimated rate of lateral retreat in this landscape is 0.02 mm/yr.
Article
Geography, Physical
Xinbo Yao, Yuntao Tian
Summary: By studying the Longmenshan-Minshan drainage divide, we found that it has reached a dynamic steady state, indicating a balance between erosion and rock uplift. This study also reveals the process of formation and evolution of the divide and raises questions about the effectiveness of divide migration metrics.
Article
Geography, Physical
Junhui Yu, Pin Yan, Yanlin Wang, Guangjian Zhong, Changliang Chen
Summary: The seafloor mounds in the Chaoshan Depression of the South China Sea are identified as mud volcanoes, with fluids coming from underlying mud-fluid diapirs. The hydrocarbon gases feeding the mud volcanoes and diapirs are reasoned to originate from deep Mesozoic source rocks, indicating significant Mesozoic hydrocarbon potential in the Chaoshan Depression.
Article
Geography, Physical
Marius Huber, Luc Scholtes, Jerome Lave
Summary: This paper investigates the relationships between hillslope stability and fabric anisotropy of brittle rock materials and the implications for landscape shaping. It explores the different stability modes and movement characteristics of anisotropic materials, and demonstrates the significant control of material anisotropy on landscape shaping.
Article
Geography, Physical
Shubhra Sharma, Anil D. Shukla
Summary: The study investigates the relationship between glacial dynamics and lake sedimentation during the mid-Holocene climate variability in the Southern Zanskar ranges. It utilizes geomorphological disposition, elemental geochemistry, and optical chronology of relict lake sediment to reconstruct the pattern of minor glacier responses to climate variability. The results indicate six centennial to millennial-scale climatic phases, with warmer phases represented by decreased mineralogical fine grain flux and increased coarse grain flux. The study highlights the potential of relict lake sediment and para/peri-glacial landforms in understanding glacial dynamics and climate change during the Holocene.
Article
Geography, Physical
Jean-Francois Bernier, Sydney W. Meury, Patrick Lajeunesse
Summary: In this study, an approach combining various data and observation methods was proposed to improve the monitoring of landfast ice dynamics and its geomorphic impact on sedimentary systems. The results demonstrate the ability of the approach to accurately measure interannual variations in landfast ice and constrain geomorphic changes. Additionally, the study found a strong relationship between the severity of freezing seasons and the response of landfast ice to hydrometeorological events, with different geomorphic responses observed under different winter conditions.
Article
Geography, Physical
Heping Shu, Fanyu Zhang
Summary: This study investigates the relationship between susceptibility of soil-water hazards and human activities, geoheritage sites in the Loess Plateau, China. Landslide and gully erosion susceptibility were obtained using gradient boosting and support vector machines, and a hazard matrix was formed to couple landslide and gully erosion susceptibility. The study found different trends in the magnification times of soil-water hazards chain under different scenarios.
Article
Geography, Physical
Guangqiang Qian, Zhuanling Yang, Xuegang Xing, Zhibao Dong, Youyuan Guo
Summary: Granule ripples are aeolian landforms armored against erosion by coarse grains. This study investigates their seasonal morphological evolution and migration in the Sanlongsha Dune Field. The findings show that wind events, especially those exceeding the threshold velocities of coarse grains, significantly influence the morphodynamics of granule ripples. The study highlights the importance of considering the reptation and saltation of coarse grains in future research on granule ripples.