Article
Chemistry, Analytical
Xiyu Zhang, Jianrong Fan, Jun Zhou, Linhua Gui, Yongqing Bi
Summary: In this study, the fire severity of a large forest fire in Southwest China was mapped at a 2m spatial resolution using random forest models and satellite imagery. The combination of Sentinel 2 and GF series images improved the classification accuracy compared to using only Sentinel 2 images. The addition of high-resolution GF series images enhanced the identification of low severity areas.
Article
Environmental Sciences
Steven E. Sesnie, Carlos I. Espinosa, Andrea K. Jara-Guerrero, Maria F. Tapia-Armijos
Summary: The increased variety of satellite remote sensing platforms provides opportunities for estimating tropical forest diversity. Only 10% of the original seasonally dry tropical forest remains in Ecuador, Peru, and Colombia. The remnant forests show high endemism rates but face degradation from various factors.
Article
Environmental Sciences
Changru Liu, Ximin Cui, Li Guo, Ling Wu, Xinming Tang, Shuhan Liu, Debao Yuan, Xia Wang
Summary: This paper presents an innovative method for combining GF-7 stereo images with laser altimetry data, which significantly improves the elevation accuracy. The validation experiments show that the root mean square error of elevation is greatly reduced after using the laser altimetry data, making it useful for reducing field survey work and improving mapping efficiency.
Article
Ecology
Demei Zhao, Jianing Zhen, Yinghui Zhang, Jing Miao, Zhen Shen, Xiapeng Jiang, Junjie Wang, Jincheng Jiang, Yuzhi Tang, Guofeng Wu
Summary: This study investigated the combined use of a radiative transfer model and a machine-learning model to estimate mangrove Leaf Area Index (LAI) using remote sensing images from different satellite sensors. The results showed that the Zhuhai-1 image had the best estimation accuracy, and newly developed three-band Vegetation Indices (VIs) proved effective in estimating mangrove LAI. Moreover, elevation and species composition were found to greatly influence the spatial distribution of mangrove LAI.
REMOTE SENSING IN ECOLOGY AND CONSERVATION
(2023)
Article
Computer Science, Artificial Intelligence
Mohamed Hosni, Juan M. Carrillo de Gea, Ali Idri, Manal El Bajta, Jose Luis Fernandez Aleman, Gines Garcia-Mateos, Ibtissam Abnane
Summary: Ensemble methods combine different techniques to overcome the limitations of single machine learning techniques, and are widely employed in research fields. In the study of cardiovascular diseases, ensemble classification techniques are commonly used, particularly for empirical research in diagnosis.
ARTIFICIAL INTELLIGENCE REVIEW
(2021)
Article
Remote Sensing
Hamid Ebrahimy, Zhou Zhang
Summary: Reliable classification of satellite images is important for various applications, and ensemble classifiers have shown success in this regard. This study introduced a new classification method called PAELM, which combines diverse extreme learning machine classifiers based on pixel-based accuracy values. Comparisons with benchmark classifiers showed that PAELM improved the accuracy of land cover and crop mapping, with an overall accuracy and F1 score of 0.811 and 0.804, respectively. PAELM was also found to be less sensitive to scene heterogeneity and class composition.
INTERNATIONAL JOURNAL OF APPLIED EARTH OBSERVATION AND GEOINFORMATION
(2023)
Article
Geosciences, Multidisciplinary
Abu Reza Md. Towfiqul Islam, Md. Mijanur Rahman Bappi, Saeed Alqadhi, Ahmed Ali Bindajam, Javed Mallick, Swapan Talukdar
Summary: This study uses machine learning techniques to predict the flash flood susceptibility in the Brahmaputra River Basin (BRB), and the results show that the hybrid model (RF-ANN) has the best prediction ability. The findings of this study are valuable for flood prevention and management, and are important for policymakers.
Article
Environmental Sciences
Seyed Vahid Razavi-Termeh, Abolghasem Sadeghi-Niaraki, Rizwan Ali Naqvi, Soo-Mi Choi
Summary: To mitigate the impact of dust on human health and the environment, this study utilized MODIS imagery to create a model and map that identifies the areas susceptible to dust in Bushehr province, Iran from 2002 to 2022. An improved ensemble machine learning model combined with evolutionary algorithms was used to prepare a dust susceptibility map (DSM). The results showed that altitude, wind speed, and land cover were the most influential factors in dust occurrence.
ENVIRONMENTAL POLLUTION
(2023)
Article
Geochemistry & Geophysics
Chia-Hsiang Lin, Man-Chun Chu, Po-Wei Tang
Summary: Mangrove mapping is a critical satellite remote sensing technology, but current benchmark methods are index-based and do not consider neighboring information or threshold setting. To improve accuracy, researchers propose a convex deep MM (CODE-MM) algorithm that incorporates deep learning and convex analysis. CODE-MM achieves small-data learning and computational efficiency by inferring a rough mangrove signature and using a convex criterion. Extensive experiments show that CODE-MM is insensitive to threshold setting and achieves state-of-the-art performance in accurate mangrove forest mapping.
IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING
(2023)
Article
Environmental Sciences
Zhen Shen, Jing Miao, Junjie Wang, Demei Zhao, Aowei Tang, Jianing Zhen
Summary: Mangrove forests are highly productive ecosystems with important ecological and economic value. Accurate mapping of mangrove forests is crucial for their management and restoration. This study utilizes multi-source remote sensing data to compare different feature selection methods and machine learning algorithms for accurate mangrove mapping. The results show that optical data performs better than SAR data, and the combination of optical and SAR data can further improve mapping accuracy. The XGBoost classification model achieves the highest overall accuracy. This research provides important insights and a reliable database for the restoration and protection of mangrove forests.
Article
Green & Sustainable Science & Technology
Juan Munizaga, Mariano Garcia, Fernando Ureta, Vanessa Novoa, Octavio Rojas, Carolina Rojas
Summary: This paper evaluates the potential of using high resolution satellite imagery to classify land cover in a coastal area in Concepcion, Chile, and finds that the random forest algorithm performs better in this type of landscape. Furthermore, incorporating Digital Terrain Model (DTM)-derived metrics and texture measures contributes significantly to the improvement of both support vector machine and random forest.
Article
Environmental Sciences
Seyed Vahid Razavi-Termeh, Abolghasem Sadeghi-Niaraki, MyoungBae Seo, Soo-Mi Choi
Summary: The purpose of this research is to analyze flood susceptibility mapping in the Sulaymaniyah province of Iraq. Machine learning algorithms, including random forest and bootstrap aggregation, were fine-tuned using a genetic algorithm. The study utilized various data sources to locate flooded areas and assess the performance of different models. The results showed that the Bagging-GA model had the highest accuracy in flood susceptibility modeling.
SCIENCE OF THE TOTAL ENVIRONMENT
(2023)
Article
Biodiversity Conservation
Yang Chen, Lixia Ma, Dongsheng Yu, Haidong Zhang, Kaiyue Feng, Xin Wang, Jie Song
Summary: This study investigates various feature selection methods for mapping soil organic matter (SOM) in restored forest land and finds that the ensemble method performs the best in improving prediction accuracy.
ECOLOGICAL INDICATORS
(2022)
Article
Chemistry, Multidisciplinary
Yidi Wei, Yongcun Cheng, Xiaobin Yin, Qing Xu, Jiangchen Ke, Xueding Li
Summary: Detailed information about mangroves is crucial for ecological and environmental protection and sustainable development. In this study, high-resolution satellite images were used to map the distribution of mangroves, and a deep-learning network, U-2-Net, was applied to extract multi-scale information. The results showed that the U-2-Net model performed well on mangrove classification, and the generated mangrove maps provided important information for monitoring dynamic changes.
APPLIED SCIENCES-BASEL
(2023)
Article
Environmental Sciences
Saeed Alqadhi, Javed Mallick, Swapan Talukdar, Ahmed Ali Bindajam, Tamal Kanti Saha, Mohd Ahmed, Roohul Abad Khan
Summary: The study constructed four optimized ensemble machine learning algorithms for landslide susceptibility mapping, with the PSO-ANN model identified as the best model and the LR model-based hybrid ensemble machine learning model performing better than the PSO-ANN model. Various resources were declared as landslide risk zones, and elevation, soil-texture, slope, rainfall, and road distance were considered the most sensitive parameters for landslide occurrences.
GEOCARTO INTERNATIONAL
(2022)
Article
Environmental Sciences
Yuanhui Zhu, Soe W. Myint, Danica Schaffer-Smith, Rebecca L. Muenich, Daoqin Tong, Yubin Li
Summary: Human-induced climate change is causing warmer conditions in the Southwestern United States, leading to more extreme urban heat island (UHI) effects. This study shows that increasing green vegetation, especially trees, can mitigate the UHI effects and lower surface temperatures. Affluent neighborhoods tend to have lower temperatures, while low-income communities experience higher temperatures. Replacing unmanaged soil with trees has the potential to significantly reduce summer daytime temperatures.
FRONTIERS IN ENVIRONMENTAL SCIENCE
(2022)
Article
Construction & Building Technology
Cho Kwong Charlie Lam, Jiafeng Weng, Kai Liu, Jian Hang
Summary: Shading effectively reduces heat stress and is a viable alternative to urban greenery. The influence of shading devices on microclimate and illuminance needs further investigation through observation and simulation.
BUILDING AND ENVIRONMENT
(2023)
Article
Environmental Sciences
Yuanhui Zhu, Soe W. Myint, Danica Schaffer-Smith, David J. Sauchyn, Xiaoyong Xu, Joseph M. Piwowar, Yubin Li
Summary: Canada's abundant water resources are crucial for its growing population and thriving economy. However, recent intense drought conditions have raised concerns about water availability. This study examined the changes in ground and surface water across all provinces in Canada and identified the factors influencing these changes, as well as predicted more frequent and severe droughts in the future.
JOURNAL OF ENVIRONMENTAL MANAGEMENT
(2022)
Article
Forestry
Ziyu Wang, Kai Liu, Jingjing Cao, Liheng Peng, Xin Wen
Summary: This study used remote sensing to monitor the dynamic changes of mangroves in China, revealing an overall increasing trend in both total area and mean patch size of mangroves, with Guangdong and Guangxi being the top two provinces in terms of mangrove area. Mangrove reserves in China, except for Dongzhaigang National Nature Reserve, showed increasing trends in mangrove area, confirming the effectiveness of the reserves.
Article
Engineering, Mechanical
Yanchun Yin, Wuwei Zheng, Xingxue Tang, Minglu Xing, Yubao Zhang, Yuanhui Zhu
Summary: Based on the stiffness theory, this study investigated the influence of loading stiffness on the mechanical behavior of sandstone specimens using an improved rock mechanics testing machine and low stiffness flexible rings. The results showed that the decrease in loading stiffness led to a change in the failure mode of the specimen, and there were correlations between the stiffness ratio and post-peak failure duration, as well as between the stress drop rate and stiffness ratio.
ENGINEERING FAILURE ANALYSIS
(2022)
Article
Geography, Physical
Yubin Li, Danica Schaffer-Smith, Soe W. Myint, Yuanhui Zhu
Summary: The Mekong River Basin has experienced extensive socio-economic development in recent decades, resulting in forest loss across the region. Cambodia faced the most significant forest loss, while Thailand and Myanmar saw net forest gain. Trends in net primary productivity and surface temperature varied starkly from north to south in the region, with a clear negative correlation between tree density decrease and surface temperature increase in degraded forest regions.
GISCIENCE & REMOTE SENSING
(2022)
Article
Environmental Studies
Kai Liu, Jingjing Cao, Minying Lu, Qian Li, Haojian Deng
Summary: This study examined the spatial and temporal dynamics of wetlands in the Guangdong-Hong Kong-Macao Greater Bay Area using remote sensing data. The results revealed the influence of natural factors and human activities on wetland evolution. The findings contribute to the scientific management and sustainable development of wetlands.
Article
Geography
Junwen Lu, Suhong Zhou, Zhong Zheng, Lin Liu, Mei-Po Kwan
Summary: This study advances the measurement of community social context by introducing a daily dynamic perspective and explores the relationship between community social context and community attachment. It measures the social context averaging or polarization effect of communities using census and cell phone data and investigates residents' community attachment in 71 communities in Guangzhou, China. The study finds that social contexts of many communities vary during the day, with distinct patterns during work hours, evening hours, and night hours. Considering variations in social contexts during evening hours significantly enhances the explanation of heterogeneity in residents' community attachment. The study suggests that context dynamics should be considered as a new dimension of community indicators and highlights the importance of targeting specific space and time periods in community governance practice.
GEOGRAFISKA ANNALER SERIES B-HUMAN GEOGRAPHY
(2023)
Article
Environmental Studies
Zhenzhi Jiao, Shaoying Li, Zhangping Lin, Zhipeng Lai, Zhuo Wu, Lin Liu
Summary: This study uses points of interest (POI) and high-resolution remote sensing images to identify and simulate future tourism land use in Xinxing County, China, considering the impact of high-speed rail (HSR). The results show that HSR-led development scenario leads to major changes in tourism land growth and provides insights for policymakers in terms of tourism sustainability and rural revitalization at the county level.
Article
Geosciences, Multidisciplinary
Yuanhui Zhu, Soe W. Myint, Xin Feng, Yubin Li
Summary: Cities in warmer climates are facing extreme heat and drought due to climate change. Developing a practical framework to balance land surface temperature reduction and water conservation is crucial for heat mitigation and resilience planning.
Article
Geography
Han Yue, Lin Liu, Luzi Xiao
Summary: This study applies a deep learning algorithm to extract on-street people and physical environment elements from street view images and constructs a spatial choice model to investigate the influence of people on the street and the streetscape's physical environments on the location choice of street theft crime offenders. The results show that the number of people on the street has a significant positive relationship with the offenders' preferences, while fences and plants have significant positive effects on attracting criminals. Grasses and sidewalks negatively affect offenders' location choices, and walls and windows do not significantly affect criminals' crime location choices.
Article
Environmental Sciences
Lin Liu, Jiayu Chang, Dongping Long, Heng Liu
Summary: Existing research suggests that COVID-19 lockdowns tend to decrease overall urban crime rates. This study conducted in ZG City, China, tested the relationship between violent crime and COVID-19 lockdown policies. The results showed a decline in violent crime during the lockdown, followed by a bounce-back post-lockdown. Violent crime moved away from the isolation location during the lockdown and continued to spread outward after the lift of the lockdown. Changes in COVID-19 risk ratings also affected the nearest crime distance during the lockdown. These findings contribute to the literature and have implications for joint crime and pandemic prevention and control.
INTERNATIONAL JOURNAL OF ENVIRONMENTAL RESEARCH AND PUBLIC HEALTH
(2022)
Article
Fisheries
Jinhao Zhou, Wu Zhou, Qiqi Zhou, Yuanhui Zhu, Fei Xie, Shen Liang, Yueming Hu
Summary: This study explores the impact of multiple pond conditions on the performance of extracting dike-pond from remote sensing images. The results show that multiple pond conditions can negatively affect the extraction performance and should be considered in applications.