Article
Environmental Sciences
Kelly L. Hondula, C. Nathan Jones, Margaret A. Palmer
Summary: Methane emissions from small freshwater ecosystems are a major uncertainty in the global methane budget. The study found that inundation extent and duration, as well as the direction of water level change, are major drivers of methane emissions in seasonally inundated forested wetlands. Modeling the hydrologic regime of these wetlands can lead to more accurate estimation of methane emissions.
ENVIRONMENTAL RESEARCH LETTERS
(2021)
Article
Environmental Sciences
Troy M. Saltiel, Philip E. Dennison, Michael J. Campbell, Tom R. Thompson, Keith R. Hambrecht
Summary: Recent advances in image classification using fine spatial resolution imagery from unoccupied aircraft systems (UASs) have shown that the spatial resolution has a significant impact on classification accuracy. Coarsening the spatial resolution can lead to a decrease in accuracy.
Article
Environmental Sciences
Jordi Cristobal, Patrick Graham, Anupma Prakash, Marcel Buchhorn, Rudi Gens, Nikki Guldager, Mark Bertram
Summary: A pilot study for mapping Arctic wetlands was conducted in the Yukon Flats National Wildlife Refuge in Alaska, using hyperspectral images and various classification methods to achieve the best classification performance. Recommendations for future work include the acquisition of LiDAR or RGB photo-derived digital surface models to improve classification efforts.
Article
Environmental Sciences
Farzane Mohseni, Meisam Amani, Pegah Mohammadpour, Mohammad Kakooei, Shuanggen Jin, Armin Moghimi
Summary: In this study, a wetland map of the GL region was created using Sentinel-1/2 datasets and the Google Earth Engine. A supervised machine learning classification workflow was used, with two main steps to accurately classify wetland and non-wetland areas. The overall accuracy and kappa coefficient of the classification results demonstrated the effectiveness of the proposed method.
Article
Environmental Sciences
Melinda Martinez, Marcelo Ardon, Mary Jane Carmichael
Summary: Wetlands are important sources of methane, and plant mediated transport is a major contributor. This study investigates the emissions from standing dead trees (snags) in wetlands and finds that the main source of methane emitted from snags is deep wetland soils, and methane is oxidized during transport.
FRONTIERS IN ENVIRONMENTAL SCIENCE
(2022)
Article
Environmental Sciences
J. Jack Kurki-Fox, Michael R. Burchell, Michael J. Vepraskas, Stephen W. Broome
Summary: The potential risk of copper and zinc buildup in wetland soils in North Carolina is relatively low, with most wetlands not posing a risk to biota from these metals. However, wetlands with a direct and significant anthropogenic source of metal contamination could be exceptions.
ENVIRONMENTAL MONITORING AND ASSESSMENT
(2021)
Article
Environmental Sciences
Chang Fan, Jilin Yang, Guosong Zhao, Junhu Dai, Mengyao Zhu, Jinwei Dong, Ruoqi Liu, Geli Zhang
Summary: This study compared ground and satellite observations and found that the 30m Landsat/Sentinel-2 data was more consistent with ground observations in wetland vegetation phenology, indicating its advantage over the 500m MODIS data. The study also highlighted the complexity of wetland phenology and its role in global climate change.
Article
Engineering, Electrical & Electronic
Hui Yang, Xiuguo Liu, Qihao Chen, Yuanye Cao
Summary: This study utilizes the capabilities of the Sentinel-1 satellite to construct time series similarity parameters and statistical texture parameters for more accurate classification of the Dongting Lake wetland. The results show that this method effectively utilizes dynamic information among different classes and can accurately identify different geomorphic features in wetlands.
IEEE JOURNAL OF SELECTED TOPICS IN APPLIED EARTH OBSERVATIONS AND REMOTE SENSING
(2022)
Article
Environmental Studies
Krista L. Noe, Christopher T. Rota, Mack W. Frantz, James T. Anderson
Summary: Wetland restoration is a common practice to offset the loss of natural wetlands due to human interference. While researchers often compare bird, amphibian, and reptile communities in restored and natural wetlands, they tend to overlook small mammals. However, small mammals play a crucial role in the ecosystem as seed dispersers and prey for larger wildlife. A study conducted in West Virginia, USA, found that restored wetlands provided sufficient habitat for small mammal communities, but the abundance of certain species differed between restored and natural wetlands.
Article
Environmental Sciences
Txomin Hermosilla, Alex Bastyr, Nicholas C. Coops, Joanne C. White, Michael A. Wulder
Summary: Knowledge of tree species is essential for forest management and monitoring, and can be achieved through remote sensing and spatial modeling. This study used National Forest Inventory data and machine learning algorithms to map and classify tree species in Canada's forest-dominated ecosystems. The overall accuracy of the classification models was 93.1%, with geographic, climatic, and topographic variables being the most influential. The most common leading tree species nationally were black spruce, trembling aspen, and lodgepole pine, while regionally there was dominance of other tree species.
REMOTE SENSING OF ENVIRONMENT
(2022)
Article
Green & Sustainable Science & Technology
Kathryn Ledford, Stephanie Ann Schmidt, Changwoo Ahn
Summary: This study assessed the soil carbon storage potential in forested wetlands in the urbanized region of Northern Virginia and found that these wetlands have the capacity to store carbon, especially in the topsoil.
Article
Geography, Physical
Xudong Hu, Penglin Zhang, Qi Zhang, Junqiang Wang
Summary: Through this study, the use of ensemble methods significantly improves the accuracy and stability of wetland cover classification, with RANN and MBANN showing the best overall performance in accuracy, resistance to data size reduction and feature variability.
GISCIENCE & REMOTE SENSING
(2021)
Article
Ecology
Allison Aldous, Steve Schill, George Raber, Marie-Claire Paiz, Emmanuel Mambela, Tariq Stevart
Summary: A novel method was used to accurately map the wetlands in the coastal region of Gabon, with results showing wetlands covering only 22% of the project area, mainly found close to water bodies. However, only 33% of the mapped wetlands fall within the boundaries of Wetlands of International Importance designated under the Ramsar Convention.
REMOTE SENSING IN ECOLOGY AND CONSERVATION
(2021)
Article
Computer Science, Interdisciplinary Applications
Rafael Blanquero, Emilio Carrizosa, Cristina Molero-Rio, Dolores Romero Morales
Summary: Classification and Regression Trees (CARTS) are commonly used techniques in modern Statistics and Machine Learning, but their classification accuracy may not be high. A new approach based on continuous optimization has been proposed recently, showing good performance in practice.
COMPUTERS & OPERATIONS RESEARCH
(2021)
Article
Management
Rafael Blanquero, Emilio Carrizosa, Cristina Molero-Rio, Dolores Romero Morales
Summary: This paper proposes an optimal regression tree model based on a continuous optimization problem, which aims to strike a balance between prediction accuracy and sparsity. The model can fulfill important properties for regression tasks and provide local explanations due to the smoothness of predictions. The computational experience demonstrates the superiority of this approach in terms of prediction accuracy compared to standard benchmark regression methods, and the scalability with respect to sample size is also illustrated.
EUROPEAN JOURNAL OF OPERATIONAL RESEARCH
(2022)