| Autor: |
Saheba Bhatnagar, Laurence Gill, Shane Regan, Owen Naughton, Paul Johnston, Steve Waldren, Bidisha Ghosh |
| Jazyk: |
angličtina |
| Rok vydání: |
2020 |
| Předmět: |
|
| Zdroj: |
International Journal of Applied Earth Observations and Geoinformation, Vol 88, Iss , Pp 102083- (2020) |
| Druh dokumentu: |
article |
| ISSN: |
1569-8432 |
| DOI: |
10.1016/j.jag.2020.102083 |
| Popis: |
Wetlands provide habitat for a wide variety of plant and animal species and contribute significantly to overall biodiversity in Ireland. Despite these known ecosystem services, the total wetland area in Ireland has reduced significantly over the past few decades leading to an ongoing need to protect such environments. The EU Habitats Directive (92/43/EEC) has recognised several wetlands types as “priority” habitats. This study concentrates on a subset of the priority habitats focussing on some groundwater dependent terrestrial ecosystems, (in particular calcareous fens and turloughs), as well as raised bogs. Monitoring these sites across the country by field visits is resource-intensive. Therefore, this study has evaluated remote sensing as a potentially cost-effective tool for monitoring the ecological health of the wetlands. Identification and presence of certain vegetation communities can indicate the condition of the wetland, which can be used for monitoring, for example, activities causing degradation or the progress of restoration attempts. The ecological composition of the wetlands has been analysed using open-source Sentinel-2 data. 10 bands of Sentinel-2 Level-2 data and 3 indices, Normalised Difference Vegetation Index (NDVI), Enhanced Vegetation Index (EVI) and Normalised Difference Water Index (NDWI) were used to create vegetation maps of each wetland using Bagged Tree (BT) ensemble classifier and graph cut segmentation also known as MAP (maximum a posteriori) estimation. The proposed methodology has been validated on five raised bogs, five turloughs, and three fens at different times during 2017 and 2018 from which three case studies are presented. An overall classification accuracy up to 87% depending on the size of the vegetation community within each wetland has been achieved which suggests that the proposed method is appropriate for wetland health monitoring. |
| Databáze: |
Directory of Open Access Journals |
| Externí odkaz: |
|
Full Text from ScienceDirect