A carregar...
Publicação
Dinâmica da neve e seus impactes na vegetação na Ilha de Deception (Shetland do Sul)
| Nome: | Descrição: | Tamanho: | Formato: | |
|---|---|---|---|---|
| 3.16 MB | Adobe PDF |
Orientador(es)
Resumo(s)
This dissertation analyses the dynamics of wet snow and its impacts on the vegetation of Deception Island (South Shetlands, Antarctica) between 2016 and 2022. To characterize the snow cover, Sentinel-1 images were processed in Google Earth Engine, QGIS, and R, allowing the calculation of the annual number of wet snow days, which ranged from 132 to 144 days throughout the study period. Spatial analysis revealed strong interannual variability, as well as the influence of altitude and local topography on snow persistence. Vegetation mapping was performed using high-resolution Kompsat-3 imagery (2020), complemented by georeferenced photographs and field data. Several supervised classification methods were tested (Spectral Angle Mapper, Random Forest, and Maximum Likelihood), with the latter performing best, achieving an overall accuracy of 98.72% and a Kappa coefficient of 0.9858. Two main vegetation classes were identified—moss communities and moss-algal communities—which distinguished themselves from other similar surface covers, such as algal ice and penguin colonies. The results show that vegetation is sparse and distributed in very small patches, located primarily in low-altitude coastal areas and areas with low snow persistence (0–34 days/year). In contrast, higher-altitude glacial regions with persistent snow cover (121–144 days/year) showed no vegetation, with the snow acting as a thermal and physical barrier to plant colonization. Transition areas, with intermediate snow cover duration (49–96 days/year), proved ecologically relevant, functioning as biological refugia where microenvironments favourable to mosses and cryptogamic algae can develop. This study contributes to the understanding of snow-vegetation interactions on Deception Island by analysing how snow cover conditions the spatial distribution of cryptogamic plant communities.
Descrição
THAWIMPACT, financiado pela Fundação para a Ciência e a Tecnologia (contrato 2022.06628.PTDC Bolsa BI-PTHAWIMPACT- 2 2023)
Palavras-chave
Antarctica Wet snow Vegetation Remote sensing Deception Island Antártida Neve húmida Vegetação Deteção remota Ilha de Deception