Space-time variations of environmental indicators of vegetation development in ice-free areas of Maritime Antarctica

• Ano de defesa: 2022
• Autor(a) principal: Ferrari, Flávia Ramos
• Tipo de documento: Tese
• Tipo de acesso: Acesso aberto
• Idioma: eng
• Instituição de defesa: Universidade Federal de Viçosa; Botânica
• Palavras-chave em Português: Comunidades vegetais - Antártica, Península; Mudanças climáticas - Antártica, Península; Geleiras - Antártica, Península; Ciências Biológicas
• Link de acesso: https://locus.ufv.br//handle/123456789/29636; https://doi.org/10.47328/ufvbbt.2022.361

The relationship between the environment and the form of vegetation growth materializes in Antarctic vegetation due to its high sensitivity to environmental changes. On a regional scale, the climate in Antarctica is the main driver that affects plant community diversity and structure, however, on a smaller scale, topography and soil properties are the more important factors. With the increase in temperature and subsequent glacier retreat, new areas are exposed. Monitoring the vegetation distribution, biogenic factors, and the glacier retreat are key factors that predict the ecological responses due to the climate changes. The study aimed to evaluate how environmental conditions influence and interact with the distribution of plant communities, identifying biotic (birds and soil microbial composition) and abiotic (soils, landscape, and greenhouses gases) factors that interfere in the dynamics of the Antarctic ecosystem. This evaluation was conducted in three distinct Antarctic regions in the South Shetland Archipelago: Thomas Point and Demay Point in King George Island, and Harmony Point in Nelson Island. For this, we measure the laboratory greenhouse gases production potential at four different temperatures (-2, 4, 6, and 22 ºC), from 11 communities in a topography gradient. To evaluate the net ecosystem CO 2 exchange from soil covers, compare the microbial community, and characterize the compounds of organic matter, we analyzed distinct vegetation covers with light chamber systems and sample soils. In three rocky outcrops, we used linear mixed effect models and soil chemistry and physical analyses to evaluate the plant frequency and coverage, richness patterns, and the main effects of potential predictors. We identify 11 plant communities through a phytosociology survey, product orthomosaic with high resolution from drone aerial images, and mapped the mean vegetation, glaciers, and total bare soil through manual classification and machine learning using the models avNNet, kknn, and rf were predicted, based in 10 classes. We compared the glacier retreating and total bare soil area in three ortho imagens from distant years. Our results showed that there is an overall trend of increasing greenhouse gases production potential with increasing temperature along a toposequence, with N 2 O sink in most communities at -2 ºC, and the highest CH 4 emissions in moss carpets. Net ecosystem exchange of CO 2 exhibits variation between sites and plant covers, microbial community, and phosphorus availability, with higher CO 2 emissions due to organic matter deposited, soil microbial diversity, and the lack of current vegetation. We found that changes in the substrate along the outcrops presumably also promote high species turnover between the three rocky fragments, where habitat-mediated filtering can determine species richness differences. The avNNet predictor had higher accuracy and when comparing the automatic with the manual method of mapping, they diverge in diverse situations, but machine learning is possible to predict ice-free areas and help the field efforts in Antarctica. Keywords: Plant communities. Glacier retreat. Climate change.