A carregar...
Publicação
Framing the forest: Assessing carbon storage and sequestration using forest species extraction methods
| Nome: | Descrição: | Tamanho: | Formato: | |
|---|---|---|---|---|
| 8.17 MB | Adobe PDF |
Autores
Orientador(es)
Resumo(s)
As florestas são importantes ecossistemas que oferecem múltiplos serviços
de ecosistemas. Desta forma, classificar com precisão espécies florestais é
crucial para uma gestão e conservação eficaz das florestas. Este estudo
apresenta uma extensa análise dos ecossistemas florestais de Portugal.
Inicialmente, foi avaliado o desempenho de diferentes classificadores do
Google Earth Engine, utilizando vários algoritmos para extrair espécies e
grupos de espécies florestais. O estudo destacou a importância da seleção
de algoritmos, enfatizando a eficácia do classificador Support Vector
Machine. Posteriormente, foi realizada uma classificação a nível nacional
para os períodos de 2018 e 2022, revelando alterações significativas a nível
das áreas florestais. Portugal registou uma redução da área florestal total,
com uma diminuição de 131,940 hectares. Em 2018, a área florestal total era
de 2 711,395 hectares e em 2022 diminuiu para 2,579,455 hectares. Por
último, os efeitos espaciais do armazenamento e sequestro de carbono foram
investigados. Os resultados mostraram uma diminuição notável de 220 kg/ha
na captura de carbono a nível nacional, enquanto que ao nível do NUTS III,
as sub-regiões sequestraram uma média de 130 kg/ha. Este estudo contribui
com dados importantes para apoiar processos de tomada de decisão,
colmatando a lacuna entre a classificação de espécies florestais e a
avaliação precisa do seu carbono sequestrado, tendo implicações práticas
para a gestão florestal, a conservação da natureza e a avaliação de créditos
de carbono.
Forests are critical ecosystems providing multiple ecological services; therefore, accurate species classification is crucial for effective forest management and conservation. This study conducted a comprehensive analysis of Portugal's forest ecosystems. First, the performance of Google Earth Engine classifiers was assessed, employing various algorithms and input configurations to extract main forest species. The study highlighted the importance of algorithm selection, particularly emphasizing the efficacy of the Support Vector Machine classifier. Second, a national-level classification for 2018 to 2022 revealed significant changes in forested areas. Portugal experienced a considerable reduction in total forested area, decreasing by 131,940 hectares. In 2018, the total forested area with the selected species and species groups was 2,711,395 hectares. By 2022, it had diminished to 2,579,455 hectares. The spatial effects of carbon storage and sequestration within these forests were investigated in the third phase. Results showed a notable decrease of 220 kg/ha in carbon capture at the national level. At NUTS III, subregions sequestered an average of 130 kg/ha. This research contributes valuable data for informed decision-making processes, bridging the gap between accurate species classification and precise carbon assessment, having practical implications for forest management strategies, conservation efforts, and carbon credit evaluations.
Forests are critical ecosystems providing multiple ecological services; therefore, accurate species classification is crucial for effective forest management and conservation. This study conducted a comprehensive analysis of Portugal's forest ecosystems. First, the performance of Google Earth Engine classifiers was assessed, employing various algorithms and input configurations to extract main forest species. The study highlighted the importance of algorithm selection, particularly emphasizing the efficacy of the Support Vector Machine classifier. Second, a national-level classification for 2018 to 2022 revealed significant changes in forested areas. Portugal experienced a considerable reduction in total forested area, decreasing by 131,940 hectares. In 2018, the total forested area with the selected species and species groups was 2,711,395 hectares. By 2022, it had diminished to 2,579,455 hectares. The spatial effects of carbon storage and sequestration within these forests were investigated in the third phase. Results showed a notable decrease of 220 kg/ha in carbon capture at the national level. At NUTS III, subregions sequestered an average of 130 kg/ha. This research contributes valuable data for informed decision-making processes, bridging the gap between accurate species classification and precise carbon assessment, having practical implications for forest management strategies, conservation efforts, and carbon credit evaluations.
Descrição
Dissertation presented as the partial requirement for obtaining a Master's degree in Geographic Information Systems and Science, specialization in Geographic Information Systems and Science
This research was part of the EXPL/CTA-AMB/0165/2021 project, supported by a grant of the Portuguese Foundation for Science and Technology ("Fundação para a Ciência e a Tecnologia"): https://doi.org/10.54499/EXPL/CTA-AMB/0165/2021
This research was part of the EXPL/CTA-AMB/0165/2021 project, supported by a grant of the Portuguese Foundation for Science and Technology ("Fundação para a Ciência e a Tecnologia"): https://doi.org/10.54499/EXPL/CTA-AMB/0165/2021
Palavras-chave
Observação da Terra Machine-Learning Espécies florestais Mapeamento de vegetação Sequestro de carbono Serviços de ecossistemas Earth Observation Machine-Learning Forest Species Vegetation Mapping Carbon Sequestration Ecosystem Services SDG 12 - Responsible Consumption and Production SDG 13 - Climate Action SDG 15 - Life on Land SDG 12 - Consumo e produção responsáveis SDG 13 - Ação contra a mudança global do clima SDG 15 - Vida terrestre