| Nome: | Descrição: | Tamanho: | Formato: | |
|---|---|---|---|---|
| 4.25 MB | Adobe PDF |
Autores
Orientador(es)
Resumo(s)
Os incêndios florestais são considerados pela proteção civil como um dos desastres ambientais mais graves existentes em Portugal devido à sua frequência e extensão. Para fazer face ao problema existe o Sistema de Defesa da Floresta Contra Incêndios, que estabelece a implementação de faixas de gestão de combustíveis como forma de auxiliar o combate aos incêndios.
A proposta desenvolvida neste trabalho foi elaborar um modelo capaz de identificar a remoção da vegetação nas faixas de gestão de combustível no entorno de aglomerados populacionais através da análise de séries temporais de imagens de satélite Sentinel 2. O modelo utilizado baseia-se na busca de diferenças estatisticamente significantes entre séries temporais de um máximo de 60 dias antes e depois da data analisada. A análise utiliza três informações da série temporal: (i) o valor do Normalized Difference Vegetation Index (NDVI) em cada pixel dentro da área de estudo; (ii) o valor da média do NDVI dos pixeis vizinhos com um raio de 500 metros, fora da área de estudo e em um mesmo uso do solo; e (iii) a diferença entre estes valores. Posteriormente a cada data da série temporal é aplicado o Welch t-test. A identificação de uma alteração é dada quando o t-test aponta a diferença significativa para o valor do ‘pixel (i)’ e da ‘diferença (iii)’ e não na ‘média (ii)’ em uma mesma data.
O modelo foi aplicado no entorno das áreas urbanas de cinco municípios de Portugal continental. Os resultados mostraram-se promissores para identificar os locais onde não foi feita a gestão das faixas de combustível, ou seja, áreas de infração a legislação, indicando que o modelo pode ser utilizado para verificar se houve ou não a gestão anual das faixas de combustível previstas na lei. O resultado obtido demonstra a potencialidade do método utilizado em servir como meio de suporte às ações de prevenção e fiscalização, além disso pode ser expandido para todo o território nacional.
Forest fires are considered by civil protection as one of the most serious environmental disasters in Portugal due to both their frequency and extent. To address the problem, there is the Sistema de Defesa da Floresta Contra Incêndios, which establishes the implementation of fuel management bands as a way to aid firefighting. The proposal developed in this study was to create a model capable of identifying the removal of vegetation in the fuel management bands around population clusters by analysing time series of satellite images Sentinel 2. The model used is based on the search for statistically significant differences between time series of a maximum of 60 days before and after the analysed date. The analysis uses three time series information: (I) the value of the Normalized Difference Vegetation Index (NDVI) in each pixel within the study area; (II) the NDVI mean value of neighbourhood pixels with a radius of 500 metres outside the study area and in the same land use; and (iii) the difference between these values. After each date of the time series, the Welch t-test is applied. The identification of a change is given when the t-test points to the significant difference for the value of ‘pixel (i)’ and ‘difference (iii)’ and not in ‘average (ii)’ on the same date. The model was applied around the urban areas of five municipalities of mainland Portugal. The results were promising to identify the places where the management of fuel bands was not performed, i.e., areas of infringement of the legislation, showing that the model can be used to verify whether there was the annual management of the fuel management bands provided for in the law. The result obtained demonstrates the potentiality of the method used to serve as a means of supporting preventive and enforcement measures and may also be expanded throughout the national territory.
Forest fires are considered by civil protection as one of the most serious environmental disasters in Portugal due to both their frequency and extent. To address the problem, there is the Sistema de Defesa da Floresta Contra Incêndios, which establishes the implementation of fuel management bands as a way to aid firefighting. The proposal developed in this study was to create a model capable of identifying the removal of vegetation in the fuel management bands around population clusters by analysing time series of satellite images Sentinel 2. The model used is based on the search for statistically significant differences between time series of a maximum of 60 days before and after the analysed date. The analysis uses three time series information: (I) the value of the Normalized Difference Vegetation Index (NDVI) in each pixel within the study area; (II) the NDVI mean value of neighbourhood pixels with a radius of 500 metres outside the study area and in the same land use; and (iii) the difference between these values. After each date of the time series, the Welch t-test is applied. The identification of a change is given when the t-test points to the significant difference for the value of ‘pixel (i)’ and ‘difference (iii)’ and not in ‘average (ii)’ on the same date. The model was applied around the urban areas of five municipalities of mainland Portugal. The results were promising to identify the places where the management of fuel bands was not performed, i.e., areas of infringement of the legislation, showing that the model can be used to verify whether there was the annual management of the fuel management bands provided for in the law. The result obtained demonstrates the potentiality of the method used to serve as a means of supporting preventive and enforcement measures and may also be expanded throughout the national territory.
Descrição
Palavras-chave
Sentinel 2 Séries temporais NDVI Alteração de Vegetação Gestão de Combustível
