Pereira, José Miguel CardosoNeves, Alana KasaharaFernandes, Bernardo Miguel Costa2025-02-072025-02-072024-03-25Fernandes, B.M.C. Análise da trajetória da vegetação pós-fogo utilizando séries temporais de imagens satélite. Lisboa: ISA, 2024, 68 p. Dissertação de Mestradohttp://hdl.handle.net/10400.5/98218Mestrado em Engenharia Florestal e dos Recursos Naturais / Instituto Superior de Agronomia. Universidade de LisboaDevido às alterações climáticas que se verificam atualmente, é possível observar-se um aumento na ocorrência de fogos de grandes dimensões, sendo este tipo de evento um dos principais responsáveis por alterações na cobertura e uso do solo. O ano de 2017 correspondeu, num passado recente, a um dos anos onde se observou uma maior área ardida no território português. Neste ano, arderam cerca de 555.222 ha, sendo que deste total, cerca de 46.586 ha correspondem à área ardida no incêndio de Pedrógão Grande, objeto de estudo desta dissertação. Na realização deste estudo pretende-se utilizar as séries temporais de imagens satélite Landsat para avaliar a recuperação da vegetação no período pós-fogo. Vão também ser analisados os fatores determinantes que afetam a recuperação da vegetação e como os padrões de recuperação pós-fogo se relacionam com estes fatores observados no pré-fogo e com a severidade do fogo. Na análise das séries temporais, será utilizado o algoritmo Continuous Change Detection and Classification (CCDC), com o suporte da plataforma Google Earth Engine (GEE), para obter informações sobre os valores de Greenness antes do fogo e em datas definidas após o fogo, assim como os valores de slope. Adicionalmente, foram extraídos diferentes dados de forma a permitir realizar uma análise do efeito de fatores determinantes na recuperação da vegetação. As diferentes classes de trajetória de recuperação foram obtidas através da classificação do tipo K-means. Os fatores determinantes a analisar neste trabalho são: a severidade do fogo, o uso e ocupação do solo, o número de anos desde o último fogo com uma área ardida considerável, o número de vezes que cada região da área de estudo ardeu no período temporal escolhido para análise e a topografia do local. Na realização deste estudo observou-se que ao longo do tempo a taxa de recuperação vai diminuindo e a densidade de vegetação vai aumentando. Foi também observado que a classe com menor severidade do fogo é constituída principalmente por florestas de eucalipto e que esta foi a primeira classe a obter valores de greenness da vegetação próximos dos observados no pré-fogo. Notou-se também que a ocorrência de fogos num passado recente terá um efeito positivo na severidade do fogo estudado, ao contrário da incidência de radiação solar que é responsável por se observar um aumento da severidade do fogo nestas regiões.Due to the current climate changes, it is possible to observe an increase in the occurrence of large-scale wildfires, with these types of events being one of the main causes of changes in land cover and land use. The year of 2017 corresponded, in a recent past, to one of the years in which a greater burned area was observed in Portuguese territory. In this year, the area burnt was around 555,222 ha, of which around 46,586 ha corresponded to the burned area in the Pedrógão Grande wildfire, the object of study in this dissertation. In this study, we intend to use time series of Landsat satellite images to assess vegetation recovery in the post-fire period. We will also analyze the driving factors that affect vegetation recovery and how post-fire recovery patterns relate to these driving factors observed pre-fire and to the fire severity. To analyze the time series, the Continuous Change Detection and Classification (CCDC) algorithm will be employed, supported by the Google Earth Engine (GEE) platform, which enables us to obtain information about Greenness values pre-fire and for the selected dates in post-fire as well as the values of slope. Additionally, different data were extracted to allow for an analysis of the effect of driving factors on vegetation recovery. The different recovery trajectory classes will be obtained through K-means classification. The driving factors to analyze in this study are: fire severity, the land use and occupation, the number of years since the last fire with a considerable burnt area, the number of times which region within the study area burnt in the time period selected and the topography. While carrying out this study it was observed that as time went by recovery rates would be slower and vegetation density would be higher. It was also observed that the class with the lowest fire severity is mainly made up of eucalyptus forests and that this was the first class to obtain greeness values close to the observed in the pre-fire. It was also noted that the occurrence of fires in the recent past will have a positive effect on the severity of the fire studied, unlike the incidence of solar radiation which is responsible for an increase in fire severity in these regions.porLandsatCCDCGEEK-meansGreennessAnálise da trajetória da vegetação pós-fogo utilizando séries temporais de imagens satélitemaster thesis203793714