Amazonian forests have exceptional biodiversity with the highest species richness on Earth, providing vital ecosystem services that regulate carbon and hydrological cycles both regionally and globally. Despite rainforests being a naturally fire-free system, increasing evidence has shown that fires existed in Amazonian forests before European colonization, where its ignition depended on a combination of drought and human activity. Nowadays, anthropogenic actions, such as land-use and land-cover changes, associated to global climate change, increasingly transform these forests into a more fire-prone environment. Fire brings several impacts to tropical forests, transforming these forests into a carbon source, altering forest dynamics, microclimate and forest structure. Despite studies on the impacts of fire on carbon dynamics in Amazonian forests, there is still a knowledge gap in how historical fires impact the current forest dynamics, especially over increased frequency of droughts, and how modern fires affects the vertical canopy structure of primary and secondary forests and their ability to recover from fires. The aim of this thesis is to investigate the impacts of historical and recent fires on current carbon dynamics and forest structure. In chapter 2, I investigate the effects of historical fires on the current response of forests to drought. For this, I used soil pyrogenic carbon (PyC) as a proxy of historical fires and field-based biomass estimates across the Amazon Basin spanning drought and non-drought years. My results show a strong positive correlation between soil PyC and soil fertility, clay and silt, and a negative correlation between soil PyC and wood density and sand. Furthermore, I found that forests with low concentrations of soil PyC were more impacted by drought. These findings support the hypothesis that soil PyC increases soil fertility and soil water holding capacity, affording higher resistance to drought, whilst also favouring the establishment of species associated with historical disturbances such as fire and drought. In chapters 3 and 4, I focus on the impacts of recent fires on primary and secondary forests, respectively. Chapter 3 investigates the effects of fire and fire reoccurrences on the canopy structure of primary forests. I used a range of forest structure attributes from airborne lidar data across the Brazilian Amazon. My findings show that forests that experienced repeated fires experience greater changes after fire and need longer to recover. In chapter 4, I used lidar data to analyse the impacts of fire on the forest structure of secondary forests. The results show that fires negatively affect canopy structure of secondary forest in early and later successional stages, however, forests in later successional stage have lower potential to recover forest structure after fire than early successional stages. Overall, the results of this thesis show that the impacts of fire on Amazonian forests affect carbon dynamics and storage, as well as altering forest structure and many related ecosystem services. Impacts caused by fire can be irreversible or may take many decades to fully recover, leaving traces behind after burns which happened centuries ago. My results indicate that forest conservation and management policies should be implemented to avoid fires and protect the long-term future of Amazonian forests.