Estimativa do fluxo de CO2 na camada limite da atmosfera sobre a RDS Uatumã/AM-Brasil por meio de imagens orbitais e dados micrometeorológicos

Abstract

In the study field of forest ecosystems, tropical forests hold a constant dynamic in relation to the atmosphere and the balances of water and energy, possessing a vast biodiversity. As the main range of tropical rainforest in the world, the Amazon Rainforest concentrates about a third of the total area of tropical rainforests in the world, working as an important climate regulator, storing and dispersing carbon flows. The Amazon Forest captures and releases large amounts of CO2 through photosynthesis and the respiration of plants and soil, acting fundamentally in the formation of carbon stocks on the planet. The net exchange embedded in the CO2 flows, as well as the land policies, dynamically affect the planet's climate, representing a central relation for a critical approach in the field of environmental science. Therefore, there is an objective relation between the rise of carbon dioxide's (CO2) concentration on the planet and the growth of human activities, as evidenced by Wohlfahrt et al. (2010) and Ramankutty et al. (2007). In order to widely understand this relation and how it influences the climate change, it is essential to hold on reliable and accessible data, as an effort to estimate the variables which affect the climate. In this context, the present research aims to estimate the carbon flux in the internal area and around the Uatumã RDS reserve (Amazonas / Brazil), using digital images from the CBERS-4 satellite (MUX sensor) and micro-meteorological data from the Amazon's High Observation Tower (ATTO). To achieve this, a carbon flow model was elaborated for each year (2017, 2018 and 2019), based on micro-meteorological data and orbital images from this region. In order to generate the models with micrometeorological data, multiple linear regression was applied on the original data for each year(2017, 2018 and 2019) (R² = 0.271; 0.278; 0.316) and on the data altered by Principal Components (R² = 0.237; 0.246; 0.288), finally transformed by polynomial (R² = 0.334; 0.318; 0.339). Then, the spectral indices for each year of NDVI (R² = 0.2596; 0.0004; 0.1197), PRI (R² = 0.4749; 0.0549; 0.5218) the proposed method by Rahman et al. (2000) (R² = 0.2998; 0.0011; 0.158) and the model proposed by this research (R² = 0.6251; 0.0083; 0.5896). The results of this research indicate that the spectral index proposed in this study provides an indirect estimation method, to measure the carbon flow in the Amazon, with a higher level of confidence in relation to the other models generated from micrometeorological data and the other spectral indexes presented.