Modelagem espectral para determinação de fluxo de CO2 em áreas de caatinga preservada e em regeneração

Abstract

There is a great need for the development of more efficient systems for monitoring the dynamics of atmospheric carbon so that there is a better understanding of the interactions between the biosphere and the atmosphere. However, it is critical that these technologies have high coverage and low cost. In the vegetation, the process of biomass generation through photosynthesis is a determining factor in the way a vegetated area will appear radiometrically in the satellite images, therefore, the remote sensing becomes an alternative to the monitoring of this dynamics, having a High coverage and have a low cost. The objective of this work is to analyze the dynamics of CO2 fluxes in the Caatinga Biome by means of multispectral remote sensing, verifying the potential of multispectral images in the detection of CO2 fluxes in areas of preserved Caatinga and in a regenerated state. The study was carried out in areas of Caatinga in the municipality of Petrolina-PE and Araripina-PE, areas that are monitored by micrometeorological stations. The methodology adopted was based on the modeling of the Carbon Forest Sequestration Index (CO2flux) that measures the efficiency of the carbon sequestration process by vegetation, and proposes the integration of the Normalized Difference Vegetation Index (NDVI) with the Index of Photochemical Reflectance (PRI). For the database, 22 OLI (Landsat-8) multispectral scenes were used together with field-measured meteorological data to verify the relationship between the variables analyzed. The results showed that the vegetation of the Caatinga has seasonal variations in CO2 flows in regions with different vegetation types. The CO2flux index can be applied to determine the CO2 fluxes, presenting better adjustments when the CO2 data are compared using the 1 pixel reading in the image, showing to be more efficient in relation to the analysis with the footprint in. However, based on spectral models of vegetation, it was possible to determine the dynamics of CO2 flows in areas of preserved Caatinga and in regeneration state using data extracted from multispectral sensors.