Análise da disponibilidade de imagens Landsat e Sentinel para o Brasil

Abstract

This paper analyzed the availability of cloud-free data from the programs Landsat (surface reflectance, 1982- 2019) and Sentinel-2 (TOA reflectance, 2015-2019) over the Brazilian territory. In the case of Landsat, the amount of information available increased considerably in 1999 with the start of Landsat 7. However, mainly due to the presence of clouds, the availability of data varies greatly in space and time. The Amazon biome, in particular, has scarcity of data with an average of 0.72 valid observations per month and five months with less than 0.4 valid observations (December to April). The Caatinga and Mata Altântica biomes also present, to a lesser extent, little data (0.96 and 1.07 valid observations per month, on average). On the contrary, other biomes, such as the pampa, present a significant number of data (1.44 valid observations per month on average for pampa) more evenly distributed over the year. Due to its better temporal resolution, Sentinel-2 is able to reach a larger number of valid observations per month (about 3 for the Amazon and 4 for the pampa). However, the Sentinel-2 satellite constellation became fully operational in 2018 and, for studies of historical periods, Landsat, eventually combined with other sensors, such as CBERS or SPOT, remains the basis of many studies