Assessment of spatial and temporal distribution of clouds in Northeast Brazil (2000-2019) using MODIS data

Abstract

The present study aims to assess the spatial and temporal distribution of cloud cover in Northeast region of Brazil, monthly and annually, from 2000 to 2019, in order to determine if there is a spatial and temporal pattern in the distribution of cloud cover and if there is a pattern between the distribution of cloud cover in relation to the amount of rainfall. For this, daily images for the analyzed period by the MODIS sensor were used, which include the quality assurance (QA) of the data produced (clear pixel, pixel with total cloud cover, or pixel mixed with clouds). For comparative purposes, QA data from OLI (Landsat-8) and MSI (Sentinel-2) sensors for 2019 were used. Besides that, rainfall data from TRMM satellite were used considering the Amazon, Atlantic Forest, Caatinga and Cerrado biomes. The results made possible to know the spatial and temporal distribution of the cloud free images and those with presence of cloud. In the Northeast region, the Amazon and Atlantic Forest biomes are strongly affected by the presence of clouds, while the Caatinga and Cerrado biomes are less affected. Annual observations did not indicate clear evidence about the reduction in the data quality in the wettest years. While monthly observations indicated that in the wettest months there was a ∼50% reduction of cloud free images. So, the most favorable months for data acquisition by optical sensors are observed in the dry period, specifically between May and October, mainly in the Cerrado biome. The results point to the need to increase the availability of temporal orbital images in the region, especially in the Amazon and Atlantic Forest biomes, as it becomes a limiting factor for land use and land monitoring in that region mainly to the lower temporal resolution of the used sensor.