Análise da cobertura de nuvens no nordeste do Brasil e seus impactos no sensoriamento remoto agrícola operacional
For the agricultural sector, the monitoring of crops throughout their development stages is required for many applications, such as the evaluation of their phytosanitary condition and production/yield. In the analysis of optical remote sensing data, cloudiness directly affects the quantity and quality of the images. Clouds or their shadows can cover the surface. In addition, there is also residual scattering effects that can alter the spectral response of the surface detected by the satellite sensors. To address this challenge, this study aims to characterize the cloud cover of the Northeast of Brazil along the year, as well as to discuss the impacts on the agricultural monitoring of the main producing areas of the region. The study area was divided into 6 homogeneous sub-regions in terms of agricultural production and production potential. To calculate the cloud cover, the MOD35 product from the Moderate Resolution Imaging Spectroradiometer (MODIS) sensor was used. The analysis was conducted, between July 2000 and June 2016. The resulting images were grouped according to the month of acquisition. For each pixel, the mean and variance of the occurrence of clouds were calculated using the homogeneous distribution of Bernoulli. Based on the set of 12 maps with monthly averages of cloud covering, it was noticed that some regions along the coast and in the northwest of Maranhão State had 80% of cloud cover in great part of the year. Despite the several months without rainfall, cloud cover in the semiarid sub-region was above 40% in the dry season. Finally, in the sub-region known as "Mapitoba", cloud cover was reduced with seasonality to values below 20% in August. In all Northeast of Brazil, cloud cover is greater than 70% between January and April