The main purpose of this study is to evaluate a technique for filtering out clouds in order to obtain vertical temperature and moisture profiles over Brazil using Advanced TIROS Operational Vertical Sounder (ATOVS). This evaluation consists of a sequence of tests using Advanced Very High Resolution Radiometer (AVHRR) infrared images and the Mask AVHRR for Inversion ATOVS (MAIA) algorithm. Two distinct versions of this algorithm were used, namely, forecast (F) and climatology (C). Precipitable water (PW) and surface temperature data as forecast by a global numerical weather prediction model were used in the former case, while a climatological atlas of PW was used in the other. In both cases PW was estimated using Advanced Microwave Sensor Unit (AMSU) data over ocean areas. The results show that MAIA does evidence coherence between the areas classified as cloudy and those ones sorted as cloudy according to a visual (subjective) comparison of images. The two versions of the cloud mask algorithm yielded similar results mainly over the southwest Atlantic Ocean. However, over continental areas it was observed a small increase in areas given as clear when the F version was used. Furthermore, it was also observed that the cloud mask is highly sensitive to changes in the threshold values based on the brightness temperature standard deviation in oceanic areas.
satellite sounding; cloud mask; infrared
