Uma ferramenta para análise automática dos Complexos Convectivos de mesoescala no Nordeste do Brasil

Abstract

The Mesoscale Convective Complexes is one of the active systems in Brazilian Northeast (BNE), which causes great economic and social losses. MCC is associated with adverse phenomena such as heavy precipitations, thunderstorms and wind gusts. An automatic method of MCC analysis and detection in the BNE not previously been developed. The aim of this work was create an automatic algorithm of MCC identification in BNE to help in study and forecasting of these systems. The algorithm was created using the GNU Octave programming language. Study area was determined as 0° - 20° S and 30° - 50° W. The algorithm processed satellite images of cloud tops temperature in 21 MCC events. Physical MCC characteristics, such as area, eccentricity and duration, were calculated. These characteristics were compared with previous studies and the results appearing to be consistent. Difference of MCC duration by automatic algorithm and previous study is equal to 1.7 h. Average MCC areas by algorithm and by previous study are around 206,000 km² and 188,000 km², respectively. Sixty-two percent of MCC have been identified. Some differences occurred due to higher frequency of used images. This automatic algorithm can calculate some new characteristics, such as, trajectory and MCC velocity.