On the morning of November 14th, 2000, areas of convective instability developed over the center of the state of Amazonas, Brazil. Some of those areas organized into a Mesoscale Convective System and moved eastward reaching Manaus, the state capital, around 1 PM. The deep convection was associated with large rain accumulations (over 74 mm), strong winds (18 m/s), intense electric activity, and caused a lot of destruction in the city: rooves were thorn out, there were floods, and more than 200 boats were overturned. Witnesses reported that when the roof tiles flew away, they seemed be sucked up and were lifted with a rotational movement.

In this paper we will present an observational description of this case, to gain understanding over the meteorological conditions that led to this event of severe weather. Several sources of data were gathered: infrared and visible GOES satellite images, model analysis, surface observations and sounding data, in addition to clippings of local and national newspapers.

The data shows that the main meteorological systems active on that day were: 1) The Bolivian high pressute system, an upper-level anti-cyclonic circulation associated with deep convection in the tropical South America, and 2) The South Atlantic Convergence Zone, (remanescent of a cold front that propagated equatorward) that organized a cloud band extending through central Brazil on the northwest-southeast direction, continuing over the Atlantic Ocean. Besides those synoptic scale systems, the early morning surface analysis showed a pre-existing wind convergence over Manaus, associated with a mesoscale region of strong temperature contrast (up to 11 C). The intensification of the convective system and the occurrence of severe weather happened when the system reached the convergence line.

We will perform further analysis of the factors described above to gain further understanding of the case. We will also explore the forecasts made with the Brazilian National Weather Service Operational Mesoscale Model over the region, since the model did not capture this event. The present study shows that understanding severe weather phenomena in the Amazon is very challenging, but also very necessary to improve model performance in the region.