This paper aims to characterize the thunderstorms observed in the Western Amazon region through the relation between cloud-to-ground lightning strokes with the time development stages of the convective systems observed during September 12th to October 10th of 2002 transition season (Dry-to-Wet) of RACCI/LBA project. In order to depict the main structures of the thunderstorms observed during the mentioned experiment, we used an algorithm that computes the morphology of the convective systems (CS), and it tracks them with time. This algorithm is known as FORTRACC and uses infrared images of GOES satellite. In addition to the life stage of the CS, we used lightning strokes measurements from the lightning detection network of TRMM/LBA cooperation project. The lightning frequency and its temporal evolution over CS might be used to evaluate the phase stage and process of development of these systems. The results computed until the present moment have shown that: a) Thunderstorms (CS with lightning) present a higher growth ratio (area expansion with time) during its initial development than storms without lightning; b) Only the large CS presented lightning; and c) Ssmall CSs are the ones that have higher lightning frequency per area, while the large CSs have the highest growth ratios. These results imply that a combination of updraft strength and convective area can determine whether a CS will become a large or small CS, therefore we might have a nowcasting tool.