Precipitation Extremes Modeling in the Brazilian Northeast via Generalized Linear Models

The Brazilian Northeast (NEB), in recent years, has recorded a gradual increase in extreme precipitation events in a short period of time, which cause damage socio-economic factors. The intensity of these events vary in scale-space. The severity of these phenomena can be associated with human activities. The initial purpose of this paper is to explore various stochastic models in order to model the extreme precipitation via Generalized Linear Models (GLM), using eight meteorological variables and nine climate proxies. One has selected seven distinct points, characterizing the three existing sub-regions (East, North and South), to analyze the extreme, and was characterized as extreme precipitation values above the threshold of 30mm. For most points, the R2 values were fairly significant and “deviance” above 50%, especially in 22 and 48 points, with values above 80%.