Influence of Sea Surface Temperature in the Event of Squall Lines on the Northern Coast and Northeastern Brazil
Datasets usually provide raw data for analysis. This raw data often comes in spreadsheet form, but can be any collection of data, on which analysis can be performed.
Abstract This work aimed to verify the possible relationships of sampling points of the sea surface temperature spatially distributed in the tropical Atlantic and the development of squall lines in North northeast coast of Brazil using techniques of generalized linear modeling. For this, we used generalized linear models from the linear regression Poisson and negative binomial, for analysis of the relationship established by modeling was applied ANOVA variance test with significance level of 0.05 probability to determine which independent variables were more significant in modeling. Also the waste generated by the adjustment of the models in order to identify the distribution that best fitted the data were analyzed. All static analysis was performed in R. software Among the 132 pairs of observations, the magnitude of the correlation coefficient of Pearson (r) ranged from 0.06 (p < 0.04) between (LI and TSM5) and 0.88 (p < 0.0001) (TSM2 and TSM5), despite relations TSM and episodes of LI present no linear relationship or perfect positive linear relationship between the observations still rather have the significant statistical point of view. The results obtained by the negative binomial model had lower waste; however, the Poisson model was also significant from a statistical point of view. The points of extraction temperature values of sea surface (TSM2, TSM5 and TSM10) had the highest contributions in explaining the variability of episodes of instability lines in the north-northeastern Brazil.