A Statistical Analysis and Exploration on Atlantic Hurricanes Database (HURDAT2); Do We Expect the Worst?

Hurricane occurrence exploration is a heavy task that requires sophisticated methods to accurately determine the occurrence and the impact of damage on the tropical environs. To control massive damage, several techniques have been developed to measure the accuracy of track forecasting. The accurate analysis and exploration of the occurrence of hurricanes are very crucial for the affected environs evaluation within a short time to minimize the loss of human life and property. Nevertheless, the exact analysis and exploration of these hurricanes is challenging and time-consuming. Therefore, this study proposed a statistical model whereby the analysis of variance (ANOVA) a linear regression scientific data statistical analysis model is applied on the dataset of Atlantic hurricane database (HURDAT2) for the exploration of hurricane occurrence and damage in the tropical environs. The best tracks of a 6 hour interval, location (latitude, and longitude), wind speed, central pressure of all identified hurricanes and subtropical typhoons from the year 2008 to 2017 (10 year period) are used to determine the R2 coefficient, which measures the goodness and fitness of the model to reveal the variability of the real data. Statistical significance and reliability of the data are tested on significance P-value <0.05 where four different parameters are considered for the analysis in order to determine the destructive of these hurricanes in the tropical environs; latitude (x1), longitude (x2), the wind speed (x3), and central pressure. The results of our model proved significant with an accuracy of 99.3%, and a mean standard error (MSE) of 1.4952 for all the hurricanes that were analyzed, and the year 2012 was established as a year that had much damage.