Ricardo Oses Rodriguez , Enrique Arango and Rigoberto Fimia Duarte
The objective of our work is to model, with the help of the ROR methodology, the earthquakes recorded in Cuba on the Richter scale in order to predict upcoming events, establish if there is any trend in the latitude and longitude where they occur, and see which are the main ones. To carry out this work, we had a database of earthquakes of all magnitudes on the Richter scale taken from 1965 to 2023, a total of 100,736 cases. These data are in the hands of the Cuban Seismology Center, CENAIS., there are no missing data in the sample. Latitude, longitude and magnitude were modeled. For the latitude, earthquakes were reported approximately close to 16 N, then the records rose to 18 N, until 2019, month 3, from now on the latitude rises to approximately 19 N, so the trend over time is to increase. The average values for this variable are 19.9 N. For longitude it occurs similar to latitude, there is a tendency to increase, values range from – 67 W to -74 W approximately on the same dates as latitude. The average values are -75.8 W. The latitude model explains 99.9% of the variance with an error of 0.651443 degrees, the trend is increasing. For longitude, perfect models are obtained, the trend for longitude is to increase as well. The error is 1.67 degrees. For both latitude, longitude and magnitude, good forecasts are presented with errors close to zero. The long-term magnitude model depends on 73 and 74 steps back and the trend is negative which means that smaller magnitudes should occur over time. Observing the small errors and the high explained variance of the models of the variables studied, we can assert that it is advisable to establish a short-medium and long- term forecasting system for the earthquakes recorded using the ROR methodology.