Forecast of Earthquakes Greater than 6 on the Richter Scale in Japan Using ROR Modeling
Ricardo Oses Rodríguez and Rigoberto Fimia Duarte
ABSTRACT
Knowing in advance the occurrence of a high intensity earthquake is an event that is predictable with a certain degree of error using the Regressive Objective Regression (ROR) methodology, this allows environmental agencies in charge of people's safety to take measures to avoid further damage to life and buildings. To carry out this work, we had a database of the largest earthquakes that occurred in Japan from March 2011 to 2024, extracted from www.nippon.com. First, we modeled the year, month, day and time in which these phenomena occurred, as well as their magnitude. We could not predict the location since the database we have only shows the name of the place and we do not have the latitude and longitude where it really is. These phenomena occurred, then a short-term forecast was made until the year 2025 of the characteristics of the predicted earthquake. The models explain more than 91% of the variance with small errors, the trend for the month is negative which means that earthquakes are more likely to occur in the first months of the year, the trend for the year is to increase logically, for the day, the hour and the magnitude the trend is not significant, regressive patterns of 1,5,8.9,11,12 and 15 steps back are shown for all variables depending on the variable to be modeled.
