A Exploratory data analysis to understand the causes of global warming and application of soft computing techniques to develop its forecasting model

Abstract

Global warming (GW) is one of the major effects of human activity where excessive use of fossil fuels as energy sources has led to an increase in the concentration of greenhouse gases (GHGs), such as CO2, CH4, and water vapour, in the atmosphere one of the main reason to increase the average surface temperature. This study analyzes the time-series data to come to a rational conclusion about the role of GW in increasing sea-water level, the reason for the increase in GHG and the correlation of GHG to GW. In this direction time-series analysis is carried out on four different datasets. The first and second dataset comprises global temperature anomalies data and the cumulative changes in seawater level for the world’s oceans since 1880. The third and fourth dataset comprises the records of concentration of GHGs in the atmosphere since 1st AD and the last 4 ice age years respectively. Finally, forecasting models are developed based on Holt’s and SARIMA techniques to predict the global temperature anomaly, the concentration of GHGs and their correlation with GW. The developed models showed 74.6%, 94.5% and 95.7% accuracy in predicting temperature anomaly, CO2, and CH4 concentration in the atmosphere respectively. The strength of the forecasting model is its ability to compute the critical values of the factors. Therefore, the forecasting models are applied to predict the year in which the critical values of the factors contributing to GW will be attained.