Forecasting Atmospheric CO2 Concentrations Using SARIMA Models: Insights from Time-Series Analytics

This record contains the article "Forecasting Atmospheric CO2 Concentrations Using SARIMA Models: Insights from Time-Series Analytics" by Hemanth Kori. The paper analyses weekly CO2 concentrations at Mauna Loa Observatory (1958-2001) using a seasonal ARIMA model. The time series is decomposed into trend, seasonal and residual components, and the model is trained on 1958-1997 data and tested on 1998-2001. Forecasts achieve mean absolute error of about 1.07 ppm and RMSE of about 1.20 ppm, and two-year projections suggest levels above 370 ppm by the end of 2003. The study discusses the relevance of classical time-series methods amid emerging AI-driven analytics. Figures and code for the analysis are included.