Forecasting Carbon Dioxide Levels Using Autoregressive Integrated Moving Average Model

Abstract

In the last few decades, the forecasting of lower atmospheric carbon dioxide (CO2) levels has been emphasized as an important topic among the atmospheric scientists and engineers for developing better predictive models in to predict the levels of CO2 keeping eye on the accelerated pollution. In the present work, we exploited the autoregressive integrated moving average (ARIMA) model capability for time-series prediction of CO2 level by considering the long-term recordings of air sample at Mauna Loa lab Observatory in Hawaii, USA, during the period from March 1958 to December 2001. The results reveal that forecasting of the parameter through ARIMA model has significantly improvements as compared to the existing techniques for such lower atmospheric parameters.