Energy and Environmental Data Science

Figure 1. The out-of-sample ARCH/ARMAX model predictions with and without the CO2 estimated effects and the actual temperature outcomes at Barrow Alaska, Jan 1, 2016 – Aug 31, 2017.

 

ABOUT ENERGY AND ENVIRONMENT DATA SCIENCE  

The failure of the COP26 to resolve the shortfall in pledges to reduce greenhouse gas emissions is a grim reminder that there is a huge gap between the scientific consensus on climate change and political support needed to achieve the goals of the Paris Accords.  As if this issue is not enough, the recent uptick in electricity operational and balancing issues in Europe represents prima facie evidence that there is room for improvement in integrating wind and solar energies into the power grid. 

Energy and Environmental Data Science (EEDS) is an independent R&D entity dedicated to addressing these issues using data science with a particular focus on the application of time-series econometric methods.  These methods are invaluable in modeling high-frequency data when the variable of interest (e.g., hourly temperature) is highly autoregressive, highly volatile, and prone to giving rise to nonGaussian errors. Indicative of this potential, a recent EEDS preprint indicates that  CO2 concentrations are a useful input in predicting hourly temperature (Figure 1).