Chapter 10 - AI for monitoring power plant emissions from space

Abstract

Effective and precise monitoring is required to estimate human emissions and predict climate impacts accurately. This chapter develops a machine learning model trained on satellite observations (Sentinel 5), ground observed data (EPA eGRID), and meteorological observations (MERRA) to predict the NO2 output of coal-fired power plants. Overfitting and generalization pose numerous challenges in developing a consistently accurate model based solely on remote sensing data, and these challenges are addressed in this chapter using a combination of preprocessing, hyperparameter tuning, and feature engineering techniques.