Validating locational marginal emissions models with wind generation

Abstract

Abstract Increasingly large amounts of electric supply and load are being deliberately operated or sited on the basis of marginal emissions factor (MEF) models. Validating and calibrating such models is therefore of growing policy importance. This paper uses a natural experiment involving variation in relative changes in wind generation potential at wind farms in the ERCOT power grid to create a benchmark MEF and examine the relative accuracy of several common classes of short term MEF models. This work focuses on MEFs at the level of a few individual generating nodes, a much smaller geographic scale than the Balancing Authority (BA) or load zone. Additionally, the use of wind generation potential as a regressor allows us to factor in wind curtailment, in contrast to previous work. We evaluate multiple prevalent existing MEF models and find that both dispatch and statistical MEF models have a high degree of agreement with the benchmark MEF, while heat rate and average emissions do not. We also find that the emissions reduction benefits of optimizing electricity with MEFs using a geographically granular model instead of a BA-wide model are 1.4, 1.3 and 1.5 times larger for dispatch, statistical and heat rate models, respectively.