Locational Marginal Emissions for Carbon-Aware Data Center Operations in Large-Scale Power Grids

TL;DR

This paper investigates the use of Locational Marginal Emissions (LMEs) to improve carbon accounting and reduction strategies for data centers within large-scale power grids, revealing regional differences and effective intervention methods.

Contribution

It provides a comprehensive analysis of LME properties in a large power grid, demonstrating their application in guiding data center emission reduction with high accuracy.

Findings

LMEs vary significantly across regions, affecting emission reduction strategies.

LME-guided interventions achieve over 85% accuracy in emission reduction.

Large-scale grid simulation is essential for accurate evaluation.

Abstract

Carbon accounting methods for electricity consumption face challenges regarding physical deliverability, double counting, additionality, and impact magnitude. Locational Marginal Emissions (LMEs) show potential to address many of these key issues. However, their use in a large-scale power grids remains understudied. We analyze the properties of LMEs from a data center's perspective in a 1493-bus Western Interconnection over one year of hourly operation. We find that LME characteristics create three distinct regions: the hydropower-dominated Pacific Northwest, with low and stable LMEs; the coal-heavy Intermountain West, containing often high LMEs; and the Sunbelt, where mixed generation leads to variable LMEs correlated with solar output. This characterization provides analytical guidance for data center emission reduction. In particular, LME-guided emission reduction interventions through data center temporal-spatial load shifting, siting, and renewable procurement display over 85% accuracy with respect to actual emission reduction. Moreover, large-scale, nodal grid simulation is shown to be critical to accurate evaluation.