This project advances EPRI modeling capabilities and methods for estimating the potential impact of changes in air temperature on energy demand for space conditioning in the United States.
Specifically, we demonstrate a structural approach within EPRI’s US-REGEN end-use demand module, which projects hourly energy demand from the bottom-up for several key end-use sectors. Electricity demand for heating and cooling is calculated using observed hourly air temperature in a building energy model emulation that projects the evolution of space conditioning technology and building stocks with exogenous assumptions about technological change, fuel prices, and socioeconomic growth. We begin by modifying the model such that the temperature parameter is indexed over time, and the distribution shifts across time periods based on the assumed climate scenario. The reference case assumes no climate change and stationary weather, accounting only for non-climatic factors affecting space conditioning demand in 2050. The baseline scenario is called ‘trend’, which extends the observed annual trend in the 40-year historical MERRA sample for each location (collection of grid cells associated with a particular heating/cooling zone within a model region). Next, we introduce a set of ten CMIP6 scenarios that are quantile-based temperature deltas for each grid cell based on 5 global climate model projections for two CMIP6 scenarios, SSP1-2.6 and SSP3-7.0. For all of these scenarios we run the end-use demand module in order to assess the impacts of higher air temperatures on energy demand across the U.S. Specifically, we develop temperature-adjusted annual load projections for future space conditioning demand under future conditions.
In the reference scenario without warming, we find that annual cooling energy requirement declines steadily between 2015 and 2050, as efficiency improvements in cooling technology and building stock efficiency are projected to far outweigh growth in cooled floorspace. For heating, we find that less final energy is required to meet growing service demand due to structural change and efficiency improvements, though electricity’s share increases relative to other fuels, and this electrification of the heating sector leads to strong growth in total and peak heating loads. The effect of warming works in the opposite direction of both mechanisms (i.e., cooling-efficiency and heating-electrification). For cooling, projected warming under SSP3-7.0 in 2050 negates the expected efficiency gains in terms of annual energy. Under the warmest climate realization, SSP3-7.0 in the UKESM1-0-LL model, residential electricity demand for cooling in 2050 is 30% greater than the baseline trend projection. This is evident across all regions, though the effect is smallest in Florida and Texas, and largest in New England and NE-Central. The potential for electrification of building heating to drive up winter electricity consumption is moderated by all warming scenarios. Moreover, even the baseline trend of warming causes final energy (electric and non-electric) demand to drop by a quarter across the US. Because of the relative size of heating versus cooling energy in most regions (Florida and Texas being exceptions), this avoided heating demand generally outweighs the additional energy for cooling.
This bottom-up approach complements the existing climate impacts literature, as many studies have relied on empirical models that estimate the statistical relationship between weather and electricity use (e.g., see reviews in Dell et al, 2014 and Auffhammer et al, 2017) – our results from a structural basis offer a point of comparison to previous estimates. Furthermore, this project helps to produce temperature-adjusted load projections that can be utilized by the broader community of energy-economy models to better assess the electric system’s vulnerability to and plan for different climate conditions. Ultimately, we will use the US-REGEN capacity expansion and dispatch model to assess these demand-side impacts on the electricity system, including generation and capacity decisions, supply cost, and emissions over time. Integrating climate warming into the US-REGEN end-use model can be applied in future studies to inform system planners and other stakeholders about electric power systems that are resilient to a range of possible climate, policy, and technology futures.
Authors Delavane Diaz and Geoffrey J. Blanford