To determine the effects of past climate trends on global energy consumption, the econometric equations providing the degree day elasticities reported previously are combined into an econometric simulation model. The endogenous variables determined by the model include energy demand in the residential, commercial, and industrial sectors of the U.S. economy and the derived demand for primary fuels used in electric power generation. The demand and supply of electricity are determined in the model. The exogenous or predetermined variables include real personal disposable income, retail sales, industrial production, energy prices, and heating and cooling degree days. Like the degree day elasticities reported previously, the equations provide estimates of price and income elasticities of demand, which were reported by Considine in 2000. Hence, the model provides a tool for estimating the contributions of each determinant of energy demand, including climate, prices, and income.
To identify the effects of past climate trends on historical energy consumption, two simulations are required. The first simulation provides a baseline using the 30-year means for heating and cooling degree days, which gives an estimate of what energy demand would have been under average climate conditions. The second simulation uses actual degree days, which yields an estimate of predicted energy demand associated with actual weather. All other exogenous variables, such as energy prices and income, are the same in the two simulations. Consequently, the changes from the base simulation represent those changes in energy demands associated with deviations of weather conditions from their 30-year means.
As discussed previously, the climate in North America has been getting warmer since the early 1980s. Indeed, cumulative cooling degree days are 2.2% higher than normal and heating degree days are 1.65% lower. The impacts of these changes in temperature are illustrated in which presents the percentage changes in energy demand from the predictions using average and actual degree days during the cooling and heating seasons.
There is an unambiguous, positive relationship between heating degree days and energy demand. With warmer winters, the demand for energy declines. Notice the string of warm winters during the late 1980s and early 1990s and the associated reductions in energy consumption. For example, the winters of 1990–1991 and 1991–1992 were approximately 8% warmer than average. As a result, the consumption of natural gas was approximately 3% lower than it would have been under normal weather conditions. Distillate and residual fuel oil use was almost 4% and more than 6% lower than normal, respectively.
The effects of cooling degree days are more ambiguous because some months of the cooling season also contain heating degree days, which may offset the effects associated with cooling degree days. Nevertheless, there are several years with warmer than normal summers and slightly higher energy use. During the entire period, energy demand under the base simulation with actual weather is 0.2% lower than that under the alternative simulation using the 30-year mean. This suggests that a warmer climate would slightly lower energy consumption because lower fuel use associated with reduced heating requirements offsets higher fuel consumption to meet increased cooling demands.
Simulating the effects of a 3-month shock to heating and cooling degree days provides a more controlled experiment. First, consider the effects of a colder than normal winter with heating degree days 10% more than the 30-year mean for 3 months. All fuel demands increase (Fig. 3). Higher residential and commercial consumption is the principal reason for the approximately 7–10% increase in natural gas consumption. In addition, electric utility fuel use, particularly oil, increases due to the simulated 2–4% increase in electricity consumption. The large estimated elasticity of electric utility oil use with respect to electricity generation accounts for most of the increase in residual fuel consumption. This result may reflect the practice of electric utilities to use oil-fired generators to service peak power demands.
The last simulation estimates the effects of a 3- month, 10% increase in cooling degree days. The impacts on fuel demands and carbon emissions are shown in Fig. 4. As expected, electricity demand increases, leading to an increase in the demand for coal and oil. Natural gas consumption increases slightly due primarily to higher consumption in the electric utility sectors. Carbon emissions increase 1.5–3.7%.
