Energy Systems

This section focuses on the drivers of changes in the North American energy system and how these drivers have influenced changes in carbon cycle dynamics. A driver is any natural or human-induced factor that directly or indirectly causes a change in the system (see, for example, Nelson 2005). Drivers often are divided into categories, such as direct versus indirect, proximate versus primary, and immediate versus underlying. These distinctions attempt to identify the speed and scale at which the driver operates and the driver’s linkage to the environmental state.

The first systematic discussion of drivers of environmental change emerged as the IPAT identity, where environmental impact (I) was estimated by multiplying the population (P) by affluence (A) and by technology (T; for a review, see Rosa and Dietz 2012). Subsequently, the drivers (PAT) were identified as primary or indirect, given that they work largely through other drivers. For example, with increasing affluence, households have more expendable income to consume energy (via air conditioning, for example) and subsequently increase their energy use (Sivak 2013; Davis and Gertler 2015). The point is that increasing affluence operates through both population units (households) and increases in energy consumption via more expendable income. The IPAT equation has expanded into a much more complex set of influences that help to explain environmental change (see, for example, Reid et al., 2005; Marcotullio et al., 2014).

The IPAT equation was the model for the Kaya Identity, named after Yoichi Kaya, which provides similar multiplicative elements to help explain the change in CO₂ emissions (Rosa and Dietz 2012; EIA 2011b).

F = P × G/P × E/G × F/E

The formula for primary drivers of carbon emissions (F) includes population (P), GDP per capita (G/P), energy per GDP output (energy intensity, E/G), and carbon emissions per energy input (carbon intensity, F/E). Often the formula also includes sectoral structural changes. The variables in the equation are factors that include a much larger number of proximate or direct influences such as fuel price, resource availability, infrastructure, behavior, policies and other processes, mechanisms, and characteristics that influence emissions (see, for example, Blanco et al., 2014; Table 3.3). The Kaya Identity accounting categories often are used in the decomposition of emissions and provide an overarching framework for examining societal influences as well as a template for scenario development (Nakicenovic 2004). This section addresses the main factors identified in the Kaya equation. For a discussion of local influences on the carbon cycle, see Ch. 4: Understanding Urban Carbon Fluxes.; for social and behavioral influences on the carbon cycle, see Ch. 6: Social Science Perspectives on Carbon; for policy influences from respective governmental policies at the international, national, and state or provincial levels, see Section 3.7.

Figure 3.10 presents the factors of the Kaya Identity, along with total energy use, in a simple decomposition analysis for the North American region. Several points become evident in this graph, including those between 2007 and 2015: 1) population and GDP per capita increased by approximately 8% and 18%, respectively; 2) energy intensity and carbon intensity decreased by about 25% and 6.4%, respectively; and 3) emissions and energy use decreased by around 11% and 4.5%, respectively. That is, since 2007, while regional population and GDP per capita increased, energy use and ­energy-related CO₂e emissions decreased. The following subsections examine the factors in more detail to explain what happened. Each subsection includes a description of the factor and how it theoretically affects energy and emissions levels, along with a review of what actually happened, at the regional scale and for each economy.

3.6.1 Population Growth

The current population of North America is almost half a billion people and growing. The most populous nation in the region, the United States, continues to grow and is projected to do so at an annual rate of 0.34% through the end of this century, when population is estimated to reach approximately 648 million (UN 2015). Although growing populations can increase energy use and subsequent carbon emissions, this is not universally true. Increases in population do not necessarily produce proportional changes in environmental stress. Thus, population may have an elastic (greater than 1) or inelastic (less than 1) effect on emissions. If the impact is elastic, greater population will produce more problems such as traffic congestion, resulting in greater emissions than expected based merely on the proportion of increased population. The larger the city, the greater the congestion, and therefore the impact may be disproportionate compared to the growth of the population. Alternatively, larger populations may induce economies of scale and enable more efficient use of resources, thereby lowering the impact on emissions levels. In this case, the impact of population growth would be inelastic.

Between 2005 and 2015, North America grew by an estimated 45 million people (approximately 1.0% annually), and yet energy use and CO₂e emissions have declined. Alternatively, Mexico’s population has increased commensurately with national energy use and carbon emissions. During this period in Mexico, however, emissions first increased with population and then decreased even as population continued to increase.

3.6.2 Financial Crisis and Declines in GDP Growth

Increasing affluence can either increase emissions levels through increased consumption per capita or mediate emissions through shifts in the scale or composition of consumption. In 2008, the world experienced the global financial crisis, which hit particularly hard in North America. Feng et al. (2015, 2016) argue that the economic crisis, through lowering GDP per capita, also decreased the volume of consumed goods and services and was responsible for 83% of the decrease in U.S. emissions from 2007 to 2009, which totaled around 0.6 Pg CO₂e (164 Tg C), or 9.9% of the nation’s total. This decrease makes up the bulk of the regional change during that period.

However, according to the World Bank (2016c), the GDP for North America in 2007 was $17.7 trillion; after declining for several years, it rebounded by 2013 to reach $18.7 trillion (all values in this paragraph are in US$ 2010). By 2016, the region’s GDP was $19.9 trillion, or over 20% higher than in 2007. The per capita GDP by country also followed the same trajectory. In 2007, the approximate GDPs per capita were $48,600 for Canada, $9,300 for Mexico, and $50,000 for the United States. After falling to lows of $46,500, $8,700, and $47,600 respectively, in 2009, each country’s GDP per capita figures had equaled or exceeded 2007 levels by 2012. By 2015, Canada’s GDP per capita was $50,300, Mexico’s was $9,600, and the United States’ was $52,000 (World Bank 2018). Despite increases in GDP combined with population growth, energy use and CO₂e emissions have remained below 2007 levels. According to Shahiduzzaman and Layton (2017), from 2010 to 2014 real GDP per capita growth and population factors (without any mitigating effects) would have resulted in yearly CO₂ emissions increases of 25.5 Tg C annually (14.8 Tg C due to increases in GDP per capita and 10.8 Tg C due to population increases). Over the 5-year period from 2010 to 2014, therefore, an increase of approximately 127 Tg C was offset by other factors. Clearly, while the economic downturn was significant for the initial change in emissions trend, it does not account for the continued reduced energy use and GHG emissions from North America’s energy systems.

3.6.3 Reduced Energy Intensity

Energy intensity is the amount of energy per GDP output (E/G). When economic growth outpaces the increase in primary energy supply, energy intensities decrease. Therefore, lowering energy intensities can represent mitigation gains, if benefits of efficiencies are not offset by greater use. Over the long term, energy intensities in Canada and the United States have been declining, due partly to increases in the efficiency of fuel and electricity use, including a shift from large synchronous generators to lighter-weight gas-fired turbines and new fuel sources (e.g., renewables; U.S. DOE 2015b; see Section 3.4.3), and partly to changes in economic structure and saturation of some key energy end uses.

In the United States, from 1950 to 2011, energy intensity decreased by 58% per real dollar of GDP and is projected to drop 2% annually to 2040 (EIA 2015c). U.S. energy intensity in 2011 was approximately 7.73 megajoules (MJ) per US$1 purchasing power parity (PPP). Since 2004, the United States experienced a 1.6% drop annually in its energy intensity. Canada has some of the highest energy intensities of the IEA countries (IEA 2010). Canada’s energy intensity remains the highest among the regional economies and in 2011 was approximately 11.2 MJ per US$1 PPP. Canada’s geography, climate, and industrial structure, including its ­export-oriented fossil fuel industry, make it a highly energy-intensive country. Like the United States, however, its energy intensities also experienced significant decreases over the last half of the past century (EIA 2016c). Over the past decade, Canadian energy intensity dropped 1.5% annually, and since 1971 it has dropped by 39%. Decreases have been attributed largely to increased contributions of low energy–using commercial activities relative to high energy–using manufacturing, as well as the rapid growth of the Canadian economy compared to population growth (Torrie et al., 2016). These economic structural changes are more important to the nation’s falling energy intensity than increasing energy efficiencies. Recently, Mexican energy intensity also has been falling, but only slightly. Mexico, an emerging economy, had been increasing its energy intensity, but over the past decade it fell by 0.04% annually. Mexico’s energy intensity is now about 5.5 MJ per US$1 PPP.

An examination of the efficiency gains across sectors of the North American energy system demonstrates structural changes in end-use energy sector components. For example, reduced energy intensity in the electricity-generation sector can be tracked by heat rates. Average operating heat rates for coal and oil power plants for 2015 in the United States are 32.5% and 31.9% efficient, respectively, for power plant type. Average U.S. operating heat rates for gas-fired plants are around 43% efficient (EIA 2016a). However, gas turbine and steam generators typically have the lowest efficiencies, while ­combined-cycle plants have the highest. For example, in 2016, gas turbines were 25.2% and 30.4% efficient for oil and gas energy sources, respectively, while combined-cycle plants reached efficiencies of 34.6% and 44.6% for oil and gas, respectively (EIA 2018d). The increased share of natural gas–fired plants and the greater use of high-efficiency combined-cycle plants have helped to reduce the overall energy intensity of the U.S. electricity-generation system (Nadel et al., 2015). Notwithstanding the importance of economic structural changes in Canada’s decline in energy intensity, business energy intensity experienced a decline from 1995 to 2010 (22% of total decline), and increases in efficiencies in power generation contributed to this decline but only slightly (5% of total decline; Torrie et al., 2016). Mexico is undergoing a major set of policy reforms to open up its power sector, including the electricity system. Actions focused on reducing generation costs include reducing heat rates and losses from transmission and distribution, all of which will improve the electricity system’s energy efficiency (CEE and ITAM 2013; Robles 2016).

Energy-efficiency improvements in appliances and utilities, residential and commercial buildings, industrial, and transportation sectors also have slowed growth in North American energy demand and helped to decouple energy demand growth from GDP. The U.S. national efficiency standards implemented since 1987 have saved consumers 9.22 GJ or 21% of household electricity usage in 2015 (deLaski and Mauer 2017). Further, these efficiencies are expected to save 74.9 EJ of energy (cumulative from 2015) by 2020 and nearly 149.8 EJ through 2030 (U.S. DOE 2017b). The cumulative utility bill savings to consumers are estimated to be more than $1 trillion by 2020 and more than $2 trillion by 2030 (U.S. DOE 2017b). Utility energy-efficiency programs for the residential sector are achieving incremental savings of about 30.6 PJ annually, equivalent to 0.7% of all electricity sales with a cumulative impact many times this value, most at a cost of US$0.030 per kWh (Hoffman et al., 2017). While these savings are impressive, energy consumption for appliances and electronics continues to rise and the increasing number of devices has offset gains in appliance efficiency (EIA 2013a).

Independently, building codes reduced residential electricity consumption in the United States by 2% to 5% in 2006 (CEC 2014). Energy savings through building codes have been supplemented by the increase in green buildings. For example, from 2003 to 2016 the number of Leadership in Energy and Environmental Design (LEED)–certified buildings in the United States increased from 116 to over 24,700, those in Canada increased from 3 to 399, and the number in Mexico increased from 0 to 172 (see Table 3.4). The United States Green Building Council estimates that green building, on average, currently reduces energy use by 30%, carbon emissions by 35%, and water use by 30% to 50%, also generating waste cost savings of 50% to 90%. A rapidly increasing market uptake of currently available and emerging advanced energy-saving technologies could result in annual reductions of 1.7 Pg CO₂e (464 Tg C) emitted to the atmosphere by 2030 in North America, compared to emissions under a “business-as-usual” approach (Commission for Environmental Cooperation 2008). In Canada from 1990 to 2013, residential- and commercial-sector energy efficiencies improved by 45% and 33%, respectively. Canadian space heating energy intensity alone was reduced by over 38% as households and commercial and institutional offices shifted from medium- to high-efficiency furnaces, improved thermal envelopes for buildings (e.g., insulation and windows), and increased efficiencies of various ­energy-consuming items such as auxiliary equipment and lighting (Natural Resources Canada 2016b). In Mexico, energy efficiency in the residential and commercial sector has focused on lighting, appliance, and equipment replacement (IEA 2015b). In the United States, the share of space heating and cooling for residential energy consumption has been falling due in part to the adoption of more efficient equipment and better insulated windows. An increasing number of residential homes are built to ENERGY STAR® specifications (U.S. EPA 2015c), lowering their energy consumption to 15% less than that for other homes. U.S. households are increasingly incorporating energy-efficient features; in 2011, ENERGY STAR® homes made up 26% of all new homes constructed (EIA 2011c, 2012a).

Industries also have experienced lower energy intensities through shifts in technologies and greater efficiencies. For example, energy use in U.S. steel production has been declining. From 1991 to 2008, there has been a 38% decline in the total energy consumption used in the industry. The largest portion, 34% of the decline in the total energy consumption, occurred between 1998 and 2006 (EIA 2017f). In Mexico, the efficiencies of thermal power generation and of the power sector as a whole have been increasing rapidly since 2002 (from 38% to 45% in 2010 in the case of thermal power generation). This recent improvement is due to a switch in the power-generation mix to natural gas and to the spread of gas combined-cycle plants. In 2010, the gas combined-cycle power capacity accounted for 43% of the total thermal capacity. The country’s chemical industry also has experienced drops in energy intensity, falling by nearly 7% per year between 1994 and 2009 (ABB 2012). In Canada, industrial oil production has been driven primarily by a rapid rise in the extraction of bitumen and synthetic crude oil from the nation’s oil sands operations, where total output has increased by 140% since 2005. This has contributed to the 37-Tg increase in CO₂e (10.1 Tg C) emissions from mining and upstream oil and gas production from 2005 to 2015. However, from 2010 to 2015 the emissions intensity of oil sands operations themselves have dropped by approximately 16% as a result of technological and efficiency improvements, less venting emissions, and reductions in the percentage of crude bitumen being upgraded to synthetic crude oil (ECCC 2017b).

In the North American transportation sector, there have been considerable improvements in efficiency over the past decade as well as reductions in fuel use in vehicle miles traveled. The on-road transportation sector, in particular, has seen reductions in fuel use for both total and per capita vehicle kilometers traveled, as well as reductions in emissions of CO₂e. According to the U.S. Department of Transportation (U.S. DOT; U.S. DOT 2016), from 2005 to 2015 total average kilometers traveled per passenger vehicle dropped from approximately 20,100 to 18,200 and total average fuel use per passenger vehicle dropped from around 2,100 liters (L) to 1,800 L. As a result, total average kilometers per liter (km/L) of fuel consumed increased from 9.4 to 10.1. These efficiencies have been driven by changes in vehicle weight and power and by corporate average fuel economy (CAFE) standards. For example, according to U.S. DOT (2014), CAFE fuel standards have increased from 11.7 km/L in 2010 to 14.5 km/L in 2014 (based on projected required average fuel economy standard values and model year [MY] reports). In 2015, while total U.S. vehicle travel distance was 4% higher than that in 2007, CO₂e emissions for transportation were 1.73 Pg CO₂e (472 Tg C), or about 8% lower compared with 1.89 Pg CO₂e (515 Tg C) in 2007 (U.S. EPA 2016). Motor gasoline consumption has not exceeded the previous 2007 peak (EIA 2016i). From 1990 to 2013, Canada also experienced energy-efficiency improvements in the transportation sector by 27%, while energy use in the sector increased during this period by 20% (Natural Resources Canada 2016b). From 2004 to 2013, Canadian transportation energy use and emissions stayed fairly level at approximately 0.17 Pg CO₂e (46.4 Tg C; ECCC 2016b). Similar to the United States, the majority of transportation emissions in Canada are related to road transportation. The growth in road transportation emissions for the country is due largely to more driving. Despite a reduction in kilometers driven per vehicle, the total vehicle fleet has increased by 19% since 2005, most notably for both light- and heavy-duty trucks, leading to more kilometers driven overall (ECCC 2017b). According to IEA (2017a), from 2007 to 2013, Mexico’s transportation CO₂e emissions increased by 2.2% annually, amounting to 10% of the total increases during this period. Emissions for this sector are expected to increase further to 2040 as demand for personal vehicles increases in Mexico (SEMARNAT-INECC 2016).

Similar trends in the United States and Canada can be seen in freight rail transport, with decreases in U.S. freight rail fuel consumption and small increases in Canada (Statistics Canada 2016; U.S. DOE 2014a). Substantial increases in fuel consumption in the international aviation sector have occurred over the past decade for both U.S. and Canadian flights (Natural Resources Canada 2016d; U.S. DOE 2014b).

Overall, in both Canada and the United States, a large portion of fuel and electricity use, associated with residential energy use and personal transportation, is weakly coupled with positive change in GDP. Research in Canada suggests that personal transportation and household energy, which compose about a third of the nation’s total energy use, are not coupled to GDP growth, resulting in an overall decrease in energy intensity when GDP rises, even if there is no economic structural change or efficiency improvement (Torrie et al., 2018). This result has been a major contributor to declining energy intensities in Canada and possibly also in the United States during recent decades.

In summary, energy-intensity decreases have been an important factor in the current trends of CO₂e emissions for North America. Shahiduzzaman and Layton (2017) calculated that, between 2005 and 2010 and between 2010 and 2014, decreases in energy intensity of output were responsible for annual reductions of 19.2 Tg C and 21.7 Tg C from the U.S. energy system, respectively. Over the 10 years of these two periods, this trend translates to about 409 Tg C, which is offset by decreases in energy intensity.

3.6.4 Decreasing Carbon Intensity

The carbon intensity (F/E in the Kaya Identity) of energy use is another factor, like energy intensity, that affects the overall level of emissions from the energy system. Different fossil fuels have different carbon intensities (e.g., per unit of energy, coal emits about 50% more CO₂ than that by refined petroleum products), and some energy forms, like solar, wind, and nuclear, do not emit CO₂ at all. The mix of fuels being used in a society changes over time and with it the carbon intensity of the energy system. Changes in the carbon intensity of the North American energy system over the past decade have been significant and mostly evident in the United States and Canada, although Mexico also has contributed to the decreasing trend.

In the United States, carbon intensities for all major energy sectors have been dropping steeply since 2005. The greatest declines were experienced by the industrial and electricity sectors. The industrial sector produced the least amount of CO₂ per unit of primary energy consumed in 2016, with emissions of 41.5 kg CO₂e per GJ. The electric power sector, which is second only to the transportation sector, produced 45.3 kg CO₂e per GJ in 2016, which is now below the commercial and residential sector’s carbon intensities (EIA 2017j). Shahiduzzaman and Layton (2017) calculate that U.S. carbon intensity reductions have offset approximately 287 Tg C from the U.S. energy system over the past 10 years.

Canada’s carbon intensities have also been decreasing. Similar to the United States, decreasing energy generation from coal and oil and increasing generation from hydropower, nuclear, and wind were the largest drivers of the 31% decrease in emissions associated with electricity production between 2005 and 2015. The permanent closure of all coal-generating stations in the province of Ontario by 2014 was an important factor in changing the national fuel mix (ECCC 2017b).

After falling during the 1990s, Mexico’s carbon intensity increased between 2000 and 2010 (OECD 2013). Mexico’s CO₂e emissions profile is heavily skewed toward transportation and the power sector. The ongoing effort to switch from oil- to gas-fired generation has reduced the carbon intensity of Mexico’s electricity sector by 23% since 2000, and further improvements are expected (IEA 2016b).

Changes in the carbon intensity in North America are related to several trends, some of which have already been discussed in detail.

• The natural gas boom, including the shift from coal to cheaper and cleaner natural gas for electricity production and industrial processes (EIA 2017j), with the critically important caveat that venting, flaring, and fugitive emissions may be underestimated (see Section 3.4.2 and Box 3.3).

• Increased renewables in the fuel mix in all North American countries, including wind, solar, and bioenergy (with caveats mentioned for this last source; see Sections 3.4.3 and 3.4.5), driven, in part, by declining costs and changing fuel prices.

• A wide range of new technologies including grid-scale electricity storage and alternative fuel vehicles.

Many new technologies affect the potential of others. For example, improvements in electric vehicle battery technology help support improvements in utility energy storage. Energy storage improves grid stabilization and buffers peak electricity demands that, in turn, help support a larger share of renewables in the electric grid.

Other important technologies include the grid-scale electricity storage (i.e., previously mentioned new battery storage for wind and solar) and alternative fuel vehicles. Grid-scale electricity storage currently includes pumped hydroelectric storage but, in the future, also may be enhanced by a wide variety of technologies that serve an array of functions within the electric power system (EIA 2011a). There are currently 40 pumped storage plants in the United States totaling more than 22 GW of capacity (about 2% of the nation’s generating capacity; EIA 2013b). Canada has one pumped storage facility in Ontario with a 174-MW capacity, and Mexico is currently exploring the possibility of developing this technology.

With the transportation sector having the highest carbon intensity in the region, use of alternative fuel vehicles can help make significant reductions. These vehicles are designed to operate on fuels other than gasoline and diesel, including compressed natural gas, propane, electricity, hydrogen, denatured ethanol, and other alcohols and methanol. An example of the increase can be seen in the electric vehicle stock. Globally, electric vehicles surpassed 1 million in 2016. In the United States, there have been recent increases in the number of electric vehicles on the road from around 23,000 in 2011 to 118,000 in 2015, and Canada’s electric vehicles jumped from fewer than 1,000 to almost 7,000 during this same period (EV-Volumes 2017). Mexico currently is focusing on increasing biofuels for its vehicle fleet. With the 2017 launch of the Tesla Model 3, the number of electric vehicles may increase (Marshall 2017).

Notwithstanding the emergence of these new technologies, an important influence that has underpinned the current decrease in carbon intensity is falling energy prices. Among different fossil fuel choices, falling prices for one fuel relative to another provide incentives to consumers to shift fuels. According to Houser et al. (2017), the surge in U.S. natural gas production due to the shale revolution made coal increasingly uncompetitive in U.S. electricity markets. Coal also faced growing competition from renewable energy.

Oil, gas, and coal prices have all dropped recently. From 2014 to 2015, world oil prices dropped dramatically and, to a lesser extent, so did natural gas and coal prices. From 2010 to mid-2014, global crude oil prices were relatively stable but historically high, at more than US$100 per barrel. In June 2014, Brent crude oil, a key global crude oil pricing benchmark, traded above US$110 per barrel. Later in 2014, oil prices began to drop, and, by January 2015, prices had declined by about 60% to under US$46 per barrel. Both Brent and West Texas Intermediate, a benchmark for U.S. crude oil, remained in the range of US$40 to US$60 per barrel for much of 2015 (National Energy Board 2016). The collapse in prices was driven by a marked slowdown in demand growth and record increases in supply, particularly tight oil (sometimes called shale oil) from North America, as well as a decision by the Organization of Petroleum Exporting Countries (OPEC) not to try to rebalance the market through cuts in output (IEA 2015a).

Differing from oil, there is no global pricing benchmark for natural gas. Instead, the three major regional markets (North America, Asia-Pacific, and Europe) have different pricing mechanisms. In North America, gas prices are determined at hubs and reflect local gas supply and demand dynamics. Notwithstanding the different market conditions, the surge in natural gas production within North America has reduced prices. While natural gas prices declined globally, the pace and extent were dramatic in North America. In the United States, for example, the average price for natural gas to power plants dropped from $10 per thousand cubic feet (ft³) in 2008 to $3 in 2016, a 71% decline (US$ 2016). During this period, despite falling coal prices, the average delivered cost of coal to power plants decreased by only 8% in real terms (Houser et al., 2017; IEA 2015a).

The increase in low-carbon energy sources also has been driven in part by falling costs of renewables. Globally, bioenergy-for-power, hydropower, geothermal, and onshore wind projects commissioned in 2017 largely fell within the range of generation costs for fossil-based electricity. Drivers of cost reductions include technological improvements, competitive procurement, and a large and growing base of experienced project developers (IRENA 2018a). In North America, between 2008 and 2016, the price of onshore wind declined by 36%, and the price of solar PV modules fell by 85% (Houser et al., 2017), prompting expansion in these PV sources. Wind prices are projected to be competitive with natural gas by 2050 (U.S. DOE 2017a). The cost of distributed generation, specifically distributed rooftop PV systems, also is declining. Median installed prices for distributed PV systems declined 6% to 12% per year from 1998 to 2015, and the decline was faster after 2009 (Barbose and Dargouth 2016).

Declining costs of renewable power generation along with increased competition from cheap natural gas are responsible for 67% of the decline in U.S. domestic coal consumption (Houser et al., 2017). Although low prices in natural gas relative to those of oil and coal have helped to reduce carbon intensities, continued low fossil fuel prices also can decrease pressure to develop renewables, possibly pushing carbon intensities in the opposite direction. IEA (2017a) suggests that this dynamic will affect conditions in the near future, unless the price of fossil fuels increases.