- Counting the Electricity Grid
Counting Carbon: Understanding Carbon Footprints of Buildings
By Nadav Malin
Historically the amount of CO2 in the atmosphere hovered just under 300 parts per million (ppm), but it’s now approaching 400 ppm. CO2 is not the most powerful of the greenhouse gases on a per-molecule basis—not by a long shot—but it is by far the most common and most significant of those generated by humans. Various targets have been proposed as acceptable levels of CO2, most famously 450 ppm, above which the resultant temperature rise would likely cause extreme disruption to Earth’s ecological and social systems. Many policy initiatives give lip service to this goal, but current actions are inadequate to reach it. Based on more recent scientific findings, author Bill McKibben has launched a campaign to reset that target at 350 ppm, a point we passed in 1988. That’s a much more ambitious goal, but one that, if achieved, would more likely lead to a future climate that resembles our own.
Regardless of the target, there is general agreement that we have to slow the growth in carbon emissions and then shrink those emissions. As researchers seek ways to reduce human-generated carbon emissions at a cost that society will accept, buildings consistently emerge as the best opportunity. “Buildings are the biggest and lowest-hanging fruit in dealing with greenhouse gases in the atmosphere,” says architect and researcher Hal Levin, who chairs the Project Committee on Carbon Emissions Tool Development of the American Society of Heating, Refrigerating, and Air-Conditioning Engineers (ASHRAE). Since carbon emissions from buildings generally follow energy use, we’ll go a long way simply by making buildings more energy efficient.
But energy use, whether it’s measured in dollars or kilowatt-hours, in absolute terms or as a percent reduction against code, is not exactly a measure of carbon emissions. How energy is generated and distributed changes how much carbon is released in the process. And energy used in the building is not the whole picture when it comes to greenhouse gas emissions. What you count and how you count it can change both the answers you get and what you do about them.
What to Count
Emissions from energy used in building operations are the obvious place to start in measuring a building’s carbon footprint, but stopping there leaves many emission sources off the table, such as emissions from transportation to and from the building, providing water to the building, and creating the building itself. These other sources grow in relative significance as the building’s operations get more efficient: “The more you improve the energy performance, the more your carbon footprint is dominated by transportation,” notes Christopher Pyke, Ph.D., director of climate change services at CTG Energetics, based in Irvine, California.
Which carbon sources you attribute to the building depends largely on why you are counting. Many carbon-accounting schemes are driven by efforts to regulate emissions or monetize emission reductions, so they focus on careful assignment of emissions to a variety of owners in a way that avoids double-counting. “Their fundamental interest is to take the entire economy and add up how much carbon is in it,” notes Pyke. “But from the point of view of reducing greenhouse gas emissions, ownership is not that important.” In its work for the U.S. Green Building Council (USGBC) on the LEED 2009 weightings tool (see EBN ), CTG’s primary interest was asking, “What does the building provide in terms of opportunities for reducing emissions?” said Pyke. Hence transportation-related impacts are included in the tool, even though the building owner can’t claim ownership of those emissions or cash in on reductions.
Carbon emission estimates based on energy models and other predictive tools are useful, but they aren’t sufficient for carbon accounting schemes, according to Pyke. “We’re going to plan based on models, but we’ll regulate and manage based on data,” he says. That doesn’t mean that models won’t play a role in regulations, however. Since the U.S. Supreme Court ruled in April 2007 that the U.S. Environmental Protection Agency (EPA) has the authority to regulate greenhouse gases as pollutants, there is a growing trend to include carbon emissions as part of the environmental impact statement that is filed for permitting and approvals, according to Sean Cryan, associate principal at Mithun Architects + Designers + Planners. These requirements, beginning in California, Massachusetts, and western Washington State, have created a demand for new tools to predict the carbon impacts of new projects and whole developments.
CO2 has captured our attention, but it is certainly not the only air-pollutant from building-related activities. Some of the tools described below also quantify emissions of sulfur dioxide (which causes acid rain), smog-generating nitrous oxides, and other pollutants.
Operational energy use
Energy used onsite is the most direct, and typically the most significant, contributor to a building’s carbon footprint. This energy usually arrives at the building in the form of electricity and natural gas or other fossil fuels, such as fuel oil or propane. Each of these fuels has a carbon footprint, so if you know the type and amount of fuel consumed, you can estimate the building’s primary contribution to greenhouse gases.
Electricity is the most common form of energy used onsite. Power plants generate this electricity with a range of fuels, each of which emits a different amount of carbon. Unfortunately, half of the electricity in the U.S. is generated by burning coal, which is the most carbon-intensive fuel (see chart). As a result, buildings in areas of the country where electricity is generated primarily from coal have a higher carbon footprint than those in other regions, even if they are equally energy efficient.
Most tools that convert electricity use to carbon emissions rely on EPA’s Emissions & Generation Resource Integrated Database (eGRID), which is also the basis for EPA’s online Power Profiler. This database provides average annual emissions of carbon and other pollutants for each power plant in the U.S. based on the mix of fuels used by that utility to generate electricity.
Macroscale models of carbon emissions consider the transportation sector separately from the buildings sector. Many building-related strategies can reduce energy use and emissions from transportation, however; they range from locating buildings in pedestrian- and transit-friendly places to designing them with pedestrians and cyclists in mind (see EBN ). Automobile use is often measured in vehicle-miles traveled (VMT), and converting VMT to gallons or liters of fuel depends on the efficiency of the vehicles—the average passenger car in the U.S. achieves 22 miles per gallon (10 liters/100 km). Gasoline contributes about 19.4 pounds of CO2 per gallon (2.3 kg/l), according to EPA. A more difficult question is whether, for carbon-accounting purposes, those miles should be pinned on where a person lives or on where she works or shops. Fortunately, we don’t have to solve that riddle to encourage carbon-reducing strategies.
The energy used to treat and transport potable water is substantial, especially in dry regions. Pumping water over the Tehachapi Mountains from Northern to Southern California represents an extreme case: the energy used to deliver water to Southern California homes amounts to about one-third as much as the average electricity used in each of those homes, according to a 2005 report from the Pacific Institute.
This energy can be translated to carbon emissions, but the relative importance varies widely. In USGBC’s LEED 2009 credit weighting tool, water use at the typical 135,000-ft2 (12,500-m2)office building accounts for anywhere from 5 to 823 metric tons of carbon dioxide equivalent (CO2e), depending on how much water the building uses—which is driven largely by the size of irrigated landscape area—and how much energy it takes to deliver that water. Compared with operating energy calculated using USGBC’s median building-systems scenario, the carbon from a building’s water use ranges from 0.17% to 29% of the carbon associated with operating energy.
These numbers do not account for methane released from wastewater treatment plants, which can be significant if the plant uses anaerobic digestion and doesn’t capture its methane. “Some sewage facilities actually have data on how much methane is released,” reports Rod Bates, environmental researcher at KieranTimberlake Associates, although in his experience they are more likely to have good data if they are capturing and reusing the methane.
Before environmental life-cycle assessment (LCA) became widespread, the environmental impact of materials in a building was often estimated based on their embodied energy, meaning the energy used to make and transport those materials. LCA tools provide a more directly relevant metric in the form of “global warming impact” CO2e. This measure adds to the energy picture by including things like carbon dioxide released from minerals as limestone is turned into portland cement, and methane released from the digestive systems of sheep as they grow wool. (Wool carpet is much worse than synthetics on a climate-change basis for this reason.)
It is also possible to estimate emission reductions that come from recycling construction waste (and from recycling by occupants). The reductions are primarily from avoiding the need to manufacture virgin materials when recycled-content materials are available instead. EPA has an online calculator and downloadable spreadsheet that can help estimate these savings. While it is not designed specifically for construction waste, it does include many of the material categories that are encountered on construction sites, including corrugated cardboard, lumber, bricks, and carpet. See.
“Carbon embodied in a building typically represents 13% to 18% of the carbon emissions over the life of the building,” says Cryan. This estimate includes CO2 emitted while manufacturing and transporting materials, as described above, as well as activities on the construction site, including emissions from running construction equipment and the release of carbon that had been sequestered in the soil. “We spent a long time figuring out the carbon associated with moving a cubic yard of soil,” reports Cryan.
Researchers have measured carbon sequestered in various soil types, but precise numbers are questionable when it comes to releases of sequestered carbon because so little is known about what really happens to soil-based carbon on a construction site. “Whatever they’re assuming is pure guess,” suggests Steven Hamburg, Ph.D., of Brown University, about efforts to include this information in a simple calculator. Nevertheless, both Hamburg and Cryan suggest the same design responses: minimizing the size of development footprints, increasing the density of neighborhoods, and protecting natural areas from development.
How to Count
Based on the simplified method described in the sidebar for estimating carbon emissions from operating energy, any energy simulation can be used to predict carbon impacts. Increasingly, energy modeling tools have this capability built in, with output reports that include predicted carbon emissions alongside the predicted energy use.
Results from energy modeling tools
In June 2008 Integrated Environmental Solutions, Ltd., released VE-Ware, a free plug-in for Autodesk’s Revit Architecture and Revit MEP software that generates a carbon footprint report for a modeled building. Carbon-emission reports are already built into EnergyPlus, which now has an energy simulation plug-in available for Google SketchUp and for Green Building Studio, a Web-based energy simulation service that was recently acquired by Autodesk. The translation from building information models (BIMs) created for design and construction purposes to models that can be used for energy simulations is rapidly improving, but it is not yet seamless, especially for complex buildings and detailed designs. As those connections improve, carbon footprint estimates should increasingly be available throughout the design process at the click of a button.
As a quick test of these tools, designers at KlingStubbins created a simple core-and-shell Revit model of a three-story building in Boston and generated an energy and carbon report using both Green Building Studio and VE-Ware. The results were “uncannily close,” according to Jason Olsen, an associate involved in the experiment. Both systems use gbXML as the protocol for transferring data to the simulation tool, but the underlying energy simulation engines are different, and the assumptions the tools made about the breakdown between electricity and natural gas were also somewhat different. Additional testing against models generated in a more controlled way by experienced engineers would be advised, but this single test indicates that these tools may already provide useful carbon-footprint feedback, at least from early design models.
A big piece of nearly every building’s carbon footprint comes from the electricity it uses, so how you translate kilowatt-hours into tons of CO2 is important. This is not a simple translation, however. Complicating factors include the way power is shared throughout large regions, variations in emission profiles from electricity generation over time, and the difference between reporting based on historical power profiles and that based on projected future conditions.
How tightly to define the grid?
EPA’s Power Profiler and many other tools estimate carbon emissions for electricity use as reported in eGRID for the electricity subregion in which the building is located. EPA divides the U.S. into 26 such subregions, and each has its own fuel mix and emissions profile. The Pacific Northwest, for example, has a lot of hydropower, so its emissions are relatively clean. The mountain states, on the other hand, rely primarily on coal, resulting in lots of carbon and other emissions. These variations make an inefficient building in Seattle look better in terms of carbon emissions than a high-performing building in Denver.
One could conclude that it’s more important to build low-energy buildings in Denver than in Seattle, but that’s not the whole story, according to Michael Deru, Ph.D., of the National Renewable Energy Laboratory, because power is transported throughout a much wider region to meet demand: “If you use more energy in Seattle, you’re using hydro, but you’re causing more coal to be burned,” Deru says, noting that if hydropower were not needed in Seattle, it would be transferred to a part of the grid where it could offset electricity generated from fossil fuels. Deru argues that we need to minimize nonrenewable energy use nationwide, so it’s better to use national average values for carbon intensity of electricity. If not national values, he suggests using values based on the large regions defined by how the nation’s utility grids are interconnected. The continental U.S. has three of these interconnected regions, one each for the eastern and western halves of the country, and a third covering most of Texas.
A related issue is the question of focusing on existing generation rather than the additional (marginal) capacity that is needed to meet new demand. Some argue that it makes more sense to estimate the carbon footprint of a new building based on this marginal capacity. In the Pacific Northwest, for example, no new hydropower dams are being built, so most new loads are being met with fossil-fuel or nuclear plants. Some would argue, then, that it is not reasonable to allow new buildings to take credit for existing hydropower in their carbon footprints.
Even within a small region, emissions from electricity generation vary by time of day, by season, and from one year to another. Because power is in short supply during peak times (usually daytime) and is readily available at other times, incentives exist specifically to shift the demand for electricity from day to night. California’s Title 24 energy code includes factors for time of use for this reason.
Ice storage is a typical example of this strategy: chillers are run at night to make ice, which is used during the day to cool a building so the chillers can sit idle. But whether these measures reduce carbon emissions depends on various factors, such as which fuels are used to generate power under which conditions. California relies heavily on natural gas for its base-load power, so the electricity it buys to meet peak loads is dirtier, notes Hal Levin. But in South Dakota, where the base load is met by coal, “they are often buying cleaner electricity for peak conditions,” Levin says.
A more difficult type of variation to track and design for is the availability of hydropower, which in many places depends on snowpack in the mountains, according to Levin. That resource changes both through the year and from one year to another.
Even power from the same fossil-fuel-powered plant can vary in its emission profile based on ambient temperatures and how that plant is operated. At night, when the air is cooler and demand is low, operations can be optimized for efficiency, according to Deru. During the day, however, the need to meet constantly changing demand leads to compromises that increase emissions.
ASHRAE is working on a tool that will take into account regional and time-of-use variations in the electricity grid, and other factors, to provide more accurate carbon-footprint estimates. Because standard energy simulation software models energy use in buildings on at least an hourly basis, this tool—due out in 2009—is intended to use that hourly data to predict carbon emissions.
Upstream emissions of fuels
It takes energy to extract, refine, and deliver fuels to a building or to a power plant. These “precombustion effects” can account for anywhere from 5% to 20% of the total emissions associated with each fuel used in a building (see table), according to the National Renewable Energy Laboratory’s U.S. Life-Cycle Inventory Database. When it comes to fuels used at power plants to generate electricity, it is not as easy to compare precombustion carbon emissions with carbon emissions from fuel content, especially for sources such as nuclear and renewable energy, which have low or no combustion-related emissions.
Of all the fuels, precombustion CO2e emissions are highest for natural gas, largely because of the high global warming potential of gas that leaks from pipelines. Overall emissions, including both precombustion and combustion effects for nonrenew-able fuels, range from about 2 pounds of CO2e per million Btus for nuclear energy to 225 pounds for coal.
The estimates for nuclear energy are hotly debated, with some arguing that as the quality of available uranium declines, more energy will be needed to process it. Also, even where estimates of precombustion effects are available, the long-term energy and environmental cost of processing and storing waste nuclear fuel are still not taken into account.
With all the complexities and efforts to refine our estimates, it’s easy to lose track of the fact that we can reduce carbon emissions dramatically by doing things that we already know about. Variations in the electrical grid and in fuel choices notwithstanding, reducing energy use also reduces emissions. Beyond that simple point, here are a few things to keep in mind.
Durability of reductions
If design and construction strategies that we implement today are going to help us meet carbon emission goals in 2020 or beyond, those strategies have to be robust and durable. Nothing is guaranteed, but there is a fundamental difference between, for example, orienting a building properly and sealing ducts. The former should last as long as the building; the latter, on the other hand, may have a dramatic short-term benefit, but “How tight do the ducts stay?” asks Pyke. This distinction won’t show up in an energy model, but it’s worth considering when evaluating which strategies to implement on a project. Rather than investing in sealing ducts, it’s better to design the building so that all ducts are contained in the conditioned space, so slight leaks are not a problem.
This distinction is even more obvious when it comes to the nebulous world of carbon offsets. An owner may decide that reducing carbon emissions from operations is too complicated and might choose instead to purchase carbon offsets or renewable energy credits with a low carbon load. But even if one accepts that buying offsets is as effective as directly reducing emissions (which few experts do), that purchasing policy can easily be reversed in the future, so it isn’t nearly as compelling a solution for long-term climate change concerns.
Other policy-related solutions are similarly vulnerable. Providing mass-transit passes to employees is certainly laudable and beneficial, but it reduces a building’s carbon footprint only if the employees use those passes and only for as long as they are available. Structural solutions like limiting the amount of parking may be more durable over time, in that they can survive changes in management philosophy.
Measures that shift electric power demand from peak to off-peak times may not directly reduce CO2 emissions, but they may do that indirectly. By reducing peak loads they help prevent or delay the construction of additional power plants. If the plants are not built, the business case for investing in conservation is much stronger than if the power is readily available as excess capacity.
That factor is one of many ways in which changes to the economy and to the environment over time may change the long-term impacts of decisions we make today. Anticipating these changes is tricky, so perhaps the smartest response is to maintain some flexibility in design solutions. This kind of thinking is not new to architects, notes Levin: “We always look at that tradeoff in building design between optimization and flexibility.”
KieranTimberlake has learned a lot by measuring and seeking to minimize its own carbon footprint, according to Bates. “We monitor our own activities, and it allows us to use ourselves as a test bed for different types of strategies,” he says. It also helps to have clients who are inspired to reduce their emissions—that provides KieranTimberlake with the incentive to do the research and improve the performance of its designs. But that conversation goes both ways: “A key factor is communicating to the clients, in a language they can understand, what emissions are associated with their buildings,” Bates suggests, adding that his firm converts carbon emissions into the equivalent vehicles on the road or acres of forest needed to offset those emissions. When clients can relate to the impacts in that way, they are more likely to support the effort. EPA has a handy online calculator for translating emissions into units that people can relate to at:.
With his 2030 Challenge, architect Edward Mazria, AIA, has invigorated the design professions to respond to climate change. While Mazria has chosen to emphasize a reduction in fossil-fuel energy use as his metric, the building industry as a whole is learning that it’s important to measure and understand more than one number as we strive for more effective, and climate-friendly, buildings. When it comes to CO2 emissions we are a long way from precision, but there is enough information to provide some guidance if you’re willing to probe a little and use the resulting data constructively.