With its dominant position defining green building in the North American market, the LEED Rating System is a popular target for critics with a wide range of axes to grind, some justified, others less so. One of the more valid concerns is that LEED’s promises of energy savings (and therefore carbon reductions) are just that—promises. With the exception of LEED for Existing Buildings, which looks at actual operations, LEED’s various rating systems assign energy points to buildings based on predictions made during design. How well those predictions hold up in reality has, until now, been subject to conjecture.
The actual energy use of LEED-certified buildings varies dramatically, even once energy-hogging labs and hospitals are excluded as in the data shown here. Note that some LEED buildings, especially those certified at the Gold or Platinum levels, are approaching or exceeding the initial target of the 2030 Challenge, which is 50% of the national (CBECS) average.
At Greenbuild 2007 in Chicago, the U.S. Green Building Council (USGBC) and the New Buildings Institute (NBI) presented a groundbreaking analysis of actual energy performance of buildings that had been certified, as of the end of 2006, under LEED for New Construction and Major Renovations (LEED-NC). The results have something for everyone. They justify the critics by revealing that, for any individual building, there is little correlation between predicted and actual energy use, especially among buildings designed to use very little energy. The difference between actual and predicted performance was not all bad news, however—many buildings are using
energy than was planned. Average energy performance across all the buildings in the study, however, was remarkably close to predicted levels. This information suggests that LEED is valuable as a policy tool because it delivers savings across a market, but that the building industry has work to do before it can ensure the life-cycle cost benefits of individual buildings.
The actual savings vary much more widely than the proposed savings that were predicted in the LEED submissions—with some projects actually showing negative savings (usage that exceeds their base case). The boxes show the quartiles (25% of projects) above and below each median. Note that LEED assigns energy points for predicted energy
cost savings, which do not always correspond to predicted energy savings.
Brendan Owens, P.E., LEED technical director at USGBC, told
that, while he strongly supports the need to verify actual performance of LEED buildings, he had fears about what such a study might find. Research that looks at actual building performance “generally leads to conclusions that you don’t want to talk about,” he said. “I wasn’t expecting to want to publish it. It’s a really nice surprise from that standpoint.”
To compile this data NBI researchers contacted people associated with 562 LEED-NC buildings. They got some response for just under half the projects and usable data for 125 of them. Based on those 125, they concluded that LEED buildings are, on average, 25%–30% more energy-efficient than non-LEED buildings as represented in the U.S. Department of Energy’s Commercial Buildings Energy Consumption Survey (CBECS) database. As might be expected, buildings certified at Gold or Platinum levels tended to achieve more energy conservation points, and their average savings were about 45% better than the CBECS baseline. The average Energy Star rating for all the LEED buildings was 68, slightly less than the 75 that is required to qualify for the Energy Star label.
This graph shows how the actual energy use of LEED buildings compares to the national average values for the same building type—although only the office category has enough buildings to represent a reliable sample.
Energy use among buildings in the study is highly variable, even within a single building type (such as offices), although the range of actual building energy use values in the CBECS database is much larger, according to Mark Frankel, technical director at NBI. Some LEED projects are actually using more energy than their base case energy model, which represents minimum performance to code in many states. NBI also found that the base-case values, against which the energy savings of LEED buildings is predicted, were on average quite close to the national average energy use values in CBECS. That surprised Frankel, who noted that ASHRAE Standard 90.1, which governs that base case, is generally assumed to require much better performance than the CBECS average. “I’m not sure that we know exactly what’s going on there yet,” Frankel said. He dismissed the suggestion that energy modelers may be distorting their base-case assumptions to earn more LEED energy points: “Every modeler in the country would have to be dishonest for that to be a major factor.”
Some buildings in the data set are laboratories, hospitals, and other building types with relatively high energy usage—over 150 thousand Btus per square foot per year (kBtu/ft2
·yr) (470 kWh/m2
·yr). Noting that labs used, on average, twice as much energy as predicted, the researchers conclude that “energy use of high-energy building types is not well understood by designers” and that “neither the LEED program nor the modeling protocol addresses these projects well.”
One might have expected commissioning to help align actual energy savings with predictions, but this analysis doesn’t prove that. In fact, projects that achieved the LEED credit for enhanced commissioning had actual savings that were, on average, slightly worse than predicted. Projects that also achieved points for either measurement and verification, daylighting, or views, on the other hand, showed slightly better actual performance than predicted.
While the study supports USGBC’s claims that LEED buildings as a group save energy, it “hardly lets us rest on our laurels,” Owens said. He intends to further analyze the results and extend the work to find ways to make LEED more effective at delivering promised energy savings.
For more information:
The presentation on which this article was based is included in the proceedings from Greenbuild 2007. A more comprehensive report on this research is due out early in 2008.
U.S. Green Building Council
New Buildings Institute
White Salmon, Washington
December 1, 2007