BEES 2.0--Building for Environmental and Economic Sustainability
Software from the Building and Fire Research Laboratory, National Institute of Standards and Technology (NIST). No charge. Download from the Web at firstname.lastname@example.org. Information: Barbara C. Lippiatt, 301/975-6133, email@example.com order from the U.S. EPA’s Pollution Prevention Information Clearinghouse, 202/260-1023,
It’s really exciting to be able to simply jump on the Web and download for free the second version of BEES—a software tool that uses detailed environmental and economic data to compare the relative performance of competing options for 15 common building components (See EBN for a detailed review of BEES 1.0). Improvements to the software tool include a doubling of the number of materials covered, four new environmental impacts, more detailed physical flow reporting, updated environmental and economic data, and lots of user-friendly changes to the workings of the software tool. The Welcome screen now makes important limitations of the program very explicit, and the new Technical Manual and User Guide is worth reading for its introduction to the science of environmental life-cycle analysis alone. There is no question that the software and manual will move any serious user into a better understanding of the issues and methods of this rapidly evolving field.
Unfortunately, there are still a number of serious weaknesses that make BEES more like an interesting work-in-progress than a published tool from a government agency known worldwide for setting standards. These weaknesses involve transparency, methodology, and data quality.
Oddly, the BEES Web site claims that it is designed to be transparent. While some data tables are accessible, they represent only one layer of information and are poorly labeled. The formulas that would explain how the data is used in the program are not provided. Also, because much of the data is proprietary, it is impossible to explore upstream impacts. Our experience was that it is very difficult to verify even simple questions about the source of particular numbers (see bulleted items below). Even people who supplied data to NIST for the program have difficulty determining how their information is being used: “It isn’t transparent, it’s a black box,” reports Martha Van Geem of Construction Technology Laboratories, Inc., who helped provide data on cement and concrete to BEES on behalf of the Portland Cement Association.
EBN’s concerns with the methods used in BEES to compare dissimilar impacts are complex and were described previously (see “BEES 2.0 on the Way, New Problems Emerge,” EBN , page 4). To be fair, there is no easy resolution to these issues, and the problem is now acknowledged on the Welcome screen and in section 2.4 of the Technical Manual.
The data quality problems range from seemingly simple mistakes to larger concerns about whole sets of data collected from different sources without peer review. Since the upstream data is proprietary, its quality is hard to assess. But the simple mistakes are discouragingly numerous. “I don’t think it has enough technical review,” notes Van Geem. Unfortunately, the lack of transparency make it very difficult to do a serious technical review of BEES.
Van Geem reviewed version 2.0 before its release and caught a number of errors, both in the cement data and elsewhere in the program. “Some of the errors we caught just didn’t make sense if you knew anything at all about buildings,” she notes. Van Geem also noted, however, that Barbara Lippiatt of NIST has been very responsive to their concerns about the program. When Van Geem and her colleagues found a significant error after 2.0 was released, a corrected version of the program, version 2.0b, was produced in short order. (BEES users who downloaded or ordered the program before August 4, 2000 should be sure to get the updated version.)
In just a short time with BEES, we twice got unexpected results that, after extensive research, we were able to trace to questionable data.
•Cost of vinyl siding and stucco: When we used BEES to compare various exterior cladding alternatives, including stucco and vinyl siding, the first costs (including materials, labor, and overhead) came up as $1.29/ft2 ($13.89/m2) for stucco and $2.08/ft2 ($22.39/m2) for vinyl. Stucco is less expensive than vinyl siding? In our review of the source reference—R. S. Means’ 2000 Building Construction Cost Data—we could not find these numbers and found vinyl siding, regardless of the application, to be less expensive per square foot than stucco (it gets a little confusing because the cost of an exterior cladding system depends on what type of wall assembly is included—block wall or stick-framing and sheathing—and what kind of stucco—elastomeric, two-coat or three coat—system is being considered). But this is part of the problem—BEES does not supply enough information to determine exactly what wall assemblies, and therefore what reference data, are being used.
•White and black asphalt roofing shingles: When setting up a comparison of various roofing materials in BEES 2.0, a nifty screen pops up asking whether you are considering roofing materials for the Sun Belt. Why? Because in cooling-dominated climates, the reflectivity of the roofing material can have a large impact on energy consumption, which affects the environmental footprint. This is great stuff, including quantitative energy impacts at the systems level in the evaluation of building materials.
The effect of this difference seemed too large, however. After tracking down BEES 2.0 reference materials, we discovered that BEES assumes that the dark shingles have an albedo of 0.1, while the light-colored shingles are at 0.7. More accurate estimates are 0.05 for black shingles and 0.25 for white—a threefold drop in the difference in albedo that would generate significantly different results in the BEES comparison.
With version 2.0, BEES has improved dramatically. Anyone interested in the environmental performance of building materials should take advantage of the free download and check it out. In terms of underlying methodology and data quality, however, it is still a prototype that has the potential for gross misuse by architects, builders, and product manufacturers less able or willing to critically assess the results. In fact, as the software gets more user-friendly, the risk of the results being misused by people who don’t understand the limitations increases.
We applaud NIST and EPA for initiating the vital and daunting task of quantitatively assessing the comparative environmental performance of building materials. We just hope that in future versions the avoidable errors will be caught, and the unavoidable shortcomings will be highlighted to prevent misinterpretation of the results.
– PY & NM