We know that Environmental Product Declarations can be both a useful tool to evaluate products’ environmental footprints, as well as source of frustration when we’re looking for accurate and comparable data. To help you navigate those challenges, our very own Monica Huang, Brook Waldman and Kate Simonen recently published a paper in the Journal Buildings & Cities that presents a new method to account for estimates of variation in EPD data.
Learn more about this research and share it among your network of informed building product purchasers!
We routinely provide embodied carbon comparisons to design teams, often use EPDs for the same.
Even when we’re not using EC3, we do try to address the inherent uncertainties in EPD comparisons, and try to minimize those by intuitively and qualitatively applying the data quality factors you have in the paper: for e.g. 1) comparing product-specific EPDs to other product-specific EPDs; 2) not focusing on the actual numbers too much when having to compare industry-wide EPDs to each other or product-specific EPDs, but the order of magnitude of differences, 3) not comparing across PCRs unless it’s a larger system-level LCA, etc
Have you found that using the uncertainty-adjusted GWP values gives you very different procurement recommendations than what you’d have gotten if you were simply using the EPD’s GWP values as is? (Of course, adjusted for units and quantities needed for same function)
Or is the uncertainty-adjustment intended more for situations when teams are trying demonstrate a specific % GWP reduction - required by say the LEED EC3 Pilot Credit or ILFI ZCC or a client goal?
As one simple example, here is a comparison we did between Tally and EC3’s uncertainty-adjusted numbers for rebar. Whether coincidence or not, it happens to give approximately the same % reduction.
If we compare the unadjusted CRSI EPD to the unadjusted Nucor Seattle EPD, we’ll see a reduction of 50%. Now the 69% reduction using the CLF method does start to look different.
That’s one of my questions - in such industry vs product comparisons - does the CLF methodology artificially inflate the CO2e savings by penalizing the industry-wide EPD for more uncertainty?
My other concern is about the absolute numbers which will get inflated upon using the CLF methodology. Right now for our core-and-shell LCAs, take offices for example, we’re seeing a GWP intensity of ~500 kgCO2e/m2 for optimized buildings, and ~630 kgCO2e/m2 for baseline buildings. If we add in uncertainty-adjustments, these may go up by 20-50%. This might then create issues when comparing against absolute benchmarks, such as ILFI Zero Carbon’s maximum of 500 kgCO2e/m2. Or even comparing embodied carbon to operational carbon. Although perhaps these aren’t ‘issues’, they’re now ‘truer’ reflections by adding in the uncertainty.
Completely agree with the questions you are asking.
The intent of the illustration I sent is to compare what a user of the software see as the GWP output of a product…suggesting that the results are not very comparable. EC3 is using A1-A3 only, and Tally uses all modules. For the Tally output, two EPDs are used: the CRSI one you referenced and one I cannot seem to find a link to (though the data is within Tally). Using the excel outputs from Tally, they average – based on the selections we made on rebar – to 1.48 MT CO2e/MT, simply by dividing the total GWP of the rebar by the mass of the rebar that Tally readily outputs. I am not sure of the algorithms within Tally that modify EPD data to produce this result because a majority of the input rebar data is based on the CSRI which is .979 MTCO2e/MT.
