Abstract
In response to growing awareness of climate change, requests to establish product carbon footprints have been increasing. Product carbon footprints are life cycle assessments restricted to just one impact category, global warming. Product carbon footprint studies generate life cycle inventory results, listing the environmental emissions of greenhouse gases from a product’s lifecycle, and characterize these by their global warming potentials, producing product carbon footprints that are commonly communicated as point values. In the present research we show that the uncertainties surrounding these point values necessitate more sophisticated ways of communicating product carbon footprints, using different sizes of catfish (Pangasius spp.) farms in Vietnam as a case study. As most product carbon footprint studies only have a comparative meaning, we used dependent sampling to produce relative results in order to increase the power for identifying environmentally superior products. We therefore argue that product carbon footprints, supported by quantitative uncertainty estimates, should be used to test hypotheses, rather than to provide point value estimates or plain confidence intervals of products’ environmental performance.
Highlights
Enthusiasm about carbon footprinting resulted in the aim of calculating product carbon footprints (PCFs) for whole product assortments [1]
PCFs of identical products can deviate by an order of magnitude between studies, even if they comply with the same methodological guidelines [9]
The fifth IPCC assessment report [52] does, to our knowledge, not specify if the uncertainty estimates in the Global warming potentials (GWPs) of greenhouse gases (GHGs) have been obtained through dependent or independent sampling, but judging the values of the uncertainties, we believe that dependent sampling has been used, as it should have been
Summary
Enthusiasm about carbon footprinting resulted in the aim of calculating product carbon footprints (PCFs) for whole product assortments [1]. PCFs of identical products can deviate by an order of magnitude between studies, even if they comply with the same methodological guidelines [9] This is largely due to data sourcing and modeling assumptions [9,10], but in some cases to different characterization factors used to translate the environmental emissions into impacts [11]. In a Monte Carlo, values are randomly sampled from the unit process distributions over a fixed number of iterations and aggregated into LCA results using an LCA matrix (step 2) This procedure produces a range of possible results, which in turn could be evaluated using different statistical tests and analyses (step 3). While the absolute overall dispersions remain large, we managed to identify significant trend differences between the different farming systems by using our proposed approach
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
