What are the current challenges and limitations of LCA impact assessment methods and tools?

Data scarcity and inappropriate data selection may lead to model very similar systems, when in reality they aren't. That's to say: a barely studied system may have a couple of data points that could be used as a reference by many researchers, therefore modeling a very similar system, when in reality each one of them are assessing a different one.

Data availability, findability and compatibility is easily the biggest challenge for LCA practitioners at the moment. When selecting background data, it's not uncommon to make compromises or to develop simplified models when data is missing, which overall lowers the confidence in the results. Using reputed sources that give a complete picture is a good start, since you can validate how close it represents the system that you're modelling. But in a short term, LCA needs to move towards more specific data (bottom-up), with different parties having connected data that they use in each others' models.

To enhance data reliability, consider employing multiple sources and verification methods. Diversifying data sources helps mitigate the risk of inaccuracies or biases from a single origin. Collaboration with industry experts, academic institutions, and relevant stakeholders can provide valuable insights and contribute to data accuracy.

The main data challenges I've encountered are: the cheer scarcity of data scarcity, lack of location-based specific data, non-public data (read: behind a paywall), outdated data, inaccuracies between datasets, ...

Conduct a comprehensive review of existing literature and best practices in your field. Learning from the experiences of similar studies or projects can offer insights into commonly accepted impact categories and indicators. Leverage established frameworks, guidelines, and standards to align your choices with widely recognized principles. Furthermore, Involve stakeholders in the decision-making process. Engaging with experts, industry representatives, and potential affected parties can provide diverse perspectives and valuable insights. This collaborative approach helps identify the most pertinent impact categories and indicators based on a collective understanding of the environmental issues at hand.

It's worth commending the role that EPDs and PEF are having in the selection of impact categories. For most studies, the selection of impact categories depends on the final purpose, so having a clear list makes the process much clearer.

For instance, stakeholders may express a higher concern for climate change impacts compared to resource depletion. Weighting is typically expressed as a factor (e.g., 2 for high importance, 1 for medium, and 0.5 for low).

I find that normalization and weighting, unless provided directly by an impact assessment method, is often a source of confusion. Most decision-makers will not have the maturity to decide the weight of different parameters, and it makes results non-standard by definition. Of course, when they're part of an impact assessment method (e.g. PEF or Ecocosts), it's a completely different story, since there's a third party setting those weights, and it just becomes a way to grant a single score.

While the uncertainty due to the LCI data is recognized, the uncertainty from the impact assessments and the methodology is largely not talked about. Any results should at minimum have ranges based on the variances in the emission factor values used for impact assessments.

Most tools (ie. OpenLCA, Sima Pro, Gabi, etc.) out there today are "made by engineers, for engineers". Completely not user-friendly, and impossible to incorporate into the strategic decision-making level of a company. I truly believe that LCA should be one of those capabilities that should be front and center at the decision-making table. Therefore, I actively look for tools with: (i) good UX, (ii) collaborative environment, (iii) good visualization, (iv) based on process flows, (v) ability to extract data easily, (vi) allows for scenario analysis, (vii) should have the ability to be leveraged at product design, etc. The only one I've come across today is Digit Mint.

To extend sustainability within a company, the subject of LCA should not be reserved for experts, but should be extended to as many people as possible. The tools need to be intuitive, easy to use and the results understandable by a human being. We also need to be properly equipped to scale up, because it's one thing to carry out an LCA on a product, but extending it to a portfolio is quite another. New products arrive on the market allowing this.

It is crucial to ensure that the LCA data we use is representative. We also need to use recognised benchmarks when assessing our processes and systems. These, unfortunately, not always exist and it may be challenging to demonstrate the environmental benefits of the system we consider.

? Adoption of consequential LCA seems to be limited. There are no LCA softwares where you can do consequential modeling without using previous modeled dataset. Currently, the LCA practitioners need to use pre-modeled datasets for consequential models. For example, if iLUC impacts need to be considered, then the values from literature need to be used, which were modeled using GTAP-AEZ, GLOBIOM or such. No way from LCA practitioner to model these.

Good selection of functional unit is known and a foundation for LCA studies. However, functional units are currently selected on only on functional basis. They don’t include scale factor. The impacts and consequences vary with scale. Currently, LCA models assume linear modeling. In order to operationalize use of LCA to design, procurement and operation aspects of asset life cycle, LCA models need to build in scale factors. Functional units that go beyond per unit of energy or per unit of function could be one solution. It will also help capture the byproducts aspects and resource availability aspects.