Page 347 - Biofuels for a More Sustainable Future

Key issue, challenges, and status quo of models for biofuel supply chain design 303

billion dry tons biomass resources in the United States (Langholtz et al.,
2016). Other examples include the U.S. National Biomass Estimator Library
(NBEL) (Wang, 2014), USDA Bioenergy Statistics (2018), IEA Energy
Access Database (IEA, 2017), USDA Wood2Energy Database (USDA,
2014), and USDA Forest Service Timber Product Output (TPO) database
(US Department of Agriculture Forest Service, 2012). All of those are good
resources to bound uncertainty, and it will be subject to BSC designers to
choose appropriate data sources based on their projects. With respect to
modeling parameter uncertainty, previous studies tried to use stochastic pro-
gramming and fuzzy programming to develop optimization models for BSC
design uncertainty as discussed in previous sections. Most of those models
are computationally intensive; the challenge is to solve those models in a rea-
sonable time with robust and reliable results.
Methodological uncertainties are those brought in by different method-
ological options. For example, when quantifying environmental impacts of
BSC, different allocation methods [e.g., mass allocation versus energy allo-
cation (Lardon et al., 2009; Wiloso et al., 2012)], environmental footprints
[e.g., GHG and water footprints (Bernardi et al., 2013; Yang et al., 2011)],
and LCIA methods [e.g., TRACI, Eco-indicator 99 (Cherubini and Strøm-
man, 2011; Kim and Dale, 2005; Morales et al., 2015; Neupane et al., 2011)]
may lead to different conclusions. It may be hard to completely address
methodological, but it is always helpful to recognize such uncertainty
sources and include sensitivity analysis to test the robustness of the results.
Another challenge is meaningful quantification of different aspects related
to sustainability. For example, job creation is widely used in BSC design
to represent social impacts of BSC. However, there are many other social
implications such as food security, environmental justice, and social welfare
benefits (Dauvergne and Neville, 2010; Bringezu et al., 2009; de Gorter and
Just, 2010). Social LCA has been developed in recent decades to evaluate the
social and socioeconomic impacts of products and their life cycles (UNEP
and SETAC Life Cycle Initiative, 2009). Similar as environmental LCA,
social LCA needs to collect intensive inventory data (e.g., number of work-
ing hours), which is challenging for biofuels that have not been industrially
commercialized in many regions. Even for environmental LCA or techno-
economic analysis that have relatively more developed methodologies and
tools than social LCA, intensive data needs, especially the need of the
process-based inventory data, are always challenging. Some researchers have
used process-based simulation models [e.g., Aspen Plus (You et al., 2012;
Zhang and Wright, 2014; Sukumara et al., 2014; Zhang et al., 2014;