Evaluation of environmental and cost tradeoffs of producing energy from soybeans for on-farm use

Authors: Adler, Paul; HUMS, MEGAN - Drexel University; McNeal, Fred; SPATARI, SABRINA - Drexel University

Submitted to: Journal of Cleaner Production
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 11/2/2018
Publication Date: 11/7/2018
Citation: Adler, P.R., Hums, M., Mcneal, F.M., Spatari, S. 2018. Evaluation of environmental and cost tradeoffs of producing energy from soybeans for on-farm use. Journal of Cleaner Production. 210:1635-1649. https://doi.org/10.1016/j.jclepro.2018.11.019.
DOI: https://doi.org/10.1016/j.jclepro.2018.11.019

Interpretive Summary: There are many opportunities to reduce nitrous oxide emissions, however the methods to estimate them have been thought to be too uncertain for these reductions to be accepted by carbon markets. We evaluate a margin of safety approach to increase the certainty of nitrous oxide emission estimates. We found that uncertainty of nitrous oxide emissions could be reduced to a low value using this approach. Adoption of this approach could increase opportunities to incorporate incentives for nitrous oxide mitigation strategies into policy frameworks.

Technical Abstract: Entity level estimations of N2O emissions are not currently incorporated into regulatory frameworks due to high levels of uncertainty. However, incorporation of entity level N2O emissions into the life cycle inventory could incentivize adoption of mitigation strategies into crop management practices. Aggregation across space has been suggested as a method to reduce uncertainty. However, it is not clear how much uncertainty is appropriately characterized as random errors that can be averaged out through aggregation. Therefore, our objectives were to clearly differentiate random errors from persistent structural uncertainty remaining even in a calibrated model and evaluate how different spatio-temporal aggregation procedures perform given under this approach. A mixed effects statistical model is used to calibrate a mechanistic biogeochemical model of nitrous oxide emissions from biofuel feedstock production within a consequential framework. Monte Carlo simulation were used to assess uncertainty in predicted emissions for scenarios of varying temporal and spatial scale to assess the potential of alternative aggregation strategies to reduce uncertainty. We show that either spatial or temporal aggregation of N2O emissions reduces the effect of random errors. The proposed framework in which uncertainty is retrospectively aggregated has promise to 1) avoid penalizing producers for uncertainty in weather at a specific time and place, 2) allow for a high margin of safety in emissions estimates, 3) effectively attenuate the uncertainty penalties borne by producers within a timescale of several years. Regardless of the uncertainty aggregation approach used, uncertainty in model calibration parameters results in an uncertainty that increases the carbon intensity of a biofuel at the 95% upper bound by 1-2 g CO2/MJ over model median predictions of roughly 3-4 CO2/MJ. With effective management it appears that the total contribution of nitrous oxide emissions from feedstock cultivation may be less than 10% of the carbon intensity of fossil fuels. Further research and model improvements offer the prospect of reducing structural uncertainty by improving parameter estimated in the calibration model; however, the uncertainty is small relative to the lost opportunity to incorporate incentives for mitigation strategies into policy frameworks.