Reliability of land-related environmental footprint of products
The use of proxy or extrapolated data sets to bridge data gaps, often without quantifying the associated uncertainty introduced, undermines the reliability of product environmental footprints given the spatial, technological and temporal variability of data sources. The goal of this project is to improve the reliability of greenhouse gas (GHG) footprints of agro-based products on a global scale using methodologies and models developed at the crop production phase for both annual and perennial crops. For annual crops, our framework hypothesizes that farm-specific greenhouse gas emissions are related to so-called predictors such as location of farm (eco-region), farm management practice, gross domestic product of country of farm residence, soil conditions and crop type. Using fruit and vegetable crop data, we calibrated and optimized regression models for the prediction of farm-specific greenhouse gas footprints, along with their associated uncertainties, with predictors representing data that are commonly available, hence investigating the possibility of using a data-driven approach to develop a streamlined LCA model. For perennial crops, biogenic emissions from land use change strongly depend on the data used for spatial assessments, such as the land cover maps and soil carbon data. Using publicly available spatial data, we quantify the impact of data variability on the uncertainty of spatially-explicit land use change assessments using InVEST models. Finally, we develop a model by combining the results from the crop level with data from the processing and manufacturing phases to fill life-cycle inventory data gaps of the GHG footprint of agro-based products in a consumer basket. The overall results serve to enable LCA practitioners to prioritize data collection efforts according to the tolerated uncertainty for the application of the results.