A regional version of a US economic input-output life-cycle assessment model

Abstract

Life cycle assessment models typically use product-specific, plant-level or national aggregate data. However, many decisions by regional policy makers would be better informed by local or regional aggregate data. This research is intended to construct and apply a regional US economic input-output analysis-based life cycle assessment (REIO-LCA) model based upon publicly available datasets. The model uses Gross State Product (GSP) estimates to calculate regional economic multipliers and then link them to regional electricity and fuel use, and air emission factors. Target audiences are governmental decision makers, industry experts and researchers concerned with the regional economic and environmental effects of public and private decisions. A regional version of the existing US EIO-LCA model was developed using regional economic multipliers and state environmental data. The national model is based on the US 491 by 491 economic input-output model, and uses sectoral energy consumption and emission factors to approximate the environmental effects of production and services. The proportion of the regional value added (Gross State Product) to the national value added for each sector was used to develop economic multipliers to allocate the output of industries to individual states and multi-state regions. Inter-sectoral transaction matrices were constructed for eight regions. Regional environmental emission and resource use factors were formed based upon publicly available data of the US Environmental Protection Agency (EPA) and Department of Energy. The Toxics Release Inventory include facility location parameters, enabling the estimation of sectoral toxic emissions for the regions. The national electricity and fuel use, air pollutants (CO, NOx, PM10, SO2 and VOC) and greenhouse gas emissions used by the EIO-LCA model were proportioned based upon state totals for each sector. A regional economic input-output model was created for US regions, and sectoral energy use and environmental emission factors were estimated for Pennsylvania, the Far West (Alaska, California, Hawaii, Nevada, Oregon and Washington) and the Mideast (Delaware, District of Columbia, Maryland, New Jersey, New York and Pennsylvania) economic areas. The use of the framework for regional IO-LCA model is demonstrated through two case studies. As a validation exercise, the regional outputs of petroleum refineries were calculated using the regional input-output matrices and the outcomes were compared to the Energy Information Administration’s (EIA) Petroleum State Profile data. The model results show that approximately 70% of the total national sectoral production takes place in three regions, i.e., South West, South East and Far West, which corresponds with the EIA statistics. The REIO-LCA model constructed for the Far West is used to conduct a second case study estimating the annual toxic air emissions of power plants in the region in 2003. The results are evaluated by comparison to data provided by the US EPA. The estimated pollutions do not differ significantly from those presented in the Toxics Release Inventory reports. The usefulness of IO LCA models can be improved through the incorporation of local economic and environmental characteristics. Wiht the lack of US regional sectoral data, the allocation of national industrial production to regions can provide a framework to create smaller scale IO models. The results of case studies support the assumption that the GSP multipliers may be used to allocate the sectoral production to the regions, and show that the framework IO LCA model provides a reasonable approximation of supply chain economic activities and environmental effects caused by production and services. The quality of data, e.g., age and level of aggregation, and the assumed linearity between sectoral outputs and environmental emissions represent the main sources of uncertainty in the model. The results show that the GSP estimates are appropriate to construct a framework for a regional economic input-output and environmental assessment model. However, further research is recommended to construct more specific state-level input-output matrices incorporating interstate commodity flows, and state environmental factors in order to mitigate the parameter uncertainties. Further, the model might be improved by updating it regularly, as more recent data become available.