You are here:
How methodological variation affects life cycle water withdrawal estimates for US industry
Citation:
Birney, C., M. Li, AND W. Ingwersen. How methodological variation affects life cycle water withdrawal estimates for US industry. American Center for Life Cycle Assessment (ACLCA), NA, September 22 - 24, 2020.
Impact/Purpose:
The purpose of this talk at the annual conference of American Center for Life Cycle Assessment (ACLCA) is to describe EPA's development of a water input methodology pertaining to the US Environmentally Extended Input-Output (USEEIO) model.
Description:
Water is an essential component in the production and consumption of goods and services. Water demand across all economic industries and by final consumers contributes to stress faced by freshwater resources in many regions of the US and the world. This research examines the impacts of methodological variation on life cycle estimates of direct and indirect water withdrawal for ~70 US goods and services from 2010-2017. In the US, industries are not required to report water withdrawals by any federal mandate, and therefore water withdrawals and use by industry must be estimated using other data sources that provide more aggregate water use estimates or by using non-water use data on industry activity that correlates with industry water use. Differences in industry-level direct water withdrawal are compared by varying datasets used and methods of attribution to industries. Indirect, or embedded, water withdrawal is estimated using USEEIO model versions constructed to incorporate annual variation in US industry structure and output along with these model variations of direct water withdrawal. Using an environmentally extended input-output (EEIO) model like USEEIO traces water use through US economic industries, by linking publicly available economic industry and environmental data (Yang et al., 2017). We perform the modeling by implementing the water attribution models in the Flow Sector Attribution tool and building the various USEEIO models and calculating their results using the USEEIO modeling framework (Ingwersen et al., 2018). All the data used and modeling performed is open source and available in the respective Github repositories for these tools. Data quality assessment is further used to compare reliability, temporal, geographic and technological correlation with the model purpose, and the data collection completeness of the results, which can be useful for interpreting the direct and life cycle water withdrawal estimates derived from the various models. The combination of quantifying the impact of methodological variation and assessing data quality provides insight into the tradeoffs of the water attribution methods and life cycle models used.