US Census Browser V2.0
US Census Browser v2.0
USEEIO v2.0 is an environmental-economic model of US goods and services that can be used for life cycle assessment, footprinting, national prioritization, and related applications. This paper describes the development of the model and accompanies the release of a full model dataset as well as various supporting datasets of national environmental totals by US industry. Novel methodological elements since USEEIO v1 models include waste sector disaggregation, final demand vectors for US consumption and production, a domestic form of the model that can be used to separate domestic and foreign impacts, and price adjustment matrices for converting outputs to purchaser price and in various US dollar years. Improvements in modeling national totals of industry and environmental flows are described. The model is validated through reproduction of national totals from input data sources and through analysis of changes from the most recent complete USEEIO model that can be explained based on data updates or method changes. The model datasets can all be reproduced with open source software packages.
USEEIO v2.0, or referred to solely as v2.0, is the latest edition of the US Environmentally-Extended Input-Output (USEEIO) model for assessing a full suite of potential life cycle impacts of US goods and services. It is the first model version since USEEIO v1.2 capable of calculating potential environmental impacts, resource use and waste generation along with economic impacts, and builds upon the creation of the USEEIO v2 GHG models, which were a series of USEEIO models used to calculate Supply Chain Greenhouse Gas Emission Factors1. This paper presents a summary of the complete v2.0 model attributes and model creation with a focus on describing methodological updates since the publication of the original USEEIO methodology.
National totals of flows (physical movements of specific resources, emissions or employment) by industries are used as the sources of environmental and employment data. Coverage of these data used in v2.0 is equivalent to that from v1.2 as seen in Table 2. These include all the types of resource use and environmental releases/losses from v1.15 plus the three additional waste generation datasets created for v1.27,8. The data for water withdrawals, criteria and hazardous air emissions, point source industrial releases to ground, point source releases to water, greenhouse gases, land use, employment, and value added have been updated in v2.0 and incorporate methodological improvements. The breakdown of these data into the records given in Table 2 is not identical to that given in Table 1 of the original USEEIO description2, but all these data are aggregated during model construction (see Model Construction section), and therefore the breakdown just describes the form of these data as they are originally processed and imported. Changes in selection of data sources and methodologies for compiling these into a standard format are described below. New procedures for preparing and integrating these datasets into the model are described in the Procedure for Model Building section.
The model indicators quantitatively relate the environmental and economic flow data to an aggregate impact through the use of characterization factors. These include values such as the carbon dioxide equivalencies of the flows that are greenhouse gases. The same indicators used in v1.113 along with the three indicators (CNHW, CNHWC, CRHW) for waste generation8 are used in v2.0 (Table 3).
New for v2.0 is the use of a standard flow list for representing elementary flows, or raw materials from or returning to the technosphere. The Federal LCA Commons Elementary Flow List (FEDEFL) v1.0.7 is used to represent the substance, environmental compartment or origin or release, and the unit in a common format with Federal LCA data14,15. Additionally, to support the use of the FEDEFL for the new environmental data, flows used in the indicators also were updated to correspond to the FEDEFL. This flow update afforded the opportunity to use more standard life cycle impact assessment (LCIA) characterization factors to populate these indicators, which were integrated with the new procedure. More information about this update is provided in the Procedure for Model Building section.
For v2.0, we derive two primary final demand vectors, a production vector and a consumption vector. We define consumption as final use within the US of all goods and services that are both produced and sold within the US or imported. We define production as final use, either within the US or abroad, of all goods and services that are produced in the US.
In v1, the Scrap commodity was removed from the model following a methodology presented by BEA for deriving a total requirements matrix11. In v2.0, Scrap is left in the model to simplify the accounting procedures, but we do not recommend use of multipliers generated from Scrap because of the lack of a clear material or functional characterization of this commodity.
The BEA IO sector codes are based on the North American Industrial Classification System (NAICS). The three zeroes at the end of the BEA code for Waste management and remediation services indicate that it is at the 3-digit NAICS level. The approach used to disaggregate this sector provides 6-digit NAICS granularity, which is the most detailed NAICS designation given in the official classification. For v2.0, Waste management and remediation services is disaggregated into the seven sectors shown in Table 5.
The US Economic Census (EC), published by the US Census Bureau, provides economic data for all sectors of the US economy and is used to estimate industry consumption of the disaggregated waste management commodities (i.e., Use table rows)20. This dataset tracks the monetary receipts by the different waste management subsectors that broadly correspond to the disaggregated sectors being introduced to the v2.0 model.
Sector correspondence between the BEA and NAICS codes, or the Sector Crosswalk, is created to connect the two classification systems and enable mapping from one system to the other. For v2.0, the Sector Crosswalk is built based on 2012 BEA and NAICS codes and includes 2007 NAICS codes according to the 2012 NAICS to 2007 NAICS concordance by Census Bureau23. The Sector Crosswalk is available as part of the primary data record24.
v2.0 builds on the flow sector attribution modeling approach taken to construct national commercial waste totals8, by estimating totals by industries defined by NAICS codes. For v2.0, national totals by sector are modeled by NAICS 6-digit codes. This approach is in contrast to the former method used in v1.1 in which flows were directly attributed to IO table industries. The modeling steps were written in Python and consolidated into a software package called flowsa. flowsa performs extraction, transforming and loading (ETL) processes for bring in original data sources, as well as the modeling steps to create a standard flow-by-sector output, which can be retrieved using the getFlowBySector function and passing the name of the flow-by-sector of interest. flowsa v1.0.126 was used for preparation of all original environmental inputs.
Increase in nonpoint emissions for manufacturing sectors. In past models, nonpoint air emissions from industrial combustion were not mapped to sectors due to insufficient data. In v2.0, these emissions have been allocated to manufacturing sectors on the basis of fuel consumption by fuel type30.
Chemical releases to water are sourced from 2017 facility reported emissions data from the Toxic Release Inventory (TRI)29 and the Discharge Monitoring Report (DMR)33. Chemical releases reported by facilities in these datasets include toxic releases, metal compounds, nutrients, and organic pollutants. Emissions are assigned to industries based on the NAICS reported by each facility to the dataset. Where particular elementary flows are reported in each dataset, flows are maintained from the DMR when a facility reports to both. In v1.1, releases from the DMR were limited to nutrient release of nitrogen and phosphorous. However, in v2.0, releases to water also include organic enrichment, sediments, and other compounds tracked within the DMR. The result is available in the National Point Source Releases to Water By Industry 2017 v1.1 dataset34.
The employment sector attribution model is created by importing and formatting the 2017 BLS Quarterly Census of Employment and Wages (QCEW) table58. QCEW publishes national annual employment at the 6-digit NAICS; no additional allocation is required for use in v2.0. Where the BLS is missing data, less aggregated NAICS are summed, or more aggregated NAICS are equally allocated. The model is available as the National Employment Totals By Industry 2017 v1.1 dataset59. The earlier versions of USEEIO used the BLS National Employment Matrix, which also publishes employment data by NAICS sector codes5. QCEW was chosen for the sector attribution model, as QCEW data is one of the primary data sources for the National Employment Matrix and as the National Employment Matrix database primary purpose is for national-level employment predictions60. Additionally, QCEW publishes state and county employment data used in other sector attribution models used in USEEIO v2.0. Using BLS QCEW for the employment model allows for a consistent data source for all employment data used throughout model construction.
Value added is a collection of the monetary benefits industries provide to government (as taxes), employees (as wages), and to their shareholders (as profits). For v2.0, total value added per industry is taken directly from the same 2012 BEA Use table that is a source for the economic data. Although it is economic data, the handling of it is identical to that of the environmental and employment data used to construct the satellite tables. This is an update from v1.1, where value added data were taken from BEA Summary level Use tables for more recent years and adjusted as described in the documentation5. 350c69d7ab