Faf5 Freight Flows By Commodity Data: Dataset Structure and Field Coverage

The Freight Analysis Framework Version 5 (FAF5) is the United States' complete origin-destination freight database, produced by the Bureau of Transportation Statistics with support from the Federal Highway Administration and developed at Oak Ridge National Laboratory. The current release, FAF5.7.1, carries three measures for every flow: tons in thousand tons, value in million dollars (2017 constant dollars, with a current-dollar column alongside), and tmiles in million ton-miles.

Temporal coverage spans a 2017 base year, recent years 2018-2024, mid-range forecasts for 2030-2050 in five-year steps (high and low growth scenarios ship in expanded files), and reprocessed state-level annual history for 1997-2012.

The commodity dimension never travels alone - FAF5 crosses sctg2 against four other axes in every release.

Mode. dms_mode distinguishes eight domestic modes: 1 Truck, 2 Rail, 3 Water, 4 Air (includes truck-air), 5 Multiple modes & mail, 6 Pipeline, 7 Other/unknown and 8 No domestic mode. Imports and exports carry their own boundary modes in fr_inmode and fr_outmode, so you can isolate, say, grain arriving by water and leaving by rail.

Trade type. trade_type separates 1 domestic flows, 2 import flows and 3 export flows, which means a single commodity query can return the domestic, inbound and outbound stories side by side.

Distance. dist_band bins each flow by weighted average shipment distance, from below 100 miles to over 2,000 miles - the dimension that turns raw tonnage into corridor analysis.

Together those axes make FAF5 the standard source for origin-destination freight flows in US ground transportation: the same commodity sliced by truck-only movements, by metro-pair corridors, or by cross-border gateways.

A repeatable recipe for a single-SCTG zone-pair table, using nothing but the published files:

1. Load the CSV variant rather than the Access file if you are scripting - the columns match the metadata XLSX data dictionary exactly.

1. Filter on sctg2. Select your 2-digit code (for example 35 for the example group used throughout the data dictionary, or 43 for mixed freight) before any aggregation, because rows are unique at zone-pair x commodity x mode x trade-type x year granularity.

1. Choose your trade types. Sum trade_type 1-3 for total flows, or keep them separate when a commodity's domestic and import stories diverge.

1. Handle the three measures correctly: tons arrive in thousand tons, value in million 2017 constant dollars (current_value holds each year's own dollars), tmiles in million ton-miles. Never mix bases.

1. Join geography through the metadata workbook, whose State and FAF Zone sheets resolve region codes to names, and cite the release number (FAF5.7.1) in any output.

They are related but not interchangeable. The Commodity Flow Survey (CFS) 2022 is the observed foundation: a joint Census Bureau and BTS survey of US establishments, run every five years, whose January 2026 PUMS microdata file holds 37,576,546 shipment records with origin, destination, SCTG commodity, mode (distinguishing for-hire from private truck), value, weight, great-circle distance and a tabulation weight - shipped as a 680 MB zipped CSV.

FAF5 integrates that survey with Census international trade data and sector-specific sources, then fills the gaps with modeling. Its own terms notes are blunt about the result: FAF figures are model-based estimates, not observations, and should be cited as such. In exchange for that caveat you get annual recency (2018-2024 versus a five-year survey cycle), full metro-to-metro coverage beyond CFS sampling areas, imports and exports that the domestic CFS does not measure, and forecasts out to 2050.

The practical rule: validate a commodity's structure against CFS microdata, then scale and project it with FAF5. The head-to-head breakdown lives on the Commodity Flow Survey (CFS) 2022 vs Freight Analysis Framework Version 5 (FAF5) comparison, and the shipment-level mechanics behind the survey are in the Commodity Flow Survey microdata walkthrough.

Three cataloged sources extend commodity flows toward the physical network and the market:

• FHWA Freight Performance Measurement Program - NPMRDS-based truck probe travel times, reliability and bottleneck delay across the National Highway System at TMC-segment level, refreshed monthly. Raw probe data requires a RITIS portal account extended to State DOTs, MPOs and registered researchers; the published Freight Mobility Trends dashboards are open. Assigning FAF5 truck flows onto highways is exactly what FHWA's own truck-assignment files (529-713 MB per year) already do for 2017, 2022 and 2050.

• USDA Agricultural Truck Rate Report (AgTransport) - 206,201 weekly refrigerated truck-rate rows since 2000-01-04, served as CSV and through an open Socrata SODA API with no registration. It prices the lanes that FAF5 volumes describe.

• Data.gov Freight Catalog - a JSON API over 552,271 federal dataset records surfacing roughly 130 cargo, freight, truck and motor-carrier titles, including BTS transborder flows for the border detail FAF foreign regions summarize.