Freight is the part of highway engineering where the measurement is weakest and the money is largest. A lane is widened, a bridge is rated, a corridor is designated and a rest area is funded on the strength of figures from four unconnected instruments: a statutory network designation, a national flow model, a truck GPS dataset, and a survey series. None was built to talk to the others, and almost every published claim about American freight infrastructure is an unacknowledged splice of at least two of them.
This section works instrument by instrument, on the premise that a specialist arguing for a project is better served by knowing what each dataset can carry than by another summary of what everyone already believes about supply chains.
Why the measurement problem is structural
Highway freight is measured by proxy at every level. The federal freight network is defined by statute with a mileage ceiling, so its extent reflects an administrative allowance as much as traffic. The national flow estimate is benchmarked to a shipper survey taken every five years and pushed forward by a commercial econometric model, so its horizon values are scenarios rather than projections. Truck delay is measured where GPS data happens to be dense. Driver rest is measured by counting parking spaces, because nobody counts drivers who wanted to stop.
Those blind spots are not symmetrical, and they push in a consistent direction. Long-haul movement on Interstates is well observed. Short-haul movement, terminal access and off-peak demand are not. The practical consequence is that the parts of the system with the best evidence attract the most investment, which is reasonable only if the evidence gaps are random. They are not.
What the articles here establish
The most important freight corridors takes the ranking question directly and finds that no federal product publishes an ordered list. It sets out what the primary highway freight system designation covers and the 3 percent redesignation cap that constrains it, why a Freight Analysis Framework tonnage means nothing until the flow set behind it is named, and why a bottleneck ranking of interchanges cannot be converted into a ranking of corridors. The piece ends with a division of labor: which of the three instruments to reach for depending on whether the question is eligibility, forecast or observed delay.
I-35 and U.S. freight is the case study where the state’s own documents are detailed enough to test the general argument. It follows the corridor from the Laredo World Trade Bridge at its southern terminus through the chain of Texas cities, and it stops on two things that get less attention than the border volumes. The state’s corridor plan reports two different truck shares for the same road in the same year, on two different denominators. And a LIDAR survey of 314 bridges along the route found that just over 15 percent meet the vertical clearance the state’s freight plan recommends, which is a limit no amount of widening addresses.
How commercial trucking shapes highway design traces the engineering back to its legal sources. Lane width exceptions are written in terms of truck counts. The pavement damage currency that underlies mixed-traffic design came out of a road test run in a single climate across a narrow range of pavement sections, and the federal agency that commissioned it stopped using the unmodified version for cost allocation in 1979 after the Congressional Budget Office attacked the assumptions. Bridge clearance on the Interstate System was set by negotiation with the Defense Department and design exceptions to it are still coordinated with an army command. Load ratings can be created by statute, as they were for emergency vehicles in 2016.
From port to highway is the strongest single finding in the section. The designated last-mile network is under one percent of National Highway System mileage, and federal analysis of a container trip from Savannah to Atlanta shows those designated miles carrying about 7 percent of the travel time and worth roughly $5 to the driver if their speed halves. The trip where connector speed genuinely governs is local drayage, where the same halving costs a quarter of fleet productivity, and those local roads are frequently not designated, because connectors are drawn to join terminals to the Interstate rather than to distribution centres.
The truck parking shortage examines the one federal measurement effort explicitly ordered to build a metric system, and finds almost everything it built is a supply ratio. Spaces per 100 miles of network, spaces per 100,000 truck miles, spaces against state economic output: none of those counts a truck that wanted to stop. The one row FHWA files under demand measures truck travel on the network, which counts trucks passing. The article also unpacks the 98 percent driver figure that circulates without its distribution.
The live questions
Three recur across the section and none is rhetorical. Nobody has published a corridor-level demand measurement for truck rest, so the shortage remains an inference from displacement. Nothing stops a project justification quoting the higher of the two truck shares a state plan publishes for the same road, even though the same plan carries the reconciling table a few pages later. And nobody funds the terminal access roads the evidence says matter most, because a volume-weighted national condition target makes low-volume, high-truck pavement the least cost-effective thing a state can repair.