A container leaving the Port of Savannah for Atlanta travels 258 miles. Eight of those miles are on the federally designated freight intermodal connectors that exist for exactly this purpose, four at the port end and four at the destination, and FHWA’s breakdown of the trip gives each of those four-mile segments 10 minutes of a 293-minute journey. The designated network handles 7 percent of the trip. Rural Interstate between the two cities handles 75 percent.
That is the awkward finding at the centre of federal work on port access, and it does not mean the connectors are unimportant. It means the connectors matter enormously to a different trip, one the designation was not drawn to serve.
What a connector is, and how few of them there are
Freight intermodal connectors are the roads that link major rail, port, airport and intermodal facilities to the National Highway System, and their extent is the first surprise. Nothing else described across the freight and transportation section is asked to carry so much on so little mileage.
The officially designated network of NHS intermodal connectors accounts for less than one percent of total NHS mileage, but these roads are critical for the timely and reliable movement of freight.
The system has a legislative origin. The National Highway System Designation Act of 1995 directed the Secretary of Transportation to develop a list of intermodal connectors and submit it to Congress; the inventory was completed in 1998 and approved as part of the Transportation Equity Act for the 21st Century, and FHWA added an assessment database in 2009. By the end of 2014 there were 950 designated connectors covering 1,407 miles of roadway and serving 798 freight intermodal facilities. The comparison figure is 616, on a date the federal record does not settle: the study’s text says 1998, while its own table and FHWA’s conditions and performance reporting both date the increase of 182 connectors, or 30 percent, to 2000 through 2014. Port facilities accounted for the largest block of that growth, rising from 252 to 329.
Designation runs on published thresholds. A port qualifies through a terminal handling more than 50,000 twenty-foot equivalent units per year, or through 100 trucks per day in each direction on the principal connecting route; a bulk commodity terminal qualifies at more than 500,000 tons annually by highway. Truck and rail facilities use the same container and truck-count tests, and airports qualify at 100 trucks per day in each direction or 100,000 tons a year moving by highway. Secondary criteria then admit terminals handling more than 20 percent of a state’s freight volume by mode.
Where the haul splits between road and rail
The modal division at a container port is the single most contested number in this subject, and the federal estimate of it is arrived at by subtraction rather than by observation. Containerised traffic at U.S. ports grew from 30 million twenty-foot equivalent units around the time port connectors were first designated to 45 million by the middle of the last decade. The Association of American Railroads reported that approximately 9 million of those units moved by rail. FHWA’s study therefore estimates that roughly 36 million units, about 80 percent, used highway connectors to reach inland destinations.
That subtraction carries the residual of every error in both inputs, and it says nothing about how far the truck went. A unit trucked four miles to a transload warehouse and a unit trucked 900 miles to a distribution centre count identically.
Rail’s own intermodal record over the same period shows growth in units and a change in kind. Containers and trailers moved by the railroads rose from about 6 million a year in 2000 to nearly 13 million in 2013, while trailers fell from half of total intermodal flows to around 15 percent. Domestic intermodal became a container business, which changes the geometry of the ramp and the drayage pattern around it.
Savannah, split into its two real trips
Georgia’s largest container facility gives the clearest published picture of what a port actually sends onto the highway. In 2013 the Port of Savannah’s annual volume exceeded 3 million twenty-foot equivalent units. Between 20 and 25 percent of imported containers left the facility by rail and the remainder went by truck, producing nearly 5,000 truck trips in each direction per day. Gate surveys then split those trucks into two populations: those carrying goods directly out of the Savannah region, and those making multiple drays a day between the port and local distribution centres. The surveys indicate the second group is two-thirds of all trucks through the gates, roughly 3,300 a day.
For the first group the connector barely registers. Atlanta is 250 miles away in the study’s text and 258 by the total of its own table, and the trip breaks into five components: four miles of designated connector at 25 mph, 20 miles of urban Interstate in the Savannah region, 200 rural Interstate miles taking 218 minutes, 30 urban Interstate miles in the Atlanta region, and four final miles partly on connectors. Average speed for the whole trip is 53 mph. Halve the connector speeds, which the study treats as severely congested conditions, and total travel time rises by 20 minutes or 7 percent, worth roughly $5 against the federal truck value of time. The driver’s net income for the trip is around $80.
For the second group the same arithmetic inverts. Drayage between the port and local warehouses runs almost entirely on local roads, and if speeds halve on half of that travel, the study calculates that drayage productivity falls by 25 percent.
Those local roads do not neatly align with the designated connectors, because connectors are designated to join intermodal facilities to the National Highway System rather than to join intermodal facilities to distribution centres. The study names a case: trucks leaving the port for I-16 may use SR 307, Bourne Avenue, which is not a designated intermodal connector. The road carrying the sensitive trip is outside the network built for port access.
The condition record, and the goal that works against it
Pavement on connectors is measurably worse than pavement on comparable roads, and the pattern inside the data is not what a casual reader would guess. Across 1,239 connectors with condition data, 37 percent rated poor on the international roughness index, above 220 inches per mile, 19 percent mediocre, 35 percent fair, 8 percent good and 1 percent very good, for an average of 211.
Sorted by length, the shortest connectors are the roughest. The 862 connectors under a mile long average an index of 233; those of 10 miles and more average 96. By mode, port connectors are the worst at an average of 230, which the study classes as poor, against 107 for rail connectors and 168 for air cargo. In the case study sample, connectors averaged 196 against 128 for roads of the same functional class in the same state, with 78 percent of them rougher than their local comparators and an average difference of 55 percent.
The explanation FHWA offers for that gap is a funding mechanism rather than neglect. The department set a national performance goal for 2013 of having 57 percent of vehicle miles traveled on the National Highway System run on pavements with good ride quality. Because the goal is weighted by traffic volume, agencies reach it most cheaply by lifting high-volume fair roads to good. Connectors carry low total volumes and high truck percentages, so they deteriorate faster and score worse on a volume-weighted measure, and moving a road from poor to good costs far more than moving one from fair to good. A state acting rationally against the published target will fix the arterial and defer the connector.
The study reports the wider trend that makes this concrete: between 2000 and 2010 the share of vehicle miles on roads in good condition improved even as the length of roads in good condition fell from 43 percent to 35 percent. Three amendments are proposed, unusually directly for a federal report: cap the percentage of roadway allowed in poor condition, weight the 57 percent goal by passenger car equivalent vehicle miles, and set connector-specific condition goals in each state. How states make that ranking is examined in how state DOTs choose which roads to rebuild.
What the money says, and one thing it does not
Bringing every designated connector to good condition would cost approximately $2.2 billion across 3,150 lane miles, with 78 percent of that concentrated in urban arterials and 48 percent attributable to port connectors alone. Delay is priced separately. Of 901 connectors with speed data, 67 percent fall below 90 percent of free-flow speed in at least one period of the day. Truck delay runs just over 4,000 hours every weekday, which annualises to 1,059,238 truck hours and 12,181,234 auto hours, and at federal values of travel time that came to $353 million in a single year, or $6.4 billion cumulatively over 30 years at a 4 percent discount rate. Port connectors carry 37 percent of the delay cost.
Against those figures the study includes a caveat that most advocacy on this subject omits. Poor pavement condition on connectors does not appear to damage freight facility operations. Drivers do not connect vehicle wear to particular connectors, and carriers do not charge more to serve customers on rough roads, because the maintenance cost is spread across the whole customer base. The externality is real and it is invisible at the point of sale, which is precisely why it does not generate the political pressure its size would suggest.
Capacity is the cheaper lever where it applies. State construction estimates put an added lane at about $2 million per mile, rising to $4 million to $6 million per lane-mile once environmental work and right-of-way are included, and on those numbers $3.2 billion would add capacity to 180 of the 540 most congested connectors. The comparison with mainline construction costs is drawn out in what a mile of Interstate costs.
None of that resolves the designation problem. A network drawn to connect terminals to the Interstate will keep performing well on the leg where connector speed is worth $5 and keep missing the leg where it is worth a quarter of a drayage fleet’s productivity. How corridor importance gets defined is the subject of the most important freight corridors, and the vehicle characteristics that wear these roads out are set out in how trucking shapes highway design. Fixing the pavement on the wrong 1,407 miles is a cheaper mistake than most, and it is still the wrong 1,407 miles.