Rail terminal dwell time is how long a railcar sits in a yard between arriving on one train and leaving on another. It is the single biggest reason a car that runs fine on the main line still shows up late. A loaded car spends most of a manifest move waiting in yards, not rolling, and every yard on the route adds another wait.
This guide covers how railroads measure dwell, how to read the numbers they file with the Surface Transportation Board every week, why cars sit, and what you can change from the shipper's side of the trip.
What Is Rail Terminal Dwell Time?
Rail terminal dwell time is the average number of hours a railcar spends at a terminal before it leaves on an outbound train. The STB's rules define it as "the average time a car resides at a specified terminal location expressed in hours," and they exclude cars on run-through trains, meaning cars that arrive and depart on the same train.
That exclusion matters. Dwell measures the cars that actually get worked: pulled off one train, sorted into a block and put on another. For a manifest car, that is the normal way to travel. A car going from a plant served by a local to a customer on a different part of the network might be handled at the local's serving yard, a large classification yard, an interchange yard and another serving yard near the destination. Each one is a dwell event.
We cover the mechanics of sorting in how railroad classification yards work. This post is about the clock: what drives it and what you can do about it.
How Dwell Is Measured and Reported
Every Class I railroad must file weekly performance data with the STB under 49 CFR Part 1250, commonly called the EP 724 data. The reporting week runs Saturday through Friday, filings are due by 5 p.m. Eastern on Wednesday, and the Board posts them publicly. Three of the data elements tell you about cars sitting still.
| EP 724 data element | What it measures | What it tells a shipper |
|---|---|---|
| Terminal dwell, 49 CFR 1250.2(a)(2) | Weekly average hours a car resides at a terminal, for the whole system and the railroad's 10 largest terminals by cars processed. Run-through cars excluded. | How long a handled car waits in the yard, and which big terminals are slow this week |
| Origin dwell for unit trains, 1250.2(a)(4) | Time from release of a unit train at origin until the receiving carrier actually moves it, by train type (grain, coal, crude, ethanol, automotive, other unit) | How long a released unit train waits for the railroad to pick it up |
| Cars not moved in 48+ hours, 1250.2(a)(6) | Daily snapshot of loaded and empty cars in normal service that have not moved in 48 hours or more, by service type (grain, coal, chemicals, fertilizer and others) | How many cars are stuck on the network, not at customer sites |
Read the 48-hour count closely. It leaves out cars placed at a customer facility, in constructive placement, placed for interchange to another carrier, in bad order status, in storage, or in railroad service. So a car stuck waiting at your plant, or one handed off at a junction, doesn't show up there. The count is the railroad's own network backlog, which is why it is a useful warning sign.
Two cautions. First, each railroad files its own methodology, so compare a carrier against its own history more than against another railroad. Second, the system number is an average across every handled car. Your car can do much better or much worse than that average.
Reading the Weekly Numbers
For the week ending October 17, 2026, the Class I railroads reported these system averages in their STB filings:
| Railroad | System terminal dwell (hours) | System train speed (mph) |
|---|---|---|
| CN | 17.3 | Not available |
| CPKC | 19.8 | 22.4 |
| UP | 19.9 | 24.5 |
| CSX | 20.8 | 24.1 |
| BNSF | 22.5 | 25.0 |
| NS | 25.3 | 19.9 |
A spread of roughly eight hours between railroads looks small until you multiply it. A car that is worked at four terminals on a 25-hour railroad sits in yards for about 100 hours on average. The same four touches at 18 hours each come to 72. That is more than a day of difference before the car turns a wheel on the main line, and an average week can hide a bad one at a single terminal on your route.
The trend matters more than any one week. A railroad whose dwell climbs three or four weeks running is usually short on crews or power, or working through a weather event or a traffic surge, and service on manifest lanes tends to slip before the railroad says so. We publish these numbers every week in Rail Pulse, with the terminal detail when it moves.
Why Railcars Sit in Yards
A railcar sits in a yard because it is waiting for something: the train that carries its block, a crew, a locomotive, clean billing, a repair or a handoff to another railroad. Most dwell comes from a few predictable causes.
Waiting for the right train
Cars move in blocks, groups headed to the same next yard. A car can only leave on a train that carries its block. If that train runs once a day, a car that arrives an hour after it leaves waits almost a full day. If it runs a few days a week, the wait can be longer. This is the biggest built-in source of dwell, and it comes from the railroad's operating plan, not from anything going wrong.
Crew and power shortages
A train that is ready but has no crew or locomotives doesn't go. The EP 724 filings report the average number of trains held per day by cause: crew, locomotive power or other. When the "held" count rises, dwell usually follows within a week or two.
Congestion and volume surges
Yards have a fixed number of tracks. When volume spikes, from a harvest, a weather recovery or a big customer ramping up, cars back up on arrival tracks waiting to be sorted. Congestion feeds on itself: a full yard sorts more slowly, which keeps it full.
Billing problems
A car without good billing has nowhere to go. A missing or wrong waybill, a bad destination code, or a route the railroads can't use leaves the car held until someone fixes it. These are among the most avoidable delays in rail, and they fall right on the shipper. The rail freight documentation guide covers what has to be on the bill of lading.
Interchange handoffs
Every time a car changes railroads it changes operating plans. The receiving railroad picks it up on its own schedule, and the car has to make that railroad's next train for its block. Two railroads can each run on time and still leave your car sitting at the junction for a day.
Bad orders and holds
A car with a mechanical defect is pulled for repair, and a car under a hold or embargo waits for it to clear. The STB's 48-hour stalled-car count leaves bad order cars out, so a weekly report can look healthy while your car sits on a repair track. To you it looks like a long dwell. Our bad order railcar guide covers what to do when that happens.
Dwell vs Demurrage vs Transit Time
Terminal dwell is time the railroad holds your car in its yards. Demurrage is time you hold the railroad's car (or a car on its tracks) past the free time. Transit time is the whole trip, and it includes both. People blur them, and the difference decides who pays.
| Terminal dwell | Demurrage | Transit time | |
|---|---|---|---|
| Where the car is | In a railroad yard | At your facility, or held for you in constructive placement | Origin to destination |
| Who controls it | The railroad's operating plan | The shipper or receiver | Both |
| Who pays for it | Nobody bills you, but you pay in time and inventory | You, under the railroad's demurrage tariff | Built into your rate and your planning |
| How you see it | Car events (arrive, depart) in tracking, EP 724 data | Monthly demurrage invoice | Shipment tracking from release to placement |
One link between them: when a railroad has high dwell, cars often arrive in bunches. Five cars that should have shown up over a week land on the same day, your crew can only unload two, and the rest start running demurrage. A slow yard upstream becomes your bill downstream. The rail demurrage fees guide covers how free time and charges work, and the demurrage course module goes deeper on managing bunching.
What Shippers Can Do About Dwell
Shippers can't run the railroad's yards, but they control the inputs that decide whether a car makes its first train and how many yards it touches. These are the steps that make a measurable difference.
- Release on time, with clean billing. Find out when your local crew works your track and release cars before that cutoff with a complete, correct bill of lading. A car released an hour late can miss the local, then miss the outbound train it was supposed to connect to.
- Learn your local's service days. Ask the railroad which days your industry is switched. Loading to that schedule beats loading as fast as possible and having cars sit on your track, or in the serving yard, until the next pull.
- Get the route right. Confirm the routing and the interchange point on the bill. A wrong junction or a route the receiving carrier doesn't use creates a hold that only gets fixed when someone notices.
- Cut yard touches where volume allows. Multi-car blocks, gathered cars, unit trains and single-line routes all reduce the number of times a car is sorted. See manifest vs unit trains. If your plant is served by one railroad and the destination is on another, compare reciprocal switching and direct service for the extra handoff it adds.
- Avoid bad orders you can prevent. Load within the car's limits, secure doors and gates, and inspect the car before release. A car pulled for a defect loses days, not hours.
- Track by events, and escalate at 48 hours. Watch the arrive and depart events, not just the ETA. If a car hasn't moved in a yard for 48 hours, the same threshold the STB uses, call the railroad's customer service with the car number and ask what it is waiting for.
- Watch your carrier's weekly data. Check system dwell and the terminals on your route each week. When dwell climbs, add buffer to customer delivery dates and ship earlier rather than promising a date the network can't support.
The last two steps are where most shippers fall down, simply because nobody has time to watch every car every day. That is the job of Steel Wheel's managed rail service: tracking, exception management and follow-up, so a car sitting in a yard gets a call to the railroad the same day it stalls. Our guide to railcar tracking vs shipment management explains the difference between watching dots on a map and actually moving stuck cars. You can also see the full list of Steel Wheel services.
Illustrative Example
A plastics shipper loads five covered hoppers a week for a customer on a second railroad. The cars are handled at the local serving yard, a hump yard, the interchange yard and the receiving railroad's serving yard: four terminal touches. One week the origin railroad's dwell climbs several hours above its usual level, and the trains-held count for crew rises with it.
The shipper sees two cars sitting at the hump yard with no departure event after 48 hours. They call the railroad with the car numbers and learn the outbound train for that block was annulled twice for crew. They tell the customer the cars are delayed at the yard rather than quoting the old ETA, ask the receiver to stage unloading for a bunched arrival, and move the next release up a day so the following week's cars make the earlier train. The cars arrive late, but the customer plans around it and avoids a surprise demurrage bill from three cars landing at once. This example is illustrative.
If you are setting up a new rail lane and want to understand how railroads plan service around their yards, the Working with Railroads course module covers the relationships and contacts behind it.
Sources: 49 CFR Part 1250, Railroad Performance Data Reporting (§1250.1 reporting period and filing deadline; §1250.2(a)(2) terminal dwell; §1250.2(a)(4) unit train origin dwell; §1250.2(a)(5) trains holding; §1250.2(a)(6) cars not moved in 48 hours), via eCFR; Surface Transportation Board, EP 724 weekly Class I performance filings for the week ending October 17, 2026, as summarized in Steel Wheel Logistics Rail Pulse, October 27, 2026. Each railroad files its own methodology, so cross-carrier comparisons are approximate. The example is illustrative.
Frequently Asked Questions
What is rail terminal dwell time?
Rail terminal dwell time is the average number of hours a railcar spends at a terminal before it leaves on another train. Under 49 CFR 1250.2, Class I railroads report it weekly to the STB for their whole system and their 10 largest terminals, excluding cars that pass through on the same train.
What is a good terminal dwell time?
There is no official target. For the week ending October 17, 2026, Class I system averages in the STB filings ranged from about 17 to 25 hours. Compare a railroad against its own recent weeks, since each railroad uses its own measuring method.
Why is my railcar sitting in a rail yard?
Most often it is waiting for the next train that carries its block. Other common causes are crew or locomotive shortages, yard congestion, missing or wrong billing, an interchange handoff to another railroad, or a mechanical bad order. Call the railroad with the car number if it hasn't moved in 48 hours.
Is terminal dwell the same as demurrage?
No. Terminal dwell is time the railroad holds your car in its yards, and you are not billed for it. Demurrage is a charge you pay when you hold a car at your facility, or in constructive placement, past the free time in the railroad's tariff.
Where can I find rail terminal dwell data?
Class I railroads file it weekly with the Surface Transportation Board under its EP 724 performance data reporting, and the Board posts the filings publicly. Many railroads also publish weekly network metrics on their own websites. Steel Wheel summarizes the numbers each week in Rail Pulse.