BNSF Reopens Transcon After Flood: 60 Trains a Day, 60-Hour Eastbound Gap

The Analysis

Laguna is a name on a milepost, not a place you stop. You pass it between Grants and Albuquerque on BNSF's Gallup Sub, with Interstate 40 running along for company, and for roughly sixty hours at the end of October it was the most expensive piece of railroad in North America. Floodwater took the Southern Transcon out on the afternoon of Oct. 24. Both main tracks came back at Laguna on Oct. 27. Call the gap sixty hours. An average of sixty trains a day run across that subdivision, the busiest freight route on the continent, and a good share of what heads east over it is international intermodal: import boxes moving inland off San Pedro Bay toward the Midwest and beyond.

Then comes the number that explains everything after it. Up to eight inches of rain fell in three days around Laguna, against a local annual average close to ten inches. Three days delivered roughly eighty per cent of a year's worth of water to ground that had no reason to expect it. A second round of rain came through while the repair crews were still working and pushed the reopening back, which is the sort of detail a freight customer files under bad luck and I file under operating doctrine for the next decade.

This line I have worked, and I have sat out more embargoes in more yards than any sensible man should. When a line closes I ask three things and I ask them in this order: which node on the map is actually blocked, what it costs to go round it by another track, another gateway or another mode, and whether the road is worth running at all. The third answer here is still yes, because nothing in North America beats Southern Transcon intermodal for boxes heading east off San Pedro Bay, but the first two answers were ugly for about two weeks. BNSF put the railroad back in sixty hours. The river of containers behind it does not straighten out that fast. Every corridor has a story, and a pothole; this pothole was sixty hours deep, sixty trains wide, and it had eight inches of rain behind it.

Keep the record straight before I start reading it. Water put the Southern Transcon out of service on the afternoon of Oct. 24 at Laguna, New Mexico, on BNSF's Gallup Subdivision. Both main tracks were back in service on Oct. 27, which makes roughly sixty hours. Sixty trains a day is the published average on that piece of railroad, and it carries import boxes inland from San Pedro Bay. Around Laguna, up to eight inches of rain fell in three days, against a local annual average near ten inches. A second round of rain delayed the reopening, and BNSF was still working through the backlog and still hardening the line at Laguna when the story went out. That is everything the record says. It says nothing about how many trains were held, how many were rerouted, what happened to dwell at the receiving ramps, or how long until the flow looks like itself again, and those are exactly the four things your business pays for.

Sixty trains a day does not sound like much until you accept what a mainline like this really is. It is not a road with lanes, it is a timetable. The Southern Transcon was engineered and signalled decades ago specifically to move double-stack intermodal at speed between Southern California and the Midwest, which means passing loops, siding lengths, crew change points and terminal slots all tuned to one plan. Trains here carry a slot, not a direction. An intermodal departure is booked like a seat: it has a departure window, a crew expiry clock, a locomotive cycle and a reserved place on the ground at the far end. When the pipe stops, that plan does not pause politely, it piles up. And a line averaging sixty a day counting both directions has almost no spare hours in a peak season week.

The rain figure is where I would put my own money. Three days delivered close to eighty per cent of a year's rainfall to one patch of New Mexico desert. Trackbed drainage, culverts and the washes the railroad runs beside were all sized against a design storm, and a design storm is a historical bet rather than a promise. Give the ground eighty per cent of a year in seventy-two hours and the water takes whatever route offers least resistance, which is usually the ballast section somebody built level and straight through the drainage. Then comes the part operators understand and shippers never see: the repair itself has to survive a second storm. That is why the second round of rain pushed the reopening back. You can pour a subgrade in dry air; you cannot race water. And BNSF is still hardening the line, which is the railroad telling you that it does not consider this spot finished either.

Why late October matters is the other half of the same coin. Late October is precisely when the holiday push is already rolling east off Los Angeles and Long Beach and the pipeline is full of freight with a date attached to it. A road crew rebuilding grade can wait three days. A replenishment run that was going to be on the shelf for a November promotion cannot. So the same sixty hours that would have drawn a shrug in a slack February becomes, in late October, a queue of committed freight with nothing slack behind it: no spare train starts, no spare crews at the terminal, no spare parking rows at the receiving ramp. Absorbing capacity is a luxury you only have when you are not busy, and this industry is never politely busy.

Run the multiplication honestly. Sixty trains a day is two and a half an hour. Multiply by sixty hours and you get about one hundred and fifty train movements that did not happen when they were planned to. That is the headline product of sixty hours: a hundred and fifty trains' worth of movement pulled out of the schedule. Now allocate it to the people who care. Assume half of the sixty run eastbound, thirty a day, and assume about half of those carry international boxes. Both are my assumptions, not published figures, and if your intermodal marketing company has better splits, redo this with theirs.

Thirty eastbound a day halved is fifteen intermodal trains a day. Across two and a half days that is about thirty-seven trains. Put a hundred and sixty forty-foot equivalents on each one, a realistic band on that kind of train running somewhere between a hundred and twenty and two hundred and twenty depending on box mix and weight, and thirty-seven trains is roughly five thousand nine hundred boxes that were going east and did not.

Here is the arithmetic nobody prints, and it is the reason sixty hours is not sixty hours: the queue drains slower than it fills. Assume the reopened line can be pushed twenty per cent above its normal eastbound intermodal clip while locomotives and crews are chased into place. Fifteen normal against about eighteen pushed through leaves an excess of three trains a day to chew through a queue of thirty-seven, which is roughly twelve days. Push the surge to forty per cent, nineteen to twenty-one a day and six of excess, and you still need about six days.

The truth sits somewhere in that band, because the surge is capped by things that do not multiply on command: locomotive availability, crew hours that expire to a clock, parking rows at the far end, and the plain fact that the segment at Laguna is being hardened and will not run at full authorized speed on the day it opens. Call it ten to fourteen days from Oct. 24 before the eastbound flow resembles itself again, with the worst of the bunching landing at the receiving end around the turn of the month.

Bunched is the word to hold on to, because a pipeline does not refill evenly. What arrives at the Chicago end is a shape rather than a flow: almost nothing for two days, then too much for two days, then a thin day while the pipe behind it refills. I once watched a three-day closure on a Central Asian crossing take the better part of three weeks to settle for exactly this reason, and nothing about North American rail operations exempts anybody from the same shape. For the receiving ramp the first two days look like a gift: empty parking rows, gates running ahead, yard crews finally breathing. Then the wall arrives, and a facility sized for the average day, staffed for the average day, with a chassis pool sized for average dwell, takes a day and a half of demand inside a day and a half of calendar, wedged into the back half of a week.

So let me say plainly what I have not read in any account of this closure, because it is the thing I would build my own response around. The node that decides your bill is not in New Mexico. It is whichever receiving ramp your boxes were aiming at, roughly fifteen hundred miles east, together with the chassis pool and the drayage fleet that serve it. Sixty hours of blockage costs you sixty hours only if nothing downstream is disturbed, but intermodal does not trick through, it arrives on a schedule, and a scheduled deficit followed by a scheduled surplus breaks every assumption the far end was built on.

Parking rows are finite. Chassis availability at any given hour is a function of dwell, not of how many chassis the pool holds on paper. Gate appointments assume freight shows up inside the window it was booked for. And the cruellest one: intermodal arrivals are planned around a Monday to Friday world. A bunching event landing on Thursday evening or Friday meets a drayage market with no weekend capacity unless somebody pre-bought it, and every hour of that gap converts into dwell, dwell converts into chassis shortage, chassis shortage converts into missed delivery appointments, and missed delivery appointments are the line item that turns up on somebody's profit and loss. That is the transmission mechanism. The washout was in New Mexico; the loss gets booked where the boxes were supposed to be.

Walk down who eats it, in the order they eat it. Receivers with contracted volumes and committed allocations go first and are served first, because railroad intermodal protects the space of whoever signed for it. Then come the intermodal marketing companies and the mid-size importers riding their pooled space, the last load served when something has to be left behind, whose customers are the ones saying it should have been here Wednesday. Then comes the ramp itself, absorbing the bunching as dwell, and after it the drayage carriers there, asked to absorb the same event a second time as weekend and night work in a market where driver hours are already the binding constraint. And one group nobody mentions in this story: export flows and empties. When eastbound capacity is short, import pulls pay better than export pulls and empty repositioning pays worse than both, so export bookings out of the Midwest and empty boxes for whoever needs them quietly stop getting chased.

Geography decides more here than size does. If your boxes come off the railhead at a Chicago-area destination and get trucked inside a three hundred mile radius, this interruption sat directly on your critical path. If your import plan already runs two gateways with real monthly volume on each, it was a phone call. If everything you own moves on one railroad through roughly sixty miles of New Mexico, then a sixty hour event is not misfortune, it is a structural exposure you have been carrying for free and never priced.

Put money on it, with every assumption written down so you can argue with me rather than take it. Take one forty foot box of general merchandise with a landed value of sixty thousand dollars; that figure is mine, substitute your own. Say it gets caught in the middle of the bunching and lands five days behind plan. Rack storage past free time at a hundred dollars a day is my working number for that line, chassis and detention another seventy, and both are planning figures rather than quotes, so check them against your own tariff.

Two days past free time comes to three hundred and forty dollars of visible charges. Inventory carrying at eight per cent a year on sixty thousand is about thirteen dollars a day, so five days adds roughly sixty five dollars. Call it three hundred and fifty to four hundred and fifty dollars of directly attributable cost per affected box: annoying, and nothing more. Which is exactly why the direct number is the least useful number in this story. The expensive failures sit inside your own operation and never on the railroad's invoice: a replenishment appointment missed at a distribution centre, a line waiting on one component, a purchase order re-cut at expedited truck rates.

Grace Yang buys this stuff for a living and can take that part apart better than I can. My point in raising it is narrower and meaner: none of it appears in your accounts labelled as flood cost, it appears as a bad week your management will ask you to explain.

Now ask what would have changed the answer, because that tells you which of my conclusions are fragile. If the outage had been twenty four hours rather than sixty, this was two days of untidiness and you would not be reading me about it. If the second round of rain had not arrived, the reopening probably lands near forty hours and the residual queue halves. If the same flood had hit in a slack February with spare train starts available, it was a service note rather than a cost.

The hardest flip sits at your end rather than theirs. If your receiving ramp has spare parking and a chassis buffer, the slug absorbs and you barely notice. If your customer's distribution centre holds four weeks of cover, none of this reaches the shelf. If you were running lean and tight and booked to the day for a seasonal build, then sixty hours of New Mexico rain cost you real money, and you should assume it happens again. It will. Eight inches in three days against a ten inch annual average is not a figure you get to file as once in a century; whether the reason is climate or luck is an argument I am not qualified to settle, but the operating posture is the same either way. Assume this piece of desert can deliver eighty per cent of a year's rain in a long weekend, and assume the next weak point between Belen and Gallup is decided by a coin flip.

What would going around have cost you. Change gateway, routing a share of boxes off another western gateway or onto another railroad's line east through El Paso, and you typically buy two to four extra days of transit and a different set of receiving terminals, and it only works if your contracts and your drayage panel already exist there. Change mode, putting boxes on rubber out of San Pedro Bay, and the honest answer is that this works for Phoenix and Las Vegas and nobody else: two thousand miles of trucking is not a substitute, it is a different business, and the driver hour market has no spare capacity either.

Change nothing and re-phase instead, pulling the next two weeks of bookings forward by seven days and accepting an early arrival, and you usually solved more of the problem for less money than either of the first two. That is the uncomfortable lesson of every closure I have worked through. The firework afterwards is always rerouting; the cheap answer was always the calendar.

Four traps, and I have fallen into most of them at least once. Assuming the reopening date restores normal service, when hardening means slow orders and maintenance windows for weeks afterwards and a piece of track under construction does not run at the velocity it was built for. Assuming an embargo excuses your demurrage, when storage relief during a service interruption is a matter of tariff and contract language rather than of fairness, and silence in your agreement means you pay. Assuming the week after reopening is representative, when it is the week in which every delayed resource has been thrown at the backlog and it flatters everybody. And assuming somebody will volunteer where your boxes are. They will tell you when you ask, and the asking is your job.

↑ Back to the key points

— By Ömer Kaya