BNSF sets July 4 holiday plan as central-US heat prompts speed restrictions

The Analysis

There is a line I keep in my head for this part of the world. It starts in the Los Angeles Basin, climbs out through Barstow and Needles, crosses the Mojave into Arizona, runs Flagstaff and Albuquerque, then turns northeast through Amarillo before it tips into the Midwest at Kansas City and rolls on to Chicago. Ask BNSF what that is and they will call it the Southern Transcon, the main route between Southern California and the Midwest. I have worked this corridor long enough to know exactly where it pinches, and in the last ten days it has been pinched in three separate places at once.

Here is what actually came out of BNSF's network update dated Thursday 2 July, which itself was delivered one day earlier than scheduled because of the holiday, a small detail that tells you how much shadow a Fourth of July throws over a North American railway. Severe thunderstorm winds put two trains on the ground on the Panhandle Subdivision, the line running between Amarillo, Texas and Wellington, Kansas, that stitches the Panhandle into the Southern Transcon. Operating crews had all main tracks back inside roughly 48 hours and flows through the area normalised.

At the same time, temperatures across much of the central US and the Desert Southwest sat in the low to mid hundreds and reached 110 degrees Fahrenheit in places, and temporary speed restrictions went on with the heat. On top of that, BNSF published its Independence Day operating plan: agricultural trains running as usual, Grain Operations fully covered, carload service continuing through the day itself, but yard, local and road switch assignments at terminals held down on Saturday 4 July in line with customer activity and demand.

Taken one at a time these read like three pieces of housekeeping. Taken together they are one corridor losing capacity in three places during the same week, and if you move containers on it that matters a great deal more than any of them alone. Stripped of everything else, this is a road problem. Bits of the road are too hot, part of the road is closed because nobody is working, and another part of the road was briefly torn up. I am going to work from BNSF's own numbers, mark every assumption I make out loud, and then give you dates and owners, because the only question worth asking this week is whether your freight moves and what it costs you if it does not.

Let me put the whole of what BNSF said on the table before I argue with any of it, because most write ups of this note take half of it out. On performance: car velocity improved compared with the prior week but was down modestly month over month; terminal dwell was consistent with the prior week and slightly higher than the prior month; local service compliance held at 90% and was steady with recent levels. On incidents: two derailments caused by severe wind associated with thunderstorms on the Panhandle Subdivision between Amarillo and Wellington, both temporarily affecting main track operations, every main track restored to service within approximately 48 hours, and traffic through the area since normalised.

On the holiday: agricultural trains running as usual, Grain Operations maintaining full coverage, carload train service continuing uninterrupted through Independence Day, but restricted yard, local and road switch assignments at terminals on Saturday 4 July, plus a warning that connecting carriers running reduced operations may cause interline delays, plus a pointer to Demurrage Tariff 6004-C for free time and charges applicable on the holiday. On weather and work: active storm activity across portions of the northern Plains and upper Midwest through the weekend with localised heavy rainfall and isolated flooding possible, significant heat persisting across much of the central US and Desert Southwest through the end of the week with temperatures in the low to mid hundreds reaching 110 degrees in some locations, and seasonal track maintenance continuing across portions of the network under the annual capital programme.

Now let us stop being a bulletin board and translate those into what actually happens to a loaded box. If local service compliance is 90%, then one placement in ten does not get spotted when it was supposed to, and that is in a week BNSF describes as steady. Assume you have 60 cars a week running through local service on this network, which is my assumption and you should put your own number in its place. Six of them miss the first attempt. Assume each miss costs one extra day plus roughly 180 dollars of wasted truck time, driver waiting and re-delivery handling, again my assumption.

Six misses is about 1,080 dollars a week. Assume the peak heat window you are planning through runs roughly eight weeks from late June into late August, and that is nearly 8,600 dollars of plain leakage before a single storm or holiday enters the story. That is the part people never budget for, because it never arrives as an event. It arrives as a phone call about one car.

Do not shrug off Amarillo. This is the habit I picked up working the rail and road corridors of Eurasia, where a village nobody has heard of regularly decides the transit time of a whole service, and it applies exactly as brutally in the Texas Panhandle. The Panhandle Subdivision runs from Amarillo, Texas up to Wellington, Kansas and it matters far beyond the grain, oil and rock it originates. This line I have worked many times in planning terms: flows out of the Los Angeles Gateway are matched against capacity past Needles, against the heat of the Mojave in July, against the geometry through the Panhandle, and against the hour they are finally delivered to a buyer who never learns the name Wellington. A derailment in the Panhandle is not a Kansas problem. It is a Los Angeles problem wearing a Kansas hat, and every importer-exporter whose nested set of boxes was quoted with a confident date inherits it.

And here is the arithmetic that makes the 48 hour figure misleading if you leave it where you found it. Forty-eight hours is the repair time. It is the time it took crews to put steel back under trains, which is genuinely quick work and deserves credit. Forty-eight hours is not the recovery time, because the trains held behind a closed main track do not evaporate. They stack up on sidings never intended to hold them, at yard ladders never intended to absorb them, behind crews whose hours of service clocks have quietly run out, and then they have to be fed through the reopened tracks one at a time.

Assume the tail runs about three times the outage, my assumption, which is a common enough outcome when the closed section sits on a main route rather than a branch. That puts roughly six days of distorted running behind a two day closure. BNSF says traffic flows through the area have since normalised and I have no reason to doubt them; what I am saying is that normalised at Amarillo is not the same sentence as normalised at your consignee's door, and the space between the two sentences is the buffer you should be carrying in your own plan.

Heat is the one everyone waves away, because a speed restriction sounds like something that only costs the railway. It does not. Here is a version of it you can argue with. Assume your intermodal move from the Los Angeles Basin to Chicago covers something like 2,200 miles of rail, my assumption and yours may differ. Assume 700 of those miles sit inside the area under temporary heat restriction, my assumption based on the geography named in the update. Assume permitted speed goes from 60 miles an hour to 45 once an order is on. Before the order, that stretch takes 11.7 hours; after, it takes 15.6.

Call it four hours, gone, from one train. Four hours sounds like nothing until you multiply it. Assume 25 loads a day of yours are riding that corridor, and four hours each is 100 loaded train hours evaporated every day. At an assumed effective 40 miles an hour across restricted territory that is 4,000 miles of lost movement a day, which is something like two full train loads gone from the plan daily. Nobody sends you an invoice for it. It simply turns into a day, slowly, and then into a late delivery you have to explain to somebody who does not care about rail temperatures.

Then there is the third squeeze, the one everybody thinks they have planned for, which is the holiday. Listen carefully to what BNSF actually wrote, because the wording is doing work. The railway is not closing. It is limiting yard, local and road switch assignments at terminals on Saturday 4 July, and it is doing so in line with customer activity and demand. That phrase means the plan responds to what you tell them you need. If you stay silent because you assume nobody works that day, you have volunteered your own car into the reduced plan.

Now put a clock on it. Assume your normal interval from placement to pull is about 24 hours, my assumption. Assume a car is placed for you at four in the afternoon on Friday 3 July. Assume free time of 48 hours and assume it begins at the first midnight after placement, which is again my assumption and yours may not be. The clock then starts at 00:01 on 4 July and runs out at 00:01 on 6 July. The yard that should have pulled it is one of the terminals where assignments were being held down.

The Monday 6 July pull arrives after your free time has already expired, not with a margin. Assume demurrage of 125 dollars a car per day and assume 15 of your cars landed in that trap: one extra day costs about 1,875 dollars and two costs about 3,750, on a weekend nobody thinks of as a risk. Whether the holiday itself counts inside your free time is written in BNSF's Demurrage Tariff 6004-C and nowhere else, and arguing about that bill is much cheaper before it is issued than after.

Put all three into one ledger, because this is the arithmetic your plan should actually be built on, and I will keep the assumptions visible. Heat costs about four hours of transit on any train running through restricted territory. A train caught in the Panhandle outage loses somewhere around half a day before it is moving again, my assumption, and then inherits part of the six day tail. A car placed into the holiday gap loses roughly 38 hours against a normal 24 hour interval, my assumption. Stacked onto one unlucky load those add to about 54 hours, which is roughly 2.3 days against plan.

Now value it. Assume exposure of 250 dollars per container per day in inventory carrying, expediting or downstream disruption, my assumption and the number you must replace with your own. Two point three days is about 575 dollars of exposure per container. Assume 40 containers a week riding this corridor, and you are looking at something like 23,000 dollars of exposure in a single rough week, made of three small things that were each reported as good news. Do not shrug at this: this is the part nobody puts in a plan, and it is where plans usually give way.

There is one more line for that ledger and it is always the one everybody forgets, because it happens on somebody else's railroad. Assume fifteen percent of your carloads leave BNSF and cross to a connecting carrier somewhere in the trip, my assumption. Assume that in a normal week those cars sit at the interchange about twelve hours before the receiving road picks them up, also my assumption. Now assume the connecting carrier runs reduced operations over the holiday, which is exactly the sentence BNSF put in its own update, and assume the first crew available to move your car is thirty hours out instead of twelve.

That is eighteen additional hours, or about three quarters of a day, on a car nobody in your building is watching because it has technically left your network. Assume twelve interline loads a week in your book and an exposure of 250 dollars a container per day, and you are carrying something like 2,250 dollars a week of delay sitting at a junction you cannot see. The reason this one bites is not the money, it is the visibility: a car in your yard generates a problem ticket, while a car at someone else's interchange generates a shrug until it rolls past its promise date.

Which is why the useful work this week is all boring, and it is mostly about timing rather than price. By close of business on Friday 3 July, freeze every placement request that would put a car on the spot on Friday afternoon or during Saturday 4 July unless extra free time has already been agreed in writing. That is one instruction to whoever places your orders, and target zero exceptions. By Monday 6 July, get the holiday free time treatment for your own accounts confirmed in writing against Demurrage Tariff 6004-C, target one hundred percent of active accounts answered before you are working anywhere near another holiday-of-record month.

By Wednesday 8 July, if you genuinely need cars worked on Saturday, ask for them, in writing, with volumes attached. BNSF's plan explicitly tracks customer activity and demand, so a named request has a real chance of being served while a silent one does not, and the honest target here is to have every critical request logged by that date rather than to guarantee the outcome.

From there the work becomes seasonal, because heat is not a July 2 story, it is a July and August story. By Friday 10 July, re-baseline the transit you quote to your own customers on the Southern California to Midwest corridor. Take your current promise, add the heat driven time you have now measured, and write down a trigger: assume more than 700 miles of your lane is under restriction, or assume any heat order covering more than thirty percent of your run has been issued, and your quoted transit automatically gains one day until the restriction lifts.

The point is to stop negotiating promises during a delay. By Friday 17 July, do the one thing people hate paying for and always wish they had: give your Midwest volume a second way in. Pick an alternative ramp or an alternative gateway and move the smaller of ten percent of volume or five containers a week through it for four weeks, with the explicit purpose of proving the second door actually opens. Target a measured on-time result of ninety percent or better on that secondary path before it is allowed to carry anything critical.

The alternatives deserve their own honest accounting, because a diversion is like a ferry: it exists, it runs on a timetable, and if you are standing on the bank when it leaves you wait for the next one. Splitting across two ramps protects you against one node having a bad week. Shifting your shipping day earlier in the week protects you against an assignment gap you can see coming two months in advance. Moving time sensitive freight onto truck for part of the journey works in July, but truckers drive on the same asphalt you are trying to escape and their own schedules degrade in the same heat, so price that realistically rather than hopefully.

What I would not do this week is chase the weather westward or northward to dodge the heat. BNSF reported expected storm activity across the northern Plains and upper Midwest with localised heavy rainfall and isolated flooding possible. There is no routing that gets you out of the July map. There is only a choice about which weather you would rather be driving into, so make that choice deliberately rather than accidentally.

Three mistakes will get made this month and they are all predictable. Somebody will read that traffic through the Panhandle area has normalised and will plan as though the six days of tail behind it never existed. Somebody will treat the Saturday plan as a full closure, stop communicating, and then pay demurrage for a holiday that was never actually closed. Somebody will route interline traffic expecting the connecting carrier to be at full strength, ignoring the sentence in BNSF's own update saying reduced operations at connecting carriers may result in delays for interline traffic, and will discover it when a car sits an extra day at an interchange nobody was watching.

There is also the quiet fourth one: seasonal track maintenance continues across the network under the annual capital programme, which is a good thing for long term reliability and a real subtraction from this summer's capacity. Lay that on top of heat orders that are already stealing hours and you get a corridor running on noticeably less effective capacity than your annual plan assumed, with nothing anywhere announcing the difference.

Here is the shape of it as I see it from here. The restrictions come off when the thermometer comes down, the derailment slides out of everyone's memory by the middle of the month, and the operating plan notes stop being read. Then the next storm comes through the northern Plains, or a heat order goes on somewhere between Barstow and Needles, and everyone is surprised again in the same way. You do not rebuild the engine in a heat wave. You check the tyres, carry water, and stop scheduling anything fragile for a Saturday when the only people working are the ones who asked to be there. That is not a philosophy, it is just how summer works on this corridor, and it is cheaper to accept it on Tuesday than to discover it on Monday.

What I would actually put in front of my own team this week is short and it is boring, which is generally a good sign. Know which of your freight is sitting in the heat zone on any given day. Know which of your placements were made into a day nobody is going to work. Know which of your cars cross to another railway and how long they normally sit at the interchange before something changes. None of that shows up in an operating metric anybody publishes, all of it is knowable by asking three questions of three people, and it is the difference between being told about a delay and having already moved around it.

↑ Back to the key points

— By Ömer Kaya