Two rails on 200 metres of track can carry as much embodied carbon as the 50 tonne machine that runs on them, and the ballast option is 19 percent lighter than the concrete one. Here is where the number comes from, why a rubber-tyred yard roughly doubles the concrete, and what to put on the civil drawing before anyone pours.
A 50 tonne gantry crane on a 200 metre runway needs 400 metres of rail, and the civil work under those rails carries about 157 tonnes of CO2e if it is a continuous concrete strip beam. The crane itself carries roughly 153 tonnes. The runway holds slightly more carbon than the machine that runs on it, and only one of the two appears on the crane quotation.
That result surprised me the first time I ran it. I had spent a week arguing about the steelmaking route of the girders and then added the runway as a footnote. The runway is not a footnote. It is half the project, and buyers rarely ask about it, because crane suppliers do not build it and civil contractors do not price carbon.
An overhead crane borrows its runway from the building. The columns and runway beams are part of the shed, they are costed with the shed, and the crane maker never touches them. A gantry crane is different. It carries its own legs, and those legs run on rails that have to sit on something engineered for the whole length of travel.
That changes how the carbon scales. For an overhead crane, the civil carbon is a share of a building that would exist anyway. For a gantry, the runway exists only because the crane exists, and its carbon grows with metres, not with tonnes. A 300 metre yard needs 600 metres of rail before the first lift happens. Two cranes in two bays double it again.
Take a 50 tonne rail-mounted gantry, 30 metre span, 200 metres of travel, QU80 rail, and work through four ways of putting that rail on the ground. All four carry the same crane. Only the civil work changes.
| Runway option | New concrete | New steel | Total CO2e |
|---|---|---|---|
| Existing slab, dowel plates | none | 33 t | 63 t |
| Ballast bed, steel sleepers | none | 64 t | 127 t |
| Strip beam, 600 x 800 mm | 192 m³ | 52 t | 157 t |
| Pile-supported beam, soft ground | 294 m³ | 64 t | 211 t |
| For reference: the 50 t crane | none | 75 t | 153 t |
Four options, a factor of 3.3 between the lightest and the heaviest, on the same machine and the same duty. Nothing about the crane changed. A buyer who specifies a pile foundation on ground that did not need one has added 148 tonnes of CO2e at the same time as adding several hundred thousand dollars of cost.
Everything downstream of the wheel follows from one number: the vertical wheel load, per rail line. Rail selection, soleplate thickness, holding-down bolt diameter and embedment, and the foundation section all key off it. Get that number wrong by 30 percent and the whole civil design is wrong by 30 percent.
For a rail-mounted gantry in the 5 to 100 tonne range, published Chinese catalogue data puts wheel load per rail line at roughly 85 to 115 kN at 5 tonnes, 380 to 580 kN at 50 tonnes, and 650 to 1,050 kN at 100 tonnes. The spread inside a single capacity is the spread between light European-style box girders and heavier general-duty machines, and it is worth 40 percent of the foundation.
| Vertical wheel load | Clips on HD bolts | HD bolts | Soleplate | Bolt diameter |
|---|---|---|---|---|
| up to 30 kN | preferred | 2 | 10 mm | M12 |
| 30 to 50 kN | possible | 2 | 10 mm | M16 |
| 50 to 120 kN | not preferred | 2 | 12 mm | M16 |
| 120 to 250 kN | do not use | 2 or 4 | 15 mm | M20 |
| above 250 kN | do not use | 2 or 4 | 20 mm or more | M20 |
This table is not ours. It comes from Gantrail's technical note on soleplates and holding-down bolts, and it is the clearest published mapping from wheel load to fixing hardware that I know of. Note the jump at 250 kN: below it you can use discontinuous soleplates, above it the note tells you to consider a continuous soleplate unless the rail is very heavy and stiff. Half the industrial gantries in the 30 to 50 tonne class sit right on that boundary.
At the top end the civil design stops being a rail detail. A 1985 pavement study for the Port of Portland designed for gantry wheel loads of up to 80,000 lb, about 356 kN, and settled on 22 inches of reinforced portland cement concrete, which is 560 mm. That is not a rail bed. That is a road structure with a rail on top.
Item by item, the strip beam is heavier because concrete is a heavy material made with a heavy binder. Here is the breakdown for the 400 metres of rail, using 300 kg CO2e per cubic metre of concrete and 1.92 kg CO2e per kg of steel, the worldsteel global average for 2024.
| Item | Quantity | Factor | t CO2e |
|---|---|---|---|
| QU80 rail | 25.5 t | 1.92 kg/kg | 48.9 |
| Soleplates, clips, bolts | 7.2 t | 1.92 kg/kg | 13.8 |
| Steel sleepers, ballast option | 30.8 t | 1.92 kg/kg | 59.1 |
| Ballast, 300 m³ | 540 t | 6 kg/t | 3.2 |
| Blanket, 200 m³ | 360 t | 6 kg/t | 2.2 |
| Beam concrete, 192 m³ | 192 m³ | 300 kg/m³ | 57.6 |
| Beam reinforcement | 19.2 t | 1.92 kg/kg | 36.9 |
The surprise is the sleepers. A ballast bed sounds like a pile of stones, and the stone itself is nearly free at 6 kg CO2e per tonne. But 615 steel sleepers at 50 kg each come to 30.8 tonnes of steel, which is more than the rails. If you swap to timber sleepers the number falls sharply, and if you swap to concrete sleepers it roughly doubles.
Published railway assessments point the same way. A 2023 lifecycle study of ballasted and slab track in Austria (Sustainability, 15(17), 12974) found 651 t CO2e per kilometre for ballasted track against 1,197 for slab track in the production phase, 11 to 20 percent lower for ballast depending on subsoil, with more than 80 percent of the total coming from material production. The line that stuck with me is this one: the slab needs a service life of 73 years to break even on carbon. Yard runways are usually designed for 30 to 50.
Quantity dominates, but the mix is the cheapest lever available because it costs nothing. Binder is the problem. In a cubic metre of C30/37 with 300 kg of CEM I and 100 kg of reinforcement, cement is about 68 percent of the carbon, reinforcement 20 percent, and aggregate, water and mixing the remaining 12 percent.
| Concrete | kg CO2e per m³ | Note |
|---|---|---|
| CEM I, 300 kg binder | 300 | our planning value |
| Measured C30 to C80 range | 262 to 402 | 2025 literature review |
| 50 percent GGBS binder | 188 | binder factor 913 to 510 kg CO2e per tonne |
| Same beam, 192 m³ | 57.6 to 36.1 t total | 21.5 t saved |
Ground granulated blast furnace slag is the standard answer and has been for twenty years. Ready-mix with 50 percent replacement shows a binder factor of about 510 kg CO2e per tonne against 913 for plain cement, a 44 percent cut on the binder, which works out at roughly 37 percent on the concrete because aggregate and haulage do not change. In most markets it does not carry a price premium. The practical limit is early strength: in cold weather a 50 percent slag mix may need three to five days before formwork comes off, which nobody plans for and everybody discovers.
Piles are not a carbon choice, they are a soil answer, and the soil decides before anyone optimises anything. Weak or reclaimed ground with soft clay near the surface will not carry a strip beam without settlement, and differential settlement is what actually kills gantry runways. Container terminal experience puts the trigger at 20 to 25 mm of differential movement between neighbouring surfaces before re-levelling is required, which is a maintenance cost long before it is a safety issue.
| Foundation type | Ground condition | USD per metre, one rail line | CO2e, 400 m of rail |
|---|---|---|---|
| Pad foundations | firm, good bearing | 120 to 350 | 75 to 95 t |
| Strip beam | medium or variable | 300 to 750 | 157 t |
| Pile and capping beam | weak or reclaimed | 700 to 2,200 | 211 t |
| Geotechnical survey | any | 2,500 to 12,000 total | negligible |
The 2026 prices above come from a Chinese gantry civil guide published in July 2026, and they cover supply and installation per linear metre of one rail line. The same source puts total civil works at 20 to 60 percent on top of the crane price for a full gantry, and warns that on weak ground the foundation can cost more than the rails and site preparation combined. That is a budget warning, and it is also a carbon warning, because the expensive option and the heavy option are the same option.
On paper, yes. An RTG has no rails, no runway beam and no piles, which is exactly why terminals buy them: you can reconfigure the yard, you can steer between blocks, and you do not wait eighteen months for a civil contract. The catch is that the load has to go somewhere, and it goes into the pavement under every wheel, everywhere the machine drives.
| Yard of 45,000 m² | Rail-mounted | Rubber-tyred |
|---|---|---|
| Runway structure | 288 m³ beam, 227 t CO2e | none |
| Pavement depth | 200 mm general | 400 mm under wheel paths |
| Pavement concrete | 9,000 m³, 3,564 t CO2e | 18,000 m³, 7,128 t CO2e |
| Total civils | 3,791 t CO2e | 7,128 t CO2e |
Roughly double, in the same yard, for the option that has no foundation. The Portland design point of 560 mm exists because a heavy RTG wheel concentrates the whole leg load on a patch the size of a dinner plate, and eight-wheel machines are worse than sixteen-wheel machines, which is why terminal operators moved to sixteen wheels to bring the wheel load down to something a pavement can survive.
None of this says RTG is the wrong choice. It says the carbon argument for RTG is weaker than the commercial argument, and if a buyer is writing a sustainability section for a tender, the pavement is the number to disclose, not the crane.
This is where the ladder inverts. A diesel RTG at 15 litres per hour over 5,000 hours burns 75,000 litres a year, and at 2.676 kg CO2e per litre that is 201 tonnes a year. The crane and its whole civil works together were 153 tonnes. One year of diesel is worth more than the entire embodied footprint of the machine and its runway.
| Configuration | Energy per year | t CO2e at 206 g/kWh | t CO2e at 600 g/kWh | 20 years |
|---|---|---|---|---|
| Diesel RTG, 15 L/h, 5,000 h | 75,000 L | 201 | 201 | 4,014 |
| Electric RTG, 33 kVA, 5,000 h | 165 MWh | 34 | 99 | 680 |
| Electric rail gantry, 27 kW, 2,000 h | 54 MWh | 11 | 32 | 222 |
The 15 litres figure is not mine. A Kalmar paper in Port Technology (edition 75, 2017) tabulates diesel-electric RTGs from 13 to 23 litres per hour at 5,000 hours a year, which is 174 to 308 tonnes of CO2 annually per machine, on a factor of 2.676 kg CO2e per litre. An EPRI study at the Port of Savannah measured 3.9 US gallons, 14.8 litres, per hour over sixteen months and reported a 69 percent CO2 reduction after electrification, with the electric machines averaging 33 kVA against a diesel equivalent of 37 kVA.
Note how much the grid matters in that table. The same electric RTG is 34 tonnes a year on the 2024 EU average of 206 g/kWh and 99 tonnes on a 600 g/kWh grid. Electrification is not a carbon strategy on its own. It is a decision to outsource the fuel choice to the grid, and the grid you are attached to decides the answer.
No, and that is worth saying because it stops people optimising the wrong end. Four cast-in anchor seats of 1.5 cubic metres each come to about 6 cubic metres of concrete, under two tonnes of CO2e, against a runway of 157. It is 1 percent. It is also the part that keeps the machine on the rails in a 42 m/s gust, so the correct approach is to specify it properly and stop worrying about its carbon.
| Device | Holding capacity | Practical note |
|---|---|---|
| Manual rail clamp | 50 to 200 kN | needs a person outdoors in the storm |
| Powered rail clamp | 100 to 800 kN | must interlock with travel |
| Cast-in anchor seat | 500 to 2,000 kN | one per corner, set with the beam |
| Concrete ballast block | 5 to 20 t each | works by friction only |
Work the numbers once and the sizing becomes obvious. A 32 tonne gantry in a 42 m/s storm generates about 180 kN of wind force. Friction at a coefficient of 0.12 covers roughly 62 kN of that. The remaining 118 kN has to come from clamps or anchors, which is four clamps of 30 kN or two anchor seats. Picking two clamps because they were cheaper leaves 60 kN unaccounted for, and that is the load that walks the crane down the runway.
This one is counter-intuitive and buyers get it wrong in both directions. The gantry crane is HS 8426, Chapter 84, and Chapter 84 is not in Annex I of Regulation (EU) 2023/956. The crane triggers nothing. The materials underneath it are a different story.
| Material | HS code | In Annex I? | Where it appears |
|---|---|---|---|
| Cement | 2523 | yes | binder in the beam |
| Rebar | 7214 | yes | beam reinforcement |
| Crane rail | 7302 | yes | the rails themselves |
| Structural steel | 7308 | yes | runway beams, anchor frames |
| Bolts and nuts | 7318 | yes | soleplate fixings |
| Ballast stone | 2517 | no | ballast bed |
| Gantry crane | 8426 | no | the machine |
The threshold is 50 tonnes of CBAM goods per calendar year per importer, counted across all CN codes together. Our 400 metres of rail carries 25.5 tonnes of rail, 7.2 tonnes of fixings and 19.2 tonnes of rebar, so 52 tonnes of steel. The runway's steel alone crosses the line. A single imported gantry crane never does.
Certificate prices for 2026 have run at 75.36, 75.28 and 82.32 euro per tonne of CO2 for the first three quarters, published by the Commission on 7 April, 6 July and 5 October, with the fourth quarter figure due on 4 January 2027 and weekly pricing from 2027. At Q3 prices, 52 tonnes of steel with roughly 2 tonnes of embedded CO2 per tonne carries about 8,560 euro of exposure, if those goods enter the EU as separate consignments. Split the crane into girders and legs to dodge the classification and it does not work: general interpretative rule 2(a) and the CJEU judgment in Case C-107/22 (27 April 2023) both treat parts that need only bolting or welding as the complete machine.
Not linearly, and not in the direction most people expect. Small gantries look cheap and light, but the runway does not shrink with them, because the rail section, the beam depth and the survey all have floor values. A 5 tonne machine on a 100 metre runway carries twice its own carbon underneath it.
| Capacity | Span | Runway length | Crane CO2e | Runway CO2e | Ratio | t CO2e per m |
|---|---|---|---|---|---|---|
| 5 t | 12 m | 100 m | 15 t | 30 t | 2.0 | 0.30 |
| 20 t | 18 m | 150 m | 54 t | 91 t | 1.7 | 0.60 |
| 50 t | 30 m | 200 m | 153 t | 157 t | 1.0 | 0.79 |
| 100 t | 40 m | 250 m | 211 t | 303 t | 1.4 | 1.21 |
The practical reading is per metre. A runway in the 50 tonne class costs about 0.79 tonnes of CO2e per metre of travel, so the difference between a 200 metre runway and a 260 metre one is 47 tonnes, which is more than the entire beam mix optimisation and more than switching to ballast. The gantry crane range publishes wheel loads per rail line for each capacity, which is the number every foundation designer asks for first.
Cheap and light travel together here. The options that cost least are also the options that emit least, which is unusual in this field and makes the conversation easier.
| Option, 400 m of rail | Indicative cost | CO2e | kg CO2e per USD |
|---|---|---|---|
| Rail and fixings on existing slab | USD 40,000 to 85,000 | 63 t | 0.74 to 1.58 |
| Ballast bed and sleepers | USD 72,000 to 149,000 | 127 t | 0.85 to 1.76 |
| Strip beam | USD 160,000 to 385,000 | 157 t | 0.41 to 0.98 |
| Pile-supported beam | USD 320,000 to 965,000 | 211 t | 0.22 to 0.66 |
The ballast numbers are our own planning figures for supply and installation, not a published quotation, and ballast pricing moves with local quarry distance more than anything else. The rail and foundation rates come from the July 2026 Chinese civil guide quoted earlier. What survives the uncertainty is the ordering: the pile foundation costs four to six times the strip beam and carries a third more carbon.
Ten lines. If a supplier or a civil contractor cannot answer all ten before concrete is poured, the carbon number is guesswork and so is the price.
| # | Item | Why it changes the answer |
|---|---|---|
| 1 | Maximum and typical wheel load, per rail line | drives rail, fixing and section |
| 2 | Rail section and steel grade | 20 to 49 t of steel each way |
| 3 | Foundation type and assumed bearing capacity | strip beam or piles, a 54 t swing |
| 4 | Concrete class and minimum cement replacement | 21.5 t for no premium |
| 5 | Reinforcement content, kg per m³ | 20 percent of the concrete carbon |
| 6 | Fixing system, bolt diameter and embedment | M12 to M20 by wheel load |
| 7 | Drainage, falls and the low point | water under a beam is a 10 year problem |
| 8 | Expansion joints and rail expansion allowance | buckling shows up in summer |
| 9 | Storm anchor and clamp capacity, and position | under 2 t of carbon, all of the safety |
| 10 | As-built survey of gauge, straightness and level | the only proof the runway is what you bought |
Ranked by tonnes saved per dollar spent, for the 50 tonne case. The last row is a counter-example, because the most common upgrade request in our inbox is also the one that adds carbon without adding capability.
| Change | t CO2e saved | Cost | Verdict |
|---|---|---|---|
| Electrify instead of diesel | 167 per year | cable reel or conductor rail | do this first, by a wide margin |
| Use the existing slab if the survey passes | 94 | one survey, 2,500 to 12,000 | best value in the table |
| Shorten the runway by 20 m | 16 | a layout decision | free, and permanent |
| Ballast instead of strip beam | 30 | lower build, higher upkeep | good if the ground is stable |
| 50 percent GGBS in the beam | 21.5 | usually zero premium | ask, then verify the batch ticket |
| One size down in rail section | 9 | cheaper rail | only if wheel load allows |
| Raise the beam from C30 to C40 | minus 8, it adds carbon | more cement | counter-example: section is set by wheel load, not strength |
That last row is the one I argue about most. A runway beam is governed by wheel load, fatigue and settlement, not by compressive strength, and a 25 N/mm² concrete has been enough in published design examples. Moving to C40 because it feels safer adds roughly 8 tonnes of CO2e to the beam and changes nothing structural. If you want margin, buy it in the section and the survey, not in the cement content.
Steel factors: worldsteel Sustainability Indicators Report 2025, using 2024 data, 1.92 kg CO2 per kg of crude steel globally, 2.34 for the blast furnace route and 0.69 for scrap-based electric arc. Concrete: a 2025 literature review of measured C30 to C80 mixes at 262 to 402 kg CO2e per cubic metre, and a slag-cement handbook giving ready-mix binder factors of 913 and 510 kg CO2e per tonne at zero and 50 percent replacement. Rail weights: QU80 at 63.69 kg per metre and QU100 at 88.96, from published crane rail tables. Fixing hardware: Gantrail technical note on soleplates and holding-down bolts. Pavement depth: the 1985 Port of Portland gantry lane study at 22 inches for 80,000 lb wheel loads, and terminal pavement literature on eight and sixteen wheel RTGs. Ballast and slab comparison: Sustainability 15(17) 12974, 2023. Diesel and electric RTG consumption: Kalmar in Port Technology edition 75, 2017, and the EPRI study at the Port of Savannah. Civil prices: a Chinese gantry foundation and rail guide published July 2026. CBAM: Regulation (EU) 2023/956 Annex I, the Commission's 2026 certificate price publications, and CJEU Case C-107/22.
Everything calculated here is a planning figure, not a published statistic. The crane masses, beam sections, travel lengths and proportion of steel in each option are ours, built to be checked rather than quoted. If your ground is soft, your runway is longer, or your wheel loads sit at the top of the catalogue range, the numbers move by 30 to 50 percent in either direction. The ordering of the options does not move, and that is the part worth taking to a budget meeting.
A rail-mounted gantry does, because the rails have to sit on something that keeps gauge, straightness and level over the whole travel length. A light machine can use an existing 200 mm slab at about 20 MPa with dowel plates. Above roughly 20 tonnes the wheel loads usually force a designed strip beam, and on soft or reclaimed ground that beam has to be carried on piles.
A 600 by 800 mm strip beam takes 0.48 cubic metres per metre of rail, so 400 metres of rail takes 192 cubic metres and about 57.6 tonnes of CO2e at 300 kg per cubic metre. A pile-supported runway adds 60 to 110 cubic metres of pile concrete over the same length, which is how the pile option reaches 211 tonnes.
Yes, by about 19 percent in our own figures, 127 tonnes against 157 for 400 metres of rail, and published railway lifecycle work puts ballasted track 11 to 20 percent below slab track. The catch is life. That study found slab track needs 73 years to break even on carbon, and yard runways are normally designed for 30 to 50.
Sometimes, and it is the largest saving on the list. You need at least 200 mm of sound concrete, a compressive strength near 20 MPa, a survey of what sits underneath, and a punching shear check at every rail seat. Use dowel plates or chemical anchors rather than expansion bolts, which crack an old slab from the inside.
To the materials, yes. Cement 2523, rebar 7214, crane rail 7302, structural steel 7308 and bolts 7318 are all in Annex I of Regulation (EU) 2023/956. The gantry crane itself is HS 8426 and is not in Annex I, and ballast stone 2517 is not either. The importer threshold is 50 tonnes a year across all CBAM goods combined.
Typically 20 to 60 percent for a full gantry on decent ground, at 2026 civil rates of 300 to 750 US dollars per metre of one rail line for a strip beam. On weak ground with piles at 700 to 2,200 dollars per metre, the civil work can exceed the crane price itself.
Related articles:
Jib Crane Foundation Carbon: What the Concrete Block Weighs Before the Crane Arrives
Traveling Jib Crane Carbon: The Rail, the Plinth and the Boom Nobody Counts
Underslung Crane Carbon: What the Roof Carries Before the Crane Hangs
Crane Embodied Carbon: What a 20 Tonne Double Girder Actually Carries
Crane Carbon Footprint: What a Single Girder Crane Emits Before It Lifts Anything
Gantry Cranes: 5 to 100 Tonne, Rail and Rubber Tyred