Bridge engineering is the branch of civil and structural engineering that plans, designs, builds, inspects and maintains bridges. It covers the whole life of a crossing: choosing where and how to cross, selecting the bridge type, calculating the loads and sizing every member against a design code, working out how the structure will be built, and then keeping it safe for seventy-five to a hundred and twenty years of traffic, weather and wear.

It draws on structural analysis (how forces travel through beams, arches, cables and trusses), materials science (concrete, steel, prestressing strand, timber, composites), geotechnical engineering (the ground under the foundations), hydraulics (rivers, scour, waves), wind engineering and earthquake engineering, and on construction management, because a bridge is built in stages and every stage is a structure in its own right. The first chapter of this book defines the vocabulary; the rest of the book follows the subject branch by branch.

What makes bridge engineering distinct from building design is scale and exposure: spans are long, the structure is fully exposed to weather and water, loads move, and there is usually no architect between the engineer and the finished form. The bridge is the engineering.

A bridge engineer spends the working week on some mix of: feasibility and type studies for a new crossing; structural analysis and design of decks, girders, piers, foundations, bearings and joints; checking a contractor's construction-stage calculations and temporary works; inspecting existing bridges and rating their load capacity; and writing reports that let an owner decide what to build, repair or replace. On the owner's side (a highway or rail authority, a port, a developer) the same engineer procures the work, reviews consultants and manages the programme. Section 1.11 walks through the life cycle and who owns each stage.

The usual route is a degree in civil or structural engineering, then several years of design-office or site experience under a chartered or licensed engineer, then the licence itself: the PE (Professional Engineer) in the United States, CEng through the Institution of Structural or Civil Engineers in the UK, P.Eng in Canada, 기술사 in Korea, and equivalents elsewhere. Bridge specialists then add a working knowledge of one or two design codes (AASHTO LRFD, the Eurocodes), one or two analysis packages, and, with time, the field experience that no code contains: how a balanced cantilever actually behaves at closure, what a scour hole looks like after a flood, why a bearing seized.

Salaries follow the general civil-engineering scale with a premium for long-span and cable-supported specialists, who are few. The profession is small (a large country has only a few thousand bridge engineers) and international, because the codes are shared and the big projects move between continents.

Bridge design runs through the same stages everywhere, whatever the code.

1. Site and requirements. Survey, ground investigation (boreholes, soil tests), river hydrology or navigation clearance, traffic and load requirements, seismic zone, design life, budget.
2. Type selection and span layout. Compare beam, box, truss, arch, cable-stayed and suspension options for the site: span lengths, pier positions, ground conditions, construction access, aesthetics and cost. Chapter 3.9 is about this choice; the span-type recommender and the Bridge Design tool make a first pass automatically.
3. Preliminary design. Girder depth from span-to-depth rules, deck width from the lanes, pier and foundation sizes from the reactions, first quantities and cost. This is the level the Bridge Design tool works at.
4. Loads and analysis. Dead load, superimposed dead load, live load (the HL-93 truck and lane in AASHTO, Load Model 1 in the Eurocodes), dynamic allowance, wind, temperature, earthquake, water and collision, combined with load factors into limit states (Chapter 4). A structural model, from a hand calculation with influence lines to a 3-D finite-element model, gives the forces in every member for every combination.
5. Detailed design. Every member and connection is checked for strength, service (deflection, cracking, vibration), fatigue and extreme events, and detailed for construction: reinforcement, prestressing, plates, welds, bolts, bearings and joints (Chapters 6, 7 and 9).
6. Construction engineering. The erection sequence, temporary works and every intermediate stage are analysed, because a half-built cantilever or a girder being launched is a different structure from the finished bridge (Chapter 8).
7. Independent check and drawings. A second engineer checks the design; drawings and specifications go to tender.

The whole process is iterative: the type chosen in step 2 is revisited when step 5 shows a pier is too heavy or step 6 shows the crane cannot reach.

Almost every bridge belongs to one of six structural families, or combines two of them.

beam / girder

A member laid across the gap that carries load in bending: slab, T-beam, I-girder, box girder. The most common type by far, from 5 m culverts to 300 m concrete boxes.

truss

A beam broken into triangles so each member works in pure tension or compression; light and stiff, the railway type of the nineteenth century, still used for long-span rail and for steel bridges built from small pieces. Up to about 550 m as a cantilever.

arch

A curved member that carries load in compression and pushes outward at its ends; stone for two thousand years, now steel or concrete, to 550–600 m.

cable-stayed

A deck held by straight cables running to one or more towers; the standard form from 300 to 1,200 m.

suspension

A deck hung from a main cable draped over towers and anchored in the ground; the longest spans in the world, past 2,000 m.

cantilever

Arms projecting from the piers and meeting in the middle, in steel (the Forth and Quebec bridges) or, in balanced-cantilever concrete construction, as the way a box girder is built.

Cross-cutting labels describe use or movement rather than structure: highway, railway, pedestrian and pipeline bridges; viaducts (long, many spans); overpasses and underpasses; movable bridges (bascule, swing, lift); floating and pontoon bridges; extradosed bridges (a box girder with short stays). Chapter 3 describes each type with its span range, construction and typical proportions, and section 1.4 explains the four ways of spanning that underlie all of them.

There is no single answer, because "strong" depends on the span and the material, and every type is strong at the span it was invented for. The honest engineering answer is: the strongest bridge is the one whose form matches its load and its material.

• For a short span in a material that is good in compression (stone, concrete), the arch is the strongest form: Roman arches are still carrying traffic after two thousand years.
• For a school project built from popsicle sticks, straws or balsa, the truss usually wins, because triangles turn bending into pure tension and compression and make the most of a little material. A Warren or Pratt truss with a deep section and glued, not pinned, joints carries the most load for its weight.
• For the longest spans, only cable forms work at all: the suspension bridge holds every span record because a cable carries load in pure tension and wastes nothing on bending.
• For everyday highway bridges the girder is not the strongest form per kilogram, but it is the cheapest, quickest and easiest to maintain, which is why most bridges are girders.

Section 1.8 explains why each form has a ceiling (its own weight) and section 1.5 why depth matters more than material in a beam. If the question is about a real project, "strongest" is replaced by "adequate with margin": every type is designed to the same safety level (section 1.9), so a well-designed girder bridge is exactly as safe as a well-designed arch.

From the top down:

deck

The surface traffic rides on, usually a reinforced-concrete slab 200–250 mm thick, sometimes an orthotropic steel plate; with the wearing surface, barriers, kerbs, drainage and lighting on top.

girders / main members

The beams, box, truss, arch or cables under (or beside) the deck that carry it across the span. Together with the deck they form the superstructure.

bearings

Small devices between the superstructure and its supports that carry the load down while letting the deck expand, contract and rotate (elastomeric pads, pot bearings, spherical bearings).

expansion joints

Gaps at the ends of the deck, sealed with a joint, so it can move with temperature without pushing on the approach.

piers

The intermediate supports: columns, walls or hammerheads with a cap beam on top.

abutments

The end supports, which also hold back the earth of the approach embankment; with wing walls and an approach slab behind.

foundations

Spread footings on good ground, or piles / drilled shafts down to it; pile caps join a group of piles. Piers, abutments and foundations are the substructure.

appurtenances

Barriers and railings, drainage, lighting, signs, utilities, inspection walkways.

Section 1.2 has the annotated drawing; Chapter 9 covers bearings, joints and the rest of the small parts that wear out first.

A bridge is designed for a list of loads and for the ways they combine. The permanent ones are its own weight (dead load) and everything fixed on it: surfacing, barriers, utilities. The main variable load is traffic: design codes replace real trucks with a notional model, the HL-93 truck plus lane load in AASHTO LRFD, Load Model 1 in the Eurocodes, KL-510 in Korea, plus a dynamic allowance for impact and braking and centrifugal forces. Then wind on the deck and towers, temperature (uniform change and gradient), earthquake, water pressure, ice and debris, scour at the foundations, ship and vehicle collision, and the loads of construction itself.

The loads are combined into limit states with factors that reflect how well each is known: Strength (about 1.25 on dead load and 1.75 on traffic in AASHTO), Service (deflection, cracking, stresses under everyday load), Fatigue (millions of truck passages) and Extreme Event (earthquake, collision, major flood). Chapter 4 explains the philosophy and the numbers; the HL-93 envelope tool computes the traffic moments and shears on this site.

Every country designs bridges to a national code, and most of them descend from a few families.

AASHTO LRFD Bridge Design Specifications

The United States, and by adoption much of the Middle East, Latin America and Asia; load-and-resistance-factor design with the HL-93 live load. The AASHTO Manual for Bridge Evaluation covers inspection and load rating. This book and the calculators on this site follow AASHTO LRFD.

Eurocodes (EN 1990–1999)

The European Union, the UK (with national annexes) and many countries in Africa and Asia; EN 1991-2 gives the traffic loads, EN 1992-2 concrete bridges, EN 1993-2 steel bridges, EN 1994-2 composite bridges, EN 1998-2 seismic.

KDS / KCS (Korea)

The Korean Design Standard for bridges (KDS 24), limit-state based since 2015, with the KL-510 live load.

Others

Japan Road Association Specifications (Japan), AS 5100 (Australia), CSA S6 (Canada), IRC codes (India), JTG (China), each with its own live-load model and factors.

Section 4.6 compares them on one page. The physics is the same everywhere; what differs is the live-load model, the factors and the detailing rules, so a design cannot simply be moved from one code to another without re-checking.

For global analysis of a whole bridge, the common packages are midas Civil, CSiBridge and SAP2000, SOFiSTiK, LUSAS, Bentley RM Bridge and LARSA 4D, all of which handle construction staging, prestressing, moving loads and cable-supported structures. For local detail and research, general finite-element codes such as ANSYS and Abaqus. For standard girders there are dedicated tools (for example for AASHTO precast girders and steel plate girders), and every design office keeps its own spreadsheets for bearings, joints, footings and piles. Drawings come from AutoCAD, MicroStation, Revit or Tekla, increasingly as a 3-D model rather than 2-D sheets.

The software does the arithmetic; it does not decide the bridge. Type selection, span layout, the construction method and the sanity check on every result are still done by hand, which is why a preliminary design can be produced from a few numbers, as the Bridge Design tool on this site does for girder bridges, and why the free calculators here are worth keeping open beside the big packages: a distribution factor, an HL-93 envelope, a scour depth or a pile capacity that can be checked in a minute is the fastest way to catch a modelling mistake.

The construction method depends on the type and on what is under the bridge. The common methods are:

• Cast in place on falsework: scaffolding from the ground; short spans over dry land.
• Precast girders: concrete or steel girders made in a yard and lifted in by crane, launching gantry or floating crane; the workhorse of highway construction.
• Balanced cantilever: a box girder built outward from each pier in segments, cast in form travellers or precast and lifted, so the river or road below stays open; for spans of 80–300 m.
• Incremental launching: the deck is cast in a yard behind the abutment and pushed out over the piers, segment by segment.
• Movable scaffolding (MSS) and precast segmental span by span: a machine casts or assembles one span at a time on long viaducts.
• Cable-supported erection: cable-stayed decks cantilever out from the tower as the stays are added; suspension decks are lifted from the water onto the main cable in sections; arches are built as two half-arches held back by temporary stays.

Chapter 8 describes each method with its cost, speed and risks, and the construction methods catalogue on this site has one page for every method, from the first borehole to the last bearing. The Bridge Master game lets you choose the method for a site and see what goes wrong.

Rough durations from contract award to opening, for a competent contractor without land or permit delays:

highway overpass, 2–3 spans

12–18 months

river bridge, 100–300 m, precast girders or a box

18–30 months; balanced cantilever adds a few months per pier

viaduct, 1–3 km

2–3 years, with the deck going up at one span a week (precast or MSS)

cable-stayed bridge, 400–1,000 m main span

4–6 years; the towers take a year, the deck cantilevers a year, foundations in water often the longest item

suspension bridge, 1–2 km main span

5–8 years; anchorages and cable spinning are the critical path

Piers in water, deep foundations, poor ground, winter or monsoon seasons and traffic management on a live road each stretch these. The Bridge Design tool gives a duration estimate for a girder bridge from its spans, method and site, and Chapter 8 explains where the time goes in each method.

Bridge cost is usually quoted per square metre of deck, because length alone says little. Typical all-in construction costs (structure only, excluding land, approaches and design fees), as world-average orders of magnitude:

precast girder overpass or viaduct

1,200–2,500 USD/m²

concrete box girder, cast in place or segmental

1,800–3,500 USD/m²

steel composite girder

2,000–3,500 USD/m²

arch or truss, 200–500 m

4,000–8,000 USD/m²

cable-stayed, 400–1,000 m

6,000–12,000 USD/m²

suspension, over 1,000 m

10,000–25,000 USD/m², driven by the anchorages and the cable

Foundations in deep water, seismic zones, high labour cost countries and marine sites push these up by 30–100 per cent; simple sites in low-cost countries come in below the ranges. Of the total, materials are typically 40–55 per cent, erection and temporary works 25–45 per cent (much more at sea) and preliminaries and contingency the rest. The budget estimator gives a quick figure by type and span; the Bridge Design tool builds a bottom-up estimate for a girder bridge, materials plus erection plus site, and section 1.8 explains why the cheapest span is an economic balance between the deck and the piers.

Design codes set a design life of 75 years (AASHTO), 100 years (Eurocodes) or 120 years (UK), which is a durability target: the concrete cover, the steel coatings, the drainage and the detailing are chosen so that the structure reaches that age with normal maintenance. The parts wear out at different rates: wearing surface 10–15 years, expansion joints 10–20, bearings 30–40, paint on steel 20–30, barriers whenever they are hit, while the girders, piers and foundations are meant to last the full life.

In practice bridges are replaced for reasons other than age: too narrow for new traffic, too weak for new legal loads, an obsolete detail that cannot be inspected, or a maintenance bill that exceeds replacement. Roman stone arches are two thousand years old; well-kept nineteenth-century iron and steel bridges are still in service; the 1960s–70s concrete bridges built with little cover and chloride-contaminated aggregate are the ones failing early. Chapter 10 covers inspection, rating and rehabilitation, which are what turn a design life into an actual one.

Complete collapses are rare, on the order of a few per year worldwide against several million bridges in service, and they cluster around a few causes:

• Scour and flooding: erosion of the soil around foundations during floods is the single largest cause of bridge collapse in most countries, roughly half of all failures in the United States.
• Collision: ships hitting piers (Sunshine Skyway 1980, Baltimore Key Bridge 2024), trucks hitting low girders, trains derailing into piers.
• Overload and deterioration: corroded steel, cracked prestressing, fatigue of welded details, undersized members that were never caught (I-35W, 2007).
• Construction accidents: falsework, launching and segment erection failures; most fatalities in bridge engineering happen during construction, not service.
• Design errors and unknown mechanisms: cast-iron beams (Dee 1847), buckling (Quebec 1907), aerodynamic flutter (Tacoma Narrows 1940), stress-corrosion cracking (Silver Bridge 1967).
• Earthquake: unseated spans, short columns and liquefied ground (Loma Prieta 1989, Kobe 1995).

Every one of these has produced a rule, and section 2.14 lists them. This site's Failures library documents 300 cases by mechanism, and the scour calculator applies the HEC-18 method written after the Schoharie Creek collapse of 1987.

Most countries inspect every public bridge on a fixed cycle. In the United States the National Bridge Inspection Standards require a routine inspection at least every 24 months, an underwater inspection every 60 months for piers in water, and hands-on inspection of fracture-critical members; similar regimes exist in Europe, Korea and Japan. Inspectors rate the deck, superstructure and substructure on a 0–9 condition scale (or the equivalent element-level states), record defects and recommend repairs, and the results feed a bridge management system that ranks the network for funding.

Load rating is the calculation that follows: given the bridge as it actually is, corrosion and all, how much traffic can it safely carry? The rating factor RF is the ratio of the capacity left after dead load to the effect of the rating vehicle; RF below 1.0 at the legal load means the bridge is posted with a weight limit, strengthened or closed. Chapter 10 covers inspection practice, condition rating, load rating, common defects and rehabilitation; the load rating tool applies the AASHTO Manual for Bridge Evaluation.

A girder is a main beam that carries the deck across the span; the word usually means a large beam, as opposed to the smaller stringers and floor beams that may sit on it. Girders come as reinforced-concrete T-beams for short spans, precast prestressed concrete I-girders, bulb-tees and U-girders for 20–55 m, and steel plate girders (a web and two flanges welded from plate) for 30–120 m. Several girders side by side share the deck, and the share each takes from a truck is the distribution factor that Chapter 4 and the distribution factor tool compute.

A box girder is a hollow beam, rectangular or trapezoidal in section, in concrete or steel. Closing the section makes it enormously stiff in torsion, which suits curved bridges, wide decks on a single support and long spans; it also puts the material at top and bottom, where bending stresses are (section 1.5). Concrete boxes are cast in place on falsework or a movable scaffold, built in balanced cantilever, launched, or assembled from precast segments; steel boxes are fabricated in a shop and lifted in. Sections 3.2 and 6.4 go into their design, and the girder depth tool gives the depth for a given span.

A bridge deck grows and shrinks with temperature (a 1 km deck by several hundred millimetres a year), rotates as it bends under traffic, and creeps and shrinks if it is concrete. Bearings sit between the deck and each pier or abutment to carry the vertical load while allowing those movements: laminated elastomeric pads (rubber and steel plates) for ordinary spans, pot bearings and spherical bearings for heavy loads and large rotations, sliding surfaces of PTFE for the movement, guides for the direction, and seismic isolators where earthquakes govern. Some modern bridges avoid bearings altogether by building the deck into the piers (integral bridges).

Expansion joints close the gap between the deck and the approach so that traffic rides over it while the deck moves: asphaltic plug joints for a few tens of millimetres, strip seals and finger joints for more, modular joints for the metre or more of movement at the end of a long bridge. Joints and bearings are the parts of a bridge that wear out first, and a leaking joint is the commonest cause of corroded girder ends and pier caps. Chapter 9 covers both, and the bearing and expansion joint tools size them.

Piers are the supports between the ends of a bridge. They come as single or multiple columns with a cap beam (a portal frame or a bent), as solid walls, as a single column with a wide hammerhead, or, for cable-supported bridges, as towers or pylons. They carry the deck reactions down to the foundation and resist the horizontal loads (braking, wind, earthquake, water and ship impact) that come with them; the taller the pier, the more slenderness and seismic behaviour govern its size.

Abutments are the supports at the two ends. They do two jobs at once: carry the end of the deck, and hold back the earth of the approach embankment as a retaining wall. A seat abutment has a stem, a bearing seat and a back wall; wing walls hold the embankment at the sides; an approach slab behind bridges the settlement of the fill. Integral and semi-integral abutments build the deck into the abutment and do without bearings and joints. Chapter 7 covers the design of both, and the pier and abutment tools carry out the checks.

The foundation takes the pier or abutment load into the ground, and the ground decides the type. Where rock or dense soil is near the surface, a spread footing (a reinforced-concrete slab a few metres square) is enough. Where good ground is deeper, piles carry the load down to it: driven steel or precast concrete piles hammered in, or drilled shafts (bored piles) of 1–3 m diameter cast in a drilled hole, joined at the top by a pile cap. In rivers and at sea, large-diameter shafts inside permanent steel casings, caissons sunk through the water, or cofferdams that let the footing be built dry. Piles resist load by friction along their shaft and bearing at their tip, and they must also resist the lateral loads and the scour that may remove the top metres of soil.

Foundations are the least predictable part of a bridge, because the ground is the one material the engineer did not specify, which is why the resistance factors are the lowest and the site investigation matters most. Chapter 7 covers footings, piles, shafts and scour; the spread footing and pile capacity tools do the sizing.

Scour is the erosion of the river bed and banks around a bridge's piers and abutments by flowing water. Three kinds add together: general degradation of the whole channel over years, contraction scour where the bridge narrows the flow, and local scour where the water accelerates around each pier and digs a hole at its base. A single major flood can remove several metres of bed and leave a footing or the top of a pile group hanging in water. Scour is the leading cause of bridge collapse worldwide.

The defence is to found the piers below the predicted scour depth, to shape piers to disturb the flow as little as possible, and to protect the bed with riprap or other armour where that is not enough; the depth itself is estimated from the flood flow, the pier width and shape and the bed material by methods such as HEC-18 in the United States. Section 7.5 explains the mechanism, and the scour depth calculator applies HEC-18.

Both hang the deck from towers with steel cables, but the geometry and the structural behaviour are different.

In a suspension bridge the deck hangs by vertical hangers from a main cable that drapes over the towers and is anchored in massive blocks (or the rock) at each end. The main cable carries the whole deck in tension; the towers only carry vertical load; the anchorages resist the huge horizontal pull. It is the only form that reaches 2,000 m, but it needs anchorages, it is flexible (which made aerodynamics a design issue after Tacoma Narrows), and the deck cannot be built until the cable is finished.

In a cable-stayed bridge straight stays run directly from the tower to points along the deck. Each stay carries its piece of deck; the horizontal components of the stay forces run into the deck as compression, so the deck is a compression member as well as a beam, and there is no anchorage. The towers are taller relative to the span, the bridge is stiffer, and it can be built by cantilevering out from the tower one stay at a time. It is the economical form between about 300 and 1,200 m; beyond that the stays become too long and flat, and the suspension bridge takes over. Sections 3.5 and 3.6 describe both in detail, and the catenary tool gives the cable geometry and forces.

Concrete is strong in compression and weak in tension, and a beam in bending has tension on one side. A prestressed concrete girder is squeezed in advance by high-strength steel strands tensioned to about 1,400 MPa, so that under the working loads the concrete stays in compression, or nearly so, and does not crack. The result is a girder about half the depth of a reinforced-concrete one, that can span 20–55 m as a precast I- or U-girder and 100–300 m as a box girder, and that stays uncracked and durable for its whole life.

There are two ways to do it. Pretensioning stretches the strands in a casting bed before the concrete is poured and releases them once it has hardened, which is how precast girders are made in a factory. Post-tensioning casts ducts into the concrete, threads tendons through them after it has hardened, jacks them against the ends and grouts the ducts; this is how box girders, balanced cantilevers and continuous decks are prestressed on site. The design must account for the losses of prestress over time (elastic shortening, creep, shrinkage, relaxation), which is why creep theory (section 2.10) came from bridges. Section 6.2 covers the design, the prestress loss tool computes the losses, and the Bridge Design tool runs the full AASHTO check on a PSC girder.

A segmental bridge is a concrete box girder built from short segments, 3–5 m long, either cast in place one after another or precast in a yard and joined with epoxy and post-tensioning. The segments let a long deck be built without falsework under it, which is the point: the road, railway or river below stays open, and the same equipment repeats the same operation hundreds of times.

Balanced cantilever construction (free cantilever method, FCM) builds the deck outward from each pier in both directions at once, so the pier is always balanced; each new segment is cast in a form traveller or lifted into place and stressed back to the last, and the two cantilevers from adjacent piers meet at a closure pour in midspan. It suits spans of 80–300 m with a deep, haunched section at the piers. Span-by-span erection assembles all the segments of one span on a gantry and stresses them together, for 30–60 m spans on long viaducts. Incremental launching casts the deck behind the abutment and pushes it out. Chapter 8 and the methods catalogue cover all of them, the FCM tool proportions a balanced-cantilever box, and section 2.11 tells how the method was invented at Worms in 1953.

"Longest" needs a qualifier, because a bridge can be long end to end or long between its supports.

longest bridge (total length)

Danyang–Kunshan Grand Bridge, China (2011): 164.8 km of high-speed railway viaduct.

longest span of any type

1915 Çanakkale Bridge, Türkiye (2022): 2,023 m suspension span; Akashi Kaikyō, Japan (1998), 1,991 m, held the record for 24 years.

longest cable-stayed span

Changtai Yangtze River Bridge, China (2025): 1,208 m, road and rail; Russky Island (2012) 1,104 m before it.

longest arch span

Pingnan Third Bridge, China (2020): 575 m steel tubular arch; concrete arch: Tian'e Longtan (2024), 600 m.

longest cantilever truss

Quebec Bridge, Canada (1917): 549 m, unbeaten for more than a century.

longest girder (beam) span

Shibanpo Bridge, Chongqing (2006): 330 m concrete box with a steel midspan section.

highest bridge deck

Huajiang Canyon Bridge, China (2025): 625 m above the river.

tallest bridge structure

Millau Viaduct, France (2004): pylon 343 m from foundation to tip.

longest sea crossing

Hong Kong–Zhuhai–Macau Bridge (2018): 55 km of bridges, tunnel and artificial islands.

Chapter 11 keeps the record tables by type, the Rankings page on this site lists them, and section 2.15 shows how the records moved through history. Records are confirmed against the opening date; bridges under construction (the 1,700 m Zhangjinggao Yangtze suspension spans, the 3,300 m Messina design) are not counted until they open.

The span is the distance between two adjacent supports; the length is the whole crossing. A 2 km bridge may be forty spans of 50 m or one span of 1,500 m with approaches, and those are entirely different structures. Span decides almost everything: the type (girders to about 300 m, arches and trusses to 550, cable-stayed to 1,200, suspension beyond), the depth of the deck (roughly span/20 for concrete girders, span/25–30 for steel, span/18–25 for boxes), and the cost per square metre.

The span layout is chosen by balancing four things: what must be cleared (a navigation channel, a road, a flood plain), where piers can stand (ground, water depth, environmental limits), the economic balance between a heavier deck and fewer piers (section 1.8), and appearance, where odd numbers of spans with a longer central span usually look best. Continuous decks are proportioned with end spans about 0.8 of the interior spans (0.6–0.7 for balanced cantilevers) so that the moments balance. The span-type recommender suggests a type for a span, and the Bridge Design tool lays out the spans for a crossing and lets you move the piers.

Live load is the load of traffic, as opposed to the permanent dead load of the structure itself. No code designs for real trucks one by one; each defines a notional model that envelopes the legal traffic with a margin. In AASHTO LRFD that model is HL-93: a design truck of three axles (35 kN, 145 kN, 145 kN, the rear axle spacing varying from 4.3 to 9.0 m) or a design tandem (two 110 kN axles 1.2 m apart), whichever governs, placed together with a lane load of 9.3 kN per metre over a 3 m width; for negative moment over a pier two trucks 15 m apart at 90 per cent. A dynamic allowance of 33 per cent is added to the truck and tandem (not the lane load), and multiple-presence factors reduce the load when several lanes are loaded at once.

The live-load moment and shear at every point of a beam are found by moving the model along an influence line, and the share each girder takes is given by the distribution factors of AASHTO 4.6.2.2. Chapter 4.3 explains the model and its relatives (the Eurocode Load Model 1, Korea's KL-510, the fatigue truck, the permit vehicles), and the HL-93 tool and the distribution factor tool do the calculation.

A footbridge carries a fraction of a highway bridge's load, so it can be far lighter and more slender, and that is exactly its problem. People walk at 1.6–2.4 steps per second and can synchronise with a deck that sways or bounces near those frequencies; London's Millennium Bridge closed two days after opening in 2000 because of lateral lock-in, and pedestrian comfort limits on acceleration, or a requirement to keep natural frequencies away from the walking range, now govern the design of most light footbridges. Wind can excite the same slender decks.

The rest is proportion: live loads of about 4–5 kPa, accessibility gradients (usually 1:20 or less for ramps), railings and guardrails at 1.1–1.4 m, lighting, and, since footbridges are often the most visible small structures in a town, architecture: stress-ribbon, cable-stayed, arch and truss footbridges are common where a highway bridge would be a plain girder. Section 6.6 covers deflection, camber and vibration; the continuous beam analyser gives deflections and the wind load tool the pressures.

bridge

Any structure carrying a path over an obstacle. Owners draw a size line: in the United States a structure over 6.1 m (20 ft) along the road is a bridge and gets its own inspection and rating programme.

culvert

A shorter structure, usually a pipe, box or arch buried under the road with fill on top, carrying water or a small road through the embankment. Structurally it is designed with the soil, not above it.

viaduct

A long bridge of many similar spans carrying a road or railway over a valley, a city or flat ground; "viaduct" describes length, not structural type.

overpass / flyover / underpass

A bridge carrying one road over another; underpass is the same structure seen from the lower road; flyover is the British and Asian term for an elevated road at a junction.

aqueduct

A bridge carrying water (a canal or pipeline) rather than traffic.

footbridge / skybridge

A bridge for pedestrians; a skybridge links two buildings above street level.

Section 1.1 starts from the definition, and the glossary has the rest of the vocabulary.

The embodied carbon of a bridge is the greenhouse gas emitted to make its materials (life-cycle stages A1–A3), transport them and build it (A4–A5), expressed in kilograms of CO₂ equivalent per square metre of deck. Concrete and steel dominate: cement is about 0.8 t CO₂e per tonne, reinforcing steel around 2 t per tonne (0.9 or less from an electric-arc furnace), structural steel 2–2.5 t per tonne, prestressing strand slightly more. Typical figures are 500–1,000 kg CO₂e/m² for an efficient girder bridge, 1,000–2,000 for a heavy concrete viaduct or a steel composite deck, and above that for long-span cable-supported bridges; the IStructE's SCORBS proposal rates bridges from A (below 500) to G on that scale.

The levers are the same as for cost: a lighter type, fewer piers, lower-carbon concrete mixes (slag or fly ash replacing 40–50 per cent of the cement), recycled-content steel, and, over the life of the bridge, durability, because a deck that lasts 120 years instead of 60 halves everything. The Bridge Design tool computes an A1–A5 footprint with a breakdown table and suggests reductions for each design it draws.