Panama Canal Expansion: the 60% water trick that made a bigger canal drink less

On June 26, 2016, a Chinese-built container ship named Cosco Shipping Panama — 299 metres long, 48.2 metres wide, drawing 10.05 metres — eased through the new Agua Clara Locks on the Atlantic side of the Panama Canal and became the first commercial vessel to transit the expanded waterway. 1 The moment was nearly two years behind the original August 2014 target, delayed by a construction stoppage, a budget dispute that halted work for a month, and a cracked concrete sill that no one could easily explain. But it worked.

What the photographs from that day did not convey was the most technically demanding aspect of what had just been built. The new Agua Clara (Atlantic) and Cocolí (Pacific) lock complexes handle vessels nearly three times the cargo capacity of anything the original 1914 locks could accept. Their chambers are the largest concrete lock chambers ever constructed. And yet, according to the Panama Canal Authority (ACP) and the engineering contractors who designed the hydraulics, the expanded canal uses 7% less fresh water per transit than the original locks — despite the far larger ships. 2 3

That is the engineering feat worth examining.

Why a third lane, and why the constraint was always water

By the mid-2000s, the original Panama Canal was running out of capacity. ACP projected it would reach maximum sustainable throughput by around 2012. 4 The reason was straightforward: global container shipping had shifted toward vessels carrying 8,000 to 12,600+ TEUs (twenty-foot equivalent units). Those ships were simply too wide for the original lock chambers, which max out at 33.5 metres. Any ship wider than 32.3 metres — the actual usable beam given mooring and fender clearances — could not transit. 3

Enlarging the existing Gatun, Pedro Miguel, and Miraflores locks was not a realistic option. Doing so would have required closing the canal for years, forfeiting roughly $2 billion in annual toll revenue per year of closure, and delivering no income while construction proceeded. The third-lane approach — building two completely new lock complexes in parallel with the operating canal — was the only path that preserved continuous revenue while expanding capacity. 5

The political case was sealed on October 22, 2006, when Panamanian citizens voted 76.8% in favour of the expansion in a national referendum. President Martín Torrijos framed it as Panama's transformation into a "First World" country; ACP projected a 12% internal rate of return and 3.5% annual toll growth over 20 years. 1 Total project cost was estimated at $5.25 billion — covering design, construction, environmental mitigation, and commissioning. 3

But behind the economics was a constraint that would shape every major engineering decision. Gatun Lake — the artificial reservoir that sits between the two sets of locks at 26.7 metres above sea level and provides the elevation difference that lifts ships across the continental divide — is simultaneously the canal's working fluid and the primary drinking-water source for roughly half of Panama's population. Every transit of the original 1914 locks releases approximately 52 million US gallons (about 200 million litres) of fresh water from the lake to the sea. 6 A larger lock chamber would require even more water per lockage, straining a lake whose level already fluctuates seasonally and was expected to face growing demand.

As Stantec, the engineering firm that led the water-saving basin design, framed it: the ACP "had to decide: modernize or risk becoming a tourist attraction." 2 The decision to modernise was straightforward. The engineering challenge — making a much larger lock system that consumed less water than the smaller one it was joining — was not.

What was built: lock dimensions and the New Panamax design vessel

The Third Set of Locks consists of two complexes: Agua Clara on the Atlantic (Caribbean) side and Cocolí on the Pacific side. Each complex has three lock chambers in series, stepping vessels up or down the ~27-metre elevation change between sea level and Gatun Lake in approximately three equal stages. 3

The chamber dimensions define the vessel class the entire project was designed around:

Parameter	Original locks (1914)	Third Set of Locks (2016)
Chamber length	304.8 m	427 m
Chamber width	33.5 m	55 m
Chamber depth	12.8 m	18.3 m
Max vessel length	~289–294 m	366 m
Max vessel beam	32.3 m	49 m
Max vessel draft	12 m	15.2 m
Max cargo capacity	~4,500–5,000 TEU	~12,600–14,500 TEU

3 6

The margins between the maximum New Panamax vessel and the lock chamber walls are narrow to a degree that demands operational precision. A 49-metre-wide ship in a 55-metre-wide chamber leaves 6 metres total clearance — 3 metres each side. 3 The original locks used electric locomotive "mules" mounted on tracks atop the lock walls to haul vessels through on cables, maintaining precise lateral position. The Third Set of Locks — whose wider chambers made the mule-track geometry impractical — uses tugboats positioned inside the chamber itself, working in more confined and dynamically complex conditions. Worldsensing, a sensor and monitoring firm that analysed the completed expansion, noted that wind conditions affecting vessel maneuverability in the narrow chamber, combined with tugboats operating at close range, elevate collision risk with the lock walls. 5

The chambers are approximately 60% larger by volume than the original, requiring 4.3 to 5.0 million cubic metres of concrete. 7 As a calibration point: the original 1914 canal required approximately 230 million cubic metres of excavation across its entire construction. The Third Set of Locks added roughly 74 million cubic metres in excavation and 7.1 million cubic metres in dredging for the approach channels. 7 3

Comparative diagram of maximum ship sizes for the Panama Canal (old Panamax vs New Panamax) alongside Suez Canal and Strait of Malacca maximums, showing hull profiles and draft values — Maximum vessel sizes by waterway: old Panamax (Length 965 ft, draft 39.5 ft) vs New Panamax (Length 1,200 ft, draft ~50 ft). Source: U.S. Energy Information Administration. 6

The water-saving basin system: recycling 60% by gravity

The single most consequential engineering decision of the Third Set of Locks was the water-saving basin (WSB) system, and it was not technically optional. Without it, a 427 × 55-metre chamber would have consumed roughly 500 million litres of fresh water per lockage — more than double the original locks' already-significant 200 million litres. 7 The WSB system reduces that to approximately 200 million litres by recycling water that would otherwise drain to the sea.

Each of the six lock chambers is flanked by three lateral basins at different elevations — 18 basins in total across both lock complexes (9 per complex). Stantec led the overall design; Tetra Tech designed the valves and bulkheads regulating the gravity-fed flows. 3

The operating principle divides the chamber's water volume into five conceptual horizontal layers of equal depth:

Emptying (ship descending):

As a ship locks down from Gatun Lake to the sea, the gate opens slightly and the chamber begins to empty. The upper three layers — layers 1, 2, and 3 — drain sequentially out of the chamber into the three lateral basins at decreasing elevation. Each basin captures one layer's worth of water and holds it. Only the lowest two layers (layers 4 and 5) drain all the way out to the lower pool and are lost to the system.

Filling (ship ascending):

As the next ship enters the lower pool and the chamber needs to be raised, the process reverses. The lowest-elevation basin (which captured layer 3) empties first back into the chamber, raising the water level by one layer's equivalent. Then the middle basin (layer 2), then the upper basin (layer 1). After all three basins have returned their water, the chamber still needs two more layers to reach Gatun Lake level — and only then does the culvert to Gatun Lake open to draw in fresh water.

The result: 3 out of 5 layers (60%) of every lockage use recycled water. 3 1

Schematic of the water-saving basin emptying operation: water from the lock chamber drains sequentially into three lateral basins at different elevations, recycling three of five equal water layers — Chamber emptying: layers 1–3 captured by basins A, B, and C; only layers 4–5 drain to sea. 3

Schematic of the water-saving basin filling operation: stored water returns from the three basins in reverse order before fresh water is drawn from Gatun Lake — Chamber filling: basins discharge in reverse order (C → B → A); fresh water from Gatun Lake supplies only the final two layers. 3

Critically, the entire system operates by gravity. No pumps are used anywhere in the Panama Canal — neither in the original 1914 locks nor in the Third Set of Locks. Water moves through main culverts (large-diameter tunnels cast inside the concrete walls) running longitudinally and distributing through lateral ducts at floor level. In the 1914 locks, water entered vertically through the floor; the new locks use horizontal floor-level distribution, which produces a more uniform flow pattern and reduces turbulence forces on the vessel. 3 Average flow rate during locking is approximately 550 m³/s, controlled by 158 valves distributed across the system. 3

To validate the design, ACP and the engineering team used two complementary methods: a 1:30 scale physical model built in Lyon, France, instrumented with over 100 sensors measuring water levels, velocities, pressures, and mooring forces on a model vessel; and supercomputer CFD (computational fluid dynamics) modelling that used 1D simplified models as inputs to 3D models at critical points such as valve locations and basin inlets. 3 The engineers described the hydraulic system in notably candid terms: "If something had not worked as expected, very little could have been done to solve it." 3 The physical model and CFD correlation was high enough to give confidence before a single cubic metre of production concrete was poured.

The rolling gates: 4,200 tonnes fabricated in Italy and sailed to Panama

The original 1914 locks use mitre gates — a V-shaped pair of hinged leaves that meet in the centre under water pressure, like an inverted wedge. They are elegant and mechanically simple, but they have one operational drawback: maintenance requires draining the affected chamber or deploying underwater divers. Neither approach is fast.

The Third Set of Locks uses rolling gates — large hollow steel parallelepipeds that slide horizontally on rails into concrete niches recessed into the lock walls. When open, a gate disappears into its niche. The niche is watertight and large enough to serve as a dry dock: a gate can be withdrawn for inspection and maintenance without removing it from the structure and without disrupting lock operations, because its sister gate at the same lock head continues working. 8 3 This was a deliberate operational decision, not a novelty for its own sake.

There are 16 rolling gates in total — 8 at Agua Clara, 8 at Cocolí. Every gate is the same length, 57.6 metres, but height, thickness, and weight vary with position:

Position	Height	Weight
Typical mid-complex gate	23–29 m	~2,300–3,319 tons
Pacific side lock head 4 (ocean-facing)	33 m	4,232 tons

9 3

The tallest, heaviest gates sit at the Pacific ocean-facing lock head because the Pacific has a tidal range of up to 6 metres — the gate must seal against the deepest possible water column. Atlantic tides are less than 0.5 metres, so the Atlantic gates are shorter and lighter.

Despite their mass, each gate operates in under five minutes. The mechanism designed by Cimolai Technology (Italy) uses twin cable winches with gearboxes delivering torque up to 330,000 Nm through upper and lower carriages running on rails. The gates are hollow — their internal buoyancy chambers reduce the effective operating weight by up to 85% of dry mass — which is why two winches can move a 4,000-tonne object in minutes rather than hours. Design life for the gates is 50 years. 3 7 As Webuild notes: "Each gate is expected to remain in service for at least another hundred years." 7

All 16 gates were fabricated by Cimolai SpA at their facility in Italy, beginning October 2011. They were transported in four shipments of four gates each from the Port of Trieste aboard specialized heavy-lift Post-Panamax vessels — a 25-day Atlantic crossing per shipment. 9

Shipment	Date	Notes
1	August 20, 2013	First 4 gates arrive
2	June 10, 2014	—
3	September 7, 2014	—
4	November 12, 2014	Final 4 gates (Atlantic side, 3,319 tons each, 29 m tall)

The Pacific-side gates could not transit the Atlantic on the same vessel as the Atlantic-side gates; they had to reach the Pacific construction site by travelling through the existing Panama Canal on barges — a logistical detail that required careful scheduling around operational transit windows. Total steel in all gates: 51,000 tonnes. 3 ACP Administrator Jorge L. Quijano marked the final arrival: "With this Expansion, the Panama Canal will remain competitive, opening new markets and possibilities for international trade through the waterway." 9

Concrete designed to survive 100 years in a salinity gradient

Every major concrete structure in the Third Set of Locks was specified to achieve a 100-year service life. That is not a marketing claim; it was a contractual requirement from ACP that drove a specific and non-trivial engineering process. 4

The fundamental challenge is chloride ingress. The lock chambers are not uniformly exposed to seawater. At Cocolí (Pacific side), the lower chamber faces full-salinity ocean water. The upper chamber, adjacent to Gatun Lake, sees nearly fresh water. The middle chamber sees a blend that shifts with each lockage. Reinforced concrete fails when chloride ions diffuse through the cover concrete and reach the steel reinforcement, initiating corrosion that eventually cracks and spalls the cover. The time to that point depends on the chloride concentration, the concrete permeability, and the cover depth — all of which vary by location within the same lock complex.

SIMCO Technologies was contracted to design location-specific concrete mixtures meeting the 100-year specification, using the U.S. Navy UFGS (Unified Facilities Guide Specifications) science-based protocol and their own STADIUM® service-life prediction software. 4 The team drilled and analysed cores from the existing Miraflores Locks — built in 1914 and still in service — to calibrate local exposure conditions and concrete durability behaviour before specifying the new mixtures. A structure already over 90 years old under comparable conditions is as good a control sample as engineers can reasonably obtain. 4

Two concrete types resulted:

Internal Mass Concrete: used for the core of the monolith walls, which at up to 30 metres tall and 27 metres wide at the base generate significant heat of hydration during curing. The mix was designed to limit temperature differentials and cracking from thermal gradients.
Structural Marine Concrete: used for all surfaces with direct exposure to water, particularly wing walls in permanent contact with saline or brackish water. Higher resistance to chloride-induced corrosion.

The seismic design requirement for structural concrete was 0.72g — a consequence of the Panama Canal sitting in a moderate-seismic zone. 2 Reinforced steel was used throughout the new locks, which was not technically available at the original 1914 construction scale. The main culverts — 18-foot (roughly 5.5-metre) diameter tunnels running longitudinally inside the wing walls — are housed within the triangular-profile wing walls whose broad base functions like a gravity dam anchoring the chamber against hydrostatic pressure. 3

Construction scale: peak concrete manufacturing rate was 540 m³/hour, with a peak daily pour of 5,000 m³. 3 Over the full project, 1.6 million tonnes of cement and 269,000–290,000 tonnes of reinforcing steel went into the lock structures. 4 3

Building it: 112 million person-hours across seven years

The main construction contract was awarded on July 15, 2009 to Grupo Unidos por el Canal (GUPC) — a consortium of Sacyr (Spain, lead partner), Salini Impregilo (Italy, now Webuild), Jan De Nul (Belgium), and Constructora Urbana S.A. (CUSA, Panama). 3

GUPC's winning bid was $3.12 billion — the only proposal below ACP's allocated budget of $3.48 billion. The next-lowest bid, from a Bechtel-Taisei-Mitsubishi consortium, was over $1 billion higher. 1 That gap, in retrospect, was already a warning sign.

At peak, 14,000 workers were on site simultaneously; the total workforce across the project's duration reached 40,000 people from 79 nationalities. 3 The design effort consumed 3 million person-hours; construction consumed 112 million. The project produced 300 operating manuals delivered to ACP, 2,000 km of electrical cable, 400 km of fibre optic cable, and over 100,000 control system signals per complex. 7

The construction panorama is large enough to be almost abstract:

Panoramic view of the Agua Clara lock complex under construction, 2013, with concrete walls reaching final height and cranes visible throughout — Agua Clara lock complex under construction, 2013: concrete monolith walls at near-final height, approach channel excavation continuing to the left. 1

Among the less-visible construction elements was the Borinquen Dam — roughly 3 km of loose-material embankment dam, 37 metres high, with a 30-metre-wide upper berm, constructed to separate the Pacific lock complex from the Miraflores Lake excavation area and connect the new Pacific locks to the existing waterway system. 3 At peak, 2,000 trucks per day moved through the Pacific construction zone alone. 3

Pietro Salini of Webuild captured the scale at the opening ceremony: "Seven years ago we began a long journey that represented the dream as well as the challenge that every entrepreneur and every person would want to have at least once in their lifetime: building a project that will change global trade." 7

The 2014 crisis and what the engineering disputes revealed

On February 5, 2014, GUPC halted all construction on the Third Set of Locks. Thousands of workers were turned away from the gates. 10

The ostensible reason was money: GUPC claimed $1.63 billion in cost overruns beyond the $3.12 billion contract, and argued that ACP had provided faulty information about site conditions. By that date, ACP had already paid $2.83 billion plus $180 million in previously approved overruns — yet the project was more than 70% complete. GUPC's position, stated publicly: "The GUPC investors are construction companies, not banks. It is unjust and impossible for the ACP and Panama to expect that private companies will finance $1.6 billion in costs on a project that was to be fully funded by the ACP." 10 ACP's administrator Jorge Quijano responded: "The Panama Canal is not going to submit to blackmail." 10

Work resumed under a partial agreement in March 2014. But the core engineering dispute never fully resolved during construction. It resurfaced nine years later in international arbitration.

In August 2018, Sacyr S.A. — GUPC's lead partner — filed a claim at ICSID (International Centre for Settlement of Investment Disputes) under the Spain-Panama bilateral investment treaty, seeking over $2.6 billion. The two core engineering allegations were: (1) that ACP had withheld geotechnical test data showing the basalt aggregate available from the Pacific Lock Excavation (PLE) site was lower quality than specified, significantly increasing concrete aggregate processing costs; and (2) that ACP knew before bid submission that seismic design requirements for the new rolling gates were more stringent than the bid documents indicated. 11

On October 31, 2025, the ICSID tribunal issued a unanimous award: all of Sacyr's claims were dismissed. Sacyr was ordered to pay Panama's arbitration costs — over $6 million. 11 The tribunal found that Sacyr had not demonstrated that ACP's conduct constituted an exercise of sovereign authority (puissance publique); the disputes were contractual in nature, and the bilateral investment treaty does not serve as a second-chance venue for contractual claims already before commercial arbitration. As the Kluwer Arbitration Blog's analysis summarised: "The award demands more than the mere presence of the State in the contract, more than the damage suffered by the investor, and more than the recasting of old contractual grievances in treaty language." 11

HFW, the maritime and infrastructure law firm, drew the practical lesson for contractors: the existence of a bilateral investment treaty "does not necessarily provide a second chance for contractual claims." They advised that contractors "should undertake their own due diligence during the tender process to verify the accuracy and reliability of site data provided by the employer." 12

The Cocolí sill crack in August 2015 — when a crack was found in the concrete sill where lock gates seat when closed — added a separate complication. Initially assessed as non-critical, it was reclassified by November 2015 as a schedule threat and required reinforcement work that pushed the opening date from April 2016 to June 2016. The repair was completed in February 2016. 1 Together with the 2014 work stoppage, these two events pushed the canal's opening nearly two years past the original August 2014 target.

Panama Canal Third Set of Locks — Key metrics

Design specifications and FY2025 operational performance

Lock chamber size

427 × 55 × 18.3 m

Side clearance (max vessel)

3 m per side

Water recycled per lockage

Rolling gates (total)

Concrete placed (est.)

4.3–5.0 M m³

Workers (total project)

Total project cost (ACP estimate)

$0.00

FY2025 toll revenue

$0.00+14.4%YoY

FY2025 total transits

0+19.3%YoY

統計カードを読み込んでいます…

Since 2016: record revenue, a drought that almost broke the model, and the fix that isn't ready yet

The expansion's first decade of operation divides cleanly into two phases: a period of strong performance that validated the investment thesis, then a climate-driven operational crisis that exposed the water dependency the engineers had always known was there.

From opening in June 2016 through mid-2023, the Third Set of Locks performed broadly as designed. Within the first 20 months of operation, 3,000 New Panamax ships had transited the expansion. Toll revenue and transit volumes grew year-on-year. 1

Then, in 2023, Panama experienced its third-driest year in 143 years of records. Rainfall during the critical May–December wet season ran about 30% below the long-term average; October 2023 was the driest October since 1950. 13 Gatun Lake fell to 79.6 feet (24.3 metres) in August 2023 — well below the normal operating range. By January 1, 2024, the lake was at its lowest recorded level for any January in the entire data series, nearly 6 feet below the January 1, 2023 level. 14

Recorded Gatun Lake water levels 1965–2024 showing 2023 and 2024 values falling below all prior years by late dry season, with 2016 and 1998 El Niño comparison years highlighted — Gatun Lake daily water levels, 1965–2024. The 2024 line (orange) starts the year lower than any prior January on record. Data: Panama Canal Authority, via Woodwell Climate Research Center. 15

ACP's response was to ration capacity rather than allow the lake to drop below safe operating minimums. Daily transits were cut from the normal 36–38 to 24 by November 2023, and to 18 per day by February 2024 — the lowest figure in the canal's history. 16 The maximum authorised draft for New Panamax vessels was reduced from 50 feet (15.2 m) to 44 feet (13.4 m), forcing shipping companies to reduce cargo loads and absorb additional fuel costs per tonne carried. 14 In FY2024, ACP estimates it lost approximately 2,211 transits to the drought restrictions. 16 One Japanese shipping consortium paid a $3.9 million auction premium to jump the queue for a single transit. 16

The scientific attribution is clear. A World Weather Attribution study found that El Niño reduced 2023 rainy-season precipitation in Panama by approximately 8%, and that the probability of such low rainfall occurring in an El Niño year is about 5% — roughly once every 40 years under current climate conditions. 13 The Smithsonian Tropical Research Institute's Steven Paton stated directly: "2023 was the third driest year ever recorded in Panama in the 143 years that we have data." 13 A September 2025 paper in AGU Geophysical Research Letters modelled Gatun Lake water levels under climate scenarios and found that under high-emissions pathways, lake levels decline meaningfully, while low-emission pathways produce relative stability — though both scenarios assume increasing variability around El Niño cycles. 17

The 2024–2025 La Niña transition brought persistent rainfall that refilled the reservoirs. By March 2026, Gatun and Alajuela reservoirs had reached their highest recorded March levels ever — Alajuela at approximately 99% capacity, Gatun above 90%. 18 ACP confirmed in May 2026 that no transit restrictions were planned for the rest of the year. Daily transits stood at 38, with peak days reaching 43. 18 FY2025 (ending September 30, 2025) produced $5.7 billion in revenue — up 14.4% year-on-year — on 13,404 transits, up 19.3%. ACP transferred $2.965 billion to the Panamanian national treasury, a 20% increase over the prior year. 19

The underlying problem, however, has not been engineered away. The water-saving basins recycle 60% of each lockage — but that 40% loss still accumulates to roughly 200 million litres per transit from Gatun Lake. In 2006, the expansion was designed around a lake that the designers expected to face growing demand. In 2024, the ACP was rationing that lake at historically low levels. The WSBs did their job; the lake still ran dry.

The structural response is the Río Indio Multipurpose Reservoir — a $1.6 billion project approved by ACP's board of directors on February 21, 2025 (Resolution No. ACP-JD-RM 25-1542). 20 21 The design: a concrete-faced rockfill dam across the Indio River, creating a reservoir with a total storage capacity of 1.577 billion cubic metres, connected to Gatun Lake via an 8,350-metre tunnel (4.5 m diameter). 20 Of the $1.6 billion, $400 million is allocated to acquiring and resettling approximately 2,000 residents and negotiating with roughly 12,000 property rights holders. 20 Construction is expected to begin in 2027, with operations by 2031–2032.

ACP hydrological manager Ayax Murillo was candid about the gap between recovery and security: "What concerns us is the 2027 dry season, given that the Río Indio reservoir is not yet operational." 18 A new El Niño watch was issued by NOAA in April 2026, forecasting possible onset by mid-2026; the probability of a "very strong" event is approximately 25%. 22

Legacy: what the third lock set changed, and what it didn't

The Third Set of Locks did not invent water-saving basins — the concept predates this project. What the Panama expansion demonstrated at scale was a specific architecture: three basins (rather than two or four), operated entirely by gravity, validated with a 1:30 physical model before any concrete was poured, and integrated into a lock chamber three times the cargo capacity of what had existed before. That combination — no pumps, 60% water recovery, physically validated hydraulics — is a template that any country building or upgrading a lock-based waterway with freshwater constraints can now study with a built reference.

The rolling gate design established a practical precedent for maintenance without operational shutdown. Niche-as-dry-dock is not a new concept, but applying it at 57.6-metre gate length, 4,200-tonne gate weight, and <5-minute cycle time — with buoyancy chambers reducing operating load by 85% — validated a specific scale range that had previously been theoretical.

The 100-year service life specification, backed by service-life prediction software and calibrated against cores from the 1914 locks, marked a shift in how concrete durability is specified for marine infrastructure. The standard practice of prescriptive mix specifications (cement type, water-cement ratio, minimum strength) gives way to predictive modelling of the actual chloride diffusion pathway in the actual exposure environment. That approach is now available as a documented project outcome.

The arbitration result — Sacyr v. Panama, ICSID UNCT/18/6, October 31, 2025 — will be cited in construction law for years. The dismissal of $2.6 billion in claims confirmed that investment treaties do not function as a second round of commercial dispute resolution, and that a contractor who wins a bid 10% below the owner's own estimate and $1 billion below the nearest competitor bears significant responsibility for validating the owner's site data before signing. The engineering disputes at the heart of the case — basalt aggregate quality and seismic design factors — were real technical disagreements, but the tribunal found they belonged in the ICC arbitration forum where contractual disputes go, not in treaty arbitration. The practical lesson for contractors on fixed-price infrastructure bids is unchanged: the site investigation you fund yourself is the one you can rely on.

The drought of 2023–2024 revealed something the water-saving basins cannot fix. The expansion reduced each lockage's water consumption from approximately 500 million litres to 200 million litres — a genuine engineering achievement. But 200 million litres per transit, multiplied by 30+ transits per day, is still a large daily withdrawal from a lake that also supplies drinking water to three million people and whose level is governed by rainfall that climate models suggest will become more variable. The WSBs bought capacity without buying water security. That purchase — the Río Indio reservoir — costs $1.6 billion more, will not be ready until 2031–2032, and will require the displacement of 2,000 people from a valley that will become a lake.

ACP administrator Ricaurte Vásquez Morales put the structural position plainly in 2024: "Aspects such as climate change and reduced rainfall as well as the increase in population have put the country in an urgent position to make decisions to ensure the future availability of the resource." 20

The 2016 expansion doubled the canal's annual capacity and opened it to the generation of container ships that dominate global trade. It did so on budget for the ACP (the overrun dispute was a contractor problem, not an owner shortfall in total programme cost), more or less on time, and with engineering systems — the water-saving basins above all — that have performed as designed. What no amount of basin engineering can change is that the canal runs on rainfall. The third set of locks solved the ship-size problem. The water-security problem is still being built.

Cover image: New Panamax tanker Maran Helios transiting Agua Clara Locks, 2019. Photo via Wikimedia Commons, CC BY-SA 4.0.