DAM
The Opening
Stand at the base of Hoover Dam. Look up. 221 meters of concrete — 6.6 million tonnes — holding back 35 billion cubic meters of water. The water pushes with 1.5 million tonnes of horizontal force. The dam pushes back. This standoff has lasted since 1936.
If it fails, a wall of water 30 meters high hits Las Vegas in 7 hours.
Requirements:
├── Hold back an entire river
├── Generate 2,080 MW of electricity
├── Survive 1,000-year floods
├── Last 100+ years
├── Don't crack
├── Don't slide
└── Don't overtop
Let's build one.
───
PHASE 1: Hold Back the Water
Fill a bathtub. Put your hand flat against the side at the bottom. You feel pressure. Now imagine the bathtub is 221 meters deep and 379 meters wide. That's what the upstream face of Hoover Dam feels every second of every day.
Water has weight. A column of water 1 meter tall, 1 meter square, weighs 1,000 kg. The deeper you go, the more water sits above you, the more it pushes outward. This is hydrostatic pressure:
P = ρgh
Where:
├── ρ = water density = 1,000 kg/m³
├── g = gravitational acceleration = 9.81 m/s²
└── h = depth below the surface (meters)
At the bottom of Hoover Dam (h = 221 m):
P = 1,000 × 9.81 × 221
P = 2,168,010 Pa ≈ 2.17 MPa ≈ 21.7 atmospheres
At the surface: zero. At the bottom: 21.7 atmospheres. The pressure increases linearly with depth — a triangle of force.
Total Force on the Dam Face
The pressure isn't uniform. It's zero at the top and maximum at the bottom. For a vertical dam face of height H and width L, the total horizontal force is the integral of pressure over the area:
F = ∫₀ᴴ ρg·h·L dh = ½ρgH²L
For Hoover Dam (H = 221 m, crest length L = 379 m):
F = ½ × 1,000 × 9.81 × (221)² × 379
F = ½ × 1,000 × 9.81 × 48,841 × 379
F = 90.8 × 10⁹ N ≈ 90.8 GN
That's 9.26 million tonnes of horizontal force. Equivalent to parking 6.2 million cars against the dam face and letting them push.
Water surface
─────────────────── ← P = 0
│░░░░░░░░░░░░░░░░│
│░░░░░░░░░░░░░░░░│ Pressure increases
│▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒│ linearly with depth
│▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒│
│▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓│ ←── triangular
│▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓│ distribution
│████████████████│
│████████████████│
│████████████████│ ← P = ρgH = 2.17 MPa
═══════════════════
Dam base
Resultant force acts at H/3 from the base
(centroid of a triangle = 1/3 from the wide end)
F_total = ½ρgH²LThe force resultant acts at 1/3 of the height from the base — not at the midpoint. This matters enormously for overturning calculations: the lever arm is shorter, but the force is concentrated low where the dam is thickest.
Why the Triangular Profile
The pressure triangle explains the dam's shape. At the top, pressure is near zero — the dam can be thin. At the bottom, pressure is enormous — the dam must be thick. This is why every gravity dam is triangular in cross-section:
Hoover Dam dimensions:
├── Crest (top): 13.7 m thick — just wide enough for a road
├── Base (bottom): 201 m thick — wider than a football field
└── The dam weighs more per meter as you go down, matching the rising pressure
The ratio of base to height: 201/221 = 0.91. Nearly 1:1. The dam is almost as wide as it is tall.
DESIGN SPEC ESTABLISHED:
├── Hydrostatic pressure: P = ρgh — linear with depth
├── Total horizontal force: F = ½ρgH²L — proportional to height SQUARED
├── Hoover: 221 m tall, force = 90.8 GN (9.26 million tonnes)
├── Pressure at base: 2.17 MPa (21.7 atmospheres)
└── Resultant acts at H/3 from base — drives triangular cross-section
───
PHASE 2: Choose Your Shape
You can fight water with brute mass — pile enough concrete that the dam is simply too heavy to push over. Or you can be clever — curve the dam, redirect the water's force into the canyon walls, and use 80% less concrete. Both have been built. Both work. The choice depends on what the valley gives you.
Gravity Dam — Win by Weight
A gravity dam resists water by sheer mass. The water pushes horizontally. The dam's weight pushes vertically. If the weight's stabilizing moment exceeds the water's overturning moment, the dam stands.
The overturning moment about the downstream toe:
M_water = F × H/3 = ½ρgH²L × H/3 = ρgH³L/6
The stabilizing moment from the dam's weight (assuming triangular cross-section, base width B):
M_weight = W × 2B/3
Where W = ½ × B × H × L × ρ_concrete × g
For stability: M_weight > M_water
Factor of safety requirement: M_weight / M_water ≥ 1.5
crest road
◄─13.7m─►
┌────────┐ ← water level
│ │
water │ │╲
░░░░░░░│ CONCRETE│ ╲
▒▒▒▒▒▒▒│ │ ╲
▓▓▓▓▓▓▓│ │ ╲
███████│ │ ╲
███████│ │ ╲ downstream
│ │ ╲ face
└──────────────┴───────┘
◄──── 201 m base ────────►
Forces:
← F_water = ½ρgH²L (at H/3 from base)
↓ W_dam = weight of concrete (at 2B/3 from toe)
↑ U_uplift = water under the dam (reduces effective weight)The dam must be heavy enough that its weight moment about the downstream toe exceeds the water's overturning moment by at least 1.5×. Hoover Dam uses 6.6 million tonnes of concrete to achieve this.
Arch Dam — Win by Geometry
An arch dam curves upstream. When water pushes against it, the curve redirects the horizontal force into compression along the arch — transferred directly into the canyon walls. The same principle as a stone arch bridge, rotated 90 degrees.
Arch stress (thin cylinder approximation):
σ = P × r / t
Where:
├── P = hydrostatic pressure at that depth
├── r = radius of the arch curve
└── t = thickness of the dam at that depth
For the arch to work, the canyon walls must be competent rock — typically granite or gneiss with compressive strength > 50 MPa.
canyon wall canyon wall
████████ ████████
████████╲ ╱████████
████████ ╲ water ╱ ████████
████████ ╲ ░░░░░░░░░░░░░ ╱ ████████
████████ ╲ ░░░░░░░░░░░░ ╱ ████████
████████ ╲░░░░░░░░░░░╱ ████████
████████ ╲─────────╱ ████████
████████ ╲ dam ╱ ████████
████████ ╲─────╱ ████████
████████ ╲───╱ ████████
████████ ████████
downstream
Water pushes → arch compresses → force goes into canyon walls
The dam itself barely needs to resist bending — it's in pure compression
Concrete comparison:
├── Gravity dam (Hoover): 6.6 million tonnes
├── Arch dam (same height): ~1.3 million tonnes
└── Savings: ~80%Hoover is actually an arch-gravity hybrid — it curves upstream with a radius of 201 m, transferring about 60% of the load into the canyon walls and carrying 40% by gravity. Pure arch dams like Vajont (262 m) use even less concrete.
Choosing: The Valley Decides
Dam type selection matrix:
├── Narrow V-shaped canyon + strong rock → Arch dam
│ (Vajont, Inguri, Xiaowan)
├── Wide valley + any foundation → Gravity dam
│ (Grand Coulee, Three Gorges)
├── Very wide valley + earth available → Embankment dam
│ (Tarbela, Oroville, Nurek)
└── Moderate canyon + strong rock → Arch-gravity hybrid
(Hoover, Glen Canyon)
The world's tallest dams are all arch: Jinping-I at 305 m. The world's largest by volume are all embankment: Tarbela at 148.5 million m³. Each shape optimizes for different geology.
DESIGN SPEC UPDATED:
├── Gravity dam: resists by weight, M_weight ≥ 1.5 × M_water
├── Arch dam: transfers load to canyon walls via compression, uses 80% less concrete
├── Arch stress: σ = Pr/t — thinner dam needs stronger rock abutments
├── Hoover: arch-gravity hybrid, 60% arch action, 40% gravity
└── Valley shape + rock quality determine dam type — not engineering preference
───
PHASE 3: Don't Slide
The water doesn't just try to push the dam over. It tries to push it downstream — slide it along its foundation like a book across a table. And it has an accomplice: water seeping under the dam, lifting it up, reducing friction. This is the failure mode that kills.
The Sliding Equation
The dam stays put if friction on the base exceeds the horizontal water force:
F_friction = μ × N
Where:
├── μ = friction coefficient between concrete and rock (typically 0.65-0.75)
└── N = normal (vertical) force = dam weight minus uplift
The sliding factor of safety:
FoS = μ × (W - U) / F_h
Where:
├── W = total dam weight
├── U = uplift force (water pressure under the dam)
└── F_h = total horizontal water force
Minimum required: FoS ≥ 1.5 for normal conditions, ≥ 1.0 for extreme flood + earthquake combined.
The Uplift Problem
Water doesn't stop at the upstream face. It seeps into cracks, joints, and pores in the rock foundation. This creates an upward pressure under the entire dam base.
upstream downstream
face face
↓ water ↓
████████████████████████████████████████████
████████████ DAM CONCRETE ████████████████
════════════════════════════════════════════
↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑
full uplift pressure zero
head (drained)
Without drainage:
U = ½ × ρgH × B × L (triangular distribution)
With drain gallery (at 1/3 from upstream face):
┌───────────┬─────────────────────────────┐
│ │drain │
│ full head │gallery reduced pressure │
│ ↑↑↑↑↑↑↑ │↑↑↑↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓ │ zero
└───────────┴─────────────────────────────┘
Effective uplift reduced by ~40-60%Without drainage, uplift can reduce the dam's effective weight by 30-50%. The drain gallery — a tunnel inside the dam near the upstream face — intercepts seepage and reduces uplift to safe levels. Every modern dam has one.
The Drain Gallery
Walk inside Hoover Dam and you find a network of tunnels running through its body. The drain gallery sits about 1/3 of the base width from the upstream face. Vertical drains drilled into the foundation below intercept seeping water before it builds pressure.
Hoover Dam's drainage system:
├── Main gallery: 1.5 m wide × 2.4 m tall — big enough to walk through
├── Vertical drain holes: 76 mm diameter, drilled every 3 m
├── Drain depth: 40% into foundation rock
├── Water collected: ~0.5 m³/s continuously
└── Monitored flow rate: increasing flow = warning sign of internal erosion
If the drains clog or fail, uplift rises. Effective weight drops. Factor of safety falls. This happened at Malpasset Dam in 1959 — uplift in fractured rock reduced friction below the sliding threshold. The dam failed in minutes. 423 people died.
DESIGN SPEC UPDATED:
├── Sliding FoS = μ(W-U)/F_h ≥ 1.5 (normal), ≥ 1.0 (extreme)
├── Uplift: water pressure under dam reduces effective weight 30-50%
├── Drain gallery at 1/3 base width from upstream face — reduces uplift 40-60%
├── Concrete-rock friction μ ≈ 0.65-0.75
└── Malpasset 1959: uplift in fractured rock → sliding failure → 423 dead
───
PHASE 4: Don't Crack
Pour a sidewalk on a hot day. Come back the next morning. Cracks. The concrete heated up from its own chemical reaction, expanded, then cooled and contracted. The tension from contraction exceeds concrete's tensile strength — which is only 3-5 MPa, about 10% of its compressive strength. Now imagine pouring 6.6 million tonnes of concrete at once.
The Hydration Heat Problem
Concrete doesn't dry — it cures. Portland cement reacts with water in an exothermic reaction:
C₃S + H₂O → C-S-H gel + Ca(OH)₂ + heat
Each cubic meter of concrete generates roughly 250-300 kJ/kg of heat from hydration. For a mass concrete pour like a dam, the interior temperature can rise 50-60°C above ambient.
The problem: the surface cools quickly, but the interior stays hot for years. This temperature difference creates internal stress:
σ = E × α × ΔT
Where:
├── E = Young's modulus of concrete ≈ 25 GPa
├── α = thermal expansion coefficient ≈ 10 × 10⁻⁶ /°C
└── ΔT = temperature difference (°C)
For ΔT = 30°C (surface vs interior):
σ = 25 × 10⁹ × 10 × 10⁻⁶ × 30
σ = 7.5 MPa
Concrete tensile strength: 3-5 MPa. The thermal stress EXCEEDS the tensile strength. It will crack.
Hoover's Solution — The Cooling Pipes
Engineers embedded 937 km of 1-inch steel pipe throughout Hoover Dam. Refrigerated water at 4°C was pumped through these pipes continuously during and after construction.
┌──────────────────────────────────────┐
│ ─────○─────────○─────────○───── │
│ │
│ ─────○─────────○─────────○───── │ ○ = cooling pipe
│ │ (1-inch steel)
│ ─────○─────────○─────────○───── │
│ │ Horizontal spacing: 1.5 m
│ ─────○─────────○─────────○───── │ Vertical spacing: 1.5 m
│ │
│ ─────○─────────○─────────○───── │ Water temp in: 4°C
│ │ Water temp out: ~25°C
│ ─────○─────────○─────────○───── │
└──────────────────────────────────────┘
Without cooling pipes: interior reaches ~70°C
With cooling pipes: interior stays below ~25°C
Time to naturally cool (no pipes): ~125 yearsWithout the cooling system, Hoover Dam's interior would still be warm today — 90 years after construction. The pipes reduced cooling time from over a century to about 2 years per block.
Block Pour Construction
Hoover Dam wasn't poured as a monolith. It was built as 230 interlocking columns, each about 15 m × 15 m, poured in lifts of 1.5 m (5 feet) at a time. Each lift was allowed to cool before the next was placed.
Why columns instead of one mass:
├── Each column shrinks independently → no crack-inducing strain between blocks
├── Contraction joints between blocks filled with grout after cooling
├── Vertical keys (shear keys) interlock the blocks for structural continuity
└── If one block cracks, the crack stops at the joint — it can't propagate
After cooling, the contraction joints were pressure-grouted at 2-4 MPa, bonding the 230 columns into a monolithic structure. The dam effectively assembles itself after construction.
DESIGN SPEC UPDATED:
├── Hydration heat: ~300 kJ/kg, interior temp rise 50-60°C
├── Thermal stress: σ = EαΔT — exceeds tensile strength at ΔT > 20°C
├── Cooling pipes: 937 km of pipe, 4°C water, reduced cooling from 125 years to ~2 years
├── Block pour: 230 columns × 1.5 m lifts, contraction joints grouted after cooling
└── Concrete tensile strength only 3-5 MPa — thermal cracking is the primary risk
───
PHASE 5: Spin the Water
You've got 35 billion cubic meters of water sitting 180 meters above the river. That's potential energy — gravitational potential energy — and you're about to convert it to electricity. A steel pipe drops the water through the dam. A turbine spins. A generator produces current. Las Vegas lights up.
The Power Equation
Hydroelectric power from first principles — it's just potential energy per unit time:
P = ρ × g × Q × H × η
Where:
├── ρ = water density = 1,000 kg/m³
├── g = 9.81 m/s²
├── Q = volumetric flow rate (m³/s)
├── H = effective head (height of water above turbine) (m)
└── η = overall efficiency (turbine × generator)
For one of Hoover's 17 generating units:
Q = 56 m³/s (per turbine)
H = 180 m (net head varies with reservoir level)
η = 0.92 (Francis turbine + generator combined)
P = 1,000 × 9.81 × 56 × 180 × 0.92
P = 91 × 10⁶ W = 91 MW per turbine
Total installed capacity: 17 units × varying sizes = 2,080 MW
Annual generation: ~4.2 TWh — enough for 450,000 homes
The Francis Turbine — Spinning at 180 RPM
Hoover uses Francis turbines — the most common type for medium-to-high head applications (40-700 m). Water enters radially (from the sides) and exits axially (downward).
penstock (high pressure pipe)
│
▼
┌──────────────────────┐
│ spiral case │ ← distributes water
│ ╔══════════╗ │ around the turbine
│ ║ guide ║ │
─────┤ ║ vanes ║ ├─────
water │ ║→ ╔═══╗ ←║ │ water
in │ ║→ ║ R ║ ←║ │ in
─────┤ ║→ ║ U ║ ←║ ├─────
│ ║→ ║ N ║ ←║ │
│ ║→ ║ N ║ ←║ │
│ ║ ║ E ║ ║ │
│ ║ ║ R ║ ║ │
│ ╚══╚═══╝══╝ │
└──────────┼───────────┘
│
▼
draft tube (decelerates flow)
│
▼
tailwater (downstream river)
Guide vanes: adjustable — control flow rate and power output
Runner diameter (Hoover): ~5 m
Rotation speed: 180 RPM
Efficiency: 93-95% at design pointWater enters the spiral case at high pressure, passes through adjustable guide vanes that control the angle and velocity, hits the runner blades which extract energy, then exits downward through the draft tube. The runner converts water's pressure and velocity into shaft rotation.
Penstock — The High-Pressure Pipe
Water reaches the turbine through a penstock — a massive steel pipe running through the dam body.
Hoover's penstocks:
├── Diameter: 9.1 m (30 feet) — big enough to drive two trucks through side by side
├── Steel thickness: ~70 mm at the bottom (where pressure is highest)
├── Water velocity inside: ~3 m/s
├── Internal pressure: ~1.8 MPa (18 atmospheres)
└── Total length: ~100 m from intake to turbine
The penstock must resist hoop stress:
σ_hoop = P × r / t
At 1.8 MPa, r = 4.55 m, t = 0.07 m:
σ_hoop = 1.8 × 10⁶ × 4.55 / 0.07 = 117 MPa
Steel yield strength: ~250 MPa. Factor of safety: 2.14. Adequate.
DESIGN SPEC UPDATED:
├── Hydroelectric power: P = ρgQHη
├── Hoover: 17 Francis turbines, 2,080 MW total, 4.2 TWh/year
├── Per turbine: Q=56 m³/s, H=180 m, η=0.92 → 91 MW
├── Penstock: 9.1 m diameter, 1.8 MPa internal pressure
└── Francis turbine efficiency: 93-95% at design point
───
PHASE 6: Survive the Flood
Your dam holds back a river. But rivers flood. The Colorado River's maximum recorded flood: 8,500 m³/s — 150 times its average flow. If that water can't get past the dam safely, it goes OVER the dam. Water flowing over an unprotected concrete dam erodes the foundation. The dam undercuts. The dam fails.
The Probable Maximum Flood (PMF)
Dam engineers don't design for the worst flood on record. They design for the Probable Maximum Flood — the theoretically largest flood that could occur based on maximum precipitation and worst-case runoff conditions.
For the Colorado River basin at Hoover:
├── PMF inflow: ~11,300 m³/s
├── 100-year flood: ~5,700 m³/s
├── Average annual peak: ~2,300 m³/s
├── Average flow: ~570 m³/s
└── Spillway capacity must handle: PMF minus what turbines + outlets can pass
The spillway is the dam's safety valve. It must pass flood water safely without letting the reservoir overtop the dam.
Spillway Design — Controlled Overflow
Hoover Dam has two spillway tunnels, one on each side of the canyon, each 15 m in diameter — carved through solid rock. Water enters through drum gates on the reservoir side and drops into the tunnels.
reservoir downstream
level ───┐ river
│ drum gate │
│ ┌───┐ │
▼ │ │ │
░░░░░░░░░│ │ │
░░░░░░░░░└───┘ │
░░░░░░░░░░░░░│ │
│ ╲ │
│ ╲ spillway tunnel │
│ ╲ (15 m diameter) │
│ ╲ │
│ ╲ │
│ ╲ │
│ ╲ ←─ 180 m vertical │
│ ╲ drop │
│ ╲ │
│ ╲──────────────────────┘
stilling basin
(energy dissipation)
Each tunnel capacity: ~5,600 m³/s
Combined: ~11,200 m³/s
Water velocity at base: ~50 m/s (180 km/h)At 50 m/s, water is destructive enough to erode solid concrete. The stilling basin at the bottom uses hydraulic jumps to dissipate the kinetic energy before the water rejoins the river.
Energy Dissipation — The Stilling Basin
Water falling 180 m reaches tremendous velocity. The kinetic energy per unit volume:
KE = ½ρv²
At v = 50 m/s:
KE = ½ × 1,000 × (50)²
KE = 1,250,000 J/m³ = 1.25 MPa
That's enough to blast concrete apart. The stilling basin forces the high-velocity flow into a hydraulic jump — an abrupt transition from fast, shallow flow (supercritical) to slow, deep flow (subcritical). The turbulence in the jump converts kinetic energy to heat.
Stilling basin dimensions at Hoover:
├── Length: ~50 m
├── Depth below tailwater: ~10 m
├── Concrete thickness: 3 m (with steel armor on the floor)
├── Energy dissipated: ~90% of incoming kinetic energy
└── Concrete erosion rate: 2-5 mm/year under flood conditions
DESIGN SPEC UPDATED:
├── PMF design: 11,300 m³/s (20× average flow)
├── Spillway: two 15 m tunnels, combined capacity 11,200 m³/s
├── Water velocity at base: 50 m/s (180 km/h)
├── Stilling basin: hydraulic jump dissipates 90% of kinetic energy
└── Overtopping = foundation erosion = dam failure
───
PHASE 7: Watch for Failure
A dam doesn't fail suddenly. It warns you — if you're listening. A crack widens 0.1 mm per year. Seepage increases by 2 liters per minute per decade. The crest shifts 3 mm downstream. These are whispers. Ignore them and they become screams.
Pendulum Monitoring — 0.1 mm Precision
Hoover Dam has plumb lines — long pendulums hanging in vertical shafts inside the dam, anchored at the top, with the weight hanging freely at the bottom in a pool of oil (for damping).
The pendulum bob doesn't move. The dam does. By measuring the position of the shaft wall relative to the stationary bob, engineers detect:
├── Crest displacement: seasonal movement of ±10-15 mm (thermal expansion/contraction)
├── Permanent drift: long-term downstream creep
├── Asymmetric movement: one side moving differently from the other
└── Sudden jump: possible foundation failure or earthquake damage
Measurement precision: 0.1 mm
Hoover's crest moves in a yearly cycle:
├── Summer (hot, low reservoir): crest moves downstream ~15 mm (thermal expansion)
├── Winter (cold, high reservoir): crest moves upstream ~10 mm (water load + contraction)
└── This cycle has been consistent since 1936
The Instrumentation Suite
Instrument Measures Precision Quantity
(Hoover)
─────────────────────────────────────────────────────────────────────
Pendulums Horizontal displacement 0.1 mm 12
Extensometers Foundation deformation 0.01 mm 40+
Piezometers Pore water pressure 0.1 kPa 100+
Seepage weirs Drainage flow rate 0.1 L/min 30+
Thermocouples Internal temperature 0.1°C 400+
Survey targets Surface movement (GPS) 1 mm 50+
Seismographs Earthquake response 0.001g 6
Joint meters Crack opening 0.01 mm 80+A large dam has 500-1,000 sensors. Data is logged continuously and analyzed for trends. The dam is one of the most heavily instrumented structures on Earth — because the consequences of missing a warning sign are measured in thousands of deaths.
Banqiao Dam — 1975: What Happens When You Don't Watch
Typhoon Nina hit Henan Province, China. In 24 hours: 1,060 mm of rain — a year's worth. The inflow to Banqiao reservoir exceeded the spillway capacity. The dam overtopped. Water eroded the downstream face. The dam breached.
The flood wave:
├── Height: 10 meters
├── Width: 10 km
├── Speed: 50 km/h
├── 62 dams failed in cascade downstream
├── Deaths: 171,000 (26,000 from flooding, 145,000 from subsequent famine and disease)
└── Displaced: 11 million people
The root causes:
├── Spillway capacity designed for only a 1,000-year flood, not PMF
├── Sluice gates were blocked by sediment — couldn't open
├── No early warning system downstream
└── Communication failures — upstream dam failures not reported
This single event killed more people than any other structural failure in human history.
DESIGN SPEC UPDATED:
├── Pendulum monitoring: 0.1 mm precision, detects seasonal cycle + long-term drift
├── 500-1,000 sensors per major dam: piezometers, extensometers, thermocouples
├── Seepage monitoring: increasing flow rate = internal erosion warning
├── Banqiao 1975: 171,000 deaths — underdesigned spillway + blocked sluice gates
└── Dam safety = continuous monitoring + adequate spillway + redundant outlets
───
PHASE 8: Let the Fish Through
Before Hoover Dam, salmon migrated 1,400 km up the Colorado River to spawn. The dam is a 221-meter concrete wall across their only path. No fish can jump 221 meters. No fish can swim through a turbine and survive. Every dam on every salmon river faces the same problem: you've blocked a migration that's been running for 10 million years.
Fish Ladders — A Staircase for Salmon
A fish ladder is a series of small pools, each slightly higher than the last, connected by slots or weirs that fish can leap through. Each step is typically 0.3 m high — within the jumping ability of most salmon species.
upstream downstream
(high) (low)
┌───┐ river
│ │ ┌───┐ │
│ │ │ │ ┌───┐ │
│ │ │ │ │ │ ┌───┐ │
│ │ │ │ │ │ │ │ ┌───┐ │
│ │ │ │ │ │ │ │ │ │ ┌───┐ │
│ │ │ │ │ │ │ │ │ │ │ │ ┌──────┘
│ └──┘ └──┘ └──┘ └──┘ └──┘ └──┘
Each step: 0.3 m rise
Pool size: ~3 m × 4 m × 1.5 m deep
Flow velocity at slot: ~2.4 m/s (salmon can burst to 6 m/s)
Steps needed for 30 m dam: 100 pools
Steps needed for 221 m dam: 737 pools ← impracticalFish ladders work for dams up to ~50 m. Beyond that, the ladder becomes so long that fish exhaust themselves climbing it, arriving at spawning grounds too spent to reproduce. For high dams like Hoover, fish passage is essentially impossible — the migration is severed permanently.
Dissolved Gas Supersaturation — Death by Bubbles
When water plunges over a spillway and crashes into the stilling basin, it entrains air. The deep, turbulent water dissolves that air at high pressure. When fish swim through this supersaturated water, dissolved nitrogen comes out of solution in their blood — exactly like the bends in scuba divers.
Total dissolved gas (TDG) levels:
├── Normal river water: 100% saturation
├── Below spillway during flood: 120-140% saturation
├── Fish mortality threshold: >110% (sustained exposure)
├── Lethal within hours: >130%
└── Maximum observed at Grand Coulee: 145%
Solutions:
├── Flip lip spillways: deflect water horizontally to reduce plunge depth
├── Surface spill: release water from higher elevation to reduce entrainment
└── Flow deflectors: break up the plunge jet before it reaches the basin
The Downstream Ecosystem
Dams don't just block fish. They change the river's fundamental character:
├── Temperature: water released from deep in the reservoir is 4-8°C year-round
│ Natural river: 0-25°C seasonal range. Cold releases sterilize warm-water species.
├── Flow regime: natural rivers flood seasonally. Dams flatten the hydrograph.
│ Species adapted to floods (cottonwoods, native fish) decline.
├── Sediment starvation: all sediment trapped behind dam (see Phase 9)
│ River below erodes its own bed — channel incision of 1-3 m per decade.
└── Nutrient trapping: phosphorus and nitrogen settle in reservoir
Downstream productivity drops — fisheries collapse.
The Colorado River below Hoover Dam lost 4 of 8 native fish species within 30 years of closure.
DESIGN SPEC UPDATED:
├── Fish ladders: work up to ~50 m, impractical for high dams (>100 m)
├── Dissolved gas supersaturation: >110% TDG kills fish, spillways cause 120-140%
├── Thermal pollution: deep releases are 4-8°C year-round, killing warm-water species
├── Sediment starvation: all sediment trapped, downstream channel erodes
└── Hoover Dam eliminated 4 of 8 native Colorado River fish species
───
PHASE 9: Handle the Sediment
Every river carries sediment — sand, silt, clay, gravel — eroded from mountains and plains upstream. The Colorado River carried 140 million tonnes per year before Hoover Dam was built. The moment the river hits the still water of Lake Mead, it slows down. Sediment drops out. The reservoir is filling up. Slowly, inevitably, your dam is burying itself.
The Sedimentation Rate
Sediment deposition follows Stokes' Law — particles settle based on size and water velocity:
v_s = (ρ_s - ρ_w) × g × d² / (18μ)
Where:
├── v_s = settling velocity
├── ρ_s = sediment density (~2,650 kg/m³ for quartz sand)
├── d = particle diameter
└── μ = dynamic viscosity of water
Particle Diameter Settling velocity Time to settle
through 100 m
──────────────────────────────────────────────────────────────────
Gravel 10 mm 1.0 m/s 1.7 minutes
Coarse sand 1 mm 0.1 m/s 17 minutes
Fine sand 0.1 mm 0.008 m/s 3.5 hours
Silt 0.01 mm 0.0001 m/s 11.6 days
Clay 0.001 mm 0.000001 m/s 3.2 yearsGravel and sand settle immediately where the river enters the reservoir, forming a delta. Silt settles throughout the reservoir body. Clay particles are so fine they may never settle — they pass through the dam in turbid flows.
Lake Mead's Countdown
Lake Mead's original storage capacity (1935): 36.7 km³
Sediment accumulated by 2024: ~2.2 km³ (~6% of original capacity)
Current sedimentation rate: ~0.025 km³/year (reduced from natural — upstream dams trap sediment first)
At the current rate:
├── 50% capacity lost in: ~700 years
├── Intake structures buried in: ~200-300 years
├── Dead storage reached in: ~150 years (water can't reach outlets)
└── Original pre-dam estimate: ~300 years to complete filling
But there's a twist — Glen Canyon Dam (built 1963, 300 km upstream) now traps 95% of the Colorado's sediment before it reaches Lake Mead. Lake Mead's sedimentation has slowed dramatically. Glen Canyon's reservoir (Lake Powell) is filling faster instead — it will be the first to die.
Fighting Sedimentation
Options for extending reservoir life:
├── Sediment flushing: open low-level outlets during floods, scour sediment downstream
│ Works for: small reservoirs with large floods. Requires sediment outlets near the bed.
│ Hoover: impossible — outlets are too high above the sediment delta.
│
├── Sediment sluicing: pass turbid flood water through before it settles
│ Effective if: reservoir has short residence time during floods.
│ Hoover: marginally effective — reservoir too large, flow too slow.
│
├── Dredging: physically excavate sediment with machinery
│ Cost: $5-15/m³. For Lake Mead's annual input: $125-375 million/year.
│ Where to put 25 million m³ of wet sediment? Indefinitely? Not economic.
│
└── Sediment bypass tunnels: route sediment-laden water around the dam entirely
Cost: $100-500 million to build. Most effective long-term solution.
Only 3 exist worldwide (all in Japan and Switzerland). None on large dams.
The honest answer: most large dams will eventually lose their reservoirs to sedimentation. The only question is when.
DESIGN SPEC UPDATED:
├── Sedimentation: all reservoirs are temporary — filling at measurable rates
├── Lake Mead: 6% filled in 89 years, ~300-year lifespan without intervention
├── Glen Canyon traps 95% of sediment before it reaches Mead
├── Settling velocity: v_s ∝ d² — gravel settles instantly, clay takes years
└── No economically viable long-term solution for large reservoir sedimentation
───
PHASE 10: Decommission or Die
───
FULL MAP
Dam
├── Phase 1: Hold Back the Water
│ ├── Hydrostatic pressure: P = ρgh — linear with depth}
│ ├── Total horizontal force: F = ½ρgH²L — proportional to height SQUARED}
│ ├── Hoover: 221 m tall, force = 90.8 GN (9.26 million tonnes)}
│ ├── Pressure at base: 2.17 MPa (21.7 atmospheres)}
│ └── Resultant acts at H/3 from base — drives triangular cross-section}
│
├── Phase 2: Choose Your Shape
│ ├── Gravity dam: resists by weight, M_weight ≥ 1.5 × M_water}
│ ├── Arch dam: transfers load to canyon walls via compression, uses 80% less concrete}
│ ├── Arch stress: σ = Pr/t — thinner dam needs stronger rock abutments}
│ ├── Hoover: arch-gravity hybrid, 60% arch action, 40% gravity}
│ └── Valley shape + rock quality determine dam type — not engineering preference}
│
├── Phase 3: Don't Slide
│ ├── Sliding FoS = μ(W-U)/F_h ≥ 1.5 (normal), ≥ 1.0 (extreme)}
│ ├── Uplift: water pressure under dam reduces effective weight 30-50%}
│ ├── Drain gallery at 1/3 base width from upstream face — reduces uplift 40-60%}
│ ├── Concrete-rock friction μ ≈ 0.65-0.75}
│ └── Malpasset 1959: uplift in fractured rock → sliding failure → 423 dead}
│
├── Phase 4: Don't Crack
│ ├── Hydration heat: ~300 kJ/kg, interior temp rise 50-60°C}
│ ├── Thermal stress: σ = EαΔT — exceeds tensile strength at ΔT > 20°C}
│ ├── Cooling pipes: 937 km of pipe, 4°C water, reduced cooling from 125 years to ~2 years}
│ ├── Block pour: 230 columns × 1.5 m lifts, contraction joints grouted after cooling}
│ └── Concrete tensile strength only 3-5 MPa — thermal cracking is the primary risk}
│
├── Phase 5: Spin the Water
│ ├── Hydroelectric power: P = ρgQHη}
│ ├── Hoover: 17 Francis turbines, 2,080 MW total, 4.2 TWh/year}
│ ├── Per turbine: Q=56 m³/s, H=180 m, η=0.92 → 91 MW}
│ ├── Penstock: 9.1 m diameter, 1.8 MPa internal pressure}
│ └── Francis turbine efficiency: 93-95% at design point}
│
├── Phase 6: Survive the Flood
│ ├── PMF design: 11,300 m³/s (20× average flow)}
│ ├── Spillway: two 15 m tunnels, combined capacity 11,200 m³/s}
│ ├── Water velocity at base: 50 m/s (180 km/h)}
│ ├── Stilling basin: hydraulic jump dissipates 90% of kinetic energy}
│ └── Overtopping = foundation erosion = dam failure}
│
├── Phase 7: Watch for Failure
│ ├── Pendulum monitoring: 0.1 mm precision, detects seasonal cycle + long-term drift}
│ ├── 500-1,000 sensors per major dam: piezometers, extensometers, thermocouples}
│ ├── Seepage monitoring: increasing flow rate = internal erosion warning}
│ ├── Banqiao 1975: 171,000 deaths — underdesigned spillway + blocked sluice gates}
│ └── Dam safety = continuous monitoring + adequate spillway + redundant outlets}
│
├── Phase 8: Let the Fish Through
│ ├── Fish ladders: work up to ~50 m, impractical for high dams (>100 m)}
│ ├── Dissolved gas supersaturation: >110% TDG kills fish, spillways cause 120-140%}
│ ├── Thermal pollution: deep releases are 4-8°C year-round, killing warm-water species}
│ ├── Sediment starvation: all sediment trapped, downstream channel erodes}
│ └── Hoover Dam eliminated 4 of 8 native Colorado River fish species}
│
├── Phase 9: Handle the Sediment
│ ├── Sedimentation: all reservoirs are temporary — filling at measurable rates}
│ ├── Lake Mead: 6% filled in 89 years, ~300-year lifespan without intervention}
│ ├── Glen Canyon traps 95% of sediment before it reaches Mead}
│ ├── Settling velocity: v_s ∝ d² — gravel settles instantly, clay takes years}
│ └── No economically viable long-term solution for large reservoir sedimentation}
│
└── Phase 10: Decommission or Die
───