Ground Effect

The two mechanisms

Ground effect is not one phenomenon but two, acting together. They are easiest to keep straight if you name them separately.

The cushion (ram) effect. As a wing moves forward close to the ground, the air beneath it is squeezed into a shrinking gap. The flow cannot escape sideways fast enough, so static pressure under the wing rises toward the stagnation value. This pressure rise is bounded by the dynamic head of the flow:

Δp_cushion  ≲  ½ ρ V²        (dynamic head sets the ceiling)

where:
  ρ = air density       (~1.225 kg/m³ at sea level)
  V = forward speed      (m/s)
  Δp = pressure rise under the wing (Pa)

The cushion dominates at very low heights and high angles of attack — the high-lift landing flare, a hovercraft, or an ekranoplan trimmed nose-up so the underside acts almost like a ram-air scoop.

The induced-drag (span) effect. A finite wing trails a vortex from each tip because high-pressure air below leaks around to the low-pressure top. These vortices induce a downward velocity — downwash — over the wing, which tilts the local lift vector backward and produces induced drag. Near the ground, the tip vortices are forced apart and shortened (as if reflected by an image wing below the surface), so the downwash angle ε falls. The induced-drag coefficient scales with the square of the lift coefficient and inversely with aspect ratio:

C_Di = C_L² / (π · AR · e)

with a ground-effect efficiency multiplier β(h/b) applied:
  C_Di,ground = β(h/b) · C_L² / (π · AR · e)

where:
  C_L = lift coefficient
  AR  = aspect ratio (b²/S)
  e   = span efficiency factor (~0.7–0.95)
  b   = span,  h = height of the quarter-chord above ground
  β   ≈ 1 in free air, falling toward ~0.5 near h/b = 0.1

A widely used approximation (after Wieselsberger) for the reduction factor is β ≈ (16 h/b)² / (1 + (16 h/b)²). At h/b = 1 it is already near unity; at h/b = 0.1 it drops to about 0.72; at h/b = 0.05 to about 0.39. Combine that with the cushion lift and the wing's effective lift-to-drag ratio can climb 30–40 percent.

Cushion versus vortex: where each wins

	Cushion (ram) effect	Induced-drag (span) effect
Physical cause	Air compressed into shrinking gap → higher static pressure	Tip vortices shortened by ground → less downwash
Scaling parameter	Height-to-chord, h/c	Height-to-span, h/b
Primary benefit	Extra lift	Less drag (higher L/D)
Dominant when	Very low, high angle of attack	Moderate heights, cruise attitude
Best exploited by	Hovercraft, low-aspect WIG, landing flare	High-aspect wings, gliders near ground
Failure mode	Cushion collapses if gap opens or air spills out	Effect vanishes as h/b → 1

Worked example: induced drag near the ground

Take a light aircraft on short final: span b = 11 m, flying at h = 1.1 m above the runway (quarter-chord height), generating a lift coefficient C_L = 1.4 in landing configuration, aspect ratio AR = 7, span efficiency e = 0.8. Compare induced drag in free air to ground effect.

h/b = 1.1 / 11 = 0.10

Free-air induced-drag coefficient:
  C_Di = C_L² / (π · AR · e)
       = 1.4² / (3.1416 · 7 · 0.8)
       = 1.96 / 17.59
       = 0.1114

Ground reduction factor (Wieselsberger):
  16 h/b = 1.6
  β = 1.6² / (1 + 1.6²) = 2.56 / 3.56 = 0.719

Induced drag in ground effect:
  C_Di,ground = 0.719 · 0.1114 = 0.0801

Reduction: about 28% less induced drag at h/b = 0.10.

That 28 percent cut in induced drag is felt directly by the pilot as float: the aircraft is suddenly more efficient, sinks reluctantly, and wants to keep flying. Carry an extra 10 knots into the flare and that float can stretch the touchdown point hundreds of metres down the runway.

Worked example: a venturi floor (ground effect inverted)

A Formula 1 car uses the ground as the throat of a venturi. Air entering under the splitter is accelerated through a shaped channel; by Bernoulli's principle the pressure drops, sucking the car down. Suppose freestream speed under the floor is V₁ = 60 m/s and the venturi throat accelerates it to V₂ = 90 m/s:

Bernoulli (incompressible) between freestream and throat:
  p₁ + ½ρV₁² = p₂ + ½ρV₂²

Pressure drop in the throat:
  Δp = ½ρ(V₂² − V₁²)
     = ½ · 1.225 · (90² − 60²)
     = 0.6125 · (8100 − 3600)
     = 0.6125 · 4500
     = 2756 Pa  (≈ 2.8 kPa suction)

Over a floor planform of 2.5 m²:
  Downforce ≈ Δp · A = 2756 · 2.5 ≈ 6.9 kN  (~700 kgf)

Lowering the ride height tightens the throat, raises V₂, and deepens the suction — so the car generates more downforce the closer it runs to the track. This is why ground-effect cars are run as low and stiff as the rules and the bumps allow, and why the same sensitivity is dangerous (see failure modes).

Real machines that live in ground effect

• Landing airliners. A 737-sized jet (span ~35 m) enters ground effect in the last ~35 m and floats noticeably in the final few metres. Pilots flare and bleed speed to overcome the reluctance to settle.
• Ekranoplans (WIG craft). The Soviet Lun-class, ~380 tonnes, cruised ~4 m above the Caspian at ~500 km/h with an L/D far above a conventional aircraft of equal size. It rode permanently in the cushion.
• Formula 1 and Le Mans prototypes. Venturi underfloors supply roughly half of total downforce on a modern F1 car — thousands of newtons that let the car corner at over 5 g.
• Racing pelicans and skimmers. Birds glide centimetres above water for tens of kilometres on the cushion, spending almost no muscular energy.
• Hovercraft. The pure cushion limit — a fan maintains an air gap so there is no contact and almost no aspect-ratio benefit, only ram pressure.

How the effect scales with height

Height ratio h/b	Induced-drag factor β	Regime	Practical meaning
1.0	≈ 0.996	Free air (edge)	Effect negligible
0.5	≈ 0.985	Onset	Just measurable
0.25	≈ 0.94	Moderate	L/D noticeably better
0.10	≈ 0.72	Strong	Float on landing; WIG cruise
0.05	≈ 0.39	Very strong	Ekranoplan operating band
0.02	≈ 0.09	Cushion-dominated	Hovercraft / extreme WIG

Note that β here is the multiplier on induced drag (lower is better); the cushion lift bonus rises in parallel as the gap closes. The two effects together explain why purpose-built WIG craft chase the lowest stable height they can hold.

Failure modes and trade-offs

• Porpoising (F1). A venturi floor sucks the car down; as ride height falls the suction grows, pulling it lower still — until the floor stalls or the gap chokes, downforce collapses, the car springs back up, and the cycle repeats at several hertz. This bouncing battered drivers in the 2022 season until floor and ride-height rules were changed.
• Pitch instability in WIG craft. The centre of pressure moves with height and pitch; a nose-up disturbance can increase lift, raising the craft out of the cushion, then drop it back — a divergent oscillation if the tail and pitch trim are not carefully tuned. Most ekranoplans use a large tailplane mounted high, out of the ground-effect zone, to stabilise pitch.
• Float and overrun on landing. The drag reduction that helps takeoff makes touchdown harder; excess approach speed turns into a long, drag-starved float and a runway overrun.
• Surface roughness. WIG craft need a smooth surface; waves higher than a fraction of the cushion gap break the seal and spike the structural loads, which is why ekranoplans were limited to calm seas.
• Loss of cushion in a turn. Banking a wing toward the ground reduces clearance on one side and can spill the cushion asymmetrically, producing a sudden rolling moment — a known hazard for low-flying agricultural aircraft.
• Compressibility ceiling. The cushion pressure cannot exceed the dynamic head ½ρV²; you cannot get arbitrary lift by flying lower, and at high subsonic speeds shock formation under the wing changes the picture entirely.