Rip Current: The Narrow River That Runs Out to Sea

Step one: breaking waves build a hill of water

A surface gravity wave carries momentum as well as energy, and the excess flux of momentum due to the waves' presence is the radiation stress, formalised by Michael Longuet-Higgins and Robert Stewart in the Journal of Fluid Mechanics (1962) and Deep-Sea Research (1964). For shore-normal waves the relevant component is the cross-shore flux of cross-shore momentum, S_xx = E(2n − ½), where E = ⅛ρgH² is the wave energy density and n = c_g/c is the ratio of group to phase speed. In deep water n = ½ and S_xx = ½E; in the shallow surf zone n → 1 and S_xx → 3E/2.

Outside the breakers, shoaling makes H grow, so S_xx grows, and the mean surface is pulled slightly down — the set-down, of order −H_b²/16h_b, a few centimetres at most. Then the wave breaks: H collapses over a few tens of metres, S_xx collapses with it, and the surrendered momentum reappears as a pressure gradient. The depth-integrated cross-shore balance in the surf zone reduces to dS_xx/dx + ρg(h + η̄) dη̄/dx = 0.

Inside the surf zone breaking is depth-limited, H ≈ γ(h + η̄) with γ ≈ 0.7–0.8, and that balance integrates to a strikingly simple rule: the mean water surface tilts up toward the beach at a fixed fraction of the bed slope, dη̄/dx = −K dh/dx with K = [1 + 8/(3γ²)]⁻¹ ≈ 0.19 for γ = 0.8. Integrate across the surf zone and subtract the set-down and you get a shoreline set-up of roughly 0.15–0.2 times the breaker height. A 1 m breaker props an extra 15–20 cm of water against the sand. Bowen, Inman and Simmons measured exactly this in a wave flume and published it in JGR in 1968.

Step two: give the hill a gap to drain through

Set-up on a perfectly uniform beach is uniform along the shore, and uniform pressure drives nothing. Rips exist because real surf zones are not uniform: a channel cut through a sandbar, a headland, a groyne, a pier piling, a submarine canyon refracting wave energy away from its axis — any of these makes breaker height vary along the beach. Where the water is deeper, waves break later, more weakly, or not at all, so less momentum is surrendered and less set-up is built.

The result is a low spot in the hill. If one stretch of bar has H_b = 1.0 m (set-up ≈ 19 cm) and the channel 100 m away has H_b = 0.5 m (set-up ≈ 10 cm), the sea surface is about 9 cm higher over the bar than over the channel. That head drives feeder currents: shallow flows running sideways along the beach, inside the bar, converging on the low spot from both directions. Where they meet they turn and accelerate offshore through the gap as the neck — the narrow, deep, dark, choppy band that swimmers see as a suspicious gap in the line of breakers.

Anthony Bowen closed the theoretical loop in JGR in 1969 (“Rip currents: 1. Theoretical investigations”), showing that an alongshore variation in breaker height forces exactly this cellular circulation; the companion paper by Bowen and Inman reproduced the cells in a laboratory basin. Past the breaker line the confining pressure gradient vanishes and the jet spreads and decelerates like a free turbulent plane jet into a mushroom-shaped rip head — typically a counter-rotating vortex pair, often stained brown by sand torn off the channel floor.

The numbers: how fast, how wide, how far

The crudest estimate treats the head as a Bernoulli conversion: u ≈ √(2gΔη). With Δη = 10 cm this gives 1.4 m/s, and with 5 cm about 1.0 m/s. That is an upper bound, because it ignores friction. A more honest steady balance sets the alongshore-fed pressure gradient against bottom stress, ρgh ∂η̄/∂y = c_fρu², giving u = √(gh(∂η̄/∂y)/c_f). Put in h = 1.5 m, a 5 cm drop over 200 m, and a drag coefficient c_f ≈ 0.003, and you get u ≈ 1.1 m/s.

That is right where the measurements sit. Francis Shepard and Douglas Inman, working off Scripps Beach in La Jolla and publishing in the Transactions of the American Geophysical Union in 1950, reported mean rip speeds of roughly 0.2–0.7 m/s with peaks above 1 m/s and short pulses approaching 2 m/s. Modern drifter and current-meter work has not overturned those figures. For scale, an ordinary swimmer sustains perhaps 0.5–0.8 m/s, so a 1 m/s rip beats almost anyone who fights it head-on — the whole drowning mechanism in one comparison.

Geometrically, necks are typically 10–30 m wide, rips recur every 100–500 m along a beach, and the spacing scales with the surf-zone width — Huntley and Short (1992) found ratios of a few times the surf-zone width in a compilation of observed rip spacings. The flow is far from steady: MacMahan and colleagues documented very-low-frequency pulsations in the Journal of Geophysical Research in 2004, with rip speed surging on periods of minutes to tens of minutes as wave groups deliver momentum in bursts and the jet itself meanders. Tides matter too, not as a driver but as a modulator: rips are usually strongest near low tide, when the bar is shallowest, breaking over it is most complete, and the alongshore contrast in set-up is largest.

Four kinds of rip, and the beaches that make them

• Channel (bathymetric) rips. The textbook case: a persistent gap in a shore-parallel bar. They can hold station for days or weeks, because the flow maintains the channel that maintains the flow — a feedback loop between water and sand.
• Boundary rips. Pinned against a headland, groyne, jetty or pier, where the structure blocks the longshore current and shelters the water beside it from breaking. The most predictable rips on any coast — and the ones people deliberately swim next to, because that water looks calm.
• Transient or flash rips. No fixed bathymetry required: vorticity injected by short-crested, finite-crest-length breaking, together with shear instabilities of the longshore current — the shear waves identified by Oltman-Shay, Howd and Birkemeier in JGR in 1989 from the SUPERDUCK experiment at Duck, North Carolina — rolls up into transient vortices that eject water seaward at essentially random times and places. They are why a featureless beach is not automatically safe.
• Mega-rips. Under storm waves on embayed beaches, Andrew Short (1985) described rips reaching about 2 m/s and extending hundreds of metres beyond the surf zone, anchored on the headlands.

Which beaches make them is well predicted by the Wright and Short (1984) beach-state model in Marine Geology: the intermediate states — transverse bar and rip (TBR), rhythmic bar and beach (RBB) — are the rip factories, while fully dissipative beaches (wide, gentle, energy shed gradually far out) and fully reflective ones (steep, coarse, barely any surf zone) have few classic rips, for opposite reasons.

How you actually measure a river you cannot see

Shepard's tools at Scripps in the late 1930s and 1940s were fluorescein dye, drift bottles, floats timed against the pier and aerial photographs — enough to establish both the phenomenon and its name (Science, 1936, in a paper titled “Undertow, rip tide, or ‘rip current’?”).

Three modern techniques do the heavy lifting. Argus video stations, the shore-based camera systems described by Rob Holman and John Stanley in Coastal Engineering (2007), take 10-minute time exposures; averaging turns the flickering breakers into a smooth white band that maps the bar, and rip channels show up as dark gaps in it. Instrument frames carrying acoustic Doppler current profilers give Eulerian velocity records at fixed points, which is how the pulsations were quantified. GPS drifters give the Lagrangian picture — where the water actually goes.

The landmark drifter study is RCEX, the Rip Current Experiment run by Jamie MacMahan and colleagues at Sand City, Monterey Bay, California, in April–May 2007, using on the order of 30 GPS-tracked surf drifters with metre-scale (order 1 m) positioning. The result overturned the popular mental model. Most rips did not behave as one-way escalators to the open sea; they behaved as semi-closed eddies, recirculating water back into the surf zone, with under roughly 20% of drifters exiting the surf zone per hour. This is why safety advice has shifted toward “float and signal” alongside “swim parallel”: the current will often bring a floating person back on its own. Forecasting followed — from the empirical Lushine index of the 1990s, through the logistic-regression probabilistic model of Dusek and Seim (2013) calibrated on lifeguard rescue records at Kill Devil Hills, North Carolina, to the rip-current probability product the US National Weather Service runs on its Nearshore Wave Prediction System.

It is not a riptide, and it is not an undertow

Three misconceptions kill people, and all three are about physics.

“Riptide.” Tides do not drive rips; waves do. Tides only modulate them by changing the depth over the bar, which is why rips typically peak near low water. Genuine tidal currents at inlets are a separate and also lethal hazard, but they run on the tidal clock, not the swell.

“Undertow.” Undertow is real, but it is a different animal: the depth-averaged seaward return of the mass that Stokes drift and breaking carry shoreward above wave-trough level. It is broad, alongshore-uniform, concentrated near the bed, and typically 0.1–0.3 m/s. It is a sheet, not a jet, and it has no confined offshore neck.

“It pulls you under.” Nothing in the momentum balance has a downward component of any consequence. A rip is a horizontal flow with a free surface; your buoyancy is untouched. People drown by swimming shoreward at 0.6 m/s against a 1 m/s current, losing ground until panic and exhaustion take over. The United States Lifesaving Association attributes more than 80% of surf-beach rescues to rip currents, and NOAA's surf-zone fatality records put US rip-related deaths at roughly 100 a year; Brighton and colleagues (Natural Hazards and Earth System Sciences, 2013) counted an average of about 21 rip-related drowning deaths a year in Australia over 2004–2011.

The named cases are instructive. Bondi Beach's persistent rip beside the southern rocks is locally nicknamed the “Backpackers' Express”, because visitors read the flat gap in the breakers as the safe place to swim. On Black Sunday, 6 February 1938, a set of large waves swept an estimated 200-plus bathers off the Bondi sandbar into the adjacent deep channel; lifesavers pulled out around 250 people and five died — still the largest mass rescue in Australian history. Panama City Beach on the Florida Panhandle ranks among the deadliest US beaches for rips, for a boringly physical reason: a long, straight bar-and-trough shoreface, frequent moderate swell, and warm water that keeps people in it.

What is still unresolved

The mean momentum balance is settled physics. Three harder questions are not.

Why that spacing? Whether rip spacing is set by a template (an inherited bar, an edge-wave pattern, a shoreline feature) or emerges spontaneously from a coupled flow–bathymetry instability is still argued. The self-organisation school — Hino (1974), later Falqués and co-workers — shows a flat bar is linearly unstable to a rip-channel mode with a preferred wavelength set by surf-zone width and sediment response; reconciling that with the scatter in field data is ongoing.

How much water actually leaves? RCEX's low exit fraction came from one gently sloping, low-energy beach. Surf-zone-to-shelf exchange sets how far pollution, larvae and swimmers travel, and the exit rate almost certainly depends on wave directional spread, longshore current strength and channel confinement. It is not a universal constant.

Can rips be forecast at the right resolution? Operational forecasts give a probability of hazardous rips over a stretch of coast on a given day. They cannot say “the rip will be 90 m north of the pier at 3 p.m. at 1.2 m/s,” because that needs nearshore bathymetry that every storm rewrites. Drone surveys, video-derived depth inversion and satellite bathymetry are closing that gap, and pairing them with phase-resolving wave models is the frontier.

Rip currents against the flows they are constantly confused with

Flow	What drives it	Shape and speed	Effect on a swimmer
Rip current	Alongshore gradient in wave setup across a bar gap	Confined seaward jet, 10-30 m wide, 0.2-2 m/s, dies out just past the breakers	Carried offshore while floating normally; danger is exhaustion and panic, not submersion
Undertow	Return of the shoreward mass flux carried above wave trough level	Broad, alongshore-uniform, near-bed sheet flow, typically 0.1-0.3 m/s	A gentle seaward tug near the feet; cannot pull a floating body under
Longshore current	Alongshore radiation stress from waves breaking at an angle	Wide river parallel to the beach inside the surf zone, 0.2-1 m/s	Drifts you steadily down the beach, not out to sea
Tidal (ebb) current	Astronomical tide draining an inlet, estuary or lagoon	Locked to the tidal clock, strongest mid-ebb, can persist for hours	Serious near inlets, but it is a tide, not the wave-driven jet people call a 'riptide'
Backwash and 'sneaker' waves	Gravity draining swash back down a steep beach face	Thin, brief, fast sheet on the beach slope itself	Knocks people off their feet at the waterline; delivers them to whatever is offshore