Sandbar wall

Two cages, ~10 m apart. The front grows to a 40% open screen, the back grows solid. Water jets through the front, sloshes in the gap out of phase with the sea, and nothing gets through. The gap is tuned to the wave so little reflects either (Jarlan chamber, built since 1962). Show up, screw in anchors, hang the cages, grow for two years.

peer-reviewed published   field measured on a real install   derived this site   input your choice

Sandbar calculator

Structure$ per mileBasis

Rock mound for the same bar, from Rock (Hudson, van der Meer, d’Angremond): 4 m high, 39 m²/m, 116,000 t/mile, Kt 0.26.

The setup

  1. Two steel egg-crate cages, 4 m tall, ~10 m apart (0.1–0.3 wavelengths). Helical anchors every 3–4 m.
  2. Titanium anodes hung in the gap. field EUDP used DSA.
  3. Rectifiers in constant-current mode, 1.5 A per m² of steel. Volts float, 2–4 V. peer-reviewed Sharker 2023 compact limit · Goreau 2012 “1–2 cm/yr, no faster”.
  4. Two years at ~2 cm/yr of limestone per side. Then drop to protection current, or off if the limestone holds alone.
  5. The gap does the work: reflection ≤0.3 only if the gap is 0.1–0.3 wavelengths. A few feet does nothing. peer-reviewed Medina 2010 · Suh 2001.

Same sea, built projects

Our wall run at the USACE Mordecai Island design conditions: Barnegat Bay NJ, 1.5 m at MHW, 0.9 m wave, T 2.5 s, crest 0.5 m above MHW. Same job, less material, no track record.

StructureWaves$ per mileMaterialRobustness

Robustness is what the table cannot show. Rock and sheet pile have been in these bays for 60 years; our wall has never been built. The risks that could make ours worse:

Risk in a NJ bayRock mound / bulkheadGrown wall
Winter ice pushing on itKnown. Stevens budgets 10% of build cost to re-set rock moved by iceUntested. 5 mm steel cage with a brittle limestone skin
Soft bay mud under itKnown. USACE lays a geocomposite and builds 1 ft high to allow settlementUntested. Held by helical anchors screwed into loose sediment; anchor capacity there is unmeasured
Scour at the toeMild. A 2:1 slope spreads the waveWorse. A vertical face scours its toe, same as a bulkhead
Boat wakesDesigned for them; USACE used wakes as the design waveModel assumes regular waves; wakes are short, steep groups. Not checked
Storm bigger than designRock loses stones gradually; start-of-failure at 1.3× design wave (van der Meer S 2 → 8). Bulkhead steel replaced at 40–60 yrWater is 1.5 m, so the biggest wave is ~1.2 m and the steel never nears yield. Whether skin or anchor gives first is unknown
Track record60+ yr, thousands of milesNone at wall scale. Biorock exists only as reef-scale mesh (History)

One test panel in Barnegat Bay through one winter answers the first four rows.

Jarlan wall — the prior art

Jarlan’s 1964 patent drawing: perforated seaward wall, chamber, and solid back wall, compared with a rubble mound of the same footprint.
US 3,118,282 (Jarlan, granted 21 Jan 1964), sheet 1 — public domain. Fig. 1 is the wall. Fig. 2 is the same footprint against a rubble mound.
sea perforated front chamber B ≈ 0.1–0.3 L (flume) · patent 0.05–0.15 L solid back sandbar / lee

What the papers say

  1. primary Jarlan 1961, patent US 3,118,282. Perforated front ~40–60% open, chamber 0.05–0.15 of an 8 s deep-water wavelength.
  2. field First build Baie-Comeau, Québec, 1962: nine concrete caissons, 3 ft holes, empty chamber. NRC strain-gauged one panel in service (Dryburgh 1967).
  3. peer-reviewed Fugazza & Natale 1992: closed-form design; one chamber is the most effective in the practical range. Suh 2001: reflection minimum near B/L ≈ 0.2, falls as steepness rises.
  4. peer-reviewed Medina / Garrido 2010–2012, 1,500 UPV flume runs: CR < 0.6 for 0.1 < B/L < 0.3; lowest at 20–30% open; slots and round holes alike at the same open %.
  5. contested Patent (50% open, B/L 0.05–0.15) vs flume (20–30% open, B/L 0.1–0.3). Both stay on the page. Linear theory goes below CR 0.3; this calculator floors Kr at 0.3. No electrolytic Jarlan has been measured.

How it is computed

derived Chamber: screen at x = 0, solid wall at x = B, propagating mode only. Screen condition ik(φ₁ − φ₂) = (F − iS) ∂φ/∂x with F = 2T|1 − R| (Kriebel’s linearised loss, T from the Physics page) and S = kD/P (fluid inertia in the holes). R = (Z − 1)/(Z + 1), Z = F + i(cot kB − S). Resonance at cot kB = S; full absorption at F = 1.

derived Loads: screen q = ρg(H/2)|F − iS||1 − R|; back wall q = ρg(H/2)|1 − R|/|sin kB|; plain wall ρgH. Search: front open 15–40%, B/L 0.06–0.30, for minimum R; then each cage is cost-optimised exactly as the single wall on Home, back cage with P = 0.

peer-reviewed Rock mound at the same sea: d’Angremond 1997 Kt = −0.4 Rc/H + 0.64 (B/H)−0.31(1 − e−0.5ξ); Zanuttigh & van der Meer 2008 Kr = tanh(0.12 ξ0.87). Mass: 2.3 m trapezoid, 3 ft crest, 2H:1V, 1.85 t/m³.

Not priced: marine installation, anodes, rectifiers and cabling, toe support, scour, design and permits. Regular waves, no breaking on the wall, no overtopping.

Sources