Constructed Wetlands for Wastewater Treatment: An Engineering Guide

What is a constructed wetland and what does it treat?

A constructed wetland is a shallow, engineered treatment basin that reproduces the physical, chemical and biological processes of a natural marsh under controlled hydraulics. Wastewater flows through a rooted matrix — open water and sediment in surface systems, or graded gravel in subsurface systems — where a consortium of processes acts in series: sedimentation and filtration capture particulates, microbial biofilms on the media and root surfaces oxidise and reduce dissolved contaminants, adsorption onto media and organic matter binds phosphorus and metals, and plant uptake assimilates a smaller fraction of nutrients.

They are used to polish municipal effluent, treat small-community and decentralised sewage, and handle industrial and agricultural streams where a low-energy, robust process is attractive. As part of a wider industrial wastewater treatment strategy, wetlands typically follow primary settlement or a package plant and act as a tertiary or secondary-polishing stage for BOD, suspended solids and nitrogen.

FWS vs subsurface flow: which wetland type do you need?

The controlling design distinction is the oxygen regime, because it dictates whether nitrification (aerobic) or denitrification (anoxic) dominates. Free water surface wetlands expose an open water column; horizontal subsurface flow keeps the water table below a gravel surface and is largely anoxic; vertical flow doses the bed intermittently so air is drawn into an unsaturated matrix, making it strongly aerobic.

Because HSSF and VF have complementary strengths, hybrid VF–HSSF trains are common where full total-nitrogen removal is required: the VF stage nitrifies, then the anoxic HSSF stage denitrifies the resulting nitrate.

How do constructed wetlands remove pollutants?

Removal is a set of parallel and sequential mechanisms rather than a single reaction:

• Physical: discrete and flocculent sedimentation of particulates, and depth filtration/interception within the gravel matrix, remove TSS and the particulate fractions of BOD, N and P.
• Microbial: attached-growth biofilm on media and roots mineralises organic carbon. Under aerobic conditions ammonium is oxidised to nitrate by autotrophic nitrifiers (Nitrosomonas, Nitrobacter); under anoxic conditions heterotrophic denitrifiers reduce nitrate to N₂ using organic carbon as electron donor.
• Sorptive/chemical: phosphorus is bound by ligand exchange and precipitation with Ca, Fe and Al in the media; metals adsorb and precipitate as (oxy)hydroxides and sulphides.
• Plant-mediated: macrophytes such as Phragmites australis (common reed) assimilate a minor fraction of N and P, but their larger role is providing biofilm surface area, releasing oxygen from roots into the rhizosphere (radial oxygen loss), and sustaining hydraulic conductivity through the root network.

The interplay of nitrification and denitrification is why total-nitrogen removal is the hardest duty: the two reactions need opposite redox conditions and rarely coexist efficiently in one bed. The same logic governs engineered tanks — see our note on the redox trade-offs behind biological treatment and the sibling article on biological nutrient removal.

How do you size a constructed wetland?

Wetlands are sized on area, not volume, because the dominant removal reactions are surface-associated (biofilm on media and roots). The standard tool is the first-order areal model of Kadlec and Knight, in which concentration declines exponentially with the ratio of an areal rate constant to the hydraulic loading rate, above a non-zero background concentration C*:

Since q = Q/A, the equation rearranges directly to give the required surface area:

A more general form, the P-k-C* model, replaces the single plug-flow exponential with a tanks-in-series expression, C − C* = (C_i − C*)(1 + k/Pq)^−P, where P is the apparent number of tanks-in-series and captures the non-ideal, dispersed hydraulics of a real bed. As P → ∞ it collapses back to the plug-flow k-C* form above. Both are areal models, so a defensible hydraulic design still turns on getting the loading rate and rate constant right.

Worked example: sizing an HSSF wetland for BOD removal

Size a horizontal subsurface flow wetland to polish a settled effluent.

• Inputs: Q = 200 m³/d; inlet BOD C_i = 150 mg/L; target outlet C = 20 mg/L; background C* = 5 mg/L; areal rate constant k = 0.10 m/d (HSSF BOD, ~37 m/yr).
• Concentration ratio: (C_i − C*)/(C − C*) = (150 − 5)/(20 − 5) = 145/15 = 9.67.
• Natural log: ln(9.67) = 2.27.
• Area: A = (Q/k) · ln(…) = (200/0.10) × 2.27 = 2000 × 2.27 = 4,540 m² (about 0.45 ha).
• Check the hydraulic loading rate: q = Q/A = 200/4,540 = 0.044 m/d = 4.4 cm/d, comfortably inside the 2–8 cm/d band for HSSF.

So a first-pass HSSF bed of roughly 4,500 m² meets the BOD target. Before committing, the same area must be re-checked against the total-nitrogen target using the TN rate constant and against the winter temperature (below), because nitrogen — not carbon — usually governs the final footprint.

Why do nitrogen and temperature complicate the design?

Two coupled constraints stop a wetland behaving like the tidy exponential above. The first is oxygen transfer. In a saturated HSSF bed, oxygen enters only by slow diffusion and root release — on the order of a few grams O₂ per m² per day — which is far below the stoichiometric demand to nitrify a typical ammonia load (4.57 g O₂ per g N oxidised). Ammonia removal in HSSF is therefore oxygen-limited, which is exactly why vertical flow beds, re-aerated convectively as each dose drains, are chosen when nitrification is required.

The second is temperature. Biological rate constants follow an Arrhenius (modified van't Hoff) relationship, and nitrogen removal is far more temperature-sensitive than BOD removal:

That 2.5-fold drop in the nitrogen rate constant through a cold winter can more than double the area needed to hold a TN consent, so wetlands serving a tight nutrient limit must be sized for the coldest operating month, not the annual mean.

What causes clogging, and how do you prevent it?

Clogging is the dominant long-term failure mode of subsurface wetlands. Suspended solids accumulate, biofilm and root biomass grow, and chemical precipitates form, progressively reducing the pore space and hydraulic conductivity of the media. As conductivity falls the water table rises, flow short-circuits over the surface, contact time collapses and treatment deteriorates — the bed effectively reverts to an uncontrolled surface-flow system.

The main defences are designed in, not retrofitted:

• Effective pre-treatment — primary settlement or screening to cap the TSS load reaching the bed, since solids are the leading clogging agent. Robust upstream solids and reuse-oriented pre-treatment materially extends bed life.
• Correct media grading — a coarse, well-sorted gravel (typically 8–16 mm) balances hydraulic conductivity against surface area for biofilm.
• Conservative organic loading — keeping areal BOD loading within accepted limits (roughly 6–10 g BOD/m²/d for HSSF) avoids excess biofilm growth.
• Intermittent dosing and rest cycles — for VF beds, alternating loaded and resting cells mineralises the accumulated organic layer and recovers conductivity.

What does a sound wetland design and operation require?

Beyond the area calculation, a defensible design fixes the hydraulics and the maintenance regime:

• Aspect ratio and inlet distribution — a length-to-width ratio and a distribution manifold that spread flow evenly and avoid the short-circuiting that a tracer test would expose.
• Liner and hydraulic control — an impermeable liner where groundwater protection is needed, plus an adjustable outlet to set water depth and residence time.
• Planting — establishing Phragmites or equivalent macrophytes to build root structure, biofilm surface and rhizosphere oxygen release.
• Monitoring — tracking inlet/outlet BOD, TSS and N, water levels (a clogging indicator) and seasonal performance against the temperature-corrected design.

Treated as a low-energy, long-life asset with the footprint set by the worst-case rate constant, a constructed wetland is one of the most robust polishing stages available; undersized on an annual-average k, it will breach its consent in the first cold winter.

How to size a constructed wetland with the k-C* areal model

1. Characterise the flow and load. Establish design flow Q (m³/d) and inlet concentrations for BOD, TSS and total nitrogen from representative sampling.
2. Select the wetland type. Choose FWS, HSSF, VF or a hybrid based on the oxygen regime the target pollutants require — VF for nitrification, HSSF for denitrification.
3. Choose rate constant and background. Pick an areal rate constant k20 and background C* for the governing pollutant from Kadlec and Knight data for that wetland type.
4. Correct for temperature. Adjust k to the coldest operating month using kT = k20·θ^(T−20), with θ ≈ 1.05–1.10 for nitrogen.
5. Compute the area. Solve A = (Q/k)·ln[(Ci − C*)/(C − C*)] for each target pollutant; the largest area governs.
6. Check hydraulic and organic loading. Confirm q = Q/A sits within the type band (e.g. 2–8 cm/d for HSSF) and areal BOD loading stays within clogging-safe limits.