Duckweed on Manure: What the Field Pilots Measured
Duckweed on manure effluent removes real nitrogen and phosphorus, then stalls near day 100. What outdoor pilots measured, and why the feed half is blocked.
Search duckweed manure and you get journal abstracts. They agree with each other, and they all promise the same two things at once: nitrogen and phosphorus out of a livestock waste stream, protein-rich biomass out of the pond.
Half of that promise has season-long outdoor numbers behind it. The other half runs into European feed law, and the project that built the first semi-industrial prototype in the EU said so in its own closing report.
This article is the operator’s version. What the pilots fed the ponds, what they measured, why growth stops around day 100, and what happens to the harvest.
Duckweed never goes on manure
The keyword is misleading and it is worth fixing before anything else. No successful pilot has grown duckweed on slurry. Raw manure is too turbid for light to reach a floating plant and far too concentrated in ammonia for one to survive in.
What the pilots feed is a treated, diluted derivative, and which derivative matters more than almost any other design choice. The Belgian systems at a working pig manure processing facility ran on a mix of 90 percent nitrification-denitrification effluent and 10 percent liquid fraction. The Spanish LIFE LEMNA prototype ran on the liquid phase of anaerobically digested pig manure, on the logic that digestion mineralises organic nitrogen and phosphorus first and so raises what the plant can actually take up.
So the pond is a polishing and nutrient-export step at the tail of a manure treatment train, not a replacement for one. The general case is covered separately in what a duckweed step actually removes.
What a full outdoor season measured
The strongest evidence for this phrase is recent. Three pilot systems of 10 square metres each, monitored at a Belgian pig manure processing facility across an entire growing season under temperate maritime conditions.
Nitrogen and phosphorus off the water
Average removal came out at 1.2 grams of nitrogen per square metre per day and 0.13 grams of phosphorus per square metre per day. The authors note those rates are consistent with the literature and comparable to other biological treatment systems on N and P.
That is the honest headline. A duckweed step is not extraordinary at nutrient removal. It is comparable, and it differs in what it does with the nutrients afterwards, which is the whole argument for it.
The Spanish work reports the same thing on a different metric. At the 250 square metre prototype in Catalonia, the LIFE LEMNA project measured uptake rates of 54 percent for nitrogen, 45 percent for phosphorus and 71 percent for potassium, noting that nitrogen is fully depleted inside the system. Percentages and areal rates are not interchangeable. Quote whichever your regulator asks for, and do not average them.
Biomass and protein
The two working Belgian systems yielded 8 tonnes of dry biomass per hectare per year and 2.8 tonnes of protein per hectare per year.
The same technology in Spain produced more than 17 tonnes of dry matter per hectare per year at 35 to 40 percent protein. More than double, on the same continent, with the same plant family.
Nothing separates those two numbers except season length and light. If you take one thing from the yield literature, take that: climate is not a modifier on a duckweed yield figure, it is the dominant term. A number measured in Flanders tells you very little about the Canary Islands, and the reverse is equally true.
The system that died
The third Belgian system was fed constructed wetland effluent, the dischargeable water at the end of the farm’s existing treatment. It suffered die-off and low protein content, with electrical conductivity reaching 12 millisiemens per centimetre.
That result is more useful than the two that worked. The water was clean enough to discharge and still killed the crop. Salinity, not nutrient shortage, was what the plant could not tolerate.
Why growth stops around day 100
The finding that deserves more attention than it gets: in the systems that grew consistently, growth ceased after approximately 100 to 120 days, despite suitable weather and despite single-nutrient concentrations that were not individually toxic.
Two mechanisms, working together.
Rising pH turns ammonium into ammonia
Ammonium and free ammonia sit in a chemical equilibrium governed by pH. A dense duckweed mat drawing carbon out of the water pushes pH upward, and as it climbs, the same total ammonium load presents more of itself as free ammonia. Free ammonia is what damages the plant.
So the pond can become toxic without anyone adding anything to it. Measure ammonium alone on a weekly grab sample and this arrives invisibly.
Salinity, and what a recirculating pond quietly concentrates
The second mechanism is slower and structural. A nine-week recirculation study by the same Flanders group tracked the full ionic composition rather than only N and P, and found that potassium, chloride, sulphur, calcium and magnesium are abundant in swine manure waters and tend to accumulate even while nitrogen and phosphorus are being removed as intended.
That is the trap in a closed loop. The system is doing exactly what it was built to do, removing the two nutrients on the discharge consent, while concentrating five others that nobody is watching. Electrical conductivity rises, and eventually the crop stalls.
The season-long paper’s own design conclusion follows from this: dimension the system on the most limiting nutrient rather than on nitrogen, and note that more surface area or more depth buys buffering capacity. Depth and area are cheap compared with a crop that stops in August, and both feed straight into what a duckweed system costs to build and run.
The feed half of the promise, and the regulation nobody links to
Every paper on this topic is framed as nutrient recovery plus feed production. The feed half deserves a plain statement of where it stands in the EU, because the ranking pages do not give one.
What the feed regulation says
Regulation (EC) No 767/2009 governs what may be placed on the market as feed. Article 6(1) points at Annex III, a list of prohibited materials, and the first entry on that list is faeces, urine and separated digestive tract content, irrespective of any form of treatment or admixture. The fifth entry excludes all waste from the treatment of urban, domestic and industrial wastewater, irrespective of any further processing.
Read those two clauses next to a duckweed pond fed on pig slurry derivatives and the difficulty is obvious. The phrase doing the work is irrespective of any form of treatment.
The project that built the prototype named it
This is not a theoretical reading. The European Sustainable Phosphorus Platform wrote to the Commission’s food and feed safety directorate in 2021 asking for a legal opinion on exactly these clauses, and its accompanying table names duckweed grown in pig manure as a specific affected case, alongside algae grown on digestate, listing the Belgian researchers by name.
And LIFE LEMNA, having built and run the first semi-industrial prototype of this kind in Europe, closes its EC project record by stating that the main barrier to going full scale is the prohibition on using duckweed grown on pig slurry digestate as animal feed, on hygiene grounds.
The reading is contested rather than settled, and none of this is legal advice. An operator has to confirm the position with their own competent authority before it becomes a business assumption.
Contaminants are the easy part
Here is what makes the block worth understanding properly. The contaminant question, the one the papers do test, comes out fine.
Across the Belgian season, cadmium, lead, arsenic and mercury in the harvested biomass were all measured well below the maxima in Directive 2002/32/EC. The 2021 study found the same for arsenic, cadmium and lead, with one practical caveat worth knowing: a starting inoculum can carry historic contamination from wherever it was collected, which then dilutes out over the first few weeks of cultivation.
So the barrier is categorical, not analytical. A clean certificate of analysis does not move it, because the exclusion attaches to what the biomass was grown on rather than what is in it. The category exists for a reason the Wageningen survey states plainly: there are concerns about pathogen transfer when animals are fed plant material grown on their own waste, and the very limited research available suggests it may not be a problem while leaving the potential for great harm unresolved.
When duckweed protein is worth discussing as feed, it is on a clean input stream, and that is duckweed as animal feed on the clean stream, a different article about a different pond.
Where SERAPH stands: remediation-grade from the first reading
Our position on manure is published and it is not a hedge. Biomass grown on wastewater, manure or contaminated water is never sold or blended as food or feed, and the two streams are separated from the first sensor reading rather than at the point of sale.
The reason is engineering rather than compliance. A stream you classify at the end is a stream you cannot trace. Containment, harvest routing and record-keeping all have to be decided before the first pond is dug, because retrofitting traceability onto a system that was built to keep its options open is not possible. Deciding the destination early costs you optionality and buys you a defensible chain of custody, and on a remediation stream that is the correct trade.
That position is built on a mapped literature, 256 indexed duckweed papers across 31 research areas, rather than on a pilot of our own. We are at TRL 3, proof of concept, with outdoor field validation still ahead, and we say so. Every number in this article belongs to somebody else and is linked to its source. What we offer on this topic is the system design and the reasoning, not a yield figure.
What to measure before you build
Six things, all of them drawn from what the pilots learned rather than from a vendor checklist.
Continuous pH and electrical conductivity, not weekly grab samples. Both are what ended the Belgian season, and both move gradually enough to hide between samples.
The full ionic composition of your feed stream. Potassium, chloride, sulphur, calcium and magnesium, not just nitrogen and phosphorus. What accumulates is what stops you.
An exchange plan, not a pure recirculation plan. Closed loops concentrate everything you are not removing. Decide the bleed rate before commissioning, not when the crop stalls.
Season length as a primary design input. The gap between the Belgian and Spanish yields is more than a factor of two on comparable technology.
Which derivative feeds the pond. Liquid fraction, digestate liquid phase and nitrification-denitrification effluent are chemically different starting points, and one of the three tested wetland effluents killed the crop outright.
The destination of the harvest, decided first. It determines containment, routing and traceability, and harvesting is the step that exports the nutrients off the site, so it is the step where the classification becomes physical. The broader treatment case is covered in phytoremediation with duckweed.
Built for the field, not the lab. On manure, that means designing for the day the pond stops growing, and for the paperwork that follows the biomass out of the gate.
FAQ
Can you put duckweed directly on a manure lagoon?
No. Raw slurry is too turbid for light to reach the fronds and far too high in ammonia for the plant to survive. Every pilot that worked fed a treated and diluted derivative: the liquid fraction after solid separation, effluent after nitrification-denitrification, or the liquid phase after anaerobic digestion.
How much nitrogen can a duckweed pond remove from manure effluent?
The Belgian season averaged 1.2 grams of nitrogen and 0.13 grams of phosphorus per square metre per day. The Spanish prototype reported uptake of 54 percent of nitrogen, 45 percent of phosphorus and 71 percent of potassium. Those are two different metrics measured under two different climates, and neither transfers cleanly to a third site.
Is duckweed grown on pig manure legal as animal feed in the EU?
It is contested and currently blocked in practice. Annex III of Regulation (EC) No 767/2009 lists faeces and urine as prohibited feed material irrespective of any form of treatment, and LIFE LEMNA named that prohibition as the main barrier to upscaling. This is not legal advice, and the position must be confirmed with the competent authority in your member state.
Does duckweed grown on manure accumulate heavy metals above feed limits?
In the published pilots it did not. Cadmium, lead, arsenic and mercury were all well below the Directive 2002/32/EC maxima. That is precisely why the regulatory question matters: the block is categorical rather than analytical, so a clean result does not resolve it.
If the biomass cannot be feed, what is it for?
Nutrient export is the point of the pond, and what physically leaves in the harvest is what leaves the nutrient balance. LIFE LEMNA assessed a biofertiliser route for the liquid fraction and a return of biomass to the on-site biogas plant, so the treatment step runs on its own energy.
Does anaerobic digestion before the pond help or hurt?
It helps the biology and does not change the legal status. Digestion mineralises organic nitrogen and phosphorus, which raises what the plant can take up, and is why the Spanish design put the pond after the digester. The feed exclusion attaches to the manure origin, so the digester does not clear it.