Phosphorus and Innovation
Lakes, rivers and wetlands in Canada hold 7% of the world’s renewable freshwater resources with 10.8% of Ontario covered by freshwater. The 250,000 lakes, 500,000 km of rivers and streams, and vast groundwater resources in Ontario provide water supply and environmental and recreational benefits. However, water quality in some of these lakes and rivers has been affected by multiple stressors that have altered the ecology of these ecosystems1. Land use change2, increased nutrient loadings3, and climate change4 are often cited as the three major culprits impacting these fragile ecosystems. Excessive phosphorus loadings reaching freshwater bodies have been the leading cause of impaired rivers and lakes in Ontario and worldwide5, 6. Impacts of excess phosphorus algal blooms, low dissolved oxygen, proliferation of invasive species, fish kills, and ultimately, ecological collapse. Economic and health impacts from phosphorus-driven algal bloom events have also been reported7, 8.
There is a need for reducing phosphorus loads through innovative practices. Several possible solutions already exist to remove excessive phosphorus discharged into a freshwater body9, 10. These solutions range from implementing best management practices at the source level to downstream integration of constructed wetlands that filter water through natural processes11, 12. However, the current technologies are prohibitively expensive, owing to large land requirements, expensive materials, and/or high capital, operating, and maintenance costs. Furthermore, costs dramatically increase when treating water at relatively low phosphorus concentrations and under different conditions (variable temperature, flows, concentrations, etc.). There is a worldwide urgent need to invest in innovation to incentivize the development of new, cost-effective approaches to remove phosphorus and achieve low concentrations13.
The $10 million George Barley Water Prize, led by The Everglades Foundation and designed by international experts, aims to drive this innovation process. The goal of the competition is to find a breakthrough total phosphorus (TP) removal technology that is radically cheaper to build and run than currently available removal technologies. The winning technology would remove phosphorus under variable conditions while not impacting the environment. The prize was designed in stages, emulating the development of any new technology, with technologies tested in a laboratory environment and at a pilot scale, before being tested at a large scale. Technology evaluation criteria consist of TP flow weighted mean (TP-FWM) at the outflow, environmental sustainability, scalability, and total costs.
Experts in chemical engineering, hydrology, water quality, and ecology selected 9 promising teams from the first two stages of the George Barley Water Prize to enter the pilot phase competition. The chosen teams were: University of Idaho-blueXgreen-Nexom (USA), Econse Water Purification Systems (Canada), ESSRE RePleNish (USA), Global Phosphate Solutions (USA), Green Water Solution (USA), MetaMateria (USA), University of Waterloo (Canada), USGS (USA), Wetsus (The Netherlands), and ZeroPhos (China).
George Barley Water Prize Pilot Phase Competition
During the pilot phase, the nine teams were challenged to treat 9,464 liters per day for two 30-day periods, and 32,176 liters per day for one 30-day period of Holland Marsh canal water for TP removal. In addition to the technology evaluation criteria, technologies had to fit within a 9 m2 footprint. Due to the significance of the Holland Marsh phosphorus loading to Lake Simcoe, the Art Janse Pumping Station (Fig. S-1) was selected as the location where the nine teams tested their technologies, treating the Holland Marsh canal water (Figs. S-2 and S-3). This type of pilot testing, where several technologies compete at the same time and from the same source of influent water to achieve an effluent TP target, while not impacting the background water quality characteristics, had never been done before at this scale. The objective of this paper is to present the TP removal performance of the nine technologies and the results from the three months of testing, including factors which may have influenced the efficiency or operation of the competing technologies.
George Barley Water Prize Pilot Phase Competing Teams
Econse Water Purification Systems technology (Team D) is an integrated electrocoagulation system (Fig. S-4) that combines various technologies into a compact mechanical based unit that is scalable and treats water from a variety of applications, including food and beverage, industrial, agricultural and small to medium sized communities. At the core of the technology is a reactor, using electrolysis (a proprietary anode) in combination with liquid/solid separators. Anodic dissolution of the metal electrode in contact with the polluted water would form a cation that would bind dissolved phosphorus and generate a precipitate settling in the tanks. No filtration was used in this process.
The ESSRE RePleNish technology (Team E) uses nano-enhanced adsorptive media, which is Hybrid Ion Exchange (HIX) resin beads dispersed with immobilized Fe(III) or Zr(IV) oxide nanoparticles with a strong affinity for phosphate ions over the other common anions in water (Fig. S-5)14, 15. These hybrid ion exchange nano resins are also amenable to regeneration and reuse for tens of cycles, thus reducing the product cost of commercially available HIX-Nano. Unlike conventional ion exchange disposal issues of a waste brine, HIX-Nano regeneration results in high-strength phosphate solutions that can be customized as N-P-K liquid plant food used by hydroponic and greenhouse growers16. Multi-media pressure and cartridge filters were used upstream of HIX-Nano resin to keep the media free of solids that could foul or plug the bed of resin.
Global Phosphate Solutions (Team C) use a “Phosphate Sponge”, which is a porous pellet composite that integrates a patented polyacrylonitrile binder with a composite material to create a system that can effectively remove phosphate contamination (Fig. S-6). The sponge can be easily regenerated to be reused. The material is most efficient at pH 6-7 and can be regenerated at pH 12-13. No filtration was used in this process.
Green Water Solution’s approach (Team I) was invented and improved over the last 7 years in collaboration with Dutch and German science institutes. This team applied a two-stage process for their solution: removal of particulate phosphorus by filtration followed by adsorption of dissolved phosphorus by the BioPhree® system (Fig. S-7). BioPhree® uses a proprietary high-affinity composite resin (a proprietary polymer with P-affine coating) combined with metal hydroxides and iron oxides to reversibly adsorb inorganic and organic phosphorus. In the regeneration cycle, the captured phosphorus is released together with co-captured organic carbon. The regeneration liquid can be re-used in following regeneration cycles to significantly lower the operating costs of the treatment, then be purified and used as a fertilizer. A pre-treatment with cartridges and bags were used in this process targeting particulates.
University of Idaho-blueXgreen-Nexom (Team G) used a hydrous ferric oxide (HFO) reactive filtration process with a high-efficiency for adsorptive removal of dissolved phosphorus in addition to particulate phosphorus removal (Fig. S-8)17, 18. Along with iron metal salt dosing, the process can also add micronized, Ca-Mg-Fe modified biochar from bioenergy pyrolyzed greenwaste to the flowing water for P recovery. The flow moves through a continuous-backwash, up-flow moving bed sand filter operated as an HFO coated-sand reactive filter. The influent water was first mixed with a recycled reactive filter continuously rejecting water containing particulates into separation tanks equipped with particle separators before continuing to the reactive filters. Periodic settling tank purged solids were collected in bag filters. For phosphorus recovery, the recovered biochar rejected solids mixture is recovered with the process water solids. The recovered phosphorus-upcycled biochar is then applicable for use as a slow-release fertilizer and soil amendment.
The University of Waterloo PhosphexTM system (Team A) incorporated a by-product of the steel industry (Basic Oxygen Furnace – BOF-slag) to adsorb and precipitate phosphorus (Fig. S-9)19, 20. Although the system was designed to be gravity driven, space limitations required the use of pumps and holding tanks. Once saturated with phosphorus, the slag can be used as a soil additive or in construction applications. During the trial, this team used bags and cartridges filters targeting particulates.
The U.S. Geological Survey- Leetown Science team (Team F) used iron oxide-based sorption media in a fixed bed process (Fig. S-10)21. The media is composed of ochres generated by the treatment of mine drainage. The process offers the possibility of significant reductions in capital and operating costs for the removal of phosphorus because of the economy of the by-product media coupled with elimination of solid-liquid separation and sludge disposal required by conventional iron or aluminum-based coagulation processes. An additional feature of the technology is the capability to recover phosphorus as a potentially marketable fertilizer, thus closing the phosphorus loop. Pre-filtration cartridges and flocculants were used in this process targeting particulates.
Wetsus NaFRAD (Team B) used a “total solution” to total phosphorus pollution, using a combination of adsorption and flocculation to remove total phosphorus and recover phosphate (Fig. S-11). In the first step, natural organic flocculants agglomerate particulates into flocs that are collected in fixed bed sand filters. The captured particulates are removed periodically through backwashing and collected from the washing liquid in a gravity separator. Soluble phosphorus is removed through adsorption on iron-based adsorbents with a high phosphorus sorption22. The phosphorus saturated adsorbent can be regenerated in an on- or offsite facility with sodium hydroxide to release the phosphate. The phosphate is recovered from the regeneration liquid in a crystallizer as pure calcium phosphate through controlled addition of calcium hydroxide. In this way, the regeneration liquid can be used multiple times.
Zerophos (Team H), from the University of Nanjing, used a Lanthanum-based polymeric nanocomposites adsorbent (denoted as “La-201”) developed specifically for deep treatment of phosphorus contaminated water (Fig. S-12)23. La-201 takes advantages of the high mechanical strength and hydraulic properties of the polymeric host, and the high activity of the embedded hydrous lanthanum oxide nanoparticles. The material could be regenerated and reused repeatedly without significant capacity loss. The small spherical beads could fit into many fast filtration systems such as fixed-bed columns. The technology is flexible and of small footprint and land-use; the whole system is stable with easy operation and maintenance. Pre-filtration cartridges and bags were used in this process targeting particulates.
Description of the monitoring protocol, site set-up and flow distribution is detailed in the supplemental information.
Over the duration of the pilot, air temperatures averaged -1.8? during the first period of testing, 1.4? during the second period of testing, and reaching 12.4? during the third period of testing. The minimum temperature was -6.6? on March 16 and the maximum temperature was 20.2? on May 2. The total precipitation during the pilot testing was around 217 mm. The maximum precipitation occurred on April 14 with 26.9 mm. Precipitation amounts greater than 10 mm per day occurred on March 29, April 3, April 13, April 14, April 15, April 16, April 25, and May 15. During and immediately after snow fall events (February and March), no visually recognizable change was observed in the influent water quality. During and immediately after rainfall events (April and May), influent water became substantially more turbid (Fig. 1). This is likely attributed to upstream agricultural practices; most fields in the marsh are cultivated close to the edge of the canals with no riparian buffers. Each spring to facilitate planting, agricultural producers actively pump water off their fields into tile drains and ditches which feed into the West Holland River. Most producers started planting around the first week of May, thus explaining the high turbidity peaks observed during the first week of May. Of particular note, the site was closed from April 15 to April 19 because of freezing rain and on May 4 because of winds (>90 km/hr) and tornado warnings.
The Holland Marsh agricultural drainage canal site chosen for the pilot phase presented a unique set of agro-environmental and water quality challenges. The Holland Marsh is a muck soil growing region containing 20-80% organic matter24. The muck-specific gravity equivalence with water inhibits settling relative to silt and clay soil particles24. Holland Marsh soil turnover, fertilization, and rainfall-driven active periods of surface/subsurface water pumping have demonstrated 4-5 times increases in TP runoff and 40-50 times increases in nitrate-N runoff25.
During the three-month testing period, the average daily inflow TP concentration was 410 ± 205 µg/L with a maximum concentration of 1,494 µg/L and a minimum concentration of 190 µg/L. The inflow ortho-phosphate was not measured as frequently as the TP with a level of 407±217 µg/L with a maximum concentration of 920 µg/L and a minimum concentration of 40 µg/L. Surprisingly, the ortho-phosphate concentrations during the same period did not differ much from the TP levels. A long-term monitoring program at the same location indicated a TP level of 520±150 µg/L and an ortho-phosphate fraction of 75% of TP. Influent total suspended solids varied between 0.25 to 26 mg/L. The daily influent pH ranged from 7.2 to 8.3 with an average of 7.84±0.24. In situ daily influent conductivity and turbidity averaged at 1,168 µS/cm and 10.15 (FNU), respectively. The inflow water had a high total organic carbon content with an average of 22.7 mg/L. The average total nitrogen was at 7.9 mg/L with the nitrate fraction representing 75% of the total nitrogen. Total alkalinity was at 271.5 mg/L (total as CaCO3). Chloride concentrations, particularly important because of de-icing, were around 121.3 mg/L during the three testing periods. The average water temperature was at 4°C, 8.5°C, and 17°C during the three testing periods, respectively.
Total Phosphorus and Orthophosphate
Collectively during the three periods of testing, the nine teams treated 10,962 m3 of water and removed 2.79 kg of TP, representing an average TP loading decrease of 65%. Some teams reached an average TP loading decrease of 86% over the three testing periods. Table S-3 summarizes the TP-FWM for all teams with an inflow TP-FWM around 394 µg/L (with a spike of TP levels during the second testing period when the rain/drainage started). When checking the TP performance, it was apparent that the nine teams can be easily separated into four categories based on the TP concentration at the outflow and on Fig. 2.
The first category presented in Fig. 2 and Fig. S-13a is the top performing class (Teams F and I) consistently achieving low TP concentrations at the outflow and high TP loading removal even when jumping from low to high flow rates. The second category is the good performing class (Teams A, B, G, and H) achieving low TP concentrations and high TP loading removal but not consistently when flows increased (Fig. 2 and Fig. S-13b). It is worth mentioning that the flow rate into Team A pod was highly variable, due to restrictions in the inflow line providing water to this pod (situation rectified near the end of the high flow period). The third category formed by Teams C and E is not capable to achieve low TP levels or high TP loading removal (Fig. 2 and Fig. S-13c). At high flows, some of these teams’ filters got clogged very fast. At one point, Team E’s filter needed to be changed every 15 minutes because their flow rate would significantly drop. Team D would belong to a fourth category with TP concentration and TP loading removal levels fluctuating from one day to another (Fig. 2 and Fig. S-13d), with high Total Suspended Solids – colloidal materials that were not settling down during the treatment process.
Even though the teams did not achieve the 10 µg/L TP-FWM criterion, note the performance of Team G achieving very low orthophosphate levels below 15 µg/L 79% of the time (Fig. 3). Additionally, Team G demonstrated consistent performance achieving a TP average of 10 µg/L TP for the 15-day period of March 13 to 27. It is worth mentioning that this team had some problems during the first few weeks of testing where inflow water was mistakenly routed into their final effluent line, resulting in high TP outflows. The team was able to correct the problems which immediately enhanced their technology performance. Teams A, D, F and I also had low orthophosphate levels when sampled (thus emphasizing the particulate-P skewing the TP levels results for Team D). Those four teams achieved orthophosphate levels lower than 30 ppb more than 75% of the time. When sampled, Teams B and H achieved 10 µg/L orthophosphate concentration multiple times but not consistently. Teams C and E performance were not to the same level when compared to the other teams. The six teams (A, B, F, G, H and I) had a TP removal ranging from 79% to 86%. Teams D and E reached a 35-37% TP removal. Team C removed 13% of the TP.
Toxicity and Byproducts
Inflows and all teams’ effluents were tested for chronic and acute toxicity to determine the IC25 and LC50 for Ceriodaphnia Dubia (CD) and Pimephales Promelas (PP). All teams passed the tests except Teams D and E. Team D failed the chronic toxicity for CD with an IC25 of 15.6% (IC25 for PP and LC50 for both species exceeded the 100% level). Team E failed the chronic and acute toxicity tests for both species. However, the inflow for this team also didn’t pass the toxicity tests. Team E IC25 for CD was 86.3% higher than the 24.1% for the influent, indicating that Team E outflow was less toxic than the influent water. Results of the NASM test showed that all teams (except Team C) generated byproducts that could be potentially land applied (Table S-5). Teams I and B were the only teams with byproducts satisfying the nutrient content condition (plants available nitrogen, phosphorus and potassium exceeding 13,000 mg/kg).
During the three months of testing and because of the high turbidity of the inflow waters, competitors experienced flow drops and spikes, impacting their treatment technologies mainly during high flow period. The onsite engineers identified several problems causing the flow drops such as a clog in the influent hose and clogging of the screen inside the supply pump with duckweed and sediment.
In order to improve parallel testing of several technologies, some important insights can be shared. First, a historical long-term water quality monitoring of the site was needed to help the teams better design their filtering process for the testing conditions. Moreover, teams needed a longer period to set-up their system and calibrate their process before launch. This would have dramatically improved the performance of the technologies as they would be better prepared for the site water quality conditions. Another important factor is the engineering set-up to be carefully considered based on the quality of the water at the site. In the present case, flow meters clogged and regular cleaning was required. The third important aspect is that the site tours were extremely useful to engage the teams with a variety of interested stakeholders. Stakeholders enjoyed the tours and peaked their interest to follow up for results/reports. The tours were also a good opportunity to bring in media and build awareness via articles/videos/social media.
Teams F, I, and G achieved the lowest TP-FWM concentration levels among all nine teams during the three-month testing period. Even though the top teams did not achieve the 10 µg/L TP-FWM criterion, the three teams achieved very low orthophosphate levels below 10µg/L when sampled. The phosphorus removal and water quality performance of the teams, as well as cost considerations, which have not been discussed in this paper, were used to identify the four teams to move to the Grand Stage. The results and lessons learned from this testing will help inform stakeholders as they explore future breakthrough and innovative phosphorus removal technology applications in waterbodies that have challenges with phosphorus pollution.
Any opinions, findings, conclusions, or recommendations expressed in the material are those of the author(s) and do not necessarily reflect the views of the Everglades Foundation. Any use of trade, product, or firm names is for descriptive purposes only and does not imply endorsement by the U.S. Government. Detailed acknowledgments are included in the Supporting Information.
- Ormerod, S. J.; Dobson, M.; Hildrew, A. G.; Townsend, C. R., Multiple stressors in freshwater ecosystems. Freshwater Biology 2010, 55, 1-4.
- Allan, J. D., Landscapes and riverscapes: The influence of land use on stream ecosystems. Annual Review of Ecology Evolution and Systematics 2004, 35, 257-284.
- Watson, S. B.; Miller, C.; Arhonditsis, G.; Boyer, G. L.; Carmichael, W.; Charlton, M. N.; Confesor, R.; Depew, D. C.; Hook, T. O.; Ludsin, S. A.; Matisoff, G.; McElmurry, S. P.; Murray, M. W.; Richards, R. P.; Rao, Y. R.; Steffen, M. M.; Wilhelm, S. W., The re-eutrophication of Lake Erie: Harmful algal blooms and hypoxia. Harmful Algae 2016, 56, 44-66.
- Robertson, D. M.; Saad, D. A.; Christiansen, D. E.; Lorenz, D. J., Simulated impacts of climate change on phosphorus loading to Lake Michigan. Journal of Great Lakes Research 2016, 42, (3), 536-548.
- Carpenter, S. R.; Caraco, N. F.; Correll, D. L.; Howarth, R. W.; Sharpley, A. N.; Smith, V. H., Nonpoint pollution of surface waters with phosphorus and nitrogen. Ecological Applications 1998, 8, (3), 559-568.
- Stammler, K. L.; Taylor, W. D.; Mohamed, M. N., Long-term decline in stream total phosphorus concentrations: A pervasive pattern in all watershed types in Ontario. Journal of Great Lakes Research 2017, 43, (5), 930-937.
- Brooks, B. W.; Lazorchak, J. M.; Howard, M. D. A.; Johnson, M. V. V.; Morton, S. L.; Perkins, D. A. K.; Reavie, E. D.; Scott, G. I.; Smith, S. A.; Steevens, J. A., Are harmful algal blooms becoming the greatest inland water quality threat to public health and aquatic ecosystems? Environmental Toxicology and Chemistry 2016, 35, (1), 6-13.
- Dodds, W. K.; Bouska, W. W.; Eitzmann, J. L.; Pilger, T. J.; Pitts, K. L.; Riley, A. J.; Schloesser, J. T.; Thornbrugh, D. J., Eutrophication of US freshwaters: Analysis of potential economic damages. Environmental Science & Technology 2009, 43, (1), 12-19.
- Benafqir, M.; Anfar, Z.; Abbaz, M.; El Haouti, R.; Lhanafi, S.; Azougarh, Y.; El Fakir, A. A.; Ez-zahery, M.; El Alem, N., Hematite-titaniferous sand as a new low-cost adsorbent for orthophosphates removal: Adsorption, mechanism and Process Capability study. Environmental Technology & Innovation 2019, 13, 153-165.
- Yamada-Ferraz, T. M.; Sueitt, A. P. E.; Oliveira, A. F.; Botta, C. M. R.; Fadini, P. S.; Nascimento, M. R. L.; Faria, B. M.; Mozeto, A. A., Assessment of Phoslock (R) application in a tropical eutrophic reservoir: An integrated evaluation from laboratory to field experiments. Environmental Technology & Innovation 2015, 4, 194-205.
- Christopher, S. F.; Tank, J. L.; Mahl, U. H.; Yen, H.; Arnold, J. G.; Trentman, M. T.; Sowa, S. P.; Herbert, M. E.; Ross, J. A.; White, M. J.; Royer, T. V., Modeling nutrient removal using watershed-scale implementation of the two-stage ditch. Ecological Engineering 2017, 108, 358-369.
- Nairn, R. W.; Mitsch, W. J., Phosphorus removal in created wetland ponds receiving river overflow. Ecological Engineering 2000, 14, (1-2), 107-126.
- Dillon, P. J.; Evans, D. O.; Gharabaghi, B.; Longstaff, B.; Molot, L.; O’Halloran, I.; Pond, B.; Wesley-Esquimaux, C.; Winter, J. G. Recommendations for Lake Simcoe and its watershed; Minister of the Environment and Climate Change: 2012.
- Sengupta, A. K.; Padungthon, S. Hybrid Anion Exchanger Impregnated with Hydrated Zirconium Oxide for Selective Removal of Contaminating Ligand and Methods of Manufacture and Use Thereof. 2013.
- Sengupta, A. K.; Cumbal, L. H. Hybrid Anion Exchanger Impregnated for Selective Removal of Contaminating Ligands from Fluids and Method of Manufacture Thereof. 2007.
- Weinberg, E.; Sengupta, A. K.; Shepsko, C. In Novel hybrid Ion Exchange resins for concurrent removal, recovery, reconcentration, reuse and recycle of pollutant nutrients (Nitrogen) N and Phosphorus (P), Nutrient Symposium, 2017; Water Environment Federation: 2017; pp 408-425.
- Newcombe, R. L.; Rule, R. A.; Hart, B. K.; Moller, G., Phosphorus removal from municipal wastewater by hydrous ferric oxide reactive filtration and coupled chemically enhanced secondary treatment: Part I – Performance. Water Environment Research 2008, 80, (3), 238-247.
- Newcombe, R. L.; Strawn, D. G.; Grant, T. M.; Childers, S. E.; Moller, G., Phosphorus removal from municipal wastewater by hydrous ferric oxide reactive filtration and coupled chemically enhanced secondary treatment: Part II – Mechanism. Water Environment Research 2008, 80, (3), 248-256.
- Baker, M. J.; Blowes, D. W.; Ptacek, C. J., Laboratory development of permeable reactive mixtures for the removal of phosphorus from onsite wastewater disposal systems. Environmental Science & Technology 1998, 32, (15), 2308-2316.
- Hussain, S. I.; Blowes, D. W.; Ptacek, C. J.; Jamieson-Hanes, J. H.; Wootton, B.; Balch, G.; Higgins, J., Mechanisms of phosphorus removal in a pilot-scale constructed wetland/BOF slag wastewater treatment system. Environmental Engineering Science 2015, 32, (4), 340-352.
- Sibrell, P. L.; Montgomery, G. A.; Ritenour, K. L.; Tucker, T. W., Removal of phosphorus from agricultural wastewaters using adsorption media prepared from acid mine drainage sludge. Water Research 2009, 43, (8), 2240-2250.
- Kumar, P. S.; Korving, L.; Keesman, K. J.; van Loosdrecht, M. C. M.; Witkamp, G. J., Effect of pore size distribution and particle size of porous metal oxides on phosphate adsorption capacity and kinetics. Chemical Engineering Journal 2019, 358, 160-169.
- Zhang, Y. Y.; Pan, B. C.; Shan, C.; Gao, X., Enhanced phosphate removal by nanosized hydrated La(III) oxide confined in cross-linked polystyrene networks. Environmental Science & Technology 2016, 50, (3), 1447-1454.
- Visser, B. Impact of Muck Soils on Water Treatment Systems; Holland Marsh Growers Association: 2015; p 4.
- Nicholls, K. H.; Maccrimmon, H. R., Nutrients in subsurface and runoff waters of Holland Marsh, Ontario. Journal of Environmental Quality 1974, 3, (1), 31-35.
- VanderMey, A., And the Swamp Flourished: The Bittersweet Story of Holland Marsh. Vanderheide Publishing Company: 1994; p 145.
- Maccoux, M. J.; Dove, A.; Backus, S. M.; Dolan, D. M., Total and soluble reactive phosphorus loadings to Lake Erie A detailed accounting by year, basin, country, and tributary. Journal of Great Lakes Research 2016, 42, (6), 1151-1165.
- McDonald, M. R. Muck Vegetable Production in Ontario. Available at: https://www.agrireseau.net/documents/Document_91278.pdf
Fig. 1. Turbidity (FNU) and precipitation (mm) levels during the pilot stage testing.
Fig. 2. Teams phosphorus removal performance.
Fig. 3. Orthophosphate levels frequency of occurrence at the outflows from the different teams.