Lettuce (Lactuca sativa, variety Salanova) production in decoupled aquaponic systems
Metadata
- Cite key: monseesLettuceLactucaSativa2019
- Item type: Journal Article
- Authors: H. Monsees, J. Suhl, M. Paul, W. Kloas, D. Dannehl, S. Würtz
- Affiliation: Leibniz-Institute of Freshwater Ecology and Inland Fisheries, Berlin, Germany (Monsees, Suhl, Paul, Kloas, Würtz); Humboldt-Universität zu Berlin, Faculty of Life Sciences, Albrecht Daniel Thaer-Institute of Agricultural and Horticultural Sciences, Division Biosystems Engineering, Berlin, Germany (Suhl, Dannehl); Humboldt-Universität zu Berlin, Faculty of Life Sciences, Institute of Biology, Department of Endocrinology, Berlin, Germany (Kloas)
- Journal: PLOS ONE 14(6) (2019) e0218368
- Date: 06/2019 (received 10 Oct 2018, accepted 2 June 2019, published 20 June 2019)
- Date added: 2026-08-09
- DOI: 10.1371/journal.pone.0218368
- Funding: European Union’s Seventh Framework Programme for research, technological development and demonstration, grant agreement no. 619137 (HM and JS); Leibniz Association’s Open Access Publishing Fund (funded the open-access publication only)
- URL: https://doi.org/10.1371/journal.pone.0218368
- PDF:
Monsees et al. - 2019 - Lettuce (Lactuca sativa, variety Salanova) product.pdf
Opinion
A clean, well-instrumented three-arm comparison (hydroponic control vs. two decoupled-aquaponic variants differing only in whether the fish water was heat-disinfected) that lands on a genuinely useful negative result: neither dissolved organic carbon/microorganisms from the RAS nor disinfecting them changed lettuce growth, dry matter, nitrate, or phenolic content at all. The fertilizer/GHG accounting (Table 3) is the paper’s most citable contribution — a directly measured 62.8% mineral-fertilizer reduction and a calculated 72.6% GHG-emissions reduction from fertilizer manufacturing — and it is refreshingly free of internal contradictions. Two things temper full confidence: (1) the paper’s own coordinate string for the greenhouse location has an internally impossible DMS format in both latitude and longitude (seconds >59 in both), which could not be recovered with any confidence rather than guessed at (see Extraction notes); (2) most water-chemistry values (Table 4’s full nutrient panel, TAN, NO3-N) are single start-of-experiment snapshots rather than repeated trial-mean measurements, which is a real limitation of the design (a static batch of nutrient solution recirculated for 7-8 weeks) rather than an extraction artifact. Fish performance itself is entirely out of scope — the RAS supplying the fish water is an already-established, continuously operating system described in a companion 2017 paper, and no fish growth/feed/survival data specific to this lettuce trial exists to extract.
Abstract
Decoupled aquaponic systems have the potential to become one of the most effective sustainable production systems for the combined production of animal protein and plant crops. Here, recirculating aquaculture systems for fish production are combined with hydroponics for soilless plant production thereby recycling dissolved nutrients derived from metabolism of the fish. The aim of the present study was to characterize hydroponic lettuce production using conventional nutrient solution in comparison with decoupled aquaponics using the nutrient rich fish water as basis for the nutrient solution being supplemented by missing nutrients. In addition, one aquaponic treatment became disinfected in order to assess any occurring advantage of the aquaponics derived fish water. For evaluation the temperature, electrical conductivity, pH, and the mineral composition of the nutrient solution, as well as colony forming units in the fish water were monitored. Additionally, plant growth (fresh and dry weight, number and area of leaves) and quality parameters of lettuce leaves (nitrate, mineral content, phenolic compounds) were examined. Carbon sources and microorganisms derived from fish water seem to have neither beneficial nor detrimental effects on plant growth in this study. Except for some differences in the mineral content of the lettuce leaves, all other quality parameters were not significantly different. The use of aquaponic fish water saved 62.8% mineral fertilizer and fully substituted the required water for the nutrient solution in comparison to the control. Additionally, the reduced fertilizer demand using decoupled aquaponics can contribute to reduce greenhouse gas emissions of an annual lettuce production site per ha by 72% due to saving the energy for fertilizer production. This study clearly demonstrates the huge potential of the innovative approach of decoupled aquaponics to foster the transformation of our conventional agriculture towards sustainable production systems saving resources and minimizing emissions.
Summary
The authors grew butterhead lettuce (Salanova, cv. Descartes RZ) in a nutrient film technique (NFT) system in a Berlin greenhouse for a ~7-8 week cycle, comparing a conventional hydroponic control (fresh tap/rain water plus mineral fertilizer) against two decoupled-aquaponic treatments built from a single batch of fish-tank water withdrawn once from an established Nile tilapia RAS: one used directly after fertilizer top-up (APunt, untreated) and one heat-disinfected (>95C for 15 min) before fertilizer top-up (APdis). All three nutrient solutions were adjusted with standard mineral fertilizers to the same target EC (2.2 dS/m) and nutrient recipe, so the aquaponic treatments received substantially less added fertilizer only because the fish water already carried part of the target nutrient load. Yield (fresh weight, leaf number, leaf area, dry matter) did not differ significantly among the three treatments, nor did leaf nitrate or the six phenolic acids/flavonoids measured; only a handful of leaf minerals (Ca, Mg, S, Na) differed significantly, and only between the two aquaponic treatments or between one aquaponic treatment and the control, with no consistent direction. Using fish water cut total mineral fertilizer input by 62.8% and fully replaced the freshwater otherwise needed to mix the nutrient solution, which the authors translate into an estimated 72.6% reduction in fertilizer-manufacturing CO2-equivalent emissions for an annual one-hectare lettuce operation. Dissolved/total organic carbon and colony-forming units were far higher in both aquaponic treatments than the control throughout, and disinfection successfully zeroed CFU at time of treatment, but neither elevated organic carbon nor the surviving/regrown microbial load produced any detectable growth or quality difference relative to the untreated aquaponic treatment or the control — a direct test, and non-replication, of an RAS-derived-growth-promotion hypothesis from an earlier paper (Delaide et al. 2016) that the authors discuss at length.
Experiment data
- Location: 75 m2 compartment of an experimental Venlo-type greenhouse, Berlin, Germany (coordinates as printed in the paper could not be resolved to decimal degrees, see Extraction notes)
- Design: 3 treatments (Control = fresh tap/rain water + fertilizer; APunt = untreated aquaponic fish water + fertilizer; APdis = heat-disinfected aquaponic fish water + fertilizer) x 3 replicate NFT gullies/treatment (9 gullies total, 10 plants/gully, 90 plants total). Treatment-to-gully assignment is shown in a non-repeating layout (Fig 1) but the word “randomized” is never used — see Extraction notes.
- Replicates / n: 3 gullies per treatment (true experimental unit); sub-sample sizes vary by measurement (n=10 heads/gully for fresh weight, n=3 heads/gully for leaf number/area/dry matter, n=4 pooled samples of 3 heads each for nitrogen/carbon/mineral/nitrate/phenolic chemistry)
- Duration: Sown ~2 weeks before transplant; transplanted into NFT gullies 18 March 2016; harvested 10 May 2016; paper states “seven weeks” of growth twice (Table 1 caption, Results 3.1) though the two stated calendar dates span 53 days (see Extraction notes)
- Organisms: Lettuce (Lactuca sativa) (butterhead, var. Salanova, cv. Descartes RZ) / Nile tilapia (Oreochromis niloticus) (fish reared in an already-established, separately-described RAS; not measured as part of this trial)
- Statistics: Univariate ANOVA (SPSS v19.0) after Shapiro-Wilk normality test; Tukey-HSD (homogeneous variance) or Dunnett-T3 (heterogeneous variance) post hoc; Kruskal-Wallis + Dunn-Bonferroni for non-normal data; CFU evaluated by Kruskal-Wallis (initial) then Friedman (weekly series); significance at p<0.05
- Yield (fresh weight, Table 2): Control 325.9 +/- 53.6 g, APunt 323.3 +/- 44.5 g, APdis 332.4 +/- 55.5 g per plant; ns across all three (Tukey-HSD)
- Tissue nitrate content: Control 5896.3 +/- 453.7, APunt 5795.3 +/- 456.3, APdis 5502.1 +/- 367.7 mg/kg FW; ns across all three (Table 6), though all three exceed the EU 4000-5000 mg/kg regulatory limit the authors cite
- Fertilizer and GHG savings (Table 3): 62.8% less total mineral fertilizer used to mix the aquaponic nutrient solutions vs. control (CAN -84.6%, KNO3 -68.4%, MgSO4 -36.4%, KH2PO4 -7.4%); 100% of the freshwater otherwise needed was replaced by fish water; calculated 72.6% reduction in fertilizer-manufacturing CO2-equivalent emissions
Yield and growth
This paper: Fresh weight, leaf number, leaf area, and dry matter were statistically indistinguishable among the hydroponic control and both aquaponic treatments (Table 2, all ns by Tukey-HSD or Dunnett-T3/Kruskal-Wallis as appropriate). The authors frame this as the headline result: at matched target nutrient concentrations, decoupled aquaponic fish water performs identically to conventional hydroponic nutrient solution, with no detectable growth penalty or growth bonus from the fish-water-derived dissolved organic carbon and microorganisms.
Compared with:
- todo Delaide et al. 2016 — lettuce (var. Sucrine), reported a 39% growth increase in complemented/decoupled aquaponics over hydroponics, attributed to RAS-derived dissolved organic matter or growth-promoting rhizobacteria/fungi; this paper explicitly designed its APunt-vs-control comparison to test that hypothesis and found no such effect, discussed at length in Discussion 4.1 (p.14-15)
- todo Suhl et al. 2016 — tomato, decoupled aquaponics vs. hydroponics, also found comparable yields, corroborating this paper’s null result (p.14, ref 16)
- todo Suhl et al. 2018 — decoupled aquaponics with a suction filter to reduce nitrogen loss, cited alongside Suhl 2016 as corroborating (p.15, ref 29)
Fertilizer savings and greenhouse gas emissions
This paper: Because all three nutrient solutions were adjusted to the same target EC/nutrient recipe (Hochmuth 2001), and the aquaponic treatments’ fish water already carried part of that nutrient load, mineral fertilizer input to the aquaponic treatments was 62.8% lower than the control overall, with fertilizer-specific reductions of 84.6% (calcium ammonium nitrate), 68.4% (KNO3), 36.4% (MgSO4), and only 7.4% (KH2PO4, since the fish water was relatively phosphorus-poor and the difference had to be made up with fertilizer). Both aquaponic treatments were mixed entirely from fish water, so freshwater demand for the nutrient solution was reduced by 100%. Using the ProBas emissions database, the authors calculate a 72.6% reduction in fertilizer-manufacturing CO2-equivalent emissions per litre of stock solution, and extrapolate this (using external lettuce yield/water-use figures from Barbosa et al. 2015, not measured in this study) to a hypothetical 1-hectare annual lettuce operation: 16.3 t CO2-eq/year for conventional hydroponic fertilizer manufacturing vs. 4.5 t CO2-eq/year using aquaponic fish water.
Compared with: (no external literature comparison given for the fertilizer/GHG accounting itself; it is original analysis by this paper using a third-party emissions database, ProBas of the German Federal Environment Agency, ref 28)
Nutrient solution chemistry and organic carbon
This paper: At the start of the experiment, total mineral nitrogen (Nmin = NO3-N + NH4-N) was comparable across all three solutions (~193-205 mg/L, Table 4), though NH4-N was significantly higher in the control (14.0 +/- 0.9 mg/L) than either aquaponic treatment (2.2-2.3 mg/L) — attributed to the choice of phosphoric acid (rather than nitric acid) to correct the higher-alkalinity tap/rain water mix down to pH 5.8. Phosphorus was almost 2-fold higher in the control than either aquaponic treatment; calcium, magnesium, sulfur, and several micronutrients (Fe, Zn, Mo, Mn, Cu, B) were higher in one or both aquaponic treatments. Total and dissolved organic carbon (TOC/DOC) were 2.9- to 4.1-fold higher in both aquaponic treatments than the control throughout the 7-week series (Table 5), attributed directly to the fish-water source; the paper explicitly concludes this elevated organic carbon had no detectable effect on plant growth (Discussion 4.1, 4.4).
Compared with:
- todo Waechter-Kristensen et al. 1999 — cited for DOC concentration ranges in the hydroponic control being consistent with prior reports (p.17, ref 44)
- todo Adani et al. 1998 — humic acids added to hydroponics reported beneficial effects on plant growth (e.g. shoot:root ratio), cited as the mechanistic backdrop against which this paper’s null organic-carbon effect is discussed (p.15, ref 30)
Colony-forming units and disinfection
This paper: Heat disinfection (>95C for 15 min, solution held >70C for >45 min total) successfully eliminated all detectable CFU in APdis at the point of treatment (Fig 3), versus 5097 CFU/mL in the untreated fish water and 177 CFU/mL in the control at the same reference point. Within 5 days of planting, CFU in both aquaponic treatments rebounded to ~2.1-2.3 x 10^4 units/mL, indicating rapid recolonization once the disinfected solution was exposed to the greenhouse/root environment. By the end of the 7-week trial, CFU in all three treatments had converged to a similar low range (502-949 units/mL), and no significant differences among treatments were detected by the Friedman test across the weekly series. The authors interpret this as showing disinfection provides no durable advantage once planting begins, consistent with the equally null growth/quality results between APunt and APdis.
Compared with:
- todo Pantanella et al. 2015 — UV sterilization effects on total coliforms and lettuce production in aquaponics; cited as reporting a similarly converged end-of-trial CFU range (p.17, ref 51)
- todo Leonard et al. 2000 — heterotrophic bacteria populations in a recirculating sea bass RAS, cited for context on typical RAS CFU loads, described as higher than observed here (p.17, ref 48)
Leaf mineral composition and phenolic/nitrate quality
This paper: Of the 9 elements measured in dried leaf tissue (Table 6), only calcium, magnesium, sulfur, and sodium differed significantly among treatments, and in every case the significant contrast was between the two aquaponic treatments themselves (or one aquaponic treatment vs. control), not a consistent aquaponic-vs-hydroponic pattern; carbon, nitrogen, phosphorus, potassium, and iron did not differ at all. None of the six phenolic acids/flavonoids measured (Table 7) differed significantly among treatments, and leaf nitrate (Table 6) likewise showed no significant difference despite all three treatments exceeding the EU nitrate limit the authors cite for summer-harvest lettuce. The authors attribute the nitrate parity to comparable total mineral nitrogen across solutions rather than to any aquaponic-specific effect, and discuss (without testing) several possible reasons the elevated organic carbon in the aquaponic treatments might not have translated into a detectable nitrate-lowering effect via humic-acid pathways reported elsewhere.
Compared with:
- todo DuPont et al. 2000, todo Llorach et al. 2008, todo Ribas-Agustí et al. 2011 — cited as establishing that this paper’s phenolic acid/flavonoid profile and concentration ranges for green lettuce match prior literature (p.18, refs 57-59)
- todo Becker et al. 2015 — nitrogen-limited lettuce accumulates more flavonoid glycosides/caffeic acid derivatives, cited as the mechanistic backdrop for why N status (not measured to differ here) could affect phenolics (p.18, ref 60)
- todo Demsar and Osvald 2003 — found a 13% nitrate reduction in lettuce at a NO3:NH4 ratio of 6.6:1 (aeroponic system); this paper’s control NO3:NH4 ratio (13.7:1) was too dilute in ammonium to reproduce that effect, offered as one explanation for the nitrate parity here (p.18, ref 69)
- todo Economakis and Koleilat 1996 — longer nitrate feeding duration increases lettuce nitrate accumulation; cited to suggest this paper’s ~7-8 week cycle and near-300g harvest weight may already be past the point where a shorter cycle/earlier harvest would have reduced nitrate content (p.18, ref 72)
Linked claims
- Decoupled aquaponics matches hydroponic growth and nutrient status while using less fertilizer
- Aquaponic lettuce tissue nitrate is not higher than hydroponic lettuce
- Recycling fish water in decoupled aquaponics substantially reduces fertilizer-related greenhouse gas emissions
- Disinfecting aquaponic fish water does not change plant growth or quality outcomes
- Dissolved organic carbon from RAS-derived water does not reliably improve plant growth in aquaponics
Citations to chase
- todo Delaide B, Goddek S, Gott J, Soyeurt H, Jijakli MH (2016) — Lettuce (Lactuca sativa L. var. Sucrine) growth performance in complemented aquaponic solution outperforms hydroponics, Water 8:467 — the 39% growth-increase finding this paper directly tests and does not replicate
- todo Suhl J, Dannehl D, Kloas W, Baganz D, Jobs S, Scheibe G, Schmidt U (2016) — Advanced aquaponics: Evaluation of intensive tomato production in aquaponics vs. conventional hydroponics, Agricultural Water Management 178:335-344
- todo Suhl J, Dannehl D, Baganz D, Schmidt U, Kloas W (2018) — An innovative suction filter device reduces nitrogen loss in double recirculating aquaponic systems, Aquacultural Engineering 82:63-72
- todo Pantanella E, Cardarelli M, Di Mattia E, Colla G (2015) — Aquaponics and food safety: Effects of UV sterilization on total coliforms and lettuce production, Acta Horticulturae 1062:71-76
- todo Waechter-Kristensen B, Caspersen S, Adalsteinsson S, Sundin P, Jensen P (1999) — Organic compounds and micro-organisms in closed, hydroponic culture: Occurrence and effects on plant growth and mineral nutrition, Acta Horticulturae 481:197-204
- todo Adani F, Genevini P, Zaccheo P, Zocchi G (1998) — The effect of commercial humic acid on tomato plant growth and mineral nutrition, Journal of Plant Nutrition 21(3):561-575
- todo Becker C, Urlić B, Špika MJ, Kläring H-P, Krumbein A, Baldermann S, et al. (2015) — Nitrogen limited red and green leaf lettuce accumulate flavonoid glycosides, caffeic acid derivatives, and sucrose while losing chlorophylls, β-carotene and xanthophylls, PLoS ONE 10(11):e0142867
- todo Demsar J, Osvald J (2003) — Influence of NO3-:NH4+ ratio on growth and nitrate accumulation in lettuce (Lactuca sativa var. capitata L.) in an aeroponic system, Agrochimica 47(3-4):112-121
- todo Economakis C, Koleilat R (1996) — Effect of nitrogen concentration on growth, water, and nutrient uptake of lettuce plants in solution culture, II International Symposium on Irrigation of Horticultural Crops 449
- todo Monsees H, Kloas W, Wuertz S (2017) — Decoupled systems on trial: Eliminating bottlenecks to improve aquaponic processes, PLoS ONE 12(9):e0183056 — describes the DRAPS/RAS unit that supplied the fish water used here in detail (companion paper, already in this vault’s pdfs/ as a separate note)
Extraction notes
Trial structure: this paper has two distinct aquaponic arms (APunt and APdis) sharing one hydroponic control (Control). Recorded as two trials.csv rows, monseesLettuceLactucaSativa2019-T1 (APunt vs Control) and -T2 (APdis vs Control), each repeating the same Control values in the HYD-labelled columns, per SCHEMA.md convention. Unlike a typical coupled-vs-decoupled comparison, the distinguishing feature between T1 and T2 is not the aquaponic loop topology (both are decoupled/DRAPS-derived, single-batch, non-recirculating-to-RAS) but whether the fish water was heat-disinfected before fertilizer supplementation; both aquaponic treatments received the same fertilizer top-up to the same target recipe as the control, i.e. this is functionally a full-remineralization (“DCAP-style”) decoupled aquaponic design in both arms, not a minimal-supplementation design.
Type classification judgment call: Recorded as experiment. Three defined treatments, true replication (3 independent physical NFT gullies per treatment), and formal statistics (ANOVA + Tukey-HSD/Dunnett-T3/Kruskal-Wallis, p<0.05) are all present. As in pantanellaAquaponicsHydroponicsProduction2012, the word “randomized” is never used for how treatments were assigned to the nine gullies; Fig 1’s layout (C-APunt-APdis / APunt-C-APdis / APdis-APunt-C across three physical positions) is consistent with randomization but not explicitly labelled as such. Given the complete, true-replicate, formally-tested design, experiment was judged the better fit over quasi-experiment, consistent with the same judgment call made for the Pantanella paper already in this vault.
⚠️WARN-BLOCK — greenhouse coordinates, Methods 2.1, p.3. The paper gives coordinates as “(52°46´82.806´´N, 13°29´88.909´´E)”. Read as literal degrees-minutes-seconds, both minutes fields (46, 29) are valid (<60), but both seconds fields (82.806, 88.909) exceed 59, which is impossible in DMS notation — and this happens simultaneously in both latitude and longitude, in the same field position. Unlike levizouCircularTriTrophicSystem2025’s coordinate issue (where only one of two coordinates was invalid and could be confidently reread as decimal-degrees-mistakenly-DMS-formatted), no single reinterpretation here reproduces both printed digit strings without assuming an unstated transcription error (e.g. a digit transposition such as 82.806->28.806 or 88.909->38.909, which would each yield a plausible decimal answer consistent with “Berlin, Germany” but cannot be confirmed from the text). Recorded Lat and Long as UNCLEAR. Affects: Lat, Long cells only (both trials). Region (Europe) and Country (Germany) are unaffected, being taken directly from the stated place name rather than the coordinate string. Added to REVIEW.md by the batch merge step.
⚠️WARN-MINOR — cultivation duration, Table 1 caption (p.4) and Results 3.1 (p.8) vs. Methods 2.1 dates (p.3). The paper twice states the trial’s growth duration in prose/caption as “seven weeks” (“the experimental period of 7 weeks”, Table 1 caption; “after seven weeks of growth”, Results 3.1) but separately states the plants were “transplanted into high gullies on 18th March 2016” and “the lettuce heads were harvested” on “10th May 2016” (Methods 2.1, 2.3) — a calendar span of 53 days (7.57 weeks), 4 days more than the stated “seven weeks” (49 days). Recorded Days Plant after transplant = 49 (UNIT CONVERSION ONLY: 7 weeks x 7 days, the paper’s own repeated, directly-stated figure) rather than the date-derived 53, since the latter would require combining two separately stated dates rather than reading a single stated duration. Does not affect any other cell; Table 2’s growth data is reported simply “at harvest,” not tied to a specific day-count.
⚠️WARN-MINOR — Table 7, quercetin-3-O-(6”-malonyl)glucoside (Q3MG), APunt value, p.14. Printed as “0.33 ± 0.07^a” with a superscript ‘a’ letter, while Control (0.25 ± 0.03) and APdis (0.36 ± 0.05) carry no superscript at all, and the table caption states “No significant differences (p<0.05) were analysed using univariate ANOVA and Tukey-HSD test (caffeoyltartaric acid, caffeoylquinic acid, and caffeoylmalic acid) or using Kruskal-Wallis test” — implying ns throughout, with no stated basis for a lone, unmatched significance letter on a single cell. Most likely a typesetting artifact (e.g. a stray footnote marker) rather than a real significant pairwise contrast, since a single letter with no partner group is not interpretable as a Tukey/Dunn grouping. Recorded in plant.csv with Significance “ns” per the caption’s own overall statement, and this anomaly noted in that row’s Notes field. Does not affect any trials.csv cell (Q3MG has no dedicated column).
No numeric contradiction, documented for completeness — water-chemistry values are single-timepoint, not repeated trial means. Table 4’s full nutrient/mineral panel (NO3-N, NH4-N, Nmin, P, K, Ca, Mg, S, Na, Fe, Mn, Mo, Zn, Cu, B) is explicitly captioned “at the beginning of the experiment” — i.e. one measurement of the freshly mixed solution before planting, never repeated later in the trial. This is not a time-series-with-no-summary situation (SCHEMA.md’s water-quality rule) so much as a single-point measurement with no repeated series at all for this panel; recorded as given (TAN/NH4-N and NO3-N cells) with this basis stated explicitly in each trial’s Experimental Remarks, since presenting it as an ordinary “trial mean” would overstate its temporal coverage. EC and pH, by contrast, genuinely were measured weekly (Fig 2) but only presented as a chart with narrative range statements in the text, not a table of numbers — recorded as explicit text-stated ranges (“2.2 to ~3.0 dS/m”, “5.5-5.8”) per the “range only, no trial mean reported” convention, since these ranges are given in running text rather than read off the figure.
NOT DERIVED, left NR (grouped by field):
- Fish: FCR, SGR, Fish size initial/final, Fish weight gain, Fish biomass created (kg), Fish survival rate, Total Feed (kg), Fish trial duration (days), Feed routine (feeding frequency/day), feed N/P/K composition beyond crude protein — the RAS supplying fish water is an already-established, continuously-operating system (fully characterized in the companion Monsees, Kloas & Würtz 2017 paper) rather than a fish cohort stocked and grown specifically for this lettuce trial; this paper gives only the RAS’s standing stocking density (90.6 kg/m3), total biomass (462.1 kg), and feed identity/ration (0.65% bodyweight/day, 37% crude protein) as background context (Methods 2.1.2), with no trial-specific fish growth, feed-total, or survival data to extract. Initial Stock density (90.6 kg/m3) was recorded since it is explicitly stated as a density, but flagged here as describing the parallel RAS operation rather than a batch tied to this trial’s 7-8 week duration.
- Water: Water recycle (L/min, not stated), Daily Water exchange rate (the 750 L nutrient-solution batch per treatment is recirculated within its own tank continuously via pump, but no top-up/exchange schedule during the trial is described), Water temperature (no nutrient-solution water temperature given; only the RAS fish-tank temperature, 25C, a different compartment, and greenhouse air temperature, Table 1, which is a different quantity), Dissolved Oxygen (only the RAS fish-tank DO is given, “always above 5.5 mg/L,” not the nutrient-solution/plant-bed side), Water classification (no categorising term distinct from “Water type” used by the paper; SCHEMA.md notes this distinction is itself undefined, see Still Open item 2), FUE AP/HYD and WUE (the paper reports fertilizer savings as a % reduction in total mineral fertilizer mass, Table 3, and a 100% freshwater substitution, not as a yield-per-input ratio matching these columns; computing one would require dividing stated yield by stated fertilizer mass, which is derivation — see the fertilizer/GHG figures preserved in Experimental Remarks below instead), Plants/m2 (only gully dimensions, 6 m x 0.35 m, and a 10-plants-per-gully count are given, not a stated density; computing one would be derivation).
- Plant: Plant height, SPAD — neither measured; Plants/m2 (see above); Plant Category beyond “butterhead lettuce” — the paper’s own consistent term throughout (abstract, Methods 2.1) was recorded as the category since it is the paper’s own descriptive wording, per SCHEMA.md’s category-fields rule.
NO COLUMN items (preserved here and in each trial row’s Experimental Remarks rather than discarded):
- Full nutrient solution mineral panel at start of experiment (Table 4, p.10): P, K, Ca, Mg, S, Na, Fe, Mn, Mo, Zn, Cu, B in mg/L for Control/APunt/APdis, with significance letters — water chemistry beyond the schema’s NH4-N/NO3-N columns has no dedicated home in trials.csv. This is a genuinely valuable, fully-quantified panel (the paper’s own headline nutrient-management narrative in Discussion 4.3 is built entirely from it) that is being routed out of both trials.csv and plant.csv per SCHEMA.md’s water-chemistry exclusion rule; flagging this explicitly per SCHEMA.md’s instruction that a too-valuable-to-discard water panel should be called out rather than silently dropped.
- Total and dissolved organic carbon (Table 5, p.12): TOC and DOC in mg/L at weeks 0, 4, and 7 for all three treatments, 2.9- to 4.1-fold higher in both aquaponic treatments throughout — again water chemistry, no schema column, and likewise flagged as too valuable to discard silently.
- Colony-forming units (Fig 3, p.13): weekly CFU/mL series for all three treatments plus an immediate post-disinfection (“ad”) measurement — this is water/system microbiology (of the nutrient solution), not plant tissue microbiology, so it does not fit plant_measurements.csv’s microbiology category either (that category is scoped to plant analytes per SCHEMA.md). No numeric values beyond the ones narrated in text (177, 5097, ~2.1-2.3x10^4, 502-949 CFU/mL at specific points) were extracted, since most of the series is chart-only (never-read-a-figure rule).
- Fertilizer savings and GHG accounting (Table 3, p.9; Discussion 4.2, p.15-16): per-fertilizer reduction percentages (CAN -84.6%, KNO3 -68.4%, MgSO4 -36.4%, KH2PO4 -7.4%), total fertilizer supply (Control 207, APunt/APdis 77 g/L stock solution), per-fertilizer CO2-eq figures, and the extrapolated 1-hectare annual GHG estimate (16.3 vs 4.5 t CO2-eq/year) — this is the paper’s headline sustainability finding but has no dedicated trials.csv column (FUE/WUE do not match its unit basis, see above); fully preserved here rather than compressed into a ratio.
- Leaf area (Table 2, p.8): Control 66.9 +/- 5.7, APunt 66.7 +/- 6.8, APdis 67.5 +/- 7.2 dm2/plant, ns — a real growth measurement but with no dedicated trials.csv column (only Plant height and Leaf count exist for morphology) and not a tissue analyte, so also excluded from plant_measurements.csv per the same reasoning used for physiological measurements in mourantianBasilFunctionalGrowth2023.
- Literature comparison figures (secondary, not this paper’s data): Delaide et al. 2016’s 39% lettuce growth increase in decoupled aquaponics; Barbosa et al. 2015’s cited lettuce yield (41 kg/m2/year) and water-use (20 L/kg/year) figures, used only to build this paper’s own hypothetical 1-ha extrapolation, not measured here.
plant_measurements.csv scope decision: Table 6’s leaf mineral panel (C, N, P, K, Ca, Mg, S, Na, Fe, %DM) and Table 7’s phenolic acid/flavonoid panel (mg/100g FW) were extracted to plant.csv (Category mineral and biochemistry respectively), duplicating the shared Control row across both T1 and T2 per the convention established in mourantianBasilFunctionalGrowth2023/pantanellaAquaponicsHydroponicsProduction2012. Table 6’s NO3 (leaf nitrate) row was deliberately NOT duplicated into plant.csv, since trials.csv already has dedicated Tissue nitrate AP/HYD columns serving exactly that purpose, following the same redundancy judgment made in pantanellaAquaponicsHydroponicsProduction2012. Carbon (C, %DM) was included under the mineral category alongside nitrogen even though SCHEMA.md’s category description does not name carbon explicitly (it names “leaf tissue N, P, K, Zn, Fe and other elemental content”) — judged the closest fit as an elemental-composition measurement, flagged here as a judgment call per CLAUDE.md’s instruction to note such calls rather than force a fit silently.
Tags judgment call: Tagged Meta/Fish/Tilapia (Nile tilapia, Oreochromis niloticus) even though no fish were reared, measured, or fed as part of this specific trial — the fish water used as the aquaponic treatments’ nutrient base came from a real, named, actively-operating RAS stocking this species, which is a materially different (and stronger) basis for the tag than a review merely mentioning tilapia in passing. Judgment call noted per CLAUDE.md’s instruction. Tagged Meta/Plant/Lettuce (reusing the vault’s existing generic tag rather than creating a butterhead- or Salanova-specific sub-facet, consistent with pantanellaAquaponicsHydroponicsProduction2012’s treatment of romaine lettuce under the same generic tag).
Quality score: caution — 1 WARN-BLOCK (coordinates) present, 0 WARN-MATERIAL, which places this at the caution threshold per SCHEMA.md’s “1 BLOCK, or 3-4 MATERIAL” rule. The BLOCK is confined entirely to Lat/Long and does not touch any biological or yield outcome in either trial row.
New wikilink targets introduced: H. Monsees, J. Suhl, M. Paul, W. Kloas, D. Dannehl, S. Würtz (no existing author notes found in the vault for any of these). Decoupled aquaponics matches hydroponic growth and nutrient status while using less fertilizer and Aquaponic lettuce tissue nitrate is not higher than hydroponic lettuce reuse claim wikilinks already established in mourantianBasilFunctionalGrowth2023 and pantanellaAquaponicsHydroponicsProduction2012 respectively, both directly corroborated by this paper. Recycling fish water in decoupled aquaponics substantially reduces fertilizer-related greenhouse gas emissions and Disinfecting aquaponic fish water does not change plant growth or quality outcomes are new claim targets specific to this paper’s fertilizer/GHG and disinfection findings. Dissolved organic carbon from RAS-derived water does not reliably improve plant growth in aquaponics is a new claim target for this paper’s direct non-replication of the Delaide et al. 2016 growth-promotion hypothesis. Reused existing canonical organism forms Nile tilapia (Oreochromis niloticus) and Lettuce (Lactuca sativa), and existing concept form Tissue nitrate content (the vault also has a second, less-used variant Tissue nitrate (NO3), noted here per CLAUDE.md rather than silently picking one). No existing Decoupled aquaponics or Nutrient film technique (NFT) concept notes were found in the vault; both are referenced in prose above as candidate new wikilink targets but were left as plain text rather than wikilinked, since neither is a load-bearing claim/concept node in this note the way the Linked claims are — listed here per CLAUDE.md’s instruction to record candidates rather than guess.
PDF quality: Clean text layer throughout (23 pages, standard single/double-column PLOS ONE typesetting), fully extractable, no OCR issues. Tables 1-7 and Figures 1-3 all have readable captions; only Figures 2 and 3 contain data (EC/pH weekly series; CFU weekly series) that is chart-only with no accompanying table, handled per the never-read-a-figure rule above.
Source: Monsees et al. - 2019 - Lettuce (Lactuca sativa, variety Salanova) product.pdf
Data Tables
Structured data extracted from this paper into the vault's
trials.csv/plant_measurements.csvdatasets. Fields the paper didn't report are omitted. Download the full datasets (measurements).
Trial Parameters
monseesLettuceLactucaSativa2019-T1
Fish
| Field | Value |
|---|---|
| Fish | Nile tilapia (Oreochromis niloticus) |
| Initial Stock density | 90.6 |
| Protein | 37 |
| % of body weight | 0.65 |
| Feed regime | Floating pellets (ALLER SANA FLOAT, Emsland Aller Aqua GmbH) containing 37% crude protein, 10% crude fat, 38.5% nitrogen-free extractive, 6% crude ash, 3.5% crude fibre, 17.9 MJ/kg digestible energy; fed at 0.65% of bodyweight per day (Methods 2.1.2, p.5) |
Water
| Field | Value |
|---|---|
| Water volume in the system | 250 (per tank; 3 tanks/treatment, 750 total per treatment, Methods 2.1, p.3-5) |
| Water type | Fish waste water (untreated), RAS-derived, supplemented with mineral fertilizer (APunt, Fig 1 caption, p.4) |
| Aq pH | 5.5-5.8 (range across the trial; no single trial mean stated, Results 3.4 p.10, Fig 2) |
| pHOptimal | 5.8 |
| EC | 2.2 (day 0, target for all treatments) to ~3.0 (week 7, approx.; AP significantly higher than control weeks 1-5, no longer significant weeks 6-7) - range only, no trial mean reported (Results 3.4 p.10, Fig 2) |
| TAN / NH4-N | 2.2 +/- 0.01 (value at start of experiment only, Table 4 p.10) |
| NO3-N | 197.1 +/- 10.5 (value at start of experiment only, Table 4 p.10) |
Plant
| Field | Value |
|---|---|
| Plant | Lettuce (Lactuca sativa L., var. Salanova, cv. Descartes RZ) |
| Details | Green open butterhead lettuce; sown in autoclaved cultivation soil, grown 2 weeks, transplanted into rockwool cubes (7.5x7.5 cm) and into NFT gullies on 18 March 2016; harvested 10 May 2016; 10 plants/gully at 50 cm spacing, 3 gullies/treatment (Methods 2.1, p.3) |
| Plant Category | Butterhead lettuce (p.3, abstract) |
| Days Plant after transplant | 49 |
| Leaf count | 214.1 +/- 42.3 |
| Plant fresh weight | 323.3 +/- 44.5 |
| Plant dry matter | 4.7 +/- 0.3 |
| Tissue nitrate AP | 5795.3 +/- 456.3 |
| Tissue nitrate HYD | 5896.3 +/- 453.7 |
System & Setup
| Field | Value |
|---|---|
| System type | Nutrient film technique (NFT) |
| Media Details | Rockwool cubes (7.5 x 7.5 cm, Cutilene) for transplant fixation; NFT gullies 6 m x 0.35 m x 0.05 m, 1% incline; black/white plastic film between plants to prevent algae; EHEIM universal 600 pumps, continuous 24 h/d flow (Methods 2.1, p.3) |
| Biological system already in use | Y (Fish water sourced from an already-operating DRAPS/RAS unit at IGB Berlin, described in detail by Monsees, Kloas & Würtz 2017 (ref. 14) (Methods 2.1.2, p.5)) |
| Iron supplemented | Y (Fe included in standard trace-element fertilizer mix (YaraTera Tenso Cocktail) targeting 2.5 mg Fe/L for all treatments (Methods 2.1.1, p.5); measured solution Fe below target and significantly lower in control (0.001+/-0.00) than APunt (0.01+/-0.00) and APdis (0.02+/-0.00) mg/L, Table 4) |
| Remineralization | Y (All nutrient solutions, including both aquaponic treatments, supplemented with standard mineral fertilizers (calcium ammonium nitrate CAN, KNO3, MgSO4, KH2PO4, trace element solution) to reach target EC 2.2 dS/m and Hochmuth (2001) target nutrient concentrations (Methods 2.1.1, p.5)) |
| pH Buffers | Y (pH adjusted to 5.8 with phosphoric acid at solution preparation; weekly monitoring, corrected with NaOH as needed (Methods 2.1.1 p.5, 2.2.2 p.6)) |
| Climate control | Y (Ventilation opened above 17C; floor-level heating target 10C night / 14C day (Methods 2.1, p.3); weekly mean greenhouse climate in Table 1) |
| Nutrient supplemented | Y (Both aquaponic treatments (fish water) supplemented with same mineral fertilizers as control to reach identical target EC/nutrient recipe (Hochmuth 2001); total fertilizer supply reduced by 62.8% overall vs control due to nutrients already present in fish water (Table 3, p.9)) |
| Equipment | EHEIM universal 600 pumps; HI9811-5 Hanna pH/EC meter; LI-3100 Area Meter (LICOR) for leaf area; RQflex 10 plus reflectometer (Merck) for nitrate; ICP-OES iCAP 6300Duo (Thermo) for macro/micronutrients; CFA San++ (Skalar) for NH4-N/NO3-N; TOC-L CPN analyzer (Shimadzu) for TOC/DOC; vario MAX elemental analyser (Elementar) for N/C; HPLC Ultimate 3000 (Thermo Fisher) for phenolics; 150 L stainless steel disinfection pot with butane gas burner (3.6 kW) + 2x2kW heating rods + plate heat exchanger |
| Control Parameters | 3 treatments (Control, APunt, APdis) x 3 replicate gullies/treatment (9 gullies, 90 plants total); EC target 2.2 dS/m at preparation (all treatments); pH target 5.8 at preparation; disinfection >95C for 15 min then solution maintained >70C for >45 min (APdis only) |
| Combination | Nile tilapia (Oreochromis niloticus) and butterhead lettuce (Lactuca sativa var. Salanova) in a decoupled, non-recirculating-to-RAS aquaponic system; fish water withdrawn once as a static batch and fertilizer-supplemented, not returned to the RAS; this row = untreated aquaponic fish water (APunt) vs Control |
Site
| Field | Value |
|---|---|
| Region | Europe |
| Country | Germany |
| Average room Temperature | 13.36-21.80 (day) / 9.26-16.28 (night) (weekly means across the trial, Table 1 p.4; no single trial mean stated) - range only, no trial mean reported |
Results & Statistics
| Field | Value |
|---|---|
| Measured Unit | g/plant (fresh weight, primary yield metric); % (dry matter); mg/kg FW (tissue nitrate); mg/100g FW (phenolics); mg/L (nutrient solution chemistry) |
| Statistic Details | Univariate ANOVA (SPSS v19.0) after Shapiro-Wilk normality test; Tukey-HSD (homogeneous variance) or Dunnett-T3 (heterogeneous variance) post hoc; Kruskal-Wallis + Dunn-Bonferroni for non-normal data; CFU: Kruskal-Wallis (initial) + Friedman (weekly); significance p<0.05 (Methods 2.6, p.8) |
| Statistically analysed | Y |
| Replicates (n) | 3 |
| AP | 323.3 |
| HYD | 325.9 |
Experimental Remarks: TRIAL DEFINITION: T1 = decoupled aquaponic treatment using untreated fish water (withdrawn once as a static batch from an established Nile tilapia DRAPS/RAS at IGB Berlin, then supplemented with mineral fertilizer to the same target EC/nutrient recipe as the control; not returned to the RAS). Paired control = Control (fresh tap/rain water 50:50 v/v + mineral fertilizer), recorded in the HYD-labelled cells. Design: 3 treatments x 3 replicate NFT gullies/treatment (9 gullies, 90 plants total), 10 plants/gully; ‘randomized’ never explicitly stated for treatment-to-gully assignment (Fig 1 layout is non-repeating across the three physical positions but not labelled as random, see Extraction notes). | Both aquaponic arms in this paper (APunt, APdis) are full-remineralization decoupled aquaponics: fish water was fertilizer-supplemented to hit the SAME target nutrient recipe as the hydroponic control, not a minimal/no-supplementation design. | WARN-BLOCK greenhouse coordinates, Methods 2.1 p.3: printed as ‘(52°46´82.806´´N, 13°29´88.909´´E)’. Both minutes fields (46, 29) are valid DMS (<60) but both seconds fields (82.806, 88.909) exceed 59, impossible in DMS notation, in both latitude and longitude simultaneously. Unlike levizouCircularTriTrophicSystem2025 (where only one coordinate was invalid and could be confidently reread as decimal-degrees-mistakenly-DMS-formatted), no single reinterpretation here reproduces both printed digit strings without assuming an unstated transcription error (e.g. digit transposition 82.806->28.806 or 88.909->38.909) that cannot be confirmed from the text. Recorded Lat and Long as UNCLEAR. Affects: Lat, Long cells only; Region (Europe) and Country (Germany) taken directly from the stated place name ‘Berlin, Germany’, unaffected. Added to REVIEW.md by the batch merge step. | WARN-MINOR cultivation duration, Table 1 caption (p.4) and Results 3.1 (p.8) vs. Methods 2.1 dates (p.3): paper twice states duration as ‘seven weeks’ (49 days by conversion) but transplant (18 March 2016) to harvest (10 May 2016) spans 53 days (7.57 weeks) by direct date subtraction, a 4-day discrepancy. Recorded Days Plant after transplant = 49 (UNIT CONVERSION ONLY: 7 weeks x 7 days, the paper’s own repeated, directly-stated figure) rather than the date-derived 53. Does not affect any other cell; Table 2’s growth data is reported simply ‘at harvest’. | Table 4’s full nutrient/mineral panel (incl. TAN/NH4-N and NO3-N) is captioned ‘at the beginning of the experiment’ — a single measurement of the freshly mixed solution before planting, never repeated later in the trial; recorded as given, with this single-timepoint basis stated explicitly rather than presented as an ordinary repeated-measures trial mean. | NOT DERIVED, left NR: FCR, SGR, Fish size initial/final, Fish weight gain, Fish biomass created (kg), Fish survival rate, Total Feed (kg), Fish trial duration (days), Feed routine (feeding frequency/day not stated), feed N/P/K composition beyond crude protein — the RAS supplying fish water is an already-established, continuously-operating system (companion Monsees, Kloas & Würtz 2017 paper) rather than a fish cohort grown for this lettuce trial; only standing stocking density (90.6 kg/m3), total biomass (462.1 kg) and feed identity/ration are given as background (Methods 2.1.2). Initial Stock density (90.6 kg/m3) recorded since explicitly stated as a density, but describes the parallel RAS operation, not a batch tied to this trial’s 7-8 week duration. | NOT DERIVED, left NR: Water recycle (L/min, not stated); Daily Water exchange rate (no top-up/exchange schedule during the trial described; static 750 L batch recirculated within its own tank); Water temperature (only RAS fish-tank temperature, 25C, a different compartment, and greenhouse air temperature, Table 1, a different quantity, are given); Dissolved Oxygen (only RAS fish-tank DO given, ‘always above 5.5 mg/L’, not the nutrient-solution/plant-bed side); Water classification (no term distinct from Water type used by the paper; SCHEMA.md notes this distinction is itself undefined); FUE AP/HYD and WUE (paper reports fertilizer savings as a % mass reduction, Table 3, and 100% freshwater substitution, not as a yield-per-input ratio; computing one would be derivation — see NO COLUMN fertilizer/GHG figures below); Plants/m2 (only gully dimensions 6 m x 0.35 m and a 10-plants/gully count given, not a stated density; computing one would be derivation). | NO COLUMN: Full nutrient solution mineral panel at start of experiment (Table 4, p.10): P, K, Ca, Mg, S, Na, Fe, Mn, Mo, Zn, Cu, B in mg/L for Control/APunt/APdis with significance letters — water chemistry beyond NH4-N/NO3-N has no dedicated trials.csv column; flagged as too valuable to discard silently per SCHEMA.md. Total and dissolved organic carbon (Table 5, p.12): TOC/DOC in mg/L at weeks 0, 4, 7 for all three treatments, 2.9- to 4.1-fold higher in both aquaponic treatments throughout. Colony-forming units (Fig 3, p.13): weekly CFU/mL series (177 control / 5097 untreated fish water at baseline; ~2.1-2.3x10^4 both aquaponic treatments 5 days after planting; 502-949 CFU/mL all treatments by week 7); no CFU detected in APdis immediately after disinfection. This is water/system microbiology, not plant tissue microbiology, so excluded from plant.csv too. Fertilizer savings and GHG accounting (Table 3, p.9; Discussion 4.2): per-fertilizer reduction (CAN -84.6%, KNO3 -68.4%, MgSO4 -36.4%, KH2PO4 -7.4%), total fertilizer supply (Control 207, APunt/APdis 77 g/L stock solution), per-fertilizer CO2-eq figures, and extrapolated 1-ha annual GHG estimate (16.3 vs 4.5 t CO2-eq/year) — this paper’s headline sustainability finding, no dedicated column, fully preserved here. Leaf area (Table 2, p.8): Control 66.9+/-5.7, APunt 66.7+/-6.8, APdis 67.5+/-7.2 dm2/plant, ns — no dedicated morphology column beyond height/leaf count, and not a tissue analyte so excluded from plant.csv per the same reasoning used for physiological measurements in mourantianBasilFunctionalGrowth2023. Literature comparison figures (secondary, not this paper’s data): Delaide et al. 2016’s 39% lettuce growth increase in decoupled aquaponics; Barbosa et al. 2015’s cited lettuce yield (41 kg/m2/year) and water-use (20 L/kg/year) figures used only to build this paper’s own hypothetical 1-ha extrapolation.
monseesLettuceLactucaSativa2019-T2
Fish
| Field | Value |
|---|---|
| Fish | Nile tilapia (Oreochromis niloticus) |
| Initial Stock density | 90.6 |
| Protein | 37 |
| % of body weight | 0.65 |
| Feed regime | Floating pellets (ALLER SANA FLOAT, Emsland Aller Aqua GmbH) containing 37% crude protein, 10% crude fat, 38.5% nitrogen-free extractive, 6% crude ash, 3.5% crude fibre, 17.9 MJ/kg digestible energy; fed at 0.65% of bodyweight per day (Methods 2.1.2, p.5) |
Water
| Field | Value |
|---|---|
| Water volume in the system | 250 (per tank; 3 tanks/treatment, 750 total per treatment, Methods 2.1, p.3-5) |
| Water type | Fish waste water (disinfected, heat-treated >95C/15min), RAS-derived, supplemented with mineral fertilizer (APdis, Fig 1 caption, p.4) |
| Aq pH | 5.5-5.8 (range across the trial; no single trial mean stated, Results 3.4 p.10, Fig 2) |
| pHOptimal | 5.8 |
| EC | 2.2 (day 0, target for all treatments) to ~3.0 (week 7, approx.; AP significantly higher than control weeks 1-5, no longer significant weeks 6-7) - range only, no trial mean reported (Results 3.4 p.10, Fig 2) |
| TAN / NH4-N | 2.3 +/- 0.01 (value at start of experiment only, Table 4 p.10) |
| NO3-N | 191.2 +/- 22.9 (value at start of experiment only, Table 4 p.10) |
Plant
| Field | Value |
|---|---|
| Plant | Lettuce (Lactuca sativa L., var. Salanova, cv. Descartes RZ) |
| Details | Green open butterhead lettuce; sown in autoclaved cultivation soil, grown 2 weeks, transplanted into rockwool cubes (7.5x7.5 cm) and into NFT gullies on 18 March 2016; harvested 10 May 2016; 10 plants/gully at 50 cm spacing, 3 gullies/treatment (Methods 2.1, p.3) |
| Plant Category | Butterhead lettuce (p.3, abstract) |
| Days Plant after transplant | 49 |
| Leaf count | 217.5 +/- 29.4 |
| Plant fresh weight | 332.4 +/- 55.5 |
| Plant dry matter | 4.7 +/- 0.2 |
| Tissue nitrate AP | 5502.1 +/- 367.7 |
| Tissue nitrate HYD | 5896.3 +/- 453.7 |
System & Setup
| Field | Value |
|---|---|
| System type | Nutrient film technique (NFT) |
| Media Details | Rockwool cubes (7.5 x 7.5 cm, Cutilene) for transplant fixation; NFT gullies 6 m x 0.35 m x 0.05 m, 1% incline; black/white plastic film between plants to prevent algae; EHEIM universal 600 pumps, continuous 24 h/d flow (Methods 2.1, p.3) |
| Biological system already in use | Y (Fish water sourced from an already-operating DRAPS/RAS unit at IGB Berlin, described in detail by Monsees, Kloas & Würtz 2017 (ref. 14) (Methods 2.1.2, p.5)) |
| Iron supplemented | Y (Fe included in standard trace-element fertilizer mix (YaraTera Tenso Cocktail) targeting 2.5 mg Fe/L for all treatments (Methods 2.1.1, p.5); measured solution Fe below target and significantly lower in control (0.001+/-0.00) than APunt (0.01+/-0.00) and APdis (0.02+/-0.00) mg/L, Table 4) |
| Remineralization | Y (All nutrient solutions, including both aquaponic treatments, supplemented with standard mineral fertilizers (calcium ammonium nitrate CAN, KNO3, MgSO4, KH2PO4, trace element solution) to reach target EC 2.2 dS/m and Hochmuth (2001) target nutrient concentrations (Methods 2.1.1, p.5)) |
| pH Buffers | Y (pH adjusted to 5.8 with phosphoric acid at solution preparation; weekly monitoring, corrected with NaOH as needed (Methods 2.1.1 p.5, 2.2.2 p.6)) |
| Climate control | Y (Ventilation opened above 17C; floor-level heating target 10C night / 14C day (Methods 2.1, p.3); weekly mean greenhouse climate in Table 1) |
| Nutrient supplemented | Y (Both aquaponic treatments (fish water) supplemented with same mineral fertilizers as control to reach identical target EC/nutrient recipe (Hochmuth 2001); total fertilizer supply reduced by 62.8% overall vs control due to nutrients already present in fish water (Table 3, p.9)) |
| Equipment | EHEIM universal 600 pumps; HI9811-5 Hanna pH/EC meter; LI-3100 Area Meter (LICOR) for leaf area; RQflex 10 plus reflectometer (Merck) for nitrate; ICP-OES iCAP 6300Duo (Thermo) for macro/micronutrients; CFA San++ (Skalar) for NH4-N/NO3-N; TOC-L CPN analyzer (Shimadzu) for TOC/DOC; vario MAX elemental analyser (Elementar) for N/C; HPLC Ultimate 3000 (Thermo Fisher) for phenolics; 150 L stainless steel disinfection pot with butane gas burner (3.6 kW) + 2x2kW heating rods + plate heat exchanger |
| Control Parameters | 3 treatments (Control, APunt, APdis) x 3 replicate gullies/treatment (9 gullies, 90 plants total); EC target 2.2 dS/m at preparation (all treatments); pH target 5.8 at preparation; disinfection >95C for 15 min then solution maintained >70C for >45 min (APdis only) |
| Combination | Nile tilapia (Oreochromis niloticus) and butterhead lettuce (Lactuca sativa var. Salanova) in a decoupled, non-recirculating-to-RAS aquaponic system; fish water withdrawn once as a static batch, heat-disinfected, and fertilizer-supplemented, not returned to the RAS; this row = disinfected aquaponic fish water (APdis) vs Control |
Site
| Field | Value |
|---|---|
| Region | Europe |
| Country | Germany |
| Average room Temperature | 13.36-21.80 (day) / 9.26-16.28 (night) (weekly means across the trial, Table 1 p.4; no single trial mean stated) - range only, no trial mean reported |
Results & Statistics
| Field | Value |
|---|---|
| Measured Unit | g/plant (fresh weight, primary yield metric); % (dry matter); mg/kg FW (tissue nitrate); mg/100g FW (phenolics); mg/L (nutrient solution chemistry) |
| Statistic Details | Univariate ANOVA (SPSS v19.0) after Shapiro-Wilk normality test; Tukey-HSD (homogeneous variance) or Dunnett-T3 (heterogeneous variance) post hoc; Kruskal-Wallis + Dunn-Bonferroni for non-normal data; CFU: Kruskal-Wallis (initial) + Friedman (weekly); significance p<0.05 (Methods 2.6, p.8) |
| Statistically analysed | Y |
| Replicates (n) | 3 |
| AP | 332.4 |
| HYD | 325.9 |
Experimental Remarks: TRIAL DEFINITION: T2 = decoupled aquaponic treatment using disinfected fish water (withdrawn once as a static batch from an established Nile tilapia DRAPS/RAS at IGB Berlin, then supplemented with mineral fertilizer to the same target EC/nutrient recipe as the control; not returned to the RAS). Paired control = Control (fresh tap/rain water 50:50 v/v + mineral fertilizer), recorded in the HYD-labelled cells. Design: 3 treatments x 3 replicate NFT gullies/treatment (9 gullies, 90 plants total), 10 plants/gully; ‘randomized’ never explicitly stated for treatment-to-gully assignment (Fig 1 layout is non-repeating across the three physical positions but not labelled as random, see Extraction notes). | Both aquaponic arms in this paper (APunt, APdis) are full-remineralization decoupled aquaponics: fish water was fertilizer-supplemented to hit the SAME target nutrient recipe as the hydroponic control, not a minimal/no-supplementation design. | WARN-BLOCK greenhouse coordinates, Methods 2.1 p.3: printed as ‘(52°46´82.806´´N, 13°29´88.909´´E)’. Both minutes fields (46, 29) are valid DMS (<60) but both seconds fields (82.806, 88.909) exceed 59, impossible in DMS notation, in both latitude and longitude simultaneously. Unlike levizouCircularTriTrophicSystem2025 (where only one coordinate was invalid and could be confidently reread as decimal-degrees-mistakenly-DMS-formatted), no single reinterpretation here reproduces both printed digit strings without assuming an unstated transcription error (e.g. digit transposition 82.806->28.806 or 88.909->38.909) that cannot be confirmed from the text. Recorded Lat and Long as UNCLEAR. Affects: Lat, Long cells only; Region (Europe) and Country (Germany) taken directly from the stated place name ‘Berlin, Germany’, unaffected. Added to REVIEW.md by the batch merge step. | WARN-MINOR cultivation duration, Table 1 caption (p.4) and Results 3.1 (p.8) vs. Methods 2.1 dates (p.3): paper twice states duration as ‘seven weeks’ (49 days by conversion) but transplant (18 March 2016) to harvest (10 May 2016) spans 53 days (7.57 weeks) by direct date subtraction, a 4-day discrepancy. Recorded Days Plant after transplant = 49 (UNIT CONVERSION ONLY: 7 weeks x 7 days, the paper’s own repeated, directly-stated figure) rather than the date-derived 53. Does not affect any other cell; Table 2’s growth data is reported simply ‘at harvest’. | Table 4’s full nutrient/mineral panel (incl. TAN/NH4-N and NO3-N) is captioned ‘at the beginning of the experiment’ — a single measurement of the freshly mixed solution before planting, never repeated later in the trial; recorded as given, with this single-timepoint basis stated explicitly rather than presented as an ordinary repeated-measures trial mean. | NOT DERIVED, left NR: FCR, SGR, Fish size initial/final, Fish weight gain, Fish biomass created (kg), Fish survival rate, Total Feed (kg), Fish trial duration (days), Feed routine (feeding frequency/day not stated), feed N/P/K composition beyond crude protein — the RAS supplying fish water is an already-established, continuously-operating system (companion Monsees, Kloas & Würtz 2017 paper) rather than a fish cohort grown for this lettuce trial; only standing stocking density (90.6 kg/m3), total biomass (462.1 kg) and feed identity/ration are given as background (Methods 2.1.2). Initial Stock density (90.6 kg/m3) recorded since explicitly stated as a density, but describes the parallel RAS operation, not a batch tied to this trial’s 7-8 week duration. | NOT DERIVED, left NR: Water recycle (L/min, not stated); Daily Water exchange rate (no top-up/exchange schedule during the trial described; static 750 L batch recirculated within its own tank); Water temperature (only RAS fish-tank temperature, 25C, a different compartment, and greenhouse air temperature, Table 1, a different quantity, are given); Dissolved Oxygen (only RAS fish-tank DO given, ‘always above 5.5 mg/L’, not the nutrient-solution/plant-bed side); Water classification (no term distinct from Water type used by the paper; SCHEMA.md notes this distinction is itself undefined); FUE AP/HYD and WUE (paper reports fertilizer savings as a % mass reduction, Table 3, and 100% freshwater substitution, not as a yield-per-input ratio; computing one would be derivation — see NO COLUMN fertilizer/GHG figures below); Plants/m2 (only gully dimensions 6 m x 0.35 m and a 10-plants/gully count given, not a stated density; computing one would be derivation). | NO COLUMN: Full nutrient solution mineral panel at start of experiment (Table 4, p.10): P, K, Ca, Mg, S, Na, Fe, Mn, Mo, Zn, Cu, B in mg/L for Control/APunt/APdis with significance letters — water chemistry beyond NH4-N/NO3-N has no dedicated trials.csv column; flagged as too valuable to discard silently per SCHEMA.md. Total and dissolved organic carbon (Table 5, p.12): TOC/DOC in mg/L at weeks 0, 4, 7 for all three treatments, 2.9- to 4.1-fold higher in both aquaponic treatments throughout. Colony-forming units (Fig 3, p.13): weekly CFU/mL series (177 control / 5097 untreated fish water at baseline; ~2.1-2.3x10^4 both aquaponic treatments 5 days after planting; 502-949 CFU/mL all treatments by week 7); no CFU detected in APdis immediately after disinfection. This is water/system microbiology, not plant tissue microbiology, so excluded from plant.csv too. Fertilizer savings and GHG accounting (Table 3, p.9; Discussion 4.2): per-fertilizer reduction (CAN -84.6%, KNO3 -68.4%, MgSO4 -36.4%, KH2PO4 -7.4%), total fertilizer supply (Control 207, APunt/APdis 77 g/L stock solution), per-fertilizer CO2-eq figures, and extrapolated 1-ha annual GHG estimate (16.3 vs 4.5 t CO2-eq/year) — this paper’s headline sustainability finding, no dedicated column, fully preserved here. Leaf area (Table 2, p.8): Control 66.9+/-5.7, APunt 66.7+/-6.8, APdis 67.5+/-7.2 dm2/plant, ns — no dedicated morphology column beyond height/leaf count, and not a tissue analyte so excluded from plant.csv per the same reasoning used for physiological measurements in mourantianBasilFunctionalGrowth2023. Literature comparison figures (secondary, not this paper’s data): Delaide et al. 2016’s 39% lettuce growth increase in decoupled aquaponics; Barbosa et al. 2015’s cited lettuce yield (41 kg/m2/year) and water-use (20 L/kg/year) figures used only to build this paper’s own hypothetical 1-ha extrapolation.
Plant Measurements
| Trial | System | Category | Analyte | Value | Unit | Sig. | Location |
|---|---|---|---|---|---|---|---|
| monseesLettuceLactucaSativa2019-T1 | HYD | mineral | Carbon (C) | 42.25 ± 0.44 | %DM | a | Table 6, p.13 |
| monseesLettuceLactucaSativa2019-T1 | AP | mineral | Carbon (C) | 42.48 ± 0.59 | %DM | a | Table 6, p.13 |
| monseesLettuceLactucaSativa2019-T1 | HYD | mineral | Nitrogen (N) | 4.12 ± 0.19 | %DM | a | Table 6, p.13 |
| monseesLettuceLactucaSativa2019-T1 | AP | mineral | Nitrogen (N) | 4.29 ± 0.12 | %DM | a | Table 6, p.13 |
| monseesLettuceLactucaSativa2019-T1 | HYD | mineral | Phosphorus (P) | 0.71 ± 0.03 | %DM | a | Table 6, p.13 |
| monseesLettuceLactucaSativa2019-T1 | AP | mineral | Phosphorus (P) | 0.73 ± 0.02 | %DM | a | Table 6, p.13 |
| monseesLettuceLactucaSativa2019-T1 | HYD | mineral | Potassium (K) | 6.33 ± 0.21 | %DM | a | Table 6, p.13 |
| monseesLettuceLactucaSativa2019-T1 | AP | mineral | Potassium (K) | 5.94 ± 0.16 | %DM | a | Table 6, p.13 |
| monseesLettuceLactucaSativa2019-T1 | HYD | mineral | Calcium (Ca) | 1.79 ± 0.19 | %DM | ab | Table 6, p.13 |
| monseesLettuceLactucaSativa2019-T1 | AP | mineral | Calcium (Ca) | 1.91 ± 0.06 | %DM | b | Table 6, p.13 |
| monseesLettuceLactucaSativa2019-T1 | HYD | mineral | Magnesium (Mg) | 0.28 ± 0.03 | %DM | ab | Table 6, p.13 |
| monseesLettuceLactucaSativa2019-T1 | AP | mineral | Magnesium (Mg) | 0.31 ± 0.0 | %DM | b | Table 6, p.13 |
| monseesLettuceLactucaSativa2019-T1 | HYD | mineral | Sulfur (S) | 0.26 ± 0.01 | %DM | a | Table 6, p.13 |
| monseesLettuceLactucaSativa2019-T1 | AP | mineral | Sulfur (S) | 0.27 ± 0.01 | %DM | ab | Table 6, p.13 |
| monseesLettuceLactucaSativa2019-T1 | HYD | mineral | Sodium (Na) | 0.27 ± 0.02 | %DM | b | Table 6, p.13 |
| monseesLettuceLactucaSativa2019-T1 | AP | mineral | Sodium (Na) | 0.22 ± 0.01 | %DM | a | Table 6, p.13 |
| monseesLettuceLactucaSativa2019-T1 | HYD | mineral | Iron (Fe) | 0.59 ± 0.06 | %DM | a | Table 6, p.13 |
| monseesLettuceLactucaSativa2019-T1 | AP | mineral | Iron (Fe) | 0.6 ± 0.1 | %DM | a | Table 6, p.13 |
| monseesLettuceLactucaSativa2019-T1 | HYD | biochemistry | Caffeoyltartaric acid (CTA) | 0.97 ± 0.23 | mg/100g FW | ns | Table 7, p.14 |
| monseesLettuceLactucaSativa2019-T1 | AP | biochemistry | Caffeoyltartaric acid (CTA) | 1.22 ± 0.34 | mg/100g FW | ns | Table 7, p.14 |
| monseesLettuceLactucaSativa2019-T1 | HYD | biochemistry | Caffeoylquinic acid (chlorogenic acid) | 0.93 ± 0.15 | mg/100g FW | ns | Table 7, p.14 |
| monseesLettuceLactucaSativa2019-T1 | AP | biochemistry | Caffeoylquinic acid (chlorogenic acid) | 1.08 ± 0.24 | mg/100g FW | ns | Table 7, p.14 |
| monseesLettuceLactucaSativa2019-T1 | HYD | biochemistry | Coumaroylquinic acid | 0.028 ± 0.005 | mg/100g FW | ns | Table 7, p.14 |
| monseesLettuceLactucaSativa2019-T1 | AP | biochemistry | Coumaroylquinic acid | 0.033 ± 0.013 | mg/100g FW | ns | Table 7, p.14 |
| monseesLettuceLactucaSativa2019-T1 | HYD | biochemistry | Caffeoylmalic acid (CMA) | 1.07 ± 0.22 | mg/100g FW | ns | Table 7, p.14 |
| monseesLettuceLactucaSativa2019-T1 | AP | biochemistry | Caffeoylmalic acid (CMA) | 1.13 ± 0.36 | mg/100g FW | ns | Table 7, p.14 |
| monseesLettuceLactucaSativa2019-T1 | HYD | biochemistry | Dicaffeoyltartaric acid (chicoric acid) | 2.45 ± 0.58 | mg/100g FW | ns | Table 7, p.14 |
| monseesLettuceLactucaSativa2019-T1 | AP | biochemistry | Dicaffeoyltartaric acid (chicoric acid) | 3.26 ± 0.74 | mg/100g FW | ns | Table 7, p.14 |
| monseesLettuceLactucaSativa2019-T1 | HYD | biochemistry | Quercetin-3-O-(6”-malonyl)glucoside (Q3MG) | 0.25 ± 0.03 | mg/100g FW | ns | Table 7, p.14 |
| monseesLettuceLactucaSativa2019-T1 | AP | biochemistry | Quercetin-3-O-(6”-malonyl)glucoside (Q3MG) | 0.33 ± 0.07 | mg/100g FW | ns | Table 7, p.14 |
| monseesLettuceLactucaSativa2019-T2 | HYD | mineral | Carbon (C) | 42.25 ± 0.44 | %DM | a | Table 6, p.13 |
| monseesLettuceLactucaSativa2019-T2 | AP | mineral | Carbon (C) | 42.0 ± 0.34 | %DM | a | Table 6, p.13 |
| monseesLettuceLactucaSativa2019-T2 | HYD | mineral | Nitrogen (N) | 4.12 ± 0.19 | %DM | a | Table 6, p.13 |
| monseesLettuceLactucaSativa2019-T2 | AP | mineral | Nitrogen (N) | 4.27 ± 0.09 | %DM | a | Table 6, p.13 |
| monseesLettuceLactucaSativa2019-T2 | HYD | mineral | Phosphorus (P) | 0.71 ± 0.03 | %DM | a | Table 6, p.13 |
| monseesLettuceLactucaSativa2019-T2 | AP | mineral | Phosphorus (P) | 0.73 ± 0.03 | %DM | a | Table 6, p.13 |
| monseesLettuceLactucaSativa2019-T2 | HYD | mineral | Potassium (K) | 6.33 ± 0.21 | %DM | a | Table 6, p.13 |
| monseesLettuceLactucaSativa2019-T2 | AP | mineral | Potassium (K) | 6.29 ± 0.32 | %DM | a | Table 6, p.13 |
| monseesLettuceLactucaSativa2019-T2 | HYD | mineral | Calcium (Ca) | 1.79 ± 0.19 | %DM | ab | Table 6, p.13 |
| monseesLettuceLactucaSativa2019-T2 | AP | mineral | Calcium (Ca) | 1.74 ± 0.06 | %DM | a | Table 6, p.13 |
| monseesLettuceLactucaSativa2019-T2 | HYD | mineral | Magnesium (Mg) | 0.28 ± 0.03 | %DM | ab | Table 6, p.13 |
| monseesLettuceLactucaSativa2019-T2 | AP | mineral | Magnesium (Mg) | 0.28 ± 0.01 | %DM | a | Table 6, p.13 |
| monseesLettuceLactucaSativa2019-T2 | HYD | mineral | Sulfur (S) | 0.26 ± 0.01 | %DM | a | Table 6, p.13 |
| monseesLettuceLactucaSativa2019-T2 | AP | mineral | Sulfur (S) | 0.28 ± 0.01 | %DM | b | Table 6, p.13 |
| monseesLettuceLactucaSativa2019-T2 | HYD | mineral | Sodium (Na) | 0.27 ± 0.02 | %DM | b | Table 6, p.13 |
| monseesLettuceLactucaSativa2019-T2 | AP | mineral | Sodium (Na) | 0.21 ± 0.01 | %DM | ab | Table 6, p.13 |
| monseesLettuceLactucaSativa2019-T2 | HYD | mineral | Iron (Fe) | 0.59 ± 0.06 | %DM | a | Table 6, p.13 |
| monseesLettuceLactucaSativa2019-T2 | AP | mineral | Iron (Fe) | 0.57 ± 0.02 | %DM | a | Table 6, p.13 |
| monseesLettuceLactucaSativa2019-T2 | HYD | biochemistry | Caffeoyltartaric acid (CTA) | 0.97 ± 0.23 | mg/100g FW | ns | Table 7, p.14 |
| monseesLettuceLactucaSativa2019-T2 | AP | biochemistry | Caffeoyltartaric acid (CTA) | 1.29 ± 0.29 | mg/100g FW | ns | Table 7, p.14 |
| monseesLettuceLactucaSativa2019-T2 | HYD | biochemistry | Caffeoylquinic acid (chlorogenic acid) | 0.93 ± 0.15 | mg/100g FW | ns | Table 7, p.14 |
| monseesLettuceLactucaSativa2019-T2 | AP | biochemistry | Caffeoylquinic acid (chlorogenic acid) | 1.12 ± 0.24 | mg/100g FW | ns | Table 7, p.14 |
| monseesLettuceLactucaSativa2019-T2 | HYD | biochemistry | Coumaroylquinic acid | 0.028 ± 0.005 | mg/100g FW | ns | Table 7, p.14 |
| monseesLettuceLactucaSativa2019-T2 | AP | biochemistry | Coumaroylquinic acid | 0.038 ± 0.01 | mg/100g FW | ns | Table 7, p.14 |
| monseesLettuceLactucaSativa2019-T2 | HYD | biochemistry | Caffeoylmalic acid (CMA) | 1.07 ± 0.22 | mg/100g FW | ns | Table 7, p.14 |
| monseesLettuceLactucaSativa2019-T2 | AP | biochemistry | Caffeoylmalic acid (CMA) | 1.34 ± 0.36 | mg/100g FW | ns | Table 7, p.14 |
| monseesLettuceLactucaSativa2019-T2 | HYD | biochemistry | Dicaffeoyltartaric acid (chicoric acid) | 2.45 ± 0.58 | mg/100g FW | ns | Table 7, p.14 |
| monseesLettuceLactucaSativa2019-T2 | AP | biochemistry | Dicaffeoyltartaric acid (chicoric acid) | 3.44 ± 0.75 | mg/100g FW | ns | Table 7, p.14 |
| monseesLettuceLactucaSativa2019-T2 | HYD | biochemistry | Quercetin-3-O-(6”-malonyl)glucoside (Q3MG) | 0.25 ± 0.03 | mg/100g FW | ns | Table 7, p.14 |
| monseesLettuceLactucaSativa2019-T2 | AP | biochemistry | Quercetin-3-O-(6”-malonyl)glucoside (Q3MG) | 0.36 ± 0.05 | mg/100g FW | ns | Table 7, p.14 |