Quartz 5

Advanced aquaponics: Evaluation of intensive tomato production in aquaponics vs. conventional hydroponics

Properties1
tagsmeta/type/quasi-experiment, meta/region/europe, meta/fish/tilapia, meta/plant/tomato

34 min read

Advanced aquaponics: Evaluation of intensive tomato production in aquaponics vs. conventional hydroponics

Metadata

  • Cite key: suhlAdvancedAquaponicsEvaluation2016
  • Item type: Journal Article
  • Authors: J. Suhl, D. Dannehl, W. Kloas, D. Baganz, S. Jobs, G. Scheibe, U. Schmidt
  • Affiliation: Department of Biology and Ecology of Fishes, Leibniz-Institute of Freshwater Ecology and Inland Fisheries, Berlin (Suhl, Baganz); Faculty of Life Sciences, Albrecht Daniel Thaer-Institute of Agricultural and Horticultural Sciences, Division Biosystems Engineering, Humboldt-Universität zu Berlin (Suhl, Dannehl, Schmidt); Department of Ecophysiology and Aquaculture, Leibniz-IGB, Berlin (Kloas); Faculty of Life Sciences, Institute of Biology, Department of Endocrinology, Humboldt-Universität zu Berlin (Kloas); PAL-Anlagenbau GmbH Abtshagen, Wittenhagen/Abtshagen, Germany (Jobs, Scheibe)
  • Journal: Agricultural Water Management 178 (2016) 335-344
  • Date: received 23 May 2016, accepted 14 Oct 2016, online 21 Oct 2016
  • Date added: [not reported]
  • DOI: 10.1016/j.agwat.2016.10.013
  • Funding: European Commission via INAPRO (grant agreement number 619137)
  • URL: https://doi.org/10.1016/j.agwat.2016.10.013
  • PDF: Suhl et al. - 2016 - Advanced aquaponics Evaluation of intensive tomat.pdf

Opinion

A methodologically transparent, single-season case study of a genuinely novel system (DRAPS — two fully independent recirculating loops for fish and plants, connected one-directionally, with the fish-derived water fertiliser-amended to a matched target recipe before reaching the crop). Despite never using the word “decoupled,” this is functionally a decoupled-aquaponics-vs-hydroponics comparison, and its result (yield parity, 25% fertiliser saving, comparable fruit quality) is a useful, early, directly-relevant data point for the DCAP literature this vault already cites it for (see mourantianBasilFunctionalGrowth2023). Two things temper confidence: (1) the design has only ONE double-gully structure per treatment (no independent replicate systems), so statistical power comes entirely from plant-level pseudoreplication (n=48) rather than true system-level replication — borderline experiment/quasi-experiment, resolved here as the latter (see Extraction notes); (2) the paper’s own arithmetic for its headline “1.55 kg tilapia + 46.1 kg tomato per m3 fresh water” claim conflates two different water-volume denominators (total RAS fresh water vs. the fraction actually reaching the crop) — each number is individually reproducible from other stated figures, but the sentence’s own framing doesn’t hold together (see WARN-MATERIAL below). The authors are commendably candid about the fish side under-performing relative to the system’s designer (Kloas et al. 2015) and about the low absolute nutrient loads this created, which limits how much of the “fertiliser saving” result can be read as a ceiling on DRAPS’s potential rather than this run’s actual outcome.

Abstract

Aquaponics for intensive crop production is a highly complex system in which three different biological systems (fish, plants, and nitrifying bacteria) with different requirements must be merged. Finding the right combination is a serious challenge and the dependencies avoid a high productivity until now. Therefore, a unique and innovative double recirculating aquaponic system (DRAPS) was developed as a prerequisite for a high productivity comparable to professional stand-alone fish/plant facilities. It consists of two independent recirculating units — a recirculating aquaculture unit for fish production and a closed hydroponic cycle for plant production — which were connected unidirectional. This allows the use of fish waste water as nutrient supply for plants in hydroponics and its optimisation for plant growth by fertilizer supply without negative effects on fish rearing. Furthermore it allows a sustainable food production.

In a new constructed DRAPS research facility, first investigations with tilapia and tomato production were conducted in 2015. During an annual production, it was demonstrated that in DRAPS comparable tomato yields were produced as obtained for conventional hydroponics. Even fruit parameters such as contents of lycopene and ß-carotene resulted in the same quantity when both systems were compared. Furthermore, the fertilizer use efficiency was increased by 23.6% in favour of the DRAPS. The total fresh water use efficiency was also increased using aquaponics.

Summary

The authors built a new research aquaponics facility in Abtshagen, Germany, implementing a “double recirculating aquaponic system” (DRAPS): a fully independent recirculating aquaculture system (RAS, 4 tanks, ~12 m3 total, stocked with all-male Nile tilapia in three staggered batches) whose water is periodically diverted through a 3-chamber pit and storage tank, topped up with mineral fertiliser to a target nutrient recipe, and then used to irrigate a separate, closed NFT tomato loop — water does not return to the fish tanks. This single aquaponic treatment (48 plants, one double-gully structure) was compared against conventional hydroponics (fresh water + the same target fertiliser recipe, also 48 plants, one double-gully structure) over an annual tomato production cycle (cv. Pureza) running January 2015-January 2016, with fruit harvested weekly for 197 days. Total cumulative yield was statistically indistinguishable between treatments (hydroponics 31.64 kg/m2, aquaponics 29.38 kg/m2), as were fruit lycopene and ß-carotene content, though aquaponic fruit had significantly lower dry matter, soluble solids and sugar-acid ratio, smaller leaf area, and higher chlorophyll-index (Chl NDI) readings than hydroponic fruit/leaves. The aquaponic treatment used 25.2% less mineral fertiliser than hydroponics for a comparable yield, improving fertiliser-use efficiency by 23.6% (54.0 vs 43.7 kg tomato per kg fertiliser), and the combined system was framed as using fresh water more efficiently overall because it also produced fish (248.84 kg total) alongside the tomato crop. The paper is candid that fish production itself underperformed relative to the DRAPS prototype’s designers (Kloas et al. 2015), attributing this to a much lower and more variable stocking density during the study, and argues that the plant-side result (yield parity, fertiliser savings, unchanged fruit quality) would likely improve further if fish productivity were optimised.

Experiment data

  • Location: Research aquaponics facility, Abtshagen, Germany (52.5200 N, 13.4050 E); Venlo-type greenhouse (139 m2) + RAS fish farm (43 m2) + technical room (14 m2), total facility 196 m2
  • Design: 2 treatments (hydroponics vs. DRAPS/aquaponics), each confined to ONE double-gully NFT structure (48 plants/treatment); no independent replicate systems per treatment and no stated randomisation of treatment-to-gully assignment — see Extraction notes on the experiment/quasi-experiment judgment call
  • Replicates / n: 48 plants/treatment (primary unit for yield/BER stats); smaller sub-samples for other measurements (n=10 fruit-quality categorisation, n=12 leaf area/leaf count, n=6 plant length, n=192 Chl NDI, n=9 fruit chemistry, n=8 water-nutrient sampling weeks)
  • Duration: Fish production 23.01.2015-08.01.2016 (351 days); tomato planted 22.01.2015, harvested weekly 14.04.2015-27.10.2015 (197 days, stated)
  • Organisms: Tomato (Solanum lycopersicum) (cv. Pureza) / Nile tilapia (Oreochromis niloticus) (“Red Natural Male Tilapia” stock, all-male)
  • Statistics: parametric t-test (after Kolmogorov-Smirnov normality check) or non-parametric Mann-Whitney U test; SPSS 19.0; significance p<0.05
  • Yield: Hydroponics 31.64 kg/m2 vs. aquaponics 29.38 kg/m2, not significantly different
  • Fertiliser use efficiency: aquaponics 54.0 vs. hydroponics 43.7 kg tomato per kg fertiliser (+23.6%), using 25.2% less total fertiliser
  • Fruit quality: lycopene (HYD 0.99, AP 0.90 mg/g DM) and ß-carotene (HYD 0.20, AP 0.19 mg/g DM) not significantly different; dry matter, SSC and SAR all significantly lower in aquaponic fruit

Yield

This paper: Total cumulative tomato yield was statistically indistinguishable between hydroponics (31.64 kg/m2) and the DRAPS aquaponic treatment (29.38 kg/m2) over the 197-day harvest period (Results 3.1, p.338; restated as “29.2 kg/m2” in the Discussion, a minor rounding difference, see Extraction notes). Marketable-fruit rate was near-identical (HYD 99.1%, AP 99.5%), with aquaponic fruit showing a slightly lower blossom-end-rot (BER) rate, which the authors attribute to higher calcium in the fish-waste-water-derived nutrient solution. The authors frame this as their central result: DRAPS, with continuous nutrient monitoring and fertiliser top-up, can match conventional hydroponic yield for an intensive fruiting crop — something the Introduction notes was previously reported as failing for single-loop (coupled) aquaponic tomato production (Graber and Junge, 2009).

Compared with:

  • todo Kloas et al. 2015 — first DRAPS prototype, 20.5 kg/m2 tomato yield over a comparable harvest duration; this paper’s 29.2/29.38 kg/m2 is framed as a 1.4-fold improvement attributed to continuous nutrient measurement/optimisation of the fish waste water rather than periodic dosing (p.341)
  • todo Roosta and Hamidpour 2011 — equal tomato yields achievable in single-loop (coupled) aquaponics via foliar macro/micronutrient application, but only over a 22-day trial, which this paper’s own Discussion calls “not meaningful” given the short duration (p.341)
  • todo Wortman 2015 — relative vegetative growth rates of basil, kale, tomato and pepper did not differ between simulated aquaponics (low EC + high pH) and hydroponics (high EC + low pH), but marketable yield of all species was significantly reduced in aquaponics; recommends fertiliser supplementation to close the gap, consistent with this paper’s own DCAP-style approach (p.341)
  • todo Graber and Junge 2009 — tomato yields significantly lower in (coupled) aquaponics than hydroponics, cited in the Introduction as the state of the evidence this paper set out to improve on (p.336)
  • todo Pantanella et al. 2012 — lettuce yields similar between aquaponics and hydroponics, but dependent on fish stocking density with nutrients supplemented, cited as a precedent for the density/nutrient-dependence logic (p.336)

Fruit quality (dry matter, soluble solids, carotenoids)

This paper: Aquaponic fruit had significantly lower dry matter (AP 5.12% vs HYD 5.37%), soluble solids content (SSC, -7% relative) and sugar-acid ratio (SAR, -5.0% relative) than hydroponic fruit (Results 3.2, p.338-339). Lycopene (HYD 0.99, AP 0.90 mg/g DM) and ß-carotene (HYD 0.20, AP 0.19 mg/g DM) did not differ significantly. The authors link the lower nitrate-nitrogen concentration in the hydroponic solution (vs. the fish-waste-derived one) to the higher DM/SSC/SAR in hydroponic fruit, citing a nitrogen-supply mechanism from Béinard et al. (2009), and note that SSC/SAR are linked to consumer-perceived flavour, recommending a future sensory analysis. Carotenoid parity between treatments is highlighted in the abstract as a headline food-quality result despite the significant NO3-N, P and Ca differences in the underlying nutrient solutions (Table 3).

Compared with:

  • todo Béinard et al. 2009 — low nitrogen supply increases tomato fruit dry matter content and improves quality in terms of SSC, cited as the mechanism for this paper’s DM/SSC results (p.341)
  • todo Baldwin et al. 1998 — SSC and SAR positively correlate with consumer-perceived tomato flavour/acceptability, cited to argue the SSC/SAR gap found here may translate to a taste difference not directly tested (p.341)
  • todo Frusciante et al. 2007 — reference lycopene/ß-carotene tomato contents against which this paper judges its own carotenoid values as “high” (p.341)
  • todo Dumas et al. 2003; Montagu and Goh 1990; Zelená et al. 2009 — lycopene positively correlated with P/S supply, negatively with N/Ca supply; ß-carotene negatively correlated with Ca — cited as background for why carotenoid parity was somewhat unexpected given this paper’s own NO3-N/P/Ca differences between treatments (p.341)

Growth (leaf area, leaf count, plant length, Chl NDI)

This paper: Aquaponic plants had significantly smaller leaf area per plant (1.15 vs 1.36 m2, -15.4%) without a corresponding reduction in leaf number (20.5 vs 21.0, ns) or total plant/vine length (10.9 vs 10.8 m, ns) (Table 2, p.340). Aquaponic leaves had a significantly higher chlorophyll normalised difference index (Chl NDI, a reflectance-based proxy, not a SPAD reading) than hydroponic leaves (0.66 vs 0.56, +17.9%). The authors interpret the combination of smaller leaf area but higher Chl NDI and unchanged yield as evidence that aquaponic plants partially compensated a reduced photosynthetic source area with higher source activity (chlorophyll content/photosynthetic efficiency per unit leaf), net-neutral for yield.

Compared with: (no external literature comparison given for the leaf-area/Chl-NDI compensation mechanism specifically; presented as this paper’s own interpretation, citing only the general source-sink framework of Engels et al. 2012)

Water and fertiliser use efficiency

This paper: Total fertiliser addition was reduced by 25.2% in the aquaponic treatment (11.6 kg vs. HYD’s 15.5 kg), improving fertiliser-use efficiency (FUE) by 23.6% (54.0 vs. 43.7 kg tomato per kg fertiliser, Table 4) — both percentages are internally consistent with Table 4’s raw figures (unlike some other DCAP papers in this vault where the abstract/results FUE percentages disagree, e.g. levizouCircularTriTrophicSystem2025). Total fresh-water use efficiency (FWUE) was reported as 47.7 kg tomato/m3 for hydroponics (677.3 kg yield / 14.2 m3 fresh water) and framed as “1.55 kg tilapia and 46.1 kg tomato… produced with one m3 fresh water applied to the RAS” for the combined aquaponic system — but see the WARN-MATERIAL in Extraction notes: these two aquaponic-side figures are computed from two different water volumes (160.4 m3 total RAS fresh water for the fish figure; 13.6 m3, the fraction of that water reaching the crop, for the tomato figure), even though the sentence presents them as sharing one denominator.

Compared with:

  • todo Kloas et al. 2015 — DRAPS prototype achieved 5 kg tilapia + 25 kg tomato per m3 fresh water, and a mean fish-water NO3-N of 127.7 mg/L (vs. this paper’s much lower, more variable 14.6 mg/L), attributed to lower/fluctuating stocking density and a technically more advanced condensation-water-reuse system in the earlier prototype (p.342)

Linked claims

Citations to chase

  • todo Kloas W, Groß R, Baganz D, Graupner J, Monsees H, Schmidt U, Staaks G, Suhl J, Tschirner M, Wittstock B, Wuertz S, Zikova A, Rennert B (2015) — A new concept for aquaponic systems to improve sustainability, increase productivity, and to reduce environmental impacts, Aquaculture Environment Interactions 7:179-192 — describes the DRAPS prototype this paper optimises
  • todo Roosta HR, Hamidpour M (2011) — Effects of foliar application of some macro- and micro-nutrients on tomato plants in aquaponic and hydroponic systems, Scientia Horticulturae 129:396-402
  • todo Wortman SE (2015) — Crop physiological response to nutrient solution electrical conductivity and pH in an ebb-and-flow hydroponic system, Scientia Horticulturae 194:34-42
  • todo Graber A, Junge R (2009) — Aquaponic systems: nutrient recycling from fish wastewater by vegetable production, Desalination 246:147-156
  • todo Pantanella E, Cardarelli M, Colla G, Rea E, Marcucci A (2012) — Aquaponics vs. hydroponics: production and quality of lettuce crop — already extracted directly, see pantanellaAquaponicsHydroponicsProduction2012
  • todo Béinard C, Gautier H, Bourgaud F, Grasselly D, Navez B, Caris-Veyrat C, Weiss M, Genard M (2009) — Effects of low nitrogen supply on tomato fruit yield and quality, Journal of Agricultural and Food Chemistry 57:4112-4123
  • todo Baldwin EA, Scott JW, Einstein MA, Malundo TMM, Carr BT, Shewfelt RL, Tandon KS (1998) — Relationship between sensory and instrumental analysis for tomato flavor, Journal of the American Society for Horticultural Science 123:906-915
  • todo Roosta HR (2014) — Comparison of the vegetative growth, eco-physiological characteristics and mineral nutrient content of basil plants in different irrigation ratios of hydroponic:aquaponic solutions, Journal of Plant Nutrition 37:1782-1803 — cited in the Introduction as evidence that “many growth parameters of mint and basil were negatively influenced by aquaponics” (p.336)

Extraction notes

Type classification judgment call: Recorded as quasi-experiment, not experiment. Two defined treatments (HYD, DRAPS-aquaponics) are compared with formal statistics (t-test/Mann-Whitney U, p<0.05) and plant-level replication (n=48 individual plants/treatment for the primary yield outcome), which would support experiment. However, Methods 2.2.2 (p.338) describes only “two different treatments with two gullies each,” and Methods 2.1 (p.337) states the greenhouse contained exactly “two double gullies” total (“only the double gullies were considered” for evaluation) — meaning each treatment occupies a single physical gully-structure with no independent replicate aquaponic/hydroponic systems, and the Methods never state that treatment-to-gully assignment was randomised. This is a true-experimental-unit-of-one-per-treatment design (system-level pseudoreplication), distinct from e.g. pantanellaAquaponicsHydroponicsProduction2012 (3 independent physical systems/treatment) or mourantianBasilFunctionalGrowth2023 (6 independent channels/treatment), both recorded as experiment in this vault. Per SCHEMA.md Part 1 decision rule 2 (“Randomised treatments with replication? … Treatments without randomisation or true replication -> quasi-experiment”), quasi-experiment was judged the better fit. Flagged here as a judgment call since the paper’s own statistical treatment (formal significance testing throughout) reads much like a standard experiment write-up.

Trial structure: ONE aquaponic treatment (DRAPS) vs. ONE hydroponic control -> one trials.csv row. This paper’s own system, despite never using the word “decoupled,” is functionally a decoupled aquaponic design: the RAS and the plant-irrigation loop are fully independent recirculating cycles connected only unidirectionally (fish waste water -> 3-chamber pit -> storage tank -> fertiliser-amended -> plants; water never returns to the fish tanks), which matches how this vault records DCAP systems elsewhere (e.g. mourantianBasilFunctionalGrowth2023’s DCAP arm). There is no coupled-aquaponics (CAP) arm or variant tested anywhere in this paper — the task brief that prompted this extraction assumed a coupled-and-decoupled design (by analogy to other papers in this vault), but on reading the actual paper this assumption does not hold: it is a single aquaponic-system-design vs. hydroponics comparison.

⚠️WARN-CHECK TAN/NH4-N and NO3-N, two candidate water-quality bases (Table 3, p.339). The paper reports both “pure fish waste water” (RAS effluent before fertiliser amendment, mean over the whole 351-day fish production period: NH4-N 24.2±20.8 mg/L, NO3-N 14.6±13.9 mg/L) and the “nutrient solution based on fish waste water” (the fertiliser-amended solution actually delivered to the aquaponic gullies, 8-week mean, n=8: NH4-N 7.0±9.0 mg/L, NO3-N 157.0±50.3 mg/L). These measure genuinely different things and the paper never states which should represent “the aquaponic system’s water quality.” Recorded the delivered-solution values in trials.csv (matches the water plants were actually irrigated with); the raw-effluent values are preserved in Experimental Remarks. Added to REVIEW.md by the batch merge step.

⚠️WARN-MATERIAL fruit dry matter percent-change (Results 3.2 vs Discussion 4.2, p.339-340). Results states “-5.6%” for the aquaponic-vs-hydroponic DM difference; Discussion gives absolute values (5.12% AP, 5.37% HYD) that recompute to -4.66%, not -5.6% — a ~20% relative discrepancy in the stated percentage, too large for simple rounding. Both agree on direction. Recorded the absolute values (5.12/5.37) in trials.csv; the -5.6% claim itself was not entered in any cell.

⚠️WARN-MATERIAL WUE/FWUE denominator conflation (Results 3.4.1 vs Methods 2.7.2, p.339-340). The paper’s headline combined-system water-use figure (“1.55 kg tilapia and 46.1 kg tomato fruit… produced with one m3 fresh water applied to the RAS”) is internally inconsistent: 1.55 kg tilapia/m3 reproduces from 248.84 kg fish yield / 160.4 m3 total RAS fresh water, but 46.1 kg tomato/m3 reproduces from 626.5 kg tomato yield / 13.6 m3 (only the fraction of that water that reached the crop specifically), not from the same 160.4 m3 denominator the sentence implies. Full arithmetic shown in trials.csv Experimental Remarks. Recorded WUE = 46.1 (matching HYD’s own FWUE methodology of yield/treatment’s-own-water-use); the fish-specific 1.55 kg/m3 figure is preserved in remarks.

⚠️WARN-MINOR items (no cell impact): BER percentage-point arithmetic (0.9-0.5=0.4 vs. the Results section’s stated “-0.37%” difference); AP yield restated as “29.2 kg/m2” in Discussion vs. 29.38 kg/m2 in Results (used); harvest-period duration restated as “28 weeks” (=196 d) in a figure caption vs. 197 days stated in Results text (used). All are small, non-interpretation-changing rounding-level restatements.

Plant height / Plant fresh weight column-fit judgment calls: This paper reports total plant/vine LENGTH in metres (10.8-10.9 m) for an indeterminate, vertically-trained-then-lowered greenhouse tomato — a materially different quantity from the standing canopy height the “Plant height” column seems designed for in leafy-vegetable papers elsewhere in this vault. Recorded anyway (unit-converted to cm) with an explicit note, since no better column exists and omitting it would lose real data; the reader should not treat this as comparable to a lettuce/basil “height” figure. Similarly, “Plant fresh weight” (schema unit g/plant) is populated with the paper’s own explicitly-stated AP per-plant total-season yield (13.05 kg/plant, cross-checked against 626.5 kg / 48 plants) rather than a single-harvest fresh weight, since that is the only per-plant mass figure this paper states in text; it represents a cumulative multi-harvest total, not a single fruit-truss weight.

[not reported] fields, grouped:

  • Fish: Fish Category, FCR, SGR, feed protein/N/P/K composition, feed product identity/routine, Fish size initial/final, Fish weight gain per fish, Fish survival rate, Total Feed (kg) — the paper gives only stocking density (kg/m3, three batches), a feeding-rate range (0.7-2.4% body weight/day), and an aggregate “total fish yield” (248.84 kg) for the whole multi-batch, stepwise-stocked/stepwise-harvested RAS; no per-fish weights, feed brand/composition, or survival data anywhere.
  • Water: Aq pH, pHOptimal, Dissolved Oxygen, Water temperature, Water type, Water classification, NO2-N, Average room Temperature — pH is discussed only as literature-cited background RANGES (fish/bacteria 7-9, hydroponics 5.5-6.5) in the Introduction, never as this paper’s own measured or target value, so was not recorded as this paper’s data; DO/water temperature/NO2-N are never measured; only greenhouse air-temperature SETPOINTS (not a measured trial-mean room temperature) are given.
  • Plant: Plant Category, SPAD — the paper never applies a categorical term to the crop, and uses a spectral Chl NDI index instead of a SPAD meter.

NO COLUMN items (full figures in trials.csv Experimental Remarks): pure fish-waste-water nutrient panel (whole production period, Table 3); delivered nutrient-solution P/K/Ca/Mg/S/Na (Table 3, both treatments); system-level fresh-water-consumption figures (160.4 m3 total RAS, 13.6 m3 AP-treatment, 14.2 m3 HYD-treatment); hypothetical whole-greenhouse water-consumption extrapolation (Discussion 4.3.1, not this trial’s actual data); BER absolute rates (HYD 0.9%, AP 0.5%); SSC/SAR relative percentage differences (no absolute values given anywhere outside a bar chart); Chl NDI (a reflectance index, not a tissue analyte or SPAD reading); leaf area per plant (Table 2); target fertiliser recipe and EC setpoint (Methods 2.2.2, identical for both treatments); facility construction volumes (3-chamber pit, storage tank, nutrient-tank-per-gully).

plant_measurements.csv scope: Only lycopene and ß-carotene (Category: biochemistry) were extracted — these are the only plant analytes given as absolute numeric values in running text (Results 3.2 / Discussion 4.2). Fruit dry matter (%) already has a dedicated trials.csv column and was not duplicated here, consistent with the vault convention established in pantanellaAquaponicsHydroponicsProduction2012. SSC and SAR have neither a trials.csv column nor an absolute in-text value (chart-only, Fig. 3) and could not be extracted anywhere; flagged in the note body and above per SCHEMA’s “too valuable to discard” instruction. No SD is given for lycopene/ß-carotene in text (Fig. 3 shows error bars only) — recorded NR per the never-read-a-figure rule, consistent with the pantanella/mourantian precedent for this exact situation.

No water panel excluded beyond what’s noted above — Table 3’s full water-nutrient panel (P/K/Ca/Mg/S/Na, both the raw fish-waste-water and the delivered-solution readings) has no trials.csv column beyond TAN/NO2-N/NO3-N and is preserved in Experimental Remarks (NO COLUMN) rather than discarded, consistent with pantanella/nicoletto/levizou precedent.

Tags judgment call: Tagged Meta/Fish/Tilapia (Nile tilapia, the sole aquaculture species, actively reared and manipulated as part of this study’s own design, unlike nicolettoExtensionAquaponicWater2018’s background-only pangasius) and Meta/Plant/Tomato (new facet — checked TAGS.md-equivalent existing facets in the vault via other tomato papers before reusing “Tomato” as the generic form). Meta/Region/Europe per Germany. Meta/Type/Quasi-experiment used to match the frontmatter type: (new tag value under the existing Meta/Type/ facet, following the same pattern as Meta/Type/Experiment elsewhere in the vault).

New wikilink targets introduced: J. Suhl, D. Dannehl, W. Kloas, D. Baganz, S. Jobs, G. Scheibe, U. Schmidt (no existing author notes found in the vault for this paper’s byline). Reused Nile tilapia (Oreochromis niloticus) and Feed Conversion Rate (FCR) (repurposed as a loose parent link for the fertiliser-use-efficiency callout line since no dedicated “Fertiliser Use Efficiency (FUE)” claim/concept note was found in the vault at extraction time — flagged here rather than silently creating a new orphan; a future pass should confirm whether a dedicated FUE note exists or should be created). Introduced Tomato (Solanum lycopersicum) as the species link, matching the vault’s existing binomial-name convention used for other crops.

PDF quality: Clean text layer throughout (10 pages, standard two-column Elsevier typesetting), fully extractable, no OCR issues.

Source: Suhl et al. - 2016 - Advanced aquaponics Evaluation of intensive tomat.pdf

Data Tables

Structured data extracted from this paper into the vault's trials.csv / plant_measurements.csv datasets. Fields the paper didn't report are omitted. Download the full datasets (measurements).

Trial Parameters

suhlAdvancedAquaponicsEvaluation2016-T1

Fish

FieldValue
FishNile tilapia (Oreochromis niloticus), ‘Red Natural Male Tilapia’ stock (all-male, p.337)
Initial Stock density2.0-2.3 (kg/m3, at each of three stocking batches: 23.1.2015, 12.3.2015, 15.7.2015; mean stocking density during whole production 20.7 kg/m3, fluctuating up to a max of 39.1 kg/m3 — see Experimental Remarks NO COLUMN)
% of body weight0.7-2.4
Fish biomass created (kg)248.84 (total fish yield of the RAS over the multi-batch production period; not a clean final-minus-initial figure — see Experimental Remarks)
Fish trial duration (days)351

Water

FieldValue
Water recycle9 (L/min, NFT hydroponic-loop flow rate, both treatments, continuous 24 h/day; RAS-side pump flow rate not separately stated)
Water volume in the system~12 (m3, whole RAS, stated as ‘around 12 m3’; components: 4 fish tanks 7.2 m3 total net production volume, mechanical filter 1.3 m3, pump sump 2.34 m3, reception tank 0.4 m3)
Daily Water exchange rate6.3
FUE AP54.0
FUE HYD43.7
WUE46.1 (kg tomato m-3 fresh water; aquaponic/DRAPS system — see WARN-MATERIAL in Experimental Remarks)
EC1.8 (dS/m, shared target/setpoint for BOTH treatments, Methods 2.2.2; not stated as a separately-measured trial mean per treatment)
TAN / NH4-N7.0 +/- 9.0 (mg/L, nutrient solution actually delivered to aquaponic gullies, 8-week mean n=8, Table 3 — see WARN-CHECK in Experimental Remarks)
NO3-N157.0 +/- 50.3 (mg/L, nutrient solution actually delivered to aquaponic gullies, 8-week mean n=8, Table 3 — see WARN-CHECK in Experimental Remarks)

Plant

FieldValue
PlantTomato (Solanum lycopersicum L., cv. Pureza)
Details144 plants total across whole greenhouse (2.3 plants/m2, 62.6 m2 net acreage); 48 plants/treatment in two gullies per treatment; planted 22.01.2015 (first truss visible); NFT recirculating gully culture; weekly fruit harvest 14.04.2015-27.10.2015
Days Plant after transplant197 (stated harvest-period duration, 14.04.2015-27.10.2015, Results 3.1; Fig. 2 caption separately gives ‘28 weeks’ = 196 d, WARN-MINOR rounding; planted 22.01.2015 (first truss visible) but no single stated duration figure covers planting-to-last-harvest)
Plants/m22.3
SPAD (aquaponics)NR (not measured; paper instead used a reflectance-based Chl NDI index — see Experimental Remarks NO COLUMN)
Plant height1090 (cm, AP; UNIT CONVERSION ONLY from stated 10.9 m plant length, n=6; NOTE this is total indeterminate-tomato stem/vine length at end of production, not standing canopy height — judgment call, see Extraction notes. HYD = 1080 cm / 10.8 m, ns; both letter ‘a’, Table 2)
Leaf count20.5 +/- 1.3 (AP, number of leaves per plant, n=12, letter ‘a’, ns vs HYD; HYD = 21.0 +/- 0.7, letter ‘a’, Table 2)
Plant fresh weight13050 (g/plant, AP; UNIT CONVERSION ONLY from stated 13.05 kg plant-1, Results 3.4.1, cross-checked: 626.5 kg total AP yield / 48 plants = 13.05, exact match. HYD per-plant figure not stated as text value — NOT DERIVED, see Experimental Remarks)
Plant dry matter5.12 (%, AP fruit dry matter, n=9; HYD = 5.37%; NOTE this is FRUIT dry matter, not whole-plant/leaf dry matter — see WARN-MATERIAL in Experimental Remarks re: stated -5.6% relative change not matching recomputed -4.66%)

System & Setup

FieldValue
System typeDouble recirculating aquaponic system (DRAPS) — two independent recirculating loops (RAS for fish; closed NFT hydroponic loop for plants) connected unidirectionally via a 3-chamber pit and storage tank, with mineral fertilizer added to the fish waste water before delivery to plants (p.336-337); vs. conventional hydroponics (fresh water + mineral fertilizer, matched target nutrient recipe). Functionally a DECOUPLED aquaponic design, though the paper never uses the word ‘decoupled’ — see Extraction notes
Media DetailsNFT (nutrient film technique) in recirculating gutters/gullies; 2 double gullies evaluated (48 plants/treatment each, indeterminate vine training); flow rate 9 L/min, 24 h/day; 300 L nutrient solution tank per gully end
Biological system already in useN (New research facility, first investigations conducted in 2015 (Abstract, Introduction); Methods 2.2.1 explicitly cites ‘activation of the biofilter’ among the reasons stocking density was kept low/fluctuating during the study, indicating the nitrifying/biological system was still being established during data collection, not already mature (p.337))
Iron supplementedY (Both HYD and AP nutrient solutions targeted the same Fe concentration (2.0 mg/L) via mineral fertilizer addition, per the shared target recipe (Methods 2.2.2, p.338))
RemineralizationY (Fish waste water (post 3-chamber-pit, stored in a 1 m3 storage tank) was adjusted with mineral fertilizer to reach the same target nutrient concentration as the hydroponic solution before delivery to plants (Methods 2.1-2.2.2, p.336-338); this is the paper’s core DRAPS/decoupled mechanism)
Climate controlY (Floor-level pipe heating; setpoints 22C day/18C night + ventilation opens at 26C (Feb-Apr), then 17C day/night + ventilation at 21C (mid-Apr-Oct); single-layer energy screen closed 1h after sunset to 1h before sunrise, opened stepwise on a max 2C roof-vs-greenhouse gradient; explicitly built WITHOUT an active cooling unit (‘light version’, Discussion 4.3.1, p.342))
Artificial LightingY (12 high-pressure sodium vapour lamps installed in greenhouse (p.337); triggered between 7 a.m.-8 p.m. when global radiation fell below 20 W/m2 (Methods 2.2.2, p.338))
Nutrient supplementedY (Both treatments’ nutrient solutions adjusted with mineral fertilizer to a shared target recipe (151 N / 37 P / 234 K / 128 Ca / 24 Mg / 110 S / 2.0 Fe / 0.3 B / 0.2 Cu / 1.2 Mn / 0.05 Mo / 0.4 Zn, all mg/L) and EC 1.8 dS/m (Methods 2.2.2, p.338); total fertilizer addition HYD 15.5 kg (100%) vs AP 11.6 kg (74.8%), a 25.2% reduction (Table 4))
EquipmentPhotodiode array spectrophotometer (Pigment Analyzer PA-1101 / MMS1 UV-vis, Carl Zeiss) for Chl NDI; digital refractometer PR101 (ATAGO) for SSC; ICP-OES (iCAP 6300 Duo MFC, Thermo) for P/K/Ca/Mg/S/Na; continuous flow analyser (San++, SKALAR) for NH4-N/NO3-N; water metre for fresh-water volume; SPSS 19.0 for statistics
Control ParametersEC target 1.8 dS/m (both treatments); shared target nutrient recipe (see Nutrient supplementedDetails); heating/ventilation setpoints (see Climate controlDetails); initial fish stocking density target 2.0-2.3 kg/m3; feeding rate 0.7-2.4% body weight/day; RAS daily water exchange 6.3% of 7.2 m3 net production volume
CombinationNile tilapia (Oreochromis niloticus) and tomato (Solanum lycopersicum cv. Pureza); double recirculating aquaponics (DRAPS, functionally decoupled) vs. conventional hydroponics, single paired comparison (no coupled-aquaponics arm in this paper)

Site

FieldValue
RegionEurope
CountryGermany
Lat52.52
Long13.405

Results & Statistics

FieldValue
Measured Unitkg/m2 (cumulative fruit yield); mg g-1 DM (lycopene, beta-carotene); % (fruit dry matter, marketable-fruit rate); m2/plant (leaf area); m (plant length); kg/plant (per-plant yield); kg kg-1 (FUE); kg m-3 (FWUE/WUE); mg/L (water nutrients, Table 3)
Statistic DetailsSPSS 19.0; parametric t-test after confirming normal distribution via Kolmogorov-Smirnov test, otherwise non-parametric Mann-Whitney U test; significance level p<0.05, small letters denote significant differences between treatments
Statistically analysedY
Replicates (n)48 (plants/treatment; primary basis for total yield, BER rate, marketable-fruit-rate statistics — see Experimental Remarks for the several smaller sub-sample n’s used for other measurements: n=10 fruit-quality categorisation, n=12 leaf area/leaf count, n=6 plant length, n=192 Chl NDI, n=9 fruit chemistry, n=8 water-nutrient sampling weeks)
AP29.38
HYD31.64

Experimental Remarks: TRIAL DEFINITION: T1 = the paper’s sole aquaponic treatment, a double recirculating aquaponic system (DRAPS: independent RAS + closed NFT hydroponic loop, connected unidirectionally, fish waste water fertilizer-amended before reaching plants — functionally decoupled, see System type and Extraction notes). Paired control = conventional hydroponics (fresh water + mineral fertilizer, same target nutrient recipe/EC), recorded in the HYD-labelled cells. This paper has only ONE aquaponic arm — no coupled-aquaponics (CAP) treatment is tested or discussed as a design variant here. One row only, and this row still gets this note per SCHEMA.md. | WARN-CHECK TAN/NH4-N and NO3-N, two candidate water-quality bases (Table 3, p.339): the paper reports BOTH ‘pure fish waste water’ (RAS effluent before fertilizer amendment, mean over the whole 351-day fish production period: NH4-N 24.2+/-20.8 mg/L, NO3-N 14.6+/-13.9 mg/L, min-max also given) AND the ‘nutrient solution based on fish waste water’ (the fertilizer-amended solution actually delivered to the aquaponic gullies, mean over an 8-week window 02.07-27.08.2015, n=8: NH4-N 7.0+/-9.0 mg/L, NO3-N 157.0+/-50.3 mg/L). These measure genuinely different things (raw RAS effluent vs. the final delivered irrigation solution) and the paper never states which one should represent ‘the aquaponic system’s water quality’ as a single summary figure. Recorded the delivered-solution values (7.0 NH4-N, 157.0 NO3-N) in the TAN/NO3-N cells, since that is what plants were actually irrigated with and matches the 8-week window during which the plant growth/leaf measurements were also taken; the pure-fish-waste-water values are preserved below under NO COLUMN as the alternative candidate. Added to REVIEW.md by the batch merge step. | WARN-MATERIAL fruit dry matter percent-change, Results 3.2 vs Discussion 4.2 (p.339-340): Results states ‘a significantly lower DM content (-5.6%)… in tomatoes ripened under the influence of aquaponics compared to hydroponics’; Discussion restates the same comparison with absolute values: ‘aquaponics developed fruits with a significantly lower DM content (5.12%) compared to those influenced by hydroponics (5.37%)’. Recomputing (5.12-5.37)/5.37 = -4.66%, not -5.6% — a roughly 20% relative discrepancy in the stated percentage, too large to be simple rounding. Both figures agree on direction (AP lower), so extraction is not blocked; recorded the absolute values (5.12 AP / 5.37 HYD) in the Plant dry matter cells since they are the more specific, directly-usable numbers. The -5.6% claim itself is not carried into any cell. Affects: interpretation of the DM effect’s exact magnitude only. | WARN-MATERIAL WUE/FWUE denominator conflation, Results 3.4.1 vs Methods 2.7.2 (p.339-340): Methods 2.7.2 states the aquaponic FWUE ‘was calculated in the same manner [as hydroponics], but under consideration of the total fresh water use for fish production in the RAS and fish waste water use for tomato production… how many kg fish and tomatoes can be produced with the same volume of fresh water.’ Results 3.4.1 then states ‘it can be derived that in the present study 1.55 kg tilapia and 46.1 kg tomato fruit were produced with one m3 fresh water applied to the RAS.’ Recomputing: 248.84 kg total fish yield / 160.4 m3 total RAS fresh-water consumption = 1.549 kg/m3, matching the stated 1.55 kg tilapia figure (uses the 160.4 m3 total-RAS-water denominator). But 626.5 kg total AP tomato yield / 160.4 m3 = 3.90 kg/m3, NOT 46.1. Instead 626.5 kg / 13.6 m3 (the fish-waste water actually consumed by the 48-plant aquaponic treatment itself, stated separately in 3.4.1: ‘the aquaponic treatment…consumed 13.6 m3, i.e. 8.5% of the available fish waste water’) = 46.05 kg/m3, matching the stated 46.1 kg tomato figure almost exactly. So the sentence describes BOTH figures as ‘per one m3 fresh water applied to the RAS’, but they are actually computed from two different denominators (160.4 m3 for fish; 13.6 m3 for tomato) — an internal inconsistency in how the combined FWUE is narrated, though each individual number is separately reproducible from other stated figures. Recorded WUE = 46.1 kg tomato m-3 (matches HYD’s own FWUE methodology of yield/treatment’s-own-water-use: 677.3/14.2=47.7, matching the paper’s own stated 47.7 kg m-3 for HYD); the 1.55 kg tilapia m-3 fish-specific figure is preserved below under NO COLUMN. Affects: whether the 46.1/1.55 pair should be read as sharing one denominator (they should not); verify against Methods 2.7.2 and Results 3.4.1 directly before citing together. | WARN-MINOR BER (blossom end rot) percentage-point arithmetic, Results 3.1 vs Discussion 4.1 (p.339 vs p.341): Results states AP had ‘a significant smaller rate (-0.37%) of fruit affected by blossom end rot (BER) compared to those ripened in hydroponics’ (a stated difference, no absolute per-treatment numbers given at that point). Discussion later gives absolute figures: HYD marketable-fruit reduction ’(-0.9%)’ i.e. HYD BER=0.9%, AP marketable-fruit reduction ’(-0.5%)’ i.e. AP BER=0.5% (consistent with Results 3.1’s marketable-fruit rates of 99.1% HYD / 99.5% AP). 0.9-0.5=0.4, not exactly the 0.37 stated as the Results-section difference — a small rounding-level mismatch. No trials.csv column exists for BER, so no cell is affected; both absolute values are preserved below under NO COLUMN. | WARN-MINOR AP yield restated, Results 3.1 (29.38 kg/m2, p.338) vs Discussion 4.1 (‘A total yield of 29.2 kg m2 was gained…’, p.341): the Results figure (29.38) is used in the AP cell as the more precise, directly-tabulated Results-section value; the Discussion’s 29.2 is a minor (~0.6%) rounding restatement, no cell impact. | WARN-MINOR harvest-period duration, Results 3.1 (‘the harvest period of 197 days’, p.338) vs Fig. 2 caption (‘produced within 28 weeks’, =196 days, p.340): essentially the same duration (1-day rounding difference); 197 used in the Days Plant after transplant cell as the more specific, explicitly-stated figure. | UNIT CONVERSION ONLY: coordinates 52 deg 31’ 12.025” N, 13 deg 24’ 17.834” E (p.336, Methods 2.1) -> both valid DMS (minutes/seconds <60) -> decimal Lat 52.5200, Long 13.4050. Plant length 10.9 m (AP) / 10.8 m (HYD) -> 1090 cm / 1080 cm recorded in the Plant height cells (see judgment-call note in Extraction notes: this is total stem/vine length for an indeterminate greenhouse tomato at end-of-production, not a standing canopy height). Per-plant AP yield 13.05 kg plant-1 (Results 3.4.1, cross-checked: 626.5 kg total AP yield / 48 plants = 13.05, exact match) -> 13050 g/plant for Plant fresh weight. | NOT DERIVED, left NR: FCR, SGR, feed protein/N/P/K composition, feed product identity, feed routine/frequency (only the 0.7-2.4% daily body-weight ration is stated — no feed brand or composition given anywhere in the paper); Fish size initial/final (only stocking DENSITY in kg/m3 is given, across three staggered batches, never a mean per-fish weight in g); Fish weight gain per fish and Fish survival rate (not reported; the paper gives only the RAS’s aggregate ‘total fish yield’ of 248.84 kg over the whole multi-batch, stepwise-stocked/stepwise-harvested production run, recorded under Fish biomass created since no cleaner final-minus-initial figure is stated); Total Feed (kg) (never stated as a summed figure); Aq pH and pHOptimal (the paper cites literature pH-optimum RANGES for fish/bacteria (7-9) and hydroponics (5.5-6.5) as Introduction background/motivation, but never states its OWN measured or target pH for either loop — not recorded as this paper’s value, to avoid misattributing cited background as this paper’s own data); Dissolved Oxygen, Water temperature of the nutrient solution/fish tanks (never measured/reported); HYD-side per-plant fresh weight (677.3 kg total HYD yield / 48 plants would be derivation, unlike the AP figure which the paper states directly as a text value — left NR). | NO COLUMN: pure fish waste water nutrient panel (whole 351-day production period, Table 3, p.339): NH4-N 24.2+/-20.8 mg/L (0.05-64.1 min-max), NO3-N 14.6+/-13.9 mg/L (bld-42.7), P 8.0+/-5.0 (0.06-15.8), K 30.2+/-16.7 (3.2-69.1), Ca 89.3+/-20.1 (54.2-119.8), Mg 13.9+/-2.4 (9.2-19.3), S 38.5+/-8.9 (15.3-50.4), Na 26.0+/-5.3 (10.8-34.7), all mg/L — see WARN-CHECK above for why the delivered-solution values were used in the TAN/NO3-N cells instead. Delivered nutrient-solution P/K/Ca/Mg/S/Na (8-week mean, n=8, Table 3, p.339): fresh-water/HYD side P 67.2+/-34.9(a), K 60.0+/-34.9(a), Ca 165.0+/-21.1(a), Mg 109.9+/-27.7(a), S 216.8+/-62.5(b), Na 130.1+/-40.7(a); fish-waste-water/AP side P 204.9+/-24.5(b), K 63.9+/-43.5(a), Ca 227.2+/-67.2(b), Mg 106.2+/-28.5(a), S 150.3+/-32.3(a), Na 126.9+/-38.7(a), mg/L; superscript letters denote significance groups within each element row (p<0.05); note S was significantly HIGHER in the HYD/fresh-water solution (+30.7%, text), the only element where HYD exceeded AP. Fresh-water/fertilizer use system-level figures (Results 3.4.1, p.339-340): total fresh-water consumption of the whole RAS 160.4 m3 (23.01.2015 initial stocking to 27.10.2015 last harvest); AP treatment’s own fish-waste-water consumption 13.6 m3 (=8.5% of the ~161 m3 fish waste water generated, from planting 05.02.2015 to 27.10.2015); HYD treatment’s own fresh-water consumption 14.2 m3; RAS total fish yield 248.84 kg -> 1.55 kg tilapia per m3 total RAS fresh water (see WARN-MATERIAL above). Hypothetical whole-greenhouse extrapolation (Discussion 4.3.1, p.342, NOT this trial’s actual data): if all 144 plants (not just the 48-plant treatment) drew fish waste water, they would consume 40.8 m3 (25.3% of the ~161 m3 generated); optimal fish-to-plant ratio calculated as 7.2 m3 RAS volume : 247 m2 greenhouse net-acreage, i.e. 1 m3 fish-tank volume per 34.3 m2 greenhouse. BER (blossom end rot) rates: HYD 0.9%, AP 0.5% of total yield (see WARN-MINOR above). SSC (soluble solids content) and SAR (sugar-acid ratio): only relative percentage differences are stated in running text (SSC -7% AP vs HYD per Results 3.2 / +6.9% HYD vs AP per Discussion 4.2 — the same comparison stated from two different bases, not a contradiction; SAR -5.0% AP vs HYD), with NO absolute values given anywhere outside Fig. 3 (bar chart only) — per the never-read-a-figure rule these are not extracted to any cell; flagging per SCHEMA’s ‘too valuable to discard’ instruction since this is real, paper-reported (if only relative) data with no home in trials.csv or plant.csv. Chl NDI (chlorophyll normalised difference index, a unitless reflectance-based proxy, Table 2, p.340): HYD 0.56+/-0.19(a), AP 0.66+/-0.16(b), n=192, AP +17.9% higher (significant) — not extracted to plant.csv (not a directly-quantified tissue analyte in the biochemistry/mineral/microbiology/proximate sense, more akin to mourantianBasilFunctionalGrowth2023’s excluded PRI/gas-exchange physiological indices) and no trials.csv column fits it either (distinct from the dedicated SPAD column, which this paper does not use). Leaf area per plant (Table 2, p.340): HYD 1.36+/-0.15 m2(b), AP 1.15+/-0.16 m2(a), n=12, AP significantly smaller (-15.4%) — no trials.csv column for leaf area. Kloas et al. (2015) first-prototype DRAPS comparison figures (secondary, not this paper’s own data): total yield 20.5 kg/m2 (vs this paper’s 29.2/29.38, a 1.4-fold increase attributed to continuous nutrient monitoring/optimisation); mean NO3-N 127.7 mg/L in fish water (vs this paper’s much lower 14.6 mg/L pure fish-waste-water mean); NH4-N 5.79 mg/L (vs this paper’s 24.2 mg/L, attributed partly to different measurement points in the water train); fish:plant ratio 5 kg tilapia + 25 kg tomato per m3 fresh water. Roosta and Hamidpour (2011) comparison: equal tomato yields achievable in SRAPS (coupled/single-loop) via foliar macro/micronutrient application, but over only a 22-day trial, called ‘not meaningful’ by this paper’s own Discussion given the short duration. Fertilizer/EC target recipe (Methods 2.2.2, p.338, applies to BOTH treatments identically, not a per-treatment measured value): N 151, P 37, K 234, Ca 128, Mg 24, S 110, Fe 2.0, B 0.3, Cu 0.2, Mn 1.2, Mo 0.05, Zn 0.4, all mg/L; EC 1.8 dS/m. System construction detail: total facility area 196 m2 (14 m2 technical room + 43 m2 RAS/fish farm + 139 m2 Venlo greenhouse); 3-chamber-pit 4.5 m3; storage tank 1 m3; nutrient solution tank 300 L per gully end. Funding: European Commission via INAPRO (grant agreement number 619137).

Plant Measurements

TrialSystemCategoryAnalyteValueUnitSig.Location
suhlAdvancedAquaponicsEvaluation2016-T1APbiochemistryLycopene0.9mg/g DMnsResults 3.2 / Discussion 4.2, p.339-340
suhlAdvancedAquaponicsEvaluation2016-T1HYDbiochemistryLycopene0.99mg/g DMnsResults 3.2 / Discussion 4.2, p.339-340
suhlAdvancedAquaponicsEvaluation2016-T1APbiochemistrybeta-Carotene0.19mg/g DMnsResults 3.2 / Discussion 4.2, p.339-340
suhlAdvancedAquaponicsEvaluation2016-T1HYDbiochemistrybeta-Carotene0.2mg/g DMnsResults 3.2 / Discussion 4.2, p.339-340

Graph View

Backlinks