Improvement of aquaponic performance through micro- and macro-nutrient addition

Metadata

Cite key: ruImprovementAquaponicPerformance2017
Item type: Journal Article
Authors: D. Ru, J. Liu, Z. Hu, Y. Zou, L. Jiang, X. Cheng, Z. Lv
Affiliation: Shandong Key Laboratory of Water Pollution Control and Resource Reuse, School of Environmental Science and Engineering, Shandong University, Jinan 250100, China (Ru, Hu, Zou, Jiang, Cheng, Lv); Environmental Management Committee Office of Dongping Lake and Nansi Lake, Jining 272000, China (Liu)
Journal: Environmental Science and Pollution Research 24(19) (2017) 16328-16335
Date: 07/2017 (print); online 25 May 2017
Date added: 2019-06-29
DOI: 10.1007/s11356-017-9273-1
Funding: National Natural Science Foundation of China (Nos. 21307076 and 41305124); Fundamental Research Funds of Shandong University (Nos. 2014TB003 and 2015JC056)
URL: https://doi.org/10.1007/s11356-017-9273-1
PDF: Ru et al. - 2017 - Improvement of aquaponic performance through micro.pdf

Opinion

A focused, mechanistically interesting study (the struvite/TP story and the chloride-inhibits-nitrification story are both genuinely useful, well-argued findings) undermined by a design the paper never quite owns up to: every table and Methods sentence is written in the singular (“each aquaponic system,” “in each aquaponics”), Table 2’s headline production numbers (feed consumption, biomass gain, FCR, yield) carry no dispersion at all, and the only stated replication (“mean of triple measurements”) reads as analytical/sampling triplicates rather than independent physical systems per treatment. That means the paper’s one-way ANOVA + Tukey test, and by extension every “significant difference” claim in the text, is most plausibly built on pseudoreplicates rather than true biological replicates — a real limitation the authors don’t discuss. The paper also isn’t a hydroponics comparison at all despite living in an aquaponics-vs-hydroponics-shaped literature: both arms are aquaponic (fish present in both), which strains this vault’s AP/HYD schema and forced a judgment call (see Extraction notes). Internally, the numbers mostly check out on recomputation (NUE, fish/plant production percentages, K enrichment ratio), with one clear exception: the FCR “14.3% higher” claim in Results is backwards relative to Table 2’s own printed values. Worth citing for the struvite/TP and Cl—nitrification mechanisms; worth treating cautiously as a “significant difference” data point given the apparent n=1-per-treatment design.

Abstract

Aquaponics is one of the “zero waste” industry in the twenty-first century, and is considered to be one of the major trends for the future development of agriculture. However, the low nitrogen utilization efficiency (NUE) restricted its widely application. To date, many attempts have been conducted to improve its NUE. In the present study, effect of micro- and macro-nutrient addition on performance of tilapia-pak choi aquaponics was investigated. Results showed that the addition of micro- and macro-nutrients improved the growth of plant directly and facilitated fish physiology indirectly, which subsequently increased NUE of aquaponics from 40.42 to 50.64%. In addition, remarkable lower total phosphorus concentration was obtained in aquaponics with micro- and macro-nutrient addition, which was attributed to the formation of struvite. Most of the added micro-nutrients were enriched in plant root, while macro-nutrients mainly existed in water. Moreover, no enrichment of micro- and macro-nutrients in aquaponic products (i.e., fish and plant leaves) was observed, indicating that it had no influence on food safety. The findings here reported manifest that appropriate addition of micro- and macro-nutrients to aquaponics is necessary, and would improve its economic feasibility.

Summary

The authors ran two parallel tilapia-pak choi floating-raft aquaponic systems in Jinan, China for a stated 2-month period: a “Treatment” system dosed weekly with micro-nutrients (B, Mn, Zn, Cu, Mo) and macro-nutrients (K, Ca, Mg) targeting roughly half-strength Hoagland micro-element concentrations, and a “Control” system with no such addition (both received weekly Fe-EDTA). There is no hydroponic arm — both systems are aquaponic, differing only in whether extra nutrients were dosed into the fish-tank water. Nutrient addition raised fish biomass gain by 52.9% and pak choi yield by 83.6%, lowered total phosphate (via struvite precipitation with the added Mg2+) from a rising 19.1 mg/L in Control to a stable ~3.39 mg/L in Treatment, and raised nitrogen use efficiency from 40.42% (Control) to 50.64% (Treatment). Chloride, added as the counter-ion of several supplemented salts, reached 1483 mg/L in Treatment water (vs 133.4 mg/L Control) and is argued to have suppressed nitrifying bacteria (both AOB and NOB, more so NOB), explaining higher nitrite accumulation in Treatment despite its better plant/fish performance. Elemental analysis (ICP-MS) of fish, plant leaf, plant root, microbial biomass and water showed added micro-nutrients concentrating mainly in plant root (a proposed detoxification/retention mechanism) and macro-nutrients remaining mainly in water, with no meaningful enrichment in the edible products (fish flesh, plant leaf) relative to Control — the paper’s food-safety argument. The design, however, appears to use a single physical system per treatment (language throughout is singular, Table 2 has no reported variance), with the stated statistical testing most likely applied to analytical/temporal pseudoreplicates rather than independent systems.

Experiment data

Location: Baihua Park, Jinan, Shandong, China (outdoor, under a rain shed), 36.6767N 117.0617E

Design: Two parallel tilapia-pak choi floating-raft aquaponic systems: “Treatment” (weekly micro- + macro-nutrient addition) vs “Control” (no addition, fish-waste nutrients only); both received weekly Fe-EDTA. No hydroponic arm. Number of independent physical replicate systems per treatment is never stated (see Extraction notes — likely n=1 per treatment based on singular phrasing and Table 2’s lack of variance).

Replicates / n: NR as a system-level count; “mean of triple measurements” stated for analytical results (Statistical analysis, p.16330), of unclear scope (see Extraction notes)

Duration: “2 months” per Methods (p.16329, recorded ~60 days); Figures 1 and 3’s time axes visually appear to run only to ~35-40 days — unresolved, flagged as an observation rather than a formal contradiction (see Extraction notes)

Organisms: Nile tilapia (Oreochromis niloticus) / Pak choi (Brassica campestris subsp. chinensis)

Statistics: One-way ANOVA + Tukey’s comparison test, p<0.05; SPSS v16.0

Fish and plant production: Treatment vs Control — fish biomass increase 1896.9 vs 1240.8 g (+52.9%); pak choi yield 5293.2 vs 2883.2 g (+83.6%); Feed Conversion Rate (FCR) 1.2 vs 1.4 (Table 2; see Extraction notes for a running-text direction error on this figure)

Nitrogen Use Efficiency (NUE): 50.64% (Treatment) vs 40.42% (Control) — no dedicated trials.csv column, recorded in Experimental Remarks

Total phosphate / Struvite formation: Control TP rose to 19.1 mg/L by end of study; Treatment TP held near 3.39 mg/L, attributed to struvite (MgNH4PO4·6H2O) precipitation from added Mg2+

Chloride and nitrification: Cl- reached 1483 mg/L (Treatment) vs 133.4 mg/L (Control); both AOB and NOB gene abundance (Q-PCR) were lower in Treatment, more markedly for NOB, argued to reflect Cl- inhibition

Fish and plant production response to nutrient addition

This paper: Treatment (nutrient-supplemented) aquaponics produced 52.9% more fish biomass (1896.9 vs 1240.8 g) and 83.6% more pak choi yield (5293.2 vs 2883.2 g) than Control over the study period, with a lower (better) FCR (1.2 vs 1.4, Table 2). Both percentage figures reproduce exactly from Table 2’s raw numbers. The FCR figure, however, is where the paper contradicts itself: Results text states feed consumption and FCR were both “higher” in Treatment by 31.5% and 14.3% respectively, but Table 2’s own printed FCR values (1.2 Treatment vs 1.4 Control) show Treatment’s FCR was 14.3% lower, i.e. better feed efficiency, not worse, consistent with the paper’s overall “improvement” narrative. Recorded from Table 2 in trials.csv; full detail in Experimental Remarks.

Compared with:

todo Endut et al. 2009 — found optimum flow rate 1.6 L/min for plant/fish growth vs 2.4 L/min for water quality in water-spinach aquaponics; cited as prior NUE-improvement work (p.16328)
todo Hu et al. 2015 — found fruity plants had higher NUE than leafy plants in floating-raft aquaponics; this paper’s own prior system description and Fe-EDTA protocol source (p.16328-16330)
todo Zou et al. 2016a — nitrifier addition and filler gradation improved NUE of media-based aquaponics by 8.8% and 16.0% respectively (p.16328)
todo Zou et al. 2016b — found higher NUE achievable under acidic conditions (pH 6.0) in media-based aquaponics (p.16328)
todo Roosta and Hamidpour 2011 — foliar micro-/macro-nutrient application on tomato in aquaponics; found (except Cu) significant yield increases, importance order K>Fe>Mn>Zn>Mg>B; this paper’s water-column addition is framed as a more root-absorption-efficient alternative (p.16328-16329, 16333)
todo Dordas 2009 — review on nutrient roles in plant disease/growth; cited generally for the vigor-at-sufficient-nutrient-level mechanism (p.16330)
todo Shilta et al. 2016 — reduced ammonia/nitrite in substrate-based aquaponics significantly increased fish (Etroplus suratensis) growth; cited as a parallel water-quality-to-fish-growth mechanism (p.16330)

Nitrogen budget and Nitrogen Use Efficiency (NUE)

This paper: NUE (% of total N input recovered by fish + plants) rose from 40.42% (Control) to 50.64% (Treatment). The nitrogen budget (Fig. 2) attributes this mainly to lower residual N in the water column of Treatment (~22.02% of input vs ~34.29% in Control), consistent with the plants absorbing more of the available NO3-. Gaseous N losses were similar between systems (Treatment 23.3%, Control 21.67%), within the 5.2-36.0% range the authors cite from other intensive-aquaculture literature. No dedicated schema column exists for NUE; recorded in Experimental Remarks. This is the paper’s headline metric and arguably its most citable single number, flagged here for visibility since it has no cell of its own in trials.csv.

Compared with:

todo Endut et al. 2014 — 5.2-36.0% of nitrogen input lost in gaseous form in intensive aquaculture systems, corroborating this paper’s N-loss range (p.16332)
todo Rakocy et al. 2003 — source of the one-half Hoagland micro-element solution target and the batch/staggered aquaponic cropping comparison; also the N-loss range citation (p.16329, 16332)

Total phosphate, struvite, and food safety

This paper: Control’s total phosphate rose continuously to 19.1 mg/L by the end of the study, while Treatment’s TP stabilized around 3.39 mg/L despite receiving additional phosphorus-adjacent inputs. The authors attribute this to struvite (MgNH4PO4·6H2O) formation, driven by the extra Mg2+ dosed as part of the macro-nutrient mix; a mass-balance check (~7.41 g Mg2+ removed vs a ~6.68 g theoretical calculation) supports the mechanism. Separately, ICP-MS elemental analysis of fish flesh and plant leaf (the two edible aquaponic products) showed no significant enrichment of any added micro- or macro-nutrient in Treatment vs Control, which the authors use to argue the practice does not compromise food safety; added micro-nutrients instead concentrated in plant root (a proposed barrier/retention mechanism) and macro-nutrients remained mainly in the water column.

Compared with:

todo Le Corre et al. 2005 — struvite crystal formation chemistry and the Mg2+-PO4(3-)-NH4+ stoichiometry (Eq. 2) used here (p.16332)
todo Lee et al. 2003 — improved Mg2+ concentration facilitates struvite formation and phosphorus removal in wastewater treatment, cited as the mechanistic basis (p.16331)
todo Clement and Lovell 1994 — suitable tissue micro-nutrient ranges for cultured Nile tilapia/channel catfish (Fe 17.5-41.4, Mn 0.368-2.000, Zn 14.3-53.4, Cu 1.17-3.05 µg/g); used here to confirm fish tissue values in both arms fall within a safe range (p.16333, mis-cited in-text as “Table 2” — see Extraction notes)
todo Clemens et al. 2013 — root-retention/exodermis mechanism for protecting shoot tissue from excess metal ions, cited to explain micro-nutrient accumulation in plant root here (p.16333-16334)

Chloride accumulation and nitrification inhibition

This paper: Chloride, introduced incidentally as the counter-ion of KCl, CaCl2·2H2O and several micro-nutrient salts, reached 1483 mg/L in Treatment water by the end of the study vs 133.4 mg/L in Control. Both ammonia-oxidizing bacteria (AOB) and nitrite-oxidizing bacteria (NOB) gene abundance (Q-PCR of amoA and Nitrobacter/Nitrospira 16S rRNA) were lower in Treatment than Control, with a larger gap for NOB — consistent with NOB being more Cl—sensitive than AOB per the cited literature, and explaining Treatment’s higher NO2- accumulation despite its otherwise-better system performance. The authors note metal sulfate or chelate forms could avoid this Cl- side effect in future designs.

Compared with:

todo Megda et al. 2014 — chloride ion inhibits nitrification in soil, with NO3- decreasing 21.95% as Cl- rose from 260 to 520 mg/kg; the quantitative basis for this paper’s Cl—inhibition argument (p.16333)
todo Radniecki et al. 2009a — heavy metal ion exposure (Zn) effects on Nitrosomonas europaea gene expression; cited alongside 2009b for NOB’s greater sensitivity to inhibitors than AOB (p.16333)
todo Radniecki et al. 2009b — companion study (Cd) to 2009a, same NOB-vs-AOB sensitivity argument (p.16333)
todo Yanbo et al. 2006 — 96-h LC50 of nitrite to Nile tilapia = 44.67 mg/L; used to confirm this study’s NO2- levels did not reach a fish-toxic threshold (p.16331)

Linked claims

Citations to chase

Extraction notes

Type classification judgment call: quasi-experiment. The paper reports a controlled manipulation (defined nutrient-addition treatment vs a defined no-addition control) and states a formal statistical test (one-way ANOVA + Tukey, p<0.05), which would suggest experiment. However, per SCHEMA.md’s decision rule 2 (“Randomised treatments with replication? Yes -> experiment. Treatments without randomisation or replication -> quasi-experiment”), this paper never states randomization of treatment-to-system assignment, and — more importantly — never states the number of independent physical aquaponic systems per treatment. All Methods language is singular (“Each aquaponic system…”, “In each aquaponics, 48 tilapias…”), and Table 2 (the paper’s headline production comparison: water replenishment, feed consumption, biomass increase, FCR, yield) reports a single number per treatment with no dispersion at all, which is what a one-system-per-treatment total looks like, not a replicated mean. Table 1 and Table 3 do report means ± SD, but the only stated basis is “all date [data] presented were mean of triple measurements” (Statistical analysis, p.16330) — most plausibly triplicate analytical/sampling measurements from a single system, not triplicate independent systems (see the REPLICATION/DESIGN note below, carried into trials.csv Experimental Remarks). Given the apparent absence of true system-level replication, quasi-experiment was judged the better fit over experiment, despite the paper’s use of a named statistical test.

Trial structure and schema mismatch (no hydroponic control). This paper has no hydroponic arm at all: “Treatment” and “Control” are both aquaponic systems (fish present, plants nourished by fish-tank effluent in both), differing only in whether extra nutrients were dosed into the water. This does not fit the vault schema’s implicit AP-vs-HYD structure. Per SCHEMA.md’s explicit, literal cell-convention rule — “NA… Use for the entire fish block in a plant-only study, and for the HYD columns when there is no hydroponic control” — HYD-labelled trials.csv cells (FUE HYD, Tissue nitrate HYD, HYD) are recorded NA throughout, and every AP-side/unpaired column (Fish, water quality, FCR, yield, etc.) holds the “Treatment” (nutrient-supplemented) arm’s values, since that is the paper’s actual tested “aquaponic treatment.” The paired “Control” arm’s values are fully preserved in trials.csv Experimental Remarks under a dedicated CONTROL VALUES block rather than being dropped, since they are the paper’s central comparison.

This is a judgment call worth flagging for consistency: a different resolution was used in abbeyBasilOcimumBasilicum2022, where a non-hydroponic (“soilless, fertigated, non-aquaponic”) paired control was recorded directly in the HYD columns rather than NA, on the reasoning that preserving the paired-control structure mattered more than the label being literally hydroponic. This note instead follows SCHEMA.md’s literal NA instruction, reasoned as the stronger/more defensible source (an explicit written rule vs. an inferred precedent from one other paper’s judgment call), and because this paper’s “Control” is unambiguously aquaponics (fish-waste-fed) rather than the abbey paper’s intermediate soilless case. Flagging the inconsistency between these two notes for the user’s awareness — this may be worth resolving as an explicit SCHEMA.md addendum (e.g., “when the paired control is non-hydroponic but still aquaponic, do X; when it’s a non-aquaponic non-hydroponic control, do Y”) rather than leaving it to per-paper judgment calls.

⚠️MATERIAL — FCR direction, Results p.16330 vs Table 2 p.16330. Results text: “The total fish feed consumption and feed conversation ratio of treatment aquaponics were 31.5 and 14.3% higher than that of the control.” Table 2 gives FCR: Treatment 1.2, Control 1.4. Recomputed: feed consumption (2278.5 vs 1732.3 g) is indeed 31.53% higher in Treatment, matching the text. FCR (1.2 vs 1.4) differs by 14.29% in magnitude, matching the text’s “14.3%” — but Treatment’s FCR is lower, not higher, than Control’s. The running text’s “higher” is backwards for FCR (correct for feed consumption, then seemingly reused by sentence-parallelism for FCR). Table 2’s printed values were used directly in the trials.csv FCR cell (Treatment=1.2, Control=1.4 recorded in remarks) since they are the structured, tabulated source and align with the paper’s own “improvement” narrative (better/lower FCR from better water quality). Full detail in trials.csv Experimental Remarks.

⚠️MINOR — mistaken table cross-reference, p.16333. Running text says “According to Table 2, micro-nutrient contents in this study were all in the range” (of Clement & Lovell 1994’s suitable fish-tissue values) — but that elemental data is in Table 3, not Table 2 (which has no elemental data). Confirmed the underlying claim is correct against Table 3; no cell affected, simple citation-number typo.

Duration observation, not formally flagged. Methods states “During the 2 months study period…” (p.16329) — the only explicit written duration, recorded as ~60 days. Figures 1(a-d) and 3(a-b), however, both plot “Time (Day)” on an axis that visually appears to end around day 35-40, not day 60. Per the never-read-a-figure-values rule, this was not treated as a competing written value or given a severity tag (there is no second written duration to conflict with “2 months”) — it is noted here only so a human reviewer can visually check the actual figure axis range if exact duration matters for downstream use.

REPLICATION / DESIGN limitation (not severity-tagged — a methodological caveat, not a numeric contradiction). See detailed reasoning in trials.csv Experimental Remarks. In short: the paper never states a system-level replicate count; Table 2’s headline production numbers show no dispersion at all (consistent with n=1 system per treatment); and the stated “mean of triple measurements” most plausibly describes analytical/sampling triplicates from a single system rather than independent biological replicates. This means the paper’s ANOVA/Tukey significance claims should be read cautiously. Replicates (n) recorded NR (not inferred as 1); Statistically analysed recorded Y (the paper does state a test was used) with this caveat carried in remarks.

[not reported] / NR fields, grouped:

Fish: Fish Category, SGR, N/P/K feed composition (proximate composition given instead: fat/ash/fiber/moisture), Fish size initial/final, Fish survival rate (paper states qualitatively “no fish…died or got illness” but gives no numeric %; recording “100%” was judged too close to fabricating an unstated number), Fish weight gain (only aggregate population biomass increase given, would require dividing by 48 fish — derivation), pHOptimal, Water recycle (L/min), Water type, Water classification, EC, TAN/NH4-N, NO2-N, NO3-N (all reported only as an unlabelled time series in Fig. 1 with no stated trial mean — see NO COLUMN note), FUE AP/HYD, WUE, Average room Temperature, Biological system already in use, pH Buffers, Climate control, Artificial Lighting, Replicates (n).
Plant: SPAD, Plant height, Leaf count, Plant fresh weight (only a total system-level bulk yield is given, no per-plant or per-m2 figure since bed area/plant count aren’t stated — recorded instead in the general AP yield field), Plant dry matter, Tissue nitrate AP/HYD (this paper never measures tissue nitrate, only Ca/Mg/K/Fe/Mn/Zn/Cu).

NO COLUMN items (full detail and values in trials.csv Experimental Remarks): Nitrogen Use Efficiency (NUE, the paper’s headline metric, 50.64% Treatment / 40.42% Control); full nitrogen budget percentages (Fig. 2); total phosphate (TP) values and the struvite mass-balance calculation; chloride accumulation values; AOB/NOB nitrifying-bacteria gene-abundance qualitative comparison (Fig. 3); feed proximate composition beyond crude protein (fat/ash/fiber/moisture); Table 3’s Fish and Microbes elemental panels (excluded from plant.csv, which is scoped to plant analytes only per SCHEMA.md, and has no trials.csv home either) — full values preserved in trials.csv remarks.

plant_measurements.csv scope decision. Table 3 (“Distribution of micro- and macro-nutrients in aquaponics”) reports Ca/Mg/K/Fe/Mn/Zn/Cu content (µg/g) for five compartments (Fish, Plant leaf, Plant root, Microbes, Water) in both Treatment and Control systems. Only Plant leaf and Plant root rows were extracted to plant.csv (Category mineral, 7 elements x 2 tissues x 2 systems = 28 rows), since plant_measurements.csv is explicitly scoped to plant analytes only (SCHEMA.md: “Plant analytes only. Water chemistry does not belong here regardless of how well it fits the column shape”). Fish, Microbes and Water elemental rows from the same table were excluded from plant.csv on the same basis and preserved instead in trials.csv Experimental Remarks (NO COLUMN), since neither file has a proper home for fish-tissue, microbial-biomass, or water-column elemental panels outside the water-quality columns trials.csv already defines (TAN/NO2/NO3 only, no elemental-metals columns).

Because this paper has no hydroponic arm, plant.csv’s System column uses AP for the Treatment (nutrient-supplemented) arm and Control for the paired non-hydroponic control arm — reusing SCHEMA.md’s explicit allowance that System may be “AP, HYD, or another treatment label used in that paper,” per the paper’s own “Control” terminology. Significance is recorded NR throughout Table 3’s rows since no p-value or letter-grouping is printed in the table itself; the one qualitative significance claim in running text (leaf K was “the only” element differing between arms, p.16333) is preserved in that row’s Notes field rather than promoted into Significance, since it isn’t the paper’s own explicit numeric/letter marking.

Recompute check (not a flag — confirms no contradiction): text states leaf K in Treatment was “1.76 times higher” than Control. Table 3: Treatment leaf K 4.22±0.56, Control 1.54±0.44. (4.22-1.54)/1.54 = 1.74, consistent with “1.76 times higher” read as the fractional-increase phrasing common in this kind of running text (i.e., Treatment = Control + 1.76×Control), not a straight multiplicative ratio (which would read 2.74x). No contradiction, not flagged.

Fish Category, Water type, Water classification NR — paper never categorises beyond species name (“similar sizes”) and generic “plastic tanks.”

Plant Category recorded as the paper’s own descriptive phrase, “a popular leafy plant in northern China” (p.16329), per SCHEMA.md’s own-wording rule — no formal categorical term is used in the paper.

Tags judgment call: Tagged Meta/Fish/Tilapia (Nile tilapia, Oreochromis niloticus, the sole aquaculture species, present and manipulated in both arms) and Meta/Plant/Pak-Choi (reusing the existing canonical tag from leeComparativeStudyGrowth2019, rather than inventing a new “Pak choi” spelling). Meta/Region/East-Asia + Meta/Region/China (reusing the paired region+country tag pattern already used in kaburagiAquaponicsUsingSaline2020). No new tag facets introduced.

New wikilink targets introduced: D. Ru, J. Liu, Z. Hu, Y. Zou, L. Jiang, X. Cheng, Z. Lv (no existing vault notes found for these authors; note an existing but unrelated B. Hu wikilink already exists in deerGraftedNongraftedCherokee2023, different person, not a conflict). Nitrogen Use Efficiency (NUE) and Struvite are new concept-note targets, not found elsewhere in the vault. Reused existing canonical forms: Feed Conversion Rate (FCR) (the vault’s most common of three fragmented FCR spellings, per CLAUDE.md’s existing-note-check rule), ANOVA.

PDF quality: Clean, fully extractable text layer throughout (8 pages, standard two-column Springer ESPR typesetting). Tables 1-3 and Figures 1-3 all legible in the source PDF; no OCR issues.

Source: Ru et al. - 2017 - Improvement of aquaponic performance through micro.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

ruImprovementAquaponicPerformance2017-T1

Fish

Field	Value
Fish	Nile tilapia (Oreochromis niloticus)
Initial Stock density	20.25
FCR	1.2
Protein	25
% of body weight	1.0-1.5
Feed routine	Once daily at 9:00 AM; uneaten feed removed 10 min after feeding; feed consumption calculated and recorded daily (p.16329)
Feed regime	Commercial fish feed: 25% crude protein, 3% crude fat, 12% crude ash, 6.5% crude fiber, 2% acid insoluble, 11% moisture; fed at 1.0-1.5% of fish body weight daily (p.16329)
Total Feed (kg)	2.2785
Fish biomass created (kg)	1.8969
Fish trial duration (days)	~60

Water

Field	Value
Water volume in the system	550 (200 L aquaculture unit + 350 L hydroponic unit; excludes 20 L clarifier bucket, p.16329)
Daily Water exchange rate	0 (closed recirculating system; no water exchange performed beyond replenishing evaporation/transpiration/sampling losses, p.16329)
Aq pH	6.93 ± 0.19
Dissolved Oxigen	4.68 ± 0.37
Water temperature	29.6 ± 1.11

Plant

Field	Value
Plant	Pak choi (Brassica campestris L. subsp. chinensis)
Details	Seedlings germinated 2 weeks (per Ako & Baker 2009), transplanted to floating foam raft at 20 plants/m2 (supplier’s guideline); harvested at end of study period (~2 months per text; see duration note in remarks)
Plant Category	Leafy plant (p.16329, paper’s own wording: ‘a popular leafy plant in northern China’)
Days Plant after transplant	~60
Plants/m2	20

System & Setup

Field	Value
System type	Floating raft (foam raft, p.16329)
Media Details	Foam raft (floating), no substrate media beyond raft material stated
Air supplement	Y (Air compressors + gas flow meters maintaining DO above 4.5 mg/L in aquaculture unit (p.16329-16330))
Iron supplemented	Y (Fe-EDTA added weekly to ALL systems (both Treatment and Control) to maintain Fe2+ >=2 mg/L, per method of Hu et al. 2015 (p.16330))
Remineralization	Y (Treatment aquaponics only: weekly macro-nutrients (112.0 g KCl, 183.8 g CaCl2·2H2O, 121.7 g MgSO4) + 200 mL micro-nutrient stock solution (per L: 2.86 g H3BO3, 1.55 g MnSO4·H2O, 0.08 g ZnSO4·7H2O, 0.08 g CuSO4·5H2O, 0.12 g Na2MoO4·2H2O), targeting concentrations similar to one-half Hoagland micro-element solution (Rakocy et al. 2003); Control received none of this (p.16329-16330))
Nutrient supplemented	Y (Core experimental treatment (Treatment arm only): weekly micro-nutrient stock solution (200 mL) + macro-nutrients (KCl, CaCl2·2H2O, MgSO4) added directly to aquaculture-unit water, targeting ~1/2-strength Hoagland micro-element solution; Control received fish-waste nutrients only, no addition (p.16329-16330))
Equipment	Peristaltic pump (clarifier lift); air compressors + gas flow meters; DO meter (HQ30d 53LEDTM, HACH); pH meter (PHS-3C, Leici); HACH TNTplus reaction kits (TNT 830 ammonia, TNT 839 nitrite, TNT 836 nitrate); ICP-MS (Thermo Fisher Scientific) for metals; ion chromatography (ICS-1600, Dionex) for Cl-; analytical balance (FA1004N, Jinghai); LECO TruSpec C/N analyzer; Q-PCR system (LightCycler-480, Roche) for AOB/NOB gene abundance; NanoVue Plus Spectrophotometer (K5500, Kaiao) for DNA quant; PowerSoil DNA Isolation Kit (MOBIO); SPSS v16.0
Control Parameters	DO maintained >4.5 mg/L (air compressors); Fe2+ maintained >=2 mg/L via weekly Fe-EDTA (all systems); Treatment water nutrient-amended weekly to ~1/2 Hoagland micro-element solution strength; feed at 1.0-1.5% body weight/day; initial stocking density 20.25 kg/m3 (both systems)
Combination	Nile tilapia (Oreochromis niloticus) and pak choi (Brassica campestris L. subsp. chinensis) floating-raft aquaponics; micro/macro-nutrient-supplemented Treatment vs standard fish-waste-only Control - both arms aquaponic, no hydroponic-only arm in this study

Site

Field	Value
Region	East-Asia
Country	China
Lat	36.6767
Long	117.0617

Results & Statistics

Field	Value
Measured Unit	g (total system-level feed/biomass/yield); kg/m3 (fish stocking density); mg/L (water TAN/NO2/NO3/TP, time series only); mg/L (DO); dimensionless (pH); °C (temperature); µg/g (elemental content, Table 3); % (NUE)
Statistic Details	One-way ANOVA + Tukey’s test, p<0.05 (SPSS v16.0); Table 1 and Table 3 values are means of triplicate measurements; Table 2 values are single system-level totals with no stated variance; number of independent replicate physical systems per treatment never stated (see REPLICATION/DESIGN note in remarks)
Statistically analysed	Y
AP	5293.2

Experimental Remarks: TRIAL DEFINITION: T1 = the paper’s single tested aquaponic manipulation - weekly micro-nutrient (H3BO3, MnSO4·H2O, ZnSO4·7H2O, CuSO4·5H2O, Na2MoO4·2H2O) and macro-nutrient (KCl, CaCl2·2H2O, MgSO4) addition to a tilapia-pak choi floating-raft aquaponic system (“Treatment”), targeting concentrations similar to one-half Hoagland micro-element solution (Rakocy et al. 2003). Paired control = a second, physically identical tilapia-pak choi aquaponic system receiving NO micro/macro-nutrient addition (“Control”) - i.e. standard fish-waste-only aquaponics. Both systems Fe-EDTA-supplemented weekly to maintain Fe2+ >=2 mg/L. Only one aquaponic manipulation tested in this paper, so one row. | SCHEMA MISMATCH (no hydroponic control): this paper has NO hydroponic arm at all - both “Treatment” and “Control” are aquaponic systems (fish present, plants fed by fish-tank effluent in both). Per SCHEMA.md’s explicit cell-convention rule (“NA… for the HYD columns when there is no hydroponic control”), HYD-labelled cells are recorded NA throughout, and the fish/water/plant-result columns hold the Treatment (nutrient-supplemented) arm’s values as the paper’s “aquaponic treatment.” The Control arm’s paired values are preserved in full below under CONTROL VALUES rather than dropped, since they are the paper’s core comparison. Judgment-call flag: a different approach was taken in abbeyBasilOcimumBasilicum2022, which used the HYD columns to hold a non-hydroponic “soilless” paired control instead of NA, on grounds that the paired-control structure should be preserved even when the pairing isn’t hydroponics. This note follows SCHEMA.md’s literal written NA-instruction instead, since this paper’s “Control” is unambiguously aquaponics (fish-waste-fed), not an intermediate soilless case - flagging the inconsistency between these two vault notes for reviewer awareness/future SCHEMA clarification. | ⚠️MATERIAL FCR direction, Results p.16330 vs Table 2 p.16330. Results text: “The total fish feed consumption and feed conversation ratio of treatment aquaponics were 31.5 and 14.3% higher than that of the control.” Table 2 gives FCR: Treatment 1.2, Control 1.4. Feed-consumption check: (2278.5-1732.3)/1732.3 = 31.53%, matches the stated “31.5% higher” for feed consumption. FCR check: (1.4-1.2)/1.4 = 14.29%, matching the stated “14.3%” in MAGNITUDE, but Treatment’s FCR (1.2) is LOWER than Control’s (1.4), not higher as the sentence claims; the text’s “higher” is backwards for FCR specifically (correct for feed consumption, then mistakenly reused by parallelism for FCR). Table 2’s printed values (1.2/1.4) are used directly in the FCR cell (Treatment=1.2) since they are structured, unambiguous, and consistent with the paper’s own headline direction of improvement (lower/better FCR under nutrient addition, matching higher biomass gain per unit feed). Control’s Table 2 FCR (1.4) is recorded under CONTROL VALUES below. Affects: interpretation of whether nutrient addition helped or hurt feed efficiency (Table 2 says helped; the running-text wording says hurt) - a reader citing only the running-text sentence would get the direction backwards. | ⚠️MINOR mistaken table cross-reference, p.16333. Running text: “According to Table 2, micro-nutrient contents in this study were all in the range [of Clement & Lovell 1994’s suitable fish-tissue values].” The micro-nutrient content data (Fe/Mn/Zn/Cu in fish, µg/g) is actually in TABLE 3 (“Distribution of micro- and macro-nutrients in aquaponics”), not Table 2 (“Performance of different aquaponics”, which has no elemental data). Values checked against Table 3 confirm the claim (Treatment/Control fish Fe/Mn/Zn/Cu all fall within the stated 17.5-41.4 / 0.368-2.000 / 14.3-53.4 / 1.17-3.05 µg/g ranges). No cell affected - simple citation/reference-number typo. | DURATION - unconfirmed observation, not formally flagged (no second WRITTEN value exists to compare against, so not treated as a data contradiction per the never-read-a-figure rule). Methods p.16329 states “During the 2 months study period…” - the only explicit written duration in the paper, recorded as ~60 days (UNIT CONVERSION ONLY, no exact start/end calendar dates given). However, Figures 1(a-d) and 3(a-b) both plot “Time (Day)” on an axis that visually appears to run only to approximately day 35-40, not day 60 - a structural observation about the figures, not a value read off a curve, so not entered as a competing data value or severity-tagged. Flagged here only so a human reviewer can visually check Figures 1/3’s actual axis range against the “2 months” text if exact duration matters downstream. | UNIT CONVERSION ONLY: coordinates “36°40’36″N, 117°03’42″E” (p.16329, Baihua Park, Jinan, China) - both DMS components valid (minutes/seconds <60) -> decimal Lat 36.6767, Long 117.0617 (consistent with Jinan, Shandong, China). Total feed 2278.5 g (Treatment, Table 2) -> 2.2785 kg. Fish biomass increase 1896.9 g (Treatment, Table 2) -> 1.8969 kg. Duration “2 months” -> ~60 days (see DURATION note above). | NOT DERIVED, left NR: Fish size initial/final (paper gives 48 fish, 20.25 kg/m3 initial density, 200 L tank volume, and final densities 29.71/26.46 kg/m3 - a mean per-fish weight is calculable but this is derivation, not recorded); Fish weight gain per fish (only the aggregate population biomass increase is stated - 1896.9 g Treatment / 1240.8 g Control for 48 fish - recorded instead as Fish biomass created; per-fish gain would require dividing by 48, derivation); SGR (initial/final density and duration are present but SGR itself is never stated; computing it would be derivation); Fish survival rate (paper states qualitatively “no fish and plant died or got illness” during the study, p.16330, but never gives a numeric %; translating this into “100%” was judged too close to printing an un-stated number and left NR - the qualitative statement is preserved here instead); N, P, K feed composition beyond crude protein (Methods gives crude protein 25%, crude fat 3%, crude ash 12%, crude fiber 6.5%, acid insoluble 2%, moisture 11% - proximate composition, not N/P/K minerals; protein recorded, rest routed to NO COLUMN below); Replicates (n) - see REPLICATION/DESIGN note below. | REPLICATION / DESIGN note (methodological limitation, not a numeric contradiction so not severity-tagged): Methods never states how many independent physical aquaponic systems (“replicates”) were run per treatment - all language is singular (“Each aquaponic system…”, “In each aquaponics, 48 tilapias…”). Table 2 (water replenishment, feed consumption, biomass increase, FCR, yield - the paper’s headline production outcomes) gives a SINGLE number per treatment with no dispersion at all, consistent with one system per treatment rather than a replicated mean. Table 1 and Table 3 values ARE given as means ± SD, but Methods’ only replication statement is “all date [data] presented were mean of triple measurements” (Statistical analysis, p.16330) - most plausibly triplicate ANALYTICAL/sampling measurements from each single system, not triplicate independent systems. If so, the stated one-way ANOVA + Tukey test (p<0.05) would have been run on technical/pseudo-replicates rather than true biological/system-level replicates for Table 1/3, and Table 2’s headline production comparison would have NO replication or significance test at all (Results narrates Table 2 only as percentage differences, never attaching a p-value to it). Recorded ‘Replicates (n)’ as NR rather than inferring ‘1’, since the paper never states the number explicitly either way; ‘Statistically analysed’ recorded Y since the paper does state ANOVA/Tukey was used, but this design caveat is important context for interpreting that Y. | CONTROL (paired, non-hydroponic) VALUES - no HYD-equivalent slot in this schema, see SCHEMA MISMATCH note above. All from Table 1/2/3 unless noted, p.16330-16334. Water quality: pH 7.08±0.18, DO 4.65±0.41 mg/L, Temperature 29.8±1.05 °C (Table 1). Production: Water replenishment 177.8 L, Fish feed consumption 1732.3 g (1.7323 kg), Fish biomass increase 1240.8 g (1.2408 kg), FCR 1.4, Pak choi yield 2883.2 g (Table 2). Fish elemental content (µg/g, Table 3): Ca 1.04±0.07, Mg 0.04±0.01, K 0.25±0.02, Fe 18.98±3.25, Mn 0.72±0.22, Zn 16.68±4.26, Cu 1.32±0.38 (plant leaf/root values duplicated into plant.csv under System=Control). Initial stocking density 20.25 kg/m3 (same as Treatment); final stocking density 26.46 kg/m3 (vs Treatment’s 29.71 kg/m3). No nutrient/remineralization addition (only Fe-EDTA, same as Treatment). NUE 40.42% (vs Treatment’s 50.64%, see NO COLUMN below). | NO COLUMN: Nitrogen Use Efficiency (NUE) - the paper’s headline metric (Abstract, Results p.16332) - Treatment 50.64%, Control 40.42% (% of total N input recovered by fish+plants); no dedicated trials.csv column exists for this ratio, and it is conceptually distinct from FUE (fertilizer-mass-based) as defined elsewhere in this schema. Nitrogen budget (Fig. 2, not read numerically per the never-read-a-figure rule): Treatment ~22.02% of N input remained in water vs Control ~34.29%; both showed similar gaseous N losses (Treatment 23.3%, Control 21.67%). Total phosphate (TP, Fig. 1d + Results text, p.16332): Control’s TP rose uninterrupted to a final 19.1 mg/L; Treatment’s TP was “maintained at around 3.39 mg/L” - attributed by the authors to struvite (MgNH4PO4·6H2O) formation from the added Mg2+ (~7.41 g Mg2+ removed as struvite vs a ~6.68 g theoretical value, Eq. 2). No TP or struvite column exists. Chloride: Cl- reached 1483 mg/L in Treatment water vs 133.4 mg/L in Control by end of study (added incidentally as the counter-ion of KCl/CaCl2/MnSO4 etc.), discussed as a likely inhibitor of nitrifying bacteria (AOB/NOB, Q-PCR-quantified, Fig. 3 - not read numerically; both AOB and NOB lower in Treatment than Control, more pronounced for NOB). No Cl- or nitrifier-abundance column exists. Feed proximate composition beyond crude protein: 3% crude fat, 12% crude ash, 6.5% crude fiber, 2% acid insoluble, 11% moisture (Methods p.16329; same feed used in both systems). Table 3’s Fish and Microbes elemental panels (µg/g, both systems) - excluded from plant.csv per its plant-analytes-only scope, and no trials.csv column exists for fish/microbial tissue elemental content either; preserved here in full: Fish (Treatment) Ca 0.81±0.09, Mg 0.03±0.00, K 0.28±0.02, Fe 21.52±3.78, Mn 0.78±0.27, Zn 18.38±3.61, Cu 1.22±0.26; Fish (Control) Ca 1.04±0.07, Mg 0.04±0.01, K 0.25±0.02, Fe 18.98±3.25, Mn 0.72±0.22, Zn 16.68±4.26, Cu 1.32±0.38; Microbes (Treatment) Ca 4.07±0.05, Mg 0.33±0.04, K 0.52±0.02, Fe 5476.27±388.28, Mn 518.19±48.02, Zn 2212.71±105.74, Cu 466.70±57.65; Microbes (Control) Ca 3.87±0.12, Mg 0.31±0.02, K 0.46±0.02, Fe 5753.76±385.41, Mn 494.80±42.46, Zn 2461.03±177.90, Cu 484.33±66.79; Water (Treatment) Ca 537.34±13.01, Mg 146.97±1.61, K 642.91±11.51, Fe 6.44±0.12, Mn 1.37±0.01, Zn 2.86±0.04, Cu 0.06±0.00; Water (Control) Ca 28.58±0.23, Mg 14.26±0.17, K 53.63±0.80, Fe 6.28±0.09, Mn 1.42±0.02, Zn 3.05±0.09, Cu 0.08±0.00 (Table 3, p.16334). Recompute check (not a flag, confirms no contradiction): text states leaf K in Treatment was “1.76 times higher” than Control; Table 3 gives Treatment leaf K 4.22±0.56 vs Control 1.54±0.44. (4.22-1.54)/1.54 = 1.74, consistent with “1.76 times higher” read as a fractional-increase statement (not a straight multiplicative ratio, which would be 2.74x) - no contradiction, not flagged. | Fish Category and Water type/classification NR - paper doesn’t categorise beyond species name and “similar sizes”/“plastic tanks”. Plant Category recorded from the paper’s own descriptive phrase “a popular leafy plant in northern China” (p.16329) rather than a formal category term, since none is given. Funding: National Natural Science Foundation of China (Nos. 21307076 and 41305124) and the Fundamental Research Funds of Shandong University (Nos. 2014TB003 and 2015JC056) (Acknowledgements, p.16334). PDF quality: clean, fully extractable text layer throughout (8 pages, standard two-column Springer ESPR typesetting); Tables 1-3 and Figures 1-3 all legible; no OCR issues.

Plant Measurements

Trial	System	Category	Analyte	Value	Unit	Sig.	Location
ruImprovementAquaponicPerformance2017-T1	AP	mineral	Leaf Ca	1.77 ± 0.33	µg/g	NR	Table 3, p.16334
ruImprovementAquaponicPerformance2017-T1	AP	mineral	Leaf Mg	0.28 ± 0.03	µg/g	NR	Table 3, p.16334
ruImprovementAquaponicPerformance2017-T1	AP	mineral	Leaf K	4.22 ± 0.56	µg/g	NR	Table 3, p.16334
ruImprovementAquaponicPerformance2017-T1	AP	mineral	Leaf Fe	53.83 ± 2.81	µg/g	NR	Table 3, p.16334
ruImprovementAquaponicPerformance2017-T1	AP	mineral	Leaf Mn	53.58 ± 5.62	µg/g	NR	Table 3, p.16334
ruImprovementAquaponicPerformance2017-T1	AP	mineral	Leaf Zn	35.92 ± 2.62	µg/g	NR	Table 3, p.16334
ruImprovementAquaponicPerformance2017-T1	AP	mineral	Leaf Cu	2.83 ± 0.62	µg/g	NR	Table 3, p.16334
ruImprovementAquaponicPerformance2017-T1	Control	mineral	Leaf Ca	1.4 ± 0.13	µg/g	NR	Table 3, p.16334
ruImprovementAquaponicPerformance2017-T1	Control	mineral	Leaf Mg	0.34 ± 0.01	µg/g	NR	Table 3, p.16334
ruImprovementAquaponicPerformance2017-T1	Control	mineral	Leaf K	1.54 ± 0.44	µg/g	NR	Table 3, p.16334
ruImprovementAquaponicPerformance2017-T1	Control	mineral	Leaf Fe	54.08 ± 4.28	µg/g	NR	Table 3, p.16334
ruImprovementAquaponicPerformance2017-T1	Control	mineral	Leaf Mn	56.97 ± 6.72	µg/g	NR	Table 3, p.16334
ruImprovementAquaponicPerformance2017-T1	Control	mineral	Leaf Zn	39.48 ± 4.59	µg/g	NR	Table 3, p.16334
ruImprovementAquaponicPerformance2017-T1	Control	mineral	Leaf Cu	1.99 ± 0.54	µg/g	NR	Table 3, p.16334
ruImprovementAquaponicPerformance2017-T1	AP	mineral	Root Ca	1.08 ± 0.25	µg/g	NR	Table 3, p.16334
ruImprovementAquaponicPerformance2017-T1	AP	mineral	Root Mg	0.21 ± 0.03	µg/g	NR	Table 3, p.16334
ruImprovementAquaponicPerformance2017-T1	AP	mineral	Root K	1.5 ± 0.09	µg/g	NR	Table 3, p.16334
ruImprovementAquaponicPerformance2017-T1	AP	mineral	Root Fe	3595.66 ± 370.93	µg/g	NR	Table 3, p.16334
ruImprovementAquaponicPerformance2017-T1	AP	mineral	Root Mn	485.51 ± 58.95	µg/g	NR	Table 3, p.16334
ruImprovementAquaponicPerformance2017-T1	AP	mineral	Root Zn	120.94 ± 25.34	µg/g	NR	Table 3, p.16334
ruImprovementAquaponicPerformance2017-T1	AP	mineral	Root Cu	67.55 ± 7.02	µg/g	NR	Table 3, p.16334
ruImprovementAquaponicPerformance2017-T1	Control	mineral	Root Ca	0.84 ± 0.09	µg/g	NR	Table 3, p.16334
ruImprovementAquaponicPerformance2017-T1	Control	mineral	Root Mg	0.18 ± 0.02	µg/g	NR	Table 3, p.16334
ruImprovementAquaponicPerformance2017-T1	Control	mineral	Root K	1.03 ± 0.11	µg/g	NR	Table 3, p.16334
ruImprovementAquaponicPerformance2017-T1	Control	mineral	Root Fe	1486.87 ± 93.75	µg/g	NR	Table 3, p.16334
ruImprovementAquaponicPerformance2017-T1	Control	mineral	Root Mn	216.98 ± 18.74	µg/g	NR	Table 3, p.16334
ruImprovementAquaponicPerformance2017-T1	Control	mineral	Root Zn	57.77 ± 10.11	µg/g	NR	Table 3, p.16334
ruImprovementAquaponicPerformance2017-T1	Control	mineral	Root Cu	34.71 ± 4.67	µg/g	NR	Table 3, p.16334

Quartz 5

Explorer

Improvement of aquaponic performance through micro- and macro-nutrient addition

Improvement of aquaponic performance through micro- and macro-nutrient addition

Metadata

Abstract

Summary

Fish and plant production response to nutrient addition

Nitrogen budget and Nitrogen Use Efficiency (NUE)

Total phosphate, struvite, and food safety

Chloride accumulation and nitrification inhibition

Linked claims

Citations to chase

Extraction notes

Data Tables

Trial Parameters

ruImprovementAquaponicPerformance2017-T1

Plant Measurements

Graph View

Table of Contents

Backlinks