Quartz 5

Improving water management in European catfish recirculating aquaculture systems through catfish-lettuce aquaponics

Properties1
tagsmeta/type/experiment, meta/region/europe, meta/fish/europeancatfish, meta/plant/lettuce

23 min read

Improving water management in European catfish recirculating aquaculture systems through catfish-lettuce aquaponics

Metadata

Opinion

A useful engineering-style water-budget study rather than a classic aquaponics performance trial: no fish growth data are reported at all (the RAS is treated as a fixed, pre-existing black box), and the “aquaponics” comparison is really a single RAS-effluent hydroponic arm (EC 1.6) benchmarked against three synthetic nutrient solutions (EC 1.6/2.0/3.0). The water-balance methodology (Section 2.2.1, Eqs. 1-6) is careful and well instrumented, but the paper’s own headline evaporation and annual-yield figures are internally inconsistent between the abstract/conclusion and the design-proposal section/Table 4 (see Extraction notes) — worth verifying against the original data before citing either the 329 vs 360 kg/year fish yield or the ~31 vs ~41.6 L/day evaporation figures. The design-scenario in Section 3.3 is a valuable back-of-envelope sizing exercise (FRR-based bed area, staggered production) but is speculative, not something the RAS actually ran.

Abstract

In the context of climate change and population growth, aquaculture plays an important role for food security, employment and economic development. Intensive recirculating aquaculture systems (RAS) allow to treat and recycle fish effluents to reduce waste concentration in outflow water thereby reducing environmental contamination. RAS sustainability may be further improved using aquaponics, a circular productive system in which RAS wastewater is recovered for crop cultivation and recycled back to the fish tanks. In this study, water metabolism of a catfish RAS was assessed and the opportunity to produce lettuce with the RAS effluent was tested. Crop growth and water consumption in aquaponics were compared to those experienced in hydroponics at three nutrient solution concentration (EC of 1.6, 2.0 and 3.0 dS∙m−1), also considering water- (WUE) and nitrogen- use efficiency (NUE). A scenario for converting the RAS in a catfish-lettuce aquaponic system was, then, proposed. The RAS water balance included an input of 555 L∙day−1, out of which 32 L∙day−1 were lost by evaporation from the tubs whereas 460 L∙day−1 were discarded. The lettuce yield, NUE and WUE in aquaponics were respectively 20.3%, 22.3% and 20.6% lower than those obtained in hydroponics. Best performances in hydroponics were achieved with EC of 2.0 dS m−1. No difference in term of water consumption arose between the treatments, with average water use of 46 mL∙plant−1∙day−1. Considering the current RAS productivity of 329 kg year−1, a 10 m2 raft system hosting 160 lettuces would satisfy the nitrogen filtration demand. Once closed the water loop between the two productive sub-units, the current water input of 532 L∙day−1 could be reduced to the amount needed to replace the water lost by evaporation (50 L∙day−1) and the RAS water output would decrease from 555 to 103 L∙day−1.

Summary

Researchers at a pre-existing experimental catfish (Silurus glanis) RAS in Soest, Germany measured the system’s full daily water balance (well-water intake, evaporation from biofilter/sedimenter/tanks/sump, fish water assimilation, discharge) over June 2016-July 2017. Separately, they grew Salanova lettuce in small deep-water-culture boxes across three seasonal experiments, comparing RAS effluent (“Aqua”, EC 1.6) against three synthetic hydroponic nutrient solutions (EC 1.6, 2.0, 3.0), tracking yield, water-use efficiency (WUE) and nitrogen-use efficiency (NUE). Lettuce grown on RAS effluent yielded less, and had lower NUE and WUE, than hydroponics — attributed to nutrient imbalances (low K, P, Fe, Mn; high Na and bicarbonate) in the fish-derived nutrient solution rather than to differences in water consumption, which did not differ between treatments. No fish growth trial was conducted; the RAS’s fish stock was treated as a fixed background system. Using the measured water flows and a literature-based Feeding-Rate-Ratio method, the authors then proposed a hypothetical design for closing the water loop by adding a 10 m2 lettuce raft bed, projecting a large reduction in well-water intake and wastewater discharge. The paper is best read as a water-engineering feasibility study for retrofitting an existing RAS into aquaponics, not as a lettuce-agronomy trial in its own right.

Experiment data

  • Location: Experimental RAS, South Westphalia University of Applied Sciences (SWUAS), Soest, Germany (p.761)
  • Design: Two components: (1) water-balance characterisation of a pre-existing 3-tank catfish RAS; (2) a lettuce cultivation trial with 4 water-source treatments (Aqua = RAS effluent EC 1.6; Hydro 1.6, 2.0, 3.0 = synthetic nutrient solutions), repeated across 3 seasonal experiments, in 8 x 30 L DWC boxes
  • Replicates / n: [not reported] — “8 boxes” and “4 treatments” are both stated but boxes-per-treatment is never labelled as n or replicates
  • Duration: Water balance: June 2016-July 2017. Lettuce experiments: exp1 Jul-Aug 2016, exp2 Sep-Oct 2016 (last days of Sept-Oct), exp3 Jun-Jul 2017. No stated day-count per cycle for the actual conducted experiments
  • Organisms: European catfish (Silurus glanis) / Lactuca sativa cv. Salanova multileaf
  • Statistics: Two-way ANOVA (95% confidence) on growing cycle x water source; means separated by LSD test, P<=0.05; CoStat software
  • Water Use Efficiency (WUE): Aqua lowest in every experiment — 69.8, 98.5, 48.1 g FW L-1 H2O for exp 1, 2, 3 respectively
  • Nitrogen Use Efficiency (NUE): Aqua 0.34 g FW mg-1 N, significantly lower than Hydro 1.6 & 2.0 (avg 0.5), not different from Hydro 3.0 (0.33)
  • Lettuce yield: Aqua 92.3 g fresh weight/plant (lowest); Hydro 2.0 and 3.0 pooled mean 120.3 g/plant (highest, not different from each other)

RAS water balance

This paper: Daily water flow through the system ≈555 L/day: well water input 532 L/day; evaporation from tubs (biofilter+sump+fish tanks+sedimenter) ≈31.1 L/day (Table 2, but see ⚠️CHECK below for a conflicting 41.6/41.9 L/day figure used later in the paper); water discharge 460 L/day; fish water assimilation 0.70 L/day. Water use for catfish: 0.59 m3/kg.

Compared with:

  • Eurofish (2009) — tilapia RAS water use 0.24 m3/kg, less than half this study’s catfish figure. (p.763, secondary)
  • todo Verdegem et al. (2006) — super-intensive RAS 0.5-0.7 m3/kg; extensive pond up to 45 m3/kg. (p.763)
  • todo Bregnballe (2015) — average RAS water use ~1 m3/kg. (p.763)

Lettuce yield, WUE and NUE (aquaponics vs hydroponics)

This paper: Lettuce grown on RAS effluent (Aqua) had the lowest fresh weight (92.3 g/plant), NUE (0.34 g FW/mg N) and WUE (69.8-98.5-48.1 g FW/L H2O across the three seasonal experiments) of the four treatments. Hydro 2.0 gave the best combination of yield, NUE and WUE without the diminishing returns seen at Hydro 3.0 (higher EC did not increase yield further — “luxury consumption”). Dry matter % and leaf C:N ratio were significantly higher in aquaponics than hydroponics (figure-only values, not extracted numerically — see Extraction notes), attributed to P-deficiency-driven carbohydrate accumulation and Na+-related osmotic stress. Water consumption itself did not differ significantly between treatments (~46 mL/plant/day average).

Compared with:

  • todo El Sayed and Samir (2015) — lower aquaponics yield vs hydroponics, same direction as this study. (p.763)
  • todo Johnson et al. (2017) — lower aquaponics yield vs hydroponics, same direction as this study. (p.763)
  • todo Pantanella et al. (2012) — no yield difference between aquaponics and hydroponics, contrasts with this study. (p.763)
  • todo Delaide et al. (2016) — no yield difference between aquaponics and hydroponics; also proposed supplementing aquaponic solution with synthetic minerals, a strategy this paper’s Conclusion recommends testing. (p.763, p.767)
  • todo Rakocy et al. (2007) — low K+, P+, Fe2+, Mn2+ in aquaponic solutions is a general limitation of fish-feed-derived nutrients. (p.763)
  • todo Nozzi et al. (2018) — P deficiency reduces root N uptake, inducing non-structural carbohydrate accumulation (explains the higher DM% in aquaponics here). (p.764)
  • todo Stefanelli et al. (2011) — lettuce NUE plateaus at 150 mg/L N with no further gain to 2400 mg/L, consistent with this study’s NUE decline at the highest EC (Hydro 3.0, 243 mg/L N). (p.764)
  • todo Chabite et al. (2017) — WUE values of 29.7-142.9 g/L for lettuce on different nutrient solutions, consistent range with this study’s 48.1-141.6 (Fig. 4). (p.765)

Design proposal for a closed-loop catfish-lettuce aquaponic system

This paper: Using the Feeding Rate Ratio (FRR) method (Rakocy et al., 2006; target 80 g feed/m2/day), the authors calculate that a 9.63 m2 (~10 m2) DWC lettuce bed, at 16 plants/m2, would filter the nitrogen load of the current RAS’s estimated 329 (or 360, see ⚠️MATERIAL below) kg/year catfish production. Closing the loop would cut well-water input from 532 to 49.64 L/day and total water output from 554.9 to 102.53 L/day (Table 4). This is a hypothetical sizing exercise, not an implemented or tested system — no data exist yet on whether the proposed integrated system would actually perform as calculated, and the yield/WUE/NUE penalties observed in the actual Aqua treatment (Section 3.2) suggest the proposed system’s lettuce productivity may be overestimated if nutrient supplementation is not added.

Compared with:

  • todo Rakocy et al. (2006) — source of the FRR method and the 60-100 g/m2/day optimum range used here. (p.762)

Linked claims

Citations to chase

  • todo El Sayed, G.K., Samir, A.A. (2015) — lower aquaponics vs hydroponics lettuce yield, opposite-season/system comparison
  • todo Johnson, G.E. et al. (2017) — lettuce comparison across spring water, hydroponic, and flow-through aquaponic systems
  • todo Pantanella, E. et al. (2012) — aquaponics vs hydroponics lettuce production and quality, no yield difference found
  • todo Delaide, B. et al. (2016) — complemented aquaponic solution outperforming hydroponics; synthetic mineral supplementation strategy
  • todo Rakocy, J. et al. (2007) — organic waste from aquaculture as a source of inorganic nutrients for hydroponics
  • todo Nozzi, V. et al. (2018) — nutrient management in aquaponics, three cultivation approaches compared
  • todo Stefanelli, D. et al. (2011) — nitrogen availability and lettuce leaf quality/antioxidant capacity
  • todo Chabite, I.T. et al. (2017) — nutrient-solution management based on N-use efficiency for lettuce
  • todo Rakocy, J.E., Masser, M.P., Losordo, T.M. (2006) — Feeding Rate Ratio method and DWC design parameters
  • todo Verdegem, M., Bosma, R., Verreth, J. (2006) — water use benchmarks across aquaculture system intensities
  • todo Bregnballe, J. (2015) — RAS water use guide (FAO/EUROFISH)

Extraction notes

⚠️MATERIAL — Fish/system annual yield: stated as 329 kg/year four times in running text (Abstract; Results 3.1, p.763; Discussion 3.3, p.765 and p.766) but as 360 kg/year twice in Table 4 (p.766), for both the current and proposed scenarios. Not a rounding difference (~9.4% apart). 329 recorded as the defensible figure (four independent restatements vs. two in a table that looks like a transcription slip); not entered into any trials.csv cell since no annual system-yield column exists — see NO COLUMN in the trial row’s Experimental Remarks.

⚠️CHECK — Evaporation losses from the RAS tubs: Abstract states “32 L∙day−1”, the Conclusion states “31 L∙day−1”, and Table 2’s own components (biofilter 17.7 + sump 6.7 + fish tanks 5.8 + sedimenter 0.9 L/day) sum to 31.1 L/day — all three agree. But Section 3.3 states “Qeva_f = 41.6 L∙day−1” and Fig. 5’s caption gives the formula “Qeva_f = Qst+Qft+Qb+Qs = 41.9 L d−1”, which does not match 31.1 when Table 2’s own values are substituted. The ~10.5 L/day gap closely matches Table 2’s separate “Undetermined losses” line (10.45 L/day): 31.1+10.45=41.55, near both 41.6 and 41.9. Likely the design-scenario figure silently includes undetermined losses while the abstract/conclusion figure does not — but the paper never states this, and Fig. 5’s formula doesn’t literally include that term, so it is recorded UNCLEAR/unresolved rather than silently reconciled. No trials.csv column exists for this value; full detail in the trial row’s Experimental Remarks (NO COLUMN).

⚠️MINOR — Wording of the biofilter-evaporation percentage comparison (p.763) is inverted relative to what the numbers show (see Experimental Remarks); does not change any recorded value.

⚠️MINOR — Design-proposal lettuce cycle length stated as both “4 weeks” (28 days, Section 2.2.3) and “30-day crop cycle” (Section 3.3) for the same hypothetical scaled-up scenario; the actual conducted experiments never state a day-count, so no cell is affected.

[not reported] / [unclear] fields, grouped:

  • Fish growth data: Fish size initial/final, FCR, SGR, survival rate, weight gain, trial duration, biomass created, Total Feed (kg), Replicates (n) — no fish-growth trial with stocking/harvest data is reported; the RAS is a fixed pre-existing background system with only standing-stock size classes (50/300/400 g fingerlings) given.
  • Water engineering: Daily Water exchange rate (%), Water recycle (L/min), Water temperature, Dissolved Oxygen, Average room Temperature — none stated in a form matching the schema’s units.
  • Plant/trial detail: Plant Category, Fish Category, Water classification, Days Plant after transplant, Plants/m2, SPAD, Plant height, Leaf count, Plant dry matter, Tissue nitrate AP/HYD, Lat/Long — not stated for the actual conducted experiment (some appear only for the hypothetical design proposal, which is a different, non-empirical scenario).

NO COLUMN items (see trial row’s Experimental Remarks for full detail): NUE values (no dedicated column); Hydro 2.0/3.0 pooled fresh weight (120.3 g/plant); nutrient-solution C:N ratio (Table 1); leaf-tissue C:N ratio (figure-only, Fig. 2C, not extracted); full Table 2 water-balance component breakdown; the entire hypothetical design-proposal scenario (FRR, bed area, staggered production schedule, projected annual outputs); per-experiment climate data; per-experiment WUE breakdown (recorded as a range in the WUE cell, but per-treatment-per-experiment detail lives here).

Water panel: No water panel was excluded from plant_measurements.csv — this paper’s only leaf-tissue analyte (C:N ratio via LECO CNS2000) is reported solely in Fig. 2C with no accompanying text/table number, so it could not be extracted under the “never read a value off a figure” rule (plant.csv has 0 rows; see below).

plant_measurements.csv = 0 rows. The only tissue-level plant analyte in this paper (leaf N/C content, presented as the C:N ratio) is shown only in Fig. 2C with no corresponding number in the text or a table; fresh weight and dry matter are not analytes (they belong in trials.csv, and dry matter itself is also figure-only, Fig. 2B). No biochemistry, additional mineral, microbiology, or proximate analyses were performed on plant tissue.

Source: Calone et al. - 2019 - Improving water management in European catfish recirculating aquaculture systems through catfish-let.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

caloneImprovingWaterManagement2019-T1

Fish

FieldValue
FishEuropean catfish (Silurus glanis), fingerlings (50, 300 and 400 g stocked separately per tank)
Initial Stock density90-100
Protein54
% of body weight4% (juveniles) to 1.5% (adults)
Feed regimeCommercial diet pellet, 2 mm diameter for small-size fish and 4.5 mm for large-size fish (Aller Aqua Group, Christiansfeld, Denmark), 54% protein, 20% fat; feeding rate 4% of biomass/day (juveniles) to 1.5%/day (adults) (p.761)

Water

FieldValue
Water volume in the system3300
Water typeRAS effluent (well-water sourced) (p.761-762)
Aq pH7.3 (initial, Table 1); 7.52-7.76 (end-of-cycle range, Table 3, cycles 1-3, no trial mean reported)
WUE48.1-98.5 (Aqua, range across exp 1-3; g FW L-1 H2O; no overall trial mean reported, Fig.4/text p.764)
EC1.6 (Aqua initial, Table 1); 1.3-1.85 (end-of-cycle range, Table 3, cycles 1-3)
TAN / NH4-N1.26 (initial, converted from 0.09 mmol/L, Table 1); <1.4 (end-of-cycle, all 3 cycles, Table 3)
NO3-N103.1 (initial, converted from 7.36 mmol/L, Table 1); 67.9-128.9 (end-of-cycle range, converted from 4.85-9.2 mmol/L, Table 3)

Plant

FieldValue
PlantLettuce (Lactuca sativa L. cv. Salanova multileaf)
DetailsTransplanted at sixth unfolded leaf stage; grown in 30 L boxes with polystyrene raft (10 plants/box); compared across 3 seasonal experiments (exp1 Jul-Aug 2016, exp2 Sep-Oct 2016, exp3 Jun-Jul 2017) (p.761-762)
Plant fresh weight92.3 (Aqua, mean across exp 1-3, no SD given in text, p.763)

System & Setup

FieldValue
System typeDeep Water Culture (DWC) (p.762)
Media DetailsPolystyrene sheet with 10 holes per 30 L box, floating raft (p.762)
Biological system already in useY (Existing RAS built in 2015 at South Westphalia University of Applied Sciences (SWUAS), Soest, Germany, including nitrifying biofilter, in continuous operation before this study’s data collection (June 2016-July 2017) (p.761))
Air supplementY (All 8 lettuce boxes connected to an air pump for nutrient-solution oxygenation (p.762, Section 2.2.2))
Iron supplementedY (Hydro nutrient solutions formulated with Fe (21.33-40 µmol/L depending on EC treatment, Table 1); Aqua (RAS effluent) unsupplemented, native Fe 0.28 µmol/L (Table 1, p.762))
pH BuffersY (Bicarbonate buffer periodically added to the RAS circulating medium to counteract pH drops from bacterial nitrification (p.763, Section 3.2.2))
Climate controlN (Explicitly no automated temperature-management system during exp 1-3, so extreme temperatures could not be prevented (p.763, Section 3.2.1); design proposal (Section 3.3) suggests installing an environment control system for future year-round production, confirming none existed during the reported experiment)
Nutrient supplementedY (Hydro 1.6/2.0/3.0 = formulated hydroponic nutrient solutions at increasing EC (1.6/2.0/3.0 dS/m, Table 1); Aqua = unsupplemented RAS effluent at native EC ~1.6 dS/m (p.762, Table 1))
EquipmentWater meter (well-water consumption); propeller anemometer (air flow speed); humidity/temperature sensors (10 s logging interval, fish tanks/sedimenter/sump); LECO CNS2000 elemental analyzer (leaf C:N); air pump (nutrient-solution oxygenation); graduated scale (evapotranspiration/water-level reading); CoStat software (statistics) (Sections 2.2.1-2.3)
Control ParametersEC of nutrient solution (4 levels: RAS-effluent ~1.6, Hydro 1.6, 2.0, 3.0 dS/m); growing cycle/season (3 repeated experiments, summer/autumn/summer)
CombinationEuropean catfish (Silurus glanis) RAS effluent used to grow lettuce (Lactuca sativa cv. Salanova) in DWC hydroponics vs three synthetic hydroponic nutrient solutions at increasing EC; no true fish-growth trial reported, water balance + lettuce yield/WUE/NUE comparison only

Site

FieldValue
RegionEurope
CountryGermany

Results & Statistics

FieldValue
Measured Unitg plant-1 (fresh weight); g FW L-1 H2O (WUE); g FW mg-1 N (NUE, NO COLUMN); % (DM, figure only)
Statistic DetailsTwo-way ANOVA (95% CI) testing growing-cycle x water-source interaction; means separated by LSD test at P<=0.05; CoStat software (Section 2.3, p.762)
Statistically analysedY
AP92.3

Experimental Remarks: TRIAL DEFINITION: T1 = aquaponic treatment ‘Aqua’ (RAS effluent, EC 1.6 dS/m) growing lettuce in DWC boxes, averaged/compared across three repeated seasonal experiments (exp1 Jul-Aug 2016, exp2 Sep-Oct 2016, exp3 Jun-Jul 2017). Paired control recorded in HYD-adjacent cells = Hydro 1.6 (hydroponic nutrient solution at the SAME EC as Aqua, the closest matched comparison); Hydro 2.0 and Hydro 3.0 are additional hydroponic arms at higher EC, not given their own row because they are variations of the HYDROPONIC control rather than distinct aquaponic treatments (schema calls for one row per aquaponic treatment; there is only one aquaponic arm in this paper). Hydro 2.0/3.0 values are recorded in these remarks instead. This paper also reports a pre-existing RAS’s water balance (Table 2) and a hypothetical design-scenario for scaling up to a full catfish-lettuce aquaponic system (Section 3.3, Table 4) — these are NOT an empirical fish-growth trial and contain no stocking/harvest weights; all fish-block growth cells reflect the paper’s SILENCE on an actual fish growth trial, not a fish-only or NA condition, hence NR rather than NA. | WARN-MATERIAL Fish/system annual yield: Abstract (‘current RAS productivity of 329 kg year-1’), Results 3.1 (p.763, ‘329 kg year-1’), and Discussion 3.3 (p.765, ‘estimated potential catfish production of the current RAS is 329 kg year-1’; p.766, ‘yearly production…will consist of 329 kg of fish meat’) state 329 kg/year FOUR times in running text. Table 4 (p.766) states ‘Yield: 360 kg year-1 fish meat’ for BOTH the current and proposed scenarios. 329 and 360 cannot both be the same annual yield figure; the ~9.4% gap is too large for rounding. 329 is recorded as the defensible value given its four independent restatements across abstract/results/discussion versus Table 4’s two instances of 360, which appear to be a transcription error carried through the table. Not recorded in any cell (no trials.csv column matches an annual system-level yield; see NO COLUMN below) but flagged here because it affects the paper’s headline sustainability claim. | WARN-CHECK Evaporation losses from RAS tubs (Qeva_f) — NO COLUMN, flagged for visibility per the ‘water management’ framing of this paper: Abstract (p.759) states ‘32 L day-1 were lost by evaporation from the tubs’; Conclusion (p.766) states ‘evaporation losses from the system’s tubs account for 31 L day-1’; Table 2 (p.763) components (biofilter 17.7 + sump 6.7 + fish tanks 5.8 + sedimenter 0.9) sum to 31.1 L/day — these three cluster together (~31-32 L/day) and are mutually consistent. By contrast, Section 3.3 (p.765) states ‘Qeva_f = 41.6 L day-1’, Fig.5’s caption (p.766) gives the explicit formula ‘Qeva_f = Qst + Qft + Qb + Qs = 41.9 L d-1’, and Table 4 (p.766) repeats ‘Evaporation losses (Qeva): 41.6’ for the current scenario — but plugging Table 2’s own Qst/Qft/Qb/Qs values into that exact formula gives 31.1, not 41.9/41.6. The ~10.5 L/day gap closely matches Table 2’s separately listed ‘Undetermined losses (10.45 L day-1)’ (31.1+10.45=41.55, within 0.05-0.35 of both 41.6 and 41.9), suggesting the design-scenario figure silently folds ‘undetermined losses’ into ‘evaporation losses’ while the abstract/conclusion figure does not, without the paper ever stating this. No evaporation-loss column exists in trials.csv; the clearer-basis reading (31.1 L/day, directly summable from Table 2’s own component values) is offered here as the value tied to raw measurements, with 41.6/41.9 as the design-scenario’s broader (and internally 41.6-vs-41.9 inconsistent) figure. Unresolved — verify before citing either figure. | WARN-MINOR Biofilter evaporation percentages: Results (p.763) states biofilter evaporation (17.7 L/day) ‘resulted 62.5%, 67.2% and 95.2% higher than’ sump/fish-tanks/sedimenter (6.7/5.8/0.9 L/day) — arithmetically these percentages are actually (17.7-X)/17.7, i.e. how much LOWER each of the other three is relative to the biofilter, not how much higher the biofilter is relative to them (which would be 164%/205%/1867%). Wording direction is inverted but Table 2’s underlying values (17.7/6.7/5.8/0.9) are used directly and are unaffected; no extracted cell changes. | WARN-MINOR Lettuce growing-cycle length in the DESIGN PROPOSAL (Section 3.3, not the conducted experiment): Section 2.2.3 (p.762) states ‘a productive growing cycle of 4 weeks was considered’ (28 days) while Section 3.3 (p.765) states ‘Assuming a 30-day crop cycle’ — both describe the same hypothetical scaled-up scenario and conflict by 2 days. Does not affect the T1 row: the actual conducted lettuce experiments (exp 1-3) never state an explicit cycle-length in days (only approximate calendar-month windows), so ‘Days Plant after transplant’ is recorded NR regardless of this design-only discrepancy. | UNIT CONVERSION ONLY: NO3-N and TAN/NH4-N converted from Table 1/Table 3’s reported mmol/L ionic concentrations using N molar mass 14.007 g/mol (mg/L N = mmol/L x 14.007). Aqua initial (Table 1): NO3- 7.36 mmol/L -> 103.1 mg/L NO3-N; NH4+ 0.09 mmol/L -> 1.26 mg/L NH4-N. Aqua end-of-cycle (Table 3, cycles 1-3): NO3- 9.2/5.7/4.85 mmol/L -> 128.9/79.8/67.9 mg/L NO3-N (range 67.9-128.9); NH4+ ‘<0.1’ mmol/L all three cycles -> ‘<1.4’ mg/L NH4-N. Water volume 3 x 1100 L fish tanks = 3300 L, matches the Highlights’ own stated ‘3.3 m3’ rearing system (p.759) — not summed with sedimenter (460 L) or heater (800 L) since the Highlights explicitly anchor the 3.3 m3 figure to the fish-tank rearing volume alone. | NOT DERIVED, left NR: Total Feed (kg) - only a % of biomass/day feeding rate given, no cycle total; Fish size initial/final, FCR, SGR, Fish survival rate, Fish weight gain, Fish trial duration, Fish biomass created for a discrete trial - no fish-growth trial with stocking/harvest weights is reported anywhere in this paper (Section 2.1 only describes the pre-existing RAS’s standing stock by size class: 50/300/400 g fingerlings farmed ‘separately’ in the three tanks, with no growth-over-time data); Daily Water exchange rate (%) - well water input (532 L/day) and system volume (3300 L fish-tank-only, or more if sedimenter/heater included) are both given but the paper never states an exchange-rate percentage, and computing one would require choosing a volume basis the paper doesn’t specify; Water recycle (L/min) - only daily volumetric flows (L/day) given, no L/min recirculation rate stated; Replicates (n) - ‘8 boxes’ and ‘4 treatments’ are both stated (Section 2.2.2, p.762) but the paper never states boxes-per-treatment as a labelled ‘n’ or ‘replicates’, so n is not recorded (8/4=2 would be derivation); Days Plant after transplant / Plants/m2 for the ACTUAL experiment - only the hypothetical design-scenario (Section 3.3) states 28-or-30 days and 16 plants/m2, not the conducted exp 1-3; Average room Temperature - only per-month min/avg/max air temperatures per individual experiment are given (Section 3.2.1), no single trial-mean room temperature. | NO COLUMN: NUE (Nitrogen Use Efficiency, g FW mg-1 N) - no trials.csv column exists for this metric. Aqua 0.34 (significantly lower than Hydro 1.6 & 2.0 average 0.5, but NOT significantly different from Hydro 3.0’s 0.33) (Fig 3, text p.764, Section 3.2.4). Hydro 2.0 and Hydro 3.0 fresh-weight: pooled mean 120.3 g/plant (text p.763), not significantly different from each other but both significantly higher than Aqua (92.3 g/plant); Hydro 1.6’s own specific fresh-weight number is only visible in Fig 2A (figure-only, not extracted). Nutrient-solution C:N ratio (Table 1): Aqua 0.083 vs. average of the three Hydro treatments ~0.0093 (8.7-fold higher in Aqua, as stated p.764) - this is the nutrient-SOLUTION C:N, distinct from the LEAF-TISSUE C:N ratio shown only in Fig 2C (figure-only, not extracted). Full water balance breakdown (Table 2, p.763): well water 532, air humidity in 22.9+/-3.0, air humidity out 52.6+/-1.41, biofilter evap 17.7+/-1.52, sump evap 6.7+/-0.26, fish-tank evap 5.8+/-0.16, sedimenter evap 0.9+/-0.09, fish water assimilation 0.70, water discharge 460, undetermined losses 10.45 (all L or kg per day) - none of these individual sub-components has a dedicated column. Design-proposal scenario figures (Section 3.3/Table 4, p.765-766, hypothetical, not this trial’s own data): FRR (Feeding Rate Ratio) target 80 g m-2 day-1; proposed 10 m2 DWC bed area (calculated as 9.63 m2, rounded), 160 lettuce heads/month, 16 plants/m2, 4-week (or 30-day, see WARN-MINOR above) staggered cycles, 40 heads harvested/week; proposed fish husbandry: 12-month cycle, fingerlings 300 g to adults 2 kg at 100 kg/m3 final density, 3 tanks staggered ~17 weeks apart, 82.3 kg/quadrimester/tank; projected annual output 329 (or 360, see WARN-MATERIAL above) kg fish meat + 1920 lettuce heads/year = 177.2 kg/m2 combined biomass; proposed water metabolism: well-water input reduced from 532 to 49.64 L/day, total output reduced from 554.9 to 102.53 L/day. Water use per kg fish: 0.59 m3/kg for this RAS’s catfish, compared in-text to 0.24 m3/kg tilapia (Eurofish 2009, secondary), 0.5-0.7 m3/kg super-intensive RAS and 45 m3/kg extensive pond (Verdegem et al. 2006, secondary), ~1 m3/kg RAS average (Bregnballe 2015, secondary) - all secondary literature comparisons, not this paper’s own additional treatments. Climate data per experiment (Section 3.2.1, p.763): exp1 (Jul-Aug 2016) min/avg/max 17.0/28.4/46.0 degC (Jul), 13.7/23.5/34.4 degC (Aug); exp2 (Sep-Oct 2016) 14.5/21.1/34.0 degC (Sep), 12.0/18.0/30.5 degC (Oct); exp3 (Jun-Jul 2017) 16.0/24.3/50.0 degC (Jun), 17.0/23.9/39.0 degC (Jul). WUE by experiment (Fig 4, text p.764): Aqua 69.8/98.5/48.1 g FW L-1 H2O for exp1/2/3 (used as the range in the WUE cell); Hydro 1.6/2.0/3.0 per-experiment values are figure-only (Fig 4 bars), not extracted. | Fish Category, Plant Category, Water classification left NR - paper does not categorise beyond species/cultivar names and treatment labels (Aqua/Hydro), per SCHEMA.md’s instruction not to substitute an external taxonomy.

Graph View

Backlinks