A circular tri-trophic system incorporating plants, fish, and insects turns waste into a resource: case study with the cultivation of cucumber
Metadata
- Cite key: levizouCircularTriTrophicSystem2025
- Item type: Journal Article
- Authors: E. Levizou, A. Mourantian, M. Chatzinikolaou, M. Feka, I. Karapanagiotidis, E. Mente, C. Athanassiou, K. Kormas, N. Katsoulas
- Affiliation: Department of Agriculture Crop Production and Rural Environment, University of Thessaly, Volos, Greece (Levizou, Mourantian, Chatzinikolaou, Feka, Athanassiou, Katsoulas); Department of Ichthyology and Aquatic Environment, University of Thessaly, Volos, Greece (Karapanagiotidis, Kormas); Department of Veterinary Medicine, Aristotle University of Thessaloniki, Thessaloniki, Greece (Mente)
- Journal: Frontiers in Plant Science 16 (2025) 1638443 (article number, Frontiers journal — not a page range)
- Date: 10/2025 (published 14 October 2025; received 02 June 2025, accepted 22 September 2025)
- Date added: 2026-07-15
- DOI: 10.3389/fpls.2025.1638443
- Funding: “The project is funded by the Green Fund, under the Programme ‘Natural Environment & Innovative Actions 2022’/P.A. 3 ‘RESEARCH AND APPLICATION.’” — the funding statement appears cut off after this phrase in the source PDF (p.14); transcribed exactly as printed, [unclear] whether more text follows.
- URL: https://doi.org/10.3389/fpls.2025.1638443
- PDF:
Levizou et al. - 2025 - A circular tri-trophic system incorporating plants, fish, and insects turns waste into a resource c.pdf
Opinion
A genuinely novel extension of the coupled/decoupled aquaponics comparisons this group has published before (Aslanidou et al. 2023, 2024; Mourantian et al. 2023; Chandrou et al. 2024) — same pilot greenhouse, same design logic, but with a third trophic layer (black soldier fly larvae) closing the fish-feed loop. The experimental design for the plant/fish comparison is solid: true randomized block replication (6 channels/treatment across 3 blocks), a real hydroponic control, and a sensible battery of growth, yield, gas-exchange and chlorophyll-fluorescence measurements. However, the “tri-trophic” framing oversells what is actually measured: the insect component is described only as an input (how BSF larvae are reared and processed into feed) and is never evaluated on its own terms — no larval growth rate, survival, feed-conversion, or waste-reduction metrics are reported for the insect loop itself. In practice this is a CAP/DCAP/HP cucumber aquaponics study whose fish feed happens to contain insect meal, not a comparative evaluation of the insect sub-system. Also worth flagging for future citation: a good number of the paper’s most interesting physiological results (gas exchange, chlorophyll fluorescence, weekly yield, leaf N/P) are reported only as bar/radar charts without a companion numbers table, so a large share of the substantive findings could not be captured as verifiable cells here.
Abstract
Introduction: Circular economy principles are key to enhancing agricultural sustainability and efficiency. In this context, a tri-trophic circular system comprising three types of organisms (plants, insects and fish) that feed each other has been implemented. The nutritional loop involved: i) fertilizing cucumbers with water containing fish metabolic waste through a recirculating water system known as aquaponics; ii) feeding black soldier fly larvae plant pruning residues and fruit; and iii) feeding tilapia fish insect larvae after they have been transformed into insect meal and incorporated into aquafeed. This study aimed to comprehensively evaluate cucumber production in this circular system by investigating key physiological, growth, and yield parameters, and assessing resource use efficiency.
Methods: We implemented in total three treatments, comparing conventional hydroponics (HP) as control, and two aquaponics variants: a) conventional coupled aquaponics (CAP), where water recirculates between crops and fish tanks, so crops receive only fish-derived nutrients; and b) decoupled aquaponics (DCAP), where fish-derived water is enriched with fertilizers to meet the crops’ nutritional needs.
Results: DCAP showed similar performance to the HP control, and both outperformed the CAP in terms of plant physiological/functional traits, fruit yield, and biomass accumulation. CAP treatment reduced total aerial biomass accumulation by 57% compared with the HP control, while DCAP increased it by 14%. The enhanced performance exhibited by DCAP can be attributed to its efficient photosynthetic apparatus and optimal leaf nutrient status. Conversely, CAP resulted in a decline in nutrient levels in irrigation water relative to HP and DCAP, which led to significantly decreased leaf concentrations of potassium and phosphorus (2.5 and 1.5 times lower than HP, respectively, by the end of the experiment). This triggered a series of responses, including a down-regulation of the photosynthetic process and a reduced photochemical activity. DCAP exhibited increased fertilizer use efficiency by 76% over HP, achieving a similar fruit yield with reduced fertilizer inputs.
Discussion: In conclusion, DCAP proved to be highly productive, overcoming the limitations observed in CAP, while offering increased environmental and economic advantages compared to HP. The circular tri-trophic system’s performance demonstrated its efficacy in harnessing synergies to optimize resource use and ensure high productivity and self-sufficiency.
Summary
The authors built a “tri-trophic” circular production system at their pilot aquaponics greenhouse (University of Thessaly, Greece) in which cucumber, red tilapia, and black soldier fly (BSF) larvae are chained together: fish water fertilizes the cucumbers (aquaponics), cucumber pruning residue and cull fruit feed the BSF larvae, and the larvae are processed into insect meal that is incorporated into the tilapia feed. Cucumber was grown for 90 days under three irrigation treatments — conventional hydroponics (HP, control), coupled aquaponics (CAP, raw pH-adjusted fish water only), and decoupled aquaponics (DCAP, fish water topped up with fertilizer to match HP nutrient targets) — in a randomized block design with 6 replicate channels per treatment. CAP plants suffered from strongly nutrient-depleted irrigation water (particularly K and P), which reduced leaf nutrient status, down-regulated photosynthesis (assessed via gas exchange and chlorophyll a fluorescence), and cut biomass accumulation and fruit yield by roughly half relative to HP. DCAP, by contrast, matched or exceeded HP on essentially every measured growth, yield, and physiological trait while using substantially less fertilizer (fertilizer-use efficiency ~74% higher than HP) and comparable water-use efficiency. The fish themselves grew well on the BSF-meal-supplemented diet (SGR 2.1%/day, FCR 1.56), which the authors interpret as indirect evidence that BSF larvae reared on plant residues make a viable feed ingredient, though the insect sub-system’s own performance (larval growth, survival, bioconversion) is not directly measured or reported in this paper.
Experiment data
- Location: Pilot-scale aquaponics greenhouse, University of Thessaly, Velestino, Central Greece
- Design: Randomized complete block; 3 blocks x 3 treatments (HP/CAP/DCAP) x 2 channels/treatment/block = 6 channel replicates/treatment; 96 plants/treatment (288 total)
- Replicates / n: 6 channels/treatment (design-level); measurement-specific subsamples: 6 plants/treatment for growth harvests (D25/D50/D90) and leaf tissue analysis, 10 plants/treatment for gas exchange (D20/D45/D70), 15 plants/treatment for chlorophyll fluorescence (D20/D45/D80)
- Duration: 90 days cultivation (fish reared over the same ~3-month period)
- Organisms: Cucumis sativus (cucumber, cv. Columbia) / Red tilapia (Oreochromis spp.) / Hermetia illucens (black soldier fly, BSF — feed-loop insect, see dedicated section below)
- Statistics: One-way ANOVA + Tukey post-hoc per measurement date; Kruskal-Wallis where ANOVA assumptions were not met; JASP 0.18.3; significance p<=0.05
- Feed Conversion Rate (FCR): 1.56 (tilapia, whole-experiment)
- Specific Growth Rate (SGR): 2.1%/day (tilapia, whole-experiment)
- Cumulative cucumber yield: HP 6.16 kg/m2; CAP 3.23 kg/m2 (47.6% less than HP); DCAP 5.46 kg/m2 (11% less than HP) — 90-day period
- Fertilizer Use Efficiency (FUE): HP 34.4; DCAP 59.9 kg cucumbers/kg fertilizer (CAP: NA, no fertilizer used)
- Water Use Efficiency (WUE): HP 100.5; CAP 57.1; DCAP 92.9 kg cucumbers/m3 water used
Experimental design and treatments
Three irrigation-solution treatments were compared for cucumber (Cucumis sativus var. Columbia), all grown in the same drip-irrigated perlite-slab channels in one greenhouse: HP (conventional hydroponic nutrient solution, Savvas et al. 2013 formula, varying by vegetative/reproductive stage — Table 2); CAP (coupled aquaponics — fish-tank water only, pH-adjusted to 6.0–6.5, no fertilizer added, then returned to the fish tanks, closing a single loop); and DCAP (decoupled aquaponics — fish-tank water analyzed weekly and topped up with high-purity fertilizers to match HP’s target concentrations per phenological stage, then pH-adjusted; water not returned to the fish tanks). The design was a randomized complete block (3 blocks, 6 channels/block, 2 channels/treatment/block), giving 6 replicate channels and 96 plants per treatment.
Fish rearing and BSF-based feed
Red tilapia (Oreochromis spp., 329 fish, 141.1 kg total initial biomass across 3 tanks of 1300 L) were fed daily at 1% of body weight with a pelleted diet containing 17% fishmeal and 10% BSF larvae meal (full formulation in Table 1; crude protein 38.15±0.28% DW). Over the 90-day trial, total fish biomass rose from 141.1 kg to 159.8 kg (a stated weight gain of 18.7 kg), giving SGR = 2.1%/day and FCR = 1.56 — “at the upper limits of the typical range for tilapia” per the authors’ comparison against Shaw et al. (2022) and Limbu et al. (2022), both of which tested BSF-substituted tilapia diets (see Citations to chase).
The insect (black soldier fly) component — how it fits the loop
This is the element that makes the system “tri-trophic” rather than standard two-way aquaponics, but it is described only as an input process, not evaluated with its own outcome metrics:
- Plant → insect: BSF larvae are reared in a separate 32 m² climate-controlled chamber (~23°C, ~40% RH). Newly hatched larvae start on chicken feed for one week, then are switched to the crop’s own pruning residues (leaves, stems) and cut/surplus cucumber fruit, with moisture adjusted to the larvae’s preference.
- Insect → fish: once larvae reach ~2–2.5 cm, they are collected, oven-dried (40°C, 5 h), vacuum-dried (24 h), milled and sieved to <1 mm, and this insect meal is incorporated at 10% into the pelleted tilapia feed alongside 17% fishmeal.
- Fish → plant: the same fish supply the metabolic-waste-laden water used in the CAP/DCAP aquaponics treatments above, closing the circle back to the crop.
A portion of the larval population is kept to complete its life cycle (love cage for mating, dark cage for pupation) to sustain the breeding colony. No larval growth rate, survival, feed/bioconversion ratio, or waste-reduction percentage is reported anywhere in the paper — the insect loop’s own performance is entirely undocumented; only its role as a feed-ingredient source is described. Consequently every insect-specific detail is a “NO COLUMN” item in trials.csv (see Extraction notes) — there is no dedicated place in the schema for larval metrics, and the paper doesn’t supply any to place there regardless.
Crop growth and yield
Cucumber aerial biomass (leaf fresh/dry weight, total aerial dry weight; Figure 3) was measured destructively at D25/D50/D90. CAP was significantly lower than both HP and DCAP at every timepoint (stated declines of 35% at D25 and 74%/57% at D50/D90 relative to HP for leaf dry weight — only percentages are given in text; absolute gram values exist only in the Figure 3 bar chart and are not extracted here). HP and DCAP were statistically indistinguishable except at the final harvest, where DCAP exceeded HP by 10% (leaves) and 14% (total aerial dry weight).
Fruit yield followed the same pattern: weekly yield (Figure 4A, 7 harvest weeks from D20) showed CAP consistently ~57–58% below HP, while HP and DCAP tracked closely (non-significant differences of 17–26% in HP’s favor in 3 of 7 weeks). Cumulative yield over the full 90-day period: HP 6.16 kg/m², CAP 3.23 kg/m² (47.6% less than HP), DCAP 5.46 kg/m² (11% less than HP). Note the weekly (57–58%) and cumulative (47.6%) CAP-vs-HP reduction figures use different bases and are not a numeric contradiction — see Extraction notes.
Plant nutrient status (leaf N, K, P)
Leaf tissue (6 plants/treatment, harvested alongside the D25/D50/D90 biomass harvests) was analyzed for N (Kjeldahl), P (vanadomolybdate colorimetry) and K (flame photometry) — Figure 7. CAP leaves were nutrient-poor relative to HP/DCAP for all three elements, most severely for K (CAP stayed in a narrow 7.4–11.6 mg/g DW range throughout, versus HP and DCAP both rising over time to 29.3 and 33.3 mg/g DW respectively by the final harvest) and P (CAP 40–60% lower than HP/DCAP throughout), with N only diverging significantly at the final harvest (CAP 35% lower). Most of these figures are given only as percentage differences in the running text; the K values above are the only ones with absolute numbers stated outside a figure, so they are the only leaf-mineral values captured in plant.csv (see below). Note: the paper’s Methods and Figure 7 axis label the third harvest “D90,” but the Results 3.6 narrative text labels it “D80” throughout — see Extraction notes.
Photosynthetic function (gas exchange and chlorophyll fluorescence) — NO COLUMN
Leaf gas exchange (net photosynthesis AN, transpiration Tr, stomatal conductance gs, intrinsic water-use efficiency iWUE; D20/D45/D70, Figure 6) and in vivo chlorophyll a fluorescence (OJIP/JIP-test parameters — Fv/Fm, PI_TOTAL, PI_ABS, DIo/RC, Sm, etc.; D20/D45/D80, Table 3, Figure 5) both told a consistent story: CAP plants showed a moderate (never exceeding ~20%) but statistically significant depression of photosynthetic rate and photochemical efficiency relative to HP and DCAP, alongside increased thermal energy dissipation (DIo/RC) — interpreted by the authors as a down-regulation to acclimate to low nutrient (especially K) availability, rather than structural damage to the photosynthetic apparatus. DCAP consistently matched or exceeded HP on these measures. None of this has a home in the trials.csv schema (no gas-exchange or fluorescence columns exist), so it is documented here in prose and flagged as NO COLUMN in the Experimental Remarks.
Resource use efficiency
Water-use efficiency (WUE, kg cucumbers per m³ irrigation water) and fertilizer-use efficiency (FUE, kg cucumbers per kg fertilizer used) were calculated for the whole cultivation period (Table 5): WUE — HP 100.5, DCAP 92.9 (7% lower than HP), CAP 57.1 (43.2% lower than HP); FUE — HP 34.4, DCAP 59.9 (a ~74% increase over HP per Results/Discussion; the Abstract states 76%, a likely drafting inconsistency — see Extraction notes), CAP not estimated (received no fertilizer at all).
Linked claims
- Coupled aquaponics produces nutrient-deficient solution relative to hydroponics
- Decoupled aquaponics systems outperform single-loop coupled systems
- Nutrient-limited aquaponic lettuce down-regulates rather than damages its photosynthetic apparatus — same down-regulation-not-damage interpretation reported here for cucumber under CAP
- Potassium deficiency in coupled aquaponics limits cucumber photosynthesis and yield (new claim — see Extraction notes)
- Black soldier fly larvae reared on plant residues support tilapia growth performance in circular aquaponics (new claim — see Extraction notes)
Citations to chase
- todo Shaw, Knopf & Kloas (2022) — tested complete fishmeal substitution with BSF meal in tilapia feed (SGR 1.76%/day, FCR 1.03; fishmeal positive control SGR 2.1%/day) — direct comparator for this paper’s fish-performance numbers
- todo Limbu et al. (2022) — tilapia diets with varying BSF larvae meal inclusion (best SGR 2.1%/day, FCR 1.01 at full replacement) — another direct fish-performance comparator
- todo Aslanidou et al. (2024) — Sci. Hortic. 337:113552 — large-scale coupled/decoupled aquaponics, reports ~40% CAP cucumber yield reduction vs HP, close to this paper’s 47.6%; also reports DCAP > HP for basil/parsley yield but similar levels for cucumber
- todo Pinho et al. (2024) — PLoS ONE 19:e0295811 — insect-based (BSF) fish feed in decoupled aquaponics with lettuce; reports 32% fertilizer reduction (no FUE estimate) and similar per-plant water volumes to this paper’s WUE finding
- todo Roosta (2014a, 2014b) — K/Fe/Mn deficiency effects and K foliar-spray remediation in aquaponic basil — cited as mechanistic support for the K-driven photosynthesis effects seen here
Extraction notes
Type classification: experiment — randomized complete block design with true replication (6 channels/treatment across 3 blocks), a hydroponic control, and formal statistical testing (ANOVA/Tukey, Kruskal-Wallis) throughout. Not a review despite extensive literature comparison in the Discussion (which is about a fishmeal/BSF-substitution sub-topic outside this study’s own design, not a substitute for its original data).
Two trials extracted (one row per aquaponic treatment, per SCHEMA.md): -T1 = CAP vs HP, -T2 = DCAP vs HP. Both share the same fish/RAS/BSF-feed infrastructure; only the irrigation-solution treatment differs.
⚠️ Contradictions found (full evidence also duplicated in trials.csv Experimental Remarks):
- WARN-MATERIAL — FUE percentage increase of DCAP over HP. Abstract: “increased fertilizer use efficiency by 76% over HP.” Results 3.7: “increased by 74.2%.” Discussion: “The 74% increase in FUE.” The underlying Table 5 values (HP FUE 34.4, DCAP FUE 59.9 kg cucumbers/kg fertilizer) compute to ≈74.1% increase, agreeing with Results/Discussion, not the Abstract. Recorded FUE AP/HYD cells use the raw Table 5 numbers directly (74.2%/76% never enters a cell); the Abstract’s 76% is judged the likely error. Affects: narrative interpretation only.
- WARN-CHECK — leaf-tissue sampling day labelled “D80” vs “D90.” Methods 2.5.4 states leaf elemental analysis used “the three plant harvests (D25, D50 and D90),” and the Figure 7 x-axis is labelled D25/D50/D90 — matching the D25/D50/D90 biomass-harvest schedule (Methods 2.5.1, Figure 3). Results 3.6 running text instead says “D80” four times when discussing the same final-harvest leaf N/K/P values. D80 is also the third timepoint used by the unrelated chlorophyll-fluorescence assay (D20/D45/D80, Methods 2.5.2), so this looks like a copy-paste conflation between sections rather than a genuinely separate sampling event. The D90 basis (Methods + Figure axis) was used to label the plant.csv rows; both labels are recorded for the record.
- WARN-CHECK — CAP yield reduction stated two ways. Results 3.3 gives a weekly-harvest-based “stable decrease of 57%-58% compared to HP” and, separately in the same section, a cumulative-total-based “47.6% less than HP” (3.23 vs 6.16 kg/m²). Different bases (weekly snapshot vs whole-period total), not a numeric conflict on one quantity. The cumulative figures were used for the AP/HYD trials.csv cells as the single clearest whole-trial number.
- Coordinates recovered, not a formal contradiction: the site is given as “39° 44′ N, 22°79′ E” (p.3). 44′ is valid (<60), giving decimal 39.7333. 79′ is impossible as DMS (>59); read as decimal degrees mistakenly typeset with DMS symbols → 22.79, consistent with Velestino, Central Greece, the paper’s own named location.
- WARN-MINOR: Table 4’s EC unit is printed as “dS m⁻²” (physically implausible; should read dS/m = dS m⁻¹, and the values 1.11–2.05 are stated in-text to be “within the typical range for cucumber soilless cultivation,” which only holds for dS/m). No cell affected.
- WARN-MINOR: Results 3.5 states net photosynthetic rate “ranging from 13 to 16 mmol CO2 m⁻² sec⁻¹,” but Methods 2.5.3 and the Figure 6 axis define AN in µmol m⁻² s⁻¹ — likely an mmol/µmol typo. No cell affected (gas exchange has no dedicated trials.csv column).
[not reported] / [unclear], grouped by field:
- Fish sizing: Initial Stock density, Fish size initial, Fish size final, Fish weight gain (per fish), Total Feed (kg), Fish survival rate — the paper gives per-tank fish counts, per-tank aggregate biomass, and per-tank weight ceilings (“up to 450/700/1000 g each”), never a mean weight, a stocking density in kg/m³, a total feed figure, or a survival/mortality statistic. Computing any of these from the given inputs would be derivation, so all are
NRwith the raw inputs cited in trials.csv remarks. - Feed composition: N, P, K (feed) — Table 1 gives only moisture/crude protein/crude lipid/ash/gross energy, no %N/%P/%K.
- Water quality unit mismatch (WARN-CHECK, judged out of scope for conversion): TAN/NH4-N and NO3-N are
NRbecause Table 4 reports NH4⁺ and NO3⁻ in mmol/L, not the schema’s mg/L; per vault precedent (see delaideEffectWastewaterPikeperch2019), a molar-mass conversion between the ionic species and the nitrogen-only mass was judged out of scope for “unit conversion only.” Raw mmol/L values are preserved in trials.csv remarks. NO2-N isNRregardless — the paper never measures nitrite. - Water system: Water recycle (L/min), Water type, Water classification, Daily Water exchange rate, Average room Temperature (distinct from the RAS/fish-tank water temperature, which is reported at 23±0.4°C) — none stated.
- Plant growth cells: Plant Category, SPAD, Plant height, Leaf count, Plant fresh weight, Plant dry matter — Figure 3’s biomass values and all SPAD/height/leaf-count-type measures are never given as absolute numbers in text or a table (only percentage differences), so all are
NRper the never-read-a-figure rule. - Artificial Lighting:
NRfor the crop — the paper’s only lighting mention (an LED mating lamp) is for the BSF breeding cage, not the greenhouse crop, and is not conflated here.
NO COLUMN items (insect metrics are — as anticipated — entirely NO COLUMN):
- All BSF/insect-specific data: rearing-chamber conditions, larval diet and harvest size, insect-meal processing steps, breeding-cage dimensions. No larval growth rate, survival, bioconversion ratio, or waste-reduction percentage is reported at all, so there was nothing quantitative to place even if a column existed.
- Full fish-feed proximate composition and ingredient list beyond crude protein (Table 1).
- Gas exchange (AN, Tr, gs, iWUE) and chlorophyll fluorescence/OJIP parameters (no dedicated trials.csv columns).
- Weekly yield time series and its week-by-week significance letters.
- Table 4’s full statistical letter-groupings across HP/CAP/DCAP.
- Full three-way WUE table (only the AP-side value went into each trial row’s single WUE cell; the other two values are cross-referenced in remarks).
- Secondary literature figures used for fish-performance comparison (Shaw 2022; Limbu 2022; Aslanidou 2024; Pinho 2024) — routed to Citations to chase, not recorded as this paper’s data.
- Ethics/approval numbers and the (apparently truncated) funding statement.
New tags/wikilinks introduced: no new Meta/ tag facet was created for the insect component. The vault’s existing facets are Meta/Type/, Meta/Region/, Meta/Fish/, Meta/Plant/ — there is no Meta/Insect/ precedent anywhere in the vault, and a single paper isn’t sufficient justification to invent one unilaterally. Documented the BSF species (Hermetia illucens) directly in the note body and here instead. New wikilink targets created: [[Fertilizer Use Efficiency (FUE)]] (no prior FUE concept note existed, unlike FCR/SGR/WUE which already have vault entries in one or more spelling variants), [[Potassium deficiency in coupled aquaponics limits cucumber photosynthesis and yield]], [[Black soldier fly larvae reared on plant residues support tilapia growth performance in circular aquaponics]]. Note for a future consolidation pass: the vault already has fragmented FCR/SGR/WUE claim spellings ([[FCR]] / [[Feed Conversion Rate (FCR)]] / [[Feed Conversion Ratio (FCR)]]; [[Water Use Efficiency (WUE)]] / [[Water use efficiency (WUE)]]) — used the more common (FCR)/(SGR) parenthetical forms here for consistency, did not attempt to merge existing fragments.
Region/tagging judgment call: tagged Meta/Fish/Tilapia and Meta/Plant/Cucumber (both directly studied); no Meta/Fish/ or Meta/Plant/ tag added for BSF since it is not a fish or plant and (per above) no insect facet was created.
Source: Levizou et al. - 2025 - A circular tri-trophic system incorporating plants, fish, and insects turns waste into a resource c.pdf
Data Tables
Structured data extracted from this paper into the vault's
trials.csv/plant_measurements.csvdatasets. Fields the paper didn't report are omitted. Download the full datasets (measurements).
Trial Parameters
levizouCircularTriTrophicSystem2025-T1
Fish
| Field | Value |
|---|---|
| Fish | Red tilapia (Oreochromis spp.) |
| FCR | 1.56 |
| SGR | 2.1 |
| Protein | 38.15 +/- 0.28 |
| % of body weight | 1 |
| Feed regime | Pelleted diet: fishmeal 17%, BSF larvae meal 10%, wheat meal 24%, rapeseed meal 19%, corn gluten meal 13%, sunflower meal 9%, soya oil 5.5%, vitamin/mineral premix 0.7%, monocalcium phosphate 0.3%, vitamin C 0.25%, vitamin E 0.25%, L-lysine 0.5%, DL-methionine 0.5% (Table 1); fed at 1% of body weight daily |
| Fish biomass created (kg) | 18.7 |
| Fish trial duration (days) | 90 |
Water
| Field | Value |
|---|---|
| Water volume in the system | 1300 (per fish tank, x3 tanks; separate 2500 L sump tank; RAS total not stated as one figure) |
| Aq pH | 6.48 +/- 0.12 |
| pHOptimal | 6.0-6.5 (target range for nutrient availability, Methods 2.4, p.6) |
| FUE HYD | 34.4 |
| WUE | 57.1 (kg cucumbers m-3 water used; CAP) |
| Dissolved Oxigen | 7.0 +/- 0.3 |
| EC | 1.11 +/- 0.02 |
| Water temperature | 23 +/- 0.4 |
Plant
| Field | Value |
|---|---|
| Plant | Cucumber (Cucumis sativus var. Columbia) |
| Details | 288 seedlings (4-true-leaf stage) transplanted at 1.175 plants/m2; 90-day cultivation; destructive growth harvests at D25/D50/D90; weekly fruit harvest from D20 (7 harvest weeks) |
| Days Plant after transplant | 90 |
| Plants/m2 | 1.175 |
System & Setup
| Field | Value |
|---|---|
| System type | Perlite slab substrate culture, drip-irrigated channels (p.3-4) |
| Media Details | Perlite slabs (particle diameter 1-5 mm, Perterra, NORDIA S.A.); 8 slabs/channel, 18 channels of 8.5 m length at 50 cm height; 5 drippers/slab |
| Biological system already in use | Y (Same pilot-scale aquaponics greenhouse/RAS facility described and used in Aslanidou et al. 2023, 2024 and Mourantian et al. 2023 (Methods 2.1)) |
| Air supplement | Y (Air blower at 100 m3/h via 22 diffusers maintaining DO 7.0 +/- 0.3 mg/L in fish tanks (Methods 2.2)) |
| Iron supplemented | N (CAP received no fertilizer amendment, pH adjustment only (Methods 2.4); for reference, HP’s Fe target = 15 umol/L at both phenological stages (Table 2)) |
| Remineralization | N (CAP water not remineralized/fertilized, only pH-adjusted (Methods 2.4)) |
| pH Buffers | Y (CAP water pH adjusted to 6.0-6.5 range before reaching crops (Methods 2.4); specific reagent/method not stated) |
| Climate control | Y (Automated greenhouse climate control (Argos Electronics software); ventilation opens >21C/85%RH, heating at 18C, fan+wet-pad cooling at 26C setpoint (Methods 2.1)) |
| Nutrient supplemented | N (CAP treatment explicitly received no fertilizer amendments (Methods 2.4: ‘fish water with no fertilizer amendments and only pH adjustment’)) |
| Equipment | FluorPen FP 110 (chlorophyll fluorescence); LI-6400/XT LI-COR (gas exchange); HQ40d Hach portable sensor (pH/EC/DO); Combo pH-EC-TDS-Temp Hanna (temperature); Rotary Drum Filter ProfIDrum; biofilter with ceramic rings + K1 Kaldnes media, Prodibio Biodigest inoculant; pH/EC/O2 transducer GHM-Greisinger; UV1900 Shimadzu photometer (leaf P); Jenway PFP7 flame photometer (leaf K); iMETOS sm meteorological station; Argos Electronics control software; JASP 0.18.3 statistics |
| Control Parameters | DO setpoint 7.0 +/- 0.3 mg/L; fish tank temperature 23 +/- 0.4 C; irrigation solution pH 6.0-6.5; greenhouse ventilation >21C/85%RH, heating 18C, fan+wet-pad cooling at 26C setpoint; insect breeding chamber ~23C/~40%RH |
| Combination | Red tilapia (Oreochromis spp.) and cucumber (Cucumis sativus var. Columbia) in a tri-trophic loop with black soldier fly (Hermetia illucens) larvae bridging plant residues to fish feed; this row = CAP vs HP comparison |
Site
| Field | Value |
|---|---|
| Region | Europe |
| Country | Greece |
| Lat | 39.7333 |
| Long | 22.79 |
Results & Statistics
| Field | Value |
|---|---|
| Measured Unit | kg/m2 (cumulative fruit yield); mg/g DW (leaf N/P/K); dS/m (EC); mmol/L (irrigation nutrient concentrations) |
| Statistic Details | One-way ANOVA + Tukey post-hoc per measurement date; Kruskal-Wallis when ANOVA assumptions unmet; JASP 0.18.3; significance p<=0.05 |
| Statistically analysed | Y |
| Replicates (n) | 6 |
| AP | 3.23 |
| HYD | 6.16 |
Experimental Remarks: TRIAL DEFINITION: T1 = coupled aquaponics (CAP) treatment - cucumber irrigated directly with fish-tank water (pH-adjusted to 6.0-6.5 only, no fertilizer amendment), water then returned to the fish tanks, closing a single coupled loop (Methods 2.4, Figure 2A). Paired control = HP (conventional hydroponic nutrient solution per Savvas et al. 2013, Table 2), recorded in the HYD-labelled cells. Design: randomized complete block, 3 blocks x 6 channels/block, 2 channels/treatment/block = 6 channel replicates/treatment, 96 plants/treatment (288 total across 3 treatments). | TRI-TROPHIC INSECT LOOP (shared infrastructure across T1 and T2): black soldier fly (BSF, Hermetia illucens) larvae are reared in a separate 32 m2 climate-controlled chamber (~23C, ~40% RH) fed initially on chicken feed then on plant pruning residues and cut/surplus cucumbers from this same crop (plant->insect link, Methods 2.3); at ~2-2.5 cm the larvae are collected, oven-dried (40C, 5h), vacuum-dried (24h), milled/sieved to <1mm and incorporated as 10% insect meal in the pelleted tilapia feed alongside 17% fishmeal (Table 1) (insect->fish link); fish, in turn, supply the metabolic waste that fertilizes the crop via CAP/DCAP aquaponics (fish->plant link), closing the tri-trophic circle (Figure 1). The insect loop itself is not evaluated with its own performance metrics (no larval growth rate, survival, bioconversion ratio, or waste-reduction % reported) - see Extraction notes and NO COLUMN below. | WARN-MATERIAL FUE % increase of DCAP over HP: Abstract states ‘increased fertilizer use efficiency by 76% over HP’; Results 3.7 states ‘increased by 74.2%’; Discussion states ‘The 74% increase in FUE’. Table 5 raw values (HP FUE 34.4, DCAP FUE 59.9 kg cucumbers/kg fertilizer) support ~74.1% ((59.9-34.4)/34.4), matching Results/Discussion, not the Abstract’s 76%. FUE AP/FUE HYD cells use the raw Table 5 values directly (unaffected by which percentage figure is correct); the percentage discrepancy itself is not carried into any cell. Abstract’s 76% judged the likely drafting/arithmetic error. | WARN-CHECK Leaf tissue sampling day label: Methods 2.5.4 and the Figure 7 axis both give the third/final leaf-tissue harvest as ‘D90’ (matching the D25/D50/D90 growth-harvest schedule of Methods 2.5.1 and Figure 3’s axis). Results 3.6 running text instead says ‘D80’ four times (‘on D80 the nitrogen content…’, ‘…on D80 compared to D50 levels’, ‘[P]…significantly lower by 40-60% (D25 and D80, respectively)’). D80 is also the third timepoint used for the UNRELATED chlorophyll-fluorescence assay (D20/D45/D80, Methods 2.5.2), so the Results-text label is plausibly a copy-paste conflation with that assay’s schedule rather than a distinct sampling event. Recorded D90 (matches Methods + Figure axis) as the basis for the plant.csv Location field; both labels given here for the record. Affects: plant.csv K rows for HP (29.3 mg/g DW) and DCAP (33.3 mg/g DW). Added to REVIEW.md by the batch merge step. | WARN-CHECK CAP yield reduction vs HP, two bases: Results 3.3 gives a weekly-harvest-based ‘stable decrease of 57%-58% compared to HP’ (per-week comparison across 7 harvest weeks, Figure 4A) and separately a cumulative-total-based ‘47.6% less than HP’ (6.16 vs 3.23 kg/m2 over the full 90-day period, Figure 4B/text). These are different bases (weekly snapshot average vs whole-period cumulative total), not a numeric conflict on the same quantity. AP/HYD cells use the cumulative figures (3.23, 6.16 kg/m2) as the single, clearest whole-trial basis; the weekly range is noted here only. Added to REVIEW.md by the batch merge step. | UNIT CONVERSION ONLY: coordinates ‘39 deg 44 min N, 22 deg 79 min E’ (p.3) -> Lat 39 deg 44’ N is valid DMS (44<60) -> decimal 39.7333; Long 22 deg 79’ E has minutes >59 (impossible as DMS), read as decimal degrees mistakenly written with DMS symbols -> 22.79, consistent with Velestino, Central Greece (the paper’s stated site). | WARN-MINOR Table 4 EC unit printed as ‘dS m-2’; physically implausible unit for EC (should be dS/m = dS m-1); values (HP 1.92, CAP 1.11, DCAP 2.05) are typical cucumber soilless-culture EC in dS/m as stated in the surrounding text (‘within the typical range for cucumber soilless cultivation’), so treated as a units-label typo. No cell affected (dS/m used as recorded). | WARN-MINOR Results 3.5 states net photosynthetic rate ‘ranging from 13 to 16 mmol CO2 m-2 sec-1’, but Methods 2.5.3 and Figure 6’s y-axis define AN in umol m-2 s-1; likely a mmol/umol typo in the running text. No cell affected (gas exchange has no dedicated column, see NO COLUMN). | NOT DERIVED, left NR: Initial Stock density (paper gives per-tank fish count, aggregate biomass, and 1300 L tank volume - e.g. tank 1: 160 fish, 43.7 kg, up to 450 g each - but no stated kg/m3 density; computing one would be derivation); Fish size initial/final (only per-tank weight CEILINGS given - ‘up to 450/700/1000 g each’ - and aggregate tank biomass, not a mean weight); Fish weight gain per fish (only the population-level aggregate ‘weight gain of 18.7 kg’ is stated, recorded instead as Fish biomass created); Total Feed (kg) (FCR and biomass gain are both stated, but total feed consumed in kg is never stated as a figure; back-calculating it would be derivation); Fish survival rate (no mortality/survival % reported for the 329 fish); Feed routine (feeding frequency per day not stated, only the 1%-of-body-weight daily ration); feed N, P, K composition beyond crude protein (Table 1 gives only moisture/crude protein/crude lipid/ash/gross energy); Water recycle (L/min, not stated); Water type/Water classification (not categorised in the paper’s own words); Daily Water exchange rate; Average room Temperature (greenhouse air setpoints given as control bands 18-26C, not a measured trial-mean room temperature - distinct from the RAS fish-tank water temperature of 23+/-0.4C recorded under Water temperature); Plant Category (no categorising term used); SPAD, Plant height, Leaf count, Plant fresh weight, Plant dry matter (Figure 3 gives only a bar chart of fresh/dry leaf and aerial-part weights, with only percentage differences in running text - ‘35% decrease…on D25, the 74% and 57% on D50 and D90’ - no absolute gram values in text or a table, so left NR per the never-read-a-figure rule; percentages given under NO COLUMN below); Artificial Lighting (no supplemental crop lighting mentioned; the LED mating lamp is for the BSF love cage, not the crop). | NO COLUMN: Fish feed full proximate composition beyond crude protein - moisture 9.68+/-0.17%, crude lipid 10.23+/-0.01%, ash 6.59+/-0.05%, gross energy 19.15+/-0.31 MJ/kg, and full ingredient list (Table 1: fishmeal 17%, BSF larvae meal 10%, wheat meal 24%, rapeseed meal 19%, corn gluten meal 13%, sunflower meal 9%, soya oil 5.5%, vitamin/mineral premix 0.7%, monocalcium phosphate 0.3%, vitamin C 0.25%, vitamin E 0.25%, L-lysine 0.5%, DL-methionine 0.5%). Insect-specific metrics (ALL): BSF larval target harvest size ~2-2.5 cm; rearing chamber 32 m2, ~23C, ~40% RH; initial larval diet chicken feed, then plant pruning residues + cut cucumbers, moisture adjusted with water; insect meal processing (oven-dry 40C/5h, vacuum-dry 24h, milled/sieved <1mm, 10% dietary inclusion); breeding infrastructure (love cage 70x70x86 cm with LED mating lamp, pupation cage 66x75x67 cm); no larval growth rate, survival, bioconversion ratio, or waste-reduction % reported anywhere in the paper. Gas exchange dynamics (AN, Tr, gs, iWUE) at D20/D45/D70 (Figure 6, no dedicated schema column). Chlorophyll a fluorescence/OJIP parameters (Fv/Fm, PItotal, PIabs, Sm, DIo/RC, ABS/RC, TRo/RC, ETo/RC, 1-Vi, 1/Vi) at D20/D45/D80 (Table 3, Figure 5, no dedicated column). Weekly fruit yield time series and week-by-week significance groupings (Figure 4A, 7 harvest weeks). Full Table 4 statistical letter-groupings for pH/EC/NO3/NH4/PO4/K/Na/Ca among HP/CAP/DCAP (a/b superscripts). WUE full three-treatment table (HP 100.5, CAP 57.1, DCAP 92.9 kg cucumbers/m3 water used) - AP-side value used in this row’s WUE cell, other two given here for reference. Literature comparison figures (secondary, not this paper’s data): Shaw et al. 2022 tilapia SGR 1.76%/day and FCR 1.03 (BSF)/0.79 (fishmeal control); Limbu et al. 2022 tilapia SGR 2.1%/day and FCR 1.01 (full BSF replacement); Aslanidou et al. 2024 ~40% CAP cucumber yield reduction vs HP; Pinho et al. 2024 32% fertilizer reduction in DCAP lettuce (FUE not estimated by that paper). Ethics/animal-welfare approval details (EU Directive 2010/63/EU; Animal Care and Use Ethics Committee approval 242627/28-05-2024; facility EL-43BIO/exp-02). Funding text appears truncated in the source PDF after ’…RESEARCH AND APPLICATION.’ with no further detail; transcribed as printed.
levizouCircularTriTrophicSystem2025-T2
Fish
| Field | Value |
|---|---|
| Fish | Red tilapia (Oreochromis spp.) |
| FCR | 1.56 |
| SGR | 2.1 |
| Protein | 38.15 +/- 0.28 |
| % of body weight | 1 |
| Feed regime | Pelleted diet: fishmeal 17%, BSF larvae meal 10%, wheat meal 24%, rapeseed meal 19%, corn gluten meal 13%, sunflower meal 9%, soya oil 5.5%, vitamin/mineral premix 0.7%, monocalcium phosphate 0.3%, vitamin C 0.25%, vitamin E 0.25%, L-lysine 0.5%, DL-methionine 0.5% (Table 1); fed at 1% of body weight daily |
| Fish biomass created (kg) | 18.7 |
| Fish trial duration (days) | 90 |
Water
| Field | Value |
|---|---|
| Water volume in the system | 1300 (per fish tank, x3 tanks; separate 2500 L sump tank; RAS total not stated as one figure) |
| Aq pH | 6.31 +/- 0.10 |
| pHOptimal | 6.0-6.5 (target range for nutrient availability, Methods 2.4, p.6) |
| FUE AP | 59.9 |
| FUE HYD | 34.4 |
| WUE | 92.9 (kg cucumbers m-3 water used; DCAP) |
| Dissolved Oxigen | 7.0 +/- 0.3 |
| EC | 2.05 +/- 0.05 |
| Water temperature | 23 +/- 0.4 |
Plant
| Field | Value |
|---|---|
| Plant | Cucumber (Cucumis sativus var. Columbia) |
| Details | 288 seedlings (4-true-leaf stage) transplanted at 1.175 plants/m2; 90-day cultivation; destructive growth harvests at D25/D50/D90; weekly fruit harvest from D20 (7 harvest weeks) |
| Days Plant after transplant | 90 |
| Plants/m2 | 1.175 |
System & Setup
| Field | Value |
|---|---|
| System type | Perlite slab substrate culture, drip-irrigated channels (p.3-4) |
| Media Details | Perlite slabs (particle diameter 1-5 mm, Perterra, NORDIA S.A.); 8 slabs/channel, 18 channels of 8.5 m length at 50 cm height; 5 drippers/slab |
| Biological system already in use | Y (Same pilot-scale aquaponics greenhouse/RAS facility described and used in Aslanidou et al. 2023, 2024 and Mourantian et al. 2023 (Methods 2.1)) |
| Air supplement | Y (Air blower at 100 m3/h via 22 diffusers maintaining DO 7.0 +/- 0.3 mg/L in fish tanks (Methods 2.2)) |
| Iron supplemented | Y (DCAP water enriched with fertilizers to reach HP target values for each phenological stage (Methods 2.4), which per Table 2 includes an Fe target of 15 umol/L (both stages); Fe not among the Table 4 measured parameters, so not independently confirmed by water analysis in this paper) |
| Remineralization | Y (Fish-tank water analyzed weekly and enriched with high-purity fertilizers to reach HP nutrient concentration targets per phenological stage (Methods 2.4; calculation method detailed in Aslanidou et al. 2023)) |
| pH Buffers | Y (DCAP water pH adjusted to 6.0-6.5 range before reaching crops (Methods 2.4); specific reagent/method not stated) |
| Climate control | Y (Automated greenhouse climate control (Argos Electronics software); ventilation opens >21C/85%RH, heating at 18C, fan+wet-pad cooling at 26C setpoint (Methods 2.1)) |
| Nutrient supplemented | Y (DCAP fish water enriched with high-purity fertilizers dosed weekly to match HP nutrient concentration targets by phenological stage (Methods 2.4)) |
| Equipment | FluorPen FP 110 (chlorophyll fluorescence); LI-6400/XT LI-COR (gas exchange); HQ40d Hach portable sensor (pH/EC/DO); Combo pH-EC-TDS-Temp Hanna (temperature); Rotary Drum Filter ProfIDrum; biofilter with ceramic rings + K1 Kaldnes media, Prodibio Biodigest inoculant; pH/EC/O2 transducer GHM-Greisinger; UV1900 Shimadzu photometer (leaf P); Jenway PFP7 flame photometer (leaf K); iMETOS sm meteorological station; Argos Electronics control software; JASP 0.18.3 statistics |
| Control Parameters | DO setpoint 7.0 +/- 0.3 mg/L; fish tank temperature 23 +/- 0.4 C; irrigation solution pH 6.0-6.5; greenhouse ventilation >21C/85%RH, heating 18C, fan+wet-pad cooling at 26C setpoint; insect breeding chamber ~23C/~40%RH |
| Combination | Red tilapia (Oreochromis spp.) and cucumber (Cucumis sativus var. Columbia) in a tri-trophic loop with black soldier fly (Hermetia illucens) larvae bridging plant residues to fish feed; this row = DCAP vs HP comparison |
Site
| Field | Value |
|---|---|
| Region | Europe |
| Country | Greece |
| Lat | 39.7333 |
| Long | 22.79 |
Results & Statistics
| Field | Value |
|---|---|
| Measured Unit | kg/m2 (cumulative fruit yield); mg/g DW (leaf N/P/K); dS/m (EC); mmol/L (irrigation nutrient concentrations) |
| Statistic Details | One-way ANOVA + Tukey post-hoc per measurement date; Kruskal-Wallis when ANOVA assumptions unmet; JASP 0.18.3; significance p<=0.05 |
| Statistically analysed | Y |
| Replicates (n) | 6 |
| AP | 5.46 |
| HYD | 6.16 |
Experimental Remarks: TRIAL DEFINITION: T2 = decoupled aquaponics (DCAP) treatment - fish-tank water enriched weekly with high-purity fertilizers (dosed based on fish-water nutrient analysis) to reach HP nutrient targets for each phenological stage, then pH-adjusted to 6.0-6.5; water is NOT returned to the fish tanks (open/decoupled loop, Methods 2.4, Figure 2A). Paired control = HP, recorded in the HYD-labelled cells (same HP values as T1’s HYD columns, repeated per one-row-per-treatment convention). Same shared RAS/fish population and BSF-supplemented feed as T1 (see insect-loop note below). Design: randomized complete block, 3 blocks x 6 channels/block, 2 channels/treatment/block = 6 channel replicates/treatment, 96 plants/treatment. | TRI-TROPHIC INSECT LOOP (shared infrastructure across T1 and T2): black soldier fly (BSF, Hermetia illucens) larvae are reared in a separate 32 m2 climate-controlled chamber (~23C, ~40% RH) fed initially on chicken feed then on plant pruning residues and cut/surplus cucumbers from this same crop (plant->insect link, Methods 2.3); at ~2-2.5 cm the larvae are collected, oven-dried (40C, 5h), vacuum-dried (24h), milled/sieved to <1mm and incorporated as 10% insect meal in the pelleted tilapia feed alongside 17% fishmeal (Table 1) (insect->fish link); fish, in turn, supply the metabolic waste that fertilizes the crop via CAP/DCAP aquaponics (fish->plant link), closing the tri-trophic circle (Figure 1). The insect loop itself is not evaluated with its own performance metrics (no larval growth rate, survival, bioconversion ratio, or waste-reduction % reported) - see Extraction notes and NO COLUMN below. | WARN-MATERIAL FUE % increase of DCAP over HP: Abstract states ‘increased fertilizer use efficiency by 76% over HP’; Results 3.7 states ‘increased by 74.2%’; Discussion states ‘The 74% increase in FUE’. Table 5 raw values (HP FUE 34.4, DCAP FUE 59.9 kg cucumbers/kg fertilizer) support ~74.1% ((59.9-34.4)/34.4), matching Results/Discussion, not the Abstract’s 76%. FUE AP/FUE HYD cells use the raw Table 5 values directly (unaffected by which percentage figure is correct); the percentage discrepancy itself is not carried into any cell. Abstract’s 76% judged the likely drafting/arithmetic error. | WARN-CHECK Leaf tissue sampling day label: Methods 2.5.4 and the Figure 7 axis both give the third/final leaf-tissue harvest as ‘D90’ (matching the D25/D50/D90 growth-harvest schedule of Methods 2.5.1 and Figure 3’s axis). Results 3.6 running text instead says ‘D80’ four times (‘on D80 the nitrogen content…’, ‘…on D80 compared to D50 levels’, ‘[P]…significantly lower by 40-60% (D25 and D80, respectively)’). D80 is also the third timepoint used for the UNRELATED chlorophyll-fluorescence assay (D20/D45/D80, Methods 2.5.2), so the Results-text label is plausibly a copy-paste conflation with that assay’s schedule rather than a distinct sampling event. Recorded D90 (matches Methods + Figure axis) as the basis for the plant.csv Location field; both labels given here for the record. Affects: plant.csv K rows for HP (29.3 mg/g DW) and DCAP (33.3 mg/g DW). Added to REVIEW.md by the batch merge step. | WARN-CHECK CAP yield reduction vs HP, two bases: Results 3.3 gives a weekly-harvest-based ‘stable decrease of 57%-58% compared to HP’ (per-week comparison across 7 harvest weeks, Figure 4A) and separately a cumulative-total-based ‘47.6% less than HP’ (6.16 vs 3.23 kg/m2 over the full 90-day period, Figure 4B/text). These are different bases (weekly snapshot average vs whole-period cumulative total), not a numeric conflict on the same quantity. AP/HYD cells use the cumulative figures (3.23, 6.16 kg/m2) as the single, clearest whole-trial basis; the weekly range is noted here only. Added to REVIEW.md by the batch merge step. | UNIT CONVERSION ONLY: coordinates ‘39 deg 44 min N, 22 deg 79 min E’ (p.3) -> Lat 39 deg 44’ N is valid DMS (44<60) -> decimal 39.7333; Long 22 deg 79’ E has minutes >59 (impossible as DMS), read as decimal degrees mistakenly written with DMS symbols -> 22.79, consistent with Velestino, Central Greece (the paper’s stated site). | WARN-MINOR Table 4 EC unit printed as ‘dS m-2’; physically implausible unit for EC (should be dS/m = dS m-1); values (HP 1.92, CAP 1.11, DCAP 2.05) are typical cucumber soilless-culture EC in dS/m as stated in the surrounding text (‘within the typical range for cucumber soilless cultivation’), so treated as a units-label typo. No cell affected (dS/m used as recorded). | WARN-MINOR Results 3.5 states net photosynthetic rate ‘ranging from 13 to 16 mmol CO2 m-2 sec-1’, but Methods 2.5.3 and Figure 6’s y-axis define AN in umol m-2 s-1; likely a mmol/umol typo in the running text. No cell affected (gas exchange has no dedicated column, see NO COLUMN). | NOT DERIVED, left NR: Initial Stock density (paper gives per-tank fish count, aggregate biomass, and 1300 L tank volume - e.g. tank 1: 160 fish, 43.7 kg, up to 450 g each - but no stated kg/m3 density; computing one would be derivation); Fish size initial/final (only per-tank weight CEILINGS given - ‘up to 450/700/1000 g each’ - and aggregate tank biomass, not a mean weight); Fish weight gain per fish (only the population-level aggregate ‘weight gain of 18.7 kg’ is stated, recorded instead as Fish biomass created); Total Feed (kg) (FCR and biomass gain are both stated, but total feed consumed in kg is never stated as a figure; back-calculating it would be derivation); Fish survival rate (no mortality/survival % reported for the 329 fish); Feed routine (feeding frequency per day not stated, only the 1%-of-body-weight daily ration); feed N, P, K composition beyond crude protein (Table 1 gives only moisture/crude protein/crude lipid/ash/gross energy); Water recycle (L/min, not stated); Water type/Water classification (not categorised in the paper’s own words); Daily Water exchange rate; Average room Temperature (greenhouse air setpoints given as control bands 18-26C, not a measured trial-mean room temperature - distinct from the RAS fish-tank water temperature of 23+/-0.4C recorded under Water temperature); Plant Category (no categorising term used); SPAD, Plant height, Leaf count, Plant fresh weight, Plant dry matter (Figure 3 gives only a bar chart of fresh/dry leaf and aerial-part weights, with only percentage differences in running text - ‘35% decrease…on D25, the 74% and 57% on D50 and D90’ - no absolute gram values in text or a table, so left NR per the never-read-a-figure rule; percentages given under NO COLUMN below); Artificial Lighting (no supplemental crop lighting mentioned; the LED mating lamp is for the BSF love cage, not the crop). | NO COLUMN: Fish feed full proximate composition beyond crude protein - moisture 9.68+/-0.17%, crude lipid 10.23+/-0.01%, ash 6.59+/-0.05%, gross energy 19.15+/-0.31 MJ/kg, and full ingredient list (Table 1: fishmeal 17%, BSF larvae meal 10%, wheat meal 24%, rapeseed meal 19%, corn gluten meal 13%, sunflower meal 9%, soya oil 5.5%, vitamin/mineral premix 0.7%, monocalcium phosphate 0.3%, vitamin C 0.25%, vitamin E 0.25%, L-lysine 0.5%, DL-methionine 0.5%). Insect-specific metrics (ALL): BSF larval target harvest size ~2-2.5 cm; rearing chamber 32 m2, ~23C, ~40% RH; initial larval diet chicken feed, then plant pruning residues + cut cucumbers, moisture adjusted with water; insect meal processing (oven-dry 40C/5h, vacuum-dry 24h, milled/sieved <1mm, 10% dietary inclusion); breeding infrastructure (love cage 70x70x86 cm with LED mating lamp, pupation cage 66x75x67 cm); no larval growth rate, survival, bioconversion ratio, or waste-reduction % reported anywhere in the paper. Gas exchange dynamics (AN, Tr, gs, iWUE) at D20/D45/D70 (Figure 6, no dedicated schema column). Chlorophyll a fluorescence/OJIP parameters (Fv/Fm, PItotal, PIabs, Sm, DIo/RC, ABS/RC, TRo/RC, ETo/RC, 1-Vi, 1/Vi) at D20/D45/D80 (Table 3, Figure 5, no dedicated column). Weekly fruit yield time series and week-by-week significance groupings (Figure 4A, 7 harvest weeks). Full Table 4 statistical letter-groupings for pH/EC/NO3/NH4/PO4/K/Na/Ca among HP/CAP/DCAP (a/b superscripts). WUE full three-treatment table (HP 100.5, CAP 57.1, DCAP 92.9 kg cucumbers/m3 water used) - AP-side value used in this row’s WUE cell, other two given here for reference. Literature comparison figures (secondary, not this paper’s data): Shaw et al. 2022 tilapia SGR 1.76%/day and FCR 1.03 (BSF)/0.79 (fishmeal control); Limbu et al. 2022 tilapia SGR 2.1%/day and FCR 1.01 (full BSF replacement); Aslanidou et al. 2024 ~40% CAP cucumber yield reduction vs HP; Pinho et al. 2024 32% fertilizer reduction in DCAP lettuce (FUE not estimated by that paper). Ethics/animal-welfare approval details (EU Directive 2010/63/EU; Animal Care and Use Ethics Committee approval 242627/28-05-2024; facility EL-43BIO/exp-02). Funding text appears truncated in the source PDF after ’…RESEARCH AND APPLICATION.’ with no further detail; transcribed as printed.
Plant Measurements
| Trial | System | Category | Analyte | Value | Unit | Sig. | Location |
|---|---|---|---|---|---|---|---|
| levizouCircularTriTrophicSystem2025-T1 | AP | mineral | Potassium (K+) | 7.4-11.6 | mg/g DW | NR | Results 3.6, p.11 |
| levizouCircularTriTrophicSystem2025-T1 | HYD | mineral | Potassium (K+) | 29.3 | mg/g DW | NR | Results 3.6, p.11 |
| levizouCircularTriTrophicSystem2025-T2 | AP | mineral | Potassium (K+) | 33.3 | mg/g DW | NR | Results 3.6, p.11 |
| levizouCircularTriTrophicSystem2025-T2 | HYD | mineral | Potassium (K+) | 29.3 | mg/g DW | NR | Results 3.6, p.11 |