Aquaponics: The Integration of Fish and Vegetable Culture in Recirculating Systems

Metadata

Cite key: rakocyAquaponicsIntegrationFish1994
Item type: Conference Paper (Caribbean Food Crops Society Proceedings; confirmed conferencePaper in zotero-export.csv)
Authors: J.E. Rakocy
Affiliation: University of the Virgin Islands, Agricultural Experiment Station, RR 2, Box 10,000, Kingshill, VI 00850
Journal: Proceedings of the Caribbean Food Crops Society, Vol. XXX / 30 (1994), pp. 101-108 (30th Annual Meeting, St. Thomas, U.S.V.I.)
Date: 1994 (no month stated in the PDF or in zotero-export.csv)
Date added: 2021-06-27 (from zotero-export.csv)
DOI: no DOI found — checked page 1 of the PDF (none printed) and zotero-export.csv (DOI field empty for this record, Key UZCVWVX6)
Funding: [not reported] — no acknowledgements or funding section anywhere in the 8-page proceedings paper
URL: NR — no URL in the PDF or in zotero-export.csv. A web search located a probable public copy at AgEconSearch (ageconsearch.umn.edu/record/258746, also mirrored on EconPapers/RePEc), but per the prime directive this was not entered in the frontmatter url field since it is neither stated in the PDF nor returned by a DOI lookup — noted here only as a pointer for the user.
PDF: Rakocy - AQUAPON1CS THE INTEGRATION OF FISH AND VEGETABLE .pdf — the “1CS” in the filename is an OCR/typo artifact for “ICS”; the PDF’s own printed title correctly reads “AQUAPONICS” (verified on p.101/p.2 of the PDF).

Opinion

A frequently-cited, historically important description of the UVI raft-aquaponics design — the same system architecture (fish tank / clarifier / raft hydroponic tanks / biofilter-in-hydroponic-tank / sump) that underlies most of the modern commercial aquaponics literature — but it is not itself a data paper. There is no Materials and Methods section, no defined treatments, no replication, and no statistical test anywhere in the eight pages; most of the quantitative claims that read like “results” are either (a) narrative production figures from operating the author’s own commercial-scale system, given once, with no SD/SE/n, or (b) explicitly attributed to the author’s own earlier, separately-published controlled studies (Rakocy 1989’s plant:fish ratio trial; Rakocy, Hargreaves and Bailey 1993’s nutrient-accumulation study), which are the papers that should actually be cited for those specific findings. Where I could check the paper’s own arithmetic (yield-per-plant vs. yield-per-m2 implying a planting density; TDS accumulation rate vs. the stated critical-TDS feed-loading range) it is internally consistent, which is reassuring given a fairly heavily OCR-corrupted text layer (see Extraction notes). Valuable to keep in the vault as the primary historical/technical reference for “the UVI system” that many later papers cite by name, but not a source of extractable trial data in its own right.

Abstract

A commercial-scale, aquaponic system for the intensive production of tilapia and hydroponic vegetables has been developed at the Virgin Islands Agricultural Experiment Station. The system is well suited for Caribbean islands and other tropical regions where fresh water is scarce or level farm land is limited. It consists of a fish rearing tank, a clarifier, two hydroponic tanks and a reservoir, and is reliable, productive and easy to operate. Water continuously circulates between the fish and hydroponic components. The fish grow rapidly on a pelleted diet that is high in protein. Waste from the fish provides most of the nutrients required by vegetables. The vegetables recover these nutrients as a valuable by-product and purify the water. The presence of both fish and plants creates a very stable growing environment, while high nutrient levels and unlimited water promote rapid vegetable growth. This system is economical because nutrient costs are reduced, the need for expensive filtration devices is eliminated, component operating and infrastructural costs are shared, land requirements are small, water is conserved, and environmental impacts are minimized.

Summary

This is a technical/design paper describing the raft-based (floating hydroponics) aquaponic system developed at the University of the Virgin Islands (UVI) Agricultural Experiment Station, presented at the 1994 Caribbean Food Crops Society meeting. Rakocy walks through the full system architecture at both experimental scale (12.3 m3 fish tank, ~1.9 m3 clarifier, two ~2.1 m3 raft hydroponic tanks, 1.4 m3 sump, 12.8 m2 growing area) and commercial scale (larger reservoir and tanks, 71.4 m2 growing area, sustaining 400 kg of tilapia), covering solids removal, biofiltration (including a finding that the raft hydroponic tanks alone can nitrify adequately once the separate rotating biological contactors were removed), pH/base management, nutrient accumulation as TDS, species selection (tilapia for fish; lettuce as the main crop, with tomato, Chinese cabbage, pak choi, peppers, basil, chives, beans and celery as other candidates), a continuous “stock-splitting” fish-rearing method, staggered vs. batch crop production scheduling, and pest/disease management under a near-total pesticide restriction. It reports one-off, unreplicated production figures for the commercial system (tomato, lettuce, Chinese cabbage and pak choi yields; ~1280 kg tilapia/year capacity) and closes with a qualitative economic/marketing outlook that points to a companion in-press economic analysis (Bailey et al. 1994) rather than presenting cost/revenue data itself. No hydroponic-only control, no treatment comparison, and no formal statistics appear anywhere in the paper — several of its most quotable numbers (the 1.9:1 optimum plant-to-fish ratio; the ~200 g TDS/kg feed accumulation rate) are explicitly sourced to the author’s own earlier, separately published experimental studies rather than derived here.

Experiment data

Location: University of the Virgin Islands (UVI) Agricultural Experiment Station, St. Croix, U.S. Virgin Islands. Both the experimental-scale and commercial-scale systems described are sited outdoors, uncovered, with only the fish rearing tank shaded by an opaque canopy.

Design: Descriptive engineering/case-study account of one research program’s system, not a controlled comparison. No defined treatments, no randomization, no replication, and no statistical test are reported in this paper itself. Several of the paper’s most specific quantitative claims are explicitly attributed to the author’s own earlier controlled studies rather than generated here: the 1.9-plants:1-fish optimum ratio to Rakocy 1989 (a six-ratio trial, 1.2-7.5), and the ~200 g TDS/kg dry feed accumulation rate to Rakocy, Hargreaves and Bailey 1993. One embedded manipulation is described narratively but without numbers: “Midway through a production trial, the RBCs [rotating biological contactors] were removed to determine if the hydroponic tanks could provide adequate nitrification at the design feeding rate… Low ammonia and nitrite levels were maintained” (p.104-105) — a real before/after comparison, but reported only as a qualitative outcome.

Replicates / n: NR — no replicate count or n is given for any figure this paper reports as its own (vegetable yields, TDS/feed relationship, tilapia production capacity, marketing-outlet counts).

Duration: Varies by subsystem and is never unified into a single “trial duration”: continuous 24-week fish stock-splitting production cycle (1,000 fingerlings stocked, 250 removed every 6 weeks); tomato “16-week production period”; Chinese cabbage and pak choi “4-week production cycle from transplant stage”; lettuce cycles of 4-5 weeks depending on cultivar; a 3- or 4-week staggered lettuce production schedule for the commercial system generally.

Organisms: Nile tilapia (Oreochromis niloticus) and Florida red tilapia hybrids / Lettuce (Lactuca sativa) (cvs. Montello, Parris Island, Sierra), tomato (cvs. Sunny, Floradade), Chinese cabbage (‘50-Day Hybrid’), pak choi (‘Le Choi’)

Statistics: None reported anywhere in the paper — no significance test, no SD/SE, no p-value.

System scale (experimental): 12.3 m3 fish rearing tank + ~1.9 m3 clarifier + two ~2.1 m3 raft hydroponic tanks (6.10 m L x 1.22 m W x 0.28 m D) + 1.4 m3 reservoir; 12.8 m2 total growing area; total operating water volume 17.3 m3 (see Extraction notes on the volume arithmetic and OCR unit corruption).

System scale (commercial): Larger reservoir (6.8 m3) and hydroponic tanks (29.59 m L x 1.27 m W x 0.41 m D); 71.4 m2 total growing area; minimum water volume ~37.1 m3; has sustained 400 kg of tilapia biomass on an 8 kg/day maximum feeding ration, with a 4-6 kg/day ration used as the actual design point for staggered bibb-lettuce production (2,112 heads at 29.6 heads/m2).

Vegetable yields (commercial system, single unreplicated figures): Tomato ‘Sunny’ 10.1 kg/plant (18.4 kg/m2), ‘Floradade’ 9.0 kg/plant (16.3 kg/m2) over 16 weeks; Chinese cabbage ‘50-Day Hybrid’ 638 g/plant (11.3 kg/m2/crop) and pak choi ‘Le Choi’ 508 g/plant (8.7 kg/m2/crop) over 4 weeks; lettuce ‘Montello’ 786 g/plant (12.7 kg/m2, 5 weeks), ‘Parris Island’ 660 g/plant (10.7 kg/m2, 4 weeks), ‘Sierra’ 522 g/plant (8.4 kg/m2, 4 weeks).

Fish production capacity: Commercial system “capable of producing 1280 kg of tilapia annually” on a complete 32%-protein floating-pellet diet under the stock-splitting method.

Nutrient accumulation: Conductivity/TDS rises ~200 g TDS/kg dry feed applied at low water exchange (<1% system volume/day) [secondary, cites Rakocy, Hargreaves and Bailey 1993]; critical TDS (~2000 mg/L, 3125 micromhos/cm) reached after 10-20 kg feed/m3 system volume depending on plant uptake — internally consistent with the 200 g/kg figure at the low (10 kg/m3, no plant removal) end (see Extraction notes).

System design: experimental vs. commercial scale

This paper: Describes two physical systems built at UVI: a smaller experimental-scale system (12.3 m3 fish tank, ~1.9 m3 clarifier, two ~2.1 m3 raft hydroponic tanks, 1.4 m3 sump, 12.8 m2 growing area, 17.3 m3 total operating water volume) and a larger commercial-scale system (larger reservoir and hydroponic tanks, 71.4 m2 growing area, ~37.1 m3 minimum water volume) that has sustained 400 kg of tilapia. The stated “optimum arrangement” of an aquaponic system (fish tank, solids-removal unit, biofilter, vegetable production unit, reservoir) is attributed to the author’s own prior review paper rather than argued fresh here.

Compared with:

todo Rakocy and Hargreaves 1993 — source of the “optimum arrangement of system components” claim (fish tank + solids removal + biofilter + vegetable unit + reservoir) that this paper’s System Design section opens with (p.102).
todo Nair, Rakocy and Hargreaves 1985 — source of the “only 25-30% of feed nutrients are retained by fish as weight gain” figure used to motivate why hydroponics is worth coupling to fish culture (p.101, Introduction). [secondary]

Plant-to-fish ratio and hydroponic sizing

This paper: States the critical determinant for sizing the hydroponic component is the daily feed (nutrient) input, and that “the optimum ratio between lettuce and tilapia was determined by evaluating six plant-to-fish ratios ranging from 1.2 to 7.5,” with maximum lettuce production (3.1 kg/m2/crop) at 1.9 plants:1 fish (equivalent to 2.4 g feed/plant/day). This entire finding is explicitly attributed to a separate, earlier paper and is not re-derived or re-tested here.

Compared with:

todo Rakocy 1989 — the six-ratio (1.2-7.5) plant:fish trial that produced the 1.9:1 optimum and 3.1 kg/m2/crop maximum lettuce yield figure this paper builds its hydroponic-sizing guidance on (p.102). [secondary, cites Rakocy 1989 — not this paper’s own data]

Water quality management and biofiltration

This paper: Describes pH management (daily monitoring, small frequent base additions of KOH/CaO to hold pH near 7.0, avoiding NaHCO3 because of Na+ toxicity to plants) and a biofiltration comparison embedded in running text rather than a table: the UVI system initially ran two ~93 m2 rotating biological contactors (RBCs) after the clarifier, in addition to ~126 m2 of surface area per hydroponic tank; partway through a production trial the RBCs were removed to test whether the hydroponic tanks alone could nitrify adequately at the 4-6 kg/day design feeding rate, and “low ammonia and nitrite levels were maintained,” which the author reads as showing raft hydroponics can eliminate the need for a separate biofilter at the optimum plant:fish ratio. No before/after ammonia or nitrite values are given, only the qualitative outcome.

Compared with: (no external comparison given for this specific RBC-removal finding; it is presented as this program’s own operational result)

Nutrient accumulation and total dissolved solids

This paper: States conductivity/TDS increases at “approximately 200 g as TDS/kg of dry weight of feed applied” under low water exchange (<1% system volume/day), with critical conductivity (~2000 mg/L TDS, 3125 micromhos/cm) reached “after the addition of 10 to 20 kg of feed/m3 of system volume depending on the quantity of plant growth.” Recomputing this as a check (not entered as a derived value anywhere): 200 g/kg x 10 kg/m3 = 2000 g/m3 = 2000 mg/L, exactly matching the stated critical TDS at the low end of the range; the paper’s own caveat (“depending on the quantity of plant growth”) explains why more feed (up to 20 kg/m3) can be added before hitting the same critical TDS when plants are actively removing nutrients along the way. The two figures are therefore internally consistent, not contradictory — recorded here as a confirmation check per SCHEMA.md, not as a cell value (this paper produces no trials.csv row).

Compared with:

todo Rakocy, Hargreaves and Bailey 1993 — source of the ~200 g TDS/kg dry feed accumulation-rate figure (p.103-104). [secondary, cites Rakocy et al. 1993 — not this paper’s own data]

Fish stock management and production capacity

This paper: Describes a “stock-splitting” continuous rearing method (Van Gorder 1991): the rearing tank is stocked with 1,000 fingerlings, and a portion (250) is removed every 6 weeks, holding the feeding rate roughly constant across a 24-week production cycle. On a complete 32%-protein floating-pellet diet with unrestricted demand-feeder access, the commercial-scale system is stated as “capable of producing 1280 kg of tilapia annually.” Fish are tested for off-flavor before sale and purged in clean water for a week if needed. No FCR, SGR, survival rate, or individual fish weight data are given anywhere in the paper.

Compared with:

todo Van Gorder 1991 — source of the stock-splitting continuous-loading method used at UVI (p.105).

Vegetable yields

This paper: Reports single, unreplicated commercial-system yield figures per crop: tomato ‘Sunny’ 10.1 kg/plant (18.4 kg/m2) and ‘Floradade’ 9.0 kg/plant (16.3 kg/m2) over a 16-week fruiting period; Chinese cabbage ‘50-Day Hybrid’ 638 g/plant (11.3 kg/m2/crop) and pak choi ‘Le Choi’ 508 g/plant (8.7 kg/m2/crop) over a 4-week cycle; lettuce ‘Montello’ 786 g/plant (12.7 kg/m2, 5 weeks), ‘Parris Island’ 660 g/plant (10.7 kg/m2, 4 weeks) and ‘Sierra’ 522 g/plant (8.4 kg/m2, 4 weeks). As an internal-consistency check (not a derived cell value — this paper has no trials.csv row): dividing each area yield by its per-plant yield recovers a planting density of ~1.8 plants/m2 for both tomato cultivars, ~17-18 plants/m2 for both cabbage-family crops, and ~16.1-16.2 plants/m2 for all three lettuce cultivars — the latter matching the low end of the paper’s own separately stated “16.2 to 29.6 plants/m2” planting-density range almost exactly. The author’s only qualitative comparison offered is that these yields are “greater than yields from local field crops and comparable to average yields of soilless culture,” without naming a specific comparator study.

Compared with: (no specific external yield study named in this section; the “comparable to soilless culture” claim is asserted without a citation)

Pest and disease management

This paper: Lists pests observed in UVI’s systems by crop (tomato: spider mite, russet mite, hornworm, fall armyworm, pinworm, aphid, leaf miner; lettuce: fall armyworm, cabbage looper; pak choi/Chinese cabbage: aphids) and states that most pesticides and fish therapeutants cannot be used because of fish toxicity or lack of aquaculture approval, restricting control to biological agents (weekly Bacillus thuringiensis spraying against armyworms/loopers on lettuce), traps, resistant varieties, screening, and cultural practices. Also asserts aquaponic systems may be more resistant than standard hydroponics to some diseases, attributed speculatively to organic matter in the culture water supporting a more diverse, stable microbial community — offered as a hypothesis, not tested here.

Compared with: (no external literature comparison given for the pest/disease observations)

Economic and marketing outlook

This paper: States a partial economic analysis of the commercial-scale system “indicates that the system has profit potential in St. Croix,” citing a companion paper for the actual cost/revenue figures; explicitly notes that analysis covered only one production unit’s capital/operating costs and revenue, not full infrastructural/administrative costs, and that a fuller evaluation awaited a subsequent 3x scale-up. Separately reports (narratively, no dates or volumes) that a year-plus marketing study sold lettuce to 42 outlets and tilapia to 21 outlets in St. Croix.

Compared with:

todo Bailey, Rakocy, Cole and Shultz 1994 — the economic analysis this paper’s profit-potential claim is based on; not this paper’s own cost/revenue data (p.107). [secondary]

Linked claims

Citations to chase

Extraction notes

Type classification (case-study vs. technical-note vs. observational vs. experiment): Recorded as case-study (“Single system or site described in depth,” SCHEMA.md Part 1, Primary research). The paper is built almost entirely around one specific site/system — the UVI Agricultural Experiment Station’s aquaponic system, at both its experimental and commercial scale — and walks through it in depth across design, component sizing, hydroponic subsystem, water-quality management, nutrient accumulation, species selection, stock management, crop scheduling, pest control, yields, and economics/marketing, all in the first person as the program’s own work. This rules out narrative-review/policy in my judgment: the paper is not primarily synthesizing or arguing from others’ literature (its handful of external citations are almost all the author’s own earlier UVI studies, or contemporaneous technical-workshop proceedings cited for a specific engineering component, not a literature synthesis), and it does report data the authors’ own program collected (vegetable yields, TDS behavior, production capacity, an RBC-removal comparison) — failing the review test in SCHEMA.md’s decision rule 1 (“did the authors collect data themselves?” — yes, for most of what is reported as this paper’s own). It also rules out experiment/quasi-experiment/field-trial: there is no Materials and Methods section, no defined treatment structure, no replication, and no statistical test anywhere in the paper; the one implied manipulation (RBC removal, p.104-105) is described only qualitatively with no numbers, n, or before/after values. exploratory was considered and rejected: exploratory connotes a pilot/proof-of-concept on a new question, whereas this describes a mature, already-commercial-scale, multi-year operating program across many facets, which is a better match for “single system… described in depth” than for a pilot study. technical-note (“short descriptive report, no hypothesis”) was also considered — and is the closest secondary-research-flavored alternative, matching the precedent set by goddekImprovingNutrientWater2020 in this vault — but that type is reserved in SCHEMA.md’s Non-research category, and this paper does report original production figures from the author’s own system (not just propose a design change, as the Goddek paper does), so case-study’s “single system described in depth” was judged the better fit within Primary research. This is a judgment call at a genuine boundary in the schema; the alternative classifications and their rationale are recorded here in full per CLAUDE.md so the decision can be revisited.

Template block choice: Used the Experiment data callout rather than Review scope, following the precedent set for hornTranslatingEnvironmentalPotential2024 (type observational, also Primary research with no manipulation, also given the Experiment data block). CLAUDE.md’s binary experiment-vs-review block mapping does not explicitly cover case-study/observational/technical-note; the working convention observed across this vault’s batches is Primary-research types (however light on formal design) get Experiment data, and Secondary/Non-research types get Review scope (e.g. technical-note in goddekImprovingNutrientWater2020).

CSV outputs: Per CLAUDE.md/SCHEMA.md, only experiment, quasi-experiment, field-trial, and exploratory papers produce trials.csv/plant.csv rows. case-study is not on that list, so out/rakocyAquaponicsIntegrationFish1994.trials.csv and out/rakocyAquaponicsIntegrationFish1994.plant.csv are written header-only, with zero data rows. This is a substantive loss for this specific paper — the vegetable-yield figures (tomato/lettuce/cabbage/pak choi, kg/m2 and kg-or-g/plant) are real, named-cultivar production numbers that would otherwise be exactly the shape of data trials.csv is built to hold — but they are not paired with any hydroponic control, replicate count, or statistical test, so a trials.csv row would overstate their evidentiary weight relative to the paper’s actual (uncontrolled, unreplicated) design. All figures are preserved narratively in the “Vegetable yields” section and the Experiment data callout above instead.

Metadata sourcing: zotero-export.csv is present in this vault and contains a matching row (Key UZCVWVX6, title matches exactly, conferencePaper item type, author “Rakocy, James E”, year 1994, empty DOI and URL fields, Date Added 2021-06-27). Per CLAUDE.md, this Zotero row was used in preference to the PDF for title/author/year/DOI; it confirms no DOI is recorded for this item. No Crossref lookup was attempted (no DOI to look up). A web search located a probable public-access copy at AgEconSearch (ageconsearch.umn.edu/record/258746, mirrored on EconPapers/RePEc as agscfcs94/258746), stating the paper was presented at the 30th CFCS meeting, 31 Jul-5 Aug 1994 — this is consistent with the PDF’s own cover page (“THIRTIETH ANNUAL MEETING 1994… Vol. XXX”) but is not itself in the PDF or in Zotero, so it was not entered into the frontmatter url field per the prime directive (every value must come from the PDF or a DOI lookup); it is recorded here only as a locator for anyone wanting to verify online.

PDF quality — OCR unit-symbol corruption: The text layer is a fairly clean OCR of a photocopied/scanned two-column proceedings volume, but superscript unit symbols (m2, m3) are corrupted in several places, almost certainly because the scanner/OCR misread superscript digits as other glyphs: “a 1.9-m5 clarifier” (p.102) is physically read as 1.9 m3 (m5 is not a real volume unit); “two 2.1-m1 hydroponic tanks” (p.102) is read as 2.1 m3, confirmed by back-calculating the stated dimensions (6.10 m x 1.22 m x 0.28 m = 2.084 m3, matching); “minimum water volume of 37.1 m!” (p.102-103) is read as 37.1 m3 (roughly consistent with 2 x [29.59 x 1.27 x 0.41 m3] + 6.8 m3 reservoir ≈ 37.6 m3, order-of-magnitude match); “two 93-rrr RBCs” (p.105) is read as 93 m2. These are reported in the note above using the corrected (m2/m3) reading, since the physical/dimensional context makes the intended unit unambiguous — this is treated as an OCR-artifact correction rather than a data judgment call, analogous to CLAUDE.md’s DMS-vs-decimal-degree correction principle, and is flagged here for transparency rather than silently fixed. The filename’s “AQUAPON1CS” (digit 1 for letter I) is the same class of OCR artifact and does not appear in the PDF’s own printed title, which reads “AQUAPONICS” correctly (p.101/p.2). One further minor OCR/typo: “Iron is supplemented every 3 weeks by adding 2/mg/L” (p.106) almost certainly means “2 mg/L” (stray slash); not flagged with a severity tag since there is no second, conflicting value in the paper to weigh it against — see SCHEMA.md’s distinction between BLOCK/MATERIAL/CHECK (competing values) and an unambiguous single-value typo.

Internal consistency checks performed (not contradictions, not entered as derived cell values — this paper has no trials.csv row to enter them into):

TDS accumulation rate (~200 g TDS/kg dry feed, secondary from Rakocy et al. 1993) x the stated 10 kg feed/m3 lower bound = 2000 g/m3 = 2000 mg/L, exactly matching the separately stated “critical conductivity… approximately 2000 mg/L as TDS” — confirms the two figures are consistent, with the paper’s own “depending on the quantity of plant growth” caveat explaining why the upper bound (20 kg/m3) can be double the arithmetic no-removal figure.
Tomato yields: 10.1 kg/plant / 18.4 kg/m2 (‘Sunny’) and 9.0 kg/plant / 16.3 kg/m2 (‘Floradade’) both imply ~1.8 plants/m2, consistent between cultivars.
Chinese cabbage (638 g/plant / 11.3 kg/m2) and pak choi (508 g/plant / 8.7 kg/m2) both imply ~17-18 plants/m2.
All three lettuce cultivars (786 g/12.7 kg/m2; 660 g/10.7 kg/m2; 522 g/8.4 kg/m2) imply ~16.1-16.2 plants/m2, matching the low end of the paper’s separately stated “16.2 to 29.6 plants/m2” planting-density range almost exactly, and the “2,112 heads… density = 29.6 heads/m2” example implies a growing area of ~71.4 m2, matching the stated commercial growing area exactly.
No arithmetic inconsistency was found between the experimental system’s summed nominal component volumes (12.3 + ~1.9 + ~2.1 + ~2.1 + 1.4 ≈ 19.8 m3) and the separately stated “total water volume during operation” of 17.3 m3; these are judged to be two different quantities (summed tank capacities vs. actual operating fill level, which is very plausibly below full nominal capacity) rather than a conflict, per SCHEMA.md’s “setpoint vs. measured” non-contradiction principle, so this was not flagged with a severity tag.
The two different commercial-system daily feeding-ration figures — “8 kg” (framed as the maximum sustained ration supporting 400 kg of fish, used to size the solids-removal/biofilter capacity) vs. “4 to 6 kg” (used three times, always in the context of sizing/operating the hydroponic component for a specific staggered lettuce production volume) — are not treated as conflicting: the paper’s own surrounding text assigns each figure to a distinct purpose (peak biofilter-design capacity vs. actual hydroponic-sizing operating point), which is the “target vs. achieved, not a conflict” case SCHEMA.md explicitly excludes from the contradiction-flagging apparatus.

No ⚠️BLOCK / ⚠️MATERIAL / ⚠️CHECK contradictions found. The paper is short, internally consistent wherever its own arithmetic could be checked, and (per the type classification above) contains no trials.csv cells for a contradiction to attach to in the first place.

[not reported], grouped:

Funding/acknowledgements — no such section anywhere in the paper.
Any FCR, SGR, survival rate, or individual fish weight (initial or final) for the tilapia — only aggregate biomass (400 kg sustained) and an annual production-capacity estimate (1280 kg/year) are given.
Any replicate count, SD, SE, or p-value for any figure in the paper.
Water temperature, dissolved-oxygen concentration (only the design threshold “>5 mg/L” is stated, not a measured value), and a numeric pH value (only “near 7.0” as a management target, and “<7.0” as the nitrification-impairment threshold).
Coordinates for the UVI Agricultural Experiment Station — not stated in the PDF; not recorded, per the prime directive, even though the site (St. Croix, USVI) is well known.
A restated month/day for “1994” in either the PDF or zotero-export.csv’s Date field.

Secondary-figure flags (SCHEMA.md: “numbers from review/other papers are secondary”; these five figures are this paper’s own citations of the author’s earlier work or a companion paper, not new data generated here): the 1.9:1 plant:fish ratio / 3.1 kg/m2 lettuce yield (Rakocy 1989); the “optimum arrangement of system components” claim (Rakocy and Hargreaves 1993); the ~200 g TDS/kg feed accumulation rate (Rakocy, Hargreaves and Bailey 1993); the 25-30% feed-nutrient-retention-by-fish figure (Nair, Rakocy and Hargreaves 1985); the profit-potential claim (Bailey et al. 1994, described as a companion “Book of Abstracts” entry, i.e. likely still unpublished/in-press at the time this paper appeared).

Reconciliation flag — same UVI research program as two other un-processed Rakocy PDFs in this vault: pdfs/ also contains “Rakocy et al. - 2004 - AQUAPONIC PRODUCTION OF TILAPIA AND BASIL COMPARI…pdf” and “Rakocy et al. - UPDATE ON TILAPIA AND VEGETABLE PRODUCTION IN THE…pdf”, neither processed in this batch. This 1994 paper describes what is very plausibly the same underlying UVI system lineage (raft hydroponics, fish-tank/clarifier/biofilter/reservoir architecture, tilapia + lettuce/tomato/cabbage) that those two later filenames suggest, and this paper is itself cited by title/finding (as “Rakocy 1989,” “Rakocy and Hargreaves 1993,” “Rakocy et al. 1993”) inside several other already-processed vault notes (e.g. buzbyScalingAquaponicSystems2014, caloneImprovingWaterManagement2019, castillocastellanosImplementationExperimentalNutrient2016, delaideLettuceLactucaSativa2016, fernandezcabanasComparativeAnalysisHorizontal2020) as an UVI-design reference point. When the two other Rakocy PDFs are eventually processed, they should be cross-checked against this note for (a) whether system dimensions/component ratios are restated consistently across all three UVI papers, and (b) whether any of the “Compared with” citations chased here (especially Rakocy 1989 and Rakocy et al. 1993, both cited internally by this paper without full bibliographic detail beyond what’s in this paper’s own References list) turn out to be the same underlying studies partially re-reported in the 2004 or “Update” papers. Not resolved here per the task’s instruction to read this paper on its own merits and only flag the overlap.

Tags — judgment calls:

Meta/Type/Case-Study — new leaf under Meta/Type/; existing leaves in this vault are Experiment, Exploratory, Field-trial, Meta-analysis, Modelling, Narrative-Review, Observational, Policy, Quasi-experiment, Review, Systematic-Review, Technical-Note — none of these is Case-Study, so this introduces a new one, Title-Case and hyphenated per the existing multi-word convention (e.g. Quasi-experiment, though note that one is not Title-Cased — Case-Study follows the Field-trial/Technical-Note capitalization pattern instead, which is the more common style in the set).
Meta/Region/Caribbean — new leaf; existing region leaves are Africa, Central-/East-/Middle-/North-/South-/Southeast-Asia-type compounds, China, Europe, Global, NorthAmerica/North-America, Oceania. None is Caribbean-specific; the U.S. Virgin Islands is sometimes folded into “North America” in other vault notes’ region tagging, but given this paper’s own framing is explicitly and repeatedly about “Caribbean islands” (title of the Introduction’s argument, and the journal itself is the Caribbean Food Crops Society), a dedicated Meta/Region/Caribbean leaf was judged the more precise, paper-faithful choice over folding it into the existing NorthAmerica leaf.
Meta/Fish/Tilapia — reused (Nile tilapia and Florida red tilapia hybrids are both actually reared and reported on).
Meta/Plant/Tomato, Meta/Plant/Lettuce, Meta/Plant/Pak-Choi — all reused from existing vault spellings (Pak-Choi matches leeComparativeStudyGrowth2019’s leaf exactly).
Meta/Plant/Chinese-Cabbage — new leaf; no existing cabbage-specific facet found in the vault (a stray textual mention of “Brassica” in one other note’s Extraction notes was not an actually-applied tag on any note, so it was not reused as a generic catch-all).
Not tagged despite being named in the paper: cucumber, pepper, basil, chives, bush beans, celery — all listed only as “other crops with potential” (Suitable Species section) with no yield or performance data reported for any of them, so per CLAUDE.md’s “only tag an organism if the paper studied it” they are excluded.

New wikilink targets introduced: J.E. Rakocy (no existing author-wikilink note found in the vault under any spelling; many other notes cite “Rakocy” only inside #todo literature-comparison bullets, never as a linked author, so this is the vault’s first author-wikilink for him). Reused existing canonical forms: Nile tilapia (Oreochromis niloticus), Lettuce (Lactuca sativa) (both matching pantanellaAquaponicsHydroponicsProduction2012’s spelling).

Source: Rakocy - AQUAPON1CS THE INTEGRATION OF FISH AND VEGETABLE .pdf

Quartz 5

Explorer

Aquaponics: The Integration of Fish and Vegetable Culture in Recirculating Systems

Aquaponics: The Integration of Fish and Vegetable Culture in Recirculating Systems

Metadata

Abstract

Summary

System design: experimental vs. commercial scale

Plant-to-fish ratio and hydroponic sizing

Water quality management and biofiltration

Nutrient accumulation and total dissolved solids

Fish stock management and production capacity

Vegetable yields

Pest and disease management

Economic and marketing outlook

Linked claims

Citations to chase

Extraction notes

Graph View

Table of Contents

Backlinks