A geoduck hatchery record follows a cohort from broodstock spawning through fertilised eggs, swimming larvae, settlement, early benthic growth and the seed counted for planting. To translate its numbers correctly, keep each biological stage and its denominator separate. Eggs released are not necessarily fertilised eggs; fertilised eggs are not all surviving larvae; larvae transferred to sand are not all recovered juveniles; and juveniles counted in a tank are not all market-ready adults. Good reporting identifies the date, unit, cohort, sampling method and measured stage for every number.
For a fictional example, imagine 2.0 million eggs counted at collection, of which an estimated 1.6 million show the hatchery’s accepted fertilisation criteria. The estimated fertilisation proportion is 1.6 divided by 2.0, or 80 percent. If 960,000 living larvae remain at a later census, survival from the estimated fertilised cohort to that census is 60 percent, while survival relative to the original collected eggs is 48 percent. Both percentages can be arithmetically correct; they describe different stages. Neither is the final seed yield.
The essential reading sequence is to identify species and broodstock group, document spawn and collection windows, separate egg count from fertilisation assessment, determine whether larval counts are concentrations or totals, record tank transfers and losses, define the competent larval and settled juvenile stages, document sand-nursery sampling, specify the shell-size threshold for usable seed, reconcile numbers across containers, and present final recoverable seed with explicit uncertainty. No adult harvest forecast should be invented merely from a promising hatchery survival percentage.
What makes geoduck hatchery records distinctive
Pacific geoduck are deeply burrowing bivalves. Adults produce eggs and sperm that are released into the water, while developing larvae initially live in the water column before settlement and benthic life. Hatcheries can condition broodstock, collect spawned gametes, rear larvae on cultured algae and move juveniles into shallow sand-based nursery systems. NOAA’s geoduck aquaculture overview describes early larval rearing and subsequent seed production. Details such as temperature, feed species and day-to-stage timing depend on the hatchery, species and conditions; do not treat a published example as a universal operating recipe.
A hatchery cohort is a traceable group produced within a defined spawning event or batch. One female may contribute to several collection containers, or gametes from multiple parents may be pooled. A label such as Batch 24A needs a documented parentage or pooling record to make later pedigree statements. If two containers share a batch number but one was transferred after an earlier mortality event, their starting denominators may differ. A table that retains cohort IDs and transfer dates prevents accidental addition of the same animals twice.
Broodstock conditioning concerns the animals before spawning. Reports may track origin, sex when known, maturity observations, holding conditions, feeding, temperature histories and response to stimulation. “Ready to spawn” is an assessed condition, not a guarantee that every adult will release gametes on schedule. “Spawned” may mean an observed release, while “contributed eggs” requires a quantified collection. Translating a readiness score as an observed egg yield would move information across stages that the evidence does not support.
An induced spawning procedure may use controlled environmental conditions or biological cues; its scientific explanation should distinguish the method used by the recorded hatchery from procedures described in older research. A research note reporting a certain temperature in one setting is not proof that the same temperature is appropriate for every geoduck stock. For an educational article, the defensible focus is reading the record: which cue was applied, when gametes were collected, how identity was preserved and what measured result followed. Unrecorded adjustments must not be invented.
Count eggs and assess fertilisation
A total egg estimate may be obtained by measuring concentration in a well-mixed suspension and multiplying by the container volume. If a carefully diluted representative subsample gives an estimated 250 eggs per millilitre in a 4,000-millilitre suspension, the estimated total is one million eggs. The calculation assumes sampling and mixing were appropriate. If the concentration was measured after a tenfold dilution, the analyst must restore the dilution factor before multiplying by the original volume, unless the lab instrument already reported corrected concentration.
Counting eggs released is different from identifying eggs that meet a laboratory’s evidence of fertilisation or normal development. A report might classify cleavage, visible embryonic stage or another validated assessment criterion. It should name the time after insemination and denominator of scored eggs. A percentage from 200 inspected eggs is a sample estimate, not proof that every one of two million eggs behaved identically. If morphology is uncertain or the scoring window is mismatched, the interpretation changes. Preserve the observation criterion rather than replacing it with a vague claim that the eggs were healthy.
Consider a synthetic assessment of 300 sampled eggs in which 246 meet the stated fertilisation criterion. The observed proportion is 246 divided by 300, or 82 percent. If the total collected egg estimate is 1.5 million, applying that proportion gives about 1.23 million criterion-positive eggs under an assumption that the sample is representative. The result is not a direct count of 1.23 million individual embryos. That distinction matters in a quality audit, because the total combines one concentration estimate with one sampled proportion and carries uncertainty from both.
In technical translation, write “estimated fertilisation percentage among eggs assessed at six hours” rather than simply “fertility 82 percent” if the study’s wording and timing support the longer form. Fertility might refer to animals, gametes or a specific assay, so it can be ambiguous. The number 82 percent also needs its denominator and sample size. A reader cannot judge whether 82 percent came from ten eggs or three hundred if the report removes that information. High percentages do not compensate for poorly documented sampling.
Concentration is not the same as total larvae
Larval counts are often recorded as organisms per millilitre or litre, whereas tank inventory is a total number. Multiplication by the actual water volume connects them. A concentration of 40 larvae per litre in a 3,000-litre tank implies about 120,000 larvae, assuming representative mixing and no special sampling correction. Reading “40 larvae/L” as forty larvae in the entire tank underestimates inventory three thousandfold. A careful table shows concentration, net culture volume, dilution factor, estimation time and derived total as distinct columns.
Water level can change when staff drain, clean or transfer tanks. If a measured concentration remains unchanged but water volume falls, the estimated total number falls proportionally in the calculation; that does not automatically mean larvae died. Some larvae may have been transferred elsewhere. The inventory should reconcile all destination containers and confirmed losses. Without a transfer ledger, a reader could call a routine tank division a biological collapse. The difference between a volume-based estimate and a physical fate is fundamental to interpreting hatchery data.
A hatchery may report larvae at multiple developmental stages. Stage names matter because feeding, behavior, settlement readiness and survival assessments can differ. “Competent larvae” means the cohort meets the study’s specified developmental criteria for settlement, not merely that larvae have survived a number of days. “Settled juveniles” describes a later stage and should not be equated with every larva stocked into a settlement tray. Methods should explain how stage was assessed, especially when sample timing is used as a substitute for individual observation.
Microalgae feed data introduce another denominator. A count of algal cells per millilitre may be a feed-stock concentration, a tank concentration after addition or a target ration. The quantity of cells added can be calculated by concentration times delivery volume, but feeding does not demonstrate ingestion. Uneaten algae, suspended cells and degraded material can change water quality. A record that states “fed 400 million cells” is not evidence that the cohort consumed all 400 million. Translate delivery, availability and measured intake as different concepts.
No conversion between eggs, larvae and seed is meaningful without dates. A sample drawn before routine transfer and a sample drawn after transfer can appear to show a sudden population change when the cohort has simply been divided. Record whether each inventory is before or after a water change, count adjustment, culling step or vessel movement. Date and stage are part of the measurement. A graph with a smooth line connecting noncomparable estimates might look scientific but conceal important operational discontinuities.
When larvae become competent to settle
Settlement is not a synonym for survival in the water column. A larva can remain alive but not yet show the developmental readiness or behavior needed to attach, metamorphose and enter a benthic nursery stage. Some protocols monitor morphology, swimming behavior and response to settlement substrates. The report should explain what competence means for that particular hatchery, the age and condition of the sampled cohort, and the observation method. A simple sentence that “all larvae settled after twenty days” should not be written unless a verified census actually supports it.
Published geoduck hatchery accounts report different timing under differing temperatures and approaches. NOAA’s aquaculture overview describes larvae settling within approximately three weeks in a general farming context, while older technical reports document specific developmental timings under defined conditions. A date interval from a publication is an example, not a law of nature. Reinterpreting a calendar-day target as a universal biological threshold can lead to false conclusions about whether an individual tank has failed. Read the actual developmental stage observations alongside the time series.
A settlement trial might stock 100,000 estimated competent larvae onto a prepared set of shallow-sand nursery containers. At a later assessment, 24,000 live settled juveniles are estimated using documented subsampling. The apparent transition proportion is 24,000 divided by 100,000, or 24 percent, but this combines two uncertain estimates and a particular definition of settled. It is not necessarily the true physiological probability of metamorphosis, since animals can be missed, transferred or lost between measurements. Keep both census dates in the percentage’s label.
Settlement plates, sand trays and tank surfaces are different sampling frames. A larva detected on a collector may still fail metamorphosis, and a juvenile that has burrowed into sand may not be visible in a surface check. If a study defines success by recovered juveniles at a specified age, do not call it instantaneous attachment rate. A useful report distinguishes attachment observed during the initial inspection, normal metamorphosis observed subsequently and survival after the first nursery interval. Each represents a different biological hurdle.
Temperature and carbonate chemistry can affect early development, but a tank’s poor outcome should not be attributed to one factor solely because it lies outside another study’s reported range. NOAA-hosted research on geoduck larvae exposed to altered carbon-dioxide conditions described developmental and physiological differences under the experiment’s specified pH treatments. That is evidence for the conditions investigated, not a calibrated operating threshold for every hatchery. A proper interpretation first describes the recorded water chemistry, developmental stage and experimental design, then carefully considers causal possibilities.
An apparent fall in larval concentration can arise through settlement, mortality, sampling changes, water replacement or movement into an unsampled part of the system. Likewise, a rise can reflect improved mixing or a correction to the dilution factor without new larvae being created. Never label every inventory difference as a death count. A production ledger should reconcile the estimate before transfer, quantities intentionally moved, quantities recovered in destination vessels and losses observed or inferred under a stated model.
The shallow-sand nursery and usable seed
After settlement, geoduck juveniles enter a benthic phase and may be reared in controlled sand environments before planting. A sand nursery is not simply a holding tank with a visible population floating in water. Juveniles may burrow and become hard to count directly. A standard operating procedure may use quadrats, sampled trays, sieves or carefully defined extraction methods. The word recovered matters: an extracted count is a measurement affected by recovery efficiency, not automatically a complete census of every animal that remains alive in the sand.
NOAA’s geoduck aquaculture overview describes seed reared to a shell dimension of roughly four to six millimetres before planting in an example production system. That does not mean every juvenile at four millimetres is healthy or approved for planting. Seed acceptance can involve size distribution, vigor, disease controls, transport conditions and local farm specifications. An educational translation should name the size dimension measured, such as shell length or another defined laboratory metric, and avoid assigning a universal quality threshold from one broad overview.
Suppose a fictional sand tray contains a measured 0.50 square metres of usable substrate. Technicians sample five separate 0.01-square-metre quadrats, recovering 90, 110, 95, 105 and 100 living juveniles. The sample total is 500 in 0.05 square metres, yielding an observed recovery density of 10,000 juveniles per square metre. Extending that density to the full 0.50 square metres gives an estimated 5,000 juveniles. The calculation is clear, but whether it is unbiased depends on the sample location rule and uniformity of burrowing animals.
That quadrat estimate can be wrong if samples were taken only from visible dense patches. Juvenile geoducks may distribute unevenly by water flow, sand depth or habitat preference. A systematic or randomized frame helps avoid a convenience-sample bias. If the sampled substrate was disturbed by routine cleaning and moved juveniles toward one side, the record should state that. A translation that omits sampling design and reports exactly 5,000 juveniles as a counted population would erase the principal uncertainty in the result.
Sieve mesh and extraction method matter because the youngest juveniles may be lost through the equipment while larger ones are retained. An observed recovery percentage should specify the method and target size class. If a retained fraction is mislabeled as overall survival, readers may infer biological mortality where mechanical recovery or size selectivity caused the difference. To assess recovery efficiency, a hatchery could conduct an independently designed test using a known group under comparable conditions. Without such a study, do not invent a correction factor.
Seed size is a distribution, not merely one average. A lot may have mean shell length 5.2 millimetres but include some juveniles below a minimum operational threshold. A useful quality table reports the number measured, the sampling method and relevant percentiles or grade counts. If acceptance requires at least four millimetres and 18 of 100 measured individuals fall below that level, the observed sample fraction meeting the size threshold is 82 percent. Applying that percentage to a lot total is an estimate whose representativeness must be defended.
Length and mass should not be casually exchanged. A measurement in millimetres describes a linear shell dimension; milligrams or grams describe weight. Some datasets record dry tissue mass, wet whole mass or combined substrate-associated material. A size class is not a feed ration and a length-frequency distribution does not give exact biomass unless a validated length–weight relationship is used. In technical translation, retain the dimension and basis for every metric, especially when the same word “size” can mean different things in successive stages.
Reading survival and yield with correct denominators
A hatchery may report egg-to-larva survival, larva-to-settlement recovery, settlement-to-seed survival and overall egg-to-usable-seed yield. These are separate ratios. If the stages use comparable cohort accounting, the product of the stage proportions can reproduce overall yield. But the multiplication is invalid if a stage denominator includes newly added animals from another batch or excludes transferred individuals without reconciliation. A traceable cohort ledger is the bridge from one ratio to the next.
Consider a complete invented chain: 2,000,000 collected eggs; 1,600,000 criterion-positive fertilised eggs; 960,000 living larvae at a defined census; 192,000 estimated settled juveniles; and 115,200 recovered seed above the specified size threshold. Stage proportions are 80 percent, 60 percent, 20 percent and 60 percent respectively. The overall yield is 115,200 divided by 2,000,000, or 5.76 percent. Multiplying 0.80 by 0.60 by 0.20 by 0.60 gives 0.0576, confirming the arithmetic under the assumptions of a single closed cohort.
The same figures support different statements depending on the denominator. Relative to fertilised eggs, usable seed is 7.2 percent. Relative to the later larval census, it is 12 percent. Relative to settled juveniles, it is 60 percent. None should be presented as the overall egg-to-seed yield. If a report supplies only one unlabeled “survival 60 percent” cell, the analyst must identify its stage interval before translating it into a public statement. Percentages with missing denominators are among the most serious errors in production dashboards.
Production yield and quality acceptance are also different. A nursery might physically recover 120,000 live juveniles while only 110,000 meet a shipment specification on the inspection day. The rejected or held 10,000 are not necessarily dead. They may be too small, in need of additional nursery time or subject to a quality hold. A translation should distinguish live recovery, graded acceptable seed, shipped seed and eventual planted seed. Those numbers measure different checkpoints; counting them all as harvest-ready animals would inflate output repeatedly.
Water-quality observations help explain variability but cannot be treated as causal proof without an appropriate design. A daily log may include temperature, salinity, dissolved oxygen, pH, total ammonia nitrogen, exchange rate and microbial observations. Missing values should remain missing, not be filled with invented normal conditions. A measured pH change can coincide with altered survival for multiple reasons. The responsible account states what was recorded and what hypothesis is being evaluated, rather than claiming that a single bad day directly caused all subsequent losses.
Worked audit: a complete imaginary hatchery batch
Start with a fictional single-cohort ledger whose initial event records 1,200,000 eggs in a measured suspension. A representative assessment identifies 900,000 meeting the hatchery’s fertilisation criterion, so the estimated proportion is 75 percent. A later living-larval inventory is 630,000. Survival from the estimated fertilised cohort to that census is 70 percent, while survival from the initial eggs to living larvae is 52.5 percent. These quantities are explicitly hypothetical and assume no additional eggs or larvae entered the cohort after the opening count.
The same imaginary batch yields 157,500 estimated settled juveniles, followed by 94,500 live individuals recovered from the sand nursery. Settlement-stage recovery relative to living larvae is 25 percent, while recovery from settled juveniles to nursery counts is 60 percent. Overall egg-to-recovered-seed yield is 94,500 divided by 1,200,000, or 7.875 percent. If only 85,050 meet the seed size requirement at shipment, the accepted seed fraction is 90 percent of recovered juveniles and 7.0875 percent of original eggs. Reporting all these as a single unqualified survival rate would conceal four distinct biological and operational transitions.
What changes if the hatchery later discovers that 10,000 previously counted nursery juveniles were included twice because two tray identifiers referred to one tray? The recovered inventory becomes 84,500 if the duplicate was a true extra count, and downstream fractions must be recomputed from this revised total. But the biological animals did not suddenly die when the error was corrected. The record needs an audit annotation stating that the change came from reconciliation of duplicate tray IDs. Calling it mortality would create a false husbandry diagnosis from a bookkeeping error.
Now suppose the juvenile count was not a complete census but an estimate expanded from substrate quadrats. A reported 94,500 may carry sampling error, and its significant figures should reflect the sampling method. A display reading of 94,500.000 animals invites a precision that is not supported by counts from a small area of sand. A responsible table can retain the calculation’s numerical value while describing it as an estimate, recording sample size and showing an uncertainty interval if the method supports one. Arithmetic correctness and statistical certainty are different ideas.
Worked audit: sample dilution and concentration
A technician counts a subsample that was diluted one volume of original suspension with nine volumes of clean diluent, making a tenfold final dilution. The diluted aliquot averages 32 larvae per millilitre. Under appropriate mixing and counting conditions, the original concentration estimate is 320 larvae per millilitre. In a 1,500-litre culture, there are 1,500,000 millilitres, so the naive expanded estimate would be 480,000,000 larvae. That very large total should trigger a plausibility and unit check, not automatic acceptance merely because multiplication was performed correctly.
For a more modest training example, suppose the original concentration after all dilution corrections is 0.32 larvae per millilitre rather than 320. In a 1,500-litre culture the total becomes 480,000 larvae. The three-orders-of-magnitude difference arises from a unit or dilution interpretation, not from a real biological event. Before publishing a translated inventory, check whether the laboratory instrument already applies the dilution, whether counts are per millilitre or per litre, and whether the tank volume was entered in litres or millilitres.
If a report uses 320 larvae per litre, the arithmetic in a 1,500-litre tank is 480,000, not 480 million. This is why repeating the unit alongside the number matters. A translator may think the letters mL and L are too minor to explain, but replacing one with the other changes the stock estimate by a factor of one thousand. A unit audit asks the reader to calculate the total independently and test whether the answer is plausible for the documented size and stage of the cohort.
Worked audit: transfers and incomplete recovery
An invented nursery has 180,000 estimated juveniles before splitting into three destinations. Staff transfer estimated totals of 70,000, 65,000 and 40,000. The allocated number is 175,000, leaving 5,000 not accounted for by those transfer figures. That remainder is a reconciliation difference, not automatically five thousand dead animals. The ledger should record handling losses, residual animals in equipment, counting variation and any intentional grading or retention. Until the discrepancy is investigated, label it unresolved rather than turning a bookkeeping gap into a precise mortality rate.
If the nursery later recovers 61,000, 57,000 and 35,000 usable juveniles from those destination groups, the combined outcome is 153,000. Relative to the 175,000 estimated transferred, the apparent usable recovery is about 87.43 percent. Relative to the 180,000 pre-transfer estimate, it is 85 percent. Both ratios have clear meanings, but neither proves that the difference consists solely of biological deaths. Some animals may be alive but below grade, and sample counts may be imperfect. Every numerator requires a consistent stage and quality definition.
If two of the three destination groups were deliberately held back for additional nursery time, a report saying all 153,000 juveniles were immediately shipped would be wrong even though the count totals are accurate. Shipment is a separate event documented by a manifest or dispatch ledger. Likewise, shipped count is not guaranteed planted count after transport and receiving inspection. An auditable production record distinguishes available seed, accepted seed, packed seed, dispatched seed, received seed and seed actually placed at an approved farm site.
Diagnostic errors and their repairs
The first common error is calling every counted egg fertilised. Repair it by retaining the fertilisation assessment criterion and denominator. The second is confusing concentration with total tank inventory. Repair it by writing the actual volume and dilution factor. The third is saying larvae died whenever water-column counts fall. Repair it by checking settlement, transfers, sampling and water changes before interpreting mortality. These corrections change the meaning of the report without requiring jargon or complicated mathematics.
The fourth error is treating attachment and normal metamorphosis as one event. Repair it by documenting the exact assessment stage and time. The fifth is equating juveniles recovered by a sieve with every live juvenile present in sand. Repair it by stating recovery method and efficiency limitations. The sixth is replacing shell length with whole-animal mass in a seed grade table. Repair it by preserving the physical dimension and sample distribution. The seventh is predicting adult harvest tonnes directly from early seed yield without a validated grow-out model.
The eighth error is hiding a batch merge. If two broodstock cohorts are combined, the later denominator becomes a pooled population, so a reported apparent survival percentage may not describe either original cohort alone. The ninth is using a trial’s temperature or timing as a universal recipe. Repair both by identifying the actual cohort definition and operating conditions. The tenth is turning an imprecise estimate into a confident exact count by adding decimals. Keep numerical precision consistent with sampling and calibration evidence.
Independent transfer problems
Problem one: A suspension contains 250 estimated eggs per millilitre and has a final measured volume of 3.2 litres. How many eggs are estimated? Answer: 3,200 millilitres times 250 equals 800,000 eggs, assuming representative mixing. Problem two: Of 240 inspected eggs, 192 meet the stated fertilisation rule. What is the observed percentage? Answer: 80 percent. Applying this proportion to all 800,000 eggs gives 640,000 estimated positive eggs, not a direct count of six hundred and forty thousand individually examined embryos.
Problem three: A tank has 90 larvae per litre in 2,400 litres of culture. Estimate inventory. Answer: 216,000 larvae. If a technician mistakenly reads the concentration as larvae per millilitre, the implied total becomes 216 million; that is a thousandfold unit mistake. Problem four: After settlement, only 43,200 juveniles are recovered. What is the apparent recovery relative to the 216,000 earlier larvae? Answer: 20 percent, provided the compared inventories refer to the same closed cohort and their estimates are compatible.
Problem five: A sand-nursery quadrat sample covering a total of 0.08 square metres recovers 640 juveniles. Estimate observed recovery density. Answer: 8,000 juveniles per square metre. A 0.75-square-metre tray extrapolates to 6,000 juveniles under a representative-sampling assumption. Problem six: The same tray’s sampling frame includes 0.15 square metres that were never suitable for burrowing. If eligible area is only 0.60 square metres, revised estimated count is 4,800. The change arises from a corrected area denominator, not observed loss of 1,200 animals.
Problem seven: A nursery recovers 50,000 live juveniles, of which 44,000 meet its specified shell-size grade. What is the accepted proportion? Answer: 88 percent of recovered animals. If 40,000 are dispatched, what is the dispatch fraction of accepted stock? Answer: about 90.91 percent. Neither percentage is total egg-to-seed yield unless an original egg denominator is also known. Problem eight: A weekly transfer makes the tank count fall by half. Does that prove half the cohort died? No. Reconcile source and destination volumes, times and counts before drawing a biological conclusion.
Frequently asked questions
Does a high fertilisation rate guarantee a high seed yield? No. Larval development, settlement, nursery survival, grading and handling remain separate processes. Does a drop in water-column larval count always indicate death? No; settlement or transfer can reduce the measured water-column inventory. Does a nursery extraction count every buried juvenile? Not necessarily; sampling and recovery efficiency matter. Can a published average settlement day replace direct observation in another tank? No. The timing depends on species, temperature, method and developmental conditions in the study.
Can a hatchery declare every live juvenile acceptable for outplanting? Only if the relevant seed acceptance rules are actually met; live count and quality grade are different. Can feed added be treated as feed ingested? No. Do all error bars describe biological variation? No; some refer to sampling or measurement estimates. Does a seed count authorize farming at any shore site? No; location, permits, health controls and the farm’s operating requirements are independent. These distinctions matter because technical vocabulary often carries conditions hidden by a short everyday synonym.
A readable summary and the right evidence
For a useful public article, begin by identifying the actual stage measured: collected eggs, criterion-positive fertilised eggs, living larvae, competent larvae, settled juveniles, recovered nursery juveniles or graded seed. Then provide one correctly labeled ratio, the date of measurement, the denominator and the principal uncertainty. Only afterward explain possible causes of changes. This keeps the answer easy to follow without pretending the process is one uninterrupted survival percentage. It also makes the calculations independently checkable by a reader.
General English words such as count, estimate, compare, measure, increase and remain are powerful because they explain technical evidence precisely. Learners can practise them using the verified eduKateSG adult core vocabulary system. That language resource does not supply marine-hatchery data; the scientific and methodological claims belong to the first-party references below. Distinguishing language practice from specialist evidence is part of an honest translation.
A complete geoduck hatchery account has a reversible chain: every final shipment should trace to an accepted seed count, every accepted count to a nursery sample or census, every nursery cohort to settlement and larval records, and every initial cohort to the spawning record. Each ratio should retain its own numerator and denominator. If the chain breaks, the report should describe the missing information rather than fill it with a plausible guess. Good teaching makes the uncertainty visible and gives the reader a method to repair mistakes.
Primary and specialist sources
NOAA Fisheries — Geoduck Aquaculture documents hatchery larval rearing, shallow-sand nursery and seed outplanting in a species overview.
NOAA Technical Report NMFS 85 provides historical geoduck hatchery method examples and developmental context; its operating details are not universal thresholds.
NOAA repository — Dynamic Response in the Larval Geoduck Proteome to Elevated pCO2 describes a controlled experimental study of larval development under specified carbonate conditions.
Washington Department of Fish and Wildlife — Commercial geoduck management differentiates wild stock and cultivated geoduck operations.