A spiny lobster water-quality sheet cannot be read safely by looking at one number and deciding that the animals were healthy or unhealthy. Dissolved oxygen, temperature, salinity, pH, total ammonia nitrogen and water movement interact. A reliable translation preserves where and when each reading was taken, its unit and measuring method, and the evidence connecting it to any animal response.
Start with the specific system: a coastal sea cage, a hatchery tank, a land-based nursery or a recirculating facility. Record species and life stage, water depth, sensor location, observation interval and recent changes such as feeding, rain, cleaning or equipment interruption. A value measured near a surface inlet is not necessarily the value experienced at a bottom shelter.
Quick summary: distinguish oxygen in milligrams per litre from oxygen percent saturation; separate TAN-N from unionised ammonia; retain pH, salinity and temperature together; document flow and exchange units; test sensor quality; compare matched time windows; and translate observations into conditional explanations rather than universal safe limits.
The phrase water quality sounds simple but refers to several measurements with different scales. A temperature is a physical state variable, dissolved oxygen is a gas concentration or saturation measure, salinity expresses dissolved salts, pH is logarithmic, and ammonia is a family of chemical species. A translation that converts them all into a single green, yellow or red category without a defensible local rule discards their meaning.
This guide is meant for readers of aquaculture logs, field notes, laboratory reports, operating summaries and translated technical articles. It teaches calculations and interpretation. It does not prescribe one numerical oxygen threshold or one emergency chemical treatment for all Panulirus species: such decisions depend on species, stage, system, local animal-health expertise and applicable requirements.
Every numeric worked example below is explicitly hypothetical unless a source is named. Hypothetical values illustrate arithmetic, not an assertion that a particular farm was sampled or that a concentration is safe. The distinguishing skill is to preserve the denominator, environmental context, observation method and uncertainty.
What the rows actually represent
Each row must specify a time, sample or sensor, depth, position and unit. An hourly surface probe is a time series at one location, not a uniform map of an entire tank. A laboratory sample may be a discrete bottle that changed during transport. A reported tank mean may hide rapid short-lived extremes. Before explaining a chart, determine which physical water parcel and observation interval its values represent. Only then can the figures support a comparison with lobster behaviour.
Dissolved oxygen concentration
Dissolved oxygen, abbreviated DO, commonly appears as mg/L, numerically equivalent to parts per million for dilute water under usual approximations but best retained in the recorded unit. If DO declines from 6.0 to 4.5 mg/L, the absolute decline is 1.5 mg/L and the relative decline from the first value is 25 percent. That arithmetic says nothing by itself about a species-specific safe endpoint. Oxygen demand and replenishment depend on animals, microorganisms, feed, temperature, water movement and gas exchange.
Oxygen percent saturation
A dissolved oxygen sensor may instead display percentage of air saturation. That is a ratio to the equilibrium concentration under the instrument’s assumptions about temperature, salinity and pressure. It is not a biological survival percentage, and 80 percent saturation is not 80 mg/L. Converting between percent saturation and mg/L requires the correct solubility or instrument model, including temperature and salinity. Do not combine a concentration record with a saturation series as if they used a common numerical scale.
Time and depth matter
Daily average oxygen may look acceptable even when brief low-oxygen intervals recur near dawn, within dense shelter structures or after an equipment failure. Surface mixing may conceal poorer bottom conditions. Translate a mean as an average over a defined set of observations, not a guaranteed minimum. A minimum based on hourly logging can miss a five-minute event. Adding probe depth, logging interval and error bars to an otherwise simple sentence may change its entire practical meaning.
Temperature as both signal and context
Water temperature influences oxygen solubility and organism metabolism, while it also affects ammonia chemistry. A warmer reading and a lower oxygen reading may coincide without proving one caused the other. If a temperature sensor sits in direct sun or a shallow unrepresentative point, its record may not reflect conditions at the lobsters. Preserve the temperature scale in degrees Celsius where used and the sensor location. A difference of two degrees Celsius is not the same as a two percent change.
Salinity, conductivity and rain
Salinity is often reported as parts per thousand, practical salinity or conductivity-derived values; these are not always interchangeable word for word. Heavy rain can create a fresher surface lens while deeper water remains saline. If a probe reports 34 and another 28, establish their units and depths before calling the event a tankwide freshening. Changes in salinity can also affect oxygen calibration and ammonia speciation calculations. Translate the method, not just the number.
pH is logarithmic
pH is related to hydrogen-ion activity, not a linear concentration percentage. A shift from pH 8.0 to pH 7.0 represents roughly a tenfold increase in hydrogen-ion activity on the same defined pH scale, not a one-unit increase in acidity that can be interpreted like one degree of temperature. In seawater, pH reporting can involve different scales and calibration conventions. Preserve the instrument protocol when comparing laboratories, and avoid presenting pH alone as proof of harmless or harmful water.
Total ammonia and unionised ammonia
Total ammonia nitrogen, often abbreviated TAN, combines ammonia and ammonium when expressed on a nitrogen basis. Unionised NH3 is a particular component whose fraction varies with temperature, pH and salinity. FAO references include seawater-specific tabulations that demonstrate the dependence. Thus 0.50 mg/L TAN-N cannot simply be announced as 0.50 mg/L NH3. Even a calculated NH3 fraction must identify the temperature, pH, salinity, method and whether masses are expressed as nitrogen or ammonia molecules.
A checked ammonia fraction example
For a purely illustrative seawater table lookup at 25 degrees Celsius and pH 8.2 in the FAO 32–40 per mille salinity band, one historical table gives about 6.61 percent unionised NH3. If TAN-N equals 0.50 mg N/L, multiplication gives 0.50 times 0.0661, or about 0.0331 mg NH3-N/L. This is a conditional arithmetic illustration, not a site-standard toxicity judgment. Converting NH3-N to NH3 molecular mass uses 17 divided by 14, giving about 0.0401 mg NH3/L.
Flow is not complete water replacement
A pump may move 300 L/min through a nominal 60 cubic metre tank. This equals 0.3 cubic metre per minute or 18 cubic metres per hour, so the nominal flow divided by volume is 0.3 per hour. That is a hydraulic turnover-rate indicator, not proof that 30 percent of every part of the tank is perfectly replaced each hour. Short-circuiting, stagnant pockets, recirculation, screen losses and mixing determine actual exchange. Describe the hydraulics without inventing uniformity.
Ammonia production and feeding
Uneaten feed, excretion and decomposition can increase nitrogen loading, but a higher TAN reading after feeding does not isolate the exact cause without input and output measurements. Protein composition, feeding mass, microbial processing and water exchange can all alter concentrations. A report that labels a single elevated measurement as feed-caused should be rewritten as an association unless independent evidence excludes other mechanisms. Translate the chronology: feeding time, sample time, flow condition, pH and temperature.
Sensor quality and calibration
A water-quality record is a measurement produced by a sensor or assay, not the water itself. Fouling, membrane condition, drift, air bubbles, reference solutions, pressure assumptions and warm-up periods can affect values. Record calibration dates, instrument IDs and quality flags. If an impossible-looking jump follows maintenance, first check whether the scale or calibration changed. A corrected reading should be distinguishable from the original value; never quietly replace the archived source.
Alkalinity versus pH
Alkalinity describes acid-neutralising capacity and is not a synonym for a high pH. Two waters can have the same measured pH and different buffering capacities. If a laboratory reports alkalinity as mg/L CaCO3 equivalents, do not translate the figure as a calcium carbonate dose or automatically as salinity. Keep the chemical basis and methods. For operational conclusions, interpret alkalinity alongside pH and other measurements, rather than making a single-variable prediction of biological condition.
Sampling versus continuous monitoring
An operator may use a handheld meter at 09:00 while a fixed logger records every five minutes. These instruments sample different time distributions, and their means are not directly equivalent. A laboratory bottle analysed hours later may represent collection conditions imperfectly if handling changes dissolved gases. Label observations as spot readings, continuous-sensor records or laboratory determinations. Check clock synchronisation, time zone and data gaps before correlating the series with feeding or mortality.
Separating environmental and biological endpoints
Water properties are exposures; feeding, activity, moulting, lesions and survival are biological responses. They should have distinct records linked by time and unit of observation. A low oxygen reading followed by inactivity is a warning to investigate, not definitive proof that oxygen was the only cause. A high mortality count with one normal sensor reading likewise does not exclude brief earlier excursions or other agents. Careful translation preserves uncertainty where the evidence does.
Replicates and independence
Three probes in one tank give measurement replication, but do not automatically create three independent treatment tanks. Similarly, a hundred animals in one shared system do not establish a hundred independent environmental experiments. A comparison of management methods needs the correct experimental unit and true replication. Translate ‘n = 3 sensors’ differently from ‘n = 3 independent tanks’. Failure to do so can make an apparent effect seem far more certain than the design justifies.
Responding to suspected harmful conditions
An abnormal value should trigger verification, an animal-welfare assessment and the site’s authorized response protocol. A technical translation must not convert a measurement into a universal chemical instruction. Check the sensor, inspect animal condition, review water supply, identify recent changes and involve appropriately qualified operators or veterinarians as needed. Decisions and times belong in a separate action log. The article explains how to read such records, not how to replace trained aquaculture judgment.
A readable final water-quality sentence
A defensible statement has this form: at a specified time and depth, a named instrument recorded a given DO concentration and related temperature, salinity or pH; the quality-control status is known; a change relative to a stated baseline was observed; and its possible relationship to animal behaviour remains either supported, tentative or unresolved. This sounds longer than ‘water was bad’, but it lets an independent reader assess the source without guessing.
Practical casebook: recognise the error before translating the claim
Case 1: The two oxygen scales
A tank chart changes from a DO concentration axis marked 5.2 mg/L to an axis marked 82 percent saturation halfway through the week. Someone calculates a huge apparent oxygen improvement by treating 82 as a concentration. These values describe different quantities and cannot be subtracted or averaged directly.
Recover the instrument configuration and its temperature-salinity compensation. Mark the unit change explicitly in the plot. If comparable concentrations cannot be recovered, show two labelled series instead of a fabricated continuous trend.
Case 2: The dawn minimum
Daily oxygen observations are 7.0 mg/L at noon and 4.0 mg/L before dawn, but a report quotes the daytime figure as the day’s oxygen condition. At minimum, that erases a 3.0 mg/L swing.
Translate the time and extrema. The mean of these two selected readings is 5.5 mg/L, but that is not the full-day average unless the sampling design supports it. A more frequent series could reveal additional troughs.
Case 3: A bottom shelter sensor
A surface instrument records 6.4 mg/L while another sensor near a sheltered bottom area records 4.8 mg/L at the same time. The difference is 1.6 mg/L between locations, not necessarily instrument drift.
Check simultaneous calibration and placement before claiming a spatial gradient. If credible, report both positions. Surface water alone should not stand for the exposure at a hidden animal shelter.
Case 4: A sensor near an air bubble
DO suddenly increases after an aerator starts, but the probe tip is positioned in a concentrated bubble plume. The signal may reflect direct local bubbles or altered sensor conditions, not an even rise in water-column oxygen.
Move assessment to validated locations and inspect sensor response. Translate the reading as a local observation and do not infer that every lobster experienced the apparent change.
Case 5: Temperature mismatch
An ammonia calculation uses pH from the tank at 28 degrees Celsius but a laboratory worksheet assumes 20 degrees. The unionised fraction depends on temperature, so a borrowed default can change the derived estimate.
Retain contemporaneous temperature and salinity and name the equilibrium model. If they are missing, report TAN and the unknown speciation rather than a falsely precise NH3 value.
Case 6: TAN measured as nitrogen
A laboratory reports total ammonia nitrogen at 0.28 mg N/L. A translated label changes it to 0.28 mg NH3/L, omitting the chemical basis. Even if all nitrogen were in NH3, molecular mass would require multiplying by 17/14.
Keep mg N/L attached to TAN. Calculate any NH3 mass concentration only after estimating the unionised fraction and changing the mass basis with an explicit equation.
Case 7: The 6.61 percent illustration
At the specified illustrative FAO seawater conditions of 25 degrees Celsius, pH 8.2 and salinity 32–40 per mille, a lookup fraction of 0.0661 converts 0.50 mg TAN-N/L to about 0.0331 mg NH3-N/L. The difference between fraction and percentage is critical.
Multiply by 0.0661, not by 6.61. State that the fraction is condition-specific and taken from a historical table, not a universal spiny lobster safety threshold.
Case 8: A pH scale shift
A sensor is recalibrated and the pH reading shifts from 8.1 to 8.3 without any documented tank intervention. An analyst attributes every part of the difference to a chemical change in seawater.
Check the calibration procedure, electrode condition, stated pH scale and reference solutions. A valid environmental conclusion cannot be stronger than measurement quality allows.
Case 9: Rain and stratification
A storm produces a fresh surface layer, with readings of 24 near the surface and 33 near the cage bottom on the stated salinity scale. A report says the entire cage water suddenly fell to 24.
Use depth-specific language. A surface lens can move and mix rapidly. A single surface sample should never be silently generalized to the entire occupied volume.
Case 10: Water flow converted to volume
A pump’s verified flow is 250 L/min for two hours. Total nominal pumped volume is 250 times 120, or 30,000 L, equivalent to 30 cubic metres. That volume is a flow integral, not a measured uniform water replacement.
Check whether the system recirculates and whether bypass flow is excluded. Translate nominal throughput differently from actual clean-water exchange or residence-time distribution.
Case 11: Nominal turnover rate
A 50 cubic metre tank receives 10 cubic metres per hour from a verified external supply. The nominal turnover rate is 0.2 per hour and nominal volume divided by supply is five hours.
Do not say that every molecule is replaced in five hours. Real mixing produces residence-time variation. This result is a simple design indicator that needs hydraulic context.
Case 12: Probe downtime hides an excursion
A logger skips the two hours during which a pump stopped. The remaining daily average DO is 5.9 mg/L. A report calls the entire day stable because the visible chart contains no low reading.
Explicitly label the missing interval. Do not interpolate a reassuring constant value without an appropriate method and uncertainty. Correlate power and equipment logs rather than assuming missing means normal.
Case 13: Unmatched measurements
DO is measured at 08:00 in the upper tank while TAN is sampled at 16:00 from a bottom discharge. An analyst describes this as a simultaneous chemical condition experienced by one lobster.
Use separate timestamps and locations. The data can describe aspects of the same system but do not form a matched observation without sufficient evidence of stability or mixing.
Case 14: An overly precise decimal
A test strip resolves pH only to roughly one decimal place, but a spreadsheet displays 8.274 after arbitrary formatting. The extra digits do not add information.
Keep precision consistent with the sensor and method. Apparent decimals from formatting should not become new scientific evidence or trigger exaggerated discussion of tiny differences.
Case 15: Ammonia and pH interact
Two samples have equal TAN-N but different pH values. Because the NH3 fraction changes with pH and temperature, identical TAN numbers need not imply identical unionised-ammonia exposure.
Translate TAN and the conditional chemistry separately. Do not rank biological risk from TAN alone or ignore that the temperature and salinity of the two samples may also differ.
Case 16: Averaging on the wrong scale
Three DO values are 3.0, 5.0 and 7.0 mg/L from equal-duration validated observations. Their arithmetic mean is 5.0 mg/L. Converting each to a saturation percent with differing temperature then averaging creates a different statistic.
Name the scale and weighting. An average concentration and an average saturation percentage answer different questions and should not be presented as interchangeable.
Case 17: Median and minimum
Seven oxygen observations are 6.0, 6.1, 6.0, 6.2, 6.1, 5.9 and 2.4 mg/L. The median is 6.0, but the minimum is 2.4. A median-only summary hides a possible short event.
Report the minimum and duration when known, plus sensor validity. If the 2.4 reading is flagged suspect, investigate it before either deleting it or treating it as definitive exposure.
Case 18: Salinity units
A handheld device shows 31 PSU while a laboratory writes a figure in grams of salt per kilogram. These scales are close in some oceanographic contexts but are not definitions to interchange automatically.
Preserve the original units and measurement principle. Avoid claiming a precise mass concentration from a practical salinity reading without appropriate conversion.
Case 19: A flow-rate denominator
A farm writes ‘exchange 20%’ but does not say whether this means water volume exchanged per hour, per day or per cycle. Operators reading different languages interpret the same percentage differently.
Restore both denominator and time. ‘Nominal flow equal to twenty percent of tank volume per hour’ is specific; ‘twenty percent renewal’ is not.
Case 20: Feed and nitrogen peak
A TAN measurement increases after a heavy feeding period, but the system also experiences lower water flow and an interrupted biofilter. The correlation with feeding is not isolated causation.
Reconstruct the sequence and use environmental controls or additional measurements. Retain words such as ‘coincided with’ until the analysis can separate feed loading from reduced processing and exchange.
Case 21: Percent saturation without pressure
A DO logger exports saturation percentages but not the pressure assumptions, then another site compares the values with a coastal station at a different elevation.
Recover temperature, salinity, pressure and sensor compensation before converting or comparing absolute concentrations. A shared percentage label does not guarantee equal mg/L.
Case 22: Oxygen load and animals
A water-quality commentary attributes an oxygen decline solely to lobster respiration. The same tank contains microbial films, decaying feed and other oxygen-consuming processes.
Use a mass balance or separate evidence if making a quantitative attribution. Otherwise report the observed decline and possible contributors without inventing a respiratory rate for the animals.
Case 23: Sample bottle holding time
A sample collected for a laboratory oxygen test spends hours in a partially filled bottle, with storage conditions unreported. Gas exchange may change the concentration between collection and analysis.
Document collection and preservation method; distinguish field DO from later laboratory estimates. Do not publish a value as if it were a faithful instantaneous field reading when the method is unknown.
Case 24: Calibration versus cleaning
A probe reads higher after its membrane is cleaned, and the dashboard labels the change ‘recovery from low oxygen’. Cleaning may alter sensor response independently of real water.
Annotate maintenance events and compare with independent checks. A data correction and an environmental change are different events and require different wording.
Case 25: pH and alkalinity
Two waters have pH 8.0, but one has greater measured total alkalinity. A summary translates both as having identical buffering properties.
Keep alkalinity in its reported units and differentiate it from pH. Acid-neutralising capacity cannot be inferred from a matching pH value alone.
Case 26: Spatial pooling
A mean of four sensor locations is reported without noting that two are adjacent to an inlet and only one is near a distant shelter. The aggregate may misrepresent where animals spend time.
Show the station map and preferably individual traces. A mean across stations describes the selected network, not a physically uniform condition.
Case 27: Time zones in equipment logs
A cloud logger stores UTC but staff record feeding and mortality in local clock time. An apparent two-hour lead or lag is created by conversion error.
Convert all events to one stated time zone while retaining original timestamps. Align clock drift and daylight-saving rules before constructing a narrative about which event preceded another.
Case 28: The false normal range
A translation says all spiny lobster species have exactly one safe DO, pH, temperature and ammonia range, citing no species-stage study. Such an assertion confuses local operating criteria with universal biological law.
Replace the phrase with the actual site’s validated monitoring plan, including source and limits, or omit the threshold. A guide for interpretation should not invent biological standards.
Case 29: A dramatic mortality headline
One tank experiences deaths after a storm while sensors show changing salinity and oxygen. A headline blames ammonia, although no corresponding TAN sample was collected.
State that mortality followed a multivariable event and that ammonia was not measured for the relevant window. It is better to disclose an unknown than misdirect animal-health decisions with an unsupported cause.
Case 30: Two sensors are not two tanks
One treatment enclosure has two DO probes; another has one. Treating three probes as three independent tank replicates inflates the claimed sample size.
Identify the experimental unit. Measurement replication improves coverage and confidence in readings but does not replace independent treatment replication.
Case 31: A corrected reading
An instrument recorded DO as 3.2 mg/L during a calibration fault. A laboratory check supports a revised estimate, but the original record is deleted.
Keep both original and corrected values with reason, timestamp and reviewer. Translation needs the data history to distinguish a corrected sensor artifact from a real harmful event.
Case 32: Unlogged intervention
The water-quality chart improves after an aerator is switched on, but the intervention is absent from the shared dataset. A reader attributes the recovery to natural tidal flushing.
Add the intervention event, its time and monitoring response. Do not infer the aerator alone caused improvement without appropriate comparison, but do not hide an important alternative mechanism.
Case 33: Sparse samples and false stability
A handheld test taken every third day shows no unusual DO, yet feeding and animal observations show short episodes of distress. A sparse sampling schedule can miss rapid fluctuations.
Translate ‘stable at sampled times’ rather than ‘continuously stable’. Propose more informative monitoring through the site’s approved protocol without fabricating the unseen values.
Case 34: Mixing two nitrogen conventions
One laboratory reports 0.4 mg TAN-N/L, another 0.4 mg NH3/L, and a spreadsheet treats them as identical concentrations of toxic unionised ammonia.
These are different analytical labels and chemical bases. Recover the speciation and conversion conventions before merging datasets or ranking locations by risk.
Case 35: A mean without measurement counts
A report averages three daily DO means and assigns equal weight, although those days contain 12, 24 and 2 valid readings. The resulting mean is a mean of means, not necessarily a mean of all observations.
Describe which weighting was used and why. If an exposure-weighted mean is intended, observations and time intervals must support it; otherwise preserve the narrower statistic.
Transfer practice with answers
Practice 1: DO percentage change
A verified concentration falls from 6.8 to 5.1 mg/L. The decrease is 1.7 mg/L and 1.7 divided by 6.8 equals 25 percent. This does not imply 25 percentage points of saturation.
Translation test: rewrite the answer for a reader who cannot inspect the raw file. Name the unit, the denominator or time interval, and whether the result is a direct observation, a conversion or a conditional estimate. Do not add an animal-health conclusion that the measurements cannot support.
Practice 2: Volume and units
At 180 L/min, a pump moves 10,800 L in an hour, or 10.8 cubic metres. State whether that flow comes from new water or merely circulates within the same reservoir.
Translation test: rewrite the answer for a reader who cannot inspect the raw file. Name the unit, the denominator or time interval, and whether the result is a direct observation, a conversion or a conditional estimate. Do not add an animal-health conclusion that the measurements cannot support.
Practice 3: Turnover indicator
A 72 cubic metre tank has 24 cubic metres per hour of measured new supply. Nominal turnover is one third per hour, and nominal volume divided by supply is three hours. Neither value proves complete replacement.
Translation test: rewrite the answer for a reader who cannot inspect the raw file. Name the unit, the denominator or time interval, and whether the result is a direct observation, a conversion or a conditional estimate. Do not add an animal-health conclusion that the measurements cannot support.
Practice 4: TAN fraction
If a conditional fraction is 4 percent and TAN-N is 0.75 mg N/L, derived NH3-N is 0.03 mg N/L. The fraction must be justified for the water’s temperature, pH and salinity.
Translation test: rewrite the answer for a reader who cannot inspect the raw file. Name the unit, the denominator or time interval, and whether the result is a direct observation, a conversion or a conditional estimate. Do not add an animal-health conclusion that the measurements cannot support.
Practice 5: Mass-basis conversion
An illustrative NH3-N concentration of 0.028 mg N/L corresponds to approximately 0.034 mg NH3/L using 17/14. A reader should be able to trace both the molecular weights and basis.
Translation test: rewrite the answer for a reader who cannot inspect the raw file. Name the unit, the denominator or time interval, and whether the result is a direct observation, a conversion or a conditional estimate. Do not add an animal-health conclusion that the measurements cannot support.
Practice 6: Two spot samples
DO measures 6.2 and 5.4 mg/L in two different places. Their difference is 0.8 mg/L but does not by itself define a tankwide time trend.
Translation test: rewrite the answer for a reader who cannot inspect the raw file. Name the unit, the denominator or time interval, and whether the result is a direct observation, a conversion or a conditional estimate. Do not add an animal-health conclusion that the measurements cannot support.
Practice 7: pH logarithm
The difference between pH 8 and pH 7 is one pH unit, corresponding approximately to a tenfold change in hydrogen-ion activity on the same scale. It is not a ten percent change.
Translation test: rewrite the answer for a reader who cannot inspect the raw file. Name the unit, the denominator or time interval, and whether the result is a direct observation, a conversion or a conditional estimate. Do not add an animal-health conclusion that the measurements cannot support.
Practice 8: Unweighted mean
For 4, 6 and 8 mg/L at three equally weighted valid observations, arithmetic mean is 6 mg/L. Report the sampling times; this is not automatically the 24-hour mean.
Translation test: rewrite the answer for a reader who cannot inspect the raw file. Name the unit, the denominator or time interval, and whether the result is a direct observation, a conversion or a conditional estimate. Do not add an animal-health conclusion that the measurements cannot support.
Practice 9: Missing interval
A sensor records for 20 of 24 hours. Coverage is about 83.3 percent. The other four hours are unknown rather than normal readings.
Translation test: rewrite the answer for a reader who cannot inspect the raw file. Name the unit, the denominator or time interval, and whether the result is a direct observation, a conversion or a conditional estimate. Do not add an animal-health conclusion that the measurements cannot support.
Practice 10: Unit preservation
A flow of 2.4 cubic metres per hour equals 2400 L/hour or 40 L/min. A decimal mistake can create a sixtyfold or thousandfold wrong conclusion.
Translation test: rewrite the answer for a reader who cannot inspect the raw file. Name the unit, the denominator or time interval, and whether the result is a direct observation, a conversion or a conditional estimate. Do not add an animal-health conclusion that the measurements cannot support.
Practice 11: Range, not certainty
Two verified DO observations are 3.8 and 6.0 mg/L. The observed range is 2.2 mg/L, but the actual minimum between readings could be lower.
Translation test: rewrite the answer for a reader who cannot inspect the raw file. Name the unit, the denominator or time interval, and whether the result is a direct observation, a conversion or a conditional estimate. Do not add an animal-health conclusion that the measurements cannot support.
Practice 12: Sensor replication
Three sensors in one tank provide three observation locations, not three independent tanks. Explain this distinction before calculating standard errors for a treatment effect.
Translation test: rewrite the answer for a reader who cannot inspect the raw file. Name the unit, the denominator or time interval, and whether the result is a direct observation, a conversion or a conditional estimate. Do not add an animal-health conclusion that the measurements cannot support.
Practice 13: Salinity depth profile
A surface observation is 29 and a lower observation is 34 on a common validated salinity scale. Their difference is five units but does not specify how much tank volume belongs to each layer.
Translation test: rewrite the answer for a reader who cannot inspect the raw file. Name the unit, the denominator or time interval, and whether the result is a direct observation, a conversion or a conditional estimate. Do not add an animal-health conclusion that the measurements cannot support.
Practice 14: Action timing
A logger flags low oxygen at 02:00 and an operator action is recorded at 02:12. The response lag is twelve minutes, not proof that the intervention succeeded; outcome readings are needed.
Translation test: rewrite the answer for a reader who cannot inspect the raw file. Name the unit, the denominator or time interval, and whether the result is a direct observation, a conversion or a conditional estimate. Do not add an animal-health conclusion that the measurements cannot support.
Frequently asked questions
What should I do if oxygen falls sharply?
Verify the instrument and location, follow the farm’s approved welfare and response procedure, and seek qualified assistance when animals may be at risk. This guide does not prescribe a universal treatment or threshold.
Is TAN the same as toxic NH3?
No. TAN is a sum on a declared basis; unionised NH3 is one chemical species. The proportion varies with water chemistry. Keep pH, temperature, salinity and the unit basis with any derived estimate.
Can a daily average demonstrate that no exposure event occurred?
No. Average values obscure short minima and missing data. Report observation frequency, coverage, credible extrema and the conditions at the animals’ locations.
Can I compare oxygen percent saturation from different instruments?
Only after checking compensation, pressure assumptions, calibration and underlying salinity and temperature. Percent saturation and mg/L need not respond identically when conditions change.
What makes a translated water-quality claim credible?
A reader can trace the exact instrument or sample, location, clock time, scale, calculation, quality controls and limits of inference. The translation carries both the number and its meaning.
Sources and relevant eduKateSG reading
FAO chemical features of water: ammonia fractions across salinities and temperatures: https://www.fao.org/4/AC175E/AC175E18.htm
FAO aquaculture guidance on DO and ammonia as contextual principles, not lobster-specific limits: https://www.fao.org/4/AC210E/AC210E09.htm
FAO AGRIS study on Panulirus weaning and water-treatment performance: https://agris.fao.org/search/fr/records/6970e917bd789e57df2e7b87
Western Australia government library: crustacean handling and stress review: https://library.dpird.wa.gov.au/fr_fja/137/
eduKateSG Spiny Lobster Feed Costs and Wet–Dry FCR: https://edukatesg.com/2026/11/01/how-to-translate-easily-spiny-lobster-feed-cost-wet-dry-fcr-break-even-margin/
eduKateSG Spiny Lobster Moulting, Cast Shells and Growth: https://edukatesg.com/2026/11/01/how-to-translate-easily-spiny-lobster-moulting-cast-shells-soft-shell-growth-intervals/
eduKateSG Spiny Lobster Cage Grow-Out and Biomass: https://edukatesg.com/2026/11/01/how-to-translate-easily-spiny-lobster-sea-cage-growout-density-biomass-harvest-yield/
eduKateSG How to Translate Easily series index: https://edukatesg.com/how-to-translate-easily-series-index/