Salinity is one of the easiest numbers to misread in a mud crab record. A nursery sheet may say 25 ppt, a conductivity logger may output a practical salinity value, and a tank operator may describe adding freshwater to reduce salinity. A translation that merely changes the language of the labels can leave the calculation wrong. This guide teaches readers to reconstruct the water and salt balance, identify what was actually measured, and avoid turning a planning calculation into a promise of animal safety.
Mud crabs are capable of living across a range of salinities, but the suitable range is not one universal number for every Scylla species and life stage. Hatchery zoeae, megalopae, nursery crablets and grow-out animals can face different husbandry requirements. Sudden change, dissolved oxygen, temperature, moult stage, handling and stock history can alter the risk even when the final displayed salinity appears acceptable. The example numbers below are hypothetical arithmetic exercises, not site-specific acclimation instructions.
When reading an acclimation sheet, distinguish initial water condition, target, water added or exchanged, planned mixing model, measured result, elapsed time and biological response. These belong in separate columns. The aim is a clear, reproducible report that says what changed and what remains uncertain rather than a confident summary that outruns its data.
1. What salinity describes and what the number does not describe
Salinity expresses the dissolved-salt content of water according to a stated measurement scale or method. Many aquaculture field records casually use parts per thousand, ppt, for a mass-style approximation such as grams of salts per kilogram of solution. Practical Salinity, derived from a standardized conductivity relationship, is conventionally dimensionless. Field readers sometimes print PSU as a convenient label, although practical salinity is a unitless quantity. The scales are not technically identical; preserve the method rather than silently converting every label as though it were an exact mass concentration.
A number written as 25 might refer to an optical refractometer reading, a temperature-compensated conductivity-based display, a hydrometer interpretation or a manually selected target. Without instrument and method information, it is premature to report that the tank contained exactly 25 grams of salt per kilogram of water. On many farms, approximate numerical comparisons suffice for operational decisions, but scientific writing should avoid giving approximate field equivalences the status of exact analytical identities.
The salinity value is not a direct observation of dissolved oxygen, pH, ammonia, temperature, disease status or crab condition. A tank could have its intended salinity and still be unsuitable for other reasons. The Southeast Asian Fisheries Development Center’s aquaculture guidance treats salinity as one component of stage-specific water quality. A useful explanatory report therefore gives salinity alongside other environmental observations when available; it must not substitute a favorable salinity reading for a general finding of water safety.
Salinity and salt mass are related but not interchangeable. A tank reading can change because water is added, removed, evaporated, spilled or exchanged. Salt may leave with the withdrawn water, but also with salt deposits, droplets or solids under some conditions. The simple calculations in this guide assume a well-mixed tank, no hidden solute sources and sufficiently similar solution densities for volume-based approximations. Where exact mass accounting matters, use measured mass and the correct solution property model instead.
2. Begin with a named tank and a real time sequence
Every numerical record needs a unit of operation. “Salinity dropped after exchange” is ambiguous if a farm has several connected tanks, a sump and a storage reservoir. Identify the measured tank, whether the pump recirculated water or changed it, and where the makeup water originated. A sample from an inlet valve does not necessarily describe the concentration in the holding compartment, especially before mixing is complete. Naming locations is part of the translation, not unnecessary detail.
Time matters because acclimation is a process rather than a before-and-after pair. A sheet with an entry at 09:00 and one at 15:00 does not tell us whether the salinity changed gradually, abruptly at 09:05, or in three pulses. An hourly reading does not capture every excursion. If the method includes a scheduled mixing period and multiple check points, retain those times. If the record lacks them, state only the observation times instead of inventing a gentle transition.
For a worked example, a holding tank contains a measured 500 liters at an indicated salinity of 28. An operator records that 100 liters were drained and 100 liters of makeup water at salinity 12 added. Under a completely mixed, constant-volume approximation, the calculated final value is 0.8 times 28 plus 0.2 times 12, or 24.8. This is a predicted mixing value, not a substitute for taking a new tank reading after circulation.
If the drain was opened while the inlet was running, mixing may occur continuously and the simple one-time replacement equation may not describe the operation. If the water was not mixed before the drain stage, the withdrawn fraction may differ from the tank average. A translation should preserve “drain then refill” and “simultaneous flow-through” as different procedures. Their time histories and, in some configurations, final values need not be the same.
3. Learn the basic constant-volume exchange equation
The standard classroom approximation for a single well-mixed replacement is final salinity equals the retained fraction multiplied by initial salinity plus the replaced fraction multiplied by source-water salinity. For a tank restored to its original volume, let f be the fraction exchanged, expressed as a decimal between zero and one. Then final approximately equals (1 minus f) times initial plus f times source. The equation states a mass-balance assumption; it is not a physiological acclimation model.
Suppose 25% of a 600-liter tank at 28 is replaced with 150 liters of water at 12. Multiply 0.75 by 28 and 0.25 by 12, giving 21 plus 3, or 24. A report may read: “The planned 25% exchange from a nominal 28 to 12 source predicts a final mixed value near 24 on the field salinity scale; the actual tank result requires confirmation.” Avoid saying the crabs experienced a uniform fall of four units without time-resolved measurements.
When the source water is nominally freshwater at near-zero salinity, a 20% replacement of a well-mixed tank initially at 30 predicts 24. The numerical change is six, but the percentage reduction in salinity is 20%. Those quantities should not be confused. A drop from 30 to 24 represents six scale points, whereas a 20% drop is relative to the starting salinity. Use “points” or state the actual values when a percent might be mistaken for water-exchange fraction.
To solve for the exchange fraction, rearrange the mixing equation: f equals (initial minus target) divided by (initial minus source), provided the denominator is not zero. For an initial 30, target 24 and source 0, f equals 6 divided by 30, or 0.20. For source 10, the same initial and target require 6 divided by 20, or 0.30. The target is unchanged, but the source-water salinity alters the volume required. Document it instead of treating all added water as fresh.
4. Understand why adding water is not replacing water
A tank that receives fresh water without first discharging liquid becomes fuller. Under an approximate conservative mass balance with negligible freshwater salinity, the original salt stays while the total volume increases. A different equation applies: final salinity approximately equals original salinity multiplied by original volume divided by final volume. Use this only when density and volume approximations are reasonable and no overflow, precipitation or leakage occurs.
For an illustrative open tank starting with 600 liters at a salinity reading of 30, adding 120 liters of freshwater gives 720 liters total. The predicted final value is 30 times 600 divided by 720, or 25. This is not the same operation as replacing 120 liters out of 600 with fresh, which predicts 24. The difference looks small in this example, but the procedure is distinct and the water-level implications can be decisive.
If a report says “20% fresh water was added,” ask whether the 20% is measured against the initial volume or the resulting total. Adding 20% of an original 600 liters means 120 liters, but freshwater making up 20% of final volume in a mixture requires 150 liters: 150 of 750. The resulting salinity predictions differ. In plain-language translation, “20% additional water relative to starting volume” is a much clearer description than an isolated percentage.
Space matters in the same equation. A tank with 600 liters and only 60 liters safe freeboard cannot accommodate a proposed 120-liter freshwater addition without discharge or overflow. If water overflows as the addition proceeds, the actual salt and water losses depend on mixing and outlet location. A correct arithmetic target does not make an infeasible procedure safe or physically realizable. Check the geometry, capacity, water level and equipment constraints.
5. Repeated exchanges produce a different history
Suppose a well-mixed tank starts at 30, and 15% is replaced with nominal freshwater. After one exchange, the predicted value is 30 times 0.85, or 25.5. If the process is repeated after complete mixing, the second predicted value is 25.5 times 0.85, or 21.675. Two 15% exchanges do not simply remove 30% of the original salt concentration; the second exchange removes a portion of water already diluted in the first.
Three identical exchange stages under those assumptions give 30 times 0.85 cubed, approximately 18.42. The predicted sequence is 30, 25.5, 21.68 and 18.42. It may help show how a planned series of operations could change a tank, but it cannot tell whether the animals tolerate the timing. Intervals, mixing, temperature and actual measurements remain necessary to interpret exposure. Avoid using a recurrence formula as biological permission to accelerate treatment.
If the source salinity is not zero, repeated exchanges approach the source concentration in the ideal model rather than reaching zero. For an initial 30 and source 10 with a constant 20% exchange, the first value is 26, the second 22.8, and the third 20.24. This follows the recurrence: next value equals 0.8 times current plus 0.2 times 10. A translator should retain which reading is before and after each exchange, or the sequence becomes impossible to verify.
Under flow-through conditions, a continuous mixing model may be more suitable. A familiar ideal stirred-tank equation gives exponential approach to inlet concentration when volume and flow are constant, but real tanks may short-circuit, stratify or have poorly mixed corners. Do not fit the ideal curve and declare it measured unless the data support the assumptions. A record should distinguish commanded pump flow, actual metered flow and the concentrations measured at specified points.
6. Solve problems involving seawater and freshwater blends
Suppose a reserve tank requires an approximate target of 24, using source waters at 32 and 0 under the same simple mixing approximation. Let x be the seawater fraction. Then 32x plus 0 times the other fraction equals 24, so x is 0.75. Three quarters of the final mixture should come from the nominal seawater source and one quarter from freshwater, before real-world corrections and direct testing. For 400 liters, that corresponds to 300 and 100 liters.
Now change the brackish source from zero to 8. For the same target 24 and seawater source 32, the seawater fraction is (24 minus 8) divided by (32 minus 8), or two thirds. A 450-liter mixture would therefore use about 300 liters of the higher-salinity source and 150 liters of the lower-salinity source. It is a blend calculation, not a recommendation for larval or broodstock water. Any real preparation needs reliable source-water measurements.
An impossible target can be spotted before pumping. If all available source waters measure 12 and 25, an ordinary mixture of only those waters cannot reach 30. A report asking for 30 from that pair either omits another source, uses a different process such as salt addition or evaporation, or contains an error. The algebra reveals the contradiction through a mixing fraction greater than one or less than zero. Interpret that as a missing assumption, not as a strange mixing instruction.
The word “seawater” also conceals variability. Salinity can change with precipitation, evaporation and freshwater inputs; a storage source may differ from the shore sample recorded yesterday. The mixture ratio should use contemporaneous measurements whenever practical. Do not substitute a generic textbook value of 35 for a recorded 31 merely because 35 is familiar to readers. The measured source conditions belong to that operation’s evidence.
7. Rainfall, evaporation and overflow in pond ledgers
Rainfall supplies water over a surface area, so the volume approximation is rainfall depth multiplied by wetted area. Twenty-five millimeters is 0.025 meters. On a pond section of 12 square meters, that depth is 0.300 cubic meters, or approximately 300 liters. If the pond initially contained 2,400 liters at 20 and no water escaped, a simple dilution estimate would be 20 times 2,400 divided by 2,700, or about 17.78.
That number is fragile if runoff, splash, seepage or overflow occurred. Rain can also be uneven, and the pond may not have mixed by the measurement time. A safe written statement calls it the predicted dilution under a no-loss, well-mixed assumption. To establish what the crabs actually experienced, compare instrument readings through the event and check water level, rainfall logs, drainage and other simultaneous conditions.
Evaporation is the reverse water-balance concept: pure water leaves more readily than dissolved salt, so concentration may rise if salts remain. If 1,000 liters at nominal salinity 20 lose 100 liters of water to evaporation, the simple prediction is 20 times 1,000 divided by 900, or about 22.22. But real systems may receive rainfall, seepage and replenishment. Describing every upward movement as evaporation without evidence is a causal leap.
Overflow complicates rainfall because water and some dissolved salt leave together. The final salinity depends on mixing during the rainfall and discharge. If a pond was already at its maximum level, 300 liters of rain might not produce 300 liters of net storage increase. Treat “rainfall received” and “net pond-volume change” as different quantities. Otherwise a translation may apply the correct formula to the wrong physical process.
8. Take instrument methods seriously
A conductivity sensor detects an electrical property that depends on dissolved ions and temperature. Its displayed practical salinity may be calculated from a standardized relation rather than measured as a mass directly. A refractometer uses an optical property; a hydrometer measures density. Their errors and calibrations differ. When reports compare values from different instruments, note the device types, calibration checks and temperature compensation rather than claiming they are exact replicates.
Suppose one log uses a displayed salinity of 26.0 and another lists conductivity in microsiemens per centimeter. They are not directly comparable as numerical values. Converting conductivity to a salinity estimate needs a suitable method, temperature information and assumptions about solution composition. A simplistic “divide by a thousand” shortcut can create a plausible-looking but unsupported figure. Preserve the original units and describe any validated conversion method used.
A measurement has a location and depth. A surface sample after rainfall may be fresher than deeper water before sufficient mixing. A sensor beside the inlet may observe a source-water plume rather than the general tank condition. A report should record sample placement, circulation state, sampling depth, time and instrument ID when the distinction affects interpretation. Two readings that disagree might both be accurate descriptions of different parts of the system.
Calibration and quality-control notes deserve space in readable prose. If a refractometer was rinsed, zeroed and checked against an appropriate standard, that supports comparability. If calibration status is unknown, do not manufacture certainty. For routine operational decisions, an approximate instrument reading may be adequate; for close physiological comparisons or published experiments, uncertainty can decide whether a small change is real. The strength of the claim should match the evidence.
9. Biological response cannot be computed from a mixing ratio
Mud crab salinity tolerance varies by species, age, stage and conditions. A salinity change that a larger, acclimated grow-out crab handles may be inappropriate for a larval stage. SEAFDEC/AQD publishes distinct hatchery, nursery and grow-out guidance, making the stage distinction explicit. A translation should preserve species name where stated, developmental stage and whether the husbandry protocol was intended for animals in that same situation.
The magnitude of a salinity change is not its rate. A fall of five units within five minutes and the same fall over many measured hours are different histories. Even a recorded average change per hour can conceal steep steps. When a paper studies “acclimation,” ask how salinity was changed, how frequently measurements were made and which biological outcomes were assessed. There is no universal safe rate that can be inferred from the arithmetic alone.
Animal response may include altered movement, feeding, survival, osmoregulatory stress or moulting behavior, but those observations require appropriate measurements. A simple survival percentage cannot describe every stress response, and a few surviving crabs do not prove the procedure was harmless. Keep mortality records, behavior notes, growth measurements and water-quality readings separate. A concise result can say the procedure reached the target concentration while leaving physiological success unestablished.
High or low salinity may co-occur with oxygen, temperature or ammonia problems. If an operator changes water, several environmental factors may change simultaneously. A claim that salinity alone caused the subsequent growth improvement is therefore weak without suitable controls. State which intervention was performed and which outcomes were observed. Distinguish a scientifically supported inference from an operational hypothesis requiring further testing.
10. Construct a complete tank-acclimation record
Consider a fictional 800-liter holding tank initially reading 30. An operator plans to move toward 25 by replacing well-mixed water with a source reading 10. Solve f=(30-25)/(30-10)=0.25; the planned constant-volume replacement is 200 liters. The operation sheet should say 200 liters removed, 200 liters added from source water reading 10, and the intended predicted mixed value of 25. It should not claim success before the post-mixing reading is available.
Imagine the actual tank measurement after circulation is 26.1. This differs by 1.1 units from the ideal prediction. Plausible explanations include source-water variability, meter uncertainty, incomplete mixing, inaccurate volume estimation or unrecorded flows. It would be unscientific to select one explanation as fact without further checking. In the revised ledger, keep predicted = 25 and observed = 26.1 in different fields, with observation time and device recorded.
Suppose 70 live crabs entered the tank, 68 were counted alive after the event, one was found dead and one was missing at census. The observed live count fraction is 68/70, about 97.1%, but attributing both differences to salinity would be unjustified, and treating the missing animal as a confirmed mortality would conflate categories. Report one documented death and one unresolved census difference. For a controlled study, predefine assessment times and outcomes before comparing procedures.
If a caretaker repeats the exchange the next day, the starting salinity for the second calculation should be the validated reading at that time, not the original 30 or the first day’s planned 25. A report that carries the target forward as though it were measured accumulates error. Separate the planned schedule from actual states and explain any departure before using a value as an input to the next calculation.
11. Environmental confounders and the water-quality envelope
Dissolved oxygen is measured independently and has its own units, often mg/L. An apparently correct salinity cannot compensate for inadequate oxygen. Temperature affects metabolism and measurement, while pH and toxic nitrogen species require their own interpretation. In a readable summary, make clear whether the environmental measurements were instantaneous, average, minimum or maximum values, and whether they covered the critical exposure interval.
Suppose an experiment compares two salinity treatments but the lower-salinity tank also runs warmer and has a different stocking density. A difference in growth cannot automatically be assigned to salinity. The appropriate report describes the multiple differences and the experimental design’s limitations. If tanks, not individual animals, were randomized, the tank is the experimental unit for many treatment comparisons. Counting crabs as independent replicates can inflate apparent confidence.
A before-and-after comparison can also be misleading when the cohort matured, when feed changed, or when handling differed. To argue that an acclimation plan caused an outcome, the evidence should address those alternatives. Translating “after” as “because of” turns chronology into causation. The repair is simple: name the temporal association first, then indicate whether the design supports a causal statement.
The absence of visible mortality within one day is not proof of longer-term welfare or performance. Some effects might appear later, and a survival endpoint alone does not describe feeding, recovery or growth. If the record stops at 24 hours, say so. If a later follow-up is planned but missing, keep that as missing. Accurate prose protects against overinterpreting an endpoint that a study did not actually measure.
12. Rounding, measurement error and approximate mass balances
Suppose a 500-liter tank volume was estimated to the nearest 25 liters and initial salinity is reported as 28 ± 1 on the instrument’s scale. A calculation producing a target of 24.800 should not be displayed to three decimals as if those inputs were exact. “Approximately 25” is more honest, with the post-mixing measurement determining whether the operational target was reached. Retain sufficient precision internally for auditing but round the reported result responsibly.
If an operator replaces 100 liters from a nominal 500 liters, the planned exchange fraction is 20%. If actual volume is somewhere from 475 to 525 liters, that fraction ranges from about 19.0% to 21.1%, even before uncertainty in flow measurement. That uncertainty directly affects the predicted salinity. Simple interval calculations can be more informative than a false display precision when the inputs are operational estimates.
Volume and mass mixing are not exact synonyms when density differs. A more rigorous dissolved-salt mass balance uses solution masses and salt mass fractions; it does not assume every liter has exactly one kilogram. For many low-precision field planning exercises, the simplified volume-weighted model is serviceable, provided it is labelled approximate and checked by measurement. For research requiring higher precision, readers should use the appropriate thermodynamic salinity scale and density information.
Temperature compensation is not an optional rhetorical flourish. Conductivity changes with temperature, and sensor algorithms attempt to standardize the relationship. A paper that reports raw conductivity from different temperatures needs the raw conditions and conversion method to compare them. Do not invent a universal correction coefficient for every water composition and instrument. Preserve the method’s calibration standard, stated uncertainty and the units in which results were actually produced.
13. Common translation and spreadsheet failures
Failure one: using ppt, PSU, grams per liter and conductivity units interchangeably. Repair by recording the original label, instrument and method, then translating to a standardized presentation only when justified. Failure two: applying constant-volume replacement math to freshwater addition. Repair by drawing the flow arrows and writing the initial and final volume. Failure three: copying a planned target as the achieved result. Repair with separate columns for calculation and later measurement.
Failure four: treating the source water as zero salinity without measuring it. Repair by including source-water observations and sensitivity to their variability. Failure five: calculating a single exchange when records describe repeated changes. Repair through the correct recurrence, using the measured starting value at each step. Failure six: summing percentages across steps as though they were additive fractions of original water. Repair by following what proportion of the prior mixture each step removes.
Failure seven: writing “safe acclimation” when only a salinity target was reached. Repair with clearly specified biological endpoints and appropriate observation time. Failure eight: omitting the particular Scylla species or life stage. Repair by preserving the identification and limiting advice to the studied cohort. Failure nine: converting every report into a causal claim. Repair by stating observations, timing and confounders before proposing mechanisms.
Failure ten: forcing absent observations into zeros. A missing pH reading is not pH zero, and a missing survival follow-up is not no mortality. Mark missing values explicitly. Failure eleven: assuming a sample beside the inlet represents an entire tank. Record location and mixing. Failure twelve: hiding overflow or continuous discharge. The true water balance may have more outflows than the simple classroom formula assumes. Build the narrative around actual flows, not just the desired answer.
14. Independent exercises with checked answers
Exercise A: A 400-liter tank reads 30 and receives a one-step 20% replacement with nominal freshwater. Under the simple constant-volume model, predicted final is 30 times 0.80 = 24. The operator must remove 80 liters and then add 80 liters to restore starting volume. If 80 liters are merely added without removal, the final volume would be 480 liters and the predicted salinity 30 times 400/480 = 25, not 24.
Exercise B: A tank reads 18 and is to approach 20 using source water reading 32 through one constant-volume exchange. The required fraction is (20-18)/(32-18) = 2/14, approximately 14.3%. For a 700-liter tank that is about 100 liters replaced. It is still only a planning estimate; the post-mixing sensor reading is needed. If source water were less saline than 20, no ordinary blend of those two waters could reach the stated target.
Exercise C: A 1,000-liter tank reading 22 loses 100 liters of essentially freshwater by evaporation, with no other exchange. The nominal new volume is 900 liters and the predicted value 22 times 1000/900 ≈24.44. Evaporation is not a recommendation or controlled acclimation procedure; it is a balance explanation. If salt deposits or exchange occurred, the assumptions no longer hold and the result must be revised.
Exercise D: A reservoir reading 32 is mixed with a source reading 12 to obtain a nominal 27. Solve x=(27-12)/(32-12)=0.75 for the high-salinity fraction. A final 600-liter mixture needs about 450 liters at 32 and 150 liters at 12. Confirm both source readings immediately before mixing if the decision depends on accuracy. The volume model is approximate, and the mixed result should be checked directly.
Exercise E: A repeated replacement removes 10% of well-mixed 30-salinity water and replaces it with zero-salinity source three times. The predicted sequence is 27, 24.3 and 21.87. Do not call the final 21.87 an exact field measurement; it assumes complete mixing, constant starting volume each time and no other flows. The salinity decline is not three times three points because each stage starts at a different concentration.
Exercise F: A 2,000-liter pond receives 20 mm of rainfall over 25 square meters. Rainfall volume is 0.020 times 25 = 0.50 cubic meters, or 500 liters. If all rainwater is retained and salinity begins at 20, nominal final is 20 times 2000/2500 = 16. If the pond spills or mixes imperfectly, the actual result differs. The data needed for a factual report include outlet behavior, measured salinity and timing.
15. Frequently asked questions
Is practical salinity the same as parts per thousand? Not exactly. Practical Salinity is defined by a conductivity-based scale, while parts per thousand may denote an approximate mass fraction in field shorthand. Numerical values are often similar enough for some practical comparisons, but technical translation should retain the method and scale. Never assert exact interchangeability without knowing how the numbers were obtained.
Can dilution arithmetic tell me how fast to acclimate crabs? No. It predicts a mixture under assumptions. Safe acclimation depends on life stage, species, past conditions and other water qualities. The method, observation period and husbandry guidance must be assessed separately. An idealized calculation neither establishes a safe rate nor replaces direct monitoring of the animals and the system.
Can a lower salinity reading prove the exchange worked? It shows that the measurement at that location and time is lower. Whether an entire tank has mixed, whether the intended volume was exchanged, and whether other factors changed require additional evidence. Use multiple readings or an appropriate mixing check if homogeneity is important. An outcome should not be overstated just because the arithmetic is plausible.
What if the source-water salinity changes after rain? Then the source value in the exchange calculation changes. Use the measured value corresponding to the actual operation. A stale source reading can undermine the plan even when every algebraic step is correct. Keep source and tank measurements separately dated. The operator’s planned target and achieved observations should remain distinguishable in any summary.
Why should I record both water volume and salinity? Because a concentration change alone cannot reveal how much salt or water moved. Addition, replacement, overflow and evaporation can produce superficially similar concentration readings through different mechanisms. With a flow ledger, readers can reconstruct the scenario and identify inconsistent notes. Without it, causal interpretation is often speculative.
How should an experiment report survival? Give stocked and surviving counts, the assessment time and the status of missing or transferred crabs. Do not call a missing animal dead without a basis. Where multiple tanks are used, distinguish replication by experimental units from individual observations. Salinity treatment claims also need concurrent environmental information and an appropriate design.
16. A reusable language-and-calculation checklist
A robust report specifies species and life stage, tank or pond identity, initial measured condition, source-water condition, exchange or addition method, volume and timing, predicted mixing result, actual post-mixing readings, sensor method, other water quality and biological observations. It also makes the unit or scale explicit. These details allow someone who does not speak the original writer’s language to reproduce the reasoning without guessing the workflow.
A useful sentence might say: “At 09:00, 200 liters were withdrawn from the 800-liter tank after circulation and replaced with 200 liters of source water reading 10 on the field salinity scale. Under a one-step mixing approximation, the planned final value was 25. After circulation, the sensor read 26.1 at 09:45.” This statement avoids claiming exact salt mass, biological safety or a cause for the discrepancy. It is specific without being overconfident.
When merging spreadsheets, give each observation a timestamp, tank ID, parameter, numerical value, unit, method and quality flag. Keep missing observations separate from zeros. If a technician corrected a volume entry, retain the corrected value with a traceable change note rather than silently overwriting the original. A multilingual record should preserve scientific names, unit conventions and uncertain points exactly enough for another reader to audit.
Once a reader can distinguish concentration from mass, planning from observation, mixing from acclimation and salinity from wider water quality, the technique transfers beyond mud crabs. It applies to brackishwater reservoirs, hatchery preparation, fish holding tanks and many environmental reports. The central habit is to ask: what entered, what left, when was it measured, what assumptions created the prediction, and what evidence supports the biological conclusion?
Sources and further reading
SEAFDEC/AQD, Farming Mangrove Crab (Mud Crab): https://www.seafdec.org.ph/mangrove-crab-mud-crab/
SEAFDEC/AQD, Biology and Hatchery of Mud Crabs, Aquaculture Extension Manual No. 34: https://repository.seafdec.org.ph/handle/10862/3104
FAO, Mud crab culture: a practical manual: https://www.fao.org/4/ba0110e/ba0110e.pdf
Google Search Central, Creating Helpful, Reliable, People-First Content: https://developers.google.com/search/docs/fundamentals/creating-helpful-content
Related eduKate guide, Mud Crab Feed Rations and Feed Conversion Ratios: https://edukatesg.com/2026/10/31/how-to-translate-easily-mud-crab-feeding-ration-wet-feed-fcr/