A partial-harvest report is a stock ledger, not just a sales receipt
A shrimp pond might report 500 kilograms removed in its first harvest, another 900 kilograms a fortnight later, and 3,850 kilograms at final clearance. The reader can add those weights to obtain 5,250 kilograms of total harvested production, but that addition does not reveal how many shrimp were originally stocked, how many animals survived, or how many remained after the first two removals. Those questions need different numbers and different denominators.
Partial harvesting means taking some marketable animals while leaving others in a culture unit for additional growth or later removal. The retained population may change in number, size distribution and biomass after every event. A report that treats the pond as one unchanged population from stocking until final harvest can confuse growth, mortality, selection and deliberate removals. An accurate translation follows the population from one event to the next.
This guide explains how to reconcile event-level harvest weights, counts per kilogram, estimated live numbers, transfers, mortalities and final production. The examples are hypothetical teaching ledgers, not observations from a farm. Real handling, netting, food safety, welfare and biosecurity decisions belong to trained operators and competent authorities. The aim is to read and check the evidence without inventing animals, weights or causes for any unexplained difference.
Start with the right question
A farmer may want to know how many kilograms were sold this month. A biologist may want cumulative survival from the original stocking. An accountant may ask how much was paid for each grade. A production manager may need the estimated number still in the pond. These are related but separate questions, even when all appear on one harvest sheet.
A strong reading begins by naming the desired output and its unit. Harvested mass is measured in kilograms or tonnes. Harvested animal count is a number of individuals, sometimes estimated. Survival is a proportion of an explicitly defined starting population. Remaining stock is a count or biomass at a particular date. Revenue is a monetary amount subject to pricing and settlement rules.
When the report says yield increased, ask which yield. It could mean kilograms per harvest event, cumulative kilograms per pond cycle, saleable kilograms after deductions or kilograms per hectare of pond area. No single number can safely substitute for the others. Precise nouns and denominators matter more than a striking headline.
Know what the published research actually concerns
First-party aquaculture references from the Food and Agriculture Organization describe complete and partial harvesting as different management choices rather than synonyms. Research records indexed by FAO AGRIS examine their economic and production consequences under particular assumptions about growth, risk, prices and scheduling. These sources establish relevant concepts, not a universal rule that one harvest pattern always outperforms another.
The whiteleg shrimp species profile, for example, describes grow-out systems where harvest may be staggered as animals reach a marketable size. That describes an operational possibility. It does not tell us the exact survival, profit, remaining count or welfare outcome for an unrelated pond. A field ledger must stand on its own observations and methods.
A translation should therefore separate a reported farm result from a modelled scenario and from a general description of farming practice. A simulation may forecast higher revenue under one assumed schedule while an actual season may produce a different outcome. Preserve the author’s conditions before carrying a numerical claim into practical advice.
Write the timeline before calculating ratios
A partial-harvest pond has a sequence: initial stocking, optional later stockings or transfers, growth, removal events, recorded mortalities, further growth and final clearance. Each state change belongs at a date. Without a timeline, a reader may add a count measured before the first harvest to a count measured after the second as though both referred to the same population.
Imagine a pond stocked on 1 March, partially harvested on 1 June and 15 June, then finally harvested on 1 July. A weight sample from 10 June describes the remaining animals after the first removal. It cannot automatically describe the animals taken on 1 June. Likewise, a stock estimate from 20 June already reflects both preceding harvests.
Use a dated event table with event type, beginning state, incoming animals, outgoing animals, observed deaths, estimated remaining animals, weighed product and method. Include an explicit distinction between measured values and calculated projections. The chronology protects the reader against counting the same biomass twice.
Separate individuals from kilograms
Animal count and total mass are connected by average individual mass, but they are not the same quantity. A harvest of 500 kilograms containing relatively small shrimp can represent more individuals than 500 kilograms containing larger shrimp. To estimate numbers from mass, a compatible average mass or count-per-kilogram measurement is required.
If a graded harvest contains approximately forty shrimp per kilogram, a 500-kilogram portion represents about 20,000 shrimp. If another 500 kilograms contains approximately twenty-five shrimp per kilogram, it represents about 12,500 shrimp. Identical harvest weight does not mean identical removal from the living population.
For stock reconciliation, count removals in individuals when survival is the question. Sum kilograms when production mass is the question. Switching between these units without the conversion data can make a report appear internally consistent while leaving its underlying animal accounting unresolved.
Count per kilogram is the inverse of average mass
The seafood trade often expresses shrimp size as pieces or count per kilogram. This is an inverse size convention: a larger numerical count usually means smaller average individuals. A group with forty shrimp per kilogram has an average mass of 1,000 grams divided by forty, or twenty-five grams per shrimp.
A group with twenty-five shrimp per kilogram has average mass forty grams. Moving from forty per kilogram to twenty-five per kilogram therefore means the average individual became larger, even though the size count number fell. An inexperienced reader may wrongly describe this as a reduction in product size.
Some markets use count per pound, broad count bands or commercial grades rather than an exact observed count. Forty per pound is not forty per kilogram. Identify the weight unit, whether the value is measured or merely a grade range, and whether shrimp are whole or processed before converting it into animal numbers.
Whole, headless and peeled mass need different labels
A mass listed on a landing or sales form might represent whole shrimp, headless shell-on shrimp, peeled product, drained product or packed product with added materials. These are distinct weight bases. Dividing a processed mass by the average whole-animal mass will not accurately recover the number of harvested animals without a justified recovery factor.
Suppose 1,000 kilograms of whole shrimp are harvested, while 650 kilograms of a particular processed product are recorded. The 650 kilograms is a product recovery outcome under an explicitly stated processing method, not evidence that only 650 kilograms were harvested alive. A reader needs both values and an account of the processing steps.
A properly translated ledger keeps biological harvest, post-harvest processing and sales settlement in separate columns. If a source supplies only one weight without identifying its basis, do not insert the words live harvest or whole shrimp. State that the product form is unspecified.
Gross, tare and net are accounting terms
A delivery vehicle may register a gross weight including containers, ice and packaging. Net shrimp mass requires subtracting the correct tare under the facility’s measurement rules. The result is not automatically saleable mass if grading rejects, drainage deductions or other adjustments occur afterward.
Imagine a hypothetical batch weighed as 650 kilograms gross with fifty kilograms of properly documented non-shrimp tare. Its net weighed material is 600 kilograms. If twenty kilograms are then excluded by a documented product-quality rule, the accepted saleable weight is 580 kilograms. Each quantity has a different description and place in the ledger.
Do not assume that every difference between gross and net represents ice, nor that every difference between net and paid weight is product loss. Audit the actual weigh tickets, tare records, receiving notes and contract terms. The arithmetic must follow the documented categories rather than a guessed explanation.
Harvesting changes the remaining count immediately
When shrimp are removed from a pond, the number remaining decreases regardless of whether the animals removed were small or large. In a simple closed stock with no new entries or deaths during the interval, remaining animals equal the previous estimated number minus the harvested animal count. In practice, uncertainty and additional movements must be recorded too.
Suppose a pond is estimated to contain 100,000 shrimp immediately before a partial harvest. A documented removal consists of an estimated 18,000 animals. The estimated number remaining is 82,000, assuming no other additions or losses occurred during the interval. That is an accounting projection, not a new physical census.
If a later cast-net observation suggests only 75,000 remain, the difference of 7,000 is an unresolved discrepancy until sampling variability, natural mortality, transfers or harvest-count errors are considered. Calling all 7,000 dead would turn a residual into an unsupported biological claim.
Create a conserved-number accounting equation
For a defined pond and time interval, the basic animal-count reconciliation is ending number equals beginning number plus entries minus harvested removals minus other exits minus deaths, with an allowance for unresolved discrepancy. Entries might include an additional stocking or documented transfer in. Other exits might include transfers to another unit.
In symbols described in words, the estimated ending count is the estimated beginning count plus animals entering, less animals harvested, transferred out and confirmed dead, adjusted when reliable measurements reveal an unresolved difference. Do not force the difference into one biological category merely to make the ledger balance.
The equation is useful as an audit structure even when actual animal counts are uncertain. Its purpose is to expose what information is missing. If an item is unobserved, label it unknown rather than zero. A reader can then understand why final survival is estimated with limitations rather than presented as a directly counted fact.
Transfers are not harvests
A movement of 10,000 live shrimp from pond A into pond B is an exit for pond A and an entry for pond B, but it is not a sale or harvest if the animals remain in culture. Treating it as harvested production for A and later again for B would double-count output across the farm.
At farm level, internal transfers should cancel when both sides are reconciled. The source pond subtracts 10,000; the destination adds 10,000. If counts differ because of transport losses or measurement uncertainty, those differences should be investigated and disclosed, not hidden by adding both totals.
A translation should distinguish moved, stocked, harvested, sold, returned and discarded. Each verb describes a different flow. The correct wording determines whether an event belongs in survival, physical inventory, total production, internal movements or sales revenue.
Partial harvests can change the size distribution
Harvest gear and market grading may selectively remove larger animals, smaller animals or certain accessible groups. After a size-selective removal, the mean weight of the remaining shrimp can be quite different from the pre-harvest mean even if no animal grows between the two observations.
Imagine a pond with 1,000 shrimp weighing twenty grams and 1,000 weighing forty grams. Its initial mean mass is thirty grams. If all 1,000 forty-gram shrimp are harvested, the 1,000 remaining shrimp have mean mass twenty grams. That apparent drop does not indicate that individual shrimp lost weight.
The reverse can also occur if a harvest removes mainly small animals. A careful report preserves size-grade information before and after removal. Otherwise it may mistakenly interpret a shift in the population composition as a biological growth or shrinkage effect.
Use a weighted size count when combining grades
A mixed harvest cannot be converted into an animal count by simply taking the arithmetic mean of its grade counts. Each grade has a different mass and count per kilogram. The number removed is the sum of each grade’s mass multiplied by that grade’s measured count per kilogram.
Suppose 200 kilograms contain approximately thirty shrimp per kilogram, while another 300 kilograms contain forty shrimp per kilogram. Estimated animal counts are 6,000 and 12,000. The combined 500-kilogram batch therefore contains approximately 18,000 animals, equivalent to an overall count of thirty-six shrimp per kilogram.
The simple average of thirty and forty is thirty-five, which is not the correct mass-weighted count for this example. The difference matters when translating sale grades back into biological removals. The report should retain both grade masses so the result can be independently checked.
A grade band is not always an exact count
A sales specification labelled thirty to forty shrimp per kilogram may describe an accepted size interval, not a sample proving exactly thirty-five shrimp per kilogram. Multiplying the mid-point by the sold mass creates an estimate resting on an untested distributional assumption. It is not a measured count.
For a 100-kilogram lot under a thirty-to-forty count band, a crude bounded scenario would represent between 3,000 and 4,000 shrimp if all individuals met the band definition and there are no additional complications. The actual number depends on how sizes are distributed and what the band means contractually.
When the farm has an observed subsample count, use that with its method and uncertainty. If only the grade band is known, present a range or explicitly labelled assumption. Avoid laundering a trade category into a falsely exact biological tally.
Estimated live stock differs from a weighed harvest
A final harvested batch can be weighed after removal. Before final clearance, the remaining stock is usually estimated from samples, feeding records, mortality observations or a model. Both figures may use kilograms, but their evidentiary status differs.
Suppose a pond is predicted to hold 2,000 kilograms on Monday. A partial harvest physically weighs 500 kilograms on Tuesday, after which a spreadsheet predicts 1,500 kilograms remain. That subtraction assumes no growth, mortality or measurement changes between the relevant times and that both masses use the same basis.
For a short interval, the estimate may be operationally useful. It should still be marked predicted remaining biomass, not weighed remaining biomass. An audit trail showing what was measured and what was inferred protects future management calculations.
Growth continues between harvest events
While some shrimp leave, the remaining animals can continue to grow. Therefore, biomass after a partial harvest is not necessarily the pre-harvest biomass minus harvest mass when observations are separated by days or weeks. Feeding, mortality, water conditions and size-specific growth alter standing biomass over time.
Imagine a hypothetical pond at 2,000 kilograms before a 500-kilogram partial harvest. Immediately afterward, a simple no-loss calculation gives about 1,500 kilograms. If remaining animals subsequently gain a combined 300 kilograms and another fifty kilograms are lost before a later sampling date, the later biomass would be around 1,750 kilograms.
A report that compares the later 1,750 kilograms to the immediate 1,500 kilograms may reasonably discuss net gain, subject to measurement uncertainty. It should not describe that difference as the weight of newly stocked shrimp. Each event requires its own category.
A complete stock ledger with three harvests
Consider a hypothetical pond stocked with 200,000 postlarvae and receiving no later entries or transfers. At the first partial harvest, 20,000 animals are estimated to be removed at twenty-five grams each, producing 500 kilograms. At the second, 30,000 animals are removed at thirty grams each, producing 900 kilograms.
At final clearance, 110,000 animals are counted or reliably estimated at thirty-five grams, producing 3,850 kilograms. Total harvested animal count is 20,000 plus 30,000 plus 110,000, or 160,000. Total harvested mass is 500 plus 900 plus 3,850, or 5,250 kilograms. These two cumulative totals answer distinct questions.
Under the simplified assumptions of a closed cohort and complete final clearance, 160,000 of the original 200,000 were harvested, giving an apparent cumulative realised harvest fraction of eighty percent. The difference of 40,000 animals is not automatically confirmed mortality: it may combine losses, undercounting and any unrecorded movements. A rigorous ledger explains the evidentiary status of that residual.
Yield per area needs a defined pond area
Production can be normalised by the area used for culture, commonly reported as kilograms per hectare per crop. One hectare contains 10,000 square metres. The calculation is total relevant production mass divided by pond area expressed in hectares, not divided by a net-cast area or a processing-factory floor area.
Using the hypothetical 5,250-kilogram cumulative harvest from a pond with an area of 0.75 hectares, crop yield is 5,250 divided by 0.75, or 7,000 kilograms per hectare per crop. This is cumulative output across all partial and final harvests, not just the last event.
The pond area should be defined consistently as productive water surface or another explicitly declared basis. Comparing one study’s actual water area with another’s property boundary can distort results. Preserve the area denominator and the crop duration when reporting performance.
Per crop and per year are different rates
A report giving 7,000 kilograms per hectare per crop is not automatically claiming 7,000 kilograms per hectare per year. Annualised productivity depends on how many crops actually occurred or can responsibly be completed, including preparation and downtime.
For example, a facility with two completed equivalent crops per year at 7,000 kilograms per hectare per crop would produce 14,000 kilograms per hectare per year, assuming each crop achieves that output and the land is consistently used. A single shorter crop does not prove that twice the output can be achieved every year.
An annualisation based on an assumed cycle frequency should be labelled a projection. Real production records with two or more completed cycles provide stronger evidence. Season, disease management, maintenance, weather and downtime can change the feasible rate.
Gross production and accepted production differ
A harvest ledger might show 5,250 kilograms physically removed, but the receiver may accept less after grading, quality rejection or processing. Biological production and marketable production therefore can diverge. A report should state what happened between the pond and the paid product.
Suppose 5,250 kilograms are weighed at the pond, while accepted saleable mass is 5,050 kilograms after documented exclusions and adjustments. The difference is 200 kilograms under the stated weighing and acceptance rules. It should not automatically be assigned to shrinkage, disease, spoilage or fraud without evidence.
Track pond-weighed output, landed material, saleable grades, rejected or returned material and settlement weights separately. The distinction matters for both animal counting and financial reporting. A saleable mass figure cannot silently replace a biological harvest figure in a survival calculation.
Price and revenue need grade-specific multiplication
More kilograms do not always mean more revenue per kilogram, because prices can vary by size grade, season, quality and contract. Revenue should be calculated as the sum of accepted mass for each grade multiplied by its applicable price, before any separately identified fees or deductions.
A hypothetical invoice sells 200 kilograms at twelve currency units per kilogram and 300 kilograms at ten. Gross billed value is 2,400 plus 3,000, or 5,400 currency units. The weighted average price is 5,400 divided by 500, or 10.8 units per kilogram. The simple average of the two listed prices, eleven, would be wrong for the mixed quantities.
A profitability claim needs costs as well as revenue: feed, labour, energy, gear, mortality, handling, transport and other operating expenses. A higher gross receipt from partial harvests does not, by itself, establish a higher net profit.
Timing can change sales prices and costs
Partial harvesting may allow some product to enter a market earlier while other animals remain for later growth. That creates a schedule of cash receipts and continuing costs. A whole-cycle total that ignores timing cannot show whether the strategy improves cash flow, financing risk or return on working capital.
Suppose two hypothetical systems each sell 5,000 kilograms over a season, but one receives a substantial early payment and another sells everything at the end. Their cumulative physical output is equal, while the timing of cash flows differs. A discounted cash-flow analysis may assign different values even before considering market price changes.
A fair economic interpretation must specify the assumptions behind discount rates, sales prices, mortality, handling costs and future growth. FAO AGRIS research on partial-harvest strategies treats such decisions as uncertainty-sensitive, not universally beneficial.
Different harvested sizes need different denominators
If a manager says fifty percent of the crop was harvested at the first event, the listener must ask fifty percent of what. It could mean half the estimated standing biomass, half the individuals, half the saleable product projected for the season or half of a target volume. Those percentages will generally differ.
Imagine a pond with an estimated 2,000 kilograms and 100,000 animals immediately before an event. If 500 kilograms containing 20,000 larger shrimp are removed, the share of standing biomass removed is twenty-five percent, while the share of estimated animal count removed is twenty percent.
Both fractions are valid under their assumptions. Translating either as a generic fifty-percent harvest rate would erase the reference population. In every table, put the numerator and denominator in the header or an immediately adjacent note.
Partial harvest survival is not the last harvest fraction
When a pond has multiple harvests, calculating survival as animals remaining at the final event divided by original stocking underestimates the cumulative fraction recovered. Animals taken earlier were alive at their own harvest dates and must remain in the numerator for realised cohort recovery.
Under the earlier hypothetical ledger, 110,000 animals remain for final clearance from 200,000 initially stocked. The final-harvest count alone equals fifty-five percent of stocking. But adding 20,000 and 30,000 taken alive at earlier events yields 160,000 total animals harvested, or eighty percent.
The fifty-five-percent value is the final event’s animal count relative to stocking, not the crop’s cumulative realised harvest fraction. This distinction is one of the most consequential repairs when translating partial-harvest statistics for parents, investors or scientific readers.
Reconcile late deaths without counting them as harvested
If an animal dies in the pond and is not harvested for sale, it is not a live harvested survivor. If mortality occurs during transport or post-harvest handling, the classification depends on the stated survival endpoint. Record biological status and timing separately from sales acceptance.
Consider an originally stocked cohort of 10,000, with 3,000 sold in one partial harvest, 5,000 sold at final harvest, 500 documented deaths during culture and 1,500 unaccounted at final clearance. Cumulative sale recoveries total 8,000, or eighty percent; documented culture mortality is five percent; unresolved difference is fifteen percent.
It would be incorrect to announce twenty-percent confirmed mortality, because only five percent was actually documented. The difference between accounted and unaccounted stock is a discrepancy that calls for better evidence, not a licence to assign a specific cause.
Do not count the same animal twice after sorting
During harvest, some captured animals may be measured, graded and returned alive to the pond. The fact that they entered a net does not mean they were permanently removed from the population. If the same returned animals are caught again later, counting both captures as harvested would inflate cumulative animal removals.
A useful event sheet differentiates gross caught animals, animals returned, animals transferred to another unit and final animals retained or sold. Only the relevant permanent removals should enter the cumulative harvest count. Repeated captures may be operational events rather than additional production.
For mass accounting, the same principle applies. Material weighed during sorting and returned to culture is not an incremental sold harvest. A ledger that separately records handling and final disposition prevents double-counting.
An estimated harvest count can differ from a physical count
A partial-harvest sheet may physically weigh 900 kilograms but estimate 30,000 shrimp from a thirty-three-and-one-third-grams approximate mean mass. The 900 kilograms might be directly measured while the 30,000 is derived from a subsample. They do not have the same measurement precision.
Suppose a small sample suggests thirty-three grams, but true mean mass of the harvested group were thirty grams. A 900-kilogram batch would then represent about 30,000 shrimp, rather than roughly 27,273 inferred at thirty-three grams. That is a material difference in the estimated number of animals removed.
A careful summary states which count is estimated from grade or sampled body weight. When calculating survival, propagate that uncertainty instead of treating every removal count as an exact census.
Sample selectivity carries forward into the ledger
A net sample may overrepresent larger, more active or spatially concentrated shrimp. If the sample average body weight is then used to infer removal count from a weighed harvest, the resulting population accounting can be biased. A perfect arithmetic conversion cannot correct an unrepresentative weight sample.
Imagine that a catch-based sample suggests a twenty-five-gram average, while an independently representative landing sample suggests twenty grams. For a 500-kilogram harvest, those give estimated counts of 20,000 and 25,000 animals respectively. A five-gram difference changes inferred removal count by 5,000.
Readers should inspect whether the mass sample was taken from the actual harvested product or from the live stock before the event. These are different sampling frames. The interpretation must match the population represented by the estimate.
Track standing biomass and cumulative harvest separately
Standing biomass is the estimated live mass in the pond at a particular moment. Cumulative harvested biomass is the total mass removed across events up to that moment. As time passes, cumulative harvest can only increase under a consistent accounting rule, whereas standing biomass can rise or fall.
Suppose cumulative harvest after two events is 1,400 kilograms and estimated standing biomass is 2,100 kilograms. Adding them gives 3,500 kilograms as a snapshot of past removals plus current standing stock, but does not mean the pond has physically held 3,500 kilograms all at once.
If the remaining animals grow before final clearance, cumulative final production may exceed that snapshot total. A translation should explicitly label observed removals and estimated current stock, rather than calling their sum current biomass.
Harvest frequency is not an outcome by itself
Three harvest events are more events than one, but that is not proof of greater total production. More frequent removals may reduce crowding or create earlier sales, yet they may also increase labour, disturbance or market handling costs. Their effects depend on the species, environment, strategy and animal responses.
A meaningful comparison measures cumulative yield, grade distribution, survival, growth, welfare and economic results under comparable conditions. Counting harvest events alone ignores the size and efficiency of each event. A study that models an optimal schedule is conditional on its biological and financial assumptions.
When a source claims that frequent harvest improves production, identify the exact outcome improved and the experimental or modelling evidence. Avoid replacing a qualified result with the sweeping statement that more harvests are always better.
A harvest target is not an achieved quantity
A planning spreadsheet might call for removing 800 kilograms at the first event and leaving 2,000 kilograms in the pond. Those are intended amounts, not measured results. A later actual harvest ticket of 720 kilograms supersedes the target for the realised production ledger.
Suppose a harvest plan anticipated twenty-five percent of biomass removal, while weighed output was only 720 kilograms against an assumed pre-harvest 3,200 kilograms. Realised removal was 22.5 percent of the assumed mass, not the planned twenty-five percent. Even this value remains conditional on the quality of the 3,200-kilogram estimate.
A translation must retain the words projected, targeted, estimated, measured and settled where appropriate. They mark different evidentiary stages. Confusing a target with an achievement can falsely improve an entire season’s statistics.
Review why a ledger fails to balance
A discrepancy does not necessarily mean fraud, mortality or defective mathematics. It may arise from missing transfers, estimated animal counts, inconsistent weight bases, grade misclassification, unrecorded returns, timing mismatches or samples that do not represent the whole crop.
Start by checking identities and time stamps. Do the stocking, transfer and harvest records refer to the same pond? Was the count per kilogram measured for the same grade and product form? Did one event have a corrected invoice or a returned batch? Is the final stock truly zero?
Only after those administrative and sampling questions are considered should a biological explanation be advanced. A useful audit shows the residual as a separate line and identifies which new evidence would narrow it.
A fully balanced numerical event ledger
Consider a hypothetical pond beginning with 100,000 shrimp. It receives 5,000 additional animals by a documented transfer and sends 3,000 to a nursery research unit, leaving 102,000 before harvest or mortality. The first harvest permanently removes 20,000, the second 25,000 and the final harvest 50,000.
Total harvested animals are 95,000. If 2,000 deaths were individually documented during culture and the final pond was confirmed empty, then the accounting identity is 100,000 plus 5,000 minus 3,000 minus 95,000 minus 2,000 equals 5,000. There is an unresolved deficit of 5,000 animals, so the ledger is not actually balanced despite complete physical clearance.
One should explicitly name the discrepancy rather than claim the figures balance. To produce a balanced hypothetical example, suppose an additional documented 5,000-animal transfer out occurred. Then total exits are 3,000 plus 5,000, harvested count 95,000 and deaths 2,000, which exactly account for 105,000 entrants. The correction requires documentary evidence in a real record; it cannot be made up just to eliminate a residual.
Worked example: counts from weighed grade lots
A hypothetical partial harvest produces 150 kilograms at forty shrimp per kilogram and 250 kilograms at thirty-two shrimp per kilogram. The estimated number removed is 150 times forty plus 250 times thirty-two, equalling 6,000 plus 8,000, or 14,000 shrimp.
Combined mass is 400 kilograms. Overall weighted count is 14,000 divided by 400, or thirty-five shrimp per kilogram. Average individual mass is 1,000 divided by thirty-five, or approximately 28.57 grams. All calculations are internally compatible, subject to the size-count values being representative of actual lots.
If someone averaged forty and thirty-two to obtain thirty-six and multiplied 400 by that, the result would be 14,400, an overcount of 400 animals. The wrong method gave each grade equal influence despite unequal lot masses.
Worked example: gross and paid weights
A receiving ticket records a gross 1,250 kilograms including documented 100 kilograms of bins and ice. Net delivered material is therefore 1,150 kilograms. A buyer rejects a separate 50 kilograms under an acceptance rule, and the accepted weight is 1,100 kilograms.
The three figures should be labelled gross received, net material and accepted saleable product. They are not interchangeable descriptions of biological harvested mass without proof of the weighing and product basis. The rejected material still physically existed even though it was not paid as accepted product.
A sales summary using only the 1,100-kilogram accepted figure might be correct financially, but a pond production summary should also retain upstream measurements. The arithmetic is simple; the interpretive challenge is choosing the appropriate noun.
Worked example: realised survival after partial harvests
A closed hypothetical cohort starts with 60,000 animals. It supplies 10,000 individuals at the first harvest, 15,000 at the second and 27,000 at final clearance. Total animals harvested alive are 52,000, or about 86.67 percent of the original stocking.
A reader who divides the final 27,000 by 60,000 gets forty-five percent, but that is not cumulative realised harvest survival. The earlier 25,000 animals were successfully removed alive and must be included for the original-cohort recovery measure.
The remaining 8,000 difference may include mortality, measurement error or undocumented exits. With no detailed death records, the correct term is unaccounted difference, not 13.33-percent confirmed mortality.
Worked example: animal versus mass share
A simplified pond estimate immediately before harvest is 50,000 animals averaging twenty grams, or 1,000 kilograms. A size-selective event takes 10,000 animals averaging thirty grams, giving 300 kilograms removed. Animal share is 10,000 divided by 50,000, or twenty percent.
Mass share is 300 divided by 1,000, or thirty percent. A headline saying thirty percent of the pond was harvested is therefore ambiguous. It might be correct as a biomass share but wrong as a share of individuals.
The remaining 40,000 animals account for an estimated 700 kilograms immediately afterward under this simplified ledger, implying an average of 17.5 grams. This is compatible with the removal of larger animals and demonstrates size-composition change.
Worked example: cumulative pond yield
A hypothetical 0.6-hectare pond produces three harvest weights of 400, 800 and 2,100 kilograms. Cumulative crop production is 3,300 kilograms. Dividing by 0.6 gives 5,500 kilograms per hectare per crop.
If someone reports only the final 2,100 kilograms, the derived figure would be 3,500 kilograms per hectare and would omit earlier production. Both figures might describe legitimate quantities, but the lower is the final-event yield rather than full-cycle cumulative yield.
To compare annual productivity, the number and length of cycles must also be documented. Crop yield alone does not determine how often the same pond can be prepared, stocked and harvested responsibly.
Worked example: partial receipts and weighted prices
A fictional farm sells 300 kilograms at fourteen monetary units per kilogram and 700 kilograms at eleven. Sales values are 4,200 and 7,700, making 11,900 total revenue before documented deductions.
Average realised price across these accepted masses is 11,900 divided by 1,000 kilograms, or 11.9 units per kilogram. The simple mean of fourteen and eleven is 12.5 and would misstate the weighted price because the larger lot sold at the lower unit price.
A production report and an income statement should still remain separate: kilograms can be summed for harvest yield, while revenue must reflect accepted grades, prices and contract adjustments. Neither result alone determines profit.
Worked example: an unresolved biomass residual
Imagine 2,500 kilograms of live biomass are estimated immediately before a partial harvest. The event yields 600 kilograms of weighed whole shrimp. A follow-up estimate suggests 1,750 kilograms remain. The arithmetic residual is 150 kilograms relative to simple subtraction.
That residual is not immediately evidence of mortality. The two stock estimates may have sampling errors, growth may have occurred between measurements, or the weight bases may differ. A reader should ask how the original and follow-up figures were obtained and whether times and methods align.
Use the phrase unexplained difference under the stated model rather than calling 150 kilograms of dead shrimp. A transparent residual invites investigation, whereas an invented explanation disguises the weakness of the evidence.
Independent transfer problem: one additional stocking
A fictional pond begins with 80,000 shrimp and receives a later documented additional 20,000. It supplies 15,000 at one partial harvest and 70,000 at final clearance, with no other recorded transfer. Calculate total accounted harvested animals and the apparent unaccounted difference relative to all entrants.
The number entering is 100,000. Cumulative harvested animals are 85,000. Apparent difference is 15,000, or fifteen percent of entrants, if the counts are compatible. The original 80,000 is not the appropriate denominator for an all-entrant recovery fraction.
It would be misleading to report harvest survival of 106.25 percent by dividing 85,000 by 80,000. That impossible-looking ratio is a signal that the later stocking has been omitted from the denominator.
Independent transfer problem: variable grades
A harvest sells 100 kilograms of one grade at twenty-five shrimp per kilogram and 400 kilograms of another at forty shrimp per kilogram. Determine the estimated number of animals in each lot, combined count, total mass and combined pieces per kilogram.
Answers: 2,500 and 16,000 animals, 18,500 total across 500 kilograms, or thirty-seven per kilogram overall. The mass-weighted overall count is not the simple arithmetic mean of twenty-five and forty, which would be 32.5.
The larger mass of the smaller shrimp grade contributes more animals. Check the result by dividing 18,500 by 500 rather than averaging the grade labels.
Independent transfer problem: cumulative versus final survival
A pond stocks 120,000 animals and has three verified harvest removals of 25,000, 35,000 and 40,000. What fraction of the stocking is represented by the final harvest alone, and what fraction by all harvested animals?
The final event alone is 40,000 divided by 120,000, or 33.33 percent. Combined removals equal 100,000, giving about 83.33 percent of the original stocking. The difference is fifty percentage points because earlier harvested animals are missing from the final-event calculation.
If twenty thousand animals remain unaccounted, do not immediately label them dead unless the record supports that classification. Reconciliation and biological attribution are separate tasks.
Independent transfer problem: partial harvest share
A pond contains an estimated 60,000 animals averaging twenty grams. A selective removal takes 12,000 animals averaging twenty-five grams. Find estimated pre-harvest biomass, removal mass, removal share by number and removal share by mass.
Starting biomass is 1,200 kilograms. Removal mass is 300 kilograms. The removal represents twenty percent of animals but twenty-five percent of standing mass. The two percentages differ because removed animals are heavier than the pond-wide average.
After removal, the simple immediate accounting estimate is 48,000 animals and 900 kilograms, implying about 18.75 grams per remaining animal. That lower average is a compositional result, not evidence of weight loss in individuals.
Independent transfer problem: price is not profit
A farm plans two harvests of 500 kilograms each. In one scenario both lots sell for ten currency units per kilogram. In another, the first sells for twelve and the second for nine. Calculate total gross sales in each scenario and explain why a profit conclusion is premature.
The first scenario yields 10,000 units. The second yields 6,000 plus 4,500, or 10,500 units. The second has 500 units more gross sales, but labour, additional harvesting costs, financing, losses and other expenses have not been included.
This illustrates the distinction between price-weighted revenue and profitability. A higher sales total is evidence about receipts under the stated assumptions, not automatically a better whole-farm strategy.
Independent transfer problem: find the ledger error
A hypothetical record says a pond began with 50,000 shrimp, received 10,000 more, transferred 5,000 out, partially harvested 15,000, finally harvested 35,000 and had 2,000 documented deaths. Assuming no animals remain, determine the unexplained difference.
Entries total 60,000. Recorded exits plus deaths total 5,000 plus 15,000 plus 35,000 plus 2,000, or 57,000. The unaccounted difference is 3,000 animals. It would be improper to change the documented mortality from 2,000 to 5,000 without new evidence simply to make the arithmetic balance.
A repaired report keeps 3,000 marked unresolved and investigates missing records or estimates. That honest line preserves the possibility of later correction.
Frequently asked question: Is partial harvesting always more profitable?
No. Its economics depend on growth, animal survival, size-dependent prices, handling costs, market timing, farm risks and the chosen comparison. Published analyses can support a partial-harvest strategy under particular assumptions while finding different outcomes when those assumptions change.
A responsible summary gives cumulative accepted mass, sales by grade and a complete cost basis before discussing profit. Number of harvest events is not a substitute for net financial evidence.
Frequently asked question: How should survival be calculated?
For a defined closed cohort with no later entries, cumulative realised harvested-animal fraction is the sum of distinct live animals permanently removed at all harvest events divided by the original number stocked. If later animals enter or some transfer out, the denominator and accounting pathway need to be adjusted to the population and question being measured.
This fraction is not necessarily an exact biological survival probability where removal counts or stocking numbers are estimates. State the method, uncertainty and observation endpoint.
Frequently asked question: Does a remaining-stock estimate count as evidence of mortality?
Not by itself. A modelled or sampled remaining-stock count can differ from the expected ledger because of imperfect capture, inaccurate average weights, changing animal distribution, transfers and unobserved deaths. The discrepancy calls for investigation, not an assumed cause.
A genuine documented dead-animal count should be recorded separately. Missing animals and confirmed deaths are different evidence categories.
Frequently asked question: Why does count per kilogram decrease as shrimp grow?
Count per kilogram is an inverse mass measure. Larger average individuals require fewer shrimp to make one kilogram. For example, fifty shrimp per kilogram suggests twenty grams each, while twenty-five shrimp per kilogram suggests forty grams each, if counts are representative and weight basis is consistent.
Never translate a lower piece count as a smaller animal. Preserve whether the source uses kilograms, pounds or a broad commercial count band.
Frequently asked question: Is a partial harvest the same as grading?
Not necessarily. Grading sorts animals or product by size or quality. A partial harvest permanently removes a subset from the culture unit. Some grading procedures return animals to the pond, so counting every graded animal as harvested can double-count removals.
The ledger should distinguish temporary captures, returns, transfers and permanent sales. An event’s final disposition determines its category.
Frequently asked question: Should the last harvest alone define total yield?
No, not when previous genuine harvests occurred. Cumulative physical crop yield includes all distinct harvested masses on the appropriate biological weight basis across the specified cycle. A final event can be large, but it is only part of the crop.
Similarly, saleable yield should combine all accepted product masses across events, with clear processing and deduction rules. Do not mix biological and paid product weight without qualification.
Frequently asked question: Can an estimate be reported to the nearest kilogram?
A scale may measure a landed batch to the nearest kilogram while the standing biomass estimate could have far greater uncertainty. Numerical presentation should reflect measurement quality. A projected population converted from a small sample should not inherit the precision of a directly weighed truck.
Keep estimates labelled, show meaningful ranges or sensitivity scenarios where justified, and avoid decimal places unsupported by the source.
Frequently asked question: What is the best compact harvest summary?
A defensible fictional summary could say that three harvests removed estimated animal counts of 20,000, 30,000 and 110,000 from a closed initial cohort of 200,000, giving approximately 160,000 animals recovered and 5,250 kilograms of cumulative whole-shrimp production. The 40,000-animal difference remains unassigned without a verified mortality and movement record.
That statement is longer than a headline promising eighty-percent survival, but it conveys what the numbers establish and which uncertainty remains. The purpose of translating an operational ledger is to preserve that distinction.
Original research sources and related reading
FAO cultured species profile: whiteleg shrimp (Penaeus vannamei)
FAO Shrimp Culture: Pond Design, Operation and Management — harvesting
FAO AGRIS: Assessing uncertainty in shrimp farm partial-harvesting programs
FAO AGRIS: Evaluating partial-harvesting strategies and profitability uncertainty
FAO AGRIS: Selective staggered shrimp harvesting
eduKateSG: shrimp cast-net sampling and average body weight
eduKateSG: fish size grading and market-ready harvest