estimated harvested for food globally
Calculated projection using a 2020 modelled estimate, not observed harvest or deaths. Elapsed-time counts and rates illustrate the annual pace, not measurement precision.
Based on an annual estimate of approximately 440,000,000,000, derived from Rethink Priorities — farmed shrimp slaughter estimate.
Approximately 440 billion farmed shrimp and prawns are represented in reported 2020 aquaculture harvest production, according to Rethink Priorities' probabilistic model. This is the researchers' reported mean, not the midpoint of the approximately 300–620 billion source-reported 90% subjective interval. The counter projects that annual pace across the current calendar year, not actual current-year production or a live measurement feed.
Rethink Priorities' model estimates approximately 440 billion farmed decapod shrimp and prawns represented in reported 2020 aquaculture harvest production. Production mass is converted to individuals using species/group weight distributions. The researchers describe these animals as slaughtered/killed for food; individual harvest or death events were not directly counted.
It represents farmed decapod shrimp and prawns in selected 2020 FAO aquaculture harvest-production categories, including relevant freshwater prawns. Rethink Priorities reports an approximate 440-billion probabilistic-model mean. It is not HC's calculation, the midpoint of 300–620 billion, or proof that every shrimp species and every unreported farm is represented.
Source: Rethink Priorities — Shrimp: The animals most commonly used and killed for food production (2023)
FAO FishStatJ reports aquaculture production as mass, not 440 billion counted animals. The researchers selected shrimp/prawn taxa, converted tonnes to grams, and divided by estimated live weight per animal using Mood and Brooke inputs where available. Species/group weight distributions feed probabilistic individual estimates, which are summed into the reported mean and subjective interval. Unsupported categories use a broad modelled 6–40 gram weight input range.
Source: Rethink Priorities — methods and probabilistic model (2023)
No. Waldhorn and Autric describe the estimated harvested animals as slaughtered/killed for food, but the count is inferred from harvest-production mass. Harvest does not establish the timing of death or how many animals were actually eaten. Wild shrimp, premature farm mortality, transport mortality, broodstock mortality, and shrimp-trawl bycatch are not added to this metric.
The source calls 300–620 billion a 90% subjective confidence interval. Several model inputs were literature ranges of possible values rather than genuinely calibrated 90% subjective bounds. This is not conventional sampling confidence, validated 90% coverage, or a demonstrated 90% probability that the true total lies within the range. Important uncertainties include animal-weight distributions, species aggregation, reporting coverage, generic weight assumptions, researcher-selected production variation, and probability-distribution choices.
Source: Rethink Priorities — uncertainty and appendix (2023)
Baseline: Rethink Priorities — farmed shrimp slaughter estimate
Reference year: 2020
Daniela R. Waldhorn and Elisa Autric's Rethink Priorities research, published on its public webpage in 2023, uses 2020 FAO FishStatJ Global Aquaculture Production mass. Selected shrimp/prawn taxa and species/group live-weight distributions yield probabilistic individual estimates, summed into the researchers' reported approximate 440-billion mean and 300–620 billion subjective interval. The authors call these animals slaughtered/killed for food, but the model infers the count from harvest production, not observed death events. The source calls 300–620 billion a 90% subjective confidence interval. Several model inputs were literature ranges of possible values rather than genuinely calibrated 90% subjective bounds. This is not conventional sampling confidence, validated 90% coverage, or a demonstrated 90% probability that the true total lies within the range. HC evenly distributes the reported mean across the current calendar year. Its per-second, daily, and average-month rates are illustrative arithmetic, not evidentiary precision or real-time observations.
Individual-count estimate: Rethink Priorities — Waldhorn & Autric (2020 observation; published 2023)
Daniela R. Waldhorn and Elisa Autric's original probabilistic analysis reports the approximate 440-billion mean, not an arithmetic midpoint, and 300–620 billion 90% subjective confidence interval. Exact copying of this reported value does not imply exact real-world precision.
Underlying mass data: FAO FishStatJ — Global Aquaculture Production (2020)
Authoritative aquaculture-production statistics used as mass inputs, not as a direct individual-animal count.
Live-weight inputs: Mood & Brooke — live-weight estimates cited by Rethink Priorities (2019)
Species or category live-weight estimates used where available in the mass-to-individual conversion; the Rethink Priorities model uses a broad 6–40 gram range for unsupported categories.
Uncertainty and harvest inference: Rethink Priorities — methods and appendix (2023)
The model derives estimated individuals from harvest mass, interpreted as slaughter for food. Literature ranges are not uniformly calibrated 90% subjective inputs; the resulting interval has no validated 90% coverage. Premature mortality remains separate.