Official statistics on numbers of fish caught are not available. Statistics on wild-caught and farmed fish published by the Food and Agriculture Organisation of the United Nations (FAO) are given only in tonnages. Nor, unfortunately, does the FAO publish mean weights of fish, which would enable numbers to be calculated from these tonnages.

As part of the project of writing the report Worse things happen at sea: the welfare of wild-caught fish (full report) the author attempted to estimate the number of fish caught using FAO fish capture tonnages and mean weights estimated from available fish size data.

Using estimated mean weights for fish species based on fish size data obtained from a range of sources (mainly internet), it was estimated that 0.97-2.7 trillion wild fish are caught globally each year. Recognising the limitations of the fish-size data available, it is concluded that the number of fish caught is of the order of a trillion. This estimate does not include unrecorded fish capture, such as fish caught illegally and those caught as bycatch and discarded.

The method used in this estimate is briefly outlined below. For full details see estimating the number of fish caught in global fishing each year. Results and supporting data are available from the study overview. The study is also described in chapter 19 of Worse things happen at sea: the welfare of wild-caught fish (full report) and the key results given in Appendix A.

Study to estimate numbers of fish caught

There are three main parts to this estimate:
(1) fish for which the FAO reports capture tonnages in single species categories, e.g. Atlantic mackerel (Scomber scombus), and for which a mean weight was estimated from available fish size data
(2) fish for which the FAO reports capture tonnages in multi-species categories, e.g. Anchovies, etc. nei (Engraulidae), and for which mean weights were estimated for the largest and smallest relevant species in each category
(3) fish for which the FAO reports capture tonnages in totally general categories, e.g. marine fishes nei, together with categories for which the species are given but for which a mean weight could not be estimated.

The first part comprises the single species categories for which it was possible to estimate a mean weight, and so estimate fish numbers. Where possible, estimated mean weights were obtained from average weight data but, where these were not available, various other types of data were used, including typical weights/lengths or weight/length ranges.

The second part comprises the multi-species categories for which fish size data were available. To estimate the mean weight for a multi-species category, the mean weight was estimated for the smallest and largest relevant species in the group and combined as a range. Relevant species were considered to be those in the species group that are both (according to fishbase.org) :

• fished commercially or for subsistence
• distributed in the region from which more than 20% of capture was taken.

The third part comprises the species categories for which no estimated mean weight was obtained. The numbers of fish represented by these tonnages were estimated by extrapolating mean weight data from species for which a mean weight had been estimated. Wherever possible, this was based on extrapolated mean weight data for the same taxonomic class of fish species.