A rejoinder to comments on “Catch reconstructions reveal that global marine fisheries catches are higher than reported and declining”

Daniel Pauly and Dirk Zeller
Sea Around Us
University of British Columbia

On January 19, 2016 the journal Nature Communications published our paper titled “Catch reconstructions reveal that global marine fisheries catches are higher than reported and declining” (Pauly and Zeller 2016) which was widely covered by the press, and triggered a very large number of positive comments.

Of particular interest to us was the comments by Sidney Holt, one of the founders of quantitative fish population studies, who wrote us on February 8, 2016: “I have just read your paper with [Dirk] Zeller on sizes of catches. It is excellent and very important. It has reminded me of arguments I had when in FAO in the 1950s, helping [L.P.D.] Gertenbach organise the Fisheries Yearbooks. Our statistical unit insisted that we had to accept what governments told us, but some of us – my Division – said we had, in addition, to add what we knew was unreported. Budget constraints prevented us from doing more”.

This clear endorsement of the basic principle of our catch reconstruction work summarized in Pauly and Zeller (2016) is very encouraging, particularly as it was sent by someone we respect, and who was there when the format of the FAO fisheries statistics was determined. There were many other positive comments, and we thank their authors.

There were a few negative comments, however, such as those posted on www.cfooduw.org on and after January 22, 2016. We believe scientific discourse is normally best carried out in a more collegial setting, such as the sphere of scientific publications, but we are posting the following responses online to clear up any confusion these comments may have created.

Comments by Michel J. Kaiser

Michel J. Kaiser, of Bangor University, U.K., writes: “Catch and stock status are two distinct measurement tools for evaluating a fishery, and suggesting inconsistent catch data is a definitive gauge of fishery health is an unreasonable indictment of the stock assessment process”.

Our publication makes no such suggestion.

Michel Kaiser further adds “Pauly and Zeller surmise that declining catches since 1996 could be a sign of fishery collapse. While they do acknowledge management changes as another possible factor, the context is misleading and important management efforts are not represented. The moratorium on cod landings is a good example – zero cod landings in the Northwest Atlantic does not mean there are zero cod in the water. Such distinctions are not apparent in the analysis”.

This sounds very reasonable, until you recall that the moratorium on industrial fishing of Northern cod was declared in 1992, in the Canadian province of Newfoundland and Labrador, because Northern cod stock size had consistently been overestimated by Canada’s stock assessment experts, and finally collapsed, such that it could not support any more fishing. The story is well documented (e.g., Walters and Maguire 1996). Any suggestion that the decline of cod catches (cod catches never dropped to zero) was unrelated to the decline in the underlying biomass is incorrect.

Furthermore, Michel Kaiser writes: “Another key consideration missing from this paper is varying management capacity. European fisheries are managed more effectively and provide more complete data than Indian Ocean fisheries, for example. A study that aggregates global landings data is suspect because indeed landings data from loosely managed fisheries are suspect.”

This is true of any study that attempts to compile global fisheries data, including FAO’s definitive and widely used ‘State of the World’s Fisheries and Aquaculture’ reports. That does not mean we or FAO should not undertake such an exercise, but rather that we should do so with the appropriate caveats, which we have.

Finally, Michel Kaiser writes that “the author’s estimated catch seems to mirror that of the official FAO catch data, ironically proving its legitimacy. ‘Official’ FAO data is not considered to be completely accurate, but rather a proportionate depiction of global trends. Pauly’s trend line is almost identical, just shifted up the y axis, and thus fails to significantly alter our perception of global fisheries”.

That the trend line presented by Pauly and Zeller (not by ‘Pauly’ alone) would roughly reflect the reported data is not surprising: after all, we complemented FAO’s and other reported landings data with fisheries that were omitted, which are roughly similar between countries and periods. Thus, we certainly did not challenge the “legitimacy” of official data, but rather their ‘accuracy’ (i.e., completeness). This incompleteness is incidentally the very point made by Sidney Holt. Also, to state that the official and reconstruction catch trend lines are “almost identical, just shifted up the y axis” overlooks the strong decline of reconstructed catches from 1996 (which contrast from officially reported statistics), which is a strong point of contention for other critics (see below).

Comments by David Agnew

David Agnew, Director of Standards at the Marine Stewardship Council, notes that “[t]he analysis of such a massive amount of data is a monumental task, and I suspect that the broad conclusions are correct”. We thank David Agnew for this supportive statement.

However, as is usual with these sorts of analyses, when one gets to a level of detail where the actual assumptions can be examined, in an area in which one is knowledgeable, it is difficult to follow all the arguments. The Antarctic catches ‘reconstruction’ apparently is based on [Palomares and Pauly 2015a] and a paper on fishing down ecosystems (Polar Record; Ainley and Pauly 2014). The only ‘reconstruction’ appears to be the addition of IUU and discard data, all of which are scrupulously reported by CCAMLR anyway, so they are not unknown”.

The reconstructions for Antarctic and Sub-Antarctic waters did not consist only of “the addition of IUU and discard data”. Among other things, these reconstructions (Ainley and Pauly 2014 for the Antarctic Continent; Boonzaier et al. 2012 for Prince Edward Island; Kleisner et al. 2015 for Heard and McDonald Islands; Padilla et al. 2015 for Bouvet Island; Palomares and Pauly 2011 for Kerguelen Island, Palomares and Pauly 2015b for the Falklands, South Georgia, South Sandwich and South Orkney Islands; Pruvost et al. 2015a for Crozet Island; and Pruvost et al. 2015b for St Paul and Amsterdam Island) included allocating catches reported by fishing ‘seasons’ (often spanning from October of one year to March of the next) to calendar years (to make them compatible with data sets from other areas), assigning ‘former-USSR’ catches to the likely Soviet Republics (Russia, Ukraine, etc.), adding subsistence and artisanal catches to the Falkland Islands catches, and more.

We agree that CCAMLR does a good job, as made evident in Figure 3 of our paper (Pauly and Zeller 2016), where the difference between official and reconstructed Antarctic catches is relatively small. Why this should be an argument against our paper eludes us. Finally, yes CCAMLR does include data on discards in their statistics (and ought to be congratulated for doing so). However, discards are not readily and transparently identifiable (i.e., labeled) as such, by fishing country and species in the data sets that can be accessed by the general public.

Then, David Agnew has a problem with 100,000 t of krill somewhere in Figure 3 of our contribution, writing that our data have not “been substantiated, nor referenced in the supporting documentation that I have seen (although I could not access the polar record paper)”. We do not know why he could not access the Polar Record paper (Ainley and Pauly 2014), which is available through Daniel Pauly’s publications page. Alternatively, it could have been requested from the authors, as is standard practice in science. It should not require stating that the inability of one researcher to access a paper is not sufficient basis for labeling the data therein as “not substantiated.”

Finally, David Agnew writes “It is noteworthy that the peak of the industrial catches – in the late 1990s/early 2000s – coincidentally aligns with the start of the recovery of many well managed stocks. This point of recovery has been documented previously (Costello et al. 2012; Rosenberg et al. 2006; Gutierrez et al. 2012) and particularly relates to the recovery of large numbers of stocks in the north Pacific, the north Atlantic and around Australia and New Zealand, and mostly to stocks that are assessed by analytical models. For stocks that need to begin recovery plans to achieve sustainability, this most often entails an overall reduction in fishing effort, which would be reflected in the reductions in catches seen here.”

In our paper we accounted for major management intervention by presenting, in addition to the world’s marine catch the catch that is obtained after removal of the US, Canada, the EU countries, Australia and New Zealand (see Figure 2 in Pauly and Zeller 2016), presumed to be largely regulated by quotas. This remaining catch (which is much higher than the amounts that were subtracted) declines rapidly in the last two decades as well.

This is not surprising in view of the generalized overfishing in most of the world, documented in one of the references cited by David Agnew in his comment (i.e., Costello et al. 2012). As for the low quotas, e.g., in the US, or the EU, we can only reiterate the point made above in conjunction with Northern cod: that low quotas (or fisheries closures) have been and continue to be imposed by fisheries managers only when a stock’s biomass has been unduly reduced by previous overfishing. Thus, long term catch declines are indicators of current or past overfishing (depending on the management regime), but not of healthy stocks (although continuing low catches they may indicate a stock rebuilding scenario being implemented by management).

Comments by Martin Pastoors

Martin Pastoors is Chief Science Officer, Pelagic Freezer-Trawler Association (PFA) in the Netherlands, and he comments on Dutch fishery operations:

  1. Pastoors’s first comment deals with the unreported/unallocated catches in the ICES Working Group reports, of which he writes that we “have interpreted this unallocated catch as unreported catch (whereas in fact they often consist of differences in calculation methods from dead weight to live weight). These unallocated catches were first split into unallocated catches by county (proportional to the size of the reported catch) and then extrapolated for the years where unallocated catches were not available (basically 1950-1990). So this is in fact inflating a somewhat random number to the largest part of the time series.”

Unreported/unallocated catches are, in ICES Working Group reports, the catches that the stock assessment experts that are members of Working Groups for various exploited species add to the officially reported catches of those species, based on their local knowledge of the illegal and otherwise unreported catches of the species in question, prior to assessing them. Not including these catches in our reconstructions would have defeated the entire purpose of our project. We suspect “random numbers” is an exaggeration, since ICES in fact uses them in its stock assessments. In addition, we are aware of and have accounted for these and other dead to live weight conversions in our work.

  1. Pastoors, like David Agnew before, writes that we “did not explain how they attributed the Dutch catch to the Dutch EEZ. I think this is pointless anyway because a large part of the Dutch catch is taken outsize of the EEZ and even outside of the North Sea”.

The technical report that underpins the Dutch catch reconstruction (Gibson et al. 2015), freely available on the page for The Netherlands EEZ on our website, states that we do our data work within ICES statistical areas, and thus do not move catches between such areas. Hence, the catches that Dutch fisheries report (hopefully correctly) by the statistical areas that include the Dutch EEZ-equivalent waters (Areas IVb and IVc), remain within these areas. We are aware that EEZ boundaries (formally declared or not) do not generally play a substantial role in EU fisheries policy and management.

And yes, there are “some pretty normative statements about the fishery” in the catch reconstruction for The Netherlands (Gibson et al. 2015), notably that “they are powerful and destructive due to relying heavily on beam trawling”. We believe the evidence supports this statement (see also Froese et al. 2015 for comments on North Sea fisheries). Dutch beam trawling is particularly damaging in this respect, and has wrecked the bottom communities of trawled areas, including the southern North Sea (e.g., Hiddink et al. 2006).

Also M. Pastoors writes that: “[t]he data that has been compiled is not presented in the report, nor are the conversion factors used. It is impossible to reconstruct what data they have taken from where and what ‘data’ is based on extrapolations and what is based on time series from other reports”.

Space did not allow us to give country-by-country methodological details (for 273 EEZs) in a summary paper such as the one discussed here (Pauly and Zeller 2016). However, as we point out repeatedly in our paper (p. 7, plus a dedicated section titled ‘Documentation of catch reconstructions’ on p. 8, as well as the detailed Supplementary Table S5, which gives the web link for each country’s catch reconstruction report or paper), all underlying catch reconstruction reports and data are publicly available without restriction, from each EEZ page on our website. We regularly receive queries from interested colleagues who have specific questions about one or more specific studies, which we are more than happy to answer in detail.

Comments by Karl-Michael Werner

Karl-Michael Werner, who is a PhD candidate at the University of Bergen, Norway, suggests that our “conclusion that global catches were previously underestimated is not surprising given the implicit understanding that FAO catch estimations have always omitted factors like discards and recreational fishing.”.

The omission of discards by FAO is not based on an “implicit understanding”, but on an explicitly stated policy: FAO treats their database as dealing only with ‘production’, i.e., FAO frames its contents in terms of commodities, rather than in an ecosystem context which would require discards to be included. Moreover, FAO is explicit in suggesting that recreational data should be included, and indeed Finland – which appears to be the only country doing so – has been doing so for many years (Zeller et al. 2011).

On the other hand, we agree with two other points by Karl-Michael Werner, namely that the decline of the Peruvian anchoveta catches and the decrease in global discards (see Zeller and Pauly 2005), which both contributed to the observed global decline, should have been “highlighted as a positive trend and not as part of reduced catches due to overfishing”.

Unfortunately, Karl-Michael Werner then repeated a demonstrably erroneous statement also made by others i.e., that we supposedly selected, when analyzing the time series of total reconstructed catch, the years that yielded the strongest decline. Thus, he writes “[i]t appears that the years 1996, which reported temporarily high catches, and 2010, with temporarily low catches, were arbitrarily chosen in order to produce a specific trend line. 1996 and 2010 represent a local high and a local low, which do not allow for a serious comparison of global catch. These fluctuations could be due to environmental changes and there is little evidence linking them directly to overfishing”.

In reality, and as described in the Methods section of our paper under Analysis (p. 8), we used an established statistical method called segmented regression, which fits a limited number of straight segments to time series and identifies the breakpoints between them. Thus, the year 1996 was never an input that we could choose, but rather an output of the analytical method used (see also Table S2 in the Supplementary Material of Pauly and Zeller 2016).

As for the year 2010, it was not picked as end point because it was a “local low”, but because we decided, more than a decade ago, to shoot for 2010 as preliminary end year for all reconstructions (we are now updating them to 2013). We would argue that careful reading of a paper and its supplementary materials should precede any scientific critique.

Karl-Michael Werner also writes that “[c]atches were higher than expected, which could suggest fisheries are more productive than expected. According to Pauly & Zeller, catches were high because stocks were overfished, but then catches went down due to overfishing as well. This is paradox; how are catch curves supposed to look like in order to be not considered as overfished then?”

Overfishing usually leads to very high catches for a few years, followed by low catches, with the earlier high catches being the cause of the eventual low catches. Indeed, this is what has been and is happening throughout much of the industrialized world, and is one of the reasons why a huge fraction of the fish consumed, e.g., in the EU, comes from Africa.

As for how catch curves are supposed to look: ideally, a fishery should be managed such that the quotas reflect environmental fluctuations, with lower quotas when the state of the environment can be expected to lead to reduced recruitment and vice-versa. Hence catches, ideally, should fluctuate around some mean, without upward or downward trend (assuming climate change does not affect recruitment, growth and survival). When applied globally and for a long time (after stocks have been rebuilt), this should lead to relatively high biomass for most stocks, and a stable global total catch. This is also known as ‘sustainability’, and is or should be the goal of fisheries management agencies throughout the world. Readers may wish to consider Froese et al. (2016) for a clear description as to where fisheries management should be heading.

Final remarks

What is interesting (or distressing) about this exchange, is that whatever you do that hints at problems with the manner fisheries are pursued; you can expect a number of individuals — often the same people — to object. The accusation that we cherry-picked breakpoints to exaggerate downward catch trends, or that we conjured large catches of some species to increase the differences between reported and reconstructed catch trends, is not based on any evidence (see above).

These points and our clarifications also suggest that careful reading of all material associated with a scientific paper should be the norm before considering a critique. A reading of the publicly available material associated with our paper would clearly reveal that we recommended to our catch reconstruction partners a policy of opting for the more conservative solution in cases where the available data offered a choice, which they did implement in their reports.

Readers of the available material would also have noted that the countries that clearly manufacture fisheries statistics (e.g., Myanmar) have ever increasing reported catches, which we could only partly correct (i.e., we consistently applied a conservative approach). This is the reason why, in our Figure 3, FAO area 57 (Eastern Indian Ocean) has exponentially growing catches, which even FAO doubts, as pointed out in our contribution. Yet, we did not present the even steeper decline of the world catch that omitting this FAO area (and the similarly suspect FAO area 71, Western Central Pacific would have generated, although it would have been legitimate given the very high uncertainty associated with these suspect catches, and would have strengthened our case.

These are significant points that the critics overlooked. Better data are needed to address the pressures facing fisheries around the world, and our paper has sought to address one aspect of that need – namely more comprehensive catch data that address the inherent bias in official, reported data. Such data are not only important on a global scale, but are especially important in developing countries, where local fisheries are crucial for food security.


We take this opportunity to thank all our colleagues and friends throughout the world, who freely contributed their time and expertise to describe the fisheries and their catches in their countries.


Ainley, D. and D. Pauly. 2014. Fishing down the food web of the Antarctic continental shelf and slope. Polar Record 50(1): 92-107.

Boonzaier, L., S. Harper, D. Zeller and D. Pauly. 2012. A brief history of fishing in the Prince Edward Islands, South Africa, 1950-2010. p. 95-101 In: S. Harper, K. Zylich, L. Boonzaier, F. Le Manach D. Pauly and D. Zeller (eds.) Fisheries catch reconstructions: Islands, Part III. Fisheries Centre Research Report 20(5). University of British Columbia, Vancouver, Canada.

Costello, C., Ovando, D., Hilborn, R., Gaines, S.D., Deschenes, O. and Lester, S.E. 2012. Status and solutions for the world’s unassessed fisheries. Science 338(6106): 517-520.

FAO. 2014. The State of World Fisheries and Aquaculture (SOFIA). Food and Agriculture Organization, Rome, Italy. 223 p.

Froese, R., Walters, C., Pauly, D., Winker, H., Weyl, O.L.F., Demirel, N., Tsikliras, A.C. and Holt, S.J. 2015. A critique of the balanced harvesting approach to fishing. ICES Journal of Marine Science DOI: 10.1093/icesjms/fsv122.

Froese, R., Winker, H., Gascuel, D., Sumaila, U.R. and Pauly, D. 2016. Minimizing the impact of fishing. Fish and Fisheries DOI: 10.1111/faf.12146

Gibson, D., Zylich, K. and Zeller, D. 2015. Preliminary reconstruction of total marine fisheries catches for the Netherlands in the North Sea (1950-2010). Fisheries Centre Working Paper #2015-46, University of British Columbia, Vancouver, Canada. 15 p.

Gutiérrez N.L., Valencia S.R., Branch T.A, Agnew D.J., Baum J.K., Bianchi P.L., Cornejo-Donoso J., Costello C., Defeo O., Essington T.E., Hilborn R. 2012. Eco-label conveys reliable information on fish stock health to seafood consumers. PLoS One, 7(8):e43765.

Hall, S.J. 1999. The effects of fisheries on ecosystems and communities Blackwell Publishing, Oxford. 274 p.

Hiddink, J.G., Jennings, S., Kaiser, M.J., Queirós, A.M., Duplisea, D.E. and Piet, G.J. 2006. Cumulative impacts of seabed trawl disturbance on benthic biomass, production, and species richness in different habitats. Canadian Journal of Fisheries and Aquatic Sciences 63(4):721-736.

Kleisner, K.M., Brennan, C., Garland, A., Lingard, S., Tracey, S., Sahlqvist, P., Tsolos, A., Pauly. D. and Zeller, D. 2015. Australia: Reconstructing estimates of total fisheries removals 1950-2010. Fisheries Centre Working Paper #2015-02, University of British Columbia, Vancouver, 26 p.

Padilla, A., Zeller, D. and Pauly, D. 2015. The fish and fisheries of Bouvet Island, pp. 21-30 In: M.L.D. Palomares and D. Pauly (eds.), Marine Fisheries Catches of SubAntarctic Islands, 1950-2010. Fisheries Centre Research Report 23(1). University of British Columbia, Vancouver, Canada.

Palomares, M.L.D. and Pauly, D. 2011. A brief history of fishing in the Kerguelen Island, France p. 15-20 In: S. Harper and D. Zeller (eds.) Fisheries catch reconstruction: Islands, Part II. Fisheries Centre Research Reports 19(4). University of British Columbia, Vancouver, Canada.

Palomares, M.L.D and Pauly, D. (Editors). 2015a. Marine fisheries catches of SubAntarctic Islands, 1950-2010, Fisheries Centre Research Reports 23(1), 45 p. University of British Columbia, Vancouver, Canada.

Palomares, M.L.D. and Pauly, D. 2015ba. Fisheries of the Falkland Islands and the British Antarctic Islands, p. 1-20 In: M.L.D. Palomares and D. Pauly (eds.) Marine Fisheries Catches of SubAntarctic Islands, 1950-2010. Fisheries Centre Research Reports 23(1). University of British Columbia, Vancouver, Canada.

Pauly, D. and Zeller, D. 2016. Catch reconstructions reveal that global marine fisheries catches are higher than reported and declining. Nature Communications 7: 10244.

Pruvost, P., Duhamel, G., Gasco, N. and Palomares, M.L.D. 2015a. A short history of the fisheries of Crozet Islands, p. 31-37 In: M.L.D. Palomares and D. Pauly (eds.) Marine Fisheries Catches of SubAntarctic Islands, 1950-2010. Fisheries Centre Research Reports 23(1). University of British Columbia, Vancouver, Canada.

Pruvost, P., Duhamel, G., Le Manach, F. and Palomares, M.L.D. 2015b. La pêche aux îles Saint Paul et Amsterdam, p. 39-48 In: M.L.D. Palomares and D. Pauly (eds.) Marine Fisheries Catches of SubAntarctic Islands, 1950-201. Fisheries Centre Research Reports 23(1). University of British Columbia, Vancouver, Canada [with an extended English language abstract and figure captions].

Rosenberg, A.A., Swasey, J.H. and Bowman, M., 2006. Rebuilding US fisheries: progress and problems. Frontiers in Ecology and the Environment,4(6), pp.303-308.

Walters, C.J. and Maguire, J.J. 1996. Lessons for stock assessment from the northern cod collapse. Reviews in Fish Biology and Fisheries 6(2): 125-137.

Zeller, D. and Pauly, D. 2005. Good news, bad news: Global fisheries discards are declining, but so are total catches. Fish and Fisheries 6: 156-159.

Zeller D., Rossing P., Harper S., Persson L., Booth S. and Pauly D. 2011. The Baltic Sea: estimates of total fisheries removals 1950-2007. Fisheries Research 108: 356-363.


Please follow and like us: