Q: Isn’t all of the density dependence (DD) not already covered in the S-R model?

A: No, currently only the DD happening before the age of recruitment is accounted for. We attempt to count for and implement specifically the post recruitment DD with our PROST calculations. Also, we will apply Surplus Production Models because they also count for the post recruitment dd, but they just do not disentangle this from the pre recruitment dd.


Q: When extracting information from ecosystem and multispecies models what model runs do you take? The biomass of other stocks often vary widely from run to run, and some could be predators taking their share of production.

A: When extracting information from ecosystem and multispecies models we aim for model runs where other stocks are as close to current stocks sizes as possible. That is why we say our new set of Fmsy values should be revisited every about 5 years. A good example to consider is the Baltic ecosystem where the current low cod stock means that the Fmsy for sprat is 0.40. However, if as is hoped the cod stock is build up to close to previous sizes of about 0.5 million t, they will predate  heavy on sprat and the Fmsy of sprat would have to be reduced to maybe only 0.1.


Q: If management adopt the new Fmsy values they will have to develop new management plans before they can be implemented?

A: No, they can apply ICES default HCR (Harvest Control Rule) with the new Fmsy value and that is a management plan, and by default accepted by ICES.


Q: Don’t you base your modelling on results from other models instead of on basic data?

A: Yes, we do. But this is not uncommon. Stock-Recruitment models for instance are almost always based on results from assessment models. It is a question of being practical and transparent. If we based our models on the basic data they would be unpractically large models, only transparent to very few people. We however, are testing this by a few cases where we do based our models on basic data and these confirm that in our case it is not a big issue to base our modelling on results from other models.


Q: Are you criticizing the current management for having wrong Fmsy values?

A: No, but we are helping improve the scientific advice. Of course one can always regard work intended for improving things as a critic of the existing, but this is the way progress should work. Furthermore, ICES is only a network of scientists and we are part of that network. We are trying to bridge the well-known gap between basic science and scientific advice. We have two innovative ideas for the difficult task of extracting the enormous knowledge on ecosystem functioning developed over the past about 4 decades and make it useful in fisheries management. One is using a model type called Surplus Production Models (SPM) on time series of catches and stocks biomass, and the other is using meta-analysis to spread knowledge from a few fish stocks to the rest of the stocks.


Q: If the new Fmsy values are adopted – how shall managers and scientists then make MSEs (Management Strategy Evaluations)?

A: If the new Fmsy values are adopted, the current approaches cannot be used, because they ignore density dependence in three out of the four relevant population dynamic parameters: natural mortality, growth, and maturity. Luckily, there are several other approaches. One is to use Surplus Production Models with the parameters from the present project. Another is to amend the current approaches (which are age and cohort based) with plausible models of density dependence in natural mortality, growth, and maturity. If there is no cannibalism the natural mortality parameter can probably be ignored.


Q: The current Fmsy include a precautionary element . The Fmsy are capped down to the F values that gives a less than 5% risk to get below Blim in a given year in the steady state situation. If the new Fmsy values are also capped they might end up not being different from the current Fmsy values?

A: There are two important aspects of this question. The first one is that the current approach in calculations of the 5% risk needs to be amended because it ignores density dependence in three out of the four relevant population dynamic parameters: natural mortality, growth, and maturity. The second one is that it is not a recommendable approach to include (some would say “hide”) precautionarity in the quite technical concept of Fmsy. It is fair enough that management would not like a higher than 5% risk to get below Blim, but that aim can easily be met by capping the TAC in a given year to the value which gives a less than 5% risk to get below Blim.


Q: Isn’t there a model approach to estimate Fmsy where it is not true that: “It is a mathematical fact that missing any of the four important density dependent factors: R/SSB, growth, maturity and mortality will give an underestimate of Fmsy“?

A: No, not any we have come across in the peer reviewed literature.


Q: The MYFISH and MAREFRAME projects – haven’t they not already dealt with the issue in your project?

A: No, these projects have mainly dealt with the longer term task of inclusion of ecological, economic, social, and institutional considerations when setting targets and limits for multispecies fisheries. MSY is just one aspect of this. We just aim at making the Fmsy values as close to the true values as possible. This will however also benefit the issues dealt with in MYFISH and MAREFRAME.


Q: We often hear that one cannot get MSY simultaneously on all stocks. Is that the same with the new Fmsy values?

A: Not to the same extent. The problem is that when fishing is reduced the stocks build up in size and interact with each other by competing for food and by predation. With the new Fmsy values which are generally about 50% higher than the current Fmsy values, the stocks will not build up to the same extent and thus the interactions will be smaller.


Q: You say it is not a full multi-species approach and that the new Fmsy values should be revisited in about 5 years time. Why is that?

A: This is because we still use the single stock concept in order for our results to be useful right away. A full multispecies management will need long negotiations about the priorities between stocks. That will likely take 5 years or even more. Our new Fmsy takes into account much more ecosystem functioning than the current Fmsy, because we include not only 1 (recruitment) out of the 4 important density dependent factors. We include all four. The other three are: individual fish growth, maturity and cannibalism. We assume implicitly that the current ecosystem regarding the other fish stocks are roughly constant, when doing our simulations 50 years or so into the future to obtain Fmsy in what is called the “steady state” stage. That is when things have stabilized after many years of fishing with the same fishing pressure. In about 5 years’ time for instance a cod stock might have increased a lot and for a forage species like sprat that might mean that Fmsy has to be decreased because cod is eating a larger proportion of the sprat stock, and new estimations of Fmsy will be needed.


Q: How is discards treated in your surplus production models?

A: We generally use total catch, i.e. landings plus discards because this is actually what is removed from the stock by fishing. This means that the Fmsy is then corresponding to maximum landings plus discards, which of course is not the main goal, which is maximizing the landings. We have tested this for a handful of stocks with high discard rates by use of the PROST models as well as by running separate SPMs with catch and landings, and it gives almost the same Fmsy values independent of whether maximizing the catch or the landings.


Q: The current Fmsy estimates are often updated with new data on weight-at-age etc. Isn’t that good enough to take account of ecosystem effects like density dependence?

A: No, because  to estimate Fmsy you need to make simulations that give stock sizes that greatly differ from the current size. Ignoring density dependence in these simulations gives a biased result. We correct for this. However, we still say the values are only valid in the short term and this is because the stocks size of other fish species might vary substantially in the longer term and thereby change the basic interactions between stocks. So, our analysis should be redone every say 5 years.


Q: ICES states (2016 ACOM): “BMSY is a notional value around which stock size fluctuates when fishing at FMSY. BMSY strongly depends on the interactions between the fish stock and the environment it lives in, including biological interactions between different species. Historical stock size trends may not be informative about BMSY (e.g., when F has exceeded FMSY for many years or when current ecosystem conditions and spatial stock structure are, or could be, substantially different from those in the past).” …Does Fmsy not depend on the same interactions?

A: Yes, Fmsy depends on the same interactions. ICES implicitly indicates that it does not matter so much for Fmsy as for Bmsy and therefore it is ok to use Fmsy, but not Bmsy. However, our analysis shows that often Fmsy varies as mush if not more than Bmsy due to density dependent factors, so we have the same problems with Fmsy as with Bmsy.


Q: Aren’t Surplus Production Models based on the assumption of steady state or equilibrium? This is rarely the case for real fish stocks and fisheries.

A: No, Surplus Production Models (SPMs) are only based on the assumption that the productivity of the stock is in steady state, not the fishing pressure (F) and the stock size in a given year. F can vary as much as it likes from year to year and the stock size will react to that, but not be in an equilibrium with the F (meaning at the size it will have after many years of fishing with the same F), and the SPMs still works fine, also from a theoretical point of view. However, some methods to estimate the parameters in SPMs are based on steady state, but not the methods we are using.

Q: How do you deal with regime shifts?

A: Regime shifts are many things and there are many definitions.  It is a research field in progress. For fish stocks it needs to be factors that influence the population dynamic of the stock. Already now part of regime shifts are included in the stock recruitment relationship - when the stock is large the environment for the juveniles is poor (and therefore fewer recruits are produced per spawner). When considering density dependence also in growth, maturity and mortality even more parts of regime shifts are included. Thus, for the assessment work regime shifts are mainly relevant when it challenge the population dynamic model used. This could by an outbreak of a disease like Ichthyofonus spp. in herring stocks in the 1990s. For most herring stocks in the Northeast Atlantic the poor recruitment in recent years are probably more likely than not, linked to the predation and food competition with the large mackerel stock and thus part of at least multispecies models. Another point is the possibilities to identify and estimate effects of regime shifts. The time series for each stock is limited, often less than 50 observations (data years) and with a lot of process and observation errors. This makes it difficult the extract signal from noise in the data when estimating Fmsy. Thus, shortening the time series even further by inserting regime shifts increases these estimation problems. Given that identification of regime shifts often are quite uncertain and type I errors are likely (identify a regime shift when there is none),  we think it is most appropriate for the time being to only consider regime shifts if it is very obvious, like for herring when we had the Ichthyofonus spp outbreak. It should also be noted that the present Harvest Control Rules used in management makes the risk to the stock quite low because fishing pressure is automatically reduced at poor regime periods. The only real problem for the scientific advice is when the regime is good because science will then be biased towards too low F.