A General Age-structured Model for Ecosystem Management

One important result from this project is derived from the pelagic multispecies model where we use non-linear programming to calculate the optimal catch composition for herring, mackerel and blue whiting. In this model, we analyze both technological interaction (different cost functions) and biological interaction between species. The same vessels harvest all three species, namely purse seiners. A quite robust result is that the mackerel fishery is the most profitable one and the one which contributes most to value creation. The policy implication of this is that the mackerel stock should be built up in order to accommodate significantly higher harvest quantities in the future. The results indicate room for significant catches of the other two species as well. The results are based on advanced methods for parameter estimation (ensembled Kalman filtering) and seem robust even if one takes into account that prices decrease with increased harvesting. The results from this subproject are published in Marine Resource Economics (2016). Another interesting result is related to an age-structured model where we analyze the profitability of so-called "recruitment enhancement." The results show that the value of such measures will be significantly reduced if they are not simultaneously combined with an optimal adjustment of catches and fishing mortality. This means that artificial recruitment through release of juveniles is not a sufficient remedy alone in order to increase profitability, but that it may work well in combination with other measures. These results are published in Marine Policy (2015). Results from the project have also been published in Environmental and Resource Economics and Natural Resource Modeling. In the former we use a continuous, non-linear bio-economic model to show how uncertainty in the biology affects optimal utilization of a resource when prices are stochastic. It is shown that harvesting should be conservative if the price is believed to be dragged back to the average (mean-reverting price). The simulations also show that in the long run exploitation ought to be lower than in the deterministic case. In the second we present a study based on a stochastic multispecies model. The model gives optimal harvest of cod and capelin in a three species ecosystem in the Barents Sea which also includes herring. It shows that the optimal catch in a stochastic model is more conservative than in an equivalent deterministic model. Furthermore, it was found that uncertainty has much greater effect on the optimal catch of the prey (capelin) than on the catch of the predator (cod). In 2015 and 2016 we have, among other things, published results on so-called "stochastically induced critical depensation". What this means, translated into plain English, is that if one introduces uncertainty in a bio-economic optimization model, the result may be that it is optimal harvest a stock, which has come below a certain critical level, down to extinction, even though this would never be optimal in an equivalent model without uncertainty. These results are published in two articles, one in Marine Resource Economics (2015) and another in Natural Resource Modeling (2016). We have also done work on an age-structured model which is continuous both in time and age of the fish. This has resulted in interesting theoretical results, and, in addition, we have analyzed numerically how fish patterns vary with a number of economic and biological parameters.

The background for this project is the general age-structured bioeconomic model that has been developed at SNF, which can include all kinds of technologies in fisheries from schooling (purse seine) fisheries to demersal (bottom trawl) fisheries where the fish is uniformly distributed in the water. We propose to develop this model further both for simulation and optimization purposes and in particular to extend it such that it can be applied as a multi-species model. The model has been presented both at ec onomic and biological meetings around the world and received a lot of attention. The theoretical model has already resulted in an article in AJAE. Researchers abroad have expressed interest for cooperation, and it would be a pity if the momentum of this w ork should be halted due to lack of funding. The generalized age-structured model is, to a much larger extent than existing models, able to incorporate density dependence in the stock. Traditional age-structured models can typically only handle demersal b ottom-trawl type fisheries and not schooling fisheries. The Beverton-Holt model comes out as a special case of our model. The new approach makes it possible to analyze not only the extreme cases of pelagic schooling species on one hand and demersal specie s at the other, but also the more realistic cases between these two extremes. It is also possible to do optimization and not only simulation. Further, optimization can be done on fairly large-scale real-world fish stocks and for quite long time horizons. This makes the model useful for solving practical management problems and yield realistic policy implications. In other words, it is not only a theoretical model but very much a tool for practical management. The main contribution will be detailed advic e on catch and effort allocation in order to increase value added in society along with analysis of consequences for employment and settlement etc. The advice will be based on ecosystem analysis.