| Popis: |
Single-species models have been commonly used to assess fish stocks in the past. Since these models have relatively simple data requirements, they sometimes provide the only tool available to assess the status of a stock when data are not enough to develop more complex models. However, these models have been criticized for several reasons since they provide reference points independently for each species assessed ignoring their interactions. For example, several studies suggest that even more substantial reductions in fishing mortality may be necessary to ensure MSY is reached when taking into consideration multiespecies interactions. Therefore, and as Pauly et al. (1998) stated, single-species analysis may mislead researchers and managers into neglecting the gear and trophic interactions which ultimately determine stocks long-term yields and ecosystem health. Ecosystem or multispecies models offer a number of advantages over single-species models. As stated in the workshop “Incorporating ecosystem considerations into stock assessments and management advice” (Mace, 2000) two general improvements are: a better appreciation of the fishing on ecosystem structure and function, and a better appreciation of the need to consider de value of marine ecosystems for functions other than harvesting fish. As disadvantages, multispecies models are statistically complex and include trophic relationships requiring more information (e.g. good estimations of biological parameters of each species and generally a full quantification of the diet sometimes available though the analysis of stomach contents). To reduce the number of species and therefore the amount of information needed, Minimum Realistic Models (MRMs) represent an intermediate level of complexity, where only the subset of the ecosystem, important for the issue under consideration, is modeled. This approach offers the advantage of allowing a refinement of our estimates and can help answer more targeted questions. Most multispecies models include interactions between commercially exploited species, since those data are more readily available. However, information is needed on at least both the main preys and predators of a selected stock. In the case of European Hake, the species we have focus our research on, cetaceans are their main predator, particularly common and bottlenose dolphins, which have been estimated to remove annually in the Atlantic shelf waters of the Iberian Peninsula, an amount similar to that caught by Spanish and Portuguese fleets (Santos et al., 2013). The European hake is one of the main fishing species of the Spanish and Portuguese fleets operating in the area, and one where more research activity has been concentrated, hence there is plenty of available biological information on growth, reproduction and trophic interactions. As a result, a population model has been built which uses trophic interactions to investigate the relationships between hake and other species. The European hake population is currently divided into two stocks, north and south. The southern hake stock, distributed along the Atlantic coast of the Iberian Peninsula, is annually assessed by the International Council for the Exploration of the Sea (ICES) and the Spanish Institute of Oceanography (IEO). For the assessment of this stock, “Gadget” a multi-specific modeling framework is used. Gadget allows the building of minimum realistic models that integrating the main trophic relationships among selected species considered to reflect the main processes in the system. Modeling cetacean populations can allow us to include complex trophic relationships in multispecies models. Furthermore, it will also be a tool to help cetaceans conservation by guiding possible management measures that ensure their viability or recovery. All cetacean species are protected by national and international legislation (e.g. Habitats Directive). However, modeling cetaceans dynamics has a number of problems due to lack of good information because, unlike the case of commercial fish stocks, usually there are no timeseries of abundance estimates available and the biological information available has been derived from the examination of stranded and/or bycatch individuals which could generate biased estimates. Common dolphin is the more abundant species of small cetacean in the Atlantic coast of the Iberian Peninsula and for which more biological information is available. For this reason, we have used this species as an example to explain our approach to obtain estimates of all the biological parameters needed to build a cetacean population model from strandings data. Southern European hake stock is distributed from Gulf of Biscay to Strait of Gibraltar, along Atlantic coast of the Iberian Peninsula, and thus the common dolphin population we will try to model is the one which lies in this area, although it is considered that there is a single common dolphin population is continuous throughout the European Atlantic coast (Natoli, 2008). Due to their protected status, no timeseries of catches neither fishing biological samples exist in the area, in contrast as in commercial fish species. Therefore, the major source of biological information is derived from strandings, mainly collected by experienced personnel from the “Coordinadora para o Estudio dos Mamíferos Mariños” (CEMMA) operating in Galicia (Northwest of Spain) since 1990. From these data we have derived the direct and indirect information required for our model. However, only the following parameters: maximum and minimum dolphin length and weight, sex-ratio and proportion of dead dolphins exhibiting bycatch signs can be used without further data processing. From these parameters, the length-weight relationship can also be easily calculated for both males and females, in our case, by fitting a logarithmic regression. Nonetheless, to fit a growth curve is necessary to know the age of each stranded dolphin and therefore, ageing the teeth (and sampling them from the carcasses) is required. In odontocetes teeth are commonly used for age estimations since dentine is laid down annually. Growth layer groups can be counted in tooth sections that have been stained and an estimate of dolphin age can thus be obtained. Once we have both length and age for each dolphin we can fit a Laird-Gompertz curve to model common dolphin growth. With length and age data we can build a maturation ogive estimating the L50 and the A50 (length and age at which half of the dolphins have already achieved maturity, respectively). For this purpose, we also require that a necropsy has been performed, so that gonads have been sampled and analyzed to estimate the maturity status. To define the trophic relationships in the model, we need to know the amount and proportion of each prey species in the dolphin diet. This information can be derived from the analysis of stomach contents that can back calculate the original prey size and weight by measuring the prey hard structures (e.g. fish otoliths, bones, cephalopod mandibles) found in the stomachs. From these data we can obtain estimates of daily dolphin consumption of prey by applying different energy requirement levels as Santos et al. (2013) did. All this basic biological parameters are relatively easy to calculate providing samples are collected and necropsies performed on stranded/bycaught individuals. However, other crucial parameters for the model such as mortality-at-age are harder to estimate from this source of information. We can assume that the proportion of observed ages in the strandings is representative of the absolute number of dolphins dead at each age. If we build a life table with the total number of stranded dolphins at each age we have a theoretical population structure which can be used to derive the proportion of mortality-at-age. Thus, we have an estimate of the total mortality experienced by each age class of the population. We are also interested in the natural mortality value and the approach we have followed to calculate this parameter is to use the evidence of bycatch in the carcasses, to subtract bycatch mortality from the total mortality for each age class. The value obtained is the “natural” mortality. Natural mortality is assumed to be the same for the whole study area; however fishing mortality is expected to be different because fishing effort is not equal in the whole study area. By relating the bycatch mortality obtained in Galicia with the fishing effort in the area, a relationship can be derived that allows us to obtain potential bycatch mortality for other areas where we only know the fishing effort. The last parameters that we have to estimate to finish building the dolphin population model are the population abundance and its trends. The best available abundance estimates for common dolphin and other cetaceans have been obtained from a sighting survey carried out in July 2005 covering the whole EU Atlantic continental shelf (Hammond et al., 2013). These estimates provide the only absolute estimate of common dolphin abundance in the shelf waters of the Iberian Peninsula. However, because the survey covered the study are only once, there are no information on trends in population size. To address this shortage, our approach has been to use again the time series of strandings. This time we model the influence of oceanographic variables such as rate of upwelling and intensity and direction of prevalent wind, as well as changes in fishing pressure, on numbers of strandings to disentangle the effects on stranding variability of oceanography, fishing bycatch and potential trends in population abundance. The uncertainty present in our estimates of some of these parameters arises from the lack of data. Thus the results should be treated with caution and keep working to obtain more biological data as well as further refinement to the data analysis needed to estimate these parameters from the existing data. |
