Resumen de Multispecies population modelling of the common dolphin ("Delphinus delphis"), the bottlenose dolphin ("Tursiops truncatus") and the southern stock of European hake ("Merluccius merluccius"), in Atlantic waters of the Iberian Peninsula
Maritime policies demand the use of an ecosystem approach to manage human impacts on marine ecosystems. Single-species population models have been widely used for the assessment and management of fish stocks, due to their simplicity and ease of use. However, ecosystem or multi-species models offer a number of advantages over single-species models (e.g., Pauly et al. 1998; Mace 2000; Pikitch et al. 2004). In this context, Models of Intermediate Complexity for Ecosystem assessment (MICE) such as the Globally applicable Area Disaggregated General Ecosystem Toolbox (GADGET) have been considered the most appropriate tools to assess multispecies interactions both in the case of fish stocks and non-target species such as cetaceans since they are capable of producing outputs that can be used for tactical decision-making (Plagányi 2007; Plagányi et al. 2014).
This thesis develops a multi-species GADGET model on the Atlantic coast of the Iberian Peninsula (ICES subdivisions VIIIc and IXa) that includes the southern stock of European Hake (Merluccius merluccius), the fishery targeting hake and two species of cetaceans: common dolphin (Delphinus delphis) and bottlenose dolphin (Tursiops truncatus), which have been shown to be important predators of hake (Santos et al. 2007, 2013).
Since 2100, the southern stock of the European hake has been assessed annually by the International Council for the Exploration of the Sea (ICES) using a single-species GADGET model (ICES 2010b). The existing model, as used for the 2015 assessment of this stock (ICES 2015a), has been used as the basis to build a multi-species model, to which were added the two species of cetaceans already mentioned. The southern hake stock is caught in a mixed fishery by the Spanish and Portuguese fleets, which also present high levels of cetacean bycatch (e.g., López et al. 2003; Goetz et al. 2014b). To be able to determine both the impact of cetaceans on the hake population and the impact of the fisheries targeting hake on the cetacean populations, the status and dynamics of cetaceans populations under study need to be known with a level of detail that is not currently available for many marine mammals (Murphy et al. 2009; Lassalle et al. 2012). Therefore, an important part of the present study has been focused on gathering and analysing data on common and bottlenose dolphins in order to estimate the parameters necessary for population modelling, alongside some secondary objectives.
In the first chapter, the prey consumption of the four most common cetacean species along the Atlantic coast of the Iberian Peninsula (common dolphin, bottlenose dolphin, striped dolphin Stenella coeruleoalba and harbour porpoise Phocoena phocoena) was calculated. Estimates of the annual consumption of each predator were provided, while highlighting the uncertainties and biases inherent in the information presently available on energy requirements, diet and population size. Common dolphins and striped dolphins consumed mainly sardine (Sardina pilchardus), followed by several species of gadids (mainly blue whiting Micromesistius poutassou), European hake and scads (Trachurus sp.). Bottlenose dolphins and harbour porpoises consumed mainly gadids and European hake. Total amounts of fish taken by harbour porpoises and bottlenose dolphins were much lower than those estimated for common and striped dolphins, reflecting their low abundance in the area. However, when comparing the consumption of each prey species separately, the highest takes of hake correspond to the bottlenose dolphin, followed by the common dolphin.
Cetacean predation on sardine represents 2 - 8% of the natural mortality (M) estimated for the stock and used in the assessment models, indicating that cetaceans probably have little influence on sardine population dynamics. For the southern European hake stock, estimated average removal by cetaceans often exceeds the M estimated by the assessment models. This suggests that cetaceans could have a significant impact on the European hake population and on the model used for its assessment. The lack of good estimates of abundance and field metabolic rate for most cetacean species probably represents the most serious barrier to reliably quantifying the interactions of cetaceans with other species (such as hake) and their role in the ecosystem.
In the second chapter, trends in the abundance of the common dolphin population in the north and northwest shelf waters of the Iberian Peninsula were evaluated. The rationale for this analysis was the need to estimate abundance indices for population modelling and to address the requirements for the assessment of the status of marine mammal populations under the European Marine Strategy Framework Directive (MSFD). A 10-year time series of data collected from multidisciplinary oceanographic surveys in shelf waters of north and northwest Spain were used. This approach provides a valuable addition to decadal large-scale dedicated surveys, offering a shorter interval between surveys and hence offering the possibility to track abundance changes at a regional scale with a finer temporal resolution. Trends in the number of common dolphins present in the study area over the last 10 years showed an annual increase, which results in a positive evaluation of Good Environmental Status (GES), under the framework of the MSFD for the common dolphin in the study area using the abundance indicator and threshold adopted for this species group. Data obtained from dedicated dual-platform surveys were used to obtain absolute abundance estimates for calculating bycatch limits. Comparing the estimated safe bycatch limits with estimates reported in the literature, the percentage of bycaught common dolphins in the study area was higher than the maximum limit allowable for bycatch mortality of cetaceans, selected under the MSFD. The observed bycatch values should cause a decline in the population not observable in the trends of their abundance, this may be due to possible displacements or migratory flow of new individuals coming from outside the study area.
Trends in abundance of cetaceans calculated using abundance surveys could only be estimated for the common dolphin, the most abundant cetacean species in the study area. Therefore, in chapter three, a different approach was developed to estimate trends in abundance using the number of cetacean stranded in Galicia (NW Spain) from 2000 to 2013. The incidence of cetacean strandings is expected to depend on a combination of several factors, including the distribution and abundance of cetaceans, their prey, and causes of mortality (e.g., natural, fishery bycatch), as well as currents and winds which affect whether carcases reach the shore. In this chapter, the spatiotemporal patterns and trends in the numbers of strandings of common dolphin, bottlenose dolphin and harbour porpoises in Galicia were investigated. Meteorological, oceanographic, prey abundance and fishing-related variables were analysed to disentangle the relationships that may exist between them, cetacean abundance and cetacean mortality off the coast. Only bottlenose dolphin showed significant fluctuations in local abundance over the study period. There was no evidence of long-term trends for any of the species and their abundances were therefore considered to have been relatively stable during the study period. The apparent contradiction with the positive trends of common dolphin abundance observed in the previous chapter may be due to the difference in the temporal scale (greater in the case of strandings) or to the spatial scale (greater in the case of oceanic surveys). On the other hand, in the strandings data a positive trend in abundance was also observed in the last years of the series, although it was not significant.
In chapter four an in-depth analysis of cetacean mortality at age was performed. Mortality is one of the most relevant parameters for the study of population dynamics. One of main sources of information to calculate the mortality experienced by a marine mammal population derives from the observed age structure of stranded animals. We developed a new method that takes into account the difference between natural and fishing mortality. For this purpose, a new package of functions (strandCet) for use in the statistical software R was developed. This package includes all the functions necessary to apply a Heligman-Pollard model adapted for cetaceans to estimate the different mortalities, and includes the possibility of applying a Leslie matrix to calculate population demographic parameters. This methodology was applied to the time series of stranded common dolphins in Galicia. Total, natural, and bycatch mortalities were estimated. Considering natural mortality only, population projections indicated a slight annual growth whereas if the model also accounted for fishing mortality, projection showed a strong annual population decline. The results indicate that bycatch mortality may be unsustainable for this population, however, this contradicts the observed abundance trends, so a possible explanation could be that geographic movements and migratory flows could bring in new individuals that counterbalance the local decline.
In chapter five, two GADGET dynamic population models were constructed for the common and the bottlenose dolphin with information derived from the analyses carried out in previous chapters as well as data obtained from new analyses. Using these models, bycatch limit reference points for the common and the bottlenose dolphin populations in the continental shelf Atlantic waters of the Iberian Peninsula were calculated for the first time. The estimated bycatch limits for the common and bottlenose dolphin are similar to those previously estimated for the North Sea harbour porpoise (i.e., ≈ 1.4% of the population), however, given the uncertainty of some data and several internationally agreed criteria, the potential bycatch of these species should not exceed 1% of their population. In addition, these dynamic models developed in GADGET were developed to be used as add-on pieces to multi-species model with the purpose of applying an ecosystem approach for managing the impact of human activities, such as fisheries, in the marine environment.
In chapter six a multi-species model was developed using GADGET. The model includes the southern European hake stock, the common dolphin, the bottlenose dolphin and the fisheries that target hake. Other non-modelled species, such as sardines, were also included in the model acting as prey for the dolphin populations. The existing hake model, which covers the period between 1982 and 2014 and used for the 2015 assessment of the stock by ICES (ICES 2015a), was used as the base for the multi-species model. Components of the model were linked by their trophic relationships, using the length distribution and the percentage of each prey type obtained from the analyses of the stomach contents of both cetacean species. For determining the daily consumption of prey by dolphins, a combination of the available energy and consumption models was applied. The abundances of cetacean populations were assumed to be constant throughout the time period modelled because the developed works did not identify a clear trend in their abundance over the modelling period. The level of fishery bycatch needed to maintain stable cetacean abundance was estimated and this value was set as constant over the simulated period. The natural mortality (M) of hake was recalculated by fitting a set of multi-species models with different possible values for M. The best model was selected using a comparative approach, relating the longevity of hake to its natural mortality (see Hewitt & Hoenig 2005). Long- and short-term projections were run to re-estimate new reference points for the multi-species model. The mean hake biomass and the precautionary biomass (B_pa) of the multi-species model were higher, while the fishing mortality which delivers maximum sustainable yield (F_msy) was considerably smaller than the one currently used for the single-species model, and also the current fishing mortality (F_sq). A medium-term projection exercise was also carried out using different scenarios of fishing pressure reduction to reach F_msy at different times (2016, 2017, 2018, 2019 and 2020). The largest differences were in the recovery time needed to reach the current catch levels after reducing the current F to F_msy. With the multi-species model, the current catch levels recovered more slowly than when using the single-species model. This is a consequence of cetacean predation (not included in the single-species model), that increases when F decreases due to a reduced bycatch level that impacts positively on cetacean populations. The results presented in this last chapter highlight the potential negative consequences of using single-species models for fisheries management that do not take into account interactions with other ecosystem components. This single species approach, by allowing catch levels that are too high or recovery scenarios that are too optimistic, can lead to potentially adverse consequences such as depletion of the fished stocks and unsustainable bycatch mortality of cetaceans. It is important to consider the assumptions involved when developing the model as well as the uncertainty in some parameters. The main assumption is that populations are closed. The model assumes that there is no migratory flow between the southern and northern stocks of hake, or between dolphin stocks with animal in the rest of their distribution range. Moreover, the abundance of common and bottlenose dolphins has been considered constant over the modelled series and only varied when the model was projected under different scenarios. The abundance used was calculated only in shelf waters, assuming that dolphins inhabiting this area have complete access to the prey species distributed within this area, not deeper than 1000 m. The natural mortality of hake has been considered constant for every age class as in the original single-species model. Natural mortality of dolphins has been also constant for the study period although different for every age class. However, bycatch mortality was estimated as an equal proportion for every age class and sex. Uncertainty of input data as fecundity, growth, length-weight relationships, density-dependency, etc., should be analysed in depth if new information could be included in the future.
The present study demonstrated the possibilities and advantages of the use of multispecies models in the management of fisheries and protected species. The multi-species model improved the knowledge of the modelled system and the interactions between its components. In addition to improving the assessment of the southern European hake stock in some respects (i.e., providing a more realistic estimate of natural mortality and varying the total mortality according to the evolution of the populations of its predators), the work also demonstrated a mechanism to include non-target species (such as cetaceans) into assessments. Because cetacean bycatch mortality is effort-related, effort reduction may represent a management solution. This work provides the first multi-species model of cetaceans and fish stocks developed in Europe. Despite of this, there are still many facets of the model and inputs that can be improved in the future and it would be advisable to continue advancing in this field of research so that the use of multispecies models in the management of populations can be a reality in the near future.
