Search

Many ecological studies rely on count data and involve manual counting of objects of interest, which is time-consuming and especially disadvantageous when time in the field or lab is limited. However, an increasing number of works uses digital imagery, which opens opportunities to automatise counting tasks. In this study, we use machine learning to automate counting objects of interest without the need to label individual objects. By leveraging already existing image-level annotations, this approach can also give value to historical data that were collected and annotated over longer time series (typical for many ecological studies), without the aim of deep learning applications. We demonstrate deep learning regression on two fundamentally different counting tasks: (i) daily growth rings from microscopic images of fish otolith (i.e., hearing stone) and (ii) hauled out seals from highly variable aerial imagery. In the otolith images, our deep learning-based regressor yields an RMSE of 3.40 day-rings and an $$R^2$$ R 2 of 0.92. Initial performance in the seal images is lower (RMSE of 23.46 seals and $$R^2$$ R 2 of 0.72), which can be attributed to a lack of images with a high number of seals in the initial training set, compared to the test set. We then show how to improve performance substantially (RMSE of 19.03 seals and $$R^2$$ R 2 of 0.77) by carefully selecting and relabelling just 100 additional training images based on initial model prediction discrepancy. The regression-based approach used here returns accurate counts ($$R^2$$ R 2 of 0.92 and 0.77 for the rings and seals, respectively), directly usable in ecological research.

To support sustainable management of apex predator populations, it is important to estimate population size and understand the drivers of population trends to anticipate the consequences of human decisions. Robust population models are needed, which must be based on realistic biological principles and validated with the best available data. A team of international experts reviewed age-structured models of North Atlantic pinniped populations, including Grey seal (Halichoerus grypus), Harp seal (Pagophilus groenlandicus), and Hooded seal (Cystophora cristata). Statistical methods used to fit such models to data were compared and contrasted. Differences in biological assumptions and model equations were driven by the data available from separate studies, including observation methodology and pre-processing. Counts of pups during the breeding season were used in all models, with additional counts of adults and juveniles available in some. The regularity and frequency of data collection, including survey counts and vital rate estimates, varied. Important differences between the models concerned the nature and causes of variation in vital rates (age-dependent survival and fecundity). Parameterisation of age at maturity was detailed and time-dependent in some models and simplified in others. Methods for estimation of model parameters were reviewed and compared. They included Bayesian and maximum likelihood (ML) approaches, implemented via bespoke coding in C, C++, TMB or JAGS. Comparative model runs suggested that as expected, ML-based implementations were rapid and computationally efficient, while Bayesian approaches, which used MCMC or sequential importance sampling, required longer for inference. For grey seal populations in the Netherlands, where preliminary ML-based TMB results were compared with the outputs of a Bayesian JAGS implementation, some differences in parameter estimates were apparent. For these seal populations, further investigations are recommended to explore differences that might result from the modelling framework and model-fitting methodology, and their importance for inference and management advice. The group recommended building on the success of this workshop via continued collaboration with ICES and NAMMCO assessment groups, as well as other experts in the marine mammal modelling community. Specifically, for Northeast Atlantic harp and hooded seal populations, the workshop represents the initial step towards a full ICES benchmark process aimed at revising and evaluating new assessment models. Publisher PDF

Climate warming in the Arctic has led to warmer and earlier springs, and as a result, many food resources for migratory animals become available earlier in the season, as well as become distributed further northwards. To optimally profit from these resources, migratory animals are expected to arrive earlier in the Arctic, as well as shift their own spatial distributions northwards. Here, we review literature to assess whether Arctic migra-tory birds and mammals already show shifts in migration timing or distribution in response to the warming climate. Distribution shifts were most prominent in marine mammals, as expected from observed northward shifts of their resources. At least for many bird species, the ability to shift distributions is likely constrained by available habitat further north. Shifts in timing have been shown in many species of terrestrial birds and ungulates, as well as for polar bears. Within species, we found strong variation in shifts in timing and distributions between populations. Ou r review thus shows that many migratory animals display shifts in migration timing and spatial distribution in reaction to a warming Arctic. Importantly, we identify large knowledge gaps especially concerning distribution shifts and timing of autumn migration, especially for marine mammals. Our understanding of how migratory animals respond to climate change appears to be mostly limited by the lack of long-term monitoring studies.

Marine Spatial Planning (MSP) is a relatively new approach to ocean management and has been widely implemented worldwide. Ideally, MSP should be established as a public process that analyzes and distributes human activities across space and time to achieve ecological, economic and social goals, which historically have been accomplished exclusively in the political arena. However, in most cases MSP seems to be driven primarily by economic interests rather than by sociocultural goals. In this paper, we discuss how integrating the missing sociocultural layers into MSP can help to reduce governance rigidity, promote adaptability in decision-making, support environmental justice, and improve MSP acceptance and uptake. In particular, we focus on identifying possible points of connection between MSP and frameworks based on social-ecological system theory, including co-management and other democratic and empowering alternatives. We conclude by proposing a new definition of the MSP process that is more inclusive, and mindful of users’ rights and sociocultural objectives. If we bridge the gap between the dominant economic rhetoric and a de facto sociocultural-ecological system approach, we are likely to improve the chances of the MSP process succeeding on both the human and nature fronts. Central place foragers often segregate in space, even without signs of direct agonistic interactions. Using parsimonious individual-based simulations, we show that for species with spatial cognitive abilities, individual-level memory of resource availability can be sufficient to cause spatial segregation in the foraging ranges of colonial animals. The shapes of the foraging distributions are governed by commuting costs, the emerging distribution of depleted resources, and the fidelity of foragers to their colonies. When colony fidelity is weak and foragers can easily switch to colonies located closer to favorable foraging grounds, this leads to space partitioning with equidistant borders between neighboring colonies. In contrast, when colony fidelity is strong—for example, because larger colonies provide safety in numbers or individuals are unable to leave—it can create a regional imbalance between resource requirements and resource availability. This leads to nontrivial space-use patterns that propagate through the landscape. Interestingly, while better spatial memory creates more defined boundaries between neighboring colonies, it can lower the average intake rate of the population, suggesting a potential trade-off between an individual’s attempt for increased intake and population growth rates. space partitioning, public information, animal movement, species distribution, animal tracking, central place foraging

Surveys of gray seals (Halichoerus grypus) during the molt period, when they are abundant on land, can be used to monitor changes in population size, but accurate interpretation of results requires an understanding of the molt process and how it may vary between years. This study investigates variability in onset (start date) and duration of visible molt by gray seals in the Wadden Sea, and the influence of environmental conditions on the onset. Molt was monitored in nine captive seals and observed molt phases were applied to wild seals over seven annual molt periods between 2004 and 2010, monitored using aerial photography. The onset of visible molt varied significantly between years, for example it differed 28 d between 2008 and 2009. Model selection by AIC retained one environmental variable that correlated with molt onset; however, its effect was inconsistent within the molt season and did not explain some of the apparent observed annual variation. Hence, the main causes of interannual variability in the onset of molt remain unclear and warrant further study. Researchers should account for annual variability in the onset of molt when interpreting survey results based on molt counts.

Gray seals were first observed breeding in the Dutch Wadden Sea in 1985, after centuries of absence. The breeding colony there is now the largest on the European continent. We describe the changes in gray seal numbers and their geographical expansion, and estimate how these processes were influenced by immigration from other colonies. Counts of hauled out animals were carried out between 1985 and 2013, monitoring three different periods of the seals’ annual cycle. Using priors determined for the UK population, a Bayesian demographic model was fitted to pup numbers to estimate the population parameters driving the growth. This included immigration of subadults into the breeding population, which contributed to an average growth rate in the pup counts of 19%/yr, much higher than expected in a closed population. This immigration may account for approximately 35% of the total annual growth. In addition, at least 200 gray seals from the UK visit the area temporarily. Recovery of the population in the Netherlands occurred more than 50 yr after gray seals were protected in the UK. These time scales should be taken into account when studying long living marine mammals, e.g., in impact and conservation studies.

1. If animals moved randomly in space, the use of different habitats would be proportional to their availability. Hence, deviations from proportionality between use and availability are considered the tell-tale sign of preference. This principle forms the basis for most habitat selection and species distribution models fitted to use-availability or count data (e.g. MaxEnt and Resource Selection Functions). 2. Yet, once an essential habitat type is sufficiently abundant to meet an individual's needs, increased availability of this habitat type may lead to a decrease in the use/availability ratio. Accordingly, habitat selection functions may estimate negative coefficients when habitats are superabundant, incorrectly suggesting an apparent avoidance. Furthermore, not accounting for the effects of availability on habitat use may lead to poor predictions, particularly when applied to habitats that differ considerably from those for which data have been collected. 3. Using simulations, we show that habitat use varies non-linearly with habitat availability, even when individuals follow simple movement rules to acquire food and avoid risk. The results show that the impact of availability strongly depends on the type of habitat (e.g. whether it is essential or substitutable) and how it interacts with the distribution and availability of other habitats. 4. We demonstrate the utility of a variety of existing and new methods that enable the influence of habitat availability to be explicitly estimated. Models that allow for non-linear effects (using b-spline smoothers) and interactions between environmental covariates defining habitats and measures of their availability were best able to capture simulated patterns of habitat use across a range of environments. 5. An appealing aspect of some of the methods we discuss is that the relative influence of availability is not defined a priori, but directly estimated by the model. This feature is likely to improve model prediction, hint at the mechanism of habitat selection, and may signpost habitats that are critical for the organism's fitness.

Anthropogenic sound in the marine environment can have negative consequences for marine fauna. Since most sound sources are intermittent or continuous, estimating how many individuals are exposed over time remains challenging, as this depends on the animals' mobility. Here we explored how animal movement influences how many, and how often, animals are impacted by sound. In a dedicated study, we estimated how different movement strategies affect the number of individual harbour porpoises Phocoena phocoena receiving temporary or permanent hearing loss due to underwater detonations of recovered explosives (mostly WWII aerial bombs). Geo-statistical distribution models were fitted to data from 4 marine mammal aerial surveys and used to simulate the distribution and movement of porpoises. Based on derived dose-response thresholds for temporary (TTS) or permanent threshold shifts (PTS), we estimated the number of animals affected in a single year. When individuals were free-roaming, an estimated 1200 and 24 000 unique individuals would suffer PTS and TTS, respectively. This equates to respectively 0.50 and 10% of the estimated North Sea population. In contrast, when porpoises remained in a local area, fewer animals would receive PTS and TTS (1100 [0.47%] and 15 000 [6.5%], respectively), but more individuals would be subjected to repeated exposures. Because most anthropogenic sound-producing activities operate continuously or intermittently, snapshot distribution estimates alone tend to underestimate the number of individuals exposed, particularly for mobile species. Hence, an understanding of animal movement is needed to estimate the impact of underwater sound or other human disturbance.

European harbour seal (Phoca vitulina) populations decreased substantially during the phocine distemper virus (PDV) outbreaks of 1988 and 2002. Different hypotheses have stated that various seals and terrestrial carnivore species might be the source of infection. To further analyse these hypotheses, grey (Halichoerus grypus) and ringed (Phoca hispida) seals, polar bears (Ursus maritimus) and minks (Mustela lutreola) were sampled from the North Sea and East Greenland coasts between 1988 and 2004 and investigated by RT-PCR using a panmorbillivirus primer pair. However, all samples were negative for morbillivirus nucleic acid.

Estimating the spatial position of organisms is essential to quantify interactions between the organism and the characteristics of its surroundings, for example, predator-prey interactions, habitat selection, and social associations. Because marine mammals spend most of their time under water and may appear at the surface only briefly, determining their exact geographic location can be challenging. Here, we developed a photogrammetric method to accurately estimate the spatial position of marine mammals or birds at the sea surface. Digital recordings containing landscape features with known geographic coordinates can be used to estimate the distance and bearing of each sighting relative to the observation point. The method can correct for frame rotation, estimates pixel size based on the reference points, and can be applied to scenarios with and without a visible horizon. A set of R functions was written to process the images and obtain accurate geographic coordinates for each sighting. The method is applied to estimate the spatiotemporal fine-scale distribution of harbour porpoises in a tidal inlet. Video recordings of harbour porpoises were made from land, using a standard digital single-lens reflex (DSLR) camera, positioned at a height of 9.59m above mean sea level. Porpoises were detected up to a distance of 3136m (mean 596m), with a mean location error of 12m. The method presented here allows for multiple detections of different individuals within a single video frame and for tracking movements of individuals based on repeated sightings. In comparison with traditional methods, this method only requires a digital camera to provide accurate location estimates. It especially has great potential in regions with ample data on local (a)biotic conditions, to help resolve functional mechanisms underlying habitat selection and other behaviors in marine mammals in coastal areas.

D.J.F.R., G.H., V.M.J., S.E.W.M. and B.M. were funded by the UK Department of Energy and Climate Change (DECC) as part of their Offshore Energy Strategic Environmental Assessment programme. The tags and their deployment were funded by GSP, NUON, RWE, Eneco and Gemini, DECC, Natural Environment Research Council, Scottish Natural Heritage and Marine Scotland. On land, species from all trophic levels have adapted to fill vacant niches in environments heavily modified by humans (e.g. [1]). In the marine environment, ocean infrastructure has led to artificial reefs, resulting in localized increases in fish and crustacean density [2]. Whether marine apex predators exhibit behavioural adaptations to utilise such a scattered potential resource is unknown. Using high resolution GPS data we show how infrastructure, including wind turbines and pipelines, shapes the movements of individuals from two seal species (Phoca vitulina and Halichoerus grypus). Using state-space models, we infer that these animals are using structures to forage. We highlight the ecological consequences of such behaviour, at a time of unprecedented developments in marine infrastructure. Postprint Postprint

The number of offshore wind farms is increasing rapidly, leading to questions about the environmental impact of such farms. In the Netherlands, an extensive monitoring programme is being executed at the first offshore wind farm (Offshore Windfarm Egmond aan Zee, OWEZ). This letter compiles the short-term (two years) results on a large number of faunal groups obtained so far. Impacts were expected from the new hard substratum, the moving rotor blades, possible underwater noise and the exclusion of fisheries. The results indicate no short-term effects on the benthos in the sandy area between the generators, while the new hard substratum of the monopiles and the scouring protection led to the establishment of new species and new fauna communities. Bivalve recruitment was not impacted by the OWEZ wind farm. Species composition of recruits in OWEZ and the surrounding reference areas is correlated with mud content of the sediment and water depth irrespective the presence of OWEZ. Recruit abundances in OWEZ were correlated with mud content, most likely to be attributed not to the presence of the farm but to the absence of fisheries. The fish community was highly dynamic both in time and space. So far, only minor effects upon fish assemblages especially near the monopiles have been observed. Some fish species, such as cod, seem to find shelter inside the farm. More porpoise clicks were recorded inside the farm than in the reference areas outside the farm. Several bird species seem to avoid the park while others are indifferent or are even attracted. The effects of the wind farm on a highly variable ecosystem are described. Overall, the OWEZ wind farm acts as a new type of habitat with a higher biodiversity of benthic organisms, a possibly increased use of the area by the benthos, fish, marine mammals and some bird species and a decreased use by several other bird species.

We present new and revised data for the phocine distemper virus (PDV) epidemics that resulted in the deaths of more than 23 000 harbour seals Phoca vitulina in 1988 and 30,000 in 2002. On both occasions the epidemics started at the Danish island of Anholt in central Kattegat, and subsequently spread to adjacent colonies in a stepwise fashion. However, this pattern was not maintained throughout the epidemics and new centres of infection appeared far from infected populations on some occasions: in 1988 early positive cases were observed in the Irish Sea, and in 2002 the epidemic appeared in the Dutch Wadden Sea, 6 wk after the initiation of the outbreak at Anholt Island. Since the harbour seal is a rather sedentary species, such 'jumps' in the spread among colonies suggest that another vector species could have been involved. We discussed the role of sympatric species as disease vectors, and suggested that grey seal populations could act as reservoirs for PDV if infection rates in sympatric species are lower than in harbour seals. Alternatively, grey seals could act as subclinical infected carriers of the virus between Arctic and North Sea seal populations. Mixed colonies of grey and harbour seal colonies are found at all locations where the jumps occurred. It seems likely that grey seals, which show long-distance movements, contributed to the spread among regions. The harbour seal populations along the Norwegian coast and in the Baltic escaped both epidemics, which could be due either to genetic differences among harbour seal populations or to immunity. Catastrophic events such as repeated epidemics should be accounted for in future models and management strategies of wildlife populations.

Perfluorinated acids (PFAs) are today widely distributed in the environment, even in remote arctic areas. Recently, perfluorooctane sulfonate (PFOS) has been identified in marine mammals all over the world, but information on the compound-specific tissue distribution remains scarce. Furthermore, although longer perfluorinated carboxylic acids (PFCAs) are used in industry and were shown to cause severe toxic effects, still little is known on potential sources or their widespread distribution. In this study, we report for the first time on levels of longer chain PFCAs, together with some short chain PFAs, perfluorobutane sulfonate (PFBS) and perfluorobutanoate (PFBA), in liver, kidney, blubber, muscle, and spleen tissues of harbor seals (Phoca vitulina) from the Dutch Wadden Sea. PFOS was the predominant compound in all seal samples measured (ranging from 89 to 2724 ng/g wet weight); however, large variations between tissues were monitored. Although these are preliminary results, it is, to our knowledge, the first time that PFBS could be found at detectable concentrations (2.3 +/- 0.7 ng/g w wt) in environmental samples. PFBS was only detected in spleen tissue. PFCA levels were much lower than PFOS concentrations. The dominant PFCA in all tissues was PFNA (perfluorononanoic acid), and concentrations generally decreased in tissues for all other PFCA homologues with increasing chain length. No clear relationship between PFOS levels in liver and kidney was observed. Furthermore, hepatic PFDA (perfluorodecanoic acid) levels increased with increasing body length, but in kidney tissue, PFDA levels showed an inverse relationship with increasing body length. These data suggest large differences in tissue distribution and accumulation patterns of perfluorinated compounds in marine organisms.

An index for the condition of a population should include a measure of the recuperative power or resilience of the population in question. This measure needs to cover both the demographic and physiological condition of the population. Applied to the harbour seal population in the Wadden Sea we therefore address respectively the population development and distribution, and its health condition, and relate these to environmental conditions. The harbour seal population has been severely depleted by hunting in the first half of the 20th century. After hunting was stopped in the mid-1970s the population recovered gradually. This recovery was twice interrupted by Phocine Distemper Virus (PDV) outbreaks in 1988 and 2002. These PDV-epizootics reduced the population by 57% and 50% respectively. They also lead to changes in age and sex structure of the population, which gradually returned to a stable age-structure. Despite the reduction in population size by respectively 57% and 50%, the population showed a strong recovery with a growth rate close to the considered maximum possible for this species. The observed changes in the distribution of the population over the 4 sub-regions indicate that distribution is not a static phenomenon. Long term field and pathological investigations point out that the general health status of the population has improved, particularly that of newborn seals (0-6months old). The increasing prevalence of parasites in lungs and intestine warrants continued monitoring of the health status of seals. This is especially relevant in view of the exponential increase of the population, which may finally approach the carrying capacity of the area. Concluding, we canstate that the condition of the population in terms of demographic and health parameters is satisfactory. The best guarantee for maintaining such a favourable conservation status is to abstain from human interferences (e.g. rescue, rehabilitation and release) with natural population processes. However, in practice there is increasing exploitation of marine waters. This requires a continuous monitoring of the potential impacts on the population, particularly of the effect on foraging and migratory behaviour.