No abstract

No abstract

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.

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 9.59 m above mean sea level. Porpoises were detected up to a distance of ∽3136 m (mean 596 m), with a mean location error of 12 m. 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.

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 migratory 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.

Impact of climate changes on species is expected to be especially visible at the extremities of the species distribution, where they meet sub-optimal conditions. In Mauritania and Iberia, two genetically isolated populations of harbor porpoises form a distinct ecotype and are presumably locally adapted to the upwelling waters. By analyzing the evolution of mitochondrial genetic variation in the Iberian population between two temporal cohorts (1990-2002 vs. 2012-2015), we report a dramatic decrease in genetic diversity. Phylogenetic analyses including neighboring populations identified two porpoises in southern Iberia carrying a divergent haplotype close to the Mauritanian population, yet forming a distinctive lineage. This suggests that Iberian porpoises may not be as isolated as previously thought with immigration from Mauritania or an unknown population in between, but none from the northern ecotype. The rapid decline in the Iberian mitochondrial diversity may be driven by either genetic drift, or by a dramatic decline in census population size possibly resulting from environmental stochasticity, prey depletion, or acute fishery bycatches. These results illustrate the value of genetics time series to inform demographic trends and emphasize the urgent need for conservation measures to ensure the viability of this small harbor porpoise population in Iberia.

On 19 July 2019 an estimated 20 bottlenose dolphins (Tursiops truncatus) were observed in the Marsdiep, a tidal inlet connecting the North Sea and the Dutch Wadden Sea, between Den Helder and the island of Texel. Photographs and video recordings were made and nine individuals were matched with known dolphins from the Moray Firth, NE Scotland. These are the first matches of this east coast of Scotland population outside the UK and Ireland. Subsequent observations of individuals from this group show that at least some of the animals have returned to Scottish waters, while others were photographed in Danish waters. Furthermore, we report on a photo identification match of a solitary bottlenose dolphin between France and the Netherlands. These matches suggest that bottlenose dolphins, in the Netherlands, originate from two different genetically distinct populations: ‘Coastal South’ and ‘Coastal North’. This evidence of previously unknown long-range movements may have important implications for the conservation and management of this species in European waters.

No abstract