62 research outputs found
Vocal repertoire of long-finned pilot whales (Globicephala melas) in northern Norway
No full textThe knowledge of the vocal repertoire of pilot whales is very limited. In this paper, the vocal repertoire of long-finned pilot whales recorded during different encounters in the Vestfjord in northern Norway between November 2006 and August 2010 are described. Sounds were analysed using two different methods: (1) an observer-based audio-visual inspection of FFT-derived spectrograms, with which, besides a general variety of clicks, buzzes, nonharmonic sounds, and whistles, 129 different distinct call types and 25 subtypes were distinguished. These call types included pulsed calls and discrete structured whistles varying from simple to highly complex structures composed of several segments and elements. In addition, ultrasonic whistles previously not described for pilot whales were found. In addition to the diversity of single calls, call sequences consisting of repetitions and combinations of specific call types were recorded and (2) a parametric approach that permitted the confirmation of the high variability in pilot whale call structures was developed. It is concluded that the pilot whale vocal repertoire is among the most complex for the mammalian species and the high structural variability, along with call repetitions and combinations, require a closer investigation to judge their importance for vocal communication
Quantifying group specificity of animal vocalizations without specific sender information.
No full textRecordings of animal vocalization can lack information about sender and context. This is often the case in studies on marine mammals or in the increasing number of automated bioacoustics monitorings. Here, we develop a framework to estimate group specificity without specific sender information. We introduce and apply a bag-of-calls-and-coefficients approach (BOCCA) to study ensembles of cepstral coefficients calculated from vocalization signals recorded from a given animal group. Comparing distributions of such ensembles of coefficients by computing relative entropies reveals group specific differences. Applying the BOCCA to ensembles of calls recorded from group of long-finned pilot whales in northern Norway, we find that differences of vocalizations within social groups of pilot whales (Globicephala melas) are significantly lower than intergroup differences
Dietary variation within and between populations of northeast Atlantic killer whales Orcinus orca inferred from δ13C and δ15N analyses
No full textEpidermal skin samples from eastern North Atlantic killer whales, Orcinus orca, were analyzed for carbon and nitrogen stable isotope ratios. From those, comparisons within a data set of 17 samples collected from Tysfjord, Norway, in November suggested that diet is relatively specialized during this time period at this location. There were significant differences between a small set of samples from Iceland and those collected from Norway, which had all been assigned to the same population by a previous population genetics study. The results would be consistent with matrilines feeding on either the Norwegian or Icelandic stocks of Atlantic herring (Clupea harengus). There was no significant difference within Icelandic samples between those assigned to the population known to feed upon herring and those assigned to a population hypothesized to follow Atlantic mackerel (Scomber scombrus). The greatest differences were between the epidermal samples analyzed in this study and tooth and bone collagen samples from the North Sea that were analyzed previously, which also showed significantly more variation in isotopic ratios than found for skin samples. These differences could reflect differences in turnover rate, differences in diet-tissue fractionation and discrimination due to the amino acid composition of the different tissues, and/or greater competition promoting dietary variation between groups in the North Sea
Vocal repertoires of two matrilineal social whale species Long-finned Pilot whales (Globicephala melas) & Killer whales (Orcinus orca) in northern Norway
No full textThe aim of this study was to describe and investigate the vocal repertoire and possible factors influencing the size and composition of two matrilineal social whale species: long-finned pilot whales and killer whales in Norway. I was not able to describe the entire vocal repertoire due to the large number of animals in this population and the limitations of fieldwork and time. In this thesis I describe a vocal repertoire subset from seven groups of pilot whales and 11 groups of killer whales recorded in the Vestfjord in northern Norway during the time period 2004 until 2011.
Using observer-based acoustic analysis I could discern 129 call types and 25 subtypes for long-finned pilot whales, and 60 call types and 25 subtypes for killer whales. Per group, pilot whales used an average of 36 call types and killer whales just 25. The general structure of call types was similar, with most call types consisting of one segment and two elements with different structures. The main element structure in pilot whale and killer whale calls was an ascending frequency band. The amount of two-voiced calls was 29% for pilot whales and 47% for killer whales. In addition, I further found different call type combinations and repetitions and investigated ultrasonic whistles, already known in killer whales, but newly described for pilot whales in this study.
The main difference between vocal repertoires of the two species appeared when I looked at call type sharing between the recorded groups. Here I found that pilot whales only shared 28% of their call types and 37% of their total calls with at least one other group, whereas killer whales shared 59% of their call types and 90% of their total calls. Average group size differed: pilot whales were found in larger groups (23 animals) than killer whales (9 animals). Overall number of calls increases with group size, however, I could not find that group size influenced the number of call types. I found rather that vocal repertoire size depends on the length of recording time and of a group’s vocal activity. This must be over one hour at minimum before it can be compared with another. During carousel feeding in killer whales, vocal activity increased. This indicates that behaviour does play a role in vocal repertoire size and composition, at least in killer whales where behaviour is easier to detect and measure than in pilot whales.
In 2011 I discovered a new foraging method for salmon by killer whales in Norway. The publication is presented in Chapter 6. For the first time I was able to observe the same two groups of killer whales for over three months and describe a full repertoire with 59 call types and 25 subtypes. It was not possible to separate the calling of the two groups, but nevertheless it shows that the vocal repertoire is larger than for my earlier, shorter observations. In addition, I found context-specific vocalisations during salmon feeding and non-feeding, and compared it to herring feeding and a food association call from the Icelandic killer whale population. Specifically, certain call type combinations contained the same beginning part (call type NKW-15) in all feeding contexts, but the combinations differed for salmon and herring feeding and between groups. I discuss the possibility of referential and arousal calling in association with food in killer whales.
This thesis is summarised in Chapter 8 with a discussion of all presented results, the limitations of the research, and potential areas for further study to increase our understanding of the two species and study vocal communication more generally in cetaceans
Call for cooperation to contain damage by Chile’s salmon farms
No full textWe suggest a cooperative approach between scientists, aquaculture industries, local fishery, politics and conservationists to protect the local ecosystem in Chile and stop existing threats for marine mammals
Narrating the soundscape of coastal tourism
No full textThis research letter concerns the soundscape along the Norwegian coastline, but the concepts can be applied to most coastal destinations. We wrote this letter in our roles as scholars of tourism and biology as well as members of a non-profit organisation located in northern Norway that is dedicated to science, conservation and the protection of marine mammals. By contrasting a promotional video with an awareness-raising video, we highlight tourism sector’s missed opportunity to foster sustainability and justice and advocate for inclusive storytelling that considers different voices, stories and soundscapes
Mediterraan ouder worden: Groepswonen in een kas
No full textIn het vergrijzende Nederland hebben veel ouderen last van eenzaamheid. Ook blijkt uit onderzoek dat ouderen 95% van hun tijd binnenshuis doorbrengen terwijl de buitenlucht juist ontzettend goed is voor de gezondheid. Door actieve ouderen samen te brengen in een groepswoning wordt de drempel tot sociale interactie verlaagd. Daarbij wordt over het gebouw een kas geplaatst waardoor de gezamenlijke binnentuin het gehele jaar bruikbaar is en een duurzaam binnen klimaat ontstaat. Het ontwerp maakt het mogelijk om in Nederland Mediterraan oud te worden: Een groot deel van de dag buiten doorbrengen met vrienden om je heen.Architectural EngineeringArchitectureArchitecture and The Built Environmen
Reinforcement learning with domain-specific relational inductive biases: Using Graph Neural Networks and domain knowledge
No full textReinforcement Learning (RL) has been used to successfully train agents for many tasks, but generalizing to a different task - or even unseen examples of the same task - remains difficult. In this thesis, Deep Reinforcement Learning (DRL) is combined with Graph Neural Networks (GNNs) and domain knowledge, with the aim of improving the generalization capabilities of RL-agents. In classical DRL setups, Convolutional Neural Networks (CNNs) and Multilayer Perceptrons (MLPs) are often applied as neural network architectures for an agent’s policy and/or value network. In this thesis, however, GNNs are used to represent the policy and value network of an agent, which allows for the application of relational inductive biases that are more domain-specific than those of MLPs and CNNs. Observations received by the agent from a simple navigation task - which requires some relational reasoning - are encoded as graphs, consisting of entities and relations between them, which are based on domain knowledge. These graphs are then used as structured input for the GNN-based architecture of the agent. This approach is inspired by human relational reasoning, which is argued to be an important factor in human generalization capabilities. Several GNN-based architectures are proposed and compared, from which two main architectures are distilled: R-GCN-domain and R-GCN-GAN. In the R-GCN-domain architecture, the graph encoding of observations is based on domain knowledge, whereas in R-GCN-GAN we aim to combine the relational encoding of a CNN with additional, learned relations, allowing for an end-to-end solution that does not require domain knowledge. Sample efficiency and both in- and out-of-distribution generalization performance of our architectures are tested on a new grid world environment called ’Key-Corridors’. We find that adding domain-specific relational inductive biases with the R-GCNdomain architecture significantly improves sample efficiency and out-of-distribution generalization, when compared to MLPs and CNNs. However, we did not succeed in learning these domain-specific relational inductive biases with R-GCN-GAN, which does not manage to significantly outperform a CNN. Overall, the results indicate that applying relational reasoning in RL - through the use of GNNs and domain knowledge - can be an important tool for improving sample efficiency and generalization performance. Computer Scienc
Sustainable material research: CoffeeBased panels in a condiment organizer
No full textPlastics that originate from renewable sources, such as potatoes or fungi, are called biobased plastics. Because of their origin, biobased plastics pose as sustainable alternative to traditional plastics, which are made from oil, a fossil substance. CoffeeBased is a small company that manufactures biobased plastic products, using their own developed materials (CB materials). To make these even more sustainable, CoffeeBased adds coffee waste as filler material to the biobased plastics. This reduces waste accumulation while generating a unique material. The coffee waste, called spent coffee grounds (SCG), adds attractive properties to the CB material. This includes a distinct coffee scent, and in combination with colorants, a coffee-like color is achieved. These unique properties suit the ambience of coffee product environments, like corporate coffee corners. CoffeeBased currently implements their CB material in (products such as) coffee machine fronts, coffee cups and coffee stations. The panels applied in the coffee stations are the most relevant to this thesis. The panel consists of a coating, adhesive, and a core that is laminated with CB material, referred to as CB laminate (EMPA, 2019). Unfortunately, the CB laminate does not function properly in the desired user context, as it discolors and distorts. To prevent this impact from reducing its aesthetic appeal, optimization is required. Next to that, the sustainable character of the panel in its entirety is questionable, as the other components have not been analyzed on this characteristic. Therefore, the first design challenge is to find an integral solution to improve both the context resistance as the sustainable value. The second challenge is to design a condiment organizer, which could form the perfect occasion to test the proposed optimized panel. The function of the condiment organizer is to display and organize additives used in coffee machine beverages, called condiments. Next to the implementation of the panel, the design has to incorporate a form language that corresponds with other CoffeeBased products. Based on the desires of the end user, limits of the manufacturing facilities and the resources available to CoffeeBased, requirements were set. The sustainability requirements are set on incorporating more waste and more biobased materials whilst ensuring the emittance of volatile organic compounds (VOCs) remains low. Keeping these requirements in mind, alternatives to the current components of the panel were looked for in market research. The theoretical best components were chosen and (physically) tested, which were CB2 as laminate material, Resysta NoWood by FiberPlast as core, BioImpact by Canect as adhesive and the clear lacquer by BioPin as coating. Four panel concepts are compared, each with different advantages. Increase in waste content is considered the most important optimization, as it suits the uniqueness of the CoffeeBased material. Therefore, the best suitable panel concept is the one composed of an EcoBoard core, CB1 laminate, BioImpact adhesive and BioPin lacquer. The optimized panel provides an increased sustainable value by a higher waste content of 83% and a biobased content increase of 12%, this while providing a potential VOC emittance reduction of 1%. Also, the concept entails edge band development, aiming to increase its aesthetic appeal and user context resistance. To tackle the second challenge, design ideas were generated based on the insights gained from brainstorm sessions. The four most promising ideas were presented to CoffeeBased, whose favorite was selected for further conceptualization. The final condiment organizer consists of a pre-assembled back frame, available with three, five or seven compartments. Condiment specific fronts can be selected by the user, based on their preference. Unfortunately, the necessary manufacturing method prevents the implementation of the optimized panel, thus a no-added formaldehyde plywood core is incorporated. Luckily, the aesthetic appearance does fit into the product portfolio of CoffeeBased, with its contour and surface engravements. Therefore, this thesis project does not only provide CoffeeBased with valuable sustainable panel solutions, but also a new product design for their product portfolio.Integrated Product Desig
- …
