Rivers act as vital arteries to the world's oceans, delivering fresh water and nutrients that sustain marine ecosystems. Globally, river flow increasingly is being altered by climate change and anthropogenic pressures; yet the significance of rivers to predatory marine species, such as seabirds, and the extent to which river-related changes affect their food webs, remains poorly understood. This review synthesises 51 studies specifically designed to examine river influences on seabird habitat selection, diet, health, and demographics, while highlighting methodological approaches and ecological patterns. Although river-related variables remain underutilised in seabird research, 88% (45/51) of studies that included them reported clear evidence of river effects for at least one type of seabird response, suggesting these ecological links are under-recognised rather than absent. When selected as the primary explanatory (most informative) variable in each study, plume-based metrics were conclusive in 95% (19/20) of cases, whereas river-specific metrics (namely number of rivers, river outflow and distance to river mouths) were conclusive in 84% (26/31) of cases, confirming that both metric types are highly reliable once aligned with the seabird response in question. River-influenced coastal waters consistently supported critical foraging hotspots across all seabird orders, whilst also exposing birds to potential pollutant burdens and altered prey dynamics. Seabird dietary data are a valuable indicator of prey variability associated with river outflows, with greater prey diversity recorded in estuarine habitats compared to marine ones. Rivers exhibited mixed effects under anthropogenic pressures but generally positive influences during climate disturbances, suggesting that seabirds may increasingly depend on riverine environments as buffers against changing marine conditions. We recommend expanding investigations into river impacts on seabird health in tropical systems, incorporating long-term hydrological influences, and prioritising the integration of river-specific and oceanographic data to predict seabird responses more effectively in a rapidly changing world.

Phenotypic responses to climate affect individual fitness, but the extent to which this translates into effects on population dynamics remains poorly understood. We assemble 213 time series on phenotypes and population sizes of wild vertebrates globally and match them with local climate data. Our meta-analysis shows that morphological traits are mostly climate insensitive. However, phenology is earlier in warmer-than-average years, which contributes positively to population growth in most species. At lower latitudes, temperature has weaker effects on phenology but stronger direct negative effects on population growth, likely because these populations are less capable of tracking climate via plasticity. Variation in the phenology-mediated effect of temperature on population growth cannot be explained by latitude, generation time, migratory mode, or diet. This suggests that simple relationships between species characteristics and population responses to warming may not occur in nature. Instead, we may need to embrace ecological complexity by considering local-scale predictors that capture intra-specific variation.

Faced with unpredictable prey in dynamic ecosystems, seabirds have developed strategies to efficiently locate food, particularly during breeding when time and space are constrained. Using 13 yr of GPS data from 422 little penguins Eudyptula minor, we tested whether they rely on consistent foraging areas to reduce search times or respond to dynamic environmental cues. Our results suggest a combination of these strategies, depending on the spatial scale. Little penguins used only 32% of the potential foraging range, concentrating in shallow waters within 70 km and mostly east of their colony, as confirmed by our spatial density models. However, finer-scale analyses using kriging maps showed no recurring foraging hotspots over 13 yr, suggesting high inter-annual variability in prey availability. Despite this variability, penguins consistently spent more time in the most productive areas and also exploited more areas with favourable environmental conditions to access prey (thermocline, current, waves), even as these conditions shifted within and across breeding seasons. In years of poor conditions, they foraged farther from the colony, resulting in lower breeding success. These results emphasise the adaptability of little penguins to dynamic environmental conditions, but underscore the vulnerability of their foraging and breeding success to oceanographic variability in a climate-impacted ecosystem.

Early environmental conditions experienced during juvenile growth are known to have marked effects on adult phenotypes in animal populations. Yet, the life-history outcomes of variable growth strategies have rarely been investigated in wild populations. The aim of this study was to examine the natural variation of growth patterns displayed within a seabird population and assess their impact on juvenile survival, age at first reproduction, lifetime reproductive outputs (LRO) and longevity. Using a 26-year study on the ecology of little penguins, we compiled over 2200 chick growth curves and defined 11 growth parameters classified by magnitude, form and rate. Although the growth curves formed a continuum according to these 11 growth parameters, non-supervised statistical clustering showed that growth trajectories clustered into three main groups: fast, slow and light. Fast chicks (n = 48%) attained the highest maximum mass in the shortest amount of time, whereas slow chicks (n = 33%) stood out by a prolonged (+7 days, i.e. +13% in comparison to fast chicks) and irregular period of juvenile growth. Finally, light chicks (n = 19%) reached low maximum and fledging masses (~-350 g; -37% and -36% of fast and slow chicks). We tested for the effects of chick growth parameters on subsequent annual vital rates estimated through capture-mark-recapture methods as well as longer term effects on life-history outcomes using Markov chain models. Fast and slow individuals had the highest survival rates from hatching to yearling age (19% and 17%, respectively), while light chicks were at a disadvantage during this initial period (3% survival). Fast individuals reproduced 12% earlier (2.6 years old) than slow individuals, had 12.5%-88% greater longevity (up to 21 years old), and produced 1.2-3.8 times more eggs over their lifespan than slow and light individuals, respectively. Fast chicks reached maturity faster and produced more offspring during their lifetime without discernible negative effects to their longevity, highlighting possible silver spoon effects.

Phenology is a major component of animals’ breeding, as they need to adjust their breeding timing to match optimal environmental conditions. While the effects of shifting phenology are well-studied on populations, few studies emphasise its ecological causes and consequences at the inter-individual level. Using a 20-year monitoring of more than 2500 breeding events from 500 breeding little penguins (Eudyptula minor), a very asynchronously breeding seabird, we investigated the consequences of late breeding on present and next breeding events. We found that individuals breeding later had reduced breeding success, lighter chicks at fledging, lower probability of laying a second clutch, and decreased parents’ post-breeding body condition. Importantly, we found important cycling effects where delayed breeding during a given year led to significantly later laying date, lower breeding probability and lower breeding success when they breed during the next season, suggesting potential carry-over effects from one season to the next. To further understand the causes of such variability in phenology while earlier breeding is associated with better individual fitness, we aimed to assess intrinsic differences amongst individuals. We showed that the heterogeneity in breeding timing was partly fixed, the laying date being a significantly repeatable behaviour (17

Passive acoustic monitoring is firmly established as an effective non-invasive technique for wildlife monitoring. The analysis of animal vocalisations recorded in their natural habitats is commonly used to monitor species occupancy, distribution mapping and community composition. However, the ability to distinguish between individual animals by their vocalisations remains underexplored and presents an exciting opportunity to study individual animal behaviour and population demographics in more detail. In this work, we investigate bioacoustic individual-level recognition. We extend on the predominant focus of existing work, where all individuals are known a priori, and additionally address situations where only a subset of the population is initially known and labelled. This is crucial because wildlife populations are constantly changing so that solutions operating only within a known set of individuals are not realistically applicable in the wild. Using two novel datasets, we show that models initially trained to classify only known individuals can also be extended to detect new and previously unknown individuals not included in the training set. We demonstrate that feature extractors pretrained on species classification can be successfully adapted for this task. Extending individual-level recognition to unknown individuals, so-called out-of-distribution classification, is a crucial step towards making individual recognition a realistic possibility in the wild.

Upon the rise in the intensity, duration and frequency of extreme events threatening life on Earth across land and ocean, there is a need to select ecologically significant areas for targeted management interventions. This study assesses the spatial overlap of multiple extreme events and calculates the cumulative magnitude faced across the Southern Hemisphere during the last decades, focusing on the 18 species of penguins as ecosystem sentinels that rely on terrestrial and marine environments. Our analysis identifies African, Snares, Emperor, Adélie and Galápagos as the penguin species experiencing the highest cumulative values of extreme events. Additionally, when looking at the trends, we identify that all penguin species, except the Galápagos penguin, will face a potential increase in extreme events if current trends are maintained. This study carries a crucial message in conservation, establishing a spatial framework that allows sounding the alert in ecologically important areas to reduce vulnerability during present and future extreme events.

ABSTRACTUnderstanding the relative contributions of environmental, behavioural and social factors to reproductive success is crucial for predicting population dynamics of seabirds. However, these factors are often studied in isolation, limiting our ability to evaluate their combined influence. This study investigates how marine environmental variables, foraging behaviour and social factors (divorce), influence reproductive success in little penguins (Eudyptula minor) over 13 breeding seasons. By examining these factors together, we aimed to identify which is the most reliable predictor of population‐level reproductive success. We found that divorce rate was the most consistent predictor of reproductive success, with lower annual rates of divorce preceding the breeding season associated with higher hatching and fledging success. Foraging trip duration also influenced reproductive success, but in contrasting ways: Longer trips during egg incubation were linked with increased hatching success, while shorter trips after hatching led to higher fledging success. Marine environmental conditions had unexpected effects, with a lower Southern Oscillation Index (SOI) correlating with improved hatching and fledging success, in contrast to previous research, while sea surface temperature (SST) had no significant effect on reproductive success. This highlights the complexity of seabird breeding responses to large‐scale oceanographic indices, suggesting SOI and SST are generally less reliable measures to use as indicators of reproductive success. Our results suggest that divorce rate could serve as a valuable, noninvasive index of reproductive success in seabirds.

As charismatic and iconic species, penguins can act as “ambassadors” or flagship species to promote the conservation of marine habitats in the Southern Hemisphere. Unfortunately, there is a lack of reliable, comprehensive, and systematic analysis aimed at compiling spatially explicit assessments of the multiple impacts that the world's 18 species of penguin are facing. We provide such an assessment by combining the available penguin occurrence information from Global Biodiversity Information Facility (>800,000 occurrences) with three main stressors: climate-driven environmental changes at sea, industrial fisheries, and human disturbances on land. Our analyses provide a quantitative assessment of how these impacts are unevenly distributed spatially within species' distribution ranges. Consequently, contrasting pressures are expected among species, and populations within species. The areas coinciding with the greatest impacts for penguins are the coast of Perú, the Patagonian Shelf, the Benguela upwelling region, and the Australian and New Zealand coasts. When weighting these potential stressors with species-specific vulnerabilities, Humboldt ( Spheniscus humboldti ), African ( Spheniscus demersus ), and Chinstrap penguin ( Pygoscelis antarcticus ) emerge as the species under the most pressure. Our approach explicitly differentiates between climate and human stressors, since the more achievable management of local anthropogenic stressors (e.g., fisheries and land-based threats) may provide a suitable means for facilitating cumulative impacts on penguins, especially where they may remain resilient to global processes such as climate change. Moreover, our study highlights some poorly represented species such as the Northern Rockhopper ( Eudyptes moseleyi ), Snares ( Eudyptes robustus ), and Erect-crested penguin ( Eudyptes sclateri ) that need internationally coordinated efforts for data acquisition and data sharing to understand their spatial distribution properly.

Passive acoustic monitoring is firmly established as an effective non-invasive technique for wildlife monitoring. The analysis of animal vocalizations recorded in their natural habitats is commonly used to monitor species occupancy, distribution mapping and community composition. The ability to distinguish between individual animals, however, remains underexplored and presents an exciting opportunity to study individual animal behavior and population demographics in more detail. In this work, we investigate bioacoustic individual-level recognition. In contrast to existing work, we focus on settings where only a subset of the existing population is known and labeled. This is crucial because wildlife populations are constantly changing so that solutions operating only within a known set of individuals are not realistically applicable in the wild. Using two novel datasets, we show that models initially trained to classify only known individuals can also be extended to detect new, previously unseen, individuals that are not part of the training set. We demonstrate that feature extractors pretrained on species classification can be successfully adapted for this task. Extending individual-level recognition to unknown individuals, so-called out-of-distribution classification, is a crucial step towards making individual recognition a realistic possibility in the wild. ### Competing Interest Statement The authors have declared no competing interest.

Semiaquatic taxa, including humans, often swim at the air-water interface where they waste energy generating surface waves. For fully marine animals however, theory predicts the most cost- efficient depth- use pattern for migrating, air- breathing species that do not feed in transit is to travel at around 2 to 3 times the depth of their body diameter, to minimize the vertical distance traveled while avoiding wave drag close to the surface. This has rarely been examined, however, due to depth measurement resolution issues at the surface. Here, we present evidence for the use of this strategy in the wild to the nearest centimeter and document the switch to shallow swimming during naturally occurring long- distance migrations. Using high- resolution depth- accelerometry and sea turtle, penguin, and whale species, we show that near- surface swimming is likely used broadly across nonforaging diving animals to minimize the cost of transport.

The raw data of the mean laying date at each site.

The COVID-19 pandemic and its lock-down measures have resulted in periods of reduced human activity, known as anthropause. While this period was expected to be favorable for the marine ecosystem, due to a probable reduction of pollution, shipping traffic, industrial activity and fishing pressure, negative counterparts such as reduced fisheries surveillance could counterbalance these positive effects. Simultaneously, on-land pressure due to human disturbance and tourism should have drastically decreased, potentially benefiting land-breeding marine animals such as seabirds. We analyzed 11 breeding seasons of data on several biological parameters of little penguins from a popular tourist attraction at Phillip Island, Australia. We investigated the impact of anthropogenic activities on penguin behavior during the breeding season measured by (1) distribution at sea, (2) colony attendance, (3) isotopic niche (4) chick meal mass, and (5) offspring investment against shipping traffic and number of tourists. The 2020 lock-downs resulted in a near absence of tourists visiting the Penguin Parade®, which was otherwise visited by 800,000+ visitors on average per breeding season. However, our long-term analysis showed no effect of the presence of visitors on little penguins' activities. Surprisingly, the anthropause did not trigger any changes in maritime traffic intensity and distribution in the region. We found inter- and intra-annual variations for most parameters, we detected a negative effect of marine traffic on the foraging efficiency. Our results suggest that environmental variations have a greater influence on the breeding behavior of little penguins compared to short-term anthropause events. Our long-term dataset was key to test whether changes in anthropogenic activities affected the wildlife during the COVID-19 pandemic.

Abstract Culturally dependent human social behaviours involving artificial light usage can potentially affect light pollution patterns and thereby impact the night‐time ecology in populated areas, although to date this has not been examined globally. By analysing continuous (monthly), highly resolved, spatially explicit data on global night lights (Visible and Infrared Imaging Radiometer Suite–Day/Night Band‐VIIRS‐DNB; 2014–2019) with circular statistical techniques, we evaluated whether macro‐cultural activities involving social aggregations and the use of artificial lights shape annual lighting patterns globally. Scheduled routines associated with cultural‐specific festivities appear to be important drivers of observed seasonal patterns in urban night‐time lights. For instance, the display of Christmas lights between Christmas and Epiphany Day celebrations (December–January) coincides with the annual peak in urban night‐time light intensity in Christian countries. Analogously, night celebrations during the Holy Month of Ramadam (from May to July) or the month‐long period of Karthika Masam (from October to November) fits with annual night light peaks in Muslim and Hindu countries. Annual peaks of urban light intensity in China and Vietnam also match with Chinese and Vietnamese (Tê't) New Year celebrations (January–February). In contrast, predominantly Buddhist countries, which do not have such prominent and prolonged celebrations involving artificial lights, show a relatively uniform distribution of night light peaks throughout the annual cycle. Social behaviour and sociocultural contexts help explain how people modify the global nightscape and contribute to light pollution globally. Understanding the cultural contexts responsible for peaks in artificial light usage is an important first step if humans are to mitigate any deleterious effects associated with global increases in night‐time light pollution. Read the free Plain Language Summary for this article on the Journal blog.

Understanding the spatial-temporal marine habits is crucial to conserving air-breathing marine animals that breed on islands and forage at sea. This study, focusing on little penguins from Phillip Island, Australia, employed tracking data to identify vital foraging areas during breeding season. Long-term data from sub-colonies and breeding stages were analysed using 50%, 75%, and 90% kernel utilisation distributions (KUDs). Breeding success, classified as low, average, or high, guided the exploration of site, year, and breeding stage-specific habitats. Using Marxan, a widely used conservation planning tool, the study proposes both static and dynamic spatial-temporal scenarios for protection based on KUDs. The dynamic approach, requiring less space than the static strategy, was more efficient and likely more acceptable to stakeholders. The study underscores the need for comprehensive data in conservation plans, as relying on one nesting site’s data might miss essential foraging areas for penguins in other locations. This study demonstrates the efficacy of animal tracking data in spatial conservation prioritisation and marine spatial planning. The dynamic areas frequented emerged as a strategy to safeguard core regions at sea, offering insights to improve the conservation of iconic species like little penguins and promoting the health of islands and the entire marine ecosystem.### Competing Interest StatementThe authors have declared no competing interest.

Raw data used for the analysis of foraging and reproductive success.

While differences in foraging and reproductive success are well studied between seabird colonies, they are less understood at a smaller subcolony scale. Working with little penguins (Eudyptula minor) at Phillip Island, Australia, we used an automated penguin monitoring system and performed regular nest checks at two subcolonies situated 2 km apart during the 2015/2016 breeding seasons. We examined whether foraging and reproductive success differed between subcolonies. We used satellite data to examine how sea surface temperature, as environmental pressure, in the foraging regions from each subcolony influenced their foraging performance. In the pre-laying and incubation breeding stages, the birds from one subcolony had a lower foraging success than birds from the other. However, this pattern was reversed between the subcolonies in the guard and post-guard stages. Breeding success data from the two subcolonies from 2004–2018 showed that reproductive success and mean egg-laying had a negative relationship with sea surface temperature. We highlighted that variation in foraging and reproductive success can arise in subcolonies, likely due to different responses to environmental conditions and prey availability. Differences at the subcolony level can help refine, develop and improve appropriate species management plans for conserving a range of colonial central place seabirds.

While the heterogeneity among individuals of a population is more and more documented, questions on the paths through which it arises, particularly whether it is linked to fixed heterogeneity or chance alone, are still widely debated. Here, we tested how individual quality, energy allocation trade-offs, and environmental stochasticity define individual fitness. To do so, we simultaneously investigated the contribution of 18 life-history traits to the fitness of breeding little penguins (Eudyptula minor), using a structural equation model. Fitness was highly variable amongst the 162 birds monitored over their entire lifespan. It increased with the individual penguin's ability to increase (a) the number of breeding events (i.e., living longer, breeding younger, breeding more often, and producing more second clutches) and (b) the breeding success per event through increased foraging performances (i.e., mass gained at sea). While all three processes (stochasticity, individual quality, and allocation trade-offs) affected fitness, interindividual variability in fitness was mainly driven by individual quality, birds consistently breeding earlier in the season and displaying higher foraging efficiency exhibiting higher fitness. Why some birds consistently can perform better at sea and breed earlier remains a question to investigate to understand how selection applies to these traits.

Protected areas are a widely adopted resource management strategy for mitigating the consequences of global change and preserve functioning ecosystems. Long-term species monitoring programmes, aided by bio-logging technology, provide insights into the extent and spatial variation of areas occupied by wild animals and inform conservation and management. High-resolution GPS-acceleration data offer a more accurate understanding of animal behavior and area use, compared to location-based inference, emphasizing the significance of specific sites amid long-term climate change. We based our case-study on the largest colony of little penguins ( Eudyptula minor ) located at Phillip Island. Based on a ten-year bio-logging dataset (247 individual tracks), we combine high-resolution bio-logging data from GPS-accelerometer loggers with proxies for resource availability (e.g. Sea Surface Tenperature, thermocline, water turbidity). Using machine learning techniques and Generalized Additive Mixed Models, we quantify the environmental factors determining spatio-temporal variability in foraging effort (defined as hunting time) across different breeding seasons and stages. Little penguins increased their hunting time by reducing spatial displacement (shorter step length) and diving deeper, with a slower increase in hunting effort below 10 m depth. In relation to environmental conditions, penguins increased hunting effort in coastal areas with high turbid and productive waters and decreased effort with increasing Sea Surface Temperature. This gives insights into how these animals allocate effort differently according to shifting environmental conditions. Our analysis offers crucial long-term insights into little penguin area usage in the Bass Strait at sufficient spatial and temporal resolution for management and conservation planning. The Bass Strait is facing intense climatic and anthripogenic pressures, and the findings here on intensity of area usage and strategy shifting according to environmental conditions, are of great relevance for the marine spatial planning currently under development along the coast. Policy implications: High-resolution behavioral information obtained from bio-logging data using GPS-accelerometer tags provides understanding of how species shift strategies in response to environmental variability. This is vital to implement climate-adaptive conservation and management strategies. Given the growing availability of long-term accelerometer datasets within the ecological community, we recommend integrating such high-resolution information into conservation programs.