Getting the bugs out of AI: Advancing ecological research on arthropods through computer vision.

Deep learning for computer vision has shown promising results in the field of entomology, however, there still remains untapped potential. Deep learning performance is enabled primarily by large quantities of annotated data which, outside of rare circumstances, are limited in ecological studies. Currently, to utilize deep learning systems, ecologists undergo extensive data collection efforts, or limit their problem to niche tasks. These solutions do not scale to region agnostic models. However, there are solutions that employ data augmentation, simulators, generative models, and self-supervised learning that can supplement limited labelled data. Here, we highlight the success of deep learning for computer vision within entomology, discuss data collection efforts, provide methodologies for optimizing learning from limited annotations, and conclude with practical guidelines for how to achieve a foundation model for entomology capable of accessible automated ecological monitoring on a global scale.

Harvesting can stabilise population fluctuations and buffer the impacts of extreme climatic events.

Harvesting can magnify the destabilising effects of environmental perturbations on population dynamics and, thereby, increase extinction risk. However, population-dynamic theory predicts that impacts of harvesting depend on the type and strength of density-dependent regulation. Here, we used logistic population growth models and an empirical reindeer case study to show that low to moderate harvesting can actually buffer populations against environmental perturbations. This occurs because of density-dependent environmental stochasticity, where negative environmental impacts on vital rates are amplified at high population density due to intra-specific resource competition. Simulations from our population models show that even low levels of harvesting may prevent overabundance, thereby dampening population fluctuations and reducing the risk of population collapse and quasi-extinction following environmental perturbations. Thus, depending on the species' life history and the strength of density-dependent environmental drivers, low to moderate harvesting can improve population resistance to increased climate variability and extreme weather expected under global warming.

Evaluating expert-based habitat suitability information of terrestrial mammals with GPS-tracking data.

Aim: Macroecological studies that require habitat suitability data for many species often derive this information from expert opinion. However, expert-based information is inherently subjective and thus prone to errors. The increasing availability of GPS tracking data offers opportunities to evaluate and supplement expert-based information with detailed empirical evidence. Here, we compared expert-based habitat suitability information from the International Union for Conservation of Nature (IUCN) with habitat suitability information derived from GPS-tracking data of 1,498 individuals from 49 mammal species. Location: Worldwide. Time period: 1998-2021. Major taxa studied: Forty-nine terrestrial mammal species. Methods: Using GPS data, we estimated two measures of habitat suitability for each individual animal: proportional habitat use (proportion of GPS locations within a habitat type), and selection ratio (habitat use relative to its availability). For each individual we then evaluated whether the GPS-based habitat suitability measures were in agreement with the IUCN data. To that end, we calculated the probability that the ranking of empirical habitat suitability measures was in agreement with IUCN's classification into suitable, marginal and unsuitable habitat types. Results: IUCN habitat suitability data were in accordance with the GPS data (> 95% probability of agreement) for 33 out of 49 species based on proportional habitat use estimates and for 25 out of 49 species based on selection ratios. In addition, 37 and 34 species had a > 50% probability of agreement based on proportional habitat use and selection ratios, respectively. Main conclusions: We show how GPS-tracking data can be used to evaluate IUCN habitat suitability data. Our findings indicate that for the majority of species included in this study, it is appropriate to use IUCN habitat suitability data in macroecological studies. Furthermore, we show that GPS-tracking data can be used to identify and prioritize species and habitat types for re-evaluation of IUCN habitat suitability data.

Landscape modification and nutrient-driven instability at a distance.

Almost 50 years ago, Michael Rosenzweig pointed out that nutrient addition can destabilise food webs, leading to loss of species and reduced ecosystem function through the paradox of enrichment. Around the same time, David Tilman demonstrated that increased nutrient loading would also be expected to cause competitive exclusion leading to deleterious changes in food web diversity. While both concepts have greatly illuminated general diversity-stability theory, we currently lack a coherent framework to predict how nutrients influence food web stability across a landscape. This is a vitally important gap in our understanding, given mounting evidence of serious ecological disruption arising from anthropogenic displacement of resources and organisms. Here, we combine contemporary theory on food webs and meta-ecosystems to show that nutrient additions are indeed expected to drive loss in stability and function in human-impacted regions. Our models suggest that destabilisation is more likely to be caused by the complete loss of an equilibrium due to edible plant species being competitively excluded. In highly modified landscapes, spatial nutrient transport theory suggests that such instabilities can be amplified over vast distances from the sites of nutrient addition. Consistent with this theoretical synthesis, the empirical frequency of these distant propagating ecosystem imbalances appears to be growing. This synthesis of theory and empirical data suggests that human modification of the Earth is strongly connecting distantly separated ecosystems, causing rapid, expansive and costly nutrient-driven instabilities over vast areas of the planet. Similar to existing food web theory, the corollary to this spatial nutrient theory is that slowing down spatial nutrient pathways can be a potent means of stabilising degraded ecosystems.

Body size and digestive system shape resource selection by ungulates: A cross-taxa test of the forage maturation hypothesis.

The forage maturation hypothesis (FMH) states that energy intake for ungulates is maximised when forage biomass is at intermediate levels. Nevertheless, metabolic allometry and different digestive systems suggest that resource selection should vary across ungulate species. By combining GPS relocations with remotely sensed data on forage characteristics and surface water, we quantified the effect of body size and digestive system in determining movements of 30 populations of hindgut fermenters (equids) and ruminants across biomes. Selection for intermediate forage biomass was negatively related to body size, regardless of digestive system. Selection for proximity to surface water was stronger for equids relative to ruminants, regardless of body size. To be more generalisable, we suggest that the FMH explicitly incorporate contingencies in body size and digestive system, with small-bodied ruminants selecting more strongly for potential energy intake, and hindgut fermenters selecting more strongly for surface water.

Habitat selection patterns are density dependent under the ideal free distribution.

Despite being widely used, habitat selection models are rarely reliable and informative when applied across different ecosystems or over time. One possible explanation is that habitat selection is context-dependent due to variation in consumer density and/or resource availability. The goal of this paper is to provide a general theoretical perspective on the contributory mechanisms of consumer and resource density-dependent habitat selection, as well as on our capacity to account for their effects. Towards this goal we revisit the ideal free distribution (IFD), where consumers are assumed to be omniscient, equally competitive and freely moving, and are hence expected to instantaneously distribute themselves across a heterogeneous landscape such that fitness is equalised across the population. Although these assumptions are clearly unrealistic to some degree, the simplicity of the structure in IFD provides a useful theoretical vantage point to help clarify our understanding of more complex spatial processes. Of equal importance, IFD assumptions are compatible with the assumptions underlying common habitat selection models. Here we show how a fitness-maximising space use model, based on IFD, gives rise to resource and consumer density-dependent shifts in consumer distribution, providing a mechanistic explanation for the context-dependent outcomes often reported in habitat selection analysis. Our model suggests that adaptive shifts in consumer distribution patterns would be expected to lead to nonlinear and often non-monotonic patterns of habitat selection. These results indicate that even under the simplest of assumptions about adaptive organismal behaviour, habitat selection strength should critically depend on system-wide characteristics. Clarifying the impact of adaptive behavioural responses may be pivotal in making meaningful ecological inferences about observed patterns of habitat selection and allow reliable transferability of habitat selection predictions across time and space.

Supply and demand drive a critical transition to dysfunctional fisheries.

There is growing awareness of the need for fishery management policies that are robust to changing environmental, social, and economic pressures. Here we use conventional bioeconomic theory to demonstrate that inherent biological constraints combined with nonlinear supply-demand relationships can generate threshold effects due to harvesting. As a result, increases in overall demand due to human population growth or improvement in real income would be expected to induce critical transitions from high-yield/low-price fisheries to low-yield/high-price fisheries, generating severe strains on social and economic systems as well as compromising resource conservation goals. As a proof of concept, we show that key predictions of the critical transition hypothesis are borne out in oceanic fisheries (cod and pollock) that have experienced substantial increase in fishing pressure over the past 60 y. A hump-shaped relationship between price and historical harvest returns, well demonstrated in these empirical examples, is particularly diagnostic of fishery degradation. Fortunately, the same heuristic can also be used to identify reliable targets for fishery restoration yielding optimal bioeconomic returns while safely conserving resource abundance.

Collective decision-making promotes fitness loss in a fusion-fission society.

While collective decision-making is recognised as a significant contributor to fitness in social species, the opposite outcome is also logically possible. We show that collective movement decisions guided by individual bison sharing faulty information about habitat quality promoted the use of ecological traps. The frequent, but short-lived, associations of bison with different spatial knowledge led to a population-wide shift from avoidance to selection of agricultural patches over 9 years in and around Prince Albert National Park, Canada. Bison were more likely to travel to an agricultural patch for the first time by following conspecifics already familiar with agricultural patches. Annual adult mortality increased by 12% due to hunting of bison on agricultural lands. Maladaptive social behaviour accordingly was a major force that contributed to a ~50% population decline in less than a decade. In human-altered landscapes, social learning by group-living species can lead to fitness losses, particularly in fusion-fission societies.

Daphnia inhibits the emergence of spatial pattern in a simple consumer-resource system.

Spatial self-organization can occur in many ecosystems with important effects on food web dynamics and the maintenance of biodiversity. The consumer-resource interaction is known to generate spatial patterning, but only a few empirical studies have investigated the effect of the consumer on resource distribution. Here we report results from a large aquatic mesocosm experiment used to investigate the effect of the consumer Daphnia magna on the distribution of its resource, the green algae Chlorella vulgaris. We maintained large tanks with capacity for 26 ,000 L with either algae or both algae and Daphnia in different temperature conditions. We found that the presence of D. magna inhibited spatial structure in algal distribution that arose as a consequence of increasing temperature. We conjecture that this homogenization effect might be caused by a combination of high mobility combined with high rates of algal consumption by Daphnia. Our study emphasizes the importance of both local constraints on growth and behavioral responses in either promoting or suppressing spatial self-organization in natural populations.

Space-use behaviour of woodland caribou based on a cognitive movement model.

Movement patterns offer a rich source of information on animal behaviour and the ecological significance of landscape attributes. This is especially useful for species occupying remote landscapes where direct behavioural observations are limited. In this study, we fit a mechanistic model of animal cognition and movement to GPS positional data of woodland caribou (Rangifer tarandus caribou; Gmelin 1788) collected over a wide range of ecological conditions. The model explicitly tracks individual animal informational state over space and time, with resulting parameter estimates that have direct cognitive and ecological meaning. Three biotic landscape attributes were hypothesized to motivate caribou movement: forage abundance (dietary digestible biomass), wolf (Canis lupus; Linnaeus, 1758) density and moose (Alces alces; Linnaeus, 1758) habitat. Wolves are the main predator of caribou in this system and moose are their primary prey. Resulting parameter estimates clearly indicated that forage abundance is an important driver of caribou movement patterns, with predator and moose avoidance often having a strong effect, but not for all individuals. From the cognitive perspective, our results support the notion that caribou rely on limited sensory inputs from their surroundings, as well as on long-term spatial memory, to make informed movement decisions. Our study demonstrates how sensory, memory and motion capacities may interact with ecological fitness covariates to influence movement decisions by free-ranging animals.

Wolves adapt territory size, not pack size to local habitat quality.

1. Although local variation in territorial predator density is often correlated with habitat quality, the causal mechanism underlying this frequently observed association is poorly understood and could stem from facultative adjustment in either group size or territory size. 2. To test between these alternative hypotheses, we used a novel statistical framework to construct a winter population-level utilization distribution for wolves (Canis lupus) in northern Ontario, which we then linked to a suite of environmental variables to determine factors influencing wolf space use. Next, we compared habitat quality metrics emerging from this analysis as well as an independent measure of prey abundance, with pack size and territory size to investigate which hypothesis was most supported by the data. 3. We show that wolf space use patterns were concentrated near deciduous, mixed deciduous/coniferous and disturbed forest stands favoured by moose (Alces alces), the predominant prey species in the diet of wolves in northern Ontario, and in proximity to linear corridors, including shorelines and road networks remaining from commercial forestry activities. 4. We then demonstrate that landscape metrics of wolf habitat quality - projected wolf use, probability of moose occupancy and proportion of preferred land cover classes - were inversely related to territory size but unrelated to pack size. 5. These results suggest that wolves in boreal ecosystems alter territory size, but not pack size, in response to local variation in habitat quality. This could be an adaptive strategy to balance trade-offs between territorial defence costs and energetic gains due to resource acquisition. That pack size was not responsive to habitat quality suggests that variation in group size is influenced by other factors such as intraspecific competition between wolf packs.

Towards an energetic landscape: broad-scale accelerometry in woodland caribou.

Energetic balance is a central driver of individual survival and population change, yet estimating energetic costs in free- and wide-ranging animals presents a significant challenge. Animal-borne activity monitors (using accelerometer technology) present a promising method of meeting this challenge and open new avenues for exploring energetics in natural settings. To determine the behaviours and estimated energetic costs associated with a given activity level, three captive reindeer (Rangifer tarandus tarandus) at the Toronto Zoo were fitted with collars and observed for 53 h. Activity patterns were then measured over 13 months for 131 free-ranging woodland caribou (R. t. caribou) spanning 450,000 km(2) in northern Ontario. The captive study revealed a positive but decelerating relationship between activity level and energetic costs inferred from previous behavioural studies. Field-based measures of activity were modelled against individual displacement, vegetation abundance (Normalized Difference Vegetation Index), snow depth and temperature, and the best fit model included all parameters and explained over half of the variation in the data. Individual displacement was positively related to activity levels, suggesting that broad differences in energetic demands are influenced by variation in movement rates. After accounting for displacement, activity was highest at intermediate levels of vegetation abundance, presumably due to foraging behaviour. Snow depth, probably associated with digging for winter forage, moderately increased activity. Activity levels increased significantly at the coldest winter temperatures, suggesting the use of behavioural thermoregulation by caribou. These interpretations of proximate causal factors should be regarded as hypotheses subject to validation under normal field conditions. These results illustrate the landscape characteristics that increase energetic demands for caribou and confirm the great potential for the use of accelerometry in studies of animal energetics.

Transgenerational epigenetic inheritance increases trait variation but is not adaptive.

Understanding processes that can produce adaptive phenotypic shifts in response to rapid environmental change is critical to reducing biodiversity loss. The ubiquity of environmentally induced epigenetic marks has led to speculation that epigenetic inheritance could potentially enhance population persistence in response to environmental change. Yet, the magnitude and fitness consequences of epigenetic marks carried beyond maternal inheritance are largely unknown. Here, we tested how transgenerational epigenetic inheritance (TEI) shapes the phenotypic response of Daphnia clones to the environmental stressor Microcystis. We split individuals from each of eight genotypes into exposure and control treatments (F0 generation) and tracked the fitness of their descendants to the F3 generation. We found transgenerational epigenetic exposure to Microcystis led to reduced rates of survival and individual growth and no consistent effect on offspring production. Increase in trait variance in the F3 relative to F0 generations suggests potential for heritable bet hedging driven by TEI, which could impact population dynamics. Our findings are counter to the working hypothesis that TEI is a generally adaptive mechanism likely to prevent extinction for populations inhabiting rapidly changing environments.

Asynchronous food-web pathways could buffer the response of Serengeti predators to El Niño Southern Oscillation.

Understanding how entire ecosystems maintain stability in the face of climatic and human disturbance is one of the most fundamental challenges in ecology. Theory suggests that a crucial factor determining the degree of ecosystem stability is simply the degree of synchrony with which different species in ecological food webs respond to environmental stochasticity. Ecosystems in which all food-web pathways are affected similarly by external disturbance should amplify variability in top carnivore abundance over time due to population interactions, whereas ecosystems in which a large fraction of pathways are nonresponsive or even inversely responsive to external disturbance will have more constant levels of abundance at upper trophic levels. To test the mechanism underlying this hypothesis, we used over half a century of demographic data for multiple species in the Serengeti (Tanzania) ecosystem to measure the degree of synchrony to variation imposed by an external environmental driver, the El Niño Southern Oscillation (ENSO). ENSO effects were mediated largely via changes in dry-season vs. wet-season rainfall and consequent changes in vegetation availability, propagating via bottom-up effects to higher levels of the Serengeti food web to influence herbivores, predators and parasites. Some species in the Serengeti food web responded to the influence of ENSO in opposite ways, whereas other species were insensitive to variation in ENSO. Although far from conclusive, our results suggest that a diffuse mixture of herbivore responses could help buffer top carnivores, such as Serengeti lions, from variability in climate. Future global climate changes that favor some pathways over others, however, could alter the effectiveness of such processes in the future.

Environmental and individual drivers of animal movement patterns across a wide geographical gradient.

Within the rapidly developing field of movement ecology, much attention has been given to studying the movement of individuals within a subset of their population's occupied range. Our understanding of the effects of landscape heterogeneity on animal movement is still fairly limited as it requires studying the movement of multiple individuals across a variety of environmental conditions. Gaining deeper understanding of the environmental drivers of movement is a crucial component of predictive models of population spread and habitat selection and may help inform management and conservation. In Ontario, woodland caribou (Rangifer tarandus caribou) occur along a wide geographical gradient ranging from the boreal forest to the Hudson Bay floodplains. We used high-resolution GPS data, collected from 114 individuals across a 450000 km(2) area in northern Ontario, to link movement behaviour to underlying local environmental variables associated with habitat permeability, predation risk and forage availability. We show that a great deal of observed variability in movement patterns across space and time can be attributed to local environmental conditions, with residual individual differences that may reflect spatial population structure. We discuss our results in the context of current knowledge of movement and caribou ecology and highlight potential applications of our approach to the study of wide-ranging animals.

Group formation stabilizes predator-prey dynamics.

Theoretical ecology is largely founded on the principle of mass action, in which uncoordinated populations of predators and prey move in a random and well-mixed fashion across a featureless landscape. The conceptual core of this body of theory is the functional response, predicting the rate of prey consumption by individual predators as a function of predator and/or prey densities. This assumption is seriously violated in many ecosystems in which predators and/or prey form social groups. Here we develop a new set of group-dependent functional responses to consider the ecological implications of sociality and apply the model to the Serengeti ecosystem. All of the prey species typically captured by Serengeti lions (Panthera leo) are gregarious, exhibiting nonlinear relationships between prey-group density and population density. The observed patterns of group formation profoundly reduce food intake rates below the levels expected under random mixing, having as strong an impact on intake rates as the seasonal migratory behaviour of the herbivores. A dynamical system model parameterized for the Serengeti ecosystem (using wildebeest (Connochaetes taurinus) as a well-studied example) shows that grouping strongly stabilizes interactions between lions and wildebeest. Our results suggest that social groups rather than individuals are the basic building blocks around which predator-prey interactions should be modelled and that group formation may provide the underlying stability of many ecosystems.

Changes in movement, habitat use, and response to human disturbance accompany parturition events in bighorn sheep (Ovis canadensis).

Parturition and the early neonatal period are critical life history stages in ungulates with considerable implications for population growth and persistence. Understanding the changes in behaviour induced by ungulate parturition is important for supporting effective population management, but reliably identifying birth sites and dates presents a challenge for managers. Rocky Mountain bighorn sheep (Ovis canadensis canadensis) are one such highly valued and ecologically important species in montane and subalpine ecosystems of Western North America. In the face of changing patterns of anthropogenic land use, wildlife managers increasingly require site-specific knowledge of the movement and habitat selection characteristics of periparturient sheep to better inform land use planning initiatives and ensure adequate protections for lambing habitat. We used movement data from GPS collared parturient (n = 13) and non-parturient (n = 8) bighorn sheep in Banff National Park, Canada to (1) identify lambing events based on changes in key movement metrics, and (2) investigate how resource selection and responses to human use change during the periparturient period. We fit a hidden Markov model (HMM) to a multivariate characterization of sheep movement (step length, daily home range area, residence time) to predict realistic lambing dates for the animals in our study system. Leave-one-out cross validation of our model resulted in a 93% success rate for parturient females. Our model, which we parameterized using data from known parturient females, also predicted lambing events in 25% of known non-parturient ewes in a test dataset. Using a latent selection difference function and resource selection functions, we tested for postpartum changes in habitat use, as well as seasonal differences in habitat selection. Immediately following lambing, ewes preferentially selected high-elevation sites on solar aspects that were more rugged, closer to escape terrain, and further from roads. Within-home range habitat selection was similar between individuals in different reproductive states, but parturient ewes had stronger selection for low snow depth, sites closer to barren ground, and sites further from trails. We propose that movement-based approaches such as HMMs are a valuable tool for identifying critical parturition habitat in species with complex movement patterns and may have particular utility in study areas without access to extensive field observations or vaginal implant transmitters. Furthermore, our results suggest that managers should minimize human disturbance in lambing areas to avoid interfering with maternal behaviour and ensure access to a broad range of suitable habitat in the periparturient period.

Intraspecific genetic variation is critical to robust toxicological predictions of aquatic contaminants.

Environmental risk assessment is a critical tool for protecting aquatic life and its effectiveness is predicated on predicting how natural populations respond to contaminants. Yet, routine toxicity testing typically examines only one genotype, which may render risk assessments inaccurate as populations are most often composed of genetically distinct individuals. To determine the importance of intraspecific variation in the translation of toxicity testing to populations, we quantified the magnitude of genetic variation within 20 Daphnia magna clones derived from one lake using whole genome sequencing and phenotypic assays. We repeated these assays across two exposure levels of microcystins, a cosmopolitan and lethal aquatic contaminant produced by harmful algal blooms. We found considerable intraspecific genetic variation in survival, growth, and reproduction, which was amplified by microcystins exposure. Finally, using simulations we demonstrate that the common practice of employing a single genotype to calculate toxicity tolerance failed to produce an estimate within the 95% confidence interval over half of the time. These results illuminate the importance of incorporating intraspecific genetic variation into toxicity testing to reliably predict how natural populations will respond to aquatic contaminants.