Collision Risk Models (CRMs) are central to environmental assessment for offshore wind but operate under data limitations and uncertain ecological assumptions. While stochastic extensions incorporate parameter uncertainty, the basic approach remains structurally unchanged, treating collision risk as a function of uncertain inputs rather than as the outcome of underlying behavioural and spatial processes. This paper argues for a shift from parameter estimation to simulation. Agent-based models (ABMs) provide a necessary extension to CRMs by simulating bird movement, behavioural responses, and interactions with turbine arrays. In this framework, collision risk emerges from generative processes rather than being imposed through fixed parameters such as avoidance rates. ABMs complement deterministic and stochastic CRMs, support scenario testing, generate synthetic data under data scarcity, and provide mechanistic insight into collision risk. ABMs have not been integrated into regulatory collision risk assessment but doing so would enhance the biological realism of collision modelling, offering a more robust basis for decision-making in offshore wind consenting.
Forests are entering an era in which future climate, disturbance regimes, and management conditions may have no close historical analogues, while society increasingly demands the simultaneous provision of timber, carbon sequestration, biodiversity conservation, and other ecosystem services. Supporting sustainable, climate-smart forest management therefore requires predictive tools capable of anticipating ecosystem responses under novel environmental conditions. Recent advances in artificial intelligence (AI), remote sensing, and large forest databases have revolutionized forest monitoring and prediction. The rapid expansion of satellite observations, sensor networks, forest inventories, and automated monitoring programs has generated unprecedented volumes of data, allowing AI to greatly improve pattern recognition, forecasting, and decision support within historically observed, data-rich domains. However, the principal limitation of purely data-driven models lies in their limited transferability to environmental conditions beyond those represented in the training data. As forests experience novel combinations of climate, disturbances, species interactions, and management practices, statistical relationships derived from historical observations may no longer remain valid, compromising extrapolative predictions. Alternatively, process-based models (PBMs) provide a complementary approach by explicitly representing the physiological and ecological mechanisms governing forest dynamics, making them better suited for exploring future scenarios outside the historical record. Therefore, integrating these mechanistic models with AI tools and Internet of Things (IoT) technologies offers a promising pathway toward next-generation hybrid models and forest Digital Twins that combine pattern-recognition capabilities with causal ecological understanding. Achieving this vision will require continued development of process-based models, improved hybrid model architectures, standardized datasets, and rigorous uncertainty analysis to support robust, climate-smart, multi-objective forest management.
Global environmental scenarios, such as the Shared Socioeconomic Pathways (SSPs), are central to decision-making under deep uncertainty but have been criticized for their lack of alternative economic futures. This study addresses these limitations by introducing selected SSPs into the Integrated Assessment Model MORDRED, which endogenizes economic, demographic, and biophysical variables through system-dynamic feedback relationships. By re-quantifying SSP2, SSP3, and SSP5 under two boundary conditions —without (SSP′) and with (SSP″) biophysical feedbacks —it was examined how the incorporation of biophysical feedbacks into economic system analyses affects the plausibility and viability of the SSP storylines. While the SSP′ implementations reproduce dynamics projected by previous SSP quantification exercises, introducing biophysical constraints in SSP″ variations generates nonlinear dynamics and historical discontinuities in socio-economic and environmental trajectories which are absent from, and incompatible with, the original SSPs. Projected developments include stagnation or collapse in final demand, demographic contractions, and major shifts in energy system evolution before the end of the 21st century. A positive feedback loop between environmental degradation, reduced productive capacity and declining energy efficiency emerges as a key driver of systemic instability. These findings challenge the implicit assumption of perpetual economic growth underpinning many global environmental scenarios and reveal that futures featuring economic stagnation and collapse are not only possible but plausible under less optimistic boundary conditions. Thus, current SSP-based assessments underestimate both the risks of climate–economy interactions and the scope of required institutional transformations to ensure societal resilience under biophysical constraints.
Spatial sampling error remains a critical yet underappreciated challenge in abundance estimation, particularly when sampling processes are not explicitly accounted for. A persistent belief holds that increasing sample size compensates for sampling bias, while overlooking how spatial sampling processes — such as preferential or convenience sampling — shape these errors.Relying on the decomposition of sampling error into components driven by sample selectivity, incompleteness, and statistical population heterogeneity, we derive analytical results linking characteristics of the realized sample to properties of the underlying sampling process.Using simulated examples, we illustrate how five sampling processes — four archetypal (equal-propensity, preferential, spatial-spread, and road-based sampling) and one composite opportunistic process combining several of these mechanisms — influence the reliability of abundance estimates.Our results show how increasing sample size alone — a common reflex in monitoring — improves precision but fails to correct for bias caused by selective sampling. This is particularly relevant for citizen science, where massive datasets may yield precise yet systematically misleading inferences if spatial sampling processes are ignored.We advocate for a shift from a reliance on sample size toward the explicit characterization of sampling processes and the prioritization of probability-based sampling designs where feasible. By treating sampling error as a process-driven phenomenon, this work provides a contribution toward more rigorous and transparent ecological inference.The resulting framework is presented from a methodological perspective, specifically targeting quantitative ecologists, biometricians, and wildlife statisticians involved in designing studies, implementing monitoring programs, or analyzing citizen science data.
Periods of low dissolved oxygen concentration — hypoxia and anoxia — threaten the health of aquatic ecosystems and the services they provide. Hypoxia is strongly influenced by temperature, but the different sensitivities and response functions of oxygen removal and production processes to temperature are not regarded in most models. Here we present OxyPOM—Oxygen and Particulate Organic Matter, a temperature-aware process-based biogeochemical model. OxyPOM incorporates process-specific temperature sensitivities for the key oxygen-related processes: photosynthesis, re-aeration, respiration, mineralization, and nitrification. Other temperature-sensitive variables, such as optimal light intensity, winter grazing inhibition, and pathogenesis, are also represented. Our model was tested in an idealized water-column experiment that represents a typical estuary with a seasonal low-oxygen environment. Differences between process-specific and uniform temperature sensitivities affect seasonal patterns, resulting in higher divergence between the two scenarios, particularly in summer months. While these changes may balance in the overall annual oxygen budget, uniform sensitivities may underestimate particulate organic carbon production by up to a factor of four over the year and overestimate nutrient concentrations. This nuanced approach to temperature sensitivity allows us to explore and test new hypotheses related to climate warming and heatwaves, addressing the ecosystem changes required by climate change models.
The push and pull is an insect pest control strategy which combines attractants and repellents to manipulate the spatial distribution of the populations. Recent studies have proven its efficacy and have identified this technique as an auspicious alternative to traditional agrochemicals. Although being promising, its implementation still faces significant challenges as, for instance, the spatial and temporal deployment of control agents. Mathematical modelling could play a pivotal role in such sense, as it might reproduce different trap/repellent configurations in silico. This study aims to apply a generalised spatially-explicit model based on ordinary differential equations to simulate the effect of traps and repellents on insect pest populations. The theoretical framework, as well as the model behaviour, has been introduced with reference to the case study of the olive fruit fly Bactrocera oleae. Eight spatial arrangements of traps and repellents were compared to further assess the model behaviour and their impact in reducing the adult population. Simulations indicate that increasing the density of control agents enhances effectiveness up to a threshold, and that the spatial configuration of push and pull components plays a critical role on control success. The model hereby introduced lays the groundwork for further formulations of optimal control problems aimed to optimise the push and pull strategy within pest management programmes, which so far are still driven by empirical methods.
Despite continuous reductions in external phosphorus (P) loading under the Great Lakes Water Quality Agreement (GLWQA), nuisance blooms of Cladophora continue to proliferate in the nearshore zones of the lower Great Lakes. This decoupling between nutrient load reductions and benthic algal responses highlights knowledge gaps regarding the interacting physical, chemical, and biological controls governing nearshore productivity. In this study, we develop a process-based nearshore ecosystem model that integrates P cycling, pelagic phytoplankton, dreissenid mussels, and Cladophora within a single framework. The model is evaluated using a combination of Morris Screening and Classification and Regression Tree (CART) analyses to identify dominant drivers, nonlinear responses, and interactions relevant to nearshore management. Sensitivity analyses reveal that Cladophora biomass is not controlled by nutrient supply alone but emerges from strong feedback among dreissenid-mediated nutrient recycling, light availability, and hydrodynamic retention. Dreissenid filtration, excretion, and ingestion parameters consistently have a strong influence on P partitioning and benthic-pelagic coupling, while light attenuation and residence time determine whether favourable nutrient conditions lead to high benthic algal biomass. Consistent with nearshore observations from Lake Ontario, CART analyses further demonstrate that high Cladophora biomass can persist even under oligotrophic conditions when pelagic shading is minimized, and internal P recycling is sustained. Scenario analyses show that even modest shifts in mussel density, thermal physiology of dreissenids, or nutrient bioavailability can substantially reconfigure P flux pathways and alter the timing and magnitude of Cladophora blooms, even without discernible changes in external loading. Collectively, these findings indicate that continued reliance on external nutrient reductions to address nearshore nuisance Cladophora proliferation may not be sufficient. Effective management strategies must account for biological feedback loops, hydrodynamics, and internal nutrient recycling processes that can refuel the nearshore autotrophic assemblage.
Dynamic Energy Budget (DEB) theory’s tetrahedron of life links the inner to the outer world of individuals, beyond energy and mass conservation. It follows mathematically from the some assumptions behind DEB theory, without using new assumptions and links the supply stress, the fraction of assimilation to reproduction and the minimum scaled functional response to reach puberty all to the fraction of mobilized reserve to soma: the cornerstone of life. The supply stress quantifies the supply–demand spectrum: supply-species eat what is available, while demand-species eat what they need. The minimum scaled functional response to reach puberty increases with the supply stress, which implies that the half-saturation coefficient needs to decrease with the supply stress to ensure that the feeding rate is close to its maximum for species near the demand end of the supply–demand spectrum. This coefficient is the ratio of the specific ingestion and food searching rates, so the food searching rate needs to increase with the supply stress. This is in harmony with the recent finding that the factorial aerobic scope increases from 3.2 for zero supply stress to 32 for maximum supply stress; a high food searching rate requires a high peak metabolism. It turns out that the supply stress needs to be low for the allocation to reproduction to be large. The supply stress has direct links with the trophic position and the precociality index, i.e. the ratio of the maturity levels at birth and puberty. The lower the fraction of assimilation to reproduction, the lower the life-time cumulated neonate mass production as well as the precociality index within a variety of taxa within the Add_my_Pet (AmP) collection of energetic data and DEB parameters. The intra-class variation of the precociality index with the fraction of assimilation to reproduction turns out to differ from the inter-class variation. Invertebrates were found to have a low supply stress, have a low trophic position, allocate a large fraction of their assimilation to reproduction and make relatively small eggs. Tetrapods, especially birds and mammals, have a high supply stress, a high trophic position, invest little in reproduction and their neonates are relatively large. This coupling of traits follows from the structure of DEB theory and is empirically confirmed by the 7334 animal species in the AmP collection for a wide range of taxa. Species with a high trophic position must have a food intake close to their maximum and cannot invest much in reproduction. The combined implications for feeding and reproduction avoids or reduces over-exploitation of resources in the food web and might contribute to the dynamic stability of food webs and to the maintenance of biodiversity in ecosystem models, challenging fitness maximization as central theme in evolution biology. The tetrahedron provides strong support for the central position of the supply–demand spectrum in animal energetics.
Spatial aggregation is widely recognized as a stabilizing feature of predator-prey systems, yet its selective consequences remain poorly understood. It is unclear whether aggregation is favored directly through encounter-rate effects or indirectly through its influence on ecological stability and proximity to instability across enrichment gradients. Using a spatially explicit, individual‑based Evolutionary Cellular Automaton (ECA), we examine how predator aggregation interacts with resource enrichment to shape dynamical regimes and evolutionary outcomes. Across 50 replicated simulations per enrichment level, we identify distinct ecological regimes ranging from persistent oscillatory coexistence to contingent exclusion and enrichment-driven collapse. Directional selection for aggregation is observed within a restricted regime characterized by low prey density and reduced ecological stability, where the system operates near a stability boundary. To clarify the structural origin of this pattern, we complement the simulations with a minimal analytical perspective showing that aggregation modifies the dominant eigenvalue of the ecological Jacobian. In predator-prey systems, increasing enrichment typically reduces local stability, potentially approaching a Hopf boundary. When aggregation increases stability—by dampening effective interaction strength—it reduces the amplification of environmental fluctuations. Because long‑term geometric growth is sensitive to variance, proximity to instability reshapes the adaptive landscape. Selection for aggregation therefore intensifies in regimes where ecological dynamics are weakly stable, but diminishes when enrichment either restores stability or drives collapse. Notably, this selective pattern is asymmetric across trophic levels: predator aggregation is favored under low effective resource availability, whereas prey aggregation shows no comparable directional selection. Importantly, directional selection does not require strong spatial covariance between predators and prey or extreme oscillatory amplitudes. Instead, it arises because enrichment alters the system's stability structure, thereby modifying how environmental variability translates into fitness differences between spatial strategies. These results demonstrate that the evolutionary consequences of predator aggregation are fundamentally regime‑dependent. Ecological stability is not merely an outcome of trait evolution but a structural mediator of selection. More broadly, our findings show that eco‑evolutionary dynamics in spatial predator-prey systems must be understood through stability‑mediated feedbacks across environmental gradients rather than solely through spatial variance.
Phytoplankton blooms can affect lake ecosystem services, particularly in large lakes where their impacts are often localized in space and time. To anticipate areas at risk and support management responses, an operational framework for monitoring and forecasting phytoplankton blooms is proposed and applied to the large deep Lake Geneva (France/Switzerland). This alert system integrates in situ observations, satellite chlorophyll-a (Chl-a) and Secchi depth imagery, and three-dimensional hydrodynamic simulations within an operational workflow. The monitoring module analyzes near-real-time satellite Chl-a to identify areas exceeding the World Health Organization risk threshold of 10 µg Chl-a L-¹, triggering local or global alerts indicative of surface bloom conditions potentially affecting water uses. The forecasting module predicts short-term surface bloom transport by advecting virtual particles within the hydrodynamic model. Model performance was evaluated using two historical surface bloom events, showing good agreement between observations and forecasts, with mean trajectory errors of 0.3-1.7 km and maximum errors below 6.1 km. Forecast skill decreased beyond three days and during bloom initiation or decay phases. To address the limitations of satellite observations, a third module detects deep Chl-a exceedances above 10 µg L-¹ not visible by satellite. This module relies on a Random Forest machine learning model trained on a 10-year in situ dataset from the central lake station and successfully detects 78% of deep bloom events. By enabling daily monitoring and forecasting, the alert system complements conventional in situ programs and helps identify short-lived or spatially localized blooms. Its modular, open-access design facilitates transferability to other lakes.
Sigmoidal functions are widely used in ecology to represent bounded growth, saturation, and asymptotic responses, yet curve form alone does not specify the ecological state relations that generate such trajectories. This paper revisits Klem’s 1933 yeast experiments, which historically contributed to the introduction of sigmoidal reasoning into fisheries surplus-production theory. The experiments are reconstructed through a minimal algebraic and resource-explicit approach that distinguishes yeast biomass, remaining nutrient resource, a biomass-equivalent growth basis, a restriction factor, a realized carrying structure, and the accessible basis remaining for further growth. From these relations, a constrained growth recursion is derived and compared with the discrete logistic form. The reconstruction shows that sigmoid-like bounded growth is reproduced when the biomass-equivalent basis is fixed, the restriction regime remains effectively stationary, and the accessible growth basis declines as biomass approaches the realized upper bound. Under these conditions, the recursion can be written in fixed-K logistic notation. Under recurrent nutrient supply, the biomass-equivalent basis and realized carrying structure change through time, and the same fixed-K interpretation no longer applies. The analysis further distinguishes the state-defined structure of growth from its interval-specific realization: α scales the transition between observed states without prescribing an explicit temporal growth trajectory. The contribution of the reconstruction is therefore a clarification of scope. Logistic notation is most directly ecologically interpretable when the state relations compressed into K remain sufficiently stationary, a condition that is relevant to both ecological growth models and fisheries surplus-production reasoning.
Accurate prediction of forest structure and composition is essential for forest management, as planning depends on anticipating future stand dynamics under changing environmental conditions. Machine learning algorithms have proven useful for this task due to their ability to capture non-linear ecological patterns in forest dynamics. However, current machine learning applications in forest ecology mostly focus on component growth rates, leaving approaches that directly predict future stand-level states like total basal area and species composition under-explored. This study develops a national-scale XGBoost model for Swiss forests, using Swiss National Forest Inventory data and CH2018 climate scenarios to predict changes in basal area and broadleaf proportion. Tested on a later inventory not used for training, XGBoost predicts the next inventory state with an R2 of 0.66 for basal area and 0.91 for broadleaf proportion, slightly above a linear Lasso model (0.65 and 0.90). When applied recursively to generate exploratory projections to 2099, the model indicates an overall increase in basal area, although unmanaged lower-elevation belts decrease, and an increase in broadleaf proportion mainly at lower elevations, especially under high emissions. These trends are similar to patterns reported by established process-based and empirical forest models. This work offers a computationally efficient approach for projecting forest dynamics over large areas, without the extensive parameterization required by process-based models.
The endemic Mediterranean bivalve Pinna nobilis has experienced severe population declines, with pathogens playing a crucial role. Haplosporidium pinnae and Mycobacterium spp. have been implicated in mass mortality events, with Vibrio spp. intensifying their impact. Here, we applied a Dynamic Energy Budget (DEB) model to assess how pathogen-induced stress, temperature, and food availability combined affect energy allocation and life history traits. We simulated pathogen effects through physiological modes of action that target maintenance and assimilation. We investigated their impact separately and combined under constant and seasonal regimes of temperature and food availability. Results show that increases in maintenance costs or reductions in assimilation decrease energetic scope, with assimilation impairment having a greater impact. Their combination further amplifies energetic deficits, particularly under high temperature and low food availability. These impacts ultimately result in delayed sexual maturation and reduced fecundity. We introduced the reproduction reserve depletion time (RRDT) as a proxy of loss of resilience under infection stress. Empirical survivor functions of RRDT, derived from simulations across ten age cohorts with varying infection onset timing and stress intensity, show that resilience is strongly affected by temperature and food availability, with temperature becoming a secondary driver under food limitation. Moreover, simulations using temperature and Chl-a data from two Greek regions, Thermaikos and Maliakos Gulfs, reveal that seasonal timing of infection emerges as a critical determinant of outcomes, with infections initiated in spring–summer leading to rapid reserve depletion. Our findings provide a mechanistic basis for interpreting patterns observed in remnant and recovering populations.
Efforts to restore freshwater habitats and support aquatic ecosystems require substantial use of resources, yet the success of such restoration efforts are often fraught with uncertainty. Formal (so-called “structured”) decision-making frameworks may help coordinate and organize complex watershed-scale efforts, prioritize conservation actions, and improve the likelihood of significant ecological benefits. Here, we focus on the federally endangered Sacramento River winter-run Chinook salmon (Oncorhynchus tshawytscha), developing a structured decision-making framework that uses optimization modeling to suggest restoration portfolios that seek to recover adult returns under a variety of available resources and hydrologies. The framework uses an expansion of the Winter-run Habitat-based Population model (WRHAP), for which additional sub-models were developed (ocean-stage, hatchery and trap-and-haul operations, reintroduction plans and recovery actions), to simulate population responses to restoration strategies (WRHAP-SEA). Defined portfolios of actions, particularly those targeting multiple stressors at different locations (e.g., enhanced off-channel access, weir notching), improved adult returns (∼3000 to ∼8000), population spatial structure, and replacement rates (0.69 to >0.85), reducing winter-run Chinook extinction risk. However, a self-sustaining and low-risk of extinction Sacramento River population was not achieved with the suite of restoration actions considered (CRRH <0.9; c^ > 90%). Results showed that frequent and sustained floodplain activation was key to improve recruitment, re-introduction programs provided important population diversity but relied on a strong Sacramento River population to avoid demographic risks, and successful two-way trap and haul programs required high capture efficiency (≥70%) at top-of-reservoir traps and low delayed mortality for transported juveniles. The developed framework, due to its assumptions and simplifications, is not intended to advocate for a specific restoration portfolio, but rather to illustrate the ability of integrated ecological modeling/optimization approaches to help organize, explore, and represent interactions among a broad set of restoration options at the watershed-scale.
Designing cost-effective Marine Protected Area (MPA) networks that balance biodiversity conservation, ecological connectivity, and socioeconomic feasibility presents significant computational challenges, particularly at basin scales involving thousands of candidate sites. Traditional systematic conservation planning approaches either rely on simplified spatial prioritization metrics or face computational intractability when incorporating explicit connectivity constraints. We present a novel computational framework integrating Graph Convolutional Networks (GCN) with constraint-based greedy optimization to address this multi-objective problem. The GCN learns compact 16-dimensional site embeddings from a spatial connectivity graph representing 10,741 candidate sites (30 km × 30 km resolution), compressing biodiversity value, cost-efficiency, and network topology into an interpretable representation for exploratory analysis of site trade-offs. A two-phase greedy algorithm, operating directly on the underlying connectivity matrix, first satisfies mandatory ecological constraints (threatened species protection, bathymetric representation, regional equity), then maximizes marginal efficiency defined as biodiversity and connectivity gains per unit cost. We demonstrate the framework using 84,721 species observations covering 120 marine taxa across the Mediterranean Sea, combined with bathymetric data and spatially explicit opportunity costs. For a 15% coverage target (1611 sites, 1449,900 km²), our approach achieves network connectivity 4.4-fold higher than random selection and 1.6-fold higher than biodiversity hotspot strategies, at total cost of €1.69 billion—42% below random selection and 58% below hotspot approaches—while guaranteeing protection of all nine threatened species (four Critically Endangered, three Endangered, two Vulnerable) alongside near-complete overall coverage (99.2% of 120 species), a guarantee not met by cost-minimizing alternatives. All ecological constraints are satisfied while maintaining computational efficiency (8–15min runtime). Network configuration remains robust across coverage-target scenarios (10%, 15%, 30%): species coverage remains stable at 99.2%, and regional distribution varies by <2 percentage points across scenarios. The modular framework is transferable to other marine regions through substitution of region-specific biodiversity, environmental, and socioeconomic data, supporting systematic conservation planning where computational efficiency is essential for stakeholder engagement.
