Assessing adaptive phenotypic plasticity by means of conditional strategies from empirical data: the latent environmental threshold model

Conditional strategies are the most common form of discrete phenotypic plasticity. In a conditional strategy, the phenotype expressed by an organism is determined by the difference between an environmental cue and a threshold, both of which may vary among individuals. The environmental threshold model (ETM) has been proposed as a mean to understand the evolution of conditional strategies, but has been surprisingly seldom applied to empirical studies. A hindrance for the application of the ETM is that often, the proximate cue triggering the phenotypic expression and the individual threshold are not measurable, and can only be assessed using a related observable cue. We describe a new statistical model that can be applied in this common situation. The Latent ETM (LETM) allows for a measurement error in the phenotypic expression of the individual environmental cue and a purely genetically determined threshold. We show that coupling our model with quantitative genetic methods allows an evolutionary approach including an estimation of the heritability of conditional strategies. We evaluate the performance of the LETM with a simulation study and illustrate its utility by applying it to empirical data on the size-dependent smolting process for stream-dwelling Atlantic salmon juveniles.     

Modeling spatial dynamics of episodic and synchronous reproduction by plant populations: the effect of small-scale pollen coupling and large-scale climate

Many plant species show masting, intermittent and synchronized reproduction at population level. In the present paper, we review the resource-based model providing a theoretically plausible physiological mechanism underlying masting. In the model, a non-linear allocation of energy reserves is considered: plants accumulate photosynthate every year, produce flowers when the energy reserve level exceeds a threshold, and set seeds at a rate limited by pollen availability. The model predicted that individual plants alter their reproductive dynamics from annual to intermittent depending on how heavily the plant invests resource in reproduction. When fruit production is limited by the availability of outcross pollen, a plant population showed diverse reproductive behavior such as completely synchronized or desynchronized reproduction. Spatial scale of reproductive synchrony tended to be a few times larger than the range of direct pollen exchange. Impact of climatic fluctuation correlated at a large spatial scale was also investigated as an alternative synchronizing factor. The variation in annual productivity and the reproductive threshold induced from climatic fluctuation was accounted for by incorporating an additional term in the model. When plants show a 2 year reproductive cycle, highly synchronized reproduction at a regional scale was induced due to correlated environmental forcing, but reproductive synchrony with long intermast periods was realized only when pollen coupling and environmental forcing were at work. These results suggest that distance-dependent processes, such as pollen exchange between nearby trees, induce synchrony at a local scale and external environmental forcing correlated at geographically large scales works to strengthen and maintain such a synchrony.     

A Model for Individual Tree Development Based on Physiological Processes

Abstract: A tree growth model is presented which calculates the 3D development of trees and stands in dependence on their individual carbon, water and nitrogen balance. The availability of energy, soil water and nutrients is estimated from field data at the scale of crown and root system fractions, taking into account the individual neighbourhood. The model includes a simple estimation of radiation distribution and the simulation of carbon and nutrient exchange. Senescence is represented by compartment-specific turnover rates. Allocation of carbon and nitrogen into foliage, fine roots, branches, coarse roots, and the stem is calculated according to functional balance and pipe model principles. Dimensional changes are calculated annually according to the distribution of net assimilation. The model describes tree development as a response to individual environmental conditions and changes environmental conditions with individual tree development. Due to this feedback loop, environmental influences can be assessed in any kind of species mixture or stand structure. Furthermore, the physiological-based approach ensures that the model can be used for investigations of complex environmental changes, e.g. CO₂ concentration, precipitation, temperature and nitrogen deposition. Thus, it is particularly suitable to analyse field investigations and to support the cognition process on the ecology of forests. It could also be used, however, to estimate forest responses to given environmental scenarios.     

A multilevel latent Markov model for the evaluation of nursing homes' performance

The periodic evaluation of health care services is a primary concern for many institutions. We consider services provided by nursing homes with the aim of ranking a set of these structures with respect to their effect on resident health status. Since the overall health status is not directly observable, and given the longitudinal and multilevel structure of the available data, we rely on latent variable models and, in particular, on a multilevel latent Markov model where residents and nursing homes are the first and the second level units, respectively. The model includes individual covariates to account for resident characteristics. The impact of nursing home membership is modelled through a pair of random effects affecting the initial distribution and the transition probabilities between different levels of health status. Through the prediction of these random effects we obtain a ranking of the nursing homes. Furthermore, the proposed model accounts for nonignorable dropout due to resident death, which typically occurs in these contexts. The motivating dataset is gathered from the Long Term Care Facilities programme, a health care protocol implemented in Umbria (Italy). Our results show that differences in performance between nursing homes are statistically significant.     

Why is body size more variable in stressful conditions: an analysis of a potential proximate mechanism

Patterns of variability in quantitative traits across environmental gradients have received relatively little attention in evolutionary ecology. A recent meta-analysis showed that relative phenotypic variability in body size tends to decrease with improving environmental conditions. This pattern was explained by introducing the concept of upper threshold size to a general optimality model of individual growth but alternative explanations certainly exist. In particular, it is frequently observed in insects that variability in individual growth rates decreases with improving environmental conditions. Here we explore the effect of this phenomenon on environment-specific variability in adult sizes. A quantitative model shows that relative variability in adult sizes is independent of environmental quality if absolute variability in growth rates remains constant across the gradient of environmental quality. Deviations from this borderline case are definitely realistic in both directions. Both negative and positive relationships between relative variability of body size and environmental quality can thus be predicted to arise as a consequence of environment-specific variability in growth rates. The variability itself can be both genetic or environmental in its nature. We present empirical data which support both the assumptions and conclusions of our model-based analysis, as well as emphasize the advantages of controlled experiments for understanding the proximate sources of phenotypic variance.     

Adaptive foraging for simulated and real robotic swarms: the dynamical response threshold approach

Developing self-organised swarm systems capable of adapting to environmental changes as well as to dynamic situations is a complex challenge. An efficient labour division model, with the ability to regulate the distribution of work among swarm robots, is an important element of this kind of system. This paper extends the popular response threshold model and proposes a new adaptive response threshold model (ARTM). Experiments were carried out in simulation and in real-robot scenarios with the aim of studying the performance of this new adaptive model. Results presented in this paper verify that the extended approach improves on the adaptability of previous systems. For example, by reducing collision duration among robots in foraging missions, our approach helps small swarms of robots to adapt more efficiently to changing environments, thus increasing their self-sustainability (survival rate). Finally, we propose a minimal version of ARTM, which is derived from the conclusions drawn through real-robot and simulation results.     

Assessing the accuracy of species distribution models to predict amphibian species richness patterns

1. Evaluating the distribution of species richness where biodiversity is high but has been insufficiently sampled is not an easy task. Species distribution modelling has become a useful approach for predicting their ranges, based on the relationships between species records and environmental variables. Overlapping predictions of individual distributions could be a useful strategy for obtaining estimates of species richness and composition in a region, but these estimates should be evaluated using a proper validation process, which compares the predicted richness values and composition with accurate data from independent sources. 2. In this study, we propose a simple approach to estimate model performance for several distributional predictions generated simultaneously. This approach is particularly suitable when species distribution modelling techniques that require only presence data are used. 3. The individual distributions for the 370 known amphibian species of Mexico were predicted using maxent to model data on their known presence (66,113 presence-only records). Distributions were subsequently overlapped to obtain a prediction of species richness. Accuracy was assessed by comparing the overall species richness values predicted for the region with observed and predicted values from 118 well-surveyed sites, each with an area of c. 100 km(2), which were identified using species accumulation curves and nonparametric estimators. 4. The derived models revealed a remarkable heterogeneity of species richness across the country, provided information about species composition per site and allowed us to obtain a measure of the spatial distribution of prediction errors. Examining the magnitude and location of model inaccuracies, as well as separately assessing errors of both commission and omission, highlights the inaccuracy of the predictions of species distribution models and the need to provide measures of uncertainty along with the model results. 5. The combination of a species distribution modelling method like maxent and species richness estimators offers a useful tool for identifying when the overall pattern provided by all model predictions might be representing the geographical patterns of species richness and composition, regardless of the particular quality or accuracy of the predictions for each individual species.     

Environmental sensitivity and heterosis for egg laying in Drosophila melanogaster

Summary Genotype × temperature interactions for egg laying were studied in Drosophila melanogaster using two sets of half diallel crosses: one between inbred lines of the same geographic origin, and the other between established laboratory, newly derived inbred lines from different geographic origins. The sensitivity of most genotypes to changes in temperature was adequately described as a linear regression of mean in temperature. The regression coefficients (linear sensitivities) were heterogeneous between genotypes. Hybrids were more affected by temperature variation than were inbreds. All the heterogeneity of linear sensitivities was accounted for by a linear function of the genotypic means, which strongly suggests that a scale effect is responsible for the differences in sensitivity to temperature. In contrast, no general relationship was found between standard error deviation (sensitivity to small environmental changes) and mean performance between genotypes, although hybrids tended to be less variable than inbreds. This shows that the sensitivity to environmental variation depends not only on the genotype, but also on the nature of the environmental variation. The variability within temperatures may be affected by the general homeostasis of individual genotypes, while the variability between temperatures could be the result of genes directly affecting the trait and their multiplicative interaction with the environment.     

Machine learning based modeling of households: A regionalized bottom‐up approach to investigate consumption‐induced environmental impacts

As major drivers of economy, households induce a large share of worldwide environmental impacts. The variability of local consumption patterns and associated environmental impacts needs to be quantified as an important starting point to devise targeted measures aimed at reducing household environmental footprints. The goal of this article is the development and appraisal of a comprehensive regionalized bottom‐up model that assesses realistic environmental profiles for individual households in a specific region. For this purpose, a physically based building energy model, the results of an agent‐based transport simulation, and a data‐driven household consumption model were interlinked within a new probability‐based classification framework and applied to the case of Switzerland. The resulting model predicts the demands in about 400 different consumption areas for each Swiss household by considering its particular circumstances and produces a realistic picture of variability in household environmental footprints. An analysis of the model results on a municipal level reveals per‐capita income, population density, buildings' age, and household structure as possible drivers of municipal carbon footprints. While higher‐emission municipalities are located in rural areas and tend to show higher shares of older buildings, lower‐emission communities have larger proportions of families and can be found in highly populated regions by trend. However, the opposing effects of various variables observed in this analysis confirm the importance of a model that is able to capture regional distinctions. The overall model constitutes a comprehensive information base supporting policymakers in understanding consumption patterns in their region and deriving environmental strategies tailored to their specific population.     

Age structure in neutral theory resolves inconsistencies related to reproductive-size threshold

Aims Neutral theory consists of a suite of models that assume ecological equivalence among individual organisms. They have been most commonly applied to tropical forest tree communities either as null models or as approximations. Neutral models typically only include reproductive adults; therefore, fitting to empirical tree community data requires defining a reproductive-size threshold, which for trees is usually set arbitrarily to a diameter at breast height (DBH) of 100 mm. The inevitable exclusion of some reproductive adults and inclusion of some saplings cause a non-random sampling bias in neutral model fits. Here, we investigate this problem and resolve it by introducing simple age structure into a neutral model.Methods We compared the performance and sensitivity of DBH threshold of three variants of a spatially explicit neutral model: the traditional model, a model incorporating random sampling and a model with two distinct age classes—reproductive adults and saplings. In the age-structured model, saplings are offspring from adults that disperse according to a Gaussian dispersal kernel around the adults. The only extra parameter is the ratio of adults to saplings, which is not a free parameter but directly measurable. We used species–area relationships (SARs) to explore the predicted effect of saplings on the species richness at different scales in our model. We then evaluated the three model variations to find the parameters required to maintain the observed level of species richness in the 50-ha plot on Barro Colorado Island (BCI). We repeated our analysis filtering the data at different minimum tree-size thresholds in order to find the effect this threshold has on our results. Lastly, we used empirical species–individual relationships (SIRs) to test the pre-existing hypothesis that environmental filtering is the primary cause of differences between the assemblage of saplings and that of adults on BCI.Important findings Our age-structured neutral model was characterized by SARs that were insensitive to the presence of saplings at large scales and highly sensitive to them at small scales. Both models without age structure were highly sensitive to the DBH threshold chosen in a way that could not be explained based on random samplings alone. The age-structured neutral model, which allowed for non-random sampling based on life stage, was consistent with species richness observations. Our analysis of empirical SIRs did not support environmental filtering as a dominant force, but it did show evidence for other differences between age classes. Age can now be easily incorporated into future studies of neutral models whenever there is a concern that a sample is not entirely composed of reproductive adult individuals. More generally, we suggest that modeling studies using tree data subject to a minimum size threshold should consider the sensitivity of their results to that threshold.     

Consequences of the genetic threshold model for observing partial migration under climate change scenarios

Migration is a widespread phenomenon across the animal kingdom as a response to seasonality in environmental conditions. Partially migratory populations are populations that consist of both migratory and residential individuals. Such populations are very common, yet their stability has long been debated. The inheritance of migratory activity is currently best described by the threshold model of quantitative genetics. The inclusion of such a genetic threshold model for migratory behavior leads to a stable zone in time and space of partially migratory populations under a wide range of demographic parameter values, when assuming stable environmental conditions and unlimited genetic diversity. Migratory species are expected to be particularly sensitive to global warming, as arrival at the breeding grounds might be increasingly mistimed as a result of the uncoupling of long‐used cues and actual environmental conditions, with decreasing reproduction as a consequence. Here, we investigate the consequences for migratory behavior and the stability of partially migratory populations under five climate change scenarios and the assumption of a genetic threshold value for migratory behavior in an individual‐based model. The results show a spatially and temporally stable zone of partially migratory populations after different lengths of time in all scenarios. In the scenarios in which the species expands its range from a particular set of starting populations, the genetic diversity and location at initialization determine the species’ colonization speed across the zone of partial migration and therefore across the entire landscape. Abruptly changing environmental conditions after model initialization never caused a qualitative change in phenotype distributions, or complete extinction. This suggests that climate change‐induced shifts in species’ ranges as well as changes in survival probabilities and reproductive success can be met with flexibility in migratory behavior at the species level, which will reduce the risk of extinction.     

Vegetation as a component of a non-nested hierarchy: a conceptual model

Abstract. A general conceptual model of vegetation based on hierarchy theory is presented. The model emphasizes that prediction of vegetation requires consideration of both mechanisms of vegetation change and the constraints within which it occurs. The mechanisms of vegetation change are the responses to and effects upon their surroundings of individual plants. The most general constraints upon vegetation are aspects of the environment not affected by vegetation over successional time, and the pool of species within dispersal range. Examples of such environmental factors include macroclimate and soil parent material. In some cases, vegetation may alter important labile environmental factors such as soil nutrient and water availability. Some vegetation compositions appear to be resistant to changes in the general constraints. Due to both sources, there are multiple possible vegetation compositions given the same general constraints. Disturbance is defined as an abrupt change in the constraints on the vegetation resulting in a change in the vegetation's state or dynamics. Both the recognition of disturbance and the distinction between independent and labile environmental factors depend on the spatial and temporal scale of observation. For example, a particular wildfire at a given stand may be a disturbance, whereas at a larger scale of observation the same event may contribute to the wildfire regime, part of the constraints at that scale. Similarly, levels of soil organic matter may constrain vegetation over short time scales, due to influencing availability of water and nutrients. Over long time scales, the vegetation itself is a primary determinant of soil organic matter content. This model contains elements of both the initial, holistic theory of vegetation and recent, reductionistic approaches. It reiterates the need to considerboth mechanisms and constraints, stressed by contemporary and earlier workers. Hierarchy theory provides new insights concerning sufficient conditions for prediction, possible limits on predictability, and appropriate research strategy.     

Scaling physiological measurements for individuals of different body size

This paper examines how selected physiological performance variables, such as maximal oxygen uptake, strength and power, might best be scaled for subject differences in body size. The apparent dilemma between using either ratio standards or a linear adjustment method to scale was investigated by considering how maximal oxygen uptake (l.min-1), peak and mean power output (W) might best be adjusted for differences in body mass (kg). A curvilinear power function model was shown to be theoretically, physiologically and empirically superior to the linear models. Based on the fitted power functions, the best method of scaling maximum oxygen uptake, peak and mean power output, required these variables to be divided by body mass, recorded in the units kg 2/3. Hence, the power function ratio standards (ml.kg-2/3.min-1) and (W.kg-2/3) were best able to describe a wide range of subjects in terms of their physiological capacity, i.e. their ability to utilise oxygen or record power maximally, independent of body size. The simple ratio standards (ml.kg-1.min-1) and (W.kg-1) were found to best describe the same subjects according to their performance capacities or ability to run which are highly dependent on body size. The appropriate model to explain the experimental design effects on such ratio standards was shown to be log-normal rather than normal. Simply by taking logarithms of the power function ratio standard, identical solutions for the design effects are obtained using either ANOVA or, by taking the unscaled physiological variable as the dependent variable and the body size variable as the covariate, ANCOVA methods.     

Variable selection for discriminant analysis with Markov random field priors for the analysis of microarray data

MOTIVATION: Discriminant analysis is an effective tool for the classification of experimental units into groups. Here, we consider the typical problem of classifying subjects according to phenotypes via gene expression data and propose a method that incorporates variable selection into the inferential procedure, for the identification of the important biomarkers. To achieve this goal, we build upon a conjugate normal discriminant model, both linear and quadratic, and include a stochastic search variable selection procedure via an MCMC algorithm. Furthermore, we incorporate into the model prior information on the relationships among the genes as described by a gene-gene network. We use a Markov random field (MRF) prior to map the network connections among genes. Our prior model assumes that neighboring genes in the network are more likely to have a joint effect on the relevant biological processes. RESULTS: We use simulated data to assess performances of our method. In particular, we compare the MRF prior to a situation where independent Bernoulli priors are chosen for the individual predictors. We also illustrate the method on benchmark datasets for gene expression. Our simulation studies show that employing the MRF prior improves on selection accuracy. In real data applications, in addition to identifying markers and improving prediction accuracy, we show how the integration of existing biological knowledge into the prior model results in an increased ability to identify genes with strong discriminatory power and also aids the interpretation of the results.     

A Stochastic Dynamic Methodology (StDM) for reservoir's water quality management: Validation of a multi-scale approach in a south European basin (Douro,Portugal)

Worldwide aquatic ecosystems have been impacted by broad-scale environmental pressures such as agriculture, point and non-point-source pollution and land-use changes overlapping in space and time, leading to the disruption of the structure and functioning of these systems. The present paper examined the applicability of a holistic Stochastic Dynamic Methodology (StDM) in predicting the tendencies of phytoplankton communities and physicochemical conditions in reservoirs as a response to the changes in the respective watershed soil use. The case of the Douro's basin (Portugal) was used to test the StDM performance in this multi-scale approach. The StDM is a sequential modelling process developed in order to predict the ecological status of changed ecosystems, from which management strategies can be designed. The data used in the dynamic model construction included true gradients of environmental changes and was sampled from 1995 to 2004. The dynamic model developed was preceded by a conventional multivariate statistical procedure performed to discriminate the significant relationships between the selected ecological components. The model validation was based on independent data, for all the state variables considered. Overall, the simulation results are encouraging since they seem to demonstrate the StDM reliability in capturing the dynamics of the studied reservoirs. The StDM model simulations were validated for the most part of the twenty-two components selected as ecological indicators, with a performance of 50% for the physicochemical variables, 75% for the phytoplankton variables, and 100% for the Carlson trophic state indices (TSI). This approach provides a useful starting point, as a contribution for the practical implementation of the European Water Framework Directive, allowing the development of a true integrated assessment tool for water quality management, both at the scale of the reservoir body and at the scale of the respective river watershed dynamics.     

A two-sample comparison for multiple ordered event data

A longitudinal study is conducted to compare the process of particular disease between two groups. The process of the disease is monitored according to which of several ordered events occur. In the paper, the sojourn time between two successive events is considered as the outcome of interest. The group effects on the sojourn times of the multiple events are parameterized by scale changes in a semiparametric accelerated failure time model where the dependence structure among the multivariate sojourn times is unspecified. Suppose that the sojourn times are subject to dependent censoring and the censoring times are observed for all subjects. A log-rank-type estimating approach by rescaling the sojourn times and the dependent censoring times into the same distribution is constructed to estimate the group effects and the corresponding estimators are consistent and asymptotically normal. Without the dependent censoring, the independent censoring times in general are not available for the uncensored data. In order to complete the censoring information, pseudo-censoring times are generated from the corresponding nonparametrically estimated survival function in each group, and we can still obtained unbiased estimating functions for the group effects. A real application and a simulation study are conducted to illustrate the proposed methods.     

A Simulation Model for Assessing Soybean Rust Epidemics

A soybean rust (causal agent Phakopsora pachyrhizt) simulation model was developed for assessing disease epidemics as a part of pest risk analysis. Equations describing environmental effects on disease components were developed by re-analyzing previous data with a view toward a systems approach. The infection rate was predicted well using dew period and temperature after inoculation as independent variables (R²=0.88, P < 0.0001). The exponential models which used physiological day as an independent variable explained 98% of the variations of latent period and senescence of disease lesions. The simulation model was validated with data from 72 sequential planting experiments during 1980 and 1981 in Taiwan. Time of onset for these epidemics varied from 25—60 days and 50—80 days after planting soybean cultivars TK 5 and G 8587, respectively. The epidemic periods were 75—95 for TK 5 and 100—120 days for G 8587. Variation of epidemics was accurately predicted by the simulator. Predicted disease curves fit well the observed disease curves for the recognized cropping seasons, spring- and autumn-seeded crops. For G 8587, which is very sensitive to photoperiod, the data from spring and autumn gave a better fit compared with data from pre-summer planting. The model underestimated disease epidemics during the winter, probably because the plant growth model failed to reflect the photoperiod rection of soybean. The simulation model was validated with data from other experiments conducted in three cropping seasons in 1979 and 1980. Determination coefficients of the regression between observed and predicted disease severity were significant.     

Relating environmental variation to selection on reaction norms: an experimental test

Theoretical models predict that selection on reaction norms should depend on the relative frequency of environmental states experienced by a population. We report a laboratory experimental test of this prediction for thermal performance curves of larval growth rate in Pieris rapae in relation to their thermal environment. We measured short-term relative growth rate (RGR) for each individual at a series of five temperatures, and then we assigned individuals randomly to warm or cool selection treatments, which differ in the frequency distributions of environmental temperatures. Selection gradient analyses of two independent experiments demonstrated significant positive selection for increasing RGR, primarily through its effects on survival to adulthood and on development rate. In both the warm and cool selection treatments, the magnitude of directional selection on RGR was consistently greater at lower (suboptimal) temperatures than at higher temperatures; differences in selection between the treatments did not match model predictions. The temporal order and duration of environmental conditions may affect patterns of selection on thermal performance curves and other continuous reaction norms, complicating the connections between variation in environment, phenotype, and fitness.     

Gypsy moth response to landscape structure differs from neutral model predictions: implications for invasion monitoring

Simulations of dispersal across computer-generated neutral landscapes have generated testable predictions about the relationship between dispersal success and landscape structure. Models predict a threshold response in dispersal success with increasing habitat fragmentation. A threshold is defined as an abrupt, disproportionate decline in dispersal success at a certain proportion of habitat in the landscape. To identify potential empirical threshold responses in invasion success to landscape structure, we quantified the relationship between progression of the gypsy moth (Lymantria dispar) invasion wavefront across Michigan (1985–1996) and the structure of the Michigan landscape using two indices of invasion success and six landscape metrics. We also examined the effect of scale of analysis and choice of land cover characterization on our results by repeating our analysis at three scales using two different land cover maps. Contrary to simulation model predictions, thresholds in invasion success did not correspond closely with thresholds in landscape structure metrics. Increased variation in invasion success indices at smaller scales of analysis also suggested that invasion success should be studied at larger spatial extents (≥75 km²) than would be appropriate for characterizing individual dispersal events. The predictions of individual dispersal models across neutral landscapes may have limited applications for the monitoring and management of vagile species with excellent dispersal capabilities such as the gypsy moth.     

Predicting global habitat suitability for stony corals on seamounts

Aim Globally, species distribution patterns in the deep sea are poorly resolved, with spatial coverage being sparse for most taxa and true absence data missing. Increasing human impacts on deep‐sea ecosystems mean that reaching a better understanding of such patterns is becoming more urgent. Cold‐water stony corals (Order Scleractinia) form structurally complex habitats (dense thickets or reefs) that can support a diversity of other associated fauna. Despite their widely accepted ecological importance, records of scleractinian corals on seamounts are patchy and simply not available for most of the global ocean. The objective of this paper is to model the global distribution of suitable habitat for stony corals on seamounts. Location Seamounts worldwide. Methods We compiled a database containing all accessible records of scleractinian corals on seamounts. Two modelling approaches developed for presence‐only data were used to predict global habitat suitability for seamount scleractinians: maximum entropy modelling (Maxent) and environmental niche factor analysis (ENFA). We generated habitat‐suitability maps and used a cross‐validation process with a threshold‐independent metric to evaluate the performance of the models. Results Both models performed well in cross‐validation, although the Maxent method consistently outperformed ENFA. Highly suitable habitat for seamount stony corals was predicted to occur at most modelled depths in the North Atlantic, and in a circumglobal strip in the Southern Hemisphere between 20° and 50° S and shallower than around 1500 m. Seamount summits in most other regions appeared much less likely to provide suitable habitat, except for small near‐surface patches. The patterns of habitat suitability largely reflect current biogeographical knowledge. Environmental variables positively associated with high predicted habitat suitability included the aragonite saturation state, and oxygen saturation and concentration. By contrast, low levels of dissolved inorganic carbon, nitrate, phosphate and silicate were associated with high predicted suitability. High correlation among variables made assessing individual drivers difficult. Main conclusions Our models predict environmental conditions likely to play a role in determining large‐scale scleractinian coral distributions on seamounts, and provide a baseline scenario on a global scale. These results present a first‐order hypothesis that can be tested by further sampling. Given the high vulnerability of cold‐water corals to human impacts, such predictions are crucial tools in developing worldwide conservation and management strategies for seamount ecosystems.