Joint Species Distribution Models Research Articles

Trait‐mediated responses to aridity and experimental drought by springtail communities across Europe

AbstractThe capacity to forecast the effects of climate change on biodiversity largely relies on identifying traits capturing mechanistic relationships with the environment through standardized field experiments distributed across relevant spatial scales. The effects of short‐term experimental manipulations on local communities may overlap with regional climate gradients that have been operating during longer time periods. However, to the best of our knowledge, there are no studies simultaneously assessing such long‐term macroecological drivers with local climate manipulations.We analysed this issue with springtails (Class Collembola), one of the dominant soil fauna groups, in a standardized climate manipulation experiment conducted across six European countries encompassing broad climate gradients. We combined community data (near 20K specimens classified into 102 species) with 22 eco‐morphological traits and reconstructed their phylogenetic relationships to track the evolution of adaptations to live at different soil depths, which is key to cope with desiccation. We then applied joint species distribution models to investigate the combined effect of the regional aridity gradient with the local experimental treatment (drought and warming) over the assembly of springtail communities and tested for significant trait–environment relationships mediating their community‐level responses.Our results show (1) a convergent evolution in all three major collembolan lineages of species adapted to inhabit at different soil strata; (2) a clear signature of aridity selecting traits of more epigeic species at a biogeographical scale and (3) the association of short‐term experimental drought with traits related to more euedaphic life‐forms.The hemiedaphic condition would be the plesiomorphic state for Collembola while the adaptations for an epigeic life would have been secondarily gained. Epigeic springtails are not only more resistant to drought, but also have a higher dispersal capacity that allows them to seek more favourable micro‐habitats after experiencing drier conditions. The observed relative edaphization of the springtail communities after short‐term experimental drought may thus be a transient community response.The disparity between macroecological trends and fast community‐level responses after climate manipulations highlights the need of simultaneously assessing long‐term and short‐term drivers at broad spatial scales to adequately interpret trait–environment relationships and better forecast biodiversity responses to climate change.Read the freePlain Language Summaryfor this article on the Journal blog.

Index to Titles and Authors/Editors of Books Reviewed

Index to Titles and Authors/Editors of Books Reviewed

Gut microbiota differences between paired intestinal wall and digesta samples in three small species of fish.

The microbial gut communities of fish are receiving increased attention for their relevance, among others, in a growing aquaculture industry. The members of these communities are often split into resident (long-term colonisers specialised to grow in and adhere to the mucus lining of the gut) and transient (short-term colonisers originated from food items and the surrounding water) microorganisms. Separating these two communities in small fish are impeded by the small size and fragility of the gastrointestinal tract. With the aim of testing whether it is possible to recover two distinct communities in small species of fish using a simple sampling technique, we used 16S amplicon sequencing of paired intestinal wall and digesta samples from three small Cyprinodontiformes fish. We examined the diversity and compositional variation of the two recovered communities, and we used joint species distribution modelling to identify microbes that are most likely to be a part of the resident community. For all three species we found that the diversity of intestinal wall samples was significantly lower compared to digesta samples and that the community composition between sample types was significantly different. Across the three species we found seven unique families of bacteria to be significantly enriched in samples from the intestinal wall, encompassing most of the 89 ASVs enriched in intestinal wall samples. We conclude that it is possible to characterise two different microbial communities and identify potentially resident microbes through separately analysing samples from the intestinal wall and digesta from small species of fish. We encourage researchers to be aware that different sampling procedures for gut microbiome characterization will capture different parts of the microbiome and that this should be taken into consideration when reporting results from such studies on small species of fish.

Traits and phylogenies modulate the environmental responses of wood‐inhabiting fungal communities across spatial scales

Abstract Identifying the spatial scales at which community assembly processes operate is fundamental for gaining a mechanistic understanding of the drivers shaping ecological communities. In this study, we examined whether and how traits and phylogenetic relationships structure fungal community assembly across spatial scales. We applied joint species distribution modelling to a European‐scale dataset on 215 wood‐inhabiting fungal species, which includes data on traits, phylogeny and environmental variables measured at the local (log‐level) and regional (site‐level) scales. At the local scale, wood‐inhabiting fungal communities were mostly structured by deadwood decay stage, and the trait and phylogenetic patterns along this environmental gradient suggested the lack of diversifying selection. At regional scales, fungal communities and their trait distributions were influenced by climatic and connectivity‐related variables. The fungal climatic niches were not phylogenetically structured, suggesting that diversifying selection or stabilizing selection for climatic niches has played a strong role in wood‐inhabiting communities. In contrast, we found a strong phylogenetic signal in the responses to connectivity‐related variables, revealing phylogenetic homogenization in small and isolated forests. Synthesis. Altogether, our results show that species‐level traits and phylogenies modulate the responses of wood‐inhabiting fungi to environmental processes acting at different scales. This result suggests that the evolutionary histories of fungal traits diverge along different environmental axes.

A taxonomic-based joint species distribution model for presence-only data.

Species distribution models (SDMs) are an important class of model for mapping taxa spatially and are a key tool for tackling biodiversity loss. However, most common SDMs depend on presence–absence data and, despite the accumulation and exponential growth of biological occurrence data across the globe, the available data are predominantly presence-only (i.e. they lack real absences). Although presence-only SDMs do exist, they inevitably require assumptions about absences of the considered taxa and they are specified mostly for single species and, thus, do not exploit fully the information in related taxa. This greatly limits the utility of global biodiversity databases such as GBIF. Here, we present a Bayesian-based SDM for multiple species that operates directly on presence-only data by exploiting the joint distribution between the multiple ecological processes and, crucially, identifies the sampling effort per taxa which allows inference on absences. The model was applied to two case studies. One, focusing on taxonomically diverse taxa over central Mexico and another focusing on the monophyletic family Cactacea over continental Mexico. In both cases, the model was able to identify the ecological and sampling effort processes for each taxon using only the presence observations, environmental and anthropological data.

Jointly modeling marine species to inform the effects of environmental change on an ecological community in the Northwest Atlantic

Single species distribution models (SSDMs) are typically used to understand and predict the distribution and abundance of marine fish by fitting distribution models for each species independently to a combination of abiotic environmental variables. However, species abundances and distributions are influenced by abiotic environmental preferences as well as biotic dependencies such as interspecific competition and predation. When species interact, a joint species distribution model (JSDM) will allow for valid inference of environmental effects. We built a joint species distribution model of marine fish and invertebrates of the Northeast US Continental Shelf, providing evidence on species relationships with the environment as well as the likelihood of species to covary. Predictive performance is similar to SSDMs but the Bayesian joint modeling approach provides two main advantages over single species modeling: (1) the JSDM directly estimates the significance of environmental effects; and (2) predicted species richness accounts for species dependencies. An additional value of JSDMs is that the conditional prediction of species distributions can use not only the environmental associations of species, but also the presence and abundance of other species when forecasting future climatic associations.

Abiotic factors are more important than land management and biotic interactions in shaping vascular plant and soil fungal communities

Human-induced disturbances to ecosystems cause a direct loss of biodiversity, and also alter the inherent processes that shape ecosystems even after the main disturbance has ceased. Therefore, is it important to understand the ongoing consequences of past and present land use practices on both above- and belowground components of agroecosystems. Our study takes a detailed investigation of what shapes communities in semi-natural grasslands under long-term management as either pastures or meadows. We investigated the relative importance of land use (land management (LM) and land use intensity (LUI)) and abiotic conditions (soil, topographic, climatic) across five grassland sites in central Germany in explaining species occurrence patterns of vascular plants and soil fungi. Analyses included a hierarchical joint-species distribution modeling approach to uncover the role of possible drivers shaping the local communities. Our results show that abiotic factors are of particularly high importance compared to LM and LUI for both vascular plant and soil fungal communities. In general, the relative importance of explanatory variables was similar across both taxon groups, however, for plant communities, local climate conditions were more important, while for fungal communities the soil variables (e.g., soil acidity and fertility including soil carbon and potassium concentrations) played a more prominent role. Species-specific responses to the respective variables showed some similarity across the taxon groups, however less so than expected given the long legacy of past LM. Here we found that approximately 50% of the plant and fungi species had clear LM preferences and responses to LUI. More plant species were positively related to pasture than meadow management, while the opposite was found for fungal species. Our findings advance the understanding of how abiotic conditions and human land use impact local species communities in managed semi-natural grasslands, aiding further research and policy development for conserving multitrophic diversity within these biodiversity rich habitats. Our results highlight the importance of controlling for soil and climate in studying the impact of land use, and of considering the environmental context at both small and larger spatial grains when making land management and biodiversity conservation decisions. In so doing, the complexity of ecological processes within managed systems are accounted for and prioritized, promoting both conservation and ecological functioning of the agroecosystem. Plant data is archived on PANGAEA (Kautzner, Auge, Roscher, Baessler & Slabbert, 2020; https://doi.org10.1594/PANGA EA.919343) and the soil microbial data on the National Center for Biotechnology Information (NCBI) Sequence Read Archive (SRA) under the accession PRJNA563995 (Kautzner, Baessler, Auge, Roscher & Wubet, 2020)

The Influence of the Interaction between Climate and Competition on the Distributional Limits of European Shrews.

Simple SummaryIt is known that species’ distributions are influenced by several ecological factors. Nonetheless, the geographical scale upon which the influence of these factors is perceived is largely undefined. We assessed the importance of competition in regulating the distributional limits of species at large geographical scales. We studied European Soricidae shrews, because their species have similar diets, and focused on how interspecific competition changes along climatic gradients. We used presence data for the seven most widespread terrestrial species of Soricidae in Europe, gathered from online repositories, European museums, and gridded climate data. Using two different methods, we analysed the correlations between species’ presences, aiming to understand the distinct roles of climate and competition in shaping species’ distributions. Our results support three key conclusions: (i) climate alone does not explain all species’ distributions at large scales; (ii) negative interactions, such as competition, seem to play a strong role in defining species’ range limits, even at large scales; and (iii) the impact of competition on a species’ distribution varies along a climatic gradient, becoming stronger at the climatic extremes. Our conclusions support previous research, highlighting the importance of considering biotic interactions when studying species’ distributions, regardless of geographical scale.It is known that species’ distributions are influenced by several ecological factors. Nonetheless, the geographical scale upon which the influence of these factors is perceived is largely undefined. We assessed the importance of competition in regulating the distributional limits of species at large geographical scales. We focus on species with similar diets, the European Soricidae shrews, and how interspecific competition changes along climatic gradients. We used presence data for the seven most widespread terrestrial species of Soricidae in Europe, gathered from GBIF, European museums, and climate data from WorldClim. We made use of two Joint Species Distribution Models to analyse the correlations between species’ presences, aiming to understand the distinct roles of climate and competition in shaping species’ distributions. Our results support three key conclusions: (i) climate alone does not explain all species’ distributions at large scales; (ii) negative interactions, such as competition, seem to play a strong role in defining species’ range limits, even at large scales; and (iii) the impact of competition on a species’ distribution varies along a climatic gradient, becoming stronger at the climatic extremes. Our conclusions support previous research, highlighting the importance of considering biotic interactions when studying species’ distributions, regardless of geographical scale.

Ecological stoichiometry and fungal community turnover reveal variation among mycorrhizal partners in their responses to warming and drought.

Symbiotic fungi mediate important energy and nutrient transfers in terrestrial ecosystems. Environmental change can lead to shifts in communities of symbiotic fungi, but the consequences of these shifts for nutrient dynamics among symbiotic partners are poorly understood. Here, we assessed variation in carbon (C), nitrogen (N) and phosphorus (P) in tissues of arbuscular mycorrhizal (AM) fungi and a host plant (Medicago sativa) in response to experimental warming and drought. We linked compositional shifts in AM fungal communities in roots and soil to variation in hyphal chemistry by using high-throughput DNA sequencing and joint species distribution modelling. Compared to plants, AM hyphae was 43% lower in (C) and 24% lower in (N) but more than nine times higher in (P), with significantly lower C:N, C:P and N:P ratios. Warming and drought resulted in increases in (P) and reduced C:P and N:P ratios in all tissues, indicating fungal P accumulation was exacerbated by climate-associated stress. Warming and drought modified the composition of AM fungal communities, and many of the AM fungal genera that were linked to shifts in mycelial chemistry were also negatively impacted by climate variation. Our study offers a unified framework to link climate change, fungal community composition, and community-level functional traits. Thus, our study provides insight into how environmental change can alter ecosystem functions via the promotion or reduction of fungal taxa with different stoichiometric characteristics and responses.

Spatiotemporal interactions of a novel mesocarnivore community in an urban environment before and during SARS-CoV-2 lockdown.

Studying species interactions and niche segregation under human pressure provides important insights into species adaptation, community functioning and ecosystem stability. Due to their high plasticity in behaviour and diet, urban mesocarnivores are ideal species for studying community assembly in novel communities.We analysed the spatial and temporal species interactions of an urban mesocarnivore community composed of the red fox Vulpes vulpes and the marten Martes sp. as native species, the raccoon Procyon lotor as invasive species, and the cat Felis catus as a domestic species in combination with human disturbance modulated by the SARS‐CoV‐2 lockdown effect that happened while the study was conducted.We analysed camera trap data and applied a joint species distribution model to understand not only the environmental variables influencing the detection of mesocarnivores and their use intensity of environmental features but also the species’ co‐occurrences while accounting for environmental variables. We then assessed whether they displayed temporal niche partitioning based on activity analyses, and finally analysed at a smaller temporal scale the time of delay after the detection of another focal species.We found that species were more often detected and displayed a higher use intensity in gardens during the SARS‐CoV‐2 lockdown period, while showing a shorter temporal delay during the same period, meaning a high human‐induced spatiotemporal overlap. All three wild species spatially co‐occurred within the urban area, with a positive response of raccoons to cats in detection and use intensity, whereas foxes showed a negative trend towards cats. When assessing the temporal partitioning, we found that all wild species showed overlapping nocturnal activities. All species displayed temporal segregation based on temporal delay. According to the temporal delay analyses, cats were the species avoided the most by all wild species. To conclude, we found that although the wild species were positively associated in space, the avoidance occurred at a smaller temporal scale, and human pressure in addition led to high spatiotemporal overlap.Our study sheds light to the complex patterns underlying the interactions in a mesocarnivore community both spatially and temporally, and the exacerbated effect of human pressure on community dynamics.

Traits mediate niches and co‐occurrences of forest beetles in ways that differ among bioclimatic regions

AbstractAimThe aim of this study was to investigate the role of traits in beetle community assembly and test for consistency in these effects among several bioclimatic regions. We asked (1) whether traits predicted species’ responses to environmental gradients (i.e. their niches), (2) whether these same traits could predict co‐occurrence patterns and (3) how consistent were niches and the role of traits among study regions.LocationBoreal forests in Norway and Finland, temperate forests in Germany.TaxonWood‐living (saproxylic) beetles.MethodsWe compiled capture records of 468 wood‐living beetle species from the three regions, along with nine morphological and ecological species traits. Eight climatic and forest covariates were also collected. We used Bayesian hierarchical joint species distribution models to estimate the influence of traits and phylogeny on species’ niches. We also tested for correlations between species associations and trait similarity. Finally, we compared species niches and the effects of traits among study regions.ResultsTraits explained some of the variability in species’ niches, but their effects differed among study regions. However, substantial phylogenetic signal in species niches implies that unmeasured but phylogenetically structured traits have a stronger effect. Degree of trait similarity was correlated with species associations but depended idiosyncratically on the trait and region. Species niches were much more consistent—widespread taxa often responded similarly to an environmental gradient in each region.Main conclusionsThe inconsistent effects of traits among regions limit their current use in understanding beetle community assembly. Phylogenetic signal in niches, however, implies that better predictive traits can eventually be identified. Consistency of species niches among regions means niches may remain relatively stable under future climate and land use changes; this lends credibility to predictive distribution models based on future climate projections but may imply that species’ scope for short‐term adaptation is limited.

Moth species richness and diversity decline in a 30-year time series in Norway, irrespective of species\u2019 latitudinal range extent and habitat

IntroductionInsects are reported to be in decline around the globe, but long-term datasets are rare. The causes of these trends are elusive, with changes in land use and climate among the top candidates. Yet if species traits can predict rates of population change, this can help identify underlying mechanisms. If climate change is important, for example, high-latitude species may decline as temperate species expand. Land use changes, however, may impact species that rely on certain habitats.Aims and methodsWe present 30 years of moth captures (comprising 97,032 individuals of 808 species) from a site in southeast Norway to test for population trends that are correlated with species traits. We use time series analyses and joint species distribution models combined with local climate and habitat data.Results and discussionSpecies richness declined by 8.2% per decade and total abundance appeared to decline as well (−9.4%, p = 0.14) but inter-annual variability was high. One-fifth of species declined, although 6% increased. Winter and summer weather were correlated with annual rates of abundance change for many species. Opposite to general expectation, many species responded negatively to higher summer and winter temperatures. Surprisingly, species’ northern range limits and the habitat in which their food plants grew were not strong predictors of their time trends or their responses to climatic variation. Complex and indirect effects of both land use and climate change may play a role in these declines.Implications for insect conservationOur results provide additional evidence for long-term declines in insect abundance. The multifaceted causes of population changes may limit the ability of species traits to reveal which species are most at risk.

Modelling temperature-driven changes in species associations across freshwater communities.

Due to global climate change-induced shifts in species distributions, estimating changes in community composition through the use of Species Distribution Models has become a key management tool. Being able to determine how species associations change along environmental gradients is likely to be pivotal in exploring the magnitude of future changes in species' distributions. This is particularly important in connectivity-limited ecosystems, such as freshwater ecosystems, where increased human translocation is creating species associations over previously unseen environmental gradients. Here, we use a large-scale presence-absence dataset of freshwater fish from lakes across the Fennoscandian region in a Joint Species Distribution Model, to measure the effect of temperature on species associations. We identified a trend of negative associations between species tolerant of cold waters and those tolerant of warmer waters, as well as positive associations between several more warm-tolerant species, with these associations often shifting depending on local temperatures. Our results confirm that freshwater ecosystems can expect to see a large-scale shift towards communities dominated by more warm-tolerant species. While there remains much work to be done to predict exactly where and when local extinctions may take place, the model implemented provides a starting-point for the exploration of climate-driven community trends. This approach is especially informative in regards to determining which species associations are most central in shaping future community composition, and which areas are most vulnerable to local extinctions.

The internal structure of metacommunities

Current analyses of metacommunity data largely focus on global attributes across the entire metacommunity, such as mean alpha, beta, and gamma diversity, as well as the partitioning of compositional variation into single estimates of contributions of space and environmental effects and, more recently, possible contributions of species interactions. However, this view neglects the fact that different species and sites in the landscape can vary widely in how they contribute to these metacommunity‐wide attributes. We argue for a new conceptual framework with matched analytics with the goals of studying the complex and interactive relations between process and pattern in metacommunities that is focused on the variation among species and among sites which we call the ‘internal structure' of the metacommunity. To demonstrate how the internal structure could be studied, we create synthetic data using a process‐based colonization–extinction metacommunity model. We then use joint species distribution models to estimate how the contributions of space, environment, and biotic interactions driving metacommunity assembly differ among species and sites. We find that this approach to the internal structure of metacommunities provides useful information about the distinct ways that different species and different sites contribute to metacommunity structure. Although it has limitations, our work points at a more general approach to understand how other possible complexities might affect internal structure and might thus be incorporated into a more cohesive metacommunity theory.

Bird species co‐occurrence patterns in an alpine environment supports the stress‐gradient hypothesis

Understanding the relative contribution of different biotic interactions in shaping species assemblages constitutes a major goal in community ecology, and consequently, multiple methods aimed at inferring the nature of these associations have emerged during the last decade. In this framework, the stress‐gradient hypothesis (SGH) predicts that prevalent biotic interactions shift from competition to facilitation as abiotic stress increases (and productivity decreases). This hypothesis originally raised by plant ecologists has been barely applied to faunal communities. Here, we take advantage of 20 years of abundance data to investigate pairwise patterns in species co‐occurrence in alpine bird communities inhabiting two contrasting habitat types: forests (high productivity) and mountain grasslands (low productivity). We also integrate functional data with presence–absence and quantitative matrices in order to detect the signature of processes driving community assembly and test for limiting similarity. We employ a null model approach, probabilistic pairwise association tests and joint species distribution models; all methods revealed a higher frequency of positive interactions in mountain grasslands in agreement with SGH predictions. Both the frequency of positive and negative interactions remained moderately stable over the study period in both habitat types. There was no significant relationship between the degree of co‐occurrence of species pairs and their functional distance in either habitat. However, when we only considered those combinations of species whose co‐occurrence patterns deviated from that expected at random, we found that co‐existing species are functionally more similar than those pairs that show segregated patterns in the forest assemblages. Such a relationship may arise via selective social information use and other processes, including microhabitat preferences. Overall, our findings suggest that interspecific competition does not constitute a major force driving the structure of bird assemblages in this mountain region. This work also shows that association analyses should not be underestimated as they can provide clues to the mechanisms responsible for community structure.

Human disturbance and shifts in vertebratecommunity compositionin a biodiversity hotspot.

Understanding how human modification of the landscape shapes vertebrate community composition is vital to understanding the current status and future trajectory of wildlife. Using a participatory approach, we deployed the largest camera-trap network in Mesoamerica to date to investigate how anthropogenic disturbance shapes the occupancy and co-occurrence of terrestrial vertebrate species in a tropical biodiversity hotspot: the Osa Peninsula, Costa Rica. We estimated species richness in different categories of land protection with rarefaction analysis and estimated the expected occupancy with a joint species distribution model that included covariates for anthropogenic disturbance, land protection, habitat quality, and habitat availability. Areas with the most stringent land-use protections (e.g., Corcovado National Park, 24 species [95% CI 23-25]) harbored significantly more species than unprotected areas (20 species [19.7-20.3]), mainly due to a reduced presence of large-bodied species of conservation concern in unprotected areas (e.g., jaguar Panthera onca and white-lipped peccary Tayassu pecari). Small-bodied generalist species, such as opossums (Didelphidae) and armadillos (Dasypus novemcinctus), in contrast, were more common at disturbed sites, resulting in a significant difference in vertebrate community composition between sites with low and high disturbance. Co-occurrence of species was also mainly associated with response to disturbance. Similar responses to disturbance create two groups of species, those whose site-level occupancy usually increased as anthropogenic disturbance increased and those whose estimated occupancy decreased. The absence of large-bodied species entails an important loss of ecological function in disturbed areas and can hinder forest development and maintenance. Efforts to protect and restore forested landscapes are likely having a positive effect on the abundance of some threatened species. These efforts, however, must be sustained and expanded to increase connectivity and ensure the long-term viability of the wildlife community.

Predicting the occurrence of riparian woody species to inform environmental water policies in an Australian tropical river

Abstract River flows are commonly altered by water resource development, with changes to the natural flow regime potentially impacting riparian vegetation. Increasingly, water resource managers seek to design policy to maintain healthy riparian ecosystems. Models that make explicit the relationship between hydrological variables and vegetation can be used by managers to assess vegetation response under different water management scenarios. We determined the potential impact of water‐takeon the spatial distribution of woody riparian plant species in the lower Fitzroy River, in north‐western Australia, an area under pressure to increase water resource development. We undertook a plant survey and developed and applied a joint species distribution model to determine the likelihood of occurrence for 26 woody riparian plant species, mapped species occurrence and assessed the change in species distribution under two water‐take scenarios. We found that the duration of inundation from flood flows was a strong predictor of species occurrence in our joint species distribution model. We identified species associated with wetter environments, as indicated by their effect size for the inundation metric. Under the 300‐Gl water‐take scenario we found little change (<2%) in species occurrence, but under the 600‐Gl scenario a decline between 5% and 7.4% was predicted for eight species associated with wetter habitats. This decline was generally confined to a localised area. Our approach highlights the usefulness of predictive modelling to identify species most likely to be impacted by water‐take, and the benefit of linking modelling to spatial mapping because it can highlight areas where change is likely to occur. This information can assist management to protect ecologically and culturally important species.

Multispecies modelling reveals potential for habitat restoration to re‐establish boreal vertebrate community dynamics

Abstract The restoration of habitats degraded by industrial disturbance is essential for achieving conservation objectives in disturbed landscapes. In boreal ecosystems, disturbances from seismic exploration lines and other linear features have adversely affected biodiversity, most notably leading to declines in threatened woodland caribou. Large‐scale restoration of disturbed habitats is needed, yet empirical assessments of restoration effectiveness on wildlife communities remain rare. We used 73 camera trap deployments from 2015 to 2019 and joint species distribution models to investigate how habitat use by the larger vertebrate community (>0.2 kg) responded to variation in key seismic line characteristics (line‐of‐sight, width, density and mounding) following restoration treatments in a landscape disturbed by oil and gas development in northeastern Alberta. The proportion of variation explained by line characteristics was low in comparison to habitat type and season, suggesting short‐term responses to restoration treatments were relatively weak. However, we found that lines with characteristics consistent with restored conditions were predicted to support an altered community composition, with reduced use by wolf and coyote, thereby indicating that line restoration will result in reduced contact rates between caribou and these key predators. Synthesis and applications. Our analysis provides a framework to assess and predict wildlife community responses to emerging restoration efforts. With the growing importance of habitat restoration for caribou and other vertebrate species, we recommend longer‐term monitoring combined with landscape‐scale comparisons of different restoration approaches to more fully understand and direct these critical conservation investments. Only by combining rigorous multispecies monitoring with large‐scale restoration, will we effectively conserve biodiversity within rapidly changing environments.

A new joint species distribution model for faster and more accurate inference of species associations from big community data

Abstract Joint species distribution models (JSDMs) explain spatial variation in community composition by contributions of the environment, biotic associations and possibly spatially structured residual covariance. They show great promise as a general analytical framework for community ecology and macroecology, but current JSDMs, even when approximated by latent variables, scale poorly on large datasets, limiting their usefulness for currently emerging big (e.g. metabarcoding and metagenomics) community datasets. Here, we present a novel, more scalable JSDM (sjSDM) that circumvents the need to use latent variables by using a Monte Carlo integration of the joint JSDM likelihood together with flexible elastic net regularization on all model components. We implemented sjSDM in PyTorch, a modern machine learning framework, which allows making use of both CPU and GPU calculations. Using simulated communities with known species–species associations and different number of species and sites, we compare sjSDM with state‐of‐the‐art JSDM implementations to determine computational runtimes and accuracy of the inferred species–species and species–environment associations. We find that sjSDM is orders of magnitude faster than existing JSDM algorithms (even when run on the CPU) and can be scaled to very large datasets. Despite the dramatically improved speed, sjSDM produces more accurate estimates of species association structures than alternative JSDM implementations. We demonstrate the applicability of sjSDM to big community data using eDNA case study with thousands of fungi operational taxonomic units (OTU). Our sjSDM approach makes the analysis of JSDMs to large community datasets with hundreds or thousands of species possible, substantially extending the applicability of JSDMs in ecology. We provide our method in an R package to facilitate its applicability for practical data analysis.

Effectiveness of joint species distribution models in the presence of imperfect detection

Abstract Joint species distribution models (JSDMs) are a recent development in biogeography and enable the spatial modelling of multiple species and their interactions and dependencies. However, most models do not consider imperfect detection, which can significantly bias estimates. This is one of the first papers to account for imperfect detection when fitting data with JSDMs and to explore the complications that may arise. A multivariate probit JSDM that explicitly accounts for imperfect detection is proposed, and implemented using a Bayesian hierarchical approach. We investigate the performance of the JSDM in the presence of imperfect detection for a range of factors, including varied levels of detection and species occupancy, and varied numbers of survey sites and replications. To understand how effective this JSDM is in practice, we also compare results to those from a JSDM that does not explicitly model detection but instead makes use of ‘collapsed data’. A case study of owls and gliders in Victoria, Australia, is also illustrated. Using simulations, we found that the JSDMs explicitly accounting for detection can accurately estimate intrinsic correlation between species with enough survey sites and replications. Reducing the number of survey sites decreases the precision of estimates, while reducing the number of survey replications can lead to biased estimates. For low probabilities of detection, the model may require a large number of survey replications to remove bias from estimates. However, JSDMs not explicitly accounting for detection may have a limited ability to disentangle detection from occupancy, which substantially reduces their ability to accurately infer the species distribution spatially. Our case study showed positive correlation between Sooty Owls and Greater Gliders, despite a low number of survey replications. To avoid biased estimates of inter‐species correlations and species distributions, imperfect detection needs to be considered. However, for low probability of detection, the JSDMs explicitly accounting for detection is data hungry. Estimates from such models may still be subject to bias. To overcome the bias, researchers need to carefully design surveys and choose appropriate modelling approaches. The survey design should ensure sufficient survey replications for unbiased inferences on species inter‐dependencies and occupancy.