Analysis of Zero‐Inflated Poisson Data Incorporating Extent of Exposure

When analyzing Poisson count data sometimes a high frequency of extra zeros is observed. The Zero‐Inflated Poisson (ZIP) model is a popular approach to handle zero‐inflation. In this paper we generalize the ZIP model and its regression counterpart to accommodate the extent of individual exposure. Empirical evidence drawn from an occupational injury data set confirms that the incorporation of exposure information can exert a substantial impact on the model fit. Tests for zero‐inflation are also considered. Their finite sample properties are examined in a Monte Carlo study.     

Spatial modeling of data with excessive zeros applied to reindeer pellet‐group counts

We analyze a real data set pertaining to reindeer fecal pellet‐group counts obtained from a survey conducted in a forest area in northern Sweden. In the data set, over 70% of counts are zeros, and there is high spatial correlation. We use conditionally autoregressive random effects for modeling of spatial correlation in a Poisson generalized linear mixed model (GLMM), quasi‐Poisson hierarchical generalized linear model (HGLM), zero‐inflated Poisson (ZIP), and hurdle models. The quasi‐Poisson HGLM allows for both under‐ and overdispersion with excessive zeros, while the ZIP and hurdle models allow only for overdispersion. In analyzing the real data set, we see that the quasi‐Poisson HGLMs can perform better than the other commonly used models, for example, ordinary Poisson HGLMs, spatial ZIP, and spatial hurdle models, and that the underdispersed Poisson HGLMs with spatial correlation fit the reindeer data best. We develop R codes for fitting these models using a unified algorithm for the HGLMs. Spatial count response with an extremely high proportion of zeros, and underdispersion can be successfully modeled using the quasi‐Poisson HGLM with spatial random effects.     

The distributions of one invasive and two native crayfishes in relation to coarse‐scale natural and anthropogenic factors

1. Native crayfishes are often extirpated from portions of their range because of interactions with invasive species, anthropogenic alterations to environmental conditions or a combination of these factors. Our goal was to identify coarse‐scale natural and anthropogenic factors related to the current distributions of the invasive crayfish, Orconectes hylas, and two endemic crayfishes, Orconectes peruncus and Orconectes quadruncus in the St. Francis River drainage, Missouri, U.S.A. and to provide wider insights into the potential role of anthropogenic factors in facilitating species displacement. 2. We used classification trees to model coarse‐scale natural and anthropogenic environmental factors and their relation to the presence or absence of each species. Model results were then used to predict probability of presence for each species within each stream segment throughout the entire St. Francis River drainage. 3. Factors related to geology and soils were the best predictors of species distributions. A dichotomy of these factors explained much of the discrete distributions of the two native species. Agricultural‐related factors were identified as the most influential anthropogenic activity related to species distributions. All associations between the invasive species and anthropogenic factors were negative which suggested the invader was not likely to establish in heavily impacted areas. Overall, our models had high correct classification rates, and we were able to reliably predict the presence of the invader in the invaded drainage. 4. Given the negative associations of the invader with anthropogenic alterations at a coarse spatial scale, we believe other mechanisms are likely to be responsible for the widespread displacement of the two native species. These findings can be used to assist in conservation activities such as creation of refugia for native species and may direct future research to identify the mechanism(s) of species displacement.     

Mapping species abundance by a spatial zero‐inflated Poisson model: a case study in the Wadden Sea,the Netherlands

The objective of the study was to provide a general procedure for mapping species abundance when data are zero‐inflated and spatially correlated counts. The bivalve species Macoma balthica was observed on a 500×500 m grid in the Dutch part of the Wadden Sea. In total, 66% of the 3451 counts were zeros. A zero‐inflated Poisson mixture model was used to relate counts to environmental covariates. Two models were considered, one with relatively fewer covariates (model “small”) than the other (model “large”). The models contained two processes: a Bernoulli (species prevalence) and a Poisson (species intensity, when the Bernoulli process predicts presence). The model was used to make predictions for sites where only environmental data are available. Predicted prevalences and intensities show that the model “small” predicts lower mean prevalence and higher mean intensity, than the model “large”. Yet, the product of prevalence and intensity, which might be called the unconditional intensity, is very similar. Cross‐validation showed that the model “small” performed slightly better, but the difference was small. The proposed methodology might be generally applicable, but is computer intensive.     

Factors influencing species selection for littoral rainforest restoration: Do environmental gradients matter?

Summary Littoral rainforest in northern New South Wales, Australia, has been severely reduced in area and is now extremely limited in extent. Factors influencing the floristics, species richness and abundance, and relationship of this coastal rainforest community type to other lowland rainforests are explored. The purpose of the study was to provide ecological information to support (i) the development of management recommendations and assist habitat expansion and restoration planning for a coastal site at Lennox Head, in northern New South Wales, and (ii) the implementation of recovery actions for an endangered ecological community and an endangered tree species. Multivariate analysis techniques were used to classify and ordinate sampled sites relative to environmental variables to provide an explanation for current floristic assemblages. Eight locations at varying distances to the coast, and representing a range of soil types, were chosen to test the influence of selected environmental variables. At the broad scale, the results show that proximity to the coast and altitude were generally correlated and represented the most influential variables; soil depth, topographic position and slope were broadly correlated but markedly less influential; disturbance was a significant but independent influence on floristics; and soil type and aspect had the least influence. The study provided insight into the ecological parameters of a range of species suited to the habitat rehabilitation and restoration project, and identified finer‐scale floristic patterns at the Lennox site that appear to reflect the influence of environmental variables. In that case, areas in closer proximity to the ocean are dominated by several littoral rainforest and disturbance‐related species, and species richness increases relative to distance from the ocean. This highlights the need to be sensitive to landscape variation, and the influence of environmental variables on plant species distributions, and population dynamics and structure, to guide final selection of appropriate plant material for littoral rainforest restoration projects.     

A comparison between Poisson and zero-inflated Poisson regression models with an application to number of black spots in Corriedale sheep

Dark spots in the fleece area are often associated with dark fibres in wool, which limits its competitiveness with other textile fibres. Field data from a sheep experiment in Uruguay revealed an excess number of zeros for dark spots. We compared the performance of four Poisson and zero-inflated Poisson (ZIP) models under four simulation scenarios. All models performed reasonably well under the same scenario for which the data were simulated. The deviance information criterion favoured a Poisson model with residual, while the ZIP model with a residual gave estimates closer to their true values under all simulation scenarios. Both Poisson and ZIP models with an error term at the regression level performed better than their counterparts without such an error. Field data from Corriedale sheep were analysed with Poisson and ZIP models with residuals. Parameter estimates were similar for both models. Although the posterior distribution of the sire variance was skewed due to a small number of rams in the dataset, the median of this variance suggested a scope for genetic selection. The main environmental factor was the age of the sheep at shearing. In summary, age related processes seem to drive the number of dark spots in this breed of sheep.     

A new multivariate zero‐adjusted Poisson model with applications to biomedicine

Recently, although advances were made on modeling multivariate count data, existing models really has several limitations: (i) The multivariate Poisson log‐normal model (Aitchison and Ho, 1989) cannot be used to fit multivariate count data with excess zero‐vectors; (ii) The multivariate zero‐inflated Poisson (ZIP) distribution (Li et al., 1999) cannot be used to model zero‐truncated/deflated count data and it is difficult to apply to high‐dimensional cases; (iii) The Type I multivariate zero‐adjusted Poisson (ZAP) distribution (Tian et al., 2017) could only model multivariate count data with a special correlation structure for random components that are all positive or negative. In this paper, we first introduce a new multivariate ZAP distribution, based on a multivariate Poisson distribution, which allows the correlations between components with a more flexible dependency structure, that is some of the correlation coefficients could be positive while others could be negative. We then develop its important distributional properties, and provide efficient statistical inference methods for multivariate ZAP model with or without covariates. Two real data examples in biomedicine are used to illustrate the proposed methods.     

Plant community composition and species richness in the High Arctic tundra: From the present to the future

Arctic plant communities are altered by climate changes. The magnitude of these alterations depends on whether species distributions are determined by macroclimatic conditions, by factors related to local topography, or by biotic interactions. Our current understanding of the relative importance of these conditions is limited due to the scarcity of studies, especially in the High Arctic. We investigated variations in vascular plant community composition and species richness based on 288 plots distributed on three sites along a coast‐inland gradient in Northeast Greenland using a stratified random design. We used an information theoretic approach to determine whether variations in species richness were best explained by macroclimate, by factors related to local topography (including soil water) or by plant‐plant interactions. Latent variable models were used to explain patterns in plant community composition. Species richness was mainly determined by variations in soil water content, which explained 35% of the variation, and to a minor degree by other variables related to topography. Species richness was not directly related to macroclimate. Latent variable models showed that 23.0% of the variation in community composition was explained by variables related to topography, while distance to the inland ice explained an additional 6.4 %. This indicates that some species are associated with environmental conditions found in only some parts of the coast–inland gradient. Inclusion of macroclimatic variation increased the model's explanatory power by 4.2%. Our results suggest that the main impact of climate changes in the High Arctic will be mediated by their influence on local soil water conditions. Increasing temperatures are likely to cause higher evaporation rates and alter the distribution of late‐melting snow patches. This will have little impact on landscape‐scale diversity if plants are able to redistribute locally to remain in areas with sufficient soil water.     

A phylogenetic analysis of the British flora sheds light on the evolutionary and ecological factors driving plant invasions

Darwin's naturalization hypothesis predicts that invasive species should perform better in their novel range in the absence of close relatives in the native flora due to reduced competition. Evidence from recent taxonomic and phylogenetic‐based studies, however, is equivocal. We test Darwin's naturalization hypothesis at two different spatial scales using a fossil‐dated molecular phylogenetic tree of the British native and alien flora (ca. 1600 species) and extensive, fine‐scale survey data from the 1998 Countryside Survey. At both landscape and local scales, invasive species were neither significantly more nor less related to the native flora than their non‐invasive alien counterparts. Species invasiveness was instead correlated with higher nitrogen and moisture preference, but not other life history traits such as life‐form and height. We argue that invasive species spread in Britain is hence more likely determined by changes in land use and other anthropogenic factors, rather than evolutionary history. Synthesis. The transition from non‐invasive to invasive is not related to phylogenetic distinctiveness to the native community, but instead to their environmental preferences. Therefore, combating biological invasions in the Britain and other industrialized countries need entirely different strategies than in more natural environments.     

Factors influencing vascular plant diversity on 22 islands off the coast of eastern North America

Aim The influence of physiographic and historical factors on species richness of native and non‐native vascular plants on 22 coastal islands was examined. Location Islands off the coast of north‐eastern USA and south‐eastern Canada between 41° and 45° N latitude were studied. Island size ranges from 3 to 26,668 ha. All islands were deglaciated between 15,000 and 11,000 yr bp ; all but the four New Brunswick islands were attached to the mainland until rising sea level isolated them between 14,000 and 3800 yr bp . Methods Island species richness was determined from floras compiled or revised since 1969. Simple and multiple regression and rank correlation analysis were employed to assess the relative influence of independent variables on species richness. Potential predictors included island area, latitude, elevation, distance from the mainland, distance from the nearest larger island, number of soil types, years since isolation, years since deglaciation, and human population density. Results Native vascular plant species richness for the 22 islands in this study is influenced most strongly by island area, latitude, and distance from the nearest larger island; richness increases with island area, but decreases with latitude and distance from the nearest larger island as hypothesized. That a similar model employing distance from the mainland does not meet the critical value of P confirms the importance of the stepping‐stone effect. Habitat diversity as measured by number of soil types is also an important predictor of native plant species richness, but at least half of its influence can be attributed to island area, with which it is correlated. Two historical factors, years since deglaciation and years since isolation, also appear to be highly correlated with native species richness, but their influence cannot be separated from that of latitude for the present sample size. Non‐native vascular plant species richness is influenced primarily by island area and present‐day human population density, although human population density may be a surrogate for the cumulative effect of several centuries of anthropogenic impacts related to agriculture, hunting, fishing, whaling, tourism, and residential development. Very high densities of ground‐nesting pelagic birds may account for the high percentage of non‐native species on several small northern islands. Main conclusions Many of the principles of island biogeography that have been applied to oceanic islands apply equally to the 22 islands in this study. Native vascular plant species richness for these islands is strongly influenced by physiographic factors. Influence of two historical factors, years since deglaciation and years since isolation, cannot be assessed with the present sample size. Non‐native vascular plant species richness is influenced by island area as well as by human population density; human population density may be a surrogate for other anthropogenic impacts.     

Reconstruction of a windborne insect invasion using a particle dispersal model,historical wind data,and Bayesian analysis of genetic data

Understanding how invasive species establish and spread is vital for developing effective management strategies for invaded areas and identifying new areas where the risk of invasion is highest. We investigated the explanatory power of dispersal histories reconstructed based on local‐scale wind data and a regional‐scale wind‐dispersed particle trajectory model for the invasive seed chalcid wasp Megastigmus schimitscheki (Hymenoptera: Torymidae) in France. The explanatory power was tested by: (1) survival analysis of empirical data on M. schimitscheki presence, absence and year of arrival at 52 stands of the wasp's obligate hosts, Cedrus (true cedar trees); and (2) Approximate Bayesian analysis of M. schimitscheki genetic data using a coalescence model. The Bayesian demographic modeling and traditional population genetic analysis suggested that initial invasion across the range was the result of long‐distance dispersal from the longest established sites. The survival analyses of the windborne expansion patterns derived from a particle dispersal model indicated that there was an informative correlation between the M. schimitscheki presence/absence data from the annual surveys and the scenarios based on regional‐scale wind data. These three very different analyses produced highly congruent results supporting our proposal that wind is the most probable vector for passive long‐distance dispersal of this invasive seed wasp. This result confirms that long‐distance dispersal from introduction areas is a likely driver of secondary expansion of alien invasive species. Based on our results, management programs for this and other windborne invasive species may consider (1) focusing effort at the longest established sites and (2) monitoring outlying populations remains critically important due to their influence on rates of spread. We also suggest that there is a distinct need for new analysis methods that have the capacity to combine empirical spatiotemporal field data, genetic data, and environmental data to investigate dispersal and invasion.     

Non‐equilibrium in plant distribution models – only an issue for introduced or dispersal limited species?

Species distribution models rely on the assumption that species' distributions are at equilibrium with environmental conditions within a region – i.e. they occur in all suitable habitats. If this assumption holds, species occurrence should be predictable from measures of the environment. Introduced species may be poor candidates for distribution models due to their presumed lack of equilibrium within the landscapes they occupy, although predicting their potential distributions is often of critical importance to natural resource managers. We determined if the accuracy of species distribution models differed between 17 native and 17 introduced riparian plant species in the western United States. We also assessed if model accuracy was associated with both environmental and biological factors that can influence dispersal. We used Random Forests to model species distributions and linear regression to determine if model accuracy was associated with dispersal‐related traits. Model accuracy for introduced species was higher than that for native species. Dispersal‐related traits did not affect model accuracy or improvement, though two other traits, family affiliation and rarity on the landscape, did have an effect. Distance‐based measures of dispersal potential improved model fit equally for both native and introduced species and for species with a variety of dispersal traits, suggesting that the importance of regional propagule pressure is relatively constant across species with different dispersal opportunities. Several lines of future questioning are suggested by our results, including why introduced species may in some cases produce more accurate distribution models than native species and how species dispersal traits relate to distribution model accuracy at different spatial scales.     

Evolution of increased cold tolerance during range expansion of the elongate hemlock scale Fiorinia externa Ferris (Hemiptera: Diaspididae)

Abstract 1. Alien species often face novel challenges to their spread and population growth. One critical hurdle often involves an organism’s ability to tolerate environmental extremes characteristic of their invaded range. Although abiotic factors often determine range limits, there is less evidence for local adaptation in invasive organisms whose initial arrival and rapid population growth is separated by a lengthy lag period. 2. The invasive elongate hemlock scale Fiorinia externa feeds on Eastern hemlock, Tsuga canadensis, on the east coast of North America. Following its 1908 arrival, it remained localised until entering a period of rapid northward range expansion in the 1970s. 3. The present study tested the survival of overwintering F. externa populations from four sites in the north and south of the invaded range (n = 8 sites total) when exposed to ?15 °C for 0–36 h. 4. This experiment was repeated on F. externa offspring that had been reared in a common‐garden environment in order to control for parental effects. 5. Northern populations were more tolerant of exposure to cold temperatures than were southern populations. This held true in both the source‐population experiment and common‐garden experiment. The common‐garden experiment demonstrates that this difference has a genetic basis and may be the consequence of local adaptation to lower winter temperatures. 6. The results provide evidence for local adaptation to extreme temperatures in F. externa. This provides one possible explanation for the lag period between the arrival of this species and its eventual northward range expansion.     

Nondetection sampling bias in marked presence‐only data

相似文献   

Invasibility and species richness of island endemic arthropods: a general model of endemic vs. exotic species

Aim This paper has two objectives. First, we examine how a variety of biotic, abiotic and anthropogenic factors influence the endemic and introduced arthropod richness on an oceanic island. Second, we look at the relationship between the endemic and introduced arthropod richness, to ask whether areas with high levels of endemic species richness deter invasions. Location The work was carried out on a young volcanic island, Terceira, in the Azores. Methods We used standard techniques to collect data on arthropod species richness. Environmental data were obtained from the CIELO climatic model and using GIS. The explanatory value of environmental variables on a small‐scale gradient of endemic and exotic arthropod species richness was examined with generalized linear models (GLMs). In addition, the impact of both endemic and exotic species richness in the communities was assessed by entering them after the environmental variable(s) to see if they contributed significantly to the final model (the hierarchical method). Results Abiotic (climatic and geomorphological) variables gave a better explanation of the variation in endemic species richness, whereas anthropogenic variables explained most of the variation in introduced species richness. Furthermore, after accounting for all environmental variables, part of the unexplained variance in the endemic species richness is explained by the introduced species richness and vice‐versa. That is, areas with high levels of endemic species richness had many introduced species. There is evidence of a somewhat inverse spatial distribution between a group of oceanic‐type, forest‐dwelling, endemic, relict arthropods and a group of more generalist endemic arthropods that are able to survive in disturbed marginal sites particularly rich in non‐indigenous species. Main conclusions Richness of endemic species is mainly driven by abiotic factors such as a climatic axis (oceanic‐type localities with lower temperatures and summer precipitations) and a binary variable CALD (location of sites in caldeiras or ravines), whereas richness of introduced species depends on disturbance related factors. However, after factoring out these major influences, there is a correlation between endemic and introduced richness, suggesting that – independent of the environmental and geographical factors that affect the distribution of endemic or introduced species – the richest endemic assemblages are more prone to invasion, due probably to a facilitation process. Inconclusive evidence suggests that non‐indigenous species are limited to those sites under anthropogenic influence located mainly near forest edges, but the rate of expansion of those species to high‐altitude, core pristine sites has still to be tested.     

中国外来入侵生物的空间分布格局及其影响因素

不同地理区域影响生物分布格局的因子不同,对外来入侵物种也是如此。在区域尺度上分析外来入侵生物的空间分布格局及其影响因子对预测生物入侵的影响及入侵种的控制管理具有重要意义。本研究应用中国外来入侵物种数据库、自然环境数据库和社会人文环境数据库,分析了我国外来入侵动植物的空间格局;并运用主成分分析（PCA）和典范对应分析（CCA）探讨了自然环境和人类活动等因素对外来入侵动植物分布格局的影响;同时研究了外来入侵物种多样性与本地物种多样性之间的关系。结果表明,我国现有外来入侵动物138种、入侵植物384种,其数量和密度都呈现出由东南沿海向西北内陆减少的趋势,且入侵动物和入侵植物空间格局基本一致;降水（MAP）是决定我国外来入侵动植物分布格局的主要自然环境因子,国民生产总值（GDP）是主要社会经济影响因子。在全国尺度上,外来入侵物种多样性与本地物种多样性之间呈显著的正相关关系,但地域间存在较大差异。不同区域外来入侵物种与本土物种多样性的相关关系表现出不同,与研究尺度有密切关系。     

Joint species distribution modelling for spatio‐temporal occurrence and ordinal abundance data

Aim

Species distribution models are important tools used to study the distribution and abundance of organisms relative to abiotic variables. Dynamic local interactions among species in a community can affect abundance. The abundance of a single species may not be at equilibrium with the environment for spreading invasive species and species that are range shifting because of climate change. Innovation : We develop methods for incorporating temporal processes into a spatial joint species distribution model for presence/absence and ordinal abundance data. We model non‐equilibrium conditions via a temporal random effect and temporal dynamics with a vector‐autoregressive process allowing for intra‐ and interspecific dependence between co‐occurring species. The autoregressive term captures how the abundance of each species can enhance or inhibit its own subsequent abundance or the subsequent abundance of other species in the community and is well suited for a ‘community modules’ approach of strongly interacting species within a food web. R code is provided for fitting multispecies models within a Bayesian framework for ordinal data with any number of locations, time points, covariates and ordinal categories.

Main conclusions

We model ordinal abundance data of two invasive insects (hemlock woolly adelgid and elongate hemlock scale) that share a host tree and were undergoing northwards range expansion in the eastern U.S.A. during the period 1997–2011. Accounting for range expansion and high inter‐annual variability in abundance led to improved estimation of the species–environment relationships. We would have erroneously concluded that winter temperatures did not affect scale abundance had we not accounted for the range expansion of scale. The autoregressive component revealed weak evidence for commensalism, in which adelgid may have predisposed hemlock stands for subsequent infestation by scale. Residual spatial dependence indicated that an unmeasured variable additionally affected scale abundance. Our robust modelling approach could provide similar insights for other community modules of co‐occurring species.     

Spatial autocorrelation in plant communities: vegetation texture versus species composition

Vegetation texture, describing plant communities by the range of characters present in them, can describe functional variation more effectively than using the identities of the species. We ask whether spatial trends, as seen in spatial autocorrelation (SAc), are stronger in vegetation texture. We also ask in what environment SAc is stronger, whether local or over longer distances, and with what measure of distance. Roadside vegetation was sampled across an orographic region of the South Island, New Zealand, from a high‐rainfall zone up to just above treeline (the “Wet” side) and then down into a rain‐shadow area (the “Dry” side). Species composition was recorded in quadrats placed in 30 clusters. Five functional plant characters were measured on all species encountered: characters that have been reported in the literature to correlate with light‐capture, heat budget and nutrient strategy. SAc was calculated using a saturation‐response type of relation in species composition, and also in texture using the five functional characters. There was significant species‐composition SAc in almost all comparisons, but the maximum percentage of variation explained was 25%. The fit was almost always higher using local, short‐distance, comparisons than using comparisons over a whole side (Wet or Dry). On average dissimilarity had reached 90% of its asymptote after ca 250 m. It is concluded that the SAc was mainly due to local factors. SAc relations almost always predicted the presence/absence of species more accurately than their abundance. Distances along the road generally gave a very slightly better fit to species composition than surface distances, suggesting that there might be dispersal limitation. On the Wet side, elevation in overall comparisons gave a still better fit, implying environmental control too. The strength of SAc in texture was strongest on the Dry side, and using overall comparisons, abundance weighting and surface distances, but in other cases texture SAc was weaker than SAc in species composition. It is concluded that the saturation‐response formula used here has theoretical advantages over previous approaches. The SAc in this vegetation seems to be caused by a combination of dispersal limitation, broad environmental trends and especially local environmental effects such as disturbance. SAc in texture may have been weaker overall on the Wet side than on the Dry because the main environmental difference across the former is in frost, and characters correlated with frost resistance have not yet been included in texture analyses.     

Processes at multiple scales affect richness and similarity of non‐native plant species in mountains around the world

Aim To investigate how species richness and similarity of non‐native plants varies along gradients of elevation and human disturbance. Location Eight mountain regions on four continents and two oceanic islands. Methods We compared the distribution of non‐native plant species along roads in eight mountainous regions. Within each region, abundance of plant species was recorded at 41–84 sites along elevational gradients using 100‐m² plots located 0, 25 and 75 m from roadsides. We used mixed‐effects models to examine how local variation in species richness and similarity were affected by processes at three scales: among regions (global), along elevational gradients (regional) and with distance from the road (local). We used model selection and information criteria to choose best‐fit models of species richness along elevational gradients. We performed a hierarchical clustering of similarity to investigate human‐related factors and environmental filtering as potential drivers at the global scale. Results Species richness and similarity of non‐native plant species along elevational gradients were strongly influenced by factors operating at scales ranging from 100 m to 1000s of km. Non‐native species richness was highest in the New World regions, reflecting the effects of colonization from Europe. Similarity among regions was low and due mainly to certain Eurasian species, mostly native to temperate Europe, occurring in all New World regions. Elevation and distance from the road explained little of the variation in similarity. The elevational distribution of non‐native species richness varied, but was always greatest in the lower third of the range. In all regions, non‐native species richness declined away from roadsides. In three regions, this decline was steeper at higher elevations, and there was an interaction between distance and elevation. Main conclusions Because non‐native plant species are affected by processes operating at global, regional and local scales, a multi‐scale perspective is needed to understand their patterns of distribution. The processes involved include global dispersal, filtering along elevational gradients and differential establishment with distance from roadsides.