Distribution Modelling Methods Research Articles

The relative contributions of competition and abiotic tolerances in determining the geographical distributions of four closely related Yucca species in Texas

AbstractAimThe determinants of species’ geographical ranges are the result of both abiotic and biotic factors. Among co‐occurring closely related species, some degree of differentiation in niche characteristics is assumed necessary to allow continued coexistence. As niche requirements are likely to show similarity or overlap, competition among such species may also restrict range limits and help maintain species boundaries. We examined the roles of niche compartmentalization and competition to assess their relative contributions in determining geographical distributions of four closely related species of Yucca co‐occurring across a large region of Texas. These species are pollinated by the same moth species, but each yucca is a distinct biological species.LocationTexas, USA.MethodsWe used species distribution modelling and niche evaluation methods based on extensive field surveys to evaluate the roles of abiotic factors and competition. We also conducted surveys of flowering activity to determine if species boundaries may be maintained by species‐specific shifts in flowering phenology that would limit interspecific gene flow.ResultsNiche evaluation methods demonstrate unique niches mostly consistent with the defined ecological regions of Texas. Geographical distributions of yuccas are mostly parapatric with some regions of sympatry. Species distribution models indicate that most species pairings display a region of potential sympatry where one of the species is numerically dominant, suggesting that competitive exclusion is delineating species ranges at parapatric margins. Flowering activity data indicate the potential for cross‐pollination among these species. Nevertheless, there is little evidence for hybridization, suggesting that strong selection due to niche requirements is mediating species boundaries.Main conclusionsResults suggest that adaptation to abiotic conditions has been a major force underpinning diversification in these plants. However, competition at the edge of species’ ranges likely limits the spread of any one species, leading to uncharacteristically high species diversity over a relatively small geographical area.

Modelling of high‐power IGBT module short‐circuit operation and current distribution by a behavioural model

This study presents current distribution modelling methods for a dual pack insulated-gate bipolar transistor (IGBT) module, which is operating under short-circuit conditions. The parasitic components of the commutation circuit current conductors and the internal busbars of the module are modelled by a three-dimensional finite-element method and the IGBT chips by a behavioural semiconductor model. The lumped IGBT chip model is characterised by using datasheet information only and without fine tuning of the model parameters. The mechanical dimensions of the IGBT module and the experimental setup are used to define the parasitic components for the commutation circuit. The model accuracy is investigated by comparing simulated dynamic and steady-state short-circuit parameters with experimental ones. Furthermore, uneven current distribution in the module main current terminals is observed in the simulations, and the behaviour is verified by experimental tests. The phenomenon may cause additional mechanical and electrical stresses in the device for instance because of thermal expansion and current couplings. The results show that the proposed methods are suitable to model the IGBT short-circuit behaviour and current distribution inside the module and can be used for example in the virtual design of power electronic inverter or switched power supply.

Do community-level models account for the effects of biotic interactions? A comparison of community-level and species distribution modeling of Rocky Mountain conifers

Community-level models (CLMs) aim to improve species distribution modeling (SDM) methods by attempting to explicitly incorporate the influences of interacting species. However, the ability of CLMs to appropriately account for biotic interactions is unclear. We applied CLM and SDM methods to predict the distributions of three dominant conifer tree species in the U.S. Rocky Mountains and compared CLM and SDM predictive accuracy as well as the ability of each approach to accurately reproduce species co-occurrence patterns. We specifically evaluated the performance of two statistical algorithms, MARS and CForest, within both CLM and SDM frameworks. Across all species, differences in SDM and CLM predictive accuracy were slight and can be attributed to differences in model structure rather than accounting for the effects of biotic interactions. In addition, CLMs generally over-predicted species co-occurrence, while SDMs under-predicted co-occurrence. Our results demonstrate no real improvement in the ability of CLMs to account for biotic interactions relative to SDMs. We conclude that alternative modeling approaches are needed in order to accurately account for the effects of biotic interactions on species distributions.

How important are choice of model selection method and spatial autocorrelation of presence data for distribution modelling by MaxEnt?

In this paper we explore new options for model selection in the distribution modelling method MaxEnt, offered by the strict maximum likelihood explanation of the method. These options, which are collectively referred to as the ‘alternative MaxEnt procedure’ (aMp), imply manual forward stepwise selection of variables and an F-ratio test to compare nested models. We compared aMp models with models obtained by the default variable transformation and model selection procedures in the popular Maxent.jar software (‘default MaxEnt procedure’; dMp), which invokes ℓ1-regularisation. We used swamp-forest presence in SE Norway as the target for our modelling. Eleven terrain variables derived from a LiDAR-based digital elevation map, making up three groups of internally correlated variables, were used as predictors in the models. An independently collected presence/absence ‘test’ data set was used for evaluation of models. Although dMp models were much more complex than aMp models, with up to 240 variables derived from the original 11, compared to a maximum number of derived variables in aMp models of 14, no significant difference in performance between dMp and aMp models was found. This indicates that the regularisation procedure is not the main reason for the good performance of MaxEnt (as implemented in Maxent.jar) in comparative studies. Advantages of the more flexible aMp procedure, which produces simpler models, are discussed. Effects of spatial autocorrelation in the response variable on modelling results were addressed by obtaining parallel MaxEnt models for two response variables that differ in the number of presence observations: the centroid of the 121 swamp forests and the 7175 grid cells, each 5m×5m, in which swamp forest was recorded as present. Our results are not conclusive with respect to the effect of spatial autocorrelation in the response variable on the predictive performance of distribution models by MaxEnt.

Methods to select areas to survey for biological control agents: An example based on growth in relation to temperature and distribution of the weed Conyza bonariensis

A novel approach for selecting areas to survey for biological control agents, incorporating climate and a hypothesised biological control agent, is demonstrated using the target weed Conyza bonariensis (Asteraceae). This weed has become important in Australian cropping regions due to its persistence and herbicide resistance, and it is also increasingly an environmental weed. Both are reasons for the investigation of biological control options. We developed a species niche model for C. bonariensis in CLIMEX based on parameters informed by plant growth and distribution of the species in the Americas. A hypothetical biological control agent (HBCA-cold) was proposed that has its ideal growth range 5°C below that of the weed, so as to favour development of the agent over that of the weed in parts of Australia. The southern part of the weed’s native distribution in Argentina, Chile and the highlands of Ecuador and Columbia were identified as the most suitable areas for surveys that take into account both the climate suitable for the HBCA-cold and the target regions in Australia. This was compared to a model (HBCA-hot) that had an ideal growth range 5°C above that of the weed, but which identified potential areas for surveys in South America that were not climatically aligned with the main regions of the weed’s economic impact in Australia. This species distribution modelling method allows for prioritisation of search areas for biological control agents in the case of widespread target species such as C. bonariensis.

Evaluating methods to quantify spatial variation in the velocity of biological invasions

Invading species rarely spread homogeneously through a landscape and invasion patterns typically display irregular frontal boundaries as the invasion progresses through space. Those irregular patterns are generally produced by local environmental factors that may slow or accelerate movement of the frontal boundary. While there is an abundant literature on species distribution modelling methods that quantify local suitability for species establishment, comparatively few studies have examined methods for measuring the local velocity of invasions that can then be statistically analysed in relation to spatially variable environmental factors. Previous studies have used simulations to compare different methods for estimating the overall rate of spread of an invasion. We adopted a similar approach of simulating invasions that resemble two real case‐studies, both in terms of their spatial resolution (i.e. considering the size of one cell as one km) and their spatial extent (> 600 000 km²). Simulations were sampled to compare how different methods used to measure local spread rate, namely the neighbouring, nearest distance and Delaunay methods, perform for spatio‐temporal comparisons. We varied the assessment using three levels of complexity of the spatio‐temporal pattern of invasion, three sample sizes (500, 1000 and 2000 points), three different spatial sampling patterns (stratified, random, aggregated), three interpolation methods (generalized linear model, kriging, thin plate spline regression) and two spatio‐temporal modelling structures (trend surface analysis and boundary displacement), resulting in a total of 486 different scenarios. The thin plate spline regression interpolation method, in combination with trend surface analysis, was found to provide the most robust local spread rate quantification as it was able to reliably accommodate different sampling conditions and invasion patterns. This best approach was successfully applied to two case‐studies, the invasion of France by the horse‐chestnut leafminerCameraria ohridellaand by the bluetongue virus, generally in agreement with previously published values of spread rates. Potential avenues for further research are discussed.

Predicted impacts of climatic change on ant functional diversity and distributions in eastern North American forests

AbstractAimsClimatic change is expected to rearrange species assemblages and ultimately affect organism‐mediated ecosystem processes. We focus on identifying patterns and relationships between common ant species (representing 99% of total ant records) richness and functional diversity; modelling how these patterns may change at local and regional scales in future climatic conditions; and interpreting how these changes might influence ant‐mediated ecosystem processes.LocationForested ecosystems of eastern North America.MethodsWe used a previously published dataset to evaluate functional diversity at 67 sites in the eastern U.S. and quantified 14 taxonomic, morphometric and natural history traits for 70 common ant species in the region. We used functional diversity metrics, functional groups and species distribution modelling methods to address our aims. We used stacked species distribution models and stacked functional group models to predict species assemblages and functional richness at the 67 sites and at a regional scale for current and future climatic conditions.ResultsSpecies richness and functional diversity are positively correlated throughout the region. Under future climate scenarios, species richness and functional group richness were predicted to decrease in southern ecoregions and increase in northern ecoregions. This may be due to increased thermal stress for species in the southern extent of their ranges and increased habitat suitability in the northern ecoregions. Decomposers, arthropod community regulators and seed dispersers are forecast to be the most threatened ant functional groups.Main ConclusionsClimate change will likely lead to major changes in ant species richness and functional group richness in the forests of the north‐eastern United States, and this may substantially alter ant‐mediated ecosystem processes and services.

Modelling the potential spatial distribution of mosquito species using three different techniques.

BackgroundModels for the spatial distribution of vector species are important tools in the assessment of the risk of establishment and subsequent spread of vector-borne diseases. The aims of this study are to define the environmental conditions suitable for several mosquito species through species distribution modelling techniques, and to compare the results produced with the different techniques.MethodsThree different modelling techniques, i.e., non-linear discriminant analysis, random forest and generalised linear model, were used to investigate the environmental suitability in the Netherlands for three indigenous mosquito species (Culiseta annulata, Anopheles claviger and Ochlerotatus punctor). Results obtained with the three statistical models were compared with regard to: (i) environmental suitability maps, (ii) environmental variables associated with occurrence, (iii) model evaluation.ResultsThe models indicated that precipitation, temperature and population density were associated with the occurrence of Cs. annulata and An. claviger, whereas land surface temperature and vegetation indices were associated with the presence of Oc. punctor. The maps produced with the three different modelling techniques showed consistent spatial patterns for each species, but differences in the ranges of the predictions. Non-linear discriminant analysis had lower predictions than other methods. The model with the best classification skills for all the species was the random forest model, with specificity values ranging from 0.89 to 0.91, and sensitivity values ranging from 0.64 to 0.95.ConclusionsWe mapped the environmental suitability for three mosquito species with three different modelling techniques. For each species, the maps showed consistent spatial patterns, but the level of predicted environmental suitability differed; NLDA gave lower predicted probabilities of presence than the other two methods. The variables selected as important in the models were in agreement with the existing knowledge about these species. All model predictions had a satisfactory to excellent accuracy; best accuracy was obtained with random forest. The insights obtained can be used to gain more knowledge on vector and non-vector mosquito species. The output of this type of distribution modelling methods can, for example, be used as input for epidemiological models of vector-borne diseases.

Evidence for cryptic northern refugia in the last glacial period in Cryptomeria japonica.

Distribution shifts and natural selection during past climatic changes are important factors in determining the genetic structure of forest species. In particular, climatic fluctuations during the Quaternary appear to have caused changes in the distribution ranges of plants, and thus strongly affected their genetic structure. This study was undertaken to identify the responses of the conifer Cryptomeria japonica, endemic to the Japanese Archipelago, to past climatic changes using a combination of phylogeography and species distribution modelling (SDM) methods. Specifically, this study focused on the locations of refugia during the last glacial maximum (LGM). Genetic diversity and structure were examined using 20 microsatellite markers in 37 populations of C. japonica. The locations of glacial refugia were assessed using STRUCTURE analysis, and potential habitats under current and past climate conditions were predicted using SDM. The process of genetic divergence was also examined using the approximate Bayesian computation procedure (ABC) in DIY ABC to test the divergence time between the gene pools detected by the STRUCTURE analysis. STRUCTURE analysis identified four gene pools: northern Tohoku district; from Chubu to Chugoku district; from Tohoku to Shikoku district on the Pacific Ocean side of the Archipelago; and Yakushima Island. DIY ABC analysis indicated that the four gene pools diverged at the same time before the LGM. SDM also indicated potential northern cryptic refugia. The combined evidence from microsatellites and SDM clearly indicates that climatic changes have shaped the genetic structure of C. japonica. The gene pool detected in northern Tohoku district is likely to have been established by cryptic northern refugia on the coast of the Japan Sea to the west of the Archipelago. The gene pool in Yakushima Island can probably be explained simply by long-term isolation from the other gene pools since the LGM. These results are supported by those of SDM and the predicted divergence time determined using ABC analysis.

Estimating the distribution of the imperiled pugnose shiner (Notropis anogenus) in the St. Lawrence River using a habitat model

The pugnose shiner, Notropis anogenus, is a small minnow that occurs in the Great Lakes basin and Upper Mississippi River basin. It was listed as ‘Endangered’ under the Canadian Species at Risk Act (SARA) and by the State of New York, largely due to its rarity, fragmented distribution, and degraded habitat. Our objective was to use species distribution modeling to better understand the spatial extent of suitable habitat for the pugnose shiner in the upper St. Lawrence River and to guide protection of habitat for this endangered species. We performed our analyses with MaxEnt, a species distribution modeling method based on maximum entropy. The pugnose shiner is associated with shallow areas of lakes or slow-moving rivers with abundant submerged aquatic vegetation. Therefore, we created a model based on depth, water velocity, and aquatic vegetation. For water depth and velocity, we used results of a two-dimensional hydrodynamic model from the Upper St. Lawrence River generated and calibrated by Environment Canada. Aquatic vegetation was estimated from a simple algorithm based on water depth and velocity. To minimize the effect of sampling bias in the analysis (i.e., sampling occurred predominantly in shallow waters), we also used restricted depth ranges in model generation. Our model produced highly significant results when depth was not restricted, and significant results for analyses with restricted depth ranges. Our results suggest an abundance of potentially suitable habitat for the pugnose shiner in the Upper St. Lawrence River, which exceeded the minimum area for viable population estimates for this species.

A new statistical framework for the quantification of covariate associations with species distributions

SummaryIdentifying processes that shape species geographical ranges is a prerequisite for understanding environmental change. Currently, species distribution modelling methods do not offer credible statistical tests of the relative influence of climate factors and typically ignore other processes (e.g. biotic interactions and dispersal limitation).We use a hierarchical model fitted with Markov Chain Monte Carlo to combine ecologically plausible niche structures using regression splines to describe unimodal but potentially skewed response terms. We apply spatially explicit error terms that account for (and may help identify) missing variables.Using three example distributions ofEuropean bird species, we map model results to show sensitivity to change in each covariate. We show that the overall strength of climatic association differs between species and that each species has considerable spatial variation in both the strength of the climatic association and the sensitivity to climate change.Our methods are widely applicable to many species distribution modelling problems and enable accurate assessment of the statistical importance of biotic and abiotic influences on distributions.

A Potential Distribution Model and Conservation Plan for the Critically Endangered Ecuadorian Capuchin, Cebus albifrons aequatorialis

Conservation actions that effectively and efficiently target single, highly threatened species require current data on the species’ geographic distribution and environmental associations. The Ecuadorian capuchin (Cebus albifrons aequatorialis) is a critically endangered primate found only in the fragmented forests of western Ecuador and northern Peru, which are among the world’s most severely threatened ecosystems. We use the MAXENT species distribution modeling method to model the potential distribution and environmental associations of Cebus albifrons aequatorialis, using all known presence localities recorded within the last 2 decades as well as 13 climate, topography, vegetation, and land-use data sets covering the entire geographic range of the subspecies. The environmental conditions that our model predicted to be ideal for supporting Cebus albifrons aequatorialis included ≥20% tree cover, mild temperature seasonality, annual precipitation <2000 mm, and low human population density. Our model identified 5028 km2 of suitable habitat remaining, although many of these forest fragments are unprotected and are unlikely to support extant populations. Using the median population density across all sites for which data are available, we estimate the total carrying capacity of the remaining habitat to be 12,500 total individuals. The true number of remaining individuals is likely to be considerably lower due to anthropogenic factors. We highlight four critical regions of high predicted suitability in western Ecuador and northern Peru on which immediate conservation actions should focus, and we lay out clear priorities to guide conservation actions for ensuring the long-term survival of this gravely threatened and little known primate.

Teaching biogeographical concepts and conservation biogeography to Brazilian undergraduate students

symposium summary ISSN 1948‐6596 Teaching biogeographic concepts and conservation biogeography 1 st Symposium of Systematics and Biogeography of Minas Gerais State — Diamantina, Bra‐ zil,15 th –17 th October 2012 Systematics and biogeography are two fields of natural sciences which have grown considerably in recent years. Both disciplines have been instru‐ mental in driving knowledge of the Brazilian Cer‐ rado, a global hotspot for biodiversity (Myers et al. 2000), and especially the Espinhaco Mountain Range. This range, at more than 1200 km, repre‐ sents the longest orogenic belt of Brazil. The land‐ scape is composed of a mosaic of bare exposed rocks and dense, mainly riparian, forests. It is a unique region housing many endemic species and its southern portion provided a spectacular back‐ drop for the 1 st Symposium of Systematics and Bio‐ geography which was convened in part to redress the still limited disciplinary training available for undergraduate students despite the rapid devel‐ opment of biogeography in Brazil. The relationship between Systematics and Biogeography was considered first by outlining the historical development of Comparative Biol‐ ogy, stressing the importance of taxonomy and its profound impact in studies of evolution and bio‐ geography. Modern biogeography was invigorated by Leon Croizat, who famously recognized that “Life and Earth evolve together” (Croizat, 1964), and thus Eduardo Almeida (University of Sao Paulo) posited that all biogeographers are, ulti‐ mately, systematic researchers of specific organ‐ isms. That topic was complemented by considera‐ tion of the historical development of biogeogra‐ phy, including the importance of biogeographical methods for investigating dispersal or vicariant events, the latter analyzed under panbio‐ geographic or cladistic perspectives (Silvio Nihei, University of Sao Paulo). Subsequently, both Almeida and Nihei discussed the current state of biogeography in Brazil, illustrating important con‐ tributions by development of computer programs for biogeographic analysis and the chronology of the highest impact publications. These accom‐ plishments contrast with the lack of recognition of biogeography as a post‐graduate subject in Brazil and the lack of support for PhD researchers, which may in the future constrain biogeographic knowl‐ edge in Brazil. Almeida and Nihei also highlighted the possible unhealthy trade‐off between quantity and quality of knowledge driven by the ‘publish‐or ‐perish’ culture that is becoming ever more appar‐ ent in Brazilian institutions. Better understanding of cutting edge biogeographical methods among Brazilian students and researchers is essential to move beyond simplistic hypotheses for distinct taxa. Discussion subsequently focused on Geo‐ graphical Distribution Areas and Endemism, a topic introduced by Marcio Bernardino da Silva (Federal University of Paraiba) talking about the development of the endemism concept. In the early 18 th Century, biogeographers considered areas of endemism as simply species ranges de‐ limited by geographical barriers. Today, areas of endemism are the basic working unit of historical biogeography, and their delimitation is of funda‐ mental importance. Silva presented his own re‐ search on endemic areas in the Atlantic Forest using harvestman. Gleyce Dutra (Federal Univer‐ sity of Jequitinhonha and Mucuri Valleys [UFVJM]) then outlined the importance of geoprocessing tools for species distribution modeling (SDM) methods, stressing the difference between the actual niche, realized niche, and fundamental niche, and that the precision of predicted distribu‐ tions is dependent on the amount and quality of the input data. She then demonstrated the effi‐ ciency of SDM compared to the still widespread practice of uniting points of occurrence. Silva and Dutra debated the use of endemic areas as a premise for delimiting areas for conservation. Silva argued that, despite the wide availability of various tools for delimiting endemic areas, they are not being incorporated into conservation pol‐ icy. Dutra cautioned about the misapplication of such tools for making conservation decisions. These concepts were illustrated by case studies in modern biogeography of the Cerrado and the Espinhaco Mountain Range. Growth in frontiers of biogeography 5.2, 2013 — © 2013 the authors; journal compilation © 2013 The International Biogeography Society

Recognising fuzzy vegetation pattern: the spatial prediction of floristically defined fuzzy communities using species distribution modelling methods

AbstractQuestionPlant communities are not necessarily spatially exclusive; a point in space can exhibit properties of multiple communities. Such variation can be described using floristically defined ‘fuzzy’ units, however these may not be easily delineated using standard remote sensing methods. Is there value in considering communities as fuzzy? Can species distribution modelling methods be used to represent fuzzy communities spatially?LocationWestern Victoria, Australia.MethodsFuzzy communities were objectively identified from vegetation census quadrats with a cluster analysis of ordinated species data. Boosted regression trees were used to create models that defined relationships between the sampled communities and environmental predictor variables. These were applied to the mapped predictors to create maps of estimated fuzzy community membership for the entire study area.ResultsFour separate fuzzy communities were identified from the sampled vegetation data. Models were created for each community and these were effectively used to generate maps of fuzzy community membership. Individual fuzzy community maps illustrated vegetation variation that could not be discerned on a discretely classified map.ConclusionsFuzzy communities were found to represent a greater proportion of species variation than discretely classified units. Species distribution modelling methods were effective in creating independent spatial maps of each floristically defined fuzzy community; however the interpretation of these maps is more complex than with a single discrete community map.

The impact of modelling choices in the predictive performance of richness maps derived from species‐distribution models: guidelines to build better diversity models

Summary The stacking of species‐distribution models (S‐SDMs) is receiving attention by conservation researchers because this approach is capable of simultaneously predicting species richness and composition. However, the steps required to build S‐SDMs implies at least two choices that influence its predictive performance which have not been extensively assessed: the selection of the modelling algorithm and the application of a threshold to transform the species‐distribution models into binary maps to be added together to build the final S‐SDM. Our goal was to provide guidelines concerning the best combinations of modelling algorithms and thresholds with which to build more accurate S‐SDMs. We generated 380 S‐SDMs of 1224 tree species in Mesoamerica by combining 19 distribution modelling methods with 20 different thresholds using presence‐only data from the Global Biodiversity Information Facility. We compared the predicted richness and composition with inventory data obtained from the BIOTREE‐NET forest plot database. We designed two indicators of predictive performance that were based on the diversity factors used to measure species turnover: a (shared species between the observed and predicted compositions), b and c (the exclusive species of the predicted and observed compositions respectively) and compared them with the Sorensen and Beta‐Simpson turnover measures. Our proposed indexes and the Sorensen index proved suitable as indicators of predictive performance for S‐SDMs, whereas the Beta‐Simpson turnover measure presented issues that would prevent its application to evaluate S‐SDMs. Some modelling methods – especially machine learning and ensemble model forecasting methods performed significantly better than others in minimizing the error in predicted richness and composition. Our results also points out that restrictive thresholds (with high omission errors) lead to more accurate S‐SDMs in terms of species richness and composition. Here, we demonstrate that particular combinations of modelling methods and thresholds provide results with higher predictive performance. These results provide clear modelling guidelines that will help S‐SDM modellers to select the appropriate combination of modelling methods and thresholds to build more accurate S‐SDMs, and therefore will have a positive impact on the quality of the diversity models used to assist conservation planning.

Evaluating the effectiveness of conservation site networks under climate change: accounting for uncertainty

We forecasted potential impacts of climate change on the ability of a network of key sites for bird conservation (Important Bird Areas; IBAs) to provide suitable climate for 370 bird species of current conservation concern in two Asian biodiversity hotspots: the Eastern Himalaya and Lower Mekong. Comparable studies have largely not accounted for uncertainty, which may lead to inappropriate conclusions. We quantified the contribution of four sources of variation (choice of general circulation models, emission scenarios and species distribution modelling methods and variation in species distribution data) to uncertainty in forecasts and tested if our projections were robust to these uncertainties. Declines in the availability of suitable climate within the IBA network by 2100 were forecast as 'extremely likely' for 45% of species, whereas increases were projected for only 2%. Thus, we predict almost 24 times as many 'losers' as 'winners'. However, for no species was suitable climate 'extremely likely' to be completely lost from the network. Considerable turnover (median=43%, 95% CI=35-69%) in species compositions of most IBAs were projected by 2100. Climatic conditions in 47% of IBAs were projected as 'extremely likely' to become suitable for fewer priority species. However, no IBA was forecast to become suitable for more species. Variation among General Circulation Models and Species Distribution Models contributed most to uncertainty among forecasts. This uncertainty precluded firm conclusions for 53% of species and IBAs because 95% confidence intervals included projections of no change. Considering this uncertainty, however, allows robust recommendations concerning the remaining species and IBAs. Overall, while the IBA network will continue to sustain bird conservation, climate change will modify which species each site will be suitable for. Thus, adaptive management of the network, including modified site conservation strategies and facilitating species' movement among sites, is critical to ensure effective future conservation.

Use of ecological niche models to predict the distribution of invasive species: a scientometric analysis

We conducted a scientometric analysis to determine the main trends and gaps of studies on the use of ecological niche models (ENMs) to predict the distribution of invasive species. We used the database of the Thomson Institute for Scientific Information (ISI). We found 190 papers published between 1991 and 2010 in 82 journals. The number of papers was low in the 1990s, but began to increase after 2003. One-third of the papers were published by researchers from the United States of America, and consequently, the USA was also the most studied region. The majority of studies were carried out in terrestrial environments, while only a few investigated aquatic systems, probably because important aquatic predictor variables are scarce or unavailable for most regions in the world. Species-occurrence records were mainly composed of presence-only records, and almost 70% of the studies were carried out with plants and insects. Twenty-three different distribution modelling methods were used. The Genetic Algorithm for Rule-set Production (GARP) was used most often. Our scientometric analysis showed a growing interest in the use of ENMs to predict the distribution of invasive species, especially in the last decade, which is probably related to the increase in species introductions worldwide. Among some important gaps that need to be filled, the relatively small number of studies conducted in developing countries and in aquatic environments deserves careful attention.

Use of predictive habitat modelling to assess the distribution and extent of the current protection of ‘listed’ deep‐sea habitats

AbstractAimTo demonstrate the application of predictive species distribution modelling methods to habitat mapping and assessment of percentage area‐based conservation targets.LocationTheNE Atlantic deep sea (UKand Irish extended continental shelf limits).MethodsMaxEnt modelling of three listed habitats (Lophelia pertusa(Linnaeus, 1758) reef (LpReef),Pheronema carpenteri(WyvilleThomson, 1869) aggregations (PcAggs) andSyringammina fragilissima(Brady, 1883) aggregations (SfAggs)), with some pre‐selection of variables by generalized additive modelling. Models are validated using repeated 70/30 build/test data splits usingAUCand threshold‐dependent assessment methods. Predicted distribution maps are used to assess the adequacy of existing area closures for the protection of listed habitats and to assess percentage representation of each community within existingMPAnetworks.ResultsModel performances are rated as fair (LpReef), excellent (PcAggs) and good (SfAggs). Current closures are focused on the protection of cold‐water coral reef and incidentally capture someSfAggs suitable environments, but largely fail to protectPcAggs. Considering the wider network ofMPAs in the study region, approximately 23% (LpReef), 2% (PcAggs) and 6% (SfAggs) of the area predicted as suitable for each habitat respectively is contained within anMPA.Main conclusionsTo date, decisions on area closures for the protection of ‘listed’ deep‐sea habitats have been based on maps of recorded presence of species that are taken as being indicative of that habitat. Predictive habitat modelling may provide a useful method of better estimating the extent of listed habitats, providing direction for futureMPAestablishment and a means of assessingMPAnetwork effectiveness against politically set percentage targets. Given the coarse resolution of the model, percentages should be taken as maximal figures, with habitat occurrence likely to be less prevalent in reality.

Multiple methods, maps, and management applications: Purpose made seafloor maps in support of ocean management

The establishment of multibeam echosounders (MBES) as a mainstream tool in ocean mapping has facilitated integrative approaches toward nautical charting, benthic habitat mapping, and seafloor geotechnical surveys. The inherent bathymetric and backscatter information generated by MBES enables marine scientists to present highly accurate bathymetric data with a spatial resolution closely matching that of terrestrial mapping. Furthermore, developments in data collection and processing of MBES backscatter, combined with the quality of the co-registered depth information, have resulted in the increasing preferential use of multibeam technology over conventional sidescan sonar for the production of benthic habitat maps. A range of post-processing approaches can generate customized map products to meet multiple ocean management needs, thus extracting maximum value from a single survey data set. Based on recent studies over German Bank off SW Nova Scotia, Canada, we show how primary MBES bathymetric and backscatter data, along with supplementary data (i.e. in situ video and stills), were processed using a variety of methods to generate a series of maps. Methods conventionally used for classification of multi-spectral data were tested for classification of the MBES data set to produce a map summarizing broad bio-physical characteristics of the seafloor (i.e. a benthoscape map), which is of value for use in many aspects of marine spatial planning. A species-specific habitat map for the sea scallop Placopecten magellanicus was also generated from the MBES data by applying a Species Distribution Modeling (SDM) method to spatially predict habitat suitability, which offers tremendous promise for use in fisheries management. In addition, we explore the challenges of incorporating benthic community data into maps based on species information derived from a large number of seafloor photographs. Through the process of applying multiple methods to generate multiple maps for management applications, we demonstrate the efficient use of survey data sets to maximize the benefit to a wide number of potential end users, and to facilitate the move toward an ecosystem-based approach to management.

Modelling commercial fish distributions: Prediction and assessment using different approaches

Species distribution models are important tools to explore the effects of future global change on biodiversity. Specifically, AquaMaps, Maxent and the Sea Around Us Project algorithm are three approaches that have been applied to predict distributions of marine fishes and invertebrates. They were designed to cope with issues of data quality and quantity common in species distribution modelling, and especially pertinent to the marine environment. However, the characteristics of model projections for marine species from these different approaches have rarely been compared. Such comparisons provide information about the robustness and uncertainty of the projections, and are thus important for spatial planning and developing management and conservation strategies. Here we apply the three commonly used species distribution modelling methods for commercial fish in the North Sea and North Atlantic, with the aim of drawing comparisons between the approaches. The effect of different assumptions within each approach on the predicted current relative habitat suitability was assessed. Predicted current distributions were tested following data partitioning and selection of pseudoabsences from within a specified distance of occurrence data. As indicated by the test statistics, each modelling method produced plausible predictions of relative habitat suitability for each species, with subsequent incorporation of expert knowledge generally improving predictions. However, because of the differences between modelling algorithms, methodologies and patterns of relative suitability, comparing models using test statistics and selecting a ‘best’ model are not recommended. We propose that a multi-model approach should be preferred and a suite of possible predictions considered if biases due to uncertainty in data and model formulation are to be minimised.