Habitat distribution models for intertidal seaweeds: responses to climatic and non‐climatic drivers

Aim Because intertidal organisms often live close to their physiological tolerance limits, they are potentially sensitive indicators of climate‐driven changes in the environment. The goals of this study were to assess the effect of climatic and non‐climatic factors on the geographical distribution of intertidal macroalgae, and to predict future distributions under different climate‐warming scenarios. Location North‐western Iberian Peninsula, southern Europe. Methods We developed distribution models for six ecologically important intertidal seaweed species. Occurrence and microhabitat data were sampled at 1‐km² resolution and analysed with climate variables measured at larger spatial scales. We used generalized linear models and applied the deviance and Bayesian information criterion to model the relationship between environmental variables and the distribution of each target species. We also used hierarchical partitioning (HP) to identify predictor variables with higher independent explanatory power. Results The distributions of Himanthalia elongata and Bifurcaria bifurcata were correlated with measures of terrestrial and marine climate, although in opposite directions. Model projections under two warming scenarios indicated the extinction of the former at a faster rate in the Cantabrian Sea (northern Spain) than in the Atlantic (west). In contrast, these models predicted an increase in the occurrence of B. bifurcata in both areas. The occurrences of Ascophyllum nodosum and Pelvetia canaliculata, species showing rather static historical distributions, were related to specific non‐climatic environmental conditions and locations, such as the location of sheltered sites. At the southernmost distributional limit, these habitats may present favourable microclimatic conditions or provide refuges from competitors or natural enemies. Model performances for Fucus vesiculosus and F. serratus were similar and poor, but several climatic variables influenced the occurrence of the latter in the HP analyses. Main conclusions The correlation between species distributions and climate was evident for two species, whereas the distributions of the others were associated with non‐climatic predictors. We hypothesize that the distribution of F. serratus responds to diverse combinations of factors in different sections of the north‐west Iberian Peninsula. Our study shows how the response of species distributions to climatic and non‐climatic variables may be complex and vary geographically. Our analyses also highlight the difficulty of making predictions based solely on variation in climatic factors measured at coarse spatial scales.     

Inferring climatic controls of rice stem borers’ spatial distributions using maximum entropy modelling

Accurate assessment of pest potential distributions is needed to identify their establishment risks that play a key role in pest management in agricultural ecosystems. We used a correlative niche modelling method (Maxent) to predict and map the spatial distributions of two important rice stem borers, Chilo suppressalis and Sesamia cretica, in paddy fields of Iran. In total, 195 presence occurrence records (101 records for C. suppressalis and 94 records for S. cretica) were compiled. A set of environmental and topographic variables, with the highest effects on the species distributions and the lowest correlations among themselves, were used. The results showed that mainly the northern parts of Iran were the most suitable areas for C. suppressalis, and north, north‐east and south‐west of Iran as the most suitable areas for S. cretica. Both models performed well, with an area under the receiver operating characteristic curve (AUC) of 0.983 and 0.786 for C. suppressalis and S. cretica, respectively. The Maxent models showed higher accuracy for predicting the distribution of the specialist pest with the small range sizes compared to the generalist species. Assessing the importance of environmental variables, which were derived from the jackknife test, showed the precipitation as the variable with the highest contribution (66%) in explaining the spatial distribution of C. suppressalis compared to the other variables. The distribution of S. cretica was influenced by a set of variables derived from both the precipitation and temperature. The Maxent predictions were useful to map the geographical distributions of the risk for both rice stem borers that is needed to develop effective management strategies.     

Predation on snails by an indigenous fish,Sargochromis codringtonii,in ponds in Zimbabwe

The extent to which Sargochromis codringtonii (a known predator of snails) can control snail numbers was investigated in cementlined ponds in order further to evaluate the fish's potential as an agent for the biological control of fresh water snails, especially Bulinus globosus which is the intermediate host for Schistosoma haematobium in Zimbabwe. Bulinus tropicus of all sizes were vulnerable to predation by fish, but both Melanoides tuberculata and Bulinus globosus with shell heights greater than 10mm seemed to be less so. Surprisingly, the density of B. globosus was lower in ponds without fish than in ponds with fish, while the density of B. tropicus was very high in ponds without fish. It was not possible to conclude whether B. globosus benefited by the presence of fish, or was less palatable to the fish than B. tropicus, or whether the high density of B. tropicus in ponds without fish exerted competitive effects on B. globosus. Consequently, further experiments are required to determine whether S. codringtonii can reduce populations of B. globosus in the absence of B. tropicus.     

The uncertain nature of absences and their importance in species distribution modelling

Species distribution models (SDM) are commonly used to obtain hypotheses on either the realized or the potential distribution of species. The reliability and meaning of these hypotheses depends on the kind of absences included in the training data, the variables used as predictors and the methods employed to parameterize the models. Information about the absence of species from certain localities is usually lacking, so pseudo‐absences are often incorporated to the training data. We explore the effect of using different kinds of pseudo‐absences on SDM results. To do this, we use presence information on Aphodius bonvouloiri, a dung beetle species of well‐known distribution. We incorporate different types of pseudo‐absences to create different sets of training data that account for absences of methodological (i.e. false absences), contingent and environmental origin. We used these datasets to calibrate SDMs with GAMs as modelling technique and climatic variables as predictors, and compare these results with geographical representations of the potential and realized distribution of the species created independently. Our results confirm the importance of the kind of absences in determining the aspect of species distribution identified through SDM. Estimations of the potential distribution require absences located farther apart in the geographic and/or environmental space than estimations of the realized distribution. Methodological absences produce overall bad models, and absences that are too far from the presence points in either the environmental or the geographic space may not be informative, yielding important overestimations. GLMs and Artificial Neural Networks yielded similar results. Synthetic discrimination measures such as the Area Under the Receiver Characteristic Curve (AUC) must be interpreted with caution, as they can produce misleading comparative results. Instead, the joint examination of ommission and comission errors provides a better understanding of the reliability of SDM results.     

Host tolerance,not symbiont tolerance,determines the distribution of coral species in relation to their environment at a Central Pacific atoll

Tolerance of environmental variables differs between corals and their dinoflagellate symbionts (Symbiodinium spp.), controlling the holobiont’s (host and symbiont combined) resilience to environmental stress. However, the ecological role that environmental variables play in holobiont distribution remains poorly understood. We compared the drivers of symbiont and coral species distributions at Palmyra Atoll, a location with a range of reef environments from low to high sediment concentrations (1–52 g dry weight m⁻² day⁻¹). We observed uniform holobiont partnerships across the atoll (e.g. Montipora spp. with Symbiodinium type C15 at all sites). Multivariate analysis revealed that field-based estimates of settling sediment predominantly explained the spatial variation of coral species among sites (P < 0.01). However, none of the environmental variables measured (sedimentation, temperature, chlorophyll concentration, salinity) affected symbiont distribution. The discord between environmental variables and symbiont distributions suggests that the symbionts are physiologically tolerant of the variable environmental regime across this location and that the distribution of different host–symbiont combinations present is largely dependent on coral rather than Symbiodinium physiology. The data highlight the importance of host tolerance to environmental stressors, which should be considered simultaneously with symbiont sensitivity when considering the impact of variations in environmental conditions on coral communities.     

Past and present potential distribution of the Iberian Abies species: a phytogeographic approach using fossil pollen data and species distribution models

Aim Quaternary palaeopalynological records collected throughout the Iberian Peninsula and species distribution models (SDMs) were integrated to gain a better understanding of the historical biogeography of the Iberian Abies species (i.e. Abies pinsapo and Abies alba). We hypothesize that SDMs and Abies palaeorecords are closely correlated, assuming a certain stasis in climatic and topographic ecological niche dimensions. In addition, the modelling results were used to assign the fossil records to A. alba or A. pinsapo, to identify environmental variables affecting their distribution, and to evaluate the ecological segregation between the two taxa. Location The Iberian Peninsula. Methods For the estimation of past Abies distributions, a hindcasting process was used. Abies pinsapo and A. alba were modelled individually, first calibrating the model for their current distributions in relation to the present climate, and then projecting it into the past—the last glacial maximum (LGM) and the Middle Holocene periods—in relation to palaeoclimate simulations. The resulting models were compared with Iberian‐wide fossil pollen records to detect areas of overlap. Results The overlap observed between past Abies refugia—inferred from fossil pollen records—and the SDMs helped to construct the Quaternary distribution of the Iberian Abies species. SDMs yielded two well‐differentiated potential distributions: A. pinsapo throughout the Baetic mountain Range and A. alba along the Pyrenees and Cantabrian Range. These results propose that the two taxa remained isolated throughout the Quaternary, indicating a significant geographical and ecological segregation. In addition, no significant differences were detected comparing the three projections (present‐day, Mid‐Holocene and LGM), suggesting a relative climate stasis in the refuge areas during the Quaternary. Main conclusions Our results confirm that SDM projections can provide a useful complement to palaeoecological studies, offering a less subjective and spatially explicit hypothesis concerning past geographic patterns of Iberian Abies species. The integration of ecological‐niche characteristics from known occurrences of Abies species in conjunction with palaeoecological studies could constitute a suitable tool to define appropriate areas in which to focus proactive conservation strategies.     

Influences of ecology and biogeography on shaping the distributions of cryptic species: three bat tales in Iberia

To determine what shapes the distributions of cryptic species, we aimed to unravel ecological niches and geographical distributions of three cryptic bat species complexes in Iberia, Plecotus auritus/begognae, Myotis mystacinus/alcathoe and Eptesicus serotinus/isabellinus (with 44, 69, 66, 27, 121 and 216 records, respectively), considering ecological interactions and biogeographical patterns. Species distribution models (SDMs) were built using a presence‐only technique (Maxent), incorporating genetically identified species records with environmental variables (climate, habitat, topography). The most relevant variables for each species’ distribution and respective response curves were then determined. SDMs for each species were overlapped to assess the contact zones within each complex. Niche analyses were performed using niche metrics and spatial principal component analyses to study niche overlap and breadth. The Plecotus complex showed a parapatric distribution, although having similar biogeographical affinities (Eurosiberian), possibly explained by competitive exclusion. The Myotis complex also showed Eurosiberian affinities, with high overlap between niches and distribution, suggesting resource partitioning between species. Finally, E. serotinus was associated with Eurosiberian areas, while E. isabellinus occurred in Mediterranean areas, suggesting possible competition in their restricted contact zone. This study highlights the relevance of considering potential ecological interactions between similarly ecological species when assessing species distributions. © 2014 The Linnean Society of London, Biological Journal of the Linnean Society, 2014, 112 ,150–162.     

Aridity could have driven the local extinction of a common and multivoltine butterfly

1. Identifying which species are being negatively impacted by climate change and the mechanisms driving their decline is essential to effectively protect biodiversity.
2. Coenonympha pamphilus is a common and generalist butterfly, widely distributed throughout the Western Palearctic, being multivoltine in southern Europe. Previous studies indicate that it will not be substantially affected by climate change; however, it has seemingly disappeared from the southeast of the Iberian Peninsula in the last decades.
3. Here, we aim to determine if it has effectively disappeared from this area, as well as identify the environmental conditions limiting its distribution and the potential causes behind this a priori local extinction.
4. We downloaded all the occurrence records of C. pamphilus and analysed their spatial and temporal trends. To identify the climatic variables driving the distribution of this butterfly in the Iberian Peninsula, we performed an ensemble species distribution model (SDM), combining 600 individual models produced with 6 algorithms.
5. We confirmed that C. pamphilus has not been observed in the southeast of the Iberian Peninsula since 2008. Aridity was the main factor limiting the distribution of C. pamphilus in our ensemble SDM, with areas with high aridity being unsuitable for this species.
6. We hypothesise that multivoltinism is the mechanism driving this local extirpation, as high aridity is causing host plants (Poaceae) to wither prematurely, precluding the development of the second and/or third generations of the butterfly. Even though generalist species are theoretically more resilient to climate change, other traits such as multivoltinism may increase their vulnerability and need to be further investigated.

     

Biogeographical analysis of species of the tribe Cytiseae (Fabaceae) in the Iberian Peninsula and Balearic Islands

Aim To analyse quantitatively the biogeographical distribution pattern of species of the Cytiseae Bercht. & J. Presl (= Genisteae Benth.) tribe in the Iberian Peninsula and the Balearic Islands, and to identify environmental variables related to the distributional patterns. Location Iberian Peninsula and Balearic Islands, using the 61 administrative provinces as operational geographical units. Methods In order to identify chorotypes (groups of species with similar geographical distribution), we performed a upgma classification based on the similarity index of Baroni‐Urbani & Buser. The method of McCoy et al. [Ecology 67 (1986), 749] enabled us to detect the significant groups and to differentiate them from those groupings that could be generated at random. Logistic regression analyses and environmental gradient analyses (DCA and CCA) were performed in order to find the relationships between the environmental variables and the observed distributional patterns. Results Sixteen chorotypes were obtained between the Cytiseae species of the Iberian Peninsula and the Balearic Islands. The thermal variables showed the greatest influence in species distribution. Specifically, temperatures (maximum, minimum and mean) of the coldest months were associated with the majority of the chorotypes. Main conclusions The species of the Cytiseae tribe were not randomly distributed in space, and can be classified in groups of species with common distributional patterns. The importance of cold tolerance in the distribution of these species, as well as their preference for acidic soils, was demonstrated. Certain general tendencies seem to exist with respect to the distribution of the biota in the Iberian Peninsula, and these seem to be independent of the taxonomic groups involved and to be determined by convergent macroclimatic factors.     

Invasive aphids of the tribe Siphini: a model of potentially suitable ecological niches

相似文献   

Eutrophication,agriculture and water level control shift aquatic plant communities from floating-leaved to submerged macrophytes in Lake Chini,Malaysia

In this study we: (1) present a quantitative spatial analysis of the macrophyte communities in Lake Chini with a focus on the biogeographical distributions of the native Nelumbo nucifera and the invasive Cabomba furcata; (2) examine the environmental changes that affect plant community composition; and (3) outline a conceptual model of the variation of ecological processes that shape the macrophyte communities. Plant species cover, biomass of C. furcata and N. nucifera, and water quality and environmental variables were measured before and after monsoonal floods in September 2009 and April 2010. Permutational multivariate analysis was used to examine the significance of the invasion of C. furcata at different spatial scales. Relationships between plant species cover and environmental variables before and after flooding were examined using principal coordinates analysis and non-parametric multivariate multiple regressions. Our findings suggest that (1) Variation in plant communities was significant at the lake scale and the distribution of plant species changed after annual floods. (2) Invasion by C. furcata significantly affected the overall plant community composition. (3) C. furcata biomass increased after the monsoonal season, which indicates that C. furcata is adapted to flooding events and that it is becoming increasingly abundant. (4) In addition to the strong monsoonal effect, total depth, nutrient concentration, and sediment type were important environmental variables that significantly affected plant community composition. The macrophyte community in Lake Chini is highly dynamic. The spatial and temporal plant community dynamics are associated with flood regime, water quality, and substrate. Human-induced changes in these parameters are likely shifting the macrophyte dominance from floating-leaved to submerged species.     

Is current climatic equilibrium a guarantee for the transferability of distribution model predictions? A case study of the spotted hyena

Aim The transferability of species distribution models requires that species show climatic equilibrium throughout their entire distribution area. We test this assumption for the case of the spotted hyena, Crocuta crocuta, a large carnivore that has shifted its distribution over the last 100,000 years from a widespread Eurasian and African range to its current geographical distribution, restricted to the Sub‐Saharan areas of the African continent. Location Western Eurasia and Africa. Methods The current realized distribution of C. crocuta was estimated using presences and reliable absences as well as climatic, land‐cover and anthropic variables as predictors. The potential distribution was estimated using presences and a set of pseudo‐absences selected from localities outside climatically suitable localities, with only climatic variables serving as predictors. The current potential distribution was transferred to the Last Interglacial period (126,000 yr bp ) using the palaeoclimatic data yielded by the GENESIS 2 general circulation model, and validated with European fossil data. Generalized linear models were used on all occasions. Results Climatic variables are able to predict the current distribution of the species with high accuracy. The geographical projection of this model indicates that the species is distributed over almost all of its potential suitable area, which allows us to suppose that the current distribution of this species is in climatic equilibrium. However, the time transference of model predictions for the western Eurasian region reveals almost no suitable conditions for hyenas, despite the widespread presence of C. crocuta fossil remains on this continent during the Last Interglacial period. Main conclusions Our results indicate that, even when model results suggest a climatic equilibrium for a species distribution, the time transferability of such models does not necessarily provide realistic results. This occurs because the current geographical range does not allow estimations of all of the environmental requirements of a species. Therefore, any model trained with current data risks underestimating the potential suitable environmental and geographical range for species in a new area or time period.     

Natural,human and spatial constraints to expanding populations of otters in the Iberian Peninsula

Aim To determine the relationships between otter (Lutra lutra) distribution dynamics and environmental and spatial constraints over a 20‐year period. Location Andalusia, southern Iberian Peninsula. Methods We synthesized otter distribution data from three otter surveys (1985, 1995 and 2005) using subcatchment areas defined by hydrological barriers. Subcatchments were characterized by two ‘natural’ (climatic and orographic variables) and two ‘human’ (land use and population density) gradients. In addition, we calculated two contagion variables (the distance to previously occupied subcatchments and the percentage of occupied subcatchments within a 50 km buffer) for consecutively surveyed subcatchments. Results Between 1985 and 2005 the percentage of subcatchments with otters present increased from 42% to 72%. Otters tended to be rare or absent from human‐dominated areas. Anthropogenic gradients were better predictors of otter distribution than natural ones. Human and natural gradients showed strong covariation, but for any value of the natural gradients otters tended to be present in subcatchments with lower human impacts. Colonization of new subcatchments was found to be strongly related to contagion variables and expansion rates were slower than those estimated in other studies. Newly colonized areas tended to be located in areas with intermediate human influence, while repeated absences occurred mainly in areas where human impact was most severe. Main conclusions Our results suggest that recent otter expansion across Andalusia is a reflection of large‐scale improvement in environmental conditions. Otter populations that survived the period of strong and generalized declines appear to be acting as sources from which neighbouring areas are colonized, probably aided by improved water quality and increases in food availability. However, the further expansion of otters into their full original range is likely to be constrained by human‐impacted landscapes.     

Combined effects of local environment and continental biogeography on the distribution of Ostracoda

Summary 1. The relationship between altitudinal gradients on small spatial scales and latitudinal gradients on broader scales has been repeatedly recognised in the biogeography of animals and plants. However, little is known about this topic in the ecology and biogeography of ostracod communities in Mediterranean flowing waters or the factors underlying these spatial patterns. 2. We analysed the ostracod assemblages of near‐natural headwater streams in the Betic and Pre‐Betic Mountains in the southern Iberian Peninsula to decipher the most important environmental gradients structuring ostracod communities on a local scale. In addition, the European altitudinal and latitudinal distributions of the most commonly found species were analysed with GIS and regression models to compare geographical effects from local to continental scales. 3. Forty sampling sites, distributed among six catchments and ranging in altitude between 150 and 1940 m a.s.l., were sampled seasonally. Limnological and geographical information was also recorded for each sample. Seventeen ostracod species were found, two of which were new findings for the Iberian Peninsula: Potamocypris fulva and Cypria reptans. The most common species were Potamocypris zschokkei, Candona neglecta, Herpetocypris brevicaudata, Cyclocypris ovum, Potamocypris villosa and Pseudocandona albicans. The distribution of these species in 918 European locations was analysed to test the hypothesised change in altitudinal distribution with varying latitude. 4. The best subset of logistic and linear regression models, selected by means of the information‐theoretic approach, found that oxygen content and the variables related with substratum and discharge were the most important variables with a negative influence on ostracod presence, abundance and species richness on a local scale. These findings suggest that the negative effect on benthic invertebrates of physical disturbances relates to high flow velocity and turbulences. 5. Multivariate ordination methods show how altitude and water chemistry are the most important variables to explain the distribution of ostracod assemblages on the small spatial scale. On a larger scale, differences in latitudinal distribution throughout Europe were significant for the six most common species found in Granada. In addition, four of these showed significant negative linear relationships between latitude and altitude in Europe, supporting the important effect of climate on local and continental scale distributions. While ostracod biogeographies are still poorly known, our results indicate the influence of Quaternary climate variability on ostracod dynamic colonisation and extinction in Europe in accordance with species‐specific temperature and water chemistry preferences.     

Accounting for multi‐scale spatial autocorrelation improves performance of invasive species distribution modelling (iSDM)

Aim Analyses of species distributions are complicated by various origins of spatial autocorrelation (SAC) in biogeographical data. SAC may be particularly important for invasive species distribution models (iSDMs) because biological invasions are strongly influenced by dispersal and colonization processes that typically create highly structured distribution patterns. We examined the efficacy of using a multi‐scale framework to account for different origins of SAC, and compared non‐spatial models with models that accounted for SAC at multiple levels. Location We modelled the spatial distribution of an invasive forest pathogen, Phytophthora ramorum, in western USA. Methods We applied one conventional statistical method (generalized linear model, GLM) and one nonparametric technique (maximum entropy, Maxent) to a large dataset on P. ramorum occurrence (n = 3787) to develop four types of model that included environmental variables and that either ignored spatial context or incorporated it at a broad scale using trend surface analysis, a local scale using autocovariates, or multiple scales using spatial eigenvector mapping. We evaluated model accuracies and amounts of explained spatial structure, and examined the changes in predictive power of the environmental and spatial variables. Results Accounting for different scales of SAC significantly enhanced the predictive capability of iSDMs. Dramatic improvements were observed when fine‐scale SAC was included, suggesting that local range‐confining processes are important in P. ramorum spread. The importance of environmental variables was relatively consistent across all models, but the explanatory power decreased in spatial models for factors with strong spatial structure. While accounting for SAC reduced the amount of residual autocorrelation for GLM but not for Maxent, it still improved the performance of both approaches, supporting our hypothesis that dispersal and colonization processes are important factors to consider in distribution models of biological invasions. Main conclusions Spatial autocorrelation has become a paradigm in biogeography and ecological modelling. In addition to avoiding the violation of statistical assumptions, accounting for spatial patterns at multiple scales can enhance our understanding of dynamic processes that explain ecological mechanisms of invasion and improve the predictive performance of static iSDMs.     

Inferring evolutionary scenarios with geostatistics and geographical information systems for the viperid snakes Vipera latastei and Vipera monticola

The present study aimed to infer evolutionary scenarios for Vipera latastei and Vipera monticola in the Iberian Peninsula and the Maghreb through the identification of spatial patterns in morphological character variation and biogeographic patterns in morphological variability distribution. Ten morphological traits from 630 vipers were analysed with geostatistic and ecological niche modelling in a geographical information system. Interpolation by Kriging was used to generate surfaces of morphological variation, which were combined with spatial principal components analysis (SPCA). Putative morphological differentiated groups generated by SPCA maps were tested with discriminant function analysis (DFA). Maximum entropy modelling and nine environmental variables were used to identify factors limiting the distribution of groups and areas for their potential occurrence. Groups supported by DFA were: Western Iberia, Eastern Iberia, Rif plus Middle Atlas, Algeria, and High Atlas. Their distribution is influenced by common environmental factors such as precipitation. Areas of probable sympatry between Iberian groups matched the morphological clines observed by geostatistics tools. Geographic variation patterns in V. latastei‐monticola are probably due to vicariant separation of Iberian and African populations during the opening of the Strait of Gibraltar, and population refugia during the Quaternary glaciations with secondary contact. The taxonomic status of northern Morocco and Algerian groups should be further investigated. We conclude that geostatistics and niche‐modelling tools are adequate to infer morphological variability across wide geographic ranges of species. © 2008 The Linnean Society of London, Biological Journal of the Linnean Society, 2008, 95 , 790–806.     

Ocean‐scale prediction of whale shark distribution

Aim Predicting distribution patterns of whale sharks (Rhincodon typus, Smith 1828) in the open ocean remains elusive owing to few pelagic records. We developed multivariate distribution models of seasonally variant whale shark distributions derived from tuna purse‐seine fishery data. We tested the hypotheses that whale sharks use a narrow temperature range, are more abundant in productive waters and select sites closer to continents than the open ocean. Location Indian Ocean. Methods We compared a 17‐year time series of observations of whale sharks associated with tuna purse‐seine sets with chlorophyll a concentration and sea surface temperature data extracted from satellite images. Different sets of pseudo‐absences based on random distributions, distance to shark locations and tuna catch were generated to account for spatiotemporal variation in sampling effort and probability of detection. We applied generalized linear, spatial mixed‐effects and Maximum Entropy models to predict seasonal variation in habitat suitability and produced maps of distribution. Results The saturated generalized linear models including bathymetric slope, depth, distance to shore, the quadratic of mean sea surface temperature, sea surface temperature variance and chlorophyll a had the highest relative statistical support, with the highest percent deviance explained when using random pseudo‐absences with fixed effect‐only models and the tuna pseudo‐absences with mixed‐effects models (e.g. 58% and 26% in autumn, respectively). Maximum Entropy results suggested that whale sharks responded mainly to variation in depth, chlorophyll a and temperature in all seasons. Bathymetric slope had only a minor influence on the presence. Main conclusions Whale shark habitat suitability in the Indian Ocean is mainly correlated with spatial variation in sea surface temperature. The relative influence of this predictor provides a basis for predicting habitat suitability in the open ocean, possibly giving insights into the migratory behaviour of the world’s largest fish. Our results also provide a baseline for temperature‐dependent predictions of distributional changes in the future.     

Environmental and geographical constraints on common swift and barn swallow spring arrival patterns throughout the Iberian Peninsula

Aim Still poorly understood, the main migratory pathways for most trans‐Saharan species pass through the Iberian Peninsula, which acts as a gateway to the European–African migratory system. Arrival patterns in this region for the common swift (Apus apus) and barn swallow (Hirundo rustica), of similar morphology and flight capabilities, were described, and the environmental and geographical factors best explaining them were examined, in a search for common ecological constraints on these two migratory species. Location Latitude ranged from 36.02 to 43.68°N, longitude from 9.05°W to 3.17°E, and altitude from 0 to 1595 m a.s.l. for 482 common swift and 812 barn swallow Spanish localities spread widely over the Iberian breeding grounds of the two species. Methods Our data set, covering the years 1960–1990, consisted of 3206 first‐arrival dates for common swifts and 6036 for barn swallows. Forty topographical, climatic, river basin, geographical and spatial variables were used as explanatory variables in general regression models (GRMs). GRMs included polynomial terms up to cubic functions in all variables when they were significant. A backward stepwise selection procedure was applied in all models until only significant terms remained. GRMs were applied in two steps. First, we searched for the best model in each one of the five types of variables (topographical, climatic, river basin, geographical and spatial). To cope with the unavoidable correlation between explanatory variables, the relative importance of each type of variable was assessed by hierarchical variance partitioning. Secondly, we searched for that model able to explain the maximum amount of the observed variability in arrival date. To obtain this model all significant explanatory variables were subjected jointly to a GRM. Spatial variables were then added to this model to take any remaining spatial structure in the data into account. Moran's I autocorrelation coefficient was used to check for spatial autocorrelation. Results Both species arrived earlier in the south‐western Iberian Peninsula, where summers are warmer and drier. From there, both species followed the main southern Iberian river basins towards the north‐east; however, several mountainous regions impede the colonization of eastern Iberia. The best models for each type of variable explained 19–47% of the variability in common swift arrival dates and 14–44% in barn swallow arrival dates. Variance partitioning indicated that climatic and geographical variables best explained variability. The best predictive models built with all variables accounted for 52% of the variability in common swift arrival dates and 50% for the barn swallow. Residuals from both models were not spatially autocorrelated, an indication that all major spatially structured variation had been accounted for. Main conclusions Spring arrival patterns are highly dependent on the geographical configuration of the Iberian Peninsula. This spatial constraint forces both species to converge very closely in their spring migration, because common swifts and barn swallows are subject to a trade‐off between optimum migratory pathways and territories ecologically suitable for breeding.