Niche overlap pdf

The mean annual rainfall is mm, ranging from mm Three main factors affect the temporal patterns of ants foraging to mm. Each transect was sampled for 11 consecutive et al. To evaluate the overlap of trophic and temporal niches of transects was not performed. Coincidence in diet or et al.

Vouchers were deposited in the Prof. Simulations for dietary overlap were conducted in EcoSim 7. Data analysis tions for overlap in temporal activity were conducted with the TimeOverlap program Castro-Arellano et al. Results diversity metric Solow, For each analysis, individuals from both seasons were combined into a single pool of individuals. Camponotus blandus Smith, Britton and Rose were the plants most visited by ants.

Ants exhibited To Table 2. The two most abundant ant species C. This null model has a desir- the dry season Table 2 or during the wet season Table 2. Average able combination of type I and type II error rates for evaluation of overlap in trophic niches was greater during the dry season 0. For temporal activity, null distributions of overlap than during the green season 0.

Of the four species that were values were generated using a randomization algorithm Rosario recorded at least 5 times each season, only C. Flowers of S.

Flowers of I. Ant codes as in Table 1. None- Shannon diversity 1. In the dry forest of Minas Gerais, Brazil, which is south of our study site and experi- palmadora and S. Nonetheless, ant species composition in mature forests exhibited distinct differences between seasons. Ants in the Caatinga exhibi- 3. Temporal overlap ted similar patterns to those in the dry forests of Minas Gerais with respect to both richness and composition Table 1.

A season Table 2 , and varied from 0. Average pair-wise overlap in temporal activity was greater during 4. Trophic niche overlap the green season 0. Of the species that were recorded at least 5 times each season, only C. Floral resources are more abundant and diverse in the Caatinga blandus exhibited seasonal differences in temporal activity during the green season than during the dry season Santos et al.

Time interval Fig. This indicates that ant density on plants is relatively season. Nonetheless, seasonal changes resources.

The most commonly captured species C. Resende and E. Mota, who helped during sampling. We thank Dr. Robert R. Activity patterns and temporal overlap an early version of the manuscript, we are especially grateful for their invaluable comments. Presley was Junker et al.

Supplementary material temporal overlap during the hot and dry season than during the cooler, green season. Our results are in direct opposition to Supplementary material related to this article can be found expectations based on a physiological constraints mechanism.

Revista Nordestina de Biologia 10, e Neotropical Entomology 36, e American Journal of Botany 92, e A year study of NDVI variability over the against nectar theft see above discussion on trophic niche overlap , northeast region of Brazil. Journal of Arid Environments 67, e Beattie, A.

Ant inhibition of pollen temporal patterns of nectar production may be a more likely function: a possible reason why ant pollination is rare.

American Journal of explanation for high temporal niche overlap. The trophic niche axis Botany 71, e Ecology 86, e Diversity of planktonic Foraminifera in deep-sea characteristics of particular times e. Science , e Competition for composition: lessons from nectar- changing behavior in response to environmental cues. The feeding ant communities. Ecology 85, e Functional complementarity and specialisation: why biodiversity is important in plant-pollinator interactions.

Basic Applied Ecology changed greatly between the seasons Appendix A. The ecological dynamics of natural selection among resources and consumers caused by both apparent and resource competition. Competitive exclusion and evolution: convergence almost never produces ecologically equivalent species. Mayfield, M. Opposing effects of competitive exclusion on the phylogenetic structure of communities. Connell, J. Diversity and the coevolution of competitors, or the ghost of competition past.

Oikos 35 , — Vukics, A. Speciation in multidimensional evolutionary space. E 68 , Dieckmann, U. On the origin of species by sympatric speciation. Competitive exclusion and limiting similarity: a unified theory. Godoy, O. Phylogenetic relatedness and the determinants of competitive outcomes. Kraft, N.

Plant functional traits and the multidimensional nature of species coexistence. Bono, L. Competition and the origins of novelty: experimental evolution of niche-width expansion in a virus. Rapid evolution of a native species following invasion by a congener. Science , — Zuppinger-Dingley, D.

Selection for niche differentiation in plant communities increases biodiversity effects. Ellis, C. Character displacement and the evolution of niche complementarity in a model biofilm community. Evolution 69 , — Peterson, M. Niche partitioning between close relatives suggests trade-offs between adaptation to local environments and competition.

Agnarsson, I. Limited role of character displacement in the coexistence of congeneric Anelosimus spiders in a Madagascan montane forest. Ecography 39 , — Morphological evolution of coexisting amphipod species pairs from sulfidic caves suggests competitive interactions and character displacement, but no environmental filtering and convergence. Miller, T. Rethinking niche evolution: experiments with natural communities of Protozoa in pitcher plants.

Tilman, D. The evolution of predator—prey interactions: theory and evidence. Hairston, N. Rapid evolution and the convergence of ecological and evolutionary time. Falster, D. Multitrait successional forest dynamics enable diverse competitive coexistence. USA , E—E Lande, R. The measurement of selection on correlated characters. Evolution 37 , — Bulmer, M.

Barton, N. The infinitesimal model: definition, derivation, and implications. Turelli, M. Species packing and competitive equilibria for many species. Falconer, D. Introduction to Quantitative Genetics Longman, Download references. We thank S. Allesina, R. Bertram, B. Inouye, S. Steppan and A. Winn for providing insightful comments on this work. This work was made possible in part by funding awarded to M. DP by the Australian Research Council. Abigail I. Pastore, Malyon D.

You can also search for this author in PubMed Google Scholar. Correspondence to Abigail I. Sakarchi and the other, anonymous, reviewer s for their contribution to the peer review of this work. Peer reviewer reports are available. Reprints and Permissions. Pastore, A. The evolution of niche overlap and competitive differences. Note that in some instances it may be difficult to define the extent of the study areas to be compared. When species The comparison of zij between two entities can be used to occur on different continents, the choice can be straightforward calculate niche overlap using the D metric Schoener, ; and should consider the complete gradient of environmental reviewed in Warren et al.

When species occur in the same region or ij on an island, the environment can be the same for all species and therefore correcting for differences in the densities of environ- where z1ij is entity 1 occupancy and z2ij is entity 2 occupancy.

This metric varies between 0 no overlap and 1 complete overlap. Note that regions of the environmental space that do not exist in geography have zij set to 0. These regions thus do Testing the framework with virtual entities not contribute to the measure of the D metric and niche A robust test of the framework described above requires entities overlap is measured among real habitats only see discussion that have distributions determined by known environmental in Warren et al.

Note also that the use of a kernel parameters and that exhibit known levels of niche overlap. To density function when calculating the density is critical for an achieve this, we simulated pairs of virtual entities with varying unbiased estimate of D. When no kernel density function is amounts of niche overlap Fig. The annual maximum temperature tmax , and annual minimum niche equivalency test determines whether niches of two entities temperature tmin.

Procedures to calculate aetpet, pet and gdd in two geographical ranges are equivalent i. In addition to a general description of the technique, an explanation of its application to the comparison of simulated niches between the European EU and North American NA continents is provided.

Depending on the type of analysis and whether a priori groups are used or not, the different areas of calibration we tested are specified. Name Description Areas of calibration PCA-occ Principal component analysis Pearson, transforms a number of correlated variables 1.

These components are the best predictors — in terms of R2 — of the original variables. In other terms, the first principal component accounts for as much of the variability in the data as possible, and each following component accounts for as much of the remaining variability as possible. For the study of niche overlap, the data used to calibrate the PCA are the climate values associated with the occurrences of the species.

Additional occurrence data can be projected in the same environmental space. When calibrating the PCA with EU and NA occurrences, differences in position along the principal components discriminate environmental differences between the two distributions.

When calibrating with EU occurrences only, differences in position along the principal components maximize the discrimination of differences among the EU distribution PCA-env Same as PCA-occ but calibrated on the entire environmental space of the two study areas, 1.

EU range including species occurrences. Here the a priori groups correspond to EU and NA. Here we use the distance in the Euclidean space. The degree of correspondence between the distances among points implied by MDS plot and the input distance structure is measured inversely by a stress function. ENFA is an ordination technique 1. This EU range method differs from other ordination techniques in that the principal components have a 2.

The next components correspond to specialization factors: axes that maximize the ratio of the variance of the global distribution to that of the species distribution CA, correspondence analysis; MCA, multiple correspondence analysis.

Since the normal density curves defin- compare the simulated level of niche overlap with the niche ing the niches of the virtual entities Appendix S2 are built overlap detected along axes calibrated using several ordination along these two gradients, we postulate that the overlap detected Table 1 and SDM techniques Table 2. For methods with by the application of the framework should be the same as the maximization criteria that do not depend on an a priori group- simulated level of niche overlap across the full range of possible ing here EU versus NA, Table 1 , we run two sets of simulations, overlaps 0 to 1.

Table 2 Species distribution modelling SDM techniques for quantifying niche overlap. MaxEnt was fitted using the dismo package in R with default settings.

For all techniques, we use pseudo-absences that were generated randomly throughout the area of calibration. Two sets of models were created using two areas of calibration: one using presence—absence data in Europe EU only and a second using presence-absence data in both EU and North America NA. The resulting predictions of occurrence of the species ranging between 0 and 1 are used as environmental axes in the niche overlap framework.

Here we use binomial presence—absence response variables with a logit link function logistic regression and allow linear and quadratic relationship between the response and explanatory variables. A stepwise procedure in both directions was used for predictor selection, based on the Akaike information criterion AIC; Akaike, MaxEnt MaxEnt Phillips et al. MaxEnt estimates species distributions by finding the distribution of maximum entropy i.

MaxEnt begins with a uniform distribution then uses an iterative approach to increase the probability value over locations with conditions similar to samples. The probability increases iteration by iteration, until the change from one iteration to the next falls below the convergence threshold. MaxEnt uses L — 1 regularization as an alternative to stepwise model selection to find parsimonious models GBM The gradient boosting machine GBM; Friedman, is an iterative computer learning algorithm.

In GBM, model fitting occurs not in parameter space but instead in function space. The GBM iteratively fits shallow regression trees, updating a base function with additional regression tree models. A randomly chosen part of the training data is used for function fitting, leaving the other part for estimating the optimal number of trees to use during prediction with the model out-of-bag estimate RF Random forests RF; Breiman, Random forests grows many classification trees.

To classify the species observations i. The forest chooses the classification having the most votes over all the trees in the forest. To R E S U LT S compare the outcomes of the methods quantitatively, for each analysis we first calculate the average absolute difference Evaluation of the framework between the simulated and measured overlap Dabs. To test for biases in the method levels of niche overlap along known gradients.

We used pairs i. A t climate gradients. The overlap we detected between each pair method that reliably measures simulated levels of niche overlap of virtual entities is almost identical to the simulated overlap should show both small errors small Dabs and low bias non- i. Because native and invaded ranges on different continents and which detected overlap cannot be larger than 1, any error in the mea- have been subjects of recent analyses of niche dynamics.

The surement of highly overlapping distribution must necessarily first case study concerns spotted knapweed Centaurea stoebe, result in underestimation. America see Broennimann et al. The second case study addresses follow a univariate normal distribution along a precipitation the fire ant Solenopsis invicta , native to South America and gradient, no underestimation was observed Fig.

When we invasive in the USA see Fitzpatrick et al. The red to blue colour scale shows the projection of the normal densities in the geographical space from low to high probabilities i.

Black dots show random occurrences. This bias is on Scores of PCA-occ and MDS are significantly biased, with the average five times larger than that of the corrected measure. Niche overlap detected by ordination and Among methods with maximization criteria based on a priori SDM methods grouping Fig.

Predictably, Fig. Note, however, that highly overlap- levels of niche overlap adequately. Both methods provide similar ping distributions are somewhat underestimated but the signifi- results in which overlap is underestimated across all simulated cance of the Wilcoxon test is unaffected. The only other levels. When calibrated on data from both ranges. For both species, niche 0. Our method is appro- Simulated overlap priate for the study of between-species differences of niches e.

Each dot corresponds to a pair of simulated entities. Broennimann means on precipitation and temperature gradients see Figure 1. Alternatively, when a record of the distribution of the climate availability density of occurrences divided by the density taxa and corresponding environment through time exists, our of climate across the entire climate space. Open circles show the approach can be used to answer the question of whether and to detected overlap when no correction for climate availability is what degree environmental niches have changed through time applied.

The average absolute difference between the simulated e. Pearman et al. First, it disentangles the dependence of species occurrences from the frequency of different climatic conditions that occur across a region. This is accomplished by brate bimodal curves that tightly fit the two distributions as a dividing the number of times a species occurs in a given envi- whole. However, when calibrated on the EU range only, all SDM ronment by the frequency of locations in the region that have methods report increasing levels of overlap along the gradient of those environmental conditions, thereby correcting for differ- simulated overlap.

Without this 0. For example, Fig. Generalized linear modelling GLM exhibits a based method using comparisons of geographical predictions of similar amount of error as GBM, but with lower reported occurrences, projections depend on a given study area. Second, application of a kernel smoother to standardized species densities makes moving from geographical space, where the species occur, to the multivariate Case studies environmental space, where analyses are performed, indepen- Analyses of spotted knapweed and fire ant occurrences using dent of both sampling effort and of the resolution in environ- PCA-env, the most accurate method in terms of niche overlap mental space Fig.

This is a critical consideration, because it detection, show that for both species the niche in the native and is unlikely that species occurrences and environmental datasets invaded ranges overlap little 0. For spotted knapweed, the invaded niche exhibits both shift same spatial resolution. Without accounting for these differ- and expansion Fig. Interest- ences, measured niche overlap will partially be a function of data ingly, two regions of dense occurrence in NA indicate two resolution.

In contrast, Although niche overlap can be detected accurately when vari- the fire ant exhibits a shift from high density in warm and wet ables driving the distribution are known e. Crosses refer to models calibrated on the European EU range only. Abs D :m indicate the average absolute difference between simulated and detected overlaps. The causes of the differences in performance among techniques remain unclear, but several Ordinations versus SDMs factors might be responsible.

Among the important factors are: 1 how the environment varies in relation to species occur- Ordinations and SDMs use contrasting approaches to reduce rences versus the study region or time period as a whole, 2 the dimensions of an environmental dataset. Black dots indicate models calibrated on both EU and NA ranges.

Abs D :m indicates the average absolute difference between simulated and detected overlaps. When conserved. It is likely that SDMs fit bimodal response curves that When calibrating SDMs using only one study area and sub- tightly match the data and artificially predict occurrences in sequently projecting the model to another area, estimated both ranges i.

SDMs model the range of each entity as a single overlap increases with simulated overlap Fig. As a result, prediction values However, the pattern of detected overlap using SDMs is irregu- for occurrences are high for both ranges. Since the overlap is lar i. Dabs:m is high , again probably because of overfitting.

Bias measured on the gradient of predicted values, measured overlap in detected overlap may also arise from the differing spatial is inevitably high. In contrast, ordinations calibrated on both structure of environments between study areas.

In general, orthogonal axes, the variable selection procedure of SDMs is no SDM method exceeded the performance of the best ordina- sensitive to collinearity. A variable that is not important for the tion method. Projection of the model to another area or con- variables according to their importance in delimiting the tinent in the present case could then be inconsistent with the niche. SDMs thus could be used to identify variables that are actual requirements of the species and lead to spurious patterns closely related to the processes driving the distribution of the of detected overlap.