Endemicity analysis, parsimony and biotic elements: a formal comparison using hypothetical distributions

Rethinking spatial history: envisioning a mechanistic historical biogeography

Historical biogeography is the study of geographic distributions of taxa through space and time. Over the last 50 years, several methods have been proposed to reconstruct these histories. However, despite their particularities, conceptually they have been most often derived from the reconstruction of area relationships. Here we advocate that area cladograms lack explanatory power and that biogeography needs to move towards a more mechanistic approach. We discuss the ontological problems related to areas of endemism and their validity as biogeographic units. Specifically, we propose that areas of endemism are not discrete historical entities and that area-based analyses are inappropriate for analytical biogeography. Instead, we suggest that biogeographic analyses should focus on those spatial-geographic elements that cause diversification, namely barriers. We discuss how barriers have more discrete boundaries in space and time than do areas of endemism, which allows the identification of homologous units and the recovery of vicariant events. Reconstructing the history of vicariant events results in a better understanding of spatial evolution within a biota because barrier formation is the relevant causal mechanism of diversification. We end by acknowledging the largely ignored views of Peter Hovenkamp and his conceptual contributions to developing a mechanistic biogeography.

Areas of endemism of Pteridaceae (Polypodiopsida) in Brazil: a first approach

Areas of endemism (AoE) comprise regions host to two or more endemic taxa, whose distributional limits are congruent and not random. These areas are important for two reasons: they comprise the smallest geographic units for biogeographic analyses and they are priority targets for conservation actions. Ferns are a monophyletic group that despite having a wide geographic distribution, concentrates great species richness and endemism in some regions (centres). The southern and southeastern regions of Brazil comprise one of these centres for the Neotropics. This study aims to verify the AoE of Pteridaceae in Brazil and examine whether the results obtained here are congruent with areas already delimited for other groups and whether there is spatial correspondence between the AoE and Conservation Units. To this end, a database was created with collection records of the 205 Pteridaceae species occurring in Brazil based on a review of herbaria. We analysed 23 815 records for 205 Pteridaceae species using Endemicity Analysis (NDM-VNDM), selecting the fill and assumed parameters, and 1°, 2° and 3° grid-cells. The consensus of 158 AoE, using different grid sizes, was calculated, and subsequently, generalized AoE were established. The Guiana Shield, southern Brazil, southeastern Brazil, and southeastern Bahia were considered generalized AoE. These areas correspond to those found for animals and angiosperms, and in previous studies with ferns. Furthermore, two areas, Acre and Mato Grosso do Sul, were recovered only on grids with 2° and 3°. It will be essential to conduct more research to confirm the persistence of both AoE (Acre and Mato Grosso do Sul), especially after expanding sampling. Most endemic species distribution points occur outside protected areas, demonstrating an alarming situation regarding the conservation of these taxa. In addition, fern distribution data could (and should) be used in conservation practices, programmes and policies, given that they are good ecological indicators and that the distribution of ferns may not reflect that of angiosperms and animals.

Distribution patterns of Triatominae (Hemiptera: Reduviidae) in the Americas: an analysis based on networks and endemicity

Triatominae, commonly known as kissing bugs, are a group of approximately 150 species of hematophagous reduviids, some of which are vectors of Trypanosoma cruzi, the etiological agent of the Chagas disease. Distributional patterns of triatomines have been studied based on macroecological and historical biogeographic approaches, but the definition of distributional patterns and areas of endemism are yet to be defined based on objective criteria. We used two methods to identify biogeographic units in the Triatominae: the endemicity analysis based on an optimality criterion (NDM/VNDM software) and a network approach aimed to simplify and highlight the underlying structure in species distributions (Infomap Bioregions). Information on species distributions was obtained from a data paper, comprising 21 815 records for 135 triatomine species occurring in the Americas. The resulting areas of each method were clustered using a meta consensus criterion based on dissimilarities and interpreted as recurrent areas. The NDM areas show a nested structure, presenting greater restrictions to the inclusion of species in a given area, requiring broad sympatry. In contrast, bioregions emphasize spatial patterns with better-delimited areas and species occurrences do not need to be highly congruent. When areas were clustered based on their species composition two clear patterns arose from both methods: (i) areas within the southern Amazon and southeast South America, especially in the Chacoan subregion, formed a cluster, and (ii) areas north of the Amazon, Pacific, Mesoamerica, Mexican transition zone and Nearctic formed another cluster. Moreover, within each of these two clusters, there was a latitudinal gradient of the areas in the Americas showing spatial similarity between the areas found in both methods. Results of both methods show well-bound areas separating the triatomine fauna in the Brazilian subregions, resulting in the recognition of areas corresponding to the biomes Chaco, Pampa, Cerrado, and Caatinga, and, to a lesser extent, the Atlantic Forest.

Assessing the Diversity and Distribution Pattern of the Speciose Genus Lycocerus (Coleoptera: Cantharidae) by the Global-Scale Data

Species richness patterns and endemism on the large-scale play a significant role in biogeography and biodiversity conservation. This study aimed to explore the diversity centers and endemic areas of a large cantharid genus Lycocerus, so as to test whether the hypothesis of montane and island systems biodiversity in previous studies was supported. In this study, a comprehensive species’ geographical database on the global scale consisting of 4,227 records for 324 Lycocerus species was compiled and analyzed. Species richness pattern was mapped into a grid-based map with a spatial resolution of 1° × 1° fishnet. An unbalanced pattern was identified, and it showed that the centers of species richness of Lycocerus were situated in Eastern Himalayas, Hengduan Mountains, Eastern Sichuan Mountains, Taiwan, and Japan. Further analyses based on two approaches, including parsimony analysis of endemicity (PAE) and endemicity analysis (EA), were applied to detect areas of endemism (AOEs) at three different grid sizes (1°, 1.5°, and 2°). Finally, a total of nine AOEs were detected, including five montane areas (Himalayan areas, Hengduan Mountains, South edge of China, Eastern China Mountains, and Eastern Sichuan Mountains), three islands (Taiwan Island, Japan, and Korean Peninsula), and one plateau (Shan Plateau), which were generally consistent with the species richness pattern. The results verify that montane and island systems have an essential role in promoting the formation of diversity centers and AOEs because of their complex topography, varied habitat and geological events.

Two centuries of distribution data: detection of areas of endemism for the Brazilian angiosperms

Brazil has high levels of biodiversity and has received strong criticism for the increasing country-wide deforestation that threatens it. Although a significant percentage of land area in Brazil is protected, the areas are insufficient and unevenly distributed. Many studies have contributed to the biogeographical knowledge of Brazilian flora, but no endemicity analysis (EA) has been conducted including all endemic angiosperms. We investigated the spatial component, drawing on a huge and taxonomically diverse dataset based on 827 016 records collected over the last two centuries. We conducted an EA for 15 034 species from 173 families using an optimality criterion with 2° and 3° grid sizes, in order to search for distributional concordance, to identify the biogeographical units and discuss the implications for conservation. Six analyses were run for basal angiosperms, monocots and eudicots. The EA recovered 66 consensus areas (CAs). The concordance of CAs enabled the identification of five best-supported areas of endemism--three in the Atlantic and Parana Forest and two areas in the Cerrado province--supported by species of 120 families. The age of divergence for some genera that contributed significantly to the identification of areas recovered in the Cerrado coincides with the recent, <10 Ma, estimated age of that province. By contrast, the areas in the Atlantic and Parana Forest are supported by genera with earlier diversification >30 Ma, supporting an ancient origin. Most areas in the Atlantic Forest are partially superimposed with the limits of the protected areas, however, big gaps were identified in the Cerrado. Protecting Brazilian biomes was at the heart of Brazil's environmental policy. Regrettably, this scenario has radically changed based on misleading divergences in conservation policy. Areas of endemism are pivotal for biodiversity conservation due to the common evolutionary history shared by their endemic taxa. Thus, we hope that these congruent patterns of endemism support the establishment of biodiversity priorities.

Endemism Patterns of Planthoppers (Fulgoroidea) in China

Studies on endemism are always of high interest in biogeography and contribute to better understanding of the evolution of species and making conservation plans. The present study aimed to investigate the endemism patterns of planthoppers in China by delimiting centers of endemism and areas of endemism. We collected 6,907 spatial distribution records for 860 endemic planthopper species from various resources. Centers of endemism were identified using weighted endemism values at 1° grid size. Parsimony analysis of endemicity and endemicity analysis were employed to detect areas of endemism at 1°, 1.5°, and 2° grid sizes. Six centers of endemism located in mountainous areas were identified: Taiwan Island, Hainan Island, eastern Yungui Plateau, Wuyi Mountains, western Qinling Mountains, and western Yunnan. We also delimited six areas of endemism, which were generally consistent with centers of endemism. Our findings demonstrated that mountainous areas have an essential role in facilitating the high level of endemism and formation of areas of endemism in planthoppers through the combined effects of complex topography, a long-term stable environment, and geological events. Dispersal ability and distribution of host plants also have important effects on the patterns of planthoppers’ endemism.

Spatial Congruence Analysis (SCAN): A method for detecting biogeographical patterns based on species range congruences

Species with congruent geographical distributions, potentially caused by common historical and ecological spatial processes, constitute biogeographical units called chorotypes. Nevertheless, the degree of spatial range congruence characterizing these groups of species is rarely used as an explicit parameter. Methods conceived for the identification of patterns of shared ranges often suffer from scale bias associated with the use of grids, or the incapacity to describe the full complexity of patterns, from core areas of high spatial congruence, to long gradients of range distributions expanding from these core areas. Here, we propose a simple analytical method, Spatial Congruence Analysis (SCAN), which identifies chorotypes by mapping direct and indirect spatial relationships among species. Assessments are made under a referential value of congruence as an explicit numerical parameter. A one-layered network connects species (vertices) using pairwise spatial congruence estimates (edges). This network is then analyzed for each species, separately, by an algorithm which searches for spatial relationships to the reference species. The method was applied to two datasets: a simulated gradient of ranges and real distributions of birds. The simulated dataset showed that SCAN can describe gradients of distribution with a high level of detail. The bird dataset showed that only a small portion of range overlaps is biogeographically meaningful, and that there is a large variation in types of patterns that can be found with real distributions. Species analyzed separately may converge on similar or identical groups, may be nested in larger chorotypes, or may even generate overlapped patterns with no species in common. Chorotypes can vary from simple ones, composed by few highly congruent species, to complex, with numerous alternative component species and spatial configurations, which offer insights about possible processes driving these patterns in distinct degrees of spatial congruence. Metrics such as congruence, depth, richness, and ratio between common and total areas can be used to describe chorotypes in detail, allowing comparisons between patterns across regions and taxa.

Statistical Modeling of Distribution Patterns: A Markov Random Field Implementation and Its Application on Areas of Endemism

A statistical framework to infer areas of endemism from geographic distributions is proposed. This novel method is based on hidden Markov random fields (HMRFs), a type of undirected graph model commonly used in computer vision. This framework assumes areas of endemism are the states of the hidden layer of the model, whereas taxon distributions are emitted values in the observed layer. Taxon distributions are associated to the observed layer through a clustering procedure based on the extent of overlap. Observations are emitted by the hidden layer according to a Gaussian distribution, whereas the joint distribution of the hidden layer follows a Potts model. State and parameter inference of the maximum a posteriori configuration is performed through a modified version of the expectation-maximization algorithm. The optimal number of areas of endemism in the data set is estimated through the pseudolikelihood information criterion, a model selection procedure that uses an approximation to likelihood. The performance of the new algorithm was assessed on simulated data, and compared with the most popular methods for delimitation of areas of endemism: biotic element analysis, parsimony analysis of endemism, and endemicity analysis. HMRFs efficiently recovered the true pattern across a wide range of uncertainty values. The performance was also examined on empirical data: South African weevils (Sciobius) and Central American ground beetles and funnel-web tarantulas (Carabidae and Dipluridae, respectively). HMRFs uncovered six areas of endemism from the weevil data set, whereas eight were estimated for the Central American arthropods (compared with 3-5 and 3-14 from the other methods, respectively).

Areas of endemism: to overlap or not to overlap, that is the question

The concept of "areas of endemism", and the assumption that these patterns are always a consequence of vicariant events, are reviewed. This assumption is related to the idea that areas of endemism have well-defined limits and never share any surface with other areas of endemism because they must represent sister areas supported by sister taxa. Based on this idea, overlapping areas have been considered rarely, or ignored completely. Using a data set of mammals of North America, we test here whether the overlapping areas are indeed sister areas supported by sister taxa, thus evaluating whether vicariant events are commonly the factor producing areas of endemism.

Hotspots within a global biodiversity hotspot - areas of endemism are associated with high mountain ranges

Conservation biology aims at identifying areas of rich biodiversity. Currently recognized global biodiversity hotspots are spatially too coarse for conservation management and identification of hotspots at a finer scale is needed. This might be achieved by identification of areas of endemism. Here, we identify areas of endemism in Iran, a major component of the Irano-Anatolian biodiversity hotspot, and address their ecological correlates. Using the extremely diverse sunflower family (Asteraceae) as our model system, five consensus areas of endemism were identified using the approach of endemicity analysis. Both endemic richness and degree of endemicity were positively related to topographic complexity and elevational range. The proportion of endemic taxa at a certain elevation (percent endemism) was not congruent with the proportion of total surface area at this elevation, but was higher in mountain ranges. While the distribution of endemic richness (i.e., number of endemic taxa) along an elevational gradient was hump-shaped peaking at mid-elevations, the percentage of endemism gradually increased with elevation. Patterns of endemic richness as well as areas of endemism identify mountain ranges as main centres of endemism, which is likely due to high environmental heterogeneity and strong geographic isolation among and within mountain ranges. The herein identified areas can form the basis for defining areas with conservation priority in this global biodiversity hotspot.

Islands conserve high species richness and areas of endemism of Hormaphidinae aphids

Patterns of biodiversity and endemism are important and form the foundation for biogeography and conservation studies. Hormaphidinae is an aphid group mainly distributed in the Sino-Japanese, Oriental, and Nearctic zoogeographic realms. To infer the areas of endemism of Hormaphidinae aphids in the Sino-Japanese and Oriental realms, we employed a geographical distribution dataset covering all 225 species in subfamily Hormaphidinae. In total, 1,245 distribution occurrence records for all species were analyzed in addition to the number of species in a certain grid cell to calculate species richness. Two methods (parsimony analysis of endemicity (PAE) and the use of an optimization criterion-NDM/VNDM) using a total of 6 grid sizes were applied to detect possible areas of endemism and to assess the probable effects of the 2 different methods and 6 grid sizes on the results of the patterns of the areas of endemism. The results revealed that islands presented most of the areas with high species richness and endemic species, particularly Japan, Taiwan Island, Java Island, the Malaysian Peninsula, southeast Himalaya, and the Hengduan Mountains. Most of these areas of endemism were located on islands or a peninsula. Islands were therefore shown to play an essential role in facilitating the formation of high species richness and endemism. Different grid sizes directly determined the regions of the areas of endemism, with small grid sizes tending to detect small and discrete areas of endemism and large grid sizes identifying continuous areas of endemism. In terms of the two methods, NDM/VNDM was found to identify more areas of endemism than PAE.

Areas of endemism in the Nearctic: a case study of 1339 species of Miridae (Insecta: Hemiptera) and their plant hosts

Areas of endemism are essential first hypotheses in investigating historical biogeography, but there is a surprising paucity of such hypotheses for the Nearctic region. Miridae, the plant bugs, are an excellent taxon to study in this context, because this group combines high species diversity, often small distribution ranges, a history of modern taxonomic revisions, and comprehensive electronic data capture and data cleaning that have resulted in an exceptionally error-free geospatial data set. Many Miridae are phytophagous and feed on only one or a small number of host plant species. The programs ndm/vndm are here used on plant bug and plant data sets to address two main objectives: (i) identify areas of endemism for plant bugs based on parameters used in a recent study that focused on Nearctic mammals; and (ii) discuss hypotheses on areas of endemism based on plant bug distributions in the context of areas identified by their host plant species. Given the narrow distribution ranges of many species of Miridae, the analytical results allow for tests of the prediction that areas of endemism for Miridae are smaller and more numerous, especially in the Western Nearctic, than are those of their host plants. Analyses of the default plant bug data set resulted in 45 areas of endemism, 35 of them north of Mexico and many located in the Western Nearctic; areas in the Nearctic are more numerous and smaller than those identified by mammals. The host plant data set resulted in ten areas of endemism, and even though the size range of areas is similar between the Miridae and plant data sets, the average area size is smaller in the Miridae data set. These results allow for the conclusion that the Miridae indeed present a valuable model system to investigate areas of endemism in the Nearctic.