Identifying complementary areas for conservation in Thailand: an example using owls, hawkmoths and tiger beetles

Setting priorities in biodiversity conservation requires that explicit, efficient and appropriate methods are developed and made available to conservation managers. The principle of complementarity is fundamental to the most efficient of these methods. Given a goal of only a single representation of each taxon, complementarity analysis of data sets of owls, hawkmoths and tiger beetles in Thailand yields near-minimum sets of 6, 14 and 34 areas, respectively. The consensus of these sets comprises 48 areas. However, when the data are combined into a single data set, complementarity analysis gives a more efficient solution of 46 areas. Over 90% of the owls, hawkmoths and tiger beetles are already represented as a minimum of one population within the current protected areas network of Thailand. However, an additional 18 areas are still required for complete representation. Several of these additional areas are discussed and their potential contribution to biodiversity conservation in Thailand examined. Khao Pok Yo and Doi Pa Hom Pok are noted as being of particular importance. The case of Sanpatong emphasizes the need to remain aware of the biology and ecology of the species under study.     

Amphibian Hotspots and Conservation Priorities in Eastern Cuba Identified by Species Distribution Modeling

The high rate of amphibian endemism and the severe habitat modification in the Caribbean islands make them an ideal place to test if the current protected areas network might protect this group. In this study, we model distribution and map species richness of the 40 amphibian species from eastern Cuba with the objectives of identify hotspots, detect gaps in species representation in protected areas, and select additional areas to fill these gaps. We used two modeling methods, Maxent and Habitat Suitability Models, to reach a consensus distribution map for each species, then calculate species richness by combining specific models and finally performed gap analyses for species and hotspots. Our results showed that the models were robust enough to predict species distributions and that most of the amphibian hotspots were represented in reserves, but 50 percent of the species were incompletely covered and Eleutherodactylus rivularis was totally uncovered by the protected areas. We identified 1441 additional km² (9.9% of the study area) that could be added to the current protected areas, allowing the representation of every species and all hotspots. Our results are relevant for the conservation planning in other Caribbean islands, since studies like this could contribute to fill the gaps in the existing protected areas and to design a future network. Both cases would benefit from modeling amphibian species distribution using available data, even if they are incomplete, rather than relying only in the protection of known or suspected hotspots.     

Identifying Mozambique's most critical areas for plant conservation: An evaluation of protected areas and Important Plant Areas

Successful protected area networks must represent biodiversity across taxonomic groups. However, too often plant species are overlooked in conservation planning, and the resulting protected areas may, as a result, fail to encompass the most important sites for plant diversity. The Mozambique Tropical Important Plant Areas project sought to promote the conservation of Mozambique's flora through the identification of Important Plant Areas (IPAs). Here, we use the Weighted Endemism including Global Endangerment (WEGE) index to identify the richest areas for rare and endemic plants in Mozambique and subsequently evaluate how well represented these hotspots are within the current protected area and IPA networks. We also examine the congruence between IPA and protected areas to identify opportunities for strengthening the conservation of plants in Mozambique. We found that high WEGE scores, representing areas rich in endemic/near-endemic and threatened species, predict the presence of IPAs in Mozambique, but do not predict the presence of protected areas. We also find that there is limited overlap between IPAs and protected areas in Mozambique. We demonstrate how IPAs could be an important tool for ensuring priority sites for plant diversity are included within protected area network expansions, particularly following the adoption of the “30 by 30” target agreed within the post-2020 Convention on Biological Diversity framework, with great potential for this method to be replicated elsewhere in the global tropics.     

Maximizing species conservation in continental Ecuador: a case of systematic conservation planning for biodiverse regions

Ecuador has the largest number of species by area worldwide, but also a low representation of species within its protected areas. Here, we applied systematic conservation planning to identify potential areas for conservation in continental Ecuador, with the aim of increasing the representation of terrestrial species diversity in the protected area network. We selected 809 terrestrial species (amphibians, birds, mammals, and plants), for which distributions were estimated via species distribution models (SDMs), using Maxent. For each species we established conservation goals based on conservation priorities, and estimated new potential protected areas using Marxan conservation planning software. For each selected area, we determined their conservation priority and feasibility of establishment, two important aspects in the decision-making processes. We found that according to our conservation goals, the current protected area network contains large conservation gaps. Potential areas for conservation almost double the surface area of currently protected areas. Most of the newly proposed areas are located in the Coast, a region with large conservation gaps and irreversible changes in land use. The most feasible areas for conservation were found in the Amazon and Andes regions, which encompass more undisturbed habitats, and already harbor most of the current reserves. Our study allows defining a viable strategy for preserving Ecuador''s biodiversity, by combining SDMs, GIS-based decision-support software, and priority and feasibility assessments of the selected areas. This approach is useful for complementing protected area networks in countries with great biodiversity, insufficient biological information, and limited resources for conservation.     

Conservation priorities for carnivores considering protected natural areas and human population density

We conducted a prioritization exercise for 47 terrestrial carnivores in North and Central America. We used 2 by 2 degree cells to explore the spatial patterns in overall richness, regionally endemic and threatened species and identified the hotspots (the top 10% of cells in each category). We obtained optimal minimum sets of cells to represent each carnivore either (1) at least once, (2) three times, or (3) in at least 10% of its regional distribution range. Our analysis considered cells with 50% or more of their area protected, and considered human population density (HPD) per grid cell, excluding the top 10% cells with higher HPD. We found low congruence among hotspots, suggesting these should not be used alone in directing conservation strategies. About 7, 18 and 84 grid cells are needed to reach each representation goal, respectively. A much higher number of protected cells are needed to achieve the same goals. Representing 10% of each species’ distribution range required optimal sets of 47 additional cells to complement the protected cells. Irreplaceable cells had a similar or higher HPD than average values for all cells. By excluding the top 10% of cells with higher HPD, irreplaceable cells in optimal sets had much lower average HPD, but three species cannot be represented at all. By defining conservation priorities and proposing optimal networks of areas needed to represent all carnivores in the region, actual conservation efforts can be reviewed and revised. Furthermore, if our results are incorporated into a general strategy, limited resources available to conserve carnivores might be directed more efficiently.     

Bats of Borneo: diversity,distributions and representation in protected areas

Protected areas are valuable in conserving tropical biodiversity, but an insufficient understanding of species diversity and distributions makes it difficult to evaluate their effectiveness. This is especially true on Borneo, a species rich island shared by three countries, and is particularly concerning for bats, a poorly known component of mammal diversity that may be highly susceptible to landscape changes. We reviewed the diversity, distributions and conservation status of 54 bat species to determine the representation of these taxa in Borneo’s protected areas, and whether these reserves complement each other in terms of bat diversity. Lower and upper bound estimates of bat species composition were characterised in 23 protected areas and the proposed boundaries of the Heart of Borneo conservation area. We used lower and upper bound estimates of species composition. By using actual inventories, species representation was highly irregular, and even if some reserves were included in the Heart of Borneo, the protected area network would still exhibit low complementarity. By inferring species presence from distributions, composition between most reserves was similar, and complementarity was much higher. Predicting species richness using abundance information suggested that bat species representation in reserves may lie between these two extremes. We recommend that researchers better sample biodiversity over the island and address the conservation threats faced in Borneo both within and outside protected areas. While the Heart of Borneo Initiative is commendable, it should not divert attention from other conservation areas.     

Solving the maximum representation problem to prioritize areas for the conservation of terrestrial mammals at risk in Oaxaca

Oaxaca, located in south‐west México within the Mesoamerican biodiversity hotspot, holds exceptionally high biodiversity for several taxa, including mammals. It has four decreed natural protected areas (NPAs) covering 5% of its total area, but only three of these, covering only 0.2% of the area, are strictly protected as National Parks. The current study develops ecological niche models for 183 terrestrial mammals for use as biodiversity surrogates in a systematic conservation planning exercise. Forty‐five of these species were selected on the basis of their being either endangered or threatened or otherwise listed under the Mexican Red List or because they were endemic to either Oaxaca or to Mexico. The niche models were constructed with a machine‐learning algorithm (GARP, Genetic Algorithm for Rule‐Set Prediction) and refined by restricting each model to sites with suitable vegetation and habitat patches contiguous with known occurrences of the species. If the entire predicted geographical distribution of each of the 45 species listed above is put under protection, the entire state of Oaxaca gets included. Therefore, we imposed different constraints on the maximum area that can be put under protection (5–30% of the area of Oaxaca) and selected nominal conservation area networks based on different percentage representation targets for the species’ modelled distributions based on their conservation status (10–100%). The area selection utilized a rarity‐ and complementarity‐based algorithm (in the ResNet software package). The goal was to have as many as possible of the 45 species at risk meet their specified representation targets in the budgeted area. The methods developed here combine ecological niche modelling and area prioritization algorithms for integrated conservation planning in a protocol that is suitable for other highly biodiverse regions.     

Low ecological representation in the protected area network of China

Protected areas are considered as an essential strategy to halt the decline of biodiversity. Ecological representation in protected areas is crucial for assessment on the progress toward conservation targets. Although China has established a large number of protected areas since the 1950s, ecological representation of protected areas is poorly understood. Here, we performed the complementarity analysis to evaluate ecological representation of protected areas in China. We used a database of the geographical distribution for 10,396 woody plant species, 2,305 fern species, 406 amphibian species, 460 reptile species, 1,364 bird species, and 590 mammal species from 2,376 counties across China. We identified complementary sets of counties for all species or threatened species of plant and vertebrate species using a complementarity algorithm. We evaluated ecological representation of 3,627 protected areas and discerned conservation gaps by comparing the distribution of protected areas with complementary sets. The results show that the spatially representative and complementary sites for biodiversity are poorly covered, and a fairly large proportion of protected areas is not designed to efficiently represent biodiversity at the national scale. Our methodology can serve as a generic framework for assessment on ecological representation of protected areas at the national scale.     

Selecting priority areas for systematic conservation of Chinese <Emphasis Type="Italic">Rhododendron</Emphasis>: hotspot versus complementarity approaches

The use of quantitative measures to select priority areas for conservation has been in practice since the early 1980s. However, the relative efficiency of different methods for identifying priority areas is still the subject of debate. Here, using the distribution data of 556 Rhododendron species in China with high spatial resolution, we evaluated the performance of the two commonly used methods, i.e. hotspot and complementarity and selected the efficient method to select priority areas for the conservation of Rhododendron in China. By overlaying the priority areas map with the locations of protected areas, we also identified the regions not covered by current protected areas (i.e. conservation gaps). We found that the complementarity method selected less number of grid cells to capture an equivalent number of species and hence had higher efficiency and representativeness than the commonly used hotspot method. Moreover, the complementarity method was better at capturing the range-restricted species than the hotspot method. Based on the complementarity method, we identified 61 grid cells of 50?×?50 km as priority areas for Rhododendron conservation in China. Among these priority areas, only about 50% grid cells were located in the hotspot areas (e.g. Hengduan Mountains), and 14% grid cells were outside the current protected area network. Our findings suggest that, despite its popularity and ease of implementation, the sites selected by hotspot algorithm may not necessarily be the best sites to allocate conservation efforts. Since the identification of priority areas in China has largely been based on the hotspot method, the current study has revived the need to reassess the priority areas for other taxonomic groups too. More importantly, our findings have emphasized the need to expand the conservation priorities from Hengduan Mountains to south and southeast China as well.     

Geographic distribution,conservation effectiveness,and gaps for national key protected wild plants in China

National key protected wild plants (NKPWPs) are species with important conservation value based on genetics, ecology, culture, and/or scientific research, which are also confronted with serious threats. However, their geographical distribution patterns and conservation status remain unclear. In this study, we compiled 1032 species of NKPWPs. We measured the diversity to identify hotspots of NKPWPs based on species richness, weighted range size rarity and a complementarity-based analysis. Comparing the distribution and hotspots of NKPWPs with the coverage of Chinese nature reserves (NRs), we assessed conservation effectiveness and identified conservation gaps. The results identified 13 diversity hotspots; only 9.5% of them were covered by NRs with >30% of the grid cell area, and even 19.5% were not covered at all by NRs. Overall, 44.7% of NKPWPs were effectively protected by national NRs. Despite this success, 571 species in Yunnan, Guizhou, Sichuan, Chongqing, Guangxi, Guangdong, southern Hainan, Taiwan, and northern Xinjiang remain unprotected by NRs. The protected proportion of plants with first-level protection was lower than that of plants with second-level protection. The low overall proportion of protected hotspots indicates that the conservation outlook for NKPWPs is not optimistic. This study identifies priority conservation areas and conservation gaps and provides a scientific reference for the conservation of wild plants in China.     

Biodiversity conservation in metacommunity networks: linking pattern and persistence

A central goal of conservation science is to identify the most important habitat patches for maintaining biodiversity on a landscape. Spatial biodiversity patterns are often used for such assessments, and patches that harbor unique diversity are generally prioritized over those with high community similarity to other areas. This places an emphasis on biodiversity representation, but removing a patch can have cascading effects on biodiversity persistence in the remaining ecological communities. Metacommunity theory provides a mechanistic route to the linking of biodiversity patterns on a landscape with the subsequent dynamics of diversity loss after habitat is degraded. Using spatially explicit neutral theory, I focus on the situation where spatial patterns of diversity and similarity are generated by the structure of dispersal networks and not environmental gradients. I find that gains in biodiversity representation are nullified by losses in persistence, and as a result the effects of removing a patch on metacommunity diversity are essentially independent of complementarity or other biodiversity patterns. In this scenario, maximizing protected area and not biodiversity representation is the key to maintaining diversity in the long term. These results highlight the need for a broader understanding of how conservation paradigms perform under different models of metacommunity dynamics.     

黄河流域濒危物种保护热点区与保护空缺识别

黄河流域具有重要的生物多样性保护意义,通过研究珍稀濒危物种分布热点区可为生物多样性保护提供依据。选取70种濒危维管植物和陆生脊椎动物,综合多来源的分布数据,运用物种分布模型Maxent模拟物种分布区,结合自然地理区划,计算保护价值,进行热点区分析,并结合国家级自然保护区和国家公园体制试点区的分布情况进行空缺分析。研究结果显示,黄河流域濒危物种分布主要呈现出南高北低、集中于山地的特征,热点区包括秦岭区域、太行山区域、子午岭-六盘山区域、陇中高原至松潘高原、祁连山、贺兰山和沿黄湿地等。在区分自然地理区后,现有的国家级自然保护区和国家公园体制试点区覆盖了热点区面积的13.89%,保护空缺主要出现于子午岭南部、六盘山南部、松潘高原南部和拉脊山等。建议在推进黄河流域生态保护和高质量发展中将濒危物种热点区考虑在内,对黄河流域自然保护地体系进行优化,并针对黄河流域的三个自然地理分区提出了相应的保护建议。此外,研究发现,在进行热点区分析时,考虑自然地理区域划分,并综合多类群叠加和单一生物类群的分析结果进行统筹考虑,可能会更好满足生物多样性就地保护需求。     

Threat patterns for endemic plants of Argentina reveal disparity of vulnerability and protection among spatially associated species groups

Given the accelerated rate of environmental degradation and climate change, there is an urgent need to protect biodiversity, especially endemic species with restricted ranges. However, which areas should be prioritized for protection remains a critical issue. A common approach to prioritizing conservation is to rank areas using species-level metrics. Nevertheless, biodiversity and threat patterns can become complex when the amounts of data increase. Here, we analyzed the distribution of 1570 Argentinean endemic plants using clustering of spatially associated species to disentangle distribution and threat patterns. We explored vulnerability levels in each cluster using mean values of species-level metrics of vulnerability, relating values obtained to the regions and environments they occupy. For each cluster we also identified its hotspots and evaluated the effectiveness of the current protected area network for their conservation. Results yielded nine main clusters, mostly differentiated by their geographic distribution and by the ecoregions they occupy. Metrics revealed disparity in vulnerability levels among clusters, with the highest values recorded for clusters related to the Central Puna in northwestern Argentina, to the Espinal, Humid Pampas, and Low Monte in the east of the country, and to the Patagonian steppe in the south. Likewise, coverage by protected areas was low for most hotspots, with the lowest values recorded for the Patagonian cluster. In particular, for hotspot of this cluster, located along the Patagonian steppe in southern Chubut and northeastern Santa Cruz provinces, analyses showed that it has both high levels of vulnerability and low levels of protection, giving it the highest conservation priority of the entire pool analyzed. Our findings identify gaps in the current protected area network and highlight key areas in need of conservation policies and strategies, both in situ and ex situ, to protect the endemic plants of Argentina.     

Identifying biodiversity hotspots for threatened mammal species in Iran

Conservation biology has much more attention for biodiversity hot spots than before. In order to recognize the hotspots for Iranian terrestrial mammal species that are listed in any red list, nationally or globally, ten Species Distribution Models (SDMs) have been applied. The SDMs evaluation results based on the TSS and AUC values showed that all ten models of habitat suitability perform significantly better than the random selection for all studied species. According to the results, biodiversity hotspots for threatened mammal species are located in north, west and central of Iran, along the Zagros and Alborz mountain range. Therefore, habitats for threatened mammal species have been limited to small parts of Iran (approximately 27% of the country). These areas are severely fragmented and only 57% of them have been announced protected by the current conservation system. The suggestion is that, as the sustainability of these habitats would strongly depend on maintaining dispersal corridors to facilitate the movement of animals among the habitat fragments, conservation efforts should focus on those hotspots which are not formally protected under conservation laws.     

Optimized monitoring sites for detection of biodiversity trends in China

Properly designed monitoring networks can generate data to understand status and trends of biodiversity, and to assess progress towards conservation targets. However, biodiversity monitoring is often affected by poor sampling design. We proposed an approach to choosing optimized monitoring sites among large areas. Based on comprehensive distribution data of 34,284 vertebrates and vascular plants from 2376 counties in China, we selected 564 optimized monitoring sites (counties) through complementarity analysis and pre-existing knowledge of nature reserves. The optimized monitoring sites are complementary to each other and reasonably distributed, to ensure that maximum species are covered while the total number of sites and monitoring costs are minimized. Incongruence of optimized monitoring sites among different taxa indicates that taxa with different ecological features should be selected for large-scale monitoring programmes. The results of this study have been applied in the design and operation of China Biodiversity Observation Network.     

Gap analyses of priority wild relatives of food crop in current ex situ and in situ conservation in Indonesia

Conservation programmes are always limited by available resources. Careful planning is therefore required to increase the efficiency of conservation and gap analysis can be used for this purpose. This method was used to assess the representativeness of current ex situ and in situ conservation actions of 234 priority crop wild relatives (CWR) in Indonesia. This analysis also included species distribution modelling, the creation of an ecogeographical land characterization map, and a complementarity analysis to identify priorities area for in situ conservation and for further collecting of ex situ conservation programmes. The results show that both current ex situ and in situ conservation actions are insufficient. Sixty-six percent of priority CWRs have no recorded ex situ collections. Eighty CWRs with ex situ collections are still under-represented in the national genebanks and 65 CWRs have no presence records within the existing protected area network although 60 are predicted to exist in several protected areas according to their potential distribution models. The complementarity analysis shows that a minimum of 61 complementary grid areas (complementary based on grid cells) are required to conserve all priority taxa and 40 complementary protected areas (complementary based on existing protected areas) are required to conserve those with known populations within the existing in situ protected area network. The top ten of complementary protected areas are proposed as the initial areas for the development of CWR genetic reserves network in Indonesia. It is recommended to enhanced coordination between ex situ and in situ conservation stakeholders for sustaining the long term conservation of CWR in Indonesia. Implementation of the research recommendations will provide for the first time an effective conservation planning of Indonesia’s CWR diversity and will significantly enhance the country’s food and nutritional security.

     

Representation of Ecological Systems within the Protected Areas Network of the Continental United States

If conservation of biodiversity is the goal, then the protected areas network of the continental US may be one of our best conservation tools for safeguarding ecological systems (i.e., vegetation communities). We evaluated representation of ecological systems in the current protected areas network and found insufficient representation at three vegetation community levels within lower elevations and moderate to high productivity soils. We used national-level data for ecological systems and a protected areas database to explore alternative ways we might be able to increase representation of ecological systems within the continental US. By following one or more of these alternatives it may be possible to increase the representation of ecological systems in the protected areas network both quantitatively (from 10% up to 39%) and geographically and come closer to meeting the suggested Convention on Biological Diversity target of 17% for terrestrial areas. We used the Landscape Conservation Cooperative framework for regional analysis and found that increased conservation on some private and public lands may be important to the conservation of ecological systems in Western US, while increased public-private partnerships may be important in the conservation of ecological systems in Eastern US. We have not assessed the pros and cons of following the national or regional alternatives, but rather present them as possibilities that may be considered and evaluated as decisions are made to increase the representation of ecological systems in the protected areas network across their range of ecological, geographical, and geophysical occurrence in the continental US into the future.     

The sensitivity of gap analysis to conservation targets

A crucial stage in systematic conservation planning is the definition of explicit conservation targets to be achieved by a network of protected areas. A wide variety of targets have been employed, including overall percentage area, uniform representation of biodiversity features, and variable targets according to conservation interest. Despite the diversity of options, most studies adopt a particular set of targets without further explanation, and few have investigated the effect of target selection on their results. Here, using a data set on the distribution of plants and terrestrial vertebrates in southern France, we investigate how variation in targets can affect both stages of a gap analysis: the assessment of the completeness of an existing reserve network, and the prioritization of areas for its expansion. Target selection had a major impact on the gap analysis results, with uniform targets (50% of each species’ range) emphasizing the representation of common species, and contrasting targets (weighted according to species’ conservation interest) concentrating attention on high conservation interest species and the areas where they occur. Systematic conservation planning exercises should thus pay close attention to the definition and justification of the representation targets employed.     

The Importance of Rotational Crops for Biodiversity Conservation in Mediterranean Areas

Nowadays we are seeing the largest biodiversity loss since the extinction of the dinosaurs. To conserve biodiversity it is essential to plan protected areas using a prioritization approach, which takes into account the current biodiversity value of the sites. Considering that in the Mediterranean Basin the agro-ecosystems are one of the most important parts of the landscape, the conservation of crops is essential to biodiversity conservation. In the framework of agro-ecosystem conservation, farmland birds play an important role because of their representativeness, and because of their steady decline in the last Century in Western Europe. The main aim of this research was to define if crop dominated landscapes could be useful for biodiversity conservation in a Mediterranean area in which the landscape was modified by humans in the last thousand years and was affected by the important biogeographical phenomenon of peninsula effect. To assess this, we identify the hotspots and the coldspots of bird diversity in southern Italy both during the winter and in the breeding season. In particular we used a scoring method, defining a biodiversity value for each cell of a 1-km grid superimposed on the study area, using data collected by fieldwork following a stratified random sampling design. This value was analysed by a multiple linear regression analysis and was predicted in the whole study area. Then we defined the hotspots and the coldspots of the study area as 15% of the cells with higher and lower value of biodiversity, respectively. Finally, we used GAP analysis to compare hotspot distribution with the current network of protected areas. This study showed that the winter hotspots of bird diversity were associated with marshes and water bodies, shrublands, and irrigated crops, whilst the breeding hotspots were associated with more natural areas (e.g. transitional wood/shrubs), such as open areas (natural grasslands, pastures and not irrigated crops). Moreover, the results underlined the negative effects of permanent crops, such as vineyards, olive groves, and orchards, in particular during the winter season. This research highlights the importance of farmland areas mainly for wintering species and the importance of open areas for breeding species in the Mediterranean Basin. This may be true even when the species’ spatial distribution could be affected by biogeography. An important result showed that the hotspots for breeding species cannot be used as a surrogate for the wintering species, which were often not considered in the planning of protected areas.     

Biodiversity hotspots are not congruent with conservation areas in the Gulf of California

As marine systems are threatened by increasing human impacts, mechanisms to maintain biodiversity and ecosystem functions and services are needed. Protecting areas of conservation importance may serve as a proxy for maintaining these functions, while also facilitating efficient use and management of limited resources. Biodiversity hotspots have been used as surrogates for spatial conservation importance; however, as many protected areas have been established opportunistically and under differing criteria, it is unclear how well they actually protect hotspots. We evaluated how well the current protected area network and priority areas selected through previous systematic conservation planning exercises preserve biodiversity hotspots in the Gulf of California, Mexico. We also determined spatial congruence between biodiversity hotspots based on different criteria, which may determine their ability to be used as surrogates for each other. We focus on the Gulf of California because it is a megadiverse system where limited information regarding species diversity and distribution has constrained development of strategies for conservation and management. We developed a species occurrence database and identified biodiversity hotspots using four different criteria: species richness, rarity, endemism, and threatened species. We interpolated species occurrence, while accounting for heterogeneous sampling efforts. We then assessed overlap of hotspots with existing protected areas and priority areas, and between hotspots derived by distinct criteria. We gathered 286,533 occurrence records belonging to 12,105 unique species, including 6388 species identified as rare, 642 as endemic, and 386 as threatened. We found that biodiversity hotspots showed little spatial overlap with areas currently under protection and previously identified priority areas. Our results highlight the importance of distinct spatial areas of biodiversity and suggest that different ecological mechanisms sustain different aspects of diversity and multiple criteria should be used when defining conservation areas.