Sampling bias and the use of ecological niche modeling in conservation planning: a field evaluation in a biodiversity hotspot

Ecological niche modeling (ENM) has become an important tool in conservation biology. Despite its recent success, several basic issues related to algorithm performance are still being debated. We assess the ability of two of the most popular algorithms, GARP and Maxent, to predict distributions when sampling is geographically biased. We use an extensive data set collected in the Brazilian Cerrado, a biodiversity hotspot in South America. We found that both algorithms give richness predictions that are very similar to other traditionally used richness estimators. Also, both algorithms correctly predicted the presence of most species collected during fieldwork, and failed to predict species collected only in very few cases (usually species with very few known localities, i.e., <5). We also found that Maxent tends to be more sensitive to sampling bias than GARP. However, Maxent performs better when sampling is poor (e.g., low number of data points). Our results indicates that ENM, even when provided with limited and geographically biased localities, is a very useful technique to estimate richness and composition of unsampled areas. We conclude that data generated by ENM maximize the utility of existing biodiversity data, providing a very useful first evaluation. However, for reliable conservation decisions ENM data must be followed by well-designed field inventories, especially for the detection of restricted range, rare species.     

Transferability and model evaluation in ecological niche modeling: a comparison of GARP and Maxent

We compared predictive success in two common algorithms for modeling species' ecological niches, GARP and Maxent, in a situation that challenged the algorithms to be general – that is, to be able to predict the species' distributions in broad unsampled regions, here termed transferability. The results were strikingly different between the two algorithms – Maxent models reconstructed the overall distributions of the species at low thresholds, but higher predictive levels of Maxent predictions reflected overfitting to the input data; GARP models, on the other hand, succeeded in anticipating most of the species' distributional potential, at the cost of increased (apparent, at least) commission error. Receiver operating characteristic (ROC) tests were weak in discerning models able to predict into broad unsampled areas from those that were not. Such transferability is clearly a novel challenge for modeling algorithms, and requires different qualities than does predicting within densely sampled landscapes – in this case, Maxent was transferable only at very low thresholds, and biases and gaps in input data may frequently affect results based on higher Maxent thresholds, requiring careful interpretation of model results.     

The effect of sample size and species characteristics on performance of different species distribution modeling methods

Species distribution models should provide conservation practioners with estimates of the spatial distributions of species requiring attention. These species are often rare and have limited known occurrences, posing challenges for creating accurate species distribution models. We tested four modeling methods (Bioclim, Domain, GARP, and Maxent) across 18 species with different levels of ecological specialization using six different sample size treatments and three different evaluation measures. Our assessment revealed that Maxent was the most capable of the four modeling methods in producing useful results with sample sizes as small as 5, 10 and 25 occurrences. The other methods compensated reasonably well (Domain and GARP) to poorly (Bioclim) when presented with datasets of small sample sizes. We show that multiple evaluation measures are necessary to determine accuracy of models produced with presence-only data. Further, we found that accuracy of models is greater for species with small geographic ranges and limited environmental tolerance, ecological characteristics of many rare species. Our results indicate that reasonable models can be made for some rare species, a result that should encourage conservationists to add distribution modeling to their toolbox.     

Effectiveness of Protected Areas for Representing Species and Populations of Terrestrial Mammals in Costa Rica

Costa Rica has one of the greatest percentages (26%) of protected land in the world. The National Protected Areas System (NPAS) of Costa Rica was established in 1976 and currently includes >190 protected areas within seven different protection categories. The effectiveness of the NPAS to represent species, populations, and areas with high species richness has not been properly evaluated. Such evaluations are fundamental to understand what is necessary to strengthen the NPAS and better protect biodiversity. We present a novel assessment of NPAS effectiveness in protecting mammal species. We compiled the geographical ranges of all terrestrial Costa Rican mammals then determined species lists for all protected areas and the estimated proportion of each species’ geographic range protected. We also classified mammal species according to their conservation status using the IUCN Red List of Threatened Species. We found almost complete representation of mammal species (98.5%) in protected areas, but low relative coverage (28.3% on average) of their geographic ranges in Costa Rica and 25% of the species were classified as underprotected according to a priori representation targets. Interestingly, many species-rich areas are not protected, and at least 43% of cells covering the entire country are not included in protected areas. Though protected areas in Costa Rica represent species richness well, strategic planning for future protected areas to improve species complementarity and range protection is necessary. Our results can help to define sites where new protected areas can have a greater impact on mammal conservation, both in terms of species richness and range protection.     

Tourism and the Conservation of Critically Endangered Frogs

Protected areas are critical for the conservation of many threatened species. Despite this, many protected areas are acutely underfunded, which reduces their effectiveness significantly. Tourism is one mechanism to promote and fund conservation in protected areas, but there are few studies analyzing its tangible conservation outcomes for threatened species. This study uses the 415 IUCN critically endangered frog species to evaluate the contribution of protected area tourism revenue to conservation. Contributions were calculated for each species as the proportion of geographic range inside protected areas multiplied by the proportion of protected area revenues derived from tourism. Geographic ranges were determined from IUCN Extent of Occurrence maps. Almost 60% (239) of critically endangered frog species occur in protected areas. Higher proportions of total range are protected in Nearctic, Australasian and Afrotopical regions. Tourism contributions to protected area budgets ranged from 5–100%. These financial contributions are highest for developing countries in the Afrotropical, Indomalayan and Neotropical regions. Data for both geographic range and budget are available for 201 critically endangered frog species with proportional contributions from tourism to species protection ranging from 0.8–99%. Tourism''s financial contributions to critically endangered frog species protection are highest in the Afrotropical region. This study uses a coarse measure but at the global scale it demonstrates that tourism has significant potential to contribute to global frog conservation efforts.     

The performance of the global protected area system in capturing vertebrate geographic ranges

Given the heavy reliance placed on and investment in protected areas for biological conservation, there has been much debate as to how effective these are in representing biodiversity features within their boundaries. The majority of studies addressing this issue have been conducted on a regional or national basis, precluding a broad picture of patterns of representation at the species level. We present a global assessment of the representation of the terrestrial geographic ranges of complete taxonomic groups: all known extant amphibians, birds and mammals (20,736 species) within the current global system of protected areas. We conclude that it is necessary substantially to improve the levels of coverage of the geographic ranges of the majority of species, even the widespread ones. This is particularly true for rare species, which might be assumed to be foci for protected area systems. To improve on the low levels of coverage of vertebrate ranges attained by the existing areas, key regions should be targeted, but heavy reliance will also have to be placed on approaches to sustaining populations in the wider, unprotected landscape.     

Vegetation dynamics and avian seasonal migration: clues from remotely sensed vegetation indices and ecological niche modelling

Aim Regional movements of tropical birds are among the least understood patterns of migration, generally assumed to be related to seasonality of vegetation and food resources. We used readily available remotely sensed data to analyse this relationship in the three‐wattled bellbird (Procnias tricarunculatus), an IUCN Vulnerable canopy frugivorous bird that undergoes localized altitudinal and latitudinal movements between several non‐breeding and breeding ranges. Location Central America. Methods We generated ecological niche models (GARP and Maxent) based on remotely sensed vegetation indices (enhanced vegetation index, EVI; red index, RI; and normalized difference water index, NDWI) that were used as proxies for canopy characteristics and phenological changes between seasons. The variation in EVI was also analysed in relation to known bellbird seasonal migration patterns. Results Remotely sensed summaries of intra‐annual vegetation variation could not explain known movement patterns of bellbirds breeding in Monteverde Cloud Forest Reserve, Costa Rica. The ecological niche modelling (ENM) framework used for exploring potential movement patterns of another population of bellbirds, observed in March 2004 in Corcovado National Park, Costa Rica, showed that the Atlantic and Pacific lowlands are suitable for the bellbirds year‐round, while middle elevations, where breeding may occur, have limited suitability. Main conclusions Our findings suggest that: (1) the Corcovado population tracks similar environmental conditions during the non‐breeding season; (2) migration to middle elevation breeding sites might be related to factors other than vegetation seasonality or breeding might not be restricted to these elevation ranges; and (3) the lowlands comprising most of the suitable areas for bellbirds are where anthropogenic pressure is highest, so conservation status should consider this threat, and conservation planning should emphasize these areas.     

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.     

Global habitat suitability models of terrestrial mammals

Detailed large-scale information on mammal distribution has often been lacking, hindering conservation efforts. We used the information from the 2009 IUCN Red List of Threatened Species as a baseline for developing habitat suitability models for 5027 out of 5330 known terrestrial mammal species, based on their habitat relationships. We focused on the following environmental variables: land cover, elevation and hydrological features. Models were developed at 300 m resolution and limited to within species' known geographical ranges. A subset of the models was validated using points of known species occurrence. We conducted a global, fine-scale analysis of patterns of species richness. The richness of mammal species estimated by the overlap of their suitable habitat is on average one-third less than that estimated by the overlap of their geographical ranges. The highest absolute difference is found in tropical and subtropical regions in South America, Africa and Southeast Asia that are not covered by dense forest. The proportion of suitable habitat within mammal geographical ranges correlates with the IUCN Red List category to which they have been assigned, decreasing monotonically from Least Concern to Endangered. These results demonstrate the importance of fine-resolution distribution data for the development of global conservation strategies for mammals.     

Land use change and the dependence of national priority species on protected areas

The establishment and maintenance of a system of protected areas is central to regional and global strategies for the conservation of biodiversity. The current global trend towards human population growth and widespread environmental degradation means that such areas are becoming increasingly isolated, fragmented habitat islands. In regions in which this process is well advanced, a high proportion of species are thus predicted to have become restricted to protected areas. Here, using uniquely detailed datasets for Britain, a region with close to the global level of percentage coverage by statutory protected areas, we determine the extent of restriction of species of conservation concern to these areas. On the basis of currently known distributions, more than a half of such species are highly dependent on protected areas for their continued persistence, occurring either entirely or largely within their bounds. Such coverage is of particular importance for those species with narrower distributions, and therefore, under the greatest threats, underlining the vital importance of adequately resourcing, maintaining, and developing protected areas to prevent these species from being lost.     

Planning the priority protected areas of endangered orchid species in northeastern China

We used endangered orchid species as a case study for establishing priority protected areas (PPAs). We conducted a detailed investigation of 2,884 plots in northeastern China and selected 6 nationally protected orchid species (the number of plots for each species ≥5) to plan PPAs in order to conserve these endangered plants. First, we estimated the potential habitat distributions of 6 orchid species using species distribution models (in the program Maxent). Second, we identified PPAs using the results of both Maxent and systematic conservation planning software (Zonation). We found that the PPAs we identified were all distributed in the eastern, northeastern, and northern parts of the study region on which we focused. We observed that existing nature reserves in those areas were much smaller than the PPAs we identified, suggesting the need for expansion of these zones. Finally, we identified some recommended protection areas based on our PPAs, which we believe are suitable areas of in situ conservation. These areas could assist in the formulation of protection policies for these endangered orchid species. Our results have broad implications for conservation biology and the redefinition of protective zones for at-risk species.     

Protected areas alleviate climate change effects on northern bird species of conservation concern

Global climate change is a major threat to biodiversity, posing increasing pressures on species to adapt in situ or shift their ranges. A protected area network is one of the main instruments to alleviate the negative impacts of climate change. Importantly, protected area networks might be expected to enhance the resilience of regional populations of species of conservation concern, resulting in slower species loss in landscapes with a significant amount of protected habitat compared to unprotected landscapes. Based on national bird atlases compiled in 1974–1989 and 2006–2010, this study examines the recent range shifts in 90 forest, mire, marshland, and Arctic mountain heath bird species of conservation concern in Finland, as well as the changes in their species richness in protected versus unprotected areas. The trends emerging from the atlas data comparisons were also related to the earlier study dealing with predictions of distributional changes for these species for the time slice of 2051–2080, developed using bioclimatic envelope models (BEMs). Our results suggest that the observed changes in bird distributions are in the same direction as the BEM‐based predictions, resulting in a decrease in species richness of mire and Arctic mountain heath species and an increase in marshland species. The patterns of changes in species richness between the two time slices are in general parallel in protected and unprotected areas. However, importantly, protected areas maintained a higher level of species richness than unprotected areas. This finding provides support for the significance and resilience provision of protected area networks in preserving species of conservation concern under climate change.     

Predicting species distributions from small numbers of occurrence records: a test case using cryptic geckos in Madagascar

Aim Techniques that predict species potential distributions by combining observed occurrence records with environmental variables show much potential for application across a range of biogeographical analyses. Some of the most promising applications relate to species for which occurrence records are scarce, due to cryptic habits, locally restricted distributions or low sampling effort. However, the minimum sample sizes required to yield useful predictions remain difficult to determine. Here we developed and tested a novel jackknife validation approach to assess the ability to predict species occurrence when fewer than 25 occurrence records are available. Location Madagascar. Methods Models were developed and evaluated for 13 species of secretive leaf‐tailed geckos (Uroplatus spp.) that are endemic to Madagascar, for which available sample sizes range from 4 to 23 occurrence localities (at 1 km² grid resolution). Predictions were based on 20 environmental data layers and were generated using two modelling approaches: a method based on the principle of maximum entropy (Maxent) and a genetic algorithm (GARP). Results We found high success rates and statistical significance in jackknife tests with sample sizes as low as five when the Maxent model was applied. Results for GARP at very low sample sizes (less than c. 10) were less good. When sample sizes were experimentally reduced for those species with the most records, variability among predictions using different combinations of localities demonstrated that models were greatly influenced by exactly which observations were included. Main conclusions We emphasize that models developed using this approach with small sample sizes should be interpreted as identifying regions that have similar environmental conditions to where the species is known to occur, and not as predicting actual limits to the range of a species. The jackknife validation approach proposed here enables assessment of the predictive ability of models built using very small sample sizes, although use of this test with larger sample sizes may lead to overoptimistic estimates of predictive power. Our analyses demonstrate that geographical predictions developed from small numbers of occurrence records may be of great value, for example in targeting field surveys to accelerate the discovery of unknown populations and species.     

Development of a scoring method to identify important areas of plant diversity in Ireland

In the face of accelerating biodiversity loss it is more important than ever to identify important areas of biodiversity and target limited resources for conservation. We developed a method to identify areas of important plant diversity using known species’ distributions and evaluations of the species importance. We collated distribution records of vascular plants and developed a scoring method of spatial prioritisation to assign conservation value to the island of Ireland at the hectad scale (10 km × 10 km) and at the tetrad scale (2 km × 2 km) for two counties where sufficient data were available. Each plant species was assigned a species conservation value based on both its conservation status and distribution in Ireland. For each cell, the species conservation values within the cell were summed, thereby differentiating between areas of high and low conservation value across the landscape. Areas with high conservation value represent the most important areas for plant conservation.The protected area cover and the number of species present in these important areas were also examined by first defining threshold values using two different criteria. Species representation was high in the important areas; the identified important areas of plant diversity maintained high representation of species of conservation concern and achieved high species representation overall, requiring a low number of sites (<8%) to do so. The coincidence of protected areas and important areas for plant diversity was found to be low and while some important areas of plant diversity might benefit from the general protection afforded by these areas, our research highlights the need for conservation outside of protected areas.     

Predicting potential distribution of chestnut phylloxerid (Hemiptera: Phylloxeridae) based on GARP and Maxent ecological niche models

The chestnut phylloxerid, Moritziella castaneivora, has been recently recorded as a forest pest in China. It heavily damaged chestnut trees and has caused serious economic losses in some main chestnut production areas. In order to effectively monitor and manage this pest, it is necessary to investigate its potential geographical distribution worldwide. In this study, we used two ecological niche models, Genetic Algorithm for Rule‐set Production (GARP) and Maximum Entropy (Maxent), along with the geographical distribution of the host plants, Japanese chestnut (Castanea crenata) and Chinese chestnut (Castanea mollissima), to predict the potential geographical distribution of M. castaneivora. The results suggested that the suitable distribution areas based on GARP were general consistent with those based on Maxent, but GARP predicted distribution areas that extended more in size than did Maxent. The results also indicated that the suitable areas for chestnut phylloxerid infestations were mainly restricted to Northeast China (northern Liaoning), East China (southern Shandong, northern Jiangsu and western Anhui), North China (southern Hebei, Beijing and Tianjin), Central China (eastern Hubei and southern Henan), Japan (Kinki, Shikoku and Tohoku) and most parts of the Korean Peninsula. In addition, some provinces of central and western China were predicted to have low suitability or unsuitable areas (e.g. Xinjiang, Qinghai and Tibet). A jackknife test in Maxent showed that the average precipitation in July was the most important environmental variable affecting the distribution of this pest species. Consequently, the study suggests several reasonable regulations and management strategies for avoiding the introduction or invasion of this high‐risk chestnut pest to these potentially suitable areas.     

Using niche modelling and human influence index to indicate conservation priorities for Atlantic forest deer species

To prioritise conservation actions and management strategies for threatened forest deer species at the Atlantic forest, we aimed to identify and describe the most suitable habitat areas for forest deer species and to indicate conservation measures for state agents and local communities. We adopt an approach based on ecological niche modelling, key variable thresholds and spatial analyses. In addition, we associated our approach with a human influence index, an invasive species dataset of occurrences, protected area cover and IUCN category. We indicate 2 % (484 km²) of the Atlantic forest cover as conservation priority areas (CPAs). Of these, 56.8 % are outside protected areas, 20.7 % are inside IUCN categories i, ii and iii protected areas, 19.9 % are inside IUCN categories iv, v, and vi protected areas, and 2.6 % are inside indigenous areas. Also, we indicate the most relevant protected areas for deer conservation in the Atlantic forest. The CPAs were classified into more human-influenced areas (MHIA) and less human-influenced areas (LHIA), and we identified 21 significant (greater than120 km²) continuous CPAs outside protected areas. We highlight actions in several perspectives of human influence, governance levels and law protection that would rationalise the use of funds and human resources.     

Modelling the geographic distributions of endemic genera in the eastern Central Asian desert

This study simulates the distributions of 13 endemic and near‐endemic genera (Ammopiptanthus, Sympegma, Iljinia, Elachanthemum, Potaninia, Tugarinovia, Kaschgaria, Sarcozygium, Timouria, Zollikoferia, Stilpnolepis, Synstemon and Tetraena) to indicate areas of plant diversity and conservation importance within the eastern Central Asian desert, and to identify the determinant environmental variables contributing to the spatial distribution patterns. Using known distribution localities and 14 environmental variables, the Maxent and Domain species distribution models were employed to map the patterns of geographic distribution. The power of predictability of the models was tested using the receiver operating characteristic method and the jackknife validation approach, according to the different number of species localities available. The estimated richness and the superimposed potential distributions of 13 genera were used to indicate endemic patterns of distribution. The comparison of Maxent and Domain further identified previously unknown areas of endemism and described the distribution for each taxon. Both observed species occurrence and the species occurrence predicted from the Maxent indicated that the eastern Alashan of Inner Mongolia is the most noticeable endemic area, and the northwestern and northern Tarim Basin of Xinjiang is the secondary center of plant diversity. These regions were then prioritized for conservation importance. Potential evapotranspiration ratio and precipitation seasonality played important roles in driving the observed patterns of endemic distribution.     

Biodiversity estimation of the western region of Ghana using arthropod mean morphospecies abundance

There is a global consensus on the fact that human activities are increasingly eroding biodiversity, especially in many ecologically rich regions of the world. Quantifying these biodiversity losses over large administrative areas, along with the impacts of the respective driving factors is necessary for biodiversity conservation. In this study, we used readily available spatial data on infrastructure, fragmentation, and land use together with ecological data on terrestrial arthropods (as species surrogate) to carry out spatial assessment and mapping of biodiversity values in the Western Region of Ghana, a highly degraded part of the West African biodiversity hotspot. Compared with its pristine condition, our results show that the region has a remaining mean species abundance (MSA) of approximately 52%, indicating a loss of 48% due to anthropogenic activities (mainly cocoa and oil palm farming). Areas of national protection (i.e. national parks and forest reserves) did not prove to have much promise for biodiversity conservation in the region since many were found to have low MSA values. Moreover, according to our arthropod data, no areas of high MSA values were found outside protected areas. However, because the thick, continuous, structurally complex canopies of these orchards responsible for the biodiversity loss provide natural connectivity between protected reserves, diversification of understory vegetation and a reduction in the use of agrochemicals particularly weedicides and pesticides may yield significant conservation outcomes.     

Towards systematic and evidence‐based conservation planning for western chimpanzees

As animal populations continue to decline, frequently driven by large‐scale land‐use change, there is a critical need for improved environmental planning. While data‐driven spatial planning is widely applied in conservation, as of yet it is rarely used for primates. The western chimpanzee (Pan troglodytes verus) declined by 80% within 24 years and was uplisted to Critically Endangered by the IUCN Red List of Threatened Species in 2016. To support conservation planning for western chimpanzees, we systematically identified geographic areas important for this taxon. We based our analysis on a previously published data set of modeled density distribution and on several scenarios that accounted for different spatial scales and conservation targets. Across all scenarios, typically less than one‐third of areas we identified as important are currently designated as high‐level protected areas (i.e., national park or IUCN category I or II). For example, in the scenario for protecting 50% of all chimpanzees remaining in West Africa (i.e., approximately 26,500 chimpanzees), an area of approximately 60,000 km² was selected (i.e., approximately 12% of the geographic range), only 24% of which is currently designated as protected areas. The derived maps can be used to inform the geographic prioritization of conservation interventions, including protected area expansion, “no‐go‐zones” for industry and infrastructure, and conservation sites outside the protected area network. Environmental guidelines by major institutions funding infrastructure and resource extraction projects explicitly require corporations to minimize the negative impact on great apes. Therefore, our results can inform avoidance and mitigation measures during the planning phases of such projects. This study was designed to inform future stakeholder consultation processes that could ultimately integrate the conservation of western chimpanzees with national land‐use priorities. Our approach may help in promoting similar work for other primate taxa to inform systematic conservation planning in times of growing threats.     

What works for threatened species recovery? An empirical evaluation for Australia

Despite the growing numbers of threatened species and high levels of spending on their recovery worldwide, there is surprisingly little evidence about which conservation approaches are effective in arresting or reversing threatened species declines. Using two government data sets, we examined associations between population trends for 841 nationally-threatened terrestrial species in Australia, and four measures of conservation effort: (a) how much their distribution overlaps with strictly protected areas (IUCN I–IV), (b) and other protected areas (IUCN V–VI), (c) the number of recovery activities directed at the species, and (d) numbers of natural resource conservation activities applied in areas where populations of the threatened species occur. We found that all populations of 606 (72%) species were in decline. Species with greater distributional overlap with strictly protected areas had proportionately more populations that were increasing or stable. This effect was robust to geographic range size, data quality differences and extent of protection. Measures other than strictly protected areas showed no positive associations with stable or increasing trends. Indeed, species from regions with more natural resource conservation activities were found to be more likely to be declining, consistent with differential targeting of such generalised conservation activities to highly disturbed landscapes. Major differences in trends were also found among the different jurisdictions in which species predominantly occurred, which may be related to different legislative protections against habitat destruction. Although we were not able to test causation, this research corroborates other evidence that protected areas contribute to the stabilization or recovery of threatened species, and provides little empirical support for other conservation approaches.