Prevalence, thresholds and the performance of presence-absence models

Predicting suitable habitats for foraging and migration in Eastern Indian Ocean pygmy blue whales from satellite tracking data

BACKGROUND

Accurate predictions of animal occurrence in time and space are crucial for informing and implementing science-based management strategies for threatened species.

METHODS

We compiled known, available satellite tracking data for pygmy blue whales in the Eastern Indian Ocean (n = 38), applied movement models to define low (foraging and reproduction) and high (migratory) move persistence underlying location estimates and matched these with environmental data. We then used machine learning models to identify the relationship between whale occurrence and environment, and predict foraging and migration habitat suitability in Australia and Southeast Asia.

RESULTS

Our model predictions were validated by producing spatially varying accuracy metrics. We identified the shelf off the Bonney Coast, Great Australian Bight, and southern Western Australia as well as the slope off the Western Australian coast as suitable habitat for migration, with predicted foraging/reproduction suitable habitat in Southeast Asia region occurring on slope and in deep ocean waters. Suitable foraging habitat occurred primarily on slope and shelf break throughout most of Australia, with use of the continental shelf also occurring, predominanly in South West and Southern Australia. Depth of the water column (bathymetry) was consistently a top predictor of suitable habitat for most regions, however, dynamic environmental variables (sea surface temperature, surface height anomaly) influenced the probability of whale occurrence.

CONCLUSIONS

Our results indicate suitable habitat is related to dynamic, localised oceanic processes that may occur at fine temporal scales or seasonally. An increase in the sample size of tagged whales is required to move towards developing more dynamic distribution models at seasonal and monthly temporal scales. Our validation metrics also indicated areas where further data collection is needed to improve model accuracy. This is of particular importance for pygmy blue whale management, since threats (e.g., shipping, underwater noise and artificial structures) from the offshore energy and shipping industries will persist or may increase with the onset of an offshore renewable energy sector in Australia.

Conservation units alone are insufficient to protect Brazilian Amazonian chelonians

The creation of protected areas (PAs) is not always based on science; consequently, some aquatic species may not receive the same level of protection as terrestrial ones. The objective of this study was to identify priority areas for the conservation of chelonians in the Brazilian Amazon basin and assess the contribution of PAs, distinguishing between Full Protection Areas, Sustainable Use Areas, and Indigenous Lands for group protection. The entire species modeling procedure was carried out using Species Distribution Models. Location records were obtained from platforms such as SpeciesLink, GBIF, the Hydroatlas database, and WorldClim for bioclimatic variables adjusted with algorithms like Maximum Entropy, Random Forest, Support Vector Machine, and Gaussian-Bayesian. Indigenous lands cover more than 50% of the distribution areas of chelonian species in the Brazilian Amazon. Protected areas with higher conservation importance (Full Protection Areas and Sustainable Use Areas) hold less than 15% of the combined species distribution. Researchers face significant challenges when making decisions with models, especially in conservation efforts involving diverse taxa that differ significantly from one another within a group of individuals.

Risk of spread of the Asian citrus psyllid Diaphorina citri Kuwayama (Hemiptera: Liviidae) in Ghana

The impact of invasive species on biodiversity, food security and economy is increasingly noticeable in various regions of the globe as a consequence of climate change. Yet, there is limited research on how climate change affects the distribution of the invasive Asian citrus psyllid Diaphorina citri Kuwayama (Hemiptera:Liviidae) in Ghana. Using maxnet package to fit the Maxent model in R software, we answered the following questions; (i) what are the main drivers for D. citri distribution, (ii) what are the D. citri-specific habitat requirements and (iii) how well do the risk maps fit with what we know to be correctly based on the available evidence?. We found that temperature seasonality (Bio04), mean temperature of warmest quarter (Bio10), precipitation of driest quarter (Bio17), moderate resolution imaging spectroradiometer land cover and precipitation seasonality (Bio15), were the most important drivers of D. citri distribution. The results follow the known distribution records of the pest with potential expansion of habitat suitability in the future. Because many invasive species, including D. citri, can adapt to the changing climates, our findings can serve as a guide for surveillance, tracking and prevention of D. citri spread in Ghana.

A species distribution model of the giant kelp Macrocystis pyrifera: Worldwide changes and a focus on the Southeast Pacific

Worldwide climate-driven shifts in the distribution of species is of special concern when it involves habitat-forming species. In the coastal environment, large Laminarian algae-kelps-form key coastal ecosystems that support complex and diverse food webs. Among kelps, Macrocystis pyrifera is the most widely distributed habitat-forming species and provides essential ecosystem services. This study aimed to establish the main drivers of future distributional changes on a global scale and use them to predict future habitat suitability. Using species distribution models (SDM), we examined the changes in global distribution of M. pyrifera under different emission scenarios with a focus on the Southeast Pacific shores. To constrain the drivers of our simulations to the most important factors controlling kelp forest distribution across spatial scales, we explored a suite of environmental variables and validated the predictions derived from the SDMs. Minimum sea surface temperature was the single most important variable explaining the global distribution of suitable habitat for M. pyrifera. Under different climate change scenarios, we always observed a decrease of suitable habitat at low latitudes, while an increase was detected in other regions, mostly at high latitudes. Along the Southeast Pacific, we observed an upper range contraction of -17.08° S of latitude for 2090-2100 under the RCP8.5 scenario, implying a loss of habitat suitability throughout the coast of Peru and poleward to -27.83° S in Chile. Along the area of Northern Chile where a complete habitat loss is predicted by our model, natural stands are under heavy exploitation. The loss of habitat suitability will take place worldwide: Significant impacts on marine biodiversity and ecosystem functioning are likely. Furthermore, changes in habitat suitability are a harbinger of massive impacts in the socio-ecological systems of the Southeast Pacific.

Unveiling of climate change-driven decline of suitable habitat for Himalayan bumblebees

Insect pollinators, especially bumblebees are rapidly declining from their natural habitat in the mountain and temperate regions of the world due to climate change and other anthropogenic activities. We still lack reliable information about the current and future habitat conditions of bumblebees in the Himalaya. In this study, we used the maximum entropy algorithm for SDM to look at current and future (in 2050 and 2070) suitable habitats for bumblebees in the Himalaya. We found that the habitat conditions in the Himalayan mountain range do not have a very promising future as suitable habitat for most species will decrease over the next 50 years. By 2050, less than 10% of the Himalayan area will remain a suitable habitat for about 72% of species, and by 2070 this number will be raised to 75%. During this time period, the existing suitable habitat of bumblebees will be declined but some species will find new suitable habitat which clearly indicates possibility of habitat range shift by Himalayan bumblebees. Overall, about 15% of the Himalayan region is currently highly suitable for bumblebees, which should be considered as priority areas for the conservation of these pollinators. Since suitable habitats for bumblebees lie between several countries, nations that share international borders in the Himalayan region should have international agreements for comprehensive pollinator diversity conservation to protect these indispensable ecosystem service providers.

A field-validated ensemble species distribution model of Eriogonum pelinophilum, an endangered subshrub in Colorado, USA

Understanding the suitable habitat of endangered species is crucial for agencies such as the Bureau of Land Management to plan management and conservation. However, few species distribution models are directly validated, potentially limiting their application in management. In preparation for a Species Status Assessment of clay-loving wild buckwheat (Eriogonum pelinophilum), an endangered subshrub found in southwest Colorado, we ran a series of species distribution models to estimate the species' potential occupied habitat and validated these models in the field. A 1-meter resolution digital elevation model derived from LiDAR and a high-resolution geology mapping helped identify biologically relevant characteristics of the species' habitat. We employed a weighted ensemble model based on two Random Forest and one Boosted Regression Tree model, and discrimination performance of the ensemble model was high (AUC-PR = 0.793). We then conducted a systematic field survey of model habitat suitability predictions, during which we discovered 55 new subpopulations of the species and demonstrated that new species observations were strongly associated with model predictions (p < .0001, Cliff's delta = 0.575). We further refined our original models by incorporating the additional species occurrences collected in the field survey, a new explanatory variable, and a more diverse set of models. These iterative changes marginally improved performance of the ensemble model (AUC-PR = 0.825). Direct validation of species distribution models is extremely rare, and our field survey provides strong validation of our model results. This helps increase confidence to utilize predictions in planning. The final model predictions greatly improve the Bureau of Land Management's understanding of the species' habitat and increase our ability to consider potential habitat in planning land use activities such as road development and travel management.

Global terrestrial invasions: Where naturalised birds, mammals, and plants might spread next and what affects this process

More species live outside their native range than at any point in human history. Yet, there is little understanding of the geographic regions that will be threatened if these species continue to spread, nor of whether they will spread. We predict the world's terrestrial regions to which 833 naturalised plants, birds, and mammals are most imminently likely to spread, and investigate what factors have hastened or slowed their spread to date. There is huge potential for further spread of naturalised birds in North America, mammals in Eastern Europe, and plants in North America, Eastern Europe, and Australia. Introduction history, dispersal, and the spatial distribution of suitable areas are more important predictors of species spread than traits corresponding to habitat usage or biotic interactions. Natural dispersal has driven spread in birds more than in plants. Whether these taxa continue to spread more widely depends partially on connectivity of suitable environments. Plants show the clearest invasion lag, and the putative importance of human transportation indicates opportunities to slow their spread. Despite strong predictive effects, questions remain, particularly why so many birds in North America do not occupy climatically suitable areas close to their existing ranges.

Identifying the environmental and anthropogenic causes, distribution, and intensity of human rhesus macaque conflict in Nepal

Reducing conflict between humans and wildlife is considered a top conservation priority. However, increasingly human-induced disturbances across natural landscapes have escalated encounters between humans and wildlife. In Nepal, forests have been destroyed, fragmented, and developed for human settlements, agricultural production, and urban centers for decades. As a result, human-wildlife conflict, in the form of crop-raiding, livestock predation, and injuries to humans and wildlife, is common throughout the country. In particular, crop-raiding by macaques is an increasingly common form of human-wildlife conflict. Rhesus macaques (Macaca mulatta) have been identified as a top ten crop-raiding wildlife species in Nepal. In order to better understand the nationwide distribution and intensity of human-rhesus macaque conflict (HRMC), we conducted an extensive literature review of reported incidences of HRMC during the period 2000 to 2021 in Nepal. We also created an online survey to obtain nationwide data on the location and severity of HRMC, and modeled the set of ecological factors that affect habitat suitability for rhesus macaques. An ensemble of three different species distribution model (SDM) algorithms were used to analyze these data. We found that almost 44% of Nepal's land area contains suitable habitat for rhesus macaques, with less than 8% of all suitable habitat located in protected national parks. As humans continue to alter and fragment natural landscapes, HRMC in Nepal has intensified. At present, nearly 15% of the country's land area in which human settlements are permitted, is characterized by moderate or high rates of HRMC. We argue that prioritizing programs of forest restoration, strategic management plans designed to connect isolated forest fragments with high rhesus macaque population densities, creating government programs that compensate farmers for income lost due to crop-raiding, and educational outreach that informs local villagers of the importance of conservation and protecting biodiversity, offer the most effective solutions to reduce HRMC in Nepal.

Present and future distribution of bat hosts of sarbecoviruses: implications for conservation and public health

Global changes in response to human encroachment into natural habitats and carbon emissions are driving the biodiversity extinction crisis and increasing disease emergence risk. Host distributions are one critical component to identify areas at risk of viral spillover, and bats act as reservoirs of diverse viruses. We developed a reproducible ecological niche modelling pipeline for bat hosts of SARS-like viruses (subgenus Sarbecovirus), given that several closely related viruses have been discovered and sarbecovirus-host interactions have gained attention since SARS-CoV-2 emergence. We assessed sampling biases and modelled current distributions of bats based on climate and landscape relationships and project future scenarios for host hotspots. The most important predictors of species distributions were temperature seasonality and cave availability. We identified concentrated host hotspots in Myanmar and projected range contractions for most species by 2100. Our projections indicate hotspots will shift east in Southeast Asia in locations greater than 2°C hotter in a fossil-fuelled development future. Hotspot shifts have implications for conservation and public health, as loss of population connectivity can lead to local extinctions, and remaining hotspots may concentrate near human populations.

iPODfish - A new method to infer the historical occurrence of diadromous fish species along river networks

Available information on diadromous fish species historical occurrences is generally biased and incomplete across species distribution range and spatial scales. This work aims to establish a new methodological framework (iPODfish - Inferring Past Occurrences of Diadromous Fish) to obtain a more complete representation of the historical occurrences of diadromous fish species over their full distribution range. The iPODfish is based on assumptions, rules and thresholds derived from the interplay between freshwater network features, diadromous fish species ecology and known historical occurrence. These are used to establish historical pseudo-occurrences at the segment scale, i.e., locations where the species was most likely to be present or absent. The methodology is expressed by a tree-like representation of a stepwise, information supported, decision process. It has five steps (separating main river segments from tributary segments; accounting for segments specificities; imposing the relative distance threshold; imposing the Strahler value threshold, and; establishing the sub-basin Strahler threshold), divides into two moments of application (main river followed by the tributaries) and establishes presences, pseudo-presences and pseudo-absences. The iPODfish can deal with multiple information sources, cope with data bias and provide a reliable consistent historical occurrence output at the segment scale for the known historical geographical range of the species. Despite its inference nature, iPODfish is still a conservative procedure leading to ecologically coherent outputs that may be applied to any diadromous fish species (with a relevant amount of historical data available) in any river network throughout the globe because the concepts and definitions used are general ecological features of diadromous and freshwater networks. The method outputs are applicable in biogeographical and/or macroecological studies using historical data and may prove useful to the management and conservation of diadromous fish species.

Comparing, evaluating and combining statistical species distribution models and CLIMEX to forecast the distributions of emerging crop pests

BACKGROUND

Forecasting the spread of emerging pests is widely requested by pest management agencies in order to prioritise and target efforts. Two widely used approaches are statistical Species Distribution Models (SDMs) and CLIMEX, which uses ecophysiological parameters. Each have strengths and weaknesses. SDMs can incorporate almost any environmental condition and their accuracy can be formally evaluated to inform managers. However, accuracy is affected by data availability and can be limited for emerging pests, and SDMs usually predict year-round distributions, not seasonal outbreaks. CLIMEX can formally incorporate expert ecophysiological knowledge and predicts seasonal outbreaks. However, the methods for formal evaluation are limited and rarely applied. We argue that both approaches can be informative and complementary, but we need tools to integrate and evaluate their accuracy. Here we develop such an approach, and test it by forecasting the potential global range of the tomato pest Tuta absoluta.

RESULTS

The accuracy of previously developed CLIMEX and new statistical SDMs were comparable on average, but the best statistical SDM techniques and environmental data substantially outperformed CLIMEX. The ensembled approach changes expectations of T. absoluta's spread. The pest's environmental tolerances and potential range in Africa, the Arabian Peninsula, Central Asia and Australia will be larger than previous estimates.

CONCLUSION

We recommend that CLIMEX be considered one of a suite of SDM techniques and thus evaluated formally. CLIMEX and statistical SDMs should be compared and ensembled if possible. We provide code that can be used to do so when employing the biomod suite of SDM techniques. © 2021 The Authors. Pest Management Science published by John Wiley & Sons Ltd on behalf of Society of Chemical Industry.

Combining Citizen Science Data and Satellite Descriptors of Ecosystem Functioning to Monitor the Abundance of a Migratory Bird during the Non-Breeding Season

Migratory birds are particularly exposed to habitat changes in their breeding and non-breeding grounds. Remote sensing technologies offer an excellent opportunity to monitor species’ habitats from space at unprecedented spatiotemporal scales. We analyzed if remotely sensed ecosystem functioning attributes (EFAs) adequately predict the spatiotemporal variation of the Woodcock’s (Scolopax rusticola) relative abundance in southwest Europe, during autumn migration and wintering periods. We used data gathered from Woodcock monitoring through citizen science (N = 355,654 hunting trips) between 2009 and 2018. We computed a comprehensive set of EFAs on a weekly basis from three MODIS satellite products: enhanced vegetation index (EVI), tasseled cap transformation (TCT), and land surface temperature (LST). We developed generalized linear mixed models to explore the predictive power of EFAs on Woodcock’s abundance during the non-breeding season. Results showed that Woodcock abundance is correlated with spatiotemporal dynamics in primary productivity (measured through the EVI), water cycle dynamics (wetness component of TCT), and surface energy balance (LST) in both periods. Our findings underline the potential of combining citizen science and remote sensing data to monitor migratory birds throughout their life cycles—an issue of critical importance to ensure adequate habitat management in the non-breeding areas.

Not all species will migrate poleward as the climate warms: The case of the seven baobab species in Madagascar

It is commonly accepted that species should move toward higher elevations and latitudes to track shifting isotherms as climate warms. However, temperature might not be the only limiting factor determining species distribution. Species might move to opposite directions to track changes in other climatic variables. Here, we used an extensive occurrence data set and an ensemble modelling approach to model the climatic niche and to predict the distribution of the seven baobab species (genus Adansonia) present in Madagascar. Using climatic projections from three global circulation models, we predicted species' future distribution and extinction risk for 2055 and 2085 under two representative concentration pathways (RCPs) and two dispersal scenarios. We disentangled the role of each climatic variable in explaining species range shift looking at relative variable importance and future climatic anomalies. Four baobab species (Adansonia rubrostipa, Adansonia madagascariensis, Adansonia perrieri¸ and Adansonia suarezensis) could experience a severe range contraction in the future (>70% for year 2085 under RCP 8.5, assuming a zero-dispersal hypothesis). For three out of the four threatened species, range contraction was mainly explained by an increase in temperature seasonality, especially in the North of Madagascar, where they are currently distributed. In tropical regions, where species are commonly adapted to low seasonality, we found that temperature seasonality will generally increase. It is, thus, very likely that many species in the tropics will be forced to move equatorward to avoid an increase in temperature seasonality. Yet, several ecological (e.g., equatorial limit, or unsuitable deforested habitat) or geographical barriers (absence of lands) could prevent species to move equatorward, thus increasing the extinction risk of many tropical species, like endemic baobab species in Madagascar.

Comprehensive marine substrate classification applied to Canada's Pacific shelf

Maps of bottom type are essential to the management of marine resources and biodiversity because of their foundational role in characterizing species’ habitats. They are also urgently needed as countries work to define marine protected areas. Current approaches are time consuming, focus largely on grain size, and tend to overlook shallow waters. Our random forest classification of almost 200,000 observations of bottom type is a timely alternative, providing maps of coastal substrate at a combination of resolution and extents not previously achieved. We correlated the observations with depth, depth-derivatives, and estimates of energy to predict marine substrate at 100 m resolution for Canada’s Pacific shelf, a study area of over 135,000 km². We built five regional models with the same data at 20 m resolution. In addition to standard tests of model fit, we used three independent data sets to test model predictions. We also tested for regional, depth, and resolution effects. We guided our analysis by asking: 1) does weighting for prevalence improve model predictions? 2) does model resolution influence model performance? And 3) is model performance influenced by depth? All our models fit the build data well with true skill statistic (TSS) scores ranging from 0.56 to 0.64. Weighting models with class prevalence improved fit and the correspondence with known spatial features. Class-based metrics showed differences across both resolutions and spatial regions, indicating non-stationarity across these spatial categories. Predictive power was lower (TSS from 0.10 to 0.36) based on independent data evaluation. Model performance was also a function of depth and resolution, illustrating the challenge of accurately representing heterogeneity. Our work shows the value of regional analyses to assessing model stationarity and how independent data evaluation and the use of error metrics can improve understanding of model performance and sampling bias.

Predicting range shifts of Davidia involucrata Ball. under future climate change

Understanding and predicting how species will respond to climate change is crucial for biodiversity conservation. Here, we assessed future climate change impacts on the distribution of a rare and endangered plant species, Davidia involucrate in China, using the most recent global circulation models developed in the sixth Assessment Report of the Intergovernmental Panel on Climate Change (IPCC6). We assessed the potential range shifts in this species by using an ensemble of species distribution models (SDMs). The ensemble SDMs exhibited high predictive ability and suggested that the temperature annual range, annual mean temperature, and precipitation of the driest month are the most influential predictors in shaping distribution patterns of this species. The projections of the ensemble SDMs also suggested that D. involucrate is very vulnerable to future climate change, with at least one-third of its suitable range expected to be lost in all future climate change scenarios and will shift to the northward of high-latitude regions. Similarly, at least one-fifth of the overlap area of the current nature reserve networks and projected suitable habitat is also expected to be lost. These findings suggest that it is of great importance to ensure that adaptive conservation management strategies are in place to mitigate the impacts of climate change on D. involucrate.

Cross-scale monitoring of habitat suitability changes using satellite time series and ecological niche models

One of the biggest challenges to deal with the global crisis of biodiversity loss is the lack of efficient and viable monitoring systems across scales. Unlike traditional in situ biodiversity monitoring, a usually costly and time-consuming enterprise, satellite remote sensing (SRS) data offer a technically feasible and sustainable in time solution. Here, we devise a cost-effective and upgradeable spatiotemporal framework for monitoring the species-specific habitat availability changes across scales by trend analysis of habitat suitability index (HSI) derived from ecological niche models (ENMs; Maxent) and using time series of SRS data (MODIS). The SRS-ENM framework was applied for a large suite of native species (911), from major taxonomic groups (flora (vascular plants), amphibians, reptiles, birds and mammals), and listed in the IUCN Red List at regional (Iberian Peninsula) and continental (Europe) scales. The HSI-trend analyses predict cumulative reductions in habitat suitability for Threatened and Non-Threatened species across scales for the period 2002-2016. Specifically, 19% and 66% of the total grid cells for both species' groups showed negative trends at both regional and continental scales, respectively. Results were similar when considering all IUCN threat categories. All taxa groups showed a decrease in habitat suitability, but amphibians and reptiles groups hosted the largest number of negative HSI-trends grid cells. Considering all groups together, 12% and 34% of both study areas have strong reductions in habitat quality. We conclude that our framework detects increases and decreases in species' habitat suitability regardless of the spatial scale, extent, and pixel size. Species' range predictions across space and time based on SRS time series represent a promising Earth observation tool to support traditional risk assessment protocols and anticipate the decision-making process, while serving as a cross-scale biodiversity monitoring system.

Predicting 1-Year Mortality after Hip Fracture Surgery: An Evaluation of Multiple Machine Learning Approaches

Postoperative death within 1 year following hip fracture surgery is reported to be up to 27%. In the current study, we benchmarked the predictive precision and accuracy of the algorithms support vector machine (SVM), naïve Bayes classifier (NB), and random forest classifier (RF) against logistic regression (LR) in predicting 1-year postoperative mortality in hip fracture patients as well as assessed the relative importance of the variables included in the LR model. All adult patients who underwent primary emergency hip fracture surgery in Sweden, between 1 January 2008 and 31 December 2017 were included in the study. Patients with pathological fractures and non-operatively managed hip fractures, as well as those who died within 30 days after surgery, were excluded from the analysis. A LR model with an elastic net regularization were fitted and compared to NB, SVM, and RF. The relative importance of the variables in the LR model was then evaluated using the permutation importance. The LR model including all the variables demonstrated an acceptable predictive ability on both the training and test datasets for predicting one-year postoperative mortality (Area under the curve (AUC) = 0.74 and 0.74 respectively). NB, SVM, and RF tended to over-predict the mortality, particularly NB and SVM algorithms. In contrast, LR only over-predicted mortality when the predicted probability of mortality was larger than 0.7. The LR algorithm outperformed the other three algorithms in predicting 1-year postoperative mortality in hip fracture patients. The most important predictors of 1-year mortality were the presence of a metastatic carcinoma, American Society of Anesthesiologists(ASA) classification, sex, Charlson Comorbidity Index (CCI) ≤ 4, age, dementia, congestive heart failure, hypertension, surgery using pins/screws, and chronic kidney disease.

Predictive Values of Preoperative Characteristics for 30-Day Mortality in Traumatic Hip Fracture Patients

Hip fracture patients have a high risk of mortality after surgery, with 30-day postoperative rates as high as 10%. This study aimed to explore the predictive ability of preoperative characteristics in traumatic hip fracture patients as they relate to 30-day postoperative mortality using readily available variables in clinical practice. All adult patients who underwent primary emergency hip fracture surgery in Sweden between 2008 and 2017 were included in the analysis. Associations between the possible predictors and 30-day mortality was performed using a multivariate logistic regression (LR) model; the bidirectional stepwise method was used for variable selection. An LR model and convolutional neural network (CNN) were then fitted for prediction. The relative importance of individual predictors was evaluated using the permutation importance and Gini importance. A total of 134,915 traumatic hip fracture patients were included in the study. The CNN and LR models displayed an acceptable predictive ability for predicting 30-day postoperative mortality using a test dataset, displaying an area under the ROC curve (AUC) of as high as 0.76. The variables with the highest importance in prediction were age, sex, hypertension, dementia, American Society of Anesthesiologists (ASA) classification, and the Revised Cardiac Risk Index (RCRI). Both the CNN and LR models achieved an acceptable performance in identifying patients at risk of mortality 30 days after hip fracture surgery. The most important variables for prediction, based on the variables used in the current study are age, hypertension, dementia, sex, ASA classification, and RCRI.

Global Invasion Risk Assessment of Prosopis juliflora at Biome Level: Does Soil Matter?

Prosopis juliflora is one of the most problematic invasive trees in tropical and subtropical regions. Understanding driving forces affecting the potential global distribution would help in managing its current and future spread. The role of climate on the global spatial distribution of P. juliflora has been well studied, but little is known about the role of soil and human impacts as potential drivers. Here, we used maximum entropy (MaxEnt) for species distribution modelling to understand the role of climate (C), soil (S) and human impacts (H), C+S, and C+S+H in controlling the potential invasion range of P. juliflora, and to project its global potential invasive risk. We defined the top threatened global biomes, as predicted by the best-selected model. The incorporation of the edaphic factors improved the model performance and enhanced the accuracy of the outcome. Our findings revealed that the potential invasion risk increases with increases in mean temperature of the driest quarter (Bio9), soil alkalinity and clay fractions. Arid and semi-arid lands are at the highest risk of invasion than other moist biomes.

Assessing biophysical and socio-economic impacts of climate change on regional avian biodiversity

Climate change threatens biodiversity directly by influencing biophysical variables that drive species' geographic distributions and indirectly through socio-economic changes that influence land use patterns, driven by global consumption, production and climate. To date, no detailed analyses have been produced that assess the relative importance of, or interaction between, these direct and indirect climate change impacts on biodiversity at large scales. Here, we apply a new integrated modelling framework to quantify the relative influence of biophysical and socio-economically mediated impacts on avian species in Vietnam and Australia and we find that socio-economically mediated impacts on suitable ranges are largely outweighed by biophysical impacts. However, by translating economic futures and shocks into spatially explicit predictions of biodiversity change, we now have the power to analyse in a consistent way outcomes for nature and people of any change to policy, regulation, trading conditions or consumption trend at any scale from sub-national to global.

Greater topoclimatic control of above- versus below-ground communities

Assessing the degree to which climate explains the spatial distributions of different taxonomic and functional groups is essential for anticipating the effects of climate change on ecosystems. Most effort so far has focused on above-ground organisms, which offer only a partial view on the response of biodiversity to environmental gradients. Here including both above- and below-ground organisms, we quantified the degree of topoclimatic control on the occurrence patterns of >1,500 taxa and phylotypes along a c. 3,000 m elevation gradient, by fitting species distribution models. Higher model performances for animals and plants than for soil microbes (fungi, bacteria and protists) suggest that the direct influence of topoclimate is stronger on above-ground species than on below-ground microorganisms. Accordingly, direct climate change effects are predicted to be stronger for above-ground than for below-ground taxa, whereas factors expressing local soil microclimate and geochemistry are likely more important to explain and forecast the occurrence patterns of soil microbiota. Detailed mapping and future scenarios of soil microclimate and microhabitats, together with comparative studies of interacting and ecologically dependent above- and below-ground biota, are thus needed to understand and realistically forecast the future distribution of ecosystems.

Multi-objective optimisation of species distribution models for river management

Environmental and measure implementation costs are two key factors to be considered by river managers in decision making. To balance effects and costs of an action, practitioners can rely on diagnostic analysis of presence/absence freshwater species distribution models (SDMs) trained to over- or underestimating species presence. Prevalence-adjusted model training aims to balance under- and overestimation depending on study objectives and training data characteristics. The objective of minimising under- and overestimation is a typical example of multi-objective optimisation (MOO). The aim of this paper is to address, for the first time, the practice of MOO-based prevalence-adjusted SDM training for freshwater decision management. In a numerical experiment, the use of Pareto-based MOO, specifically the non-dominated sorting genetic algorithm II (NSGA-II), is compared to commonly-used single-objective optimisation. SDMs for 11 pollution-sensitive freshwater macroinvertebrate species are trained with a subset of the Limnodata, a large data set holding records in the Netherlands over 30 years at 20,000 locations. An increase of two to four times is observed for the ability to identify a large range distribution of the solutions in the Pareto space, when using NSGA-II counter to repeated single-objective optimisation, this by increasing the average runtime with only four percent for a single run. In addition, the use of NSGA-II is found to be effective to identify reliable SDMs useful for diagnostic analysis. By applying and comparing a broad range of MOO methodologies for prevalence-adjusted model training, we believe a closer collaboration between model developers and freshwater managers can be facilitated and environmental standard limits can be set on a more objective basis. In conclusion, the use of MOO for prevalence-adjusted model training is assessed as a valuable tool to support river - and potentially all environmental - decision making.

A checklist for maximizing reproducibility of ecological niche models

Reporting specific modelling methods and metadata is essential to the reproducibility of ecological studies, yet guidelines rarely exist regarding what information should be noted. Here, we address this issue for ecological niche modelling or species distribution modelling, a rapidly developing toolset in ecology used across many aspects of biodiversity science. Our quantitative review of the recent literature reveals a general lack of sufficient information to fully reproduce the work. Over two-thirds of the examined studies neglected to report the version or access date of the underlying data, and only half reported model parameters. To address this problem, we propose adopting a checklist to guide studies in reporting at least the minimum information necessary for ecological niche modelling reproducibility, offering a straightforward way to balance efficiency and accuracy. We encourage the ecological niche modelling community, as well as journal reviewers and editors, to utilize and further develop this framework to facilitate and improve the reproducibility of future work. The proposed checklist framework is generalizable to other areas of ecology, especially those utilizing biodiversity data, environmental data and statistical modelling, and could also be adopted by a broader array of disciplines.

Two alternative evaluation metrics to replace the true skill statistic in the assessment of species distribution models

Model evaluation metrics play a critical role in the selection of adequate species distribution models for conservation and for any application of species distribution modelling (SDM) in general. The responses of these metrics to modelling conditions, however, are rarely taken into account. This leads to inadequate model selection, downstream analyses and uniformed decisions. To aid modellers in critically assessing modelling conditions when choosing and interpreting model evaluation metrics, we analysed the responses of the True Skill Statistic (TSS) under a variety of presence-background modelling conditions using purely theoretical scenarios. We then compared these responses with those of two evaluation metrics commonly applied in the field of meteorology which have potential for use in SDM: the Odds Ratio Skill Score (ORSS) and the Symmetric Extremal Dependence Index (SEDI). We demonstrate that (1) large cell number totals in the confusion matrix, which is strongly biased towards ‘true’ absences in presence-background SDM and (2) low prevalence both compromise model evaluation with TSS. This is since (1) TSS fails to differentiate useful from random models at extreme prevalence levels if the confusion matrix cell number total exceeds ~30,000 cells and (2) TSS converges to hit rate (sensitivity) when prevalence is lower than ~2.5%. We conclude that SEDI is optimal for most presence-background SDM initiatives. Further, ORSS may provide a better alternative if absence data are available or if equal error weighting is strictly required.

A practical guide for combining data to model species distributions

Understanding and accurately modeling species distributions lies at the heart of many problems in ecology, evolution, and conservation. Multiple sources of data are increasingly available for modeling species distributions, such as data from citizen science programs, atlases, museums, and planned surveys. Yet reliably combining data sources can be challenging because data sources can vary considerably in their design, gradients covered, and potential sampling biases. We review, synthesize, and illustrate recent developments in combining multiple sources of data for species distribution modeling. We identify five ways in which multiple sources of data are typically combined for modeling species distributions. These approaches vary in their ability to accommodate sampling design, bias, and uncertainty when quantifying environmental relationships in species distribution models. Many of the challenges for combining data are solved through the prudent use of integrated species distribution models: models that simultaneously combine different data sources on species locations to quantify environmental relationships for explaining species distribution. We illustrate these approaches using planned survey data on 24 species of birds coupled with opportunistically collected eBird data in the southeastern United States. This example illustrates some of the benefits of data integration, such as increased precision in environmental relationships, greater predictive accuracy, and accounting for sample bias. Yet it also illustrates challenges of combining data sources with vastly different sampling methodologies and amounts of data. We provide one solution to this challenge through the use of weighted joint likelihoods. Weighted joint likelihoods provide a means to emphasize data sources based on different criteria (e.g., sample size), and we find that weighting improves predictions for all species considered. We conclude by providing practical guidance on combining multiple sources of data for modeling species distributions.

A new baseline for countrywide α-diversity and species distributions: illustration using >6,000 plant species in Panama

Estimating α-diversity and species distributions provides baseline information to understand factors such as biodiversity loss and erosion of ecosystem services. Yet, species surveys typically cover a small portion of any country's landmass. Public, global databases could help, but contain biases. Thus, the magnitude of bias should be identified and ameliorated, the value of integration determined, and application to current policy issues illustrated. The ideal integrative approach should be powerful, flexible, efficient, and conceptually straightforward. We estimated distributions for >6,000 species, integrating species sightings (S) from the Global Biodiversity Information Facility (GBIF), systematic survey data (S₂ ), and "bias-adjustment kernels" (BaK) using spatial and species trait databases (S₂ BaK). We validated our approach using both locational and species holdout sets, and then applied our predictive model to Panama. Using sightings alone (the most common approach) discriminated relative probabilities of occurrences well (area under the curve [AUC] = 0.88), but underestimated actual probabilities by ~4,000%, while using survey data alone omitted over three-quarters of the >6,000 species. Comparatively, S₂ BaK had no systematic underestimation, and substantially stronger discrimination (AUC = 0.96) and predictive power (deviance explained = 47%). Our model suggested high diversity (~200% countrywide mean) where urban development is projected to occur (the Panama Canal watershed) and also suggested this is not due to higher sampling intensity. However, portions of the Caribbean coast and eastern Panama (the Darién Gap) were even higher, both for total plant biodiversity (~250% countrywide mean), and CITES listed species. Finally, indigenous territories appeared half as diverse as other regions, based on survey observations. However, our model suggested this was largely due to site selection, and that richness in and out of indigenous territories was roughly equal. In brief, we provide arguably the best estimate of countrywide plant α-diversity and species distributions in the Neotropics, and make >6,000 species distributions available. We identify regions of overlap between development and high biodiversity, and improve interpretation of biodiversity patterns, including for policy-relevant CITES species, and locations with limited access (i.e., indigenous territories). We derive a powerful, flexible, efficient and simple estimation approach for biodiversity science.

Where will threatened migratory birds go under climate change? Implications for China's national nature reserves

Climate change, regarded as one of the major threats to biodiversity and ecosystems, can impact on the distribution and survival of migratory birds. To investigate the threats of climate change to threatened migratory bird distributions, we used species distribution model (SDM) and climatic data under current and future climate scenarios to predict future changes in species distributions and how the geographic distribution of these threatened birds may respond to climate change by 2050. Our results show the hotspots for all species may remain in the lower and middle reaches of the Yangtze River, while more species may dwell in the coastal regions of the Bohai Gulf and the Yellow Sea in the future. Our findings show that the percentage of all species distributions or hotspots for all threatened species covered by national nature reserves (NNRs) in China remain low by 2050. Thus, we propose that China should increase and expand reserves in eastern China. Significantly, we emphasize the creation of protected areas to make it the Ramsar sites in the world and recommend that China should (1) strengthen the cooperation with neighboring countries to share maximum species occurrence data (especially the threatened species), (2) overlay maps of individual species for each taxon to assess the efficiency of coastal nature reserves and predict the hotspots shift under climate change, (3) trade off urban development and ecosystem stability to create new and dynamic protected areas to make it the Ramsar sites, (4) appeal for long-term protection of ecosystem stability to achieve sustainable development in the world.

Forecasting the global extent of invasion of the cereal pest Spodoptera frugiperda, the fall armyworm

Fall armyworm, Spodopterafrugiperda, is a crop pest native to the Americas, which has invaded and spread throughout sub-Saharan Africa within two years. Recent estimates of 20–50% maize yield loss in Africa suggest severe impact on livelihoods. Fall armyworm is still infilling its potential range in Africa and could spread to other continents. In order to understand fall armyworm’s year-round, global, potential distribution, we used evidence of the effects of temperature and precipitation on fall armyworm life-history, combined with data on native and African distributions to construct Species Distribution Models (SDMs). We also investigated the strength of trade and transportation pathways that could carry fall armyworm beyond Africa. Up till now, fall armyworm has only invaded areas that have a climate similar to the native distribution, validating the use of climatic SDMs. The strongest climatic limits on fall armyworm’s year-round distribution are the coldest annual temperature and the amount of rain in the wet season. Much of sub-Saharan Africa can host year-round fall armyworm populations, but the likelihoods of colonising North Africa and seasonal migrations into Europe are hard to predict. South and Southeast Asia and Australia have climate conditions that would permit fall armyworm to invade. Current trade and transportation routes reveal Australia, China, India, Indonesia, Malaysia, Philippines and Thailand face high threat of fall armyworm invasions originating from Africa.

Ecosystem Services in a Protected Mountain Range of Portugal: Satellite-Based Products for State and Trend Analysis

Mountains are facing strong environmental pressures, which may jeopardize the supply of various ecosystem services. For sustainable land management, ecosystem services and their supporting functions should thus be evaluated and monitored. Satellite products have been receiving growing attention for monitoring ecosystem functioning, mainly due to their increasing temporal and spatial resolutions. Here, we aim to illustrate the high potential of satellite products, combined with ancillary in situ and statistical data, to monitor the current state and trend of ecosystem services in the Peneda-Gerês National Park, a protected mountain range in Portugal located in a transition climatic zone (Atlantic to Mediterranean). We focused on three ecosystem services belonging to three broad categories: provisioning (reared animals), regulating (of water flows), and cultural (conservation of an endemic and iconic species). These services were evaluated using a set of different satellite products, namely grassland cover, soil moisture, and ecosystem functional attributes. In situ and statistical data were also used to compute final indicators of ecosystem services. We found a decline in the provision of reared animals since year 2000, although the area of grasslands had remained stable. The regulation of water flows had been maintained, and a strong relationship with interannual precipitation pattern was noted. In the same period, conservation of the focal iconic species might have been affected by interannual fluctuations of suitable habitat areas, with a possible influence of wildfires and precipitation. We conclude that satellite products can efficiently provide information about the current state and trend in the supply of various categories of ecosystem services, especially when combined with in situ or statistical data in robust modeling frameworks.